Mechanism and kinetics of the diffusion of gold in silicon

The diffusion of Au in Si is known to take place via the interchange of Au atoms between substitutional (Au _s ) and interstitial (Au _i ) sites. So far it has generally been believed that this interchange involves lattice vacancies (V) and that it occurs via the Frank-Turnbull mechanism V+Au _i ⇆Au _s . It is stated in the literature that this model explains the observation that the Au _s concentrationC _s ^m in the centre of Au-diffused Si wafers increases with timet according to . We show that this statement is incorrect, i.e., the Frank-Turnbull model cannot account for the law. Such a dependence is expected in the case of Si wafers with a sufficiently low density of internal sinks for self-interstitials if the Au _i −Au _s interchange is controlled by the so-called kick-out mechanism Au _i ⇆Au _s +1. Since this mechanism involves self-interstitials (I) the present result is in accordance with the fact that under high-temperature equilibrium conditions the dominating intrinsic point defects in Si are self-interstitials and not vacancies as in Ge or metals.     

On the theory of the diffusion of gold into dislocated silicon wafers

The paper presents a theoretical study of the diffusion of gold into dislocated silicon wafers in terms of the kick-out mechanism Au _i Au _s +I, where Au _i , Au _s , andI mean Au interstitials, substitutional Au atoms, and Si self-interstitials, respectively. In agreement with experiments it is found that the Au _s concentration in the centre of a wafer,C _s ^m , increases with the durationt of the diffusion anneal according toC _s ^m =C _s ^eq (k ₀ t)^1/2 except forC _s ^m values in the vicinity of the solubility limitC _s ^eq of Au _s . Approximate analytical expressions fork ₀ as a function of the densityN _I of the dislocations acting asI sinks are given for the entire regime 0N_I<+t8.     

Hydrogen implantation and diffusion in silicon and silicon dioxide

Depth profiles of hydrogen implanted into crystalline silicon in random direction at different fluences have been measured by the¹⁵N technique and by SIMS. Whereas hydrogen implanted at a fluence of 10¹⁵ ions/cm² shows some limited mobility, no such mobility is observed for higher implantation fluences. In these cases, ballistic computer codes describe the depth distributions well, within the ranges of both experimental and theoretical accuracy. Annealing up to 510 K does not change the hydrogen distributions.Furthermore, high-fluence hydrogen implantation into silicon dioxide has been examined. There is some indication for radiation-enhanced diffusion during the implantation process. Upon subsequent thermal annealing, the hydrogen is found to diffuse, probably via a trapping/detrapping mechanism associated with an OH/H₂ transformation of the hydrogen bonding.     

Passivation of laser induced defects in silicon by low energy hydrogen ion implantation

It is well established that the pulsed-laser annealing of as-implanted Si introduces electrically active defects. The aim of this paper is to show that these defects can be neutralized by low-energy hydrogen ion implantation. The techniques used for the sample characterization are current-voltage measurements and capacitance transient spectroscopy.     

Oxygen diffusion and thermal donor formation in silicon

The information available on the diffusion of oxygen and on the formation of thermal donors in silicon is critically reviewed. In this context the effects of intrinsic point defects on the diffusion-controlled growth of oxygen precipitates is investigated in some detail. Seemingly contradictory experimental results on the diffusivity of oxygen in silicon at temperatures around 400° C are explained in terms offast-diffusing gas-like molecular oxygen in silicon. The concept of molecular oxygen is also invoked in a newly suggested model of thermal donor formation in silicon. The diffusivity of molecular oxygen in silicon is estimated to be around 10^–9cm²s^–1 at 450° C, almost nine orders of magnitude higher than the diffusivity of atomic oxygen in interstitial position.     

Diffusion of gold in dislocation-free or highly dislocated silicon measured by the spreading-resistance technique

The diffusion of Au in dislocation-free or plastically deformed Si (10¹¹ to 10¹³ dislocations/m²) was measured with the aid of the spreading-resistance technique. The Au profiles produced indislocation-free Si slices by in-diffusion from both surfaces possess nonerfc-type U shapes as predicted by the so-called kick-out diffusion model. This model is used to calculate the contribution of self-interstitials to the (uncorrelated) Si self-diffusion coefficient,D _I ^SD =0.064×exp(–4.80 eV/kT)m² s^–1, from the present and previous data on the diffusivity and solubility of Au in Si in the temperature range 1073–1473 K. Inhighly dislocated Si the diffusion of Au is considerably faster than in dislocation-free Si. From the erfc-type penetration profiles found in this case, effective Au diffusion coefficients were deduced and combined with data on the solubility of Au in Si. ThusC _i ^eq D _i=0.0064 ×exp(–3.93 eV/kT)m² s^–1 was obtained in the temperature range 1180–1427 K, whereC _i ^eq andD _i are the solubility and diffusivity of interstitial Au in Si.     

Point defects,diffusion processes,and swirl defect formation in silicon

The paper consists of three parts. In the first part we review the basic experimental and theoretical results which shaped our present knowledge on point defects and diffusion processes in silicon. These results concern on one side oxidation effects which established that silicon self-interstitials and vacancies coexist in silicon and on the other side diffusion of gold into dislocation-free silicon which allowed to determine the self-interstitial contribution to silicon self-diffusion and to estimate the corresponding vacancy contribution. In the second part we discuss topics for which an understanding is just emerging within the framework of coexisting self-interstitials and vacancies: reaching of local dynamical equilibrium between self-interstitials and vacancies; rough estimates of the thermal equilibrium concentrations of self-interstitials and vacancies and their respective diffusivities, and finally, various possibilities to generate an undersaturation of self-interstitials. In the third part we examine swirl defect formation in silicon in terms of vacancies and self-interstitials.     

On the nature of point defects and the effect of oxidation on substitutional dopant diffusion in silicon

An extensive analysis of the substitutional dopant diffusion phenomena in silicon during oxidation is presented. The analysis covers qualitative as well as quantitative aspects of the oxidation-enhanced and -retarded diffusion (OED and ORD) phenomena, and examines three different possible assumptions that can be made on the nature of the silicon thermal equilibrium point defect species: silicon self-interstitials (I) only, vacancies (V) only, coexistence of I and V. The only consistent way to interpret all properly documented OED/ORD data is to assume that I and V coexist under oxidation as well as under thermal equilibrium conditions at high temperatures.     

Residual non-idealities in the almost ideal silicon p-n junction

The reproducible, mass production of almost ideal silicon p-n junctions has allowed two new phenomena to be discovered: a pure generation without recombination, and a slow capacitance-free current transient. Our present knowledge of these phenomena is reviewed and speculations about the centres responsible for them are discussed; these centres seem to be connected to ultimate, unavoidable properties of the silicon p-n junction rather than to unwanted impurities.     

Boron doping of silicon by excimer laser irradiation in a reactive atmosphere

UV excimer lasers have been used to dope semiconductors by a one-step process in which the laser serves both to melt a controlled thickness of a sample placed in dopant ambient and to photodissociate the dopant molecules themselves. Here we report the boron doping of silicon by means of an ArF (193 nm) excimer laser. Dopant atoms are obtained by photolysis of BCl₃ or pyrolysis of BF₃ molecules. The doping is performed both in gas ambient and using only an adsorbed layer. We have investigated the dependence of doping parameters such as laser pulse repetition and gas pressure on the subsequent boron impurity profiles and the dopant incorporation rate. These results indicate that the laser doping process is dopant-flux limited for BF₃ and externally rate limited for BCl₃.     

Nd: YAG laser annealing of gallium-implanted silicon

A Q-switched Nd: YAG laser with a pulse duration of 20 ns was used to investigate effects of laser annealing in gallium implanted silicon. Rutherford backscattering and Hall-effect measurements were performed to evaluate the annealed layer. Differential Hall-effect measurements were carried out to obtain carrier concentration profiles after annealing. It was found that a maximum sheet carrier concentration of 8×10¹⁵ cm⁻² can be obtained for a gallium implantation of 10¹⁶ cm⁻² by laser annealing with an energy density of more than 1.0 J cm⁻². Although the peak carrier concentration was found to be 8.0×10²⁰ cm⁻³, the annealed layer showed polycrystalline structures even after annealing with an energy density up to 4J cm⁻². The annealing took place in the solid phase in this energy density range.     

Focused ion beam gallium implantation into silicon

Samples formed of a thin metal film deposited on silicon single crystal were annealed with electron and laser (ruby and excimer) pulses over a wide range of fluences. From a comparison of the experimental results with the temperature profiles of the irradiated samples, it turns out that suicide formation starts when the metal/silicon interface reaches the lowest eutectic temperature of the binary metal/silicon system. The growth rate of reacted layers is of the order of 1 m/s.     

Disorders produced during high-current and high-dose phosphorus ion implantation in silicon

Transmission electron microscopy, optical reflection and channeling effect measurements are employed to investigate disorders in 30 keV, high dose (3×10¹⁶ions/cm²) and high current (≦5 mA) phosphorus as-implanted silicon with (111), (100), and (110) orientation as a function of temperature rise (100–850°C) by the beam heating effect during implantation. Temperature rise below 400°C results in continuous amorrphous layer formation. This contrasts with results of the recovery into single crystals for temperature rise samples above 500°C, regardless of wafer orientation. Secondary defects (black-dotted defects, dislocation loops and rodlike defects) are formed in singlecrystal recovery samples, having a deeper distribution in (110) wafers and a shallower distribution in (111) and (100) wafers. Rodlike defects observed in 850°C samples are of “vacancy” type and have the largest density in (110) wafers.     

Front and back surface cw CO2-laser annealing of arsenic ion-implanted silicon

The annealing behavior of arsenic-implanted silicon under scanned cw CO₂-laser irradiation from front and back surfaces is investigated. Ellipsometry, Hall effect, Rutherford backscattering measurements and neutron activation analysis indicate an enhancement of annealing efficiency by laser irradiation from the back surface, which provides complete recovery of crystal damage, high substitutionality and electrical activation of implanted arsenic atoms without redistribution of concentration profile. The enhancement of annealing efficiency under back-surface irradiation is explained by the difference in laser reflection from the front and back surface of silicon wafers. No differences in the results are found for scanned and static annealing.     

CW CO2-laser annealing of arsenic implanted silicon

CW CO₂-laser annealing of arsenic implanted silicon was investigated in comparison with thermal annealing. Ion channeling, ellipsometry, and Hall effect measurements were performed to characterize the annealed layers and a correlation among the different methods was made. The laser annealing was done with power densities of 100 to 640 W cm⁻² for 1 to 20 s. It was found that the lattice disorder produced during implantation can be completely annealed out by laser annealing with a power density of 500 W cm⁻² and the arsenic atoms are brought on lattice sites up to 96±2%. The maximum sheet carrier concentration of 6×10¹⁵ cm⁻² was obtained for 1×10¹⁶ cm⁻² implantation after laser annealing, which was up to 33% higher than that after thermal annealing at 600 to 900°C for 30 min.     

Dependence of anomalous phosphorus diffusion in silicon on depth position of defects created by ion implantation

Transient enhanced diffusion of phosphorus in silicon has been investigated for implants below and above the threshold for a complete amorphization. Rapid thermal processes (electron beam) and conventional furnaces have been used for the annealing. In the case of implants below amorphization, a strong enhanced diffusion, proportional to the amount of damage produced, has been observed. The extent of the phenomenon is practically independent of the damage depth position. In contrast to this, the formation of extended defects at the original amorphous-crystalline interface makes the diffusivity strongly dependent on depth in the case of post-amorphized samples. No enhanced diffusion effect is observed if the dopant is confined in the amorphous layer, while a remarkable increase in the diffusivity is detected for the dopant located in the crystalline region beyond the amorphous-crystalline interface.Damage distribution after implantation and its evolution during annealing have been determined by double crystal x-ray diffraction and correlated to anomalous P diffusivity. A qualitative distribution of the interstitial excess in solution in the silicon lattice during annealing is proposed for the two different cases. These point defects, released by the dissolution of the interstitial clusters produced by the implanted ions, have been identified as responsible for the observed enhanced P diffusion.     

Boron neutralization and hydrogen diffusion in silicon subjected to low-energy hydrogen implantation

Using the hydrogen neutralization of the boron acceptor, the diffusion of hydrogen is investigated in the temperature range 20 °–160 °C. The hydrogenation is performed by low-energy implantation. We observe a fast initial hydrogen migration, followed by a long-time diffusion phase that is described by an effective diffusion coefficientD _eff=D _{0 eff} exp(–E _a/kT) withD _{0 eff}–cm²s^–1 andE _a=(0.83±0.05) eV. No deeper hydrogen migration is detected for implantation times longer than – 30 min. Our data are explained by the build-up of a large amount of molecular hydrogen beneath the surface, which strongly hinders the transfer of the implanted hydrogen to the bulk. The thermal reactivation kinetics of the neutralized boron show a rapid initial step followed by a longtime thermally activated second order phase, which is limited by the recombination of hydrogen into molecules.     

Simulation of range profiles for boron implantation into SiO2/Si and Si3N4/SiO2/Si targets

The simple method of profile combination is shown to be applicable to the simulation of boron profiles in SiO₂/Si and Si₃N₄/Si layered targets. This is demonstrated by comparison with range distributions calculated by more sophisticated theoretical methods, i.e. TRIM Monte Carlo simulations and the algorithm of Christel et al., and with experimental data. The method of profile combination can also be applied to layered targets with a crystalline silicon substrate.     

X-ray Photoelectron Spectroscopy of silicon oxynitride layers obtained by low-energy ion implantation

Simultaneous oxygen and nitrogen low-energy ion implantation in silicon single crystals have been used to obtain surface layers of silicon oxynitride. In-depth concentration profiles, measured by XPS, showed all the possible tetrahedral configurations of silicon with the neighboring atoms. The most important feature of these profiles has been found to be the accumulation of a so-called Si³⁺ state just below the surface and about 20 nm deep. This amphoteric state has interesting properties for impurity passivation or electron trapping.     

Crystal forms of hillocks and voids formed by electromigration on ultrapure gold and silver wires

Hillocks and voids (vacancy crystals) which were formed by electromigration on ultrapure gold and silver wires, showed only the combined crystal forms {111} and {100}. These planes are identical with the equilibrium planes G_I, calculated by Stranski and Kaishew for nonpolar substances under the assumption of only nearest neighbor interaction. Hillock and void nucleation occurred on crystal imperfections (heterogeneous nucleation). Hillocks and vacancy crystals (voids) on gold showed twin formation obeying the spinel law. Dedicated to Professor I. N. Stranski on the occasion of his 80th birthday.