F. C. Unterwald

Electrical transport properties of CoSi2 and NiSi2 thin films

Transport studies have been performed on thin films of CoSi2 and NiSi2 in the temperature range 1–300 K. The conductivities are metallic with essentially the same temperature dependence; however, the residual resistivities are markedly different even though the two silicides are structurally similar (the room‐temperature resistivity of NiSi2 being at least twice that of CoSi2 of 15 μΩ cm). The difference is attributed to intrinsic defects in NiSi2. This defect has been simulated by ion bombardment of the film where it is also shown that Matthiessen’s rule is obeyed over a remarkable range of bombardment doses.

Oxidation enhanced diffusion in Si B‐doping superlattices and Si self‐interstitial diffusivities

A special thin film structure has been grown by low temperature molecular beam epitaxy for an investigation of the properties of self‐interstitials in Si. It consists of a doping superlattice made from B spikes separated from each other by 100 nm of Si. After dry oxidation, the width of each spike is directly proportional to the interstitial concentration at that depth. The superlattice as a whole thus gives a depth profile of the time‐averaged interstitial concentration, allowing the direct determination of the diffusion coefficient of interstitials. The abrupt dopant concentration transitions possible in low‐temperature molecular‐beam‐epitaxy‐grown films allow this investigation in the temperature range 750–900 °C. At 800 °C we find a value of DI=(1.4±0.4)10−13 cm2/s. Performing the experiments as a function of temperature yields DI = D0eEa/kT with D0=102±2 cm2/s and Ea=(3.1±0.4) eV.

Doping of Si thin films by low-temperature molecular beam epitaxy

Two‐dimensional doping sheets (‘‘δ doping’’) are integral parts of many novel semiconductor device concepts. Deep submicron design rules require junction depths significantly below 100 nm. This level of control is difficult to achieve with ion implantation. We discuss the application of thermal, coevaporative doping with Sb and elemental B during Si molecular beam epitaxy at growth temperatures below ≊300 °C to this problem. We show that it is possible to create structures with very high doping levels, yet with very sharp doping transitions. Delta‐doping spikes with a full width at half maximum of <2.7 nm and <4.0 nm have been obtained by secondary‐ion mass spectrometry for Sb and B, respectively, with corresponding up‐slopes of 2.5 and 0.94 nm/decade. Homogeneously doped films show full activation up to NSb≊6×1020 cm−3 and NB≳1×1021 cm−3. Mobilities agree with bulk values at corresponding concentrations. Mesa‐isolated pn junctions exhibit ideality factors of 1.05.

Design of an ultrahigh-vacuum specimen environment for high-resolution transmission electron microscopy

A JEOL 200CX high‐resolution transmission electron microscope with point‐to‐point resolution of 2.5 A has been modified to achieve an ultrahigh‐vacuum specimen environment (∼10−9τ). In situ heating and evaporation are provided in the specimen viewing position, where high resolution can be maintained at temperatures exceeding 600 °C. Our design employs complete differential pumping of the specimen chamber and the use of a He‐cooled cyroshield at the specimen for best vacuum attainment. Our design philosophy permits the instruments to be used for a wide variety of in situ experiments, including low‐pressure (<10−1τ) gas reaction.

Diffusion of dopants in B‐ and Sb‐delta‐doped Si films grown by solid‐phase epitaxy

The diffusion of δ‐function‐shaped B‐ and Sb‐dopant spikes in thin Si films grown by solid‐phase‐epitaxy [(SPE), growth of amorphous film by molecular‐beam epitaxy (MBE) at room temperature and subsequent regrowth in situ] during annealing in vacuum is compared to diffusion in films grown by low‐temperature (LT) MBE. Diffusion temperatures from 750 to 900 °C, and two‐dimensional concentrations of 0.7–1.6×1014 cm−2 have been investigated. The diffusive behavior of dopants in SPE films is found to be qualitatively different from that in films grown by LTMBE. This is related to the vacancylike defects that are intrinsic to growth by SPE but not to growth by LTMBE. Dopant profiles widen significantly during SPE regrowth, making the achievement of δ‐function dopant spikes impossible. After a vacuum anneal the diffusion coefficients for both n‐ and p‐type dopants are lower in SPE films than the corresponding values in films grown by LTMBE by up to one order of magnitude. The diffused depth profile of the dopant...

Influence of fluorine implant on boron diffusion: Determination of process modeling parameters

The effects of low‐dose ion implants with Si+, Ne+, and F+ on the transient enhanced diffusion of B in silicon after annealing at 900 °C for 30 min have been investigated. Processing conditions such as implant dose (3.5×1013 cm−2) and energy (30–60 keV) were chosen to simulate the lightly doped drain implant in a 0.35 μm complementary metal‐oxide‐semiconductor technology. An epitaxially grown B‐doping superlattice is used to extract directly depth profiles of average Si self‐interstitial concentration after processing. For Si+ the transient enhanced diffusion of B increases with the energy of the implanted ion. Ne+ implanted with the same energy as Si+ causes more transient enhanced diffusion, while Ne+ implanted with the same range as Si+ causes slightly less. Implantation of F+ enhances the B diffusivity considerably less than Si or Ne implantation. These effects were modeled using simulations of defect diffusion in the presence of traps. A trap concentration of (2.4±0.5)×1016 cm−3 gave good agreement i...

Point defects in Si thin films grown by molecular beam epitaxy

Depth profiles of vacancylike defects have been determined by positron annihilation spectroscopy in 200‐nm‐thick Si films grown by molecular beam epitaxy on Si(100) substrates at growth temperatures Tgrowth=200–560 °C. The line shape of the radiation emitted from implanted positrons annihilating in the near‐surface region of a solid gives quantitative, depth‐resolved information on defect concentrations in a nondestructive way. In particular, the method is sensitive to vacancylike defects in a concentration range inaccessible to electron microscopy or ion scattering, but important for electrical device characteristics. The sensitivity limit for these defects in the present experiments is estimated as 5×1015 cm−3. Films grown at Tgrowth≥475±20 °C are indistinguishable from virgin wafers. So are samples with Tgrowth=220±20 °C, subjected to a 2 min, TRTA≳500 °C rapid thermal anneal (RTA) after every ≊30 nm of Si growth. If TRTA=450±20 °C, part of the film contains a concentration of vacancylike defects on th...

Determination of Si self‐interstitial diffusivities from the oxidation‐enhanced diffusion in B doping‐superlattices: The influence of the marker layers

Si self‐interstitial diffusivities can be extracted from the diffusive behavior of certain metals (e.g., Au) in an inert annealing ambient or from the diffusion of dopant markers (typically B) under oxidizing conditions. Each type of experiment yields fairly consistent results; however, interstitial diffusivities obtained in these two ways differ greatly. The marker layer experiments rely on the assumption that the presence of the dopant does not disturb the diffusion of the interstitials, and the validity of this assumption is explored. A model of interstitial diffusivity in the presence of B is developed, two extreme cases of the B‐atom–interstitial interaction strength are considered, and the predictions of the model are compared with experiments of oxidation‐enhanced diffusion in B doping‐superlattices. From this comparison it is concluded that trapping of interstitials by B atoms in the markers cannot be responsible for the different values of the Si interstitial diffusivity reported in the literatur...

BEHAVIOR OF INTRINSIC SI POINT DEFECTS DURING ANNEALING IN VACUUM

Using B and Sb doped Si(100) doping superlattices (DSL) as tracers of native Si point defect behavior it is shown that vacuum annealing at 810 °C leads to a depletion of Si self‐interstitials, with their smallest concentration at the surface, but does not affect the vacancy population. At a fixed depth, the interstitial concentration drops for increasing annealing times; for a given time, the interstitial concentration increases into the sample as a function of depth. Inert anneals of a B‐DSL in Ar show flat interstitial profiles. Apparently, the vacuum anneal makes the surface a better sink for interstitials than an inert Ar anneal, leading to an equilibrium interstitial concentration below the value in the bulk and establishing a net outflow of interstitials to the surface. The absence of a response of the vacancy population yields a lower limit on the interstitial‐vacancy recombination time of 104 s at 810 °C. Process simulation of this scenario captures the essential trends of the experimental data.

Time dependence of dopant diffusion in δ‐doped Si films and properties of Si point defects

The diffusion of Sb and B in thin Si films grown by low temperature molecular beam epitaxy is investigated in the temperature range 750–900 °C for times of 0.25–60 h. The small spatial extent of the initial δ‐function‐like dopant profiles allows the detection of very small diffusional displacements. The dopant atoms are used as tracers of Si point defects (vacancies and self‐interstitials). Diffusion of Sb is found to be enhanced relative to equilibrium values, while that of B is retarded. We propose a model based on an initial supersaturation of vacancies. Matching this model to the experimental data allows the extraction of the vacancy diffusivity, the activation energy of vacancy formation, and the recombination lifetime of interstitials. The results show that interstitial and vacancy populations cannot be considered independent at low temperature, as has been previously suggested.