Boron-implanted Silicon Research Articles

Near-Field Imaging of the Photocurrent Induced by Light on Boron-Implanted Silicon

We report near-field photocurrent (PC) measurements performed on boron-implanted silicon and acquired at λ = 633, 1330, 1550 nm corresponding to photon energies of 1.96, 0.93, and 0.80 eV, respectively. The second and the third energies are below the silicon energy gap (Egap = 1.12 eV), representing incident radiation to which Si is virtually transparent. The PC images acquired at λ = 1330 and 1550 nm reveal the presence of boron clusters which are a consequence of B-implantation and rapid thermal annealing at 1100 °C. Boron clusters behave as metal clusters embedded into the silicon matrix and introduce gap states which give rise to the observed photocurrent. At energies larger than the Si energy gap the near-field PC images give information about subsurface defects not related to boron and with no correspondence in the surface topography.

Separation of vacancy and interstitial depth profiles in proton- and boron-implanted silicon

A new experimental method of studying shifts between concentration-versus-depth profiles of vacancy-type and interstitial-type defects in ion-implanted silicon is demonstrated. The concept is based on deep level transient spectroscopy (DLTS) measurements utilizing the filling pulse variation technique. The vacancy profile, represented by the vacancy-oxygen center and the interstitial profile, represented by the substitutional carbon-interstitial carbon pair, are obtained at the same sample temperature by varying the duration of the filling pulse. Thus the two profiles can be recorded with a high relative depth resolution. Point defects have been introduced in low doped float zone n-type silicon by implantation with 6 MeV boron ions and 1.3 MeV protons at room temperature, using low doses. For each implantation condition the peak of the interstitial profile is shown to be displaced by ∼ 0.5 μm towards larger depths compared to that of the vacancy profile. This shift is primarily attributed to the preferential forward momentum of recoiling Si atoms, in accordance with theoretical predictions.

Ultrashallow profiling of semiconductors by secondary ion mass spectrometry:: methods and applications

We have investigated in detail the capability of a CAMECA IMS-4f magnetic-sector SIMS instrument in terms of ultimate depth resolution, accuracy, quantification in the uppermost layers, sensitivity and dynamic range, as well as fastness and ease of measurement. A method to obtain ultrahigh depth resolutions will be demonstrated and discussed. The study has primarily been finalized to the investigation of ultrashallow boron implants in silicon. Specific applications to the above-mentioned material will be presented. In particular, it will be shown that, taking advantage of the developed measurement protocol, it was possible to perform a detailed investigation on the atomic transport properties of the sub-keV energy implanted boron in silicon.

In search of optimum conditions for the growth of sharp and shallow B-delta markers in Si by molecular beam epitaxy

Multiple boron delta spikes in silicon, with spacings between 4.3 and 20 nm, have been grown by molecular beam epitaxy at temperatures of about 100 °C (L1) and 400 °C (S4). The test samples were depth profiled by secondary ion mass spectrometry using 500 eV O2+ at normal beam incidence. The surface of S4 was quite smooth, with a root mean square roughness σ<0.1 nm. By contrast, L1 was rather rough, σ≅0.5 nm. The boron depth profiles of S4 revealed sharp peaks but pronounced tails on either side. The tails, which dominate the dopant distributions at concentrations below about 40% of the peak level, are attributed to defect-promoted boron diffusion during growth. Sample L1 showed boron spikes of larger width above the 10%–20% peak level, but a much more rapid, roughly exponential falloff on both sides. This sharpness of the dopant spikes implies the absence of boron diffusion during low-temperature growth. The “best” deltas (those with small width and sharp falloff) were obtained with boron contaminants of ambient origin that resided at the (oxidized) interface between the substrate and the silicon buffer layer. This observation suggests that boron atoms in silicon dioxide are rather immobile. Depth profile measurements on crystalline samples, either containing boron deltas or being uniformly doped with boron, revealed severe variations of the B+ signal over a depth of up to 25 nm at normal and oblique beam incidences (up to 50°, also with oxygen flooding). Silicon matrix signals measured in parallel did not show any variation beyond the transient depth. The initial overshoot of the B+ signal, observed just below the transient depth, sometimes exceeded the stationary signal by more than a factor of 2, and the signal undershoot in extended regions at larger depths was low by up to several 10%. This artifact calls for recalibration of previously reported profiles of shallow boron implantations in silicon.

Depth profiling of ultra-shallow implants using a Cameca IMS-6f

We have previously reported measurements of ultra-shallow boron implants in silicon using a magnetic sector Cameca IMS-6f, and excellent depth resolution and dynamic range were obtained for boron implants down to 1 keV using O2+ primary beam. The appropriate impact energy of the primary beam and incidence angle could be easily achieved for an analysis by the choice of several combinations of sample and source voltages, based on calculations governing the penetration depth and incidence angle of the primary ion beam. This article outlines additional low energy analysis results for boron and arsenic implants with either oxygen or cesium low energy beams. Methods for low energy beam alignment have been utilized to allow analysis conditions to be reached quickly while maintaining a good beam shape, and calculations that more accurately describe how the angle of incidence varies with the primary, sample, and deflector voltages are described.

Cluster formation during annealing of ultra-low-energy boron-implanted silicon

The clustering of low-energy ion-implanted boron has been investigated. Two 1 keV boron implantations at doses of 1×1015 and 5×1015 cm−2 were annealed for 10 s between 700 and 1100 °C. The evolution of the boron concentration profiles was monitored using secondary ion mass spectrometry. Electrical activation was measured with four-point-probe measurements and spreading resistance profiling. The displaced Si concentration profiles were determined from medium-energy ion-scattering measurements.

Diffuse X-Ray Streaks from Defects and Surface Features in Boron Implanted Silicon

In a grazing incidence X-ray diffuse scattering investigation of boron implanted silicon we have discovered narrow intensity rods along [111] directions. From a detailed analysis of three dimensional reciprocal space maps clear evidence is found that the rod-like scattering is due to extrinsic stacking faults with an average diameter of 71 nm, formed in the implanted layer after rapid thermal annealing. At the same time nanometer sized structures appear at the Si surface which are characterized by atomic force microscopy and specular reflectivity measurements.

Diffuse X-Ray Streaks from Defects and Surface Features in Boron Implanted Silicon

In a grazing incidence X-ray diffuse scattering investigation of boron implanted silicon we have discovered narrow intensity rods along [111] directions. From a detailed analysis of three dimensional reciprocal space maps clear evidence is found that the rod-like scattering is due to extrinsic stacking faults with an average diameter of 71 nm, formed in the implanted layer after rapid thermal annealing. At the same time nanometer sized structures appear at the Si surface which are characterized by atomic force microscopy and specular reflectivity measurements.

Electrical activation kinetics for shallow boron implants in silicon

Silicon implanted with boron at 1015 cm−2 dose and energies from 500 eV to 1 keV were annealed over wide variations in temperature and time to obtain process kinetics and thermal activation energies for shallow junction formation. Diffusion depths and carrier densities were determined by modeling sheet electrical transport. The thermal activation energy for the mean time to produce electrical activation is found to be 5.1±0.1 eV, while for the mean diffusivity it is found to be 4.1±0.1 eV. The 1 eV difference expresses quantitatively the particular advantage of spike thermal anneals at temperatures above 1000 °C.

Analysis of structural defects in boron-implanted silicon single crystals on the basis of the results of double-and triple-crystal x-ray diffractometry

The structural defects in Si single crystals are analyzed on the basis of diffraction reflection curves and triple-crystal spectra. The relative variation of the lattice period and its distribution as a function of depth are calculated, and the type of defects appearing and the behavior of the implanted impurity in response to high-temperature annealing are determined.

Process effects in shallow junction formation by plasma doping

We have characterized the effect of implant parameters on the characteristics of boron-implanted silicon prepared by plasma doping. The thickness of the amorphous layer was monitored using Rutherford backscattering spectrometry, transmission electron microscopy, and double crystal x-ray diffraction. At low plasma implant pulse biases (2 kV), both dose rate and pulse repetition rate (10 Hz to 1 kHz) were found to have a negligible effect on the generation of the surface amorphous layer. At higher voltages (5 kV), the thickness of the amorphous layer increased with the pulse repetition rate but was apparently not sensitive to the dose rate. The sheet resistance after annealing correlated strongly with the increasing thickness of the surface amorphous layer formed at 5 kV. From these results, we conclude that an increase in pulse repetition rate has a stronger effect on the formation of the surface amorphous layer than does dose rate.

Infrared near-field imaging of implanted semiconductors: Evidence of a pure dielectric contrast

In this letter, we demonstrate the ability of our reflection mode scanning near-field optical microscope functioning in the mid-infrared to reveal infrared dielectric contrast in absence of any topographical contrast. This contrast is induced by local structures prepared by low energy boron implantation in silicon.

Metal gettering by boron-silicide precipitates in boron-implanted silicon

We show that Fe, Co, Cu and Au impurities in Si are strongly gettered to boron-silicide precipitates formed by supersaturation boron implantation and annealing. Effective binding free energies relative to interstitial solution are extracted and range from about 1.2 to 2.2 eV. The B-Si precipitates formed at temperatures ≤ 1100°C lack long range structural order but are similar to crystalline B 3Si in composition and B chemical potential. Evidence indicates that the metal atoms go into solution in the B-Si phase, and this is interpreted in terms of the novel bonding and structural characteristics of B-rich icosahedral compounds.

Characterization of vacancy-like defects in boron-implanted silicon with slow positrons

Vacancy-like defects in Si after boron implantation were studied by means of slow positrons. Ion implantation was carried out at 300 K subsequently with 3 energies to obtain a wide, homogeneously damaged layer. The width of the defect profiles increases with the ion dose and is in good agreement with Monte-Carlo simulations (TRIM-92). The defect production rate was ∼ N , where N is the boron fluence. This behavior is typical for defect formation via homogeneous nucleation. The divacancy concentration observed by infrared absorption was nearly constant after boron implantation, whereas the overall concentration of vacancy-like defects and the S-parameter in the implanted layer increase. This led to the conclusion that the divacancy is dominated by another defect of larger volume in highly boron-implanted Si. The samples were annealed in-situ up to 1150 K. The annealing behavior depends on the implantation dose and on the sample material (i.e. FZ- or CzSi). An annealing stage at 725 K was found in all samples. A decrease of the S-parameter below the value obtained for defect-free Si was observed in CzSi only after annealing at 750 K. This must be attributed to the formation of oxygen-vacancy complexes. An increase of the S-parameter above the defect value at room temperature was observed in a region 100 nm below the surface for high boron fluences after annealing at 650 K.

The damage recovery and electrical activation of shallow boron implants in silicon: The effects of high energy implants

A study of the interaction between shallow implanted boron junctions and high-energy silicon implants is presented. The implantation induced damage structure and its evolution with annealing temperature was followed by transmission electron microscopy. The electrically active boron depth profiles have been measured using spreading resistance profiling. The results show that the boron secondary defect formation is greatly suppressed by a high-energy silicon implant. A reduction by more than 50% has been observed. Moreover, the electrical activation is affected, and is found to be very sensitive to the initial damage structure, which depends on the implant conditions and subsequent thermal processes. The interaction between the damage structure of the low energy and high energy implants has been simulated by solving the appropriate set of fully coupled diffusion equations for interaction and diffusion of point defects. A comparison of the experimental findings with the results of the simulations is presented and discussed.

A technology oriented model for transient diffusion and activation of boron in silicon

We propose a model for point-defect-assisted transient diffusion and activation of high-dose boron implants in silicon. To model transient diffusion, a nonlinear equilibrium clustering model for point defects is used. The activation of boron is modeled as a Fermi-level-dependent transformation of inactive dopant clusters to substitutional atoms. Comparison with experimental data shows that this approach can provide a description of both rapid thermal annealing and long-time furnace annealing steps, with an excellent predictive capability for both chemical and electrically active profiles.

Conductivity of boron-implanted polycrystalline thin silicon films

Sub-micron thin films of polycrystalline silicon obtained by low-pressure chemical vapor deposition, boron-implanted (amount varying from 1012 to 1016 cm−2), and annealed at 760 °C for 26 h under an oxygen ambient (dry O2) have been characterized in terms of structural and electrical properties. The results obtained are correlated first with the Hall-effect measurements, which gives the resistivity p, the concentration of the free carriers p and the mobility μ, and in the second part with the observations of transmission electron microscopy (TEM), which gives the size of grains. The variations of p, p and μ have been studied as a function of the dopant concentration. A low rate of activation has been measured (40%) and a very low value of the mobility μ has been noted (3 cm2 V−1 s−1) as well as small grain size. These results have been discussed both qualitatively and quantitatively, based on the existence of phenomena of carrier trapping and dopant segregation at the grain boundaries.

Conductivity of boron-implanted polycrystalline thin silicon films

Sub-micron thin films of polycrystalline silicon obtained by low-pressure chemical vapor deposition, boron-implanted (amount varying from 1012 to 1016 cm−2), and annealed at 760 °C for 26 h under an oxygen ambient (dry O2) have been characterized in terms of structural and electrical properties. The results obtained are correlated first with the Hall-effect measurements, which gives the resistivity p, the concentration of the free carriers p and the mobility μ, and in the second part with the observations of transmission electron microscopy (TEM), which gives the size of grains. The variations of p, p and μ have been studied as a function of the dopant concentration. A low rate of activation has been measured (40%) and a very low value of the mobility μ has been noted (3 cm2 V−1 s−1) as well as small grain size. These results have been discussed both qualitatively and quantitatively, based on the existence of phenomena of carrier trapping and dopant segregation at the grain boundaries.

Comparison of Damage Accumulation Models for Boron Implantation in Silicon

ABSTRACTTwo models of damage accumulation during boron implantation are compared: The first one analyzes the full collision cascades and lets vacancies and interstitials recombine, if they are located within some capture radius of each other. The second one uses a constant fraction of defects surviving damage annihilation within a recoil cascade and a damage saturation density to take into account recombination with defects generated in previous cascades. While the first model is more fundamental, the second one is computationally more efficient. By comparing model predictions with 20 keV boron implantations at various doses, performed into (100) and (110) silicon with 7° and 0°, respectively, we conclude that different capture radii have to be used for damage annihilation within the recoil cascades and with defects generated in previous cascades. Moreover, we show that the two models are almost equivalent, if appropriate parameters are chosen. Recombination factors determined from simulations using a capture radius are almost independent of depth and implantation energy.

Boron depth profiles in silicon and simulation of α-spectra

A method for simulating α-spectra, when measuring the boron depth profiles by the 11B(p, α) 8Be at E p=660keV broad resonance reaction, is reported. The method is applied for obtaining the boron depth profiles in boron-diffused silicon, boron-implanted silicon and boron-doped amorphous silicon film. The results show that this method is valuable for the study of boron depth profiles.