Boron Diffusion In Si Research Articles

The Lattice Kinetic Monte Carlo Simulation of Boron Diffusion in SiGe

The lattice kinetic Monte Carlo simulation (kMCS) was applied to study the boron diffusion in Si-SiGe beyond nanotechnology. Both the interstitialcy and kick-out mechanisms of boron diffusion were considered, including the effects of annealing temperatures, boron dopant concentrations, Ge compositions, and concentrations of Si self-interstitial defects (SiI). The effects on boron diffusion caused by single and double layer(s) of SiGe phase with different Ge contents and varying boron concentrations in double layers of SiGe phase were also simulated. The results show that boron diffusion in Si and between SiGe-Si both largely increase as the temperature or concentration of SiI increases, but the boron diffusion between SiGe-Si is much less than in Si. Increasing the Ge contents in SiGe alloy could retard boron diffusion heavily, while increasing the boron concentration on SiGe phase would enhance boron diffusion.

Ab initioinvestigation of boron diffusion paths in germanium

Experimental data indicate that boron diffuses very differently in Ge than in Si. To examine the kinetics of boron diffusion, density functional calculations were performed on a variety of boron diffusion mechanisms, including interstitial and vacancy-mediated paths, as well as a correlated exchange mechanism. It was found that although vacancy and correlated exchange mechanisms possess high diffusion barriers comparable with experiment, the barrier for interstitial-mediated diffusion lies around $3.8\phantom{\rule{0.3em}{0ex}}\mathrm{eV}$ and is similar to those found for boron diffusion in Si. This estimate is well below the experimental activation energy. The difference is attributed to the failure of the theory to include the effect of electronic excitations.

Simulation of boron diffusion in Si and strained SiGe layers

Boron outdiffusion from the base into the emitter and collector caused by annealing in SiGe heterobipolar transistors (HBTs) has a serious influence on the transit frequency. One solution of the problem of boron outdiffusion is the creation of intrinsic spacers between the base, emitter and collector layers to prevent diffusion of boron across the heterointerface. For optimisation of SiGe HBT properties, several simulators are used. This paper presents a quantitative analysis of a SiGe HBT by process simulators SUPREM IV.GS and ISE TCAD-DIOS. Models for simulation of a boron-doped SiGe base of HBT are discussed and compared.

Effect of point defect injection on diffusion of boron in silicon and silicon–germanium in the presence of carbon

Boron diffusion in Si and strained SiGe with and without C was studied using point defect injection. Interstitial-, vacancy- and noninjection conditions were achieved by annealing Si capping layers which were either bare, with Si3N4 film or with Si3N4+SiO2 bilayers, respectively. Concentration profiles of B, Ge, and C were obtained using secondary-ion-mass spectrometry and diffusion coefficients of B in each type of matrix were extracted by computer simulation. Under inert annealing, we find that C strongly suppresses B diffusion in SiGe:C, but the effect of C is less strong in Si:C, particularly at high temperatures. In contrast, C only weakly suppresses B diffusion in both Si:C and SiGe:C under interstitial injection. For inert anneal conditions, C reduces the B diffusion coefficient in Si:C by factors of 4.2, 5.9, and 1.9 at 940, 1000, and 1050 °C respectively, whereas for interstitial injection the factors are 2.1, 1.3, and 1.1, respectively. The equivalent factors for SiGe:C are 8.4, 5.9, and 8.0 for inert anneal conditions and 2.2, 3.4, and 1.6 for interstitial injection conditions. The degree of B diffusion suppression achieved in both Si:C and SiGe:C is dependent on the level of C retained during annealing. Diffusion of C is shown to be faster in Si:C and hence less C is retained there after annealing than in SiGe:C. Interstitial injection is shown to strongly enhance C diffusion in both Si:C and SiGe:C and hence decreases the effectiveness of C for B diffusion suppression. These findings illustrate that the retarding effect of C on B diffusion in both Si:C and SiGe:C is strongly reduced when the anneal is carried out under conditions where interstitials are injected from the surface.

Diffusion of Boron in Silicon and Silicon-Germanium in the Presence of Carbon

Boron diffusion in Si and strained SiGe with and without C was studied. Using gassource molecular beam epitaxy (MBE), B containing epitaxial layers of: (i) Si, (ii) Si containing 0.1% C, (iii) SiGe with 11% Ge and (iv) SiGe with 11% Ge and with a 0.1% C, were grown on substrates. These samples were then rapid thermal annealed (RTA) at 940, 1000 and 1050°C in an O2 ambient. Self-interstitial-, vacancy- and non-injection conditions were achieved by annealing bare, Si3N4- and Si3N4+SiO2-coated surfaces, respectively. Concentration profiles of B, Ge and C were obtained using Secondary-Ion Mass Spectrometry (SIMS). Diffusion coefficients of B in each type of matrix were extracted by computer simulation. We find that B diffusivity is reduced by both Ge and C. The suppression due to C is much larger. In all materials, a substantial enhancement of B diffusion was observed due to self-interstitial injection compared to non-injection conditions. These results indicate that B diffusion in all four types of layers is mediated primarily by interstitialcy type defects.

Suppression of boron transient enhanced diffusion in silicon and silicon germanium by fluorine implantation

In this paper, a study is made of the effect of fluorine implantation on boron transient enhanced diffusion in Si and SiGe by characterising the diffusion of boron marker layers in Si and SiGe in samples with and without a 288 keV, 6×10 13 cm −2 P + implant and with and without a 185 keV, 2.3×10 15 cm −2 F + implant. It is shown that fluorine implantation totally eliminates boron transient enhanced diffusion caused by the phosphorus implant as well as significantly reduces boron thermal diffusion in both Si and SiGe. Fluorine SIMS profiles for both Si and SiGe samples show the presence of a shallow peak in the vicinity of the boron marker layer and a deep peak close to the range of the fluorine implant. Cross-sectional transmission electron micrographs show the presence of a band of dislocation loops that correlates with the position of the deep fluorine peak, but no defects are present in the position of the shallow fluorine peak. It is proposed that the shallow fluorine peak is due to vacancy-fluorine clusters that are too small to resolve by TEM. The role of the vacancy-fluorine clusters and the dislocation loops on the suppression of boron diffusion in Si and SiGe is discussed.

Transient enhanced diffusion of boron in Si

On annealing a boron implanted Si sample at ∼800 °C, boron in the tail of the implanted profile diffuses very fast, faster than the normal thermal diffusion by a factor 100 or more. After annealing for a sufficiently long time, the enhanced diffusion saturates. The enhanced diffusion is temporary, on annealing the sample a second time after saturation, enhanced diffusion does not occur. It is therefore designated as transient enhanced diffusion (TED). The high concentration peak of the implanted boron profile, which is electrically inactive, does not diffuse. TED makes it difficult to fabricate modern Si based devices, in particular TED produces the parasitic barriers which degrade the performance of the SiGe heterostructure bipolar transistors and TED can limit the fabrication of shallow junctions required for sub-100 nm complementary metal–oxide–semiconductor technology. The mechanisms of TED have been elucidated recently. A Si interstitial “kicks out” the substitutional boron atom to an interstitial position where it can diffuse easily. Alternatively the interstitials and boron atoms form highly mobile pairs. In both cases Si interstitials are required for the diffusion of boron. Therefore the enhanced boron diffusivity is proportional to the concentration of the excess Si interstitials. The interstitials are injected during implantation with Si or dopant ions. The interstitials are also injected during oxidation of the Si surface. Therefore the diffusivity increases temporarily in both cases. Even at relatively low annealing temperatures (∼800 °C) the mobility of the interstitials is high. The TED at this temperature lasts for more than 1 h. This large TED time can be explained by the presence of interstitial clusters and interstitial–boron clusters. The interstitial clusters are the {311} extended defects and dislocation loops. The precise structure of interstitial–boron clusters is not yet known though several models have been proposed. The clusters are the reservoirs of the interstitials. When the supersaturation of interstitials becomes low, the clusters dissolve and emit interstitials. The interstitials emitted from the clusters sustain the TED. Many groups have suggested that the rate of emission of interstitials is determined by Ostwald ripening of the clusters. However, recently TED evolution has also been explained without invoking Ostwald ripening of the {311} defects. The evidence of Ostwald ripening of dislocation loops is more direct. In this case the Ostwald ripening has been confirmed by the measurements of the size distributions of the dislocation loops at different times and temperatures of annealing. At higher temperatures the extended clusters are not stable and coupling between the interstitials and boron atoms is reduced. Therefore at high temperatures TED lasts only for a short time. At high temperatures the displacement during TED is also small. This suggests that if rapid thermal annealing with high ramp rates is used, TED should be suppressed. Currently high ramp rates, 300–400 °C/s are being tried to suppress TED.

Modeling of complete suppression of boron out-diffusion in Si 1− xGe x by carbon incorporation

In this paper, we demonstrate a simple modeling of boron diffusion in Si 1− x− y Ge x C y by manipulating the strain and the intrinsic carrier concentration. We show that the diffusion of boron is strongly suppressed by a moderate concentration of substitutional C in Si 1− x Ge x . This suppression is due to an under-saturation of Si self-interstitials in the C-rich region. The results obtained from the proposed model are in better agreement with the measured values.

Boron diffusion in Si and SiC during 2.5 MeV proton irradiation at 500–850°C

Radiation enhanced diffusion (RED) of B in Si and SiC was investigated using ∼85 nA/cm 2 2.5 MeV proton bombardment at elevated temperatures. A strong concentration dependence of RED of B in Si is observed. The mobilization of boron increases with increasing irradiation temperature (with an activation energy of 1.5 eV), so that at 850°C almost 100% of the boron atoms are mobile. For ∼1 h anneals at 700°C and 500°C the boron atoms in excess of 5 × 10 17 and 1 × 10 17 B/cm 3 stay immobile, respectively. A comparison of our data with that previously reported by other authors shows that the point defect wind originated from the depth of the proton projected range plays a significant role during the RED measurements in Si. No RED of B was observed in SiC and the migration/trapping parts of the activation energy of B diffusion in SiC are estimated to be >3.5 eV.

Retarded diffusion of boron in Si due to the formation of an epitaxial CoSi 2 layer

The diffusion of boron in silicon with and without a 20-nm thick epitaxial CoSi 2 layer was investigated by secondary ion mass spectrometry (SIMS) using doping superlattices (DSLs). For specimens without CoSi 2 layer, we observed oxidation enhanced diffusion (OED) of B in Si in accordance with the literature. However, B diffusion in silicide capped specimens was found to be retarded by about a factor of 20 as compared to specimens without a silicide layer. The measured diffusivity was even less than the intrinsic thermal diffusivity of B in Si.

Enhanced Diffusion of Boron in Si during Doping from Borosilicate Glass

Boron diffusion into Si from low‐pressure chemical vapor deposition (LPCVD) borosilicate glass (BSG) by rapid thermal annealing (RTA) has been studied. The influence of native oxide between BSG and Si on doping characteristics is small for LPCVD BSG, while an unintentionally formed thin oxide between BSG and Si causes a rapid change in surface boron concentration at the initial stage of diffusion for atmospheric pressure chemical vapor deposition BSG. Boron diffusion in Si is markedly enhanced for high‐boron‐concentration BSG. This is inherent to the BSG/Si structure, because enhanced diffusion is also observed for furnace annealing (FA). However, the enhancement is more pronounced for RTA, indicating that enhanced diffusion originating from RTA occurs for the BSG/Si structure. The origin of the enhanced diffusion in RTA is not related to rapid heating but to infrared light heating.

Etching and boron diffusion of high aspect ratio Si trenches for released resonators

This article discusses etching and boron diffusion in Si for the purpose of fabricating micromachined devices with high aspect ratio features. The etch rate of Si under various doping conditions in a Cl2 plasma generated by an electron cyclotron resonance (ECR) source was measured. It was found that lightly boron and phosphorus doped Si were etched at rates of 0.17 μm/min, whereas heavily boron doped Si had a similar etch rate of 0.16 μm/min. Si doped with high phosphorus concentrations had a faster etch rate of 0.31 μm/min. These etch rates were measured for Si etched at 100 W microwave power, 100 W rf power, 3 mTorr, with 20 sccm of Cl2 flow, and an ECR source to sample distance of 8 cm. The difference between the p++ and n++Si etch rates was more significant when etched at higher microwave power, higher rf power, or higher temperature. The depth of a heavily doped boron diffusion layer was measured for different feature sizes, trench openings, and aspect ratios. The diffusion layer thickness was found to decrease from 2.9 μm at the bottom of 50-μm-wide trenches, to 1.5 μm at the bottom of 2-μm-wide trenches. Diffusion thickness at the bottom of trenches near narrow features was found to increase to 3 μm compared to 2.5 μm for trenches near large features. The diffusion layer on the sides of the trenches for a 30 min boron diffusion at 1175 °C was 3.25 μm thick and was found to be independent of the trench opening and the trench aspect ratio. A deep etch-shallow diffusion process was used to fabricate thick released resonators with a short boron diffusion time. The second dry etch in the deep etch-shallow diffusion process, which is used to etch the p++ layer at the bottom of adjacent features, was studied. It was found that this etch successfully removed the p++ layer at the bottom of Si trenches in addition to widening the slightly tapered profile near the bottom of the trenches. This process has been applied to the fabrication of micromachined comb drive resonators and micromirrors successfully.

Effects of strain on boron diffusion in Si and Si1−xGex

Boron diffusion in in situ doped Si1−xGex and Si, subjected to inert-ambient furnace annealing at 800 °C, was investigated. For Si1−xGex, the effect of biaxial compressive and tensile strain on boron diffusion was studied using Si1−xGex layers with a constant Ge content (x≊0.10 and x≊0.20) grown epitaxially on various relaxed Si1−yGey (0≤y≤0.20) substrates. Boron diffusion is primarily a function of Ge content, x in the Si1−xGex layers, and does not show a strong dependence on macroscopic biaxial strain. For Si, the effect of biaxial tension was investigated using relaxed Si1−yGey layers as substrate templates for epitaxial Si layers. As in Si1−xGex, boron diffusion in Si does not depend strongly on biaxial strain.

In-Situ Doped Multi-Layer Epitaxial Structures with Abrupt Doping Transitions by Ultra High Vacuum Rapid Thermal Chemical Vapor Deposition (UHV-RTCVD)

ABSTRACTIn this work, we have studied boron doped multi-layer epitaxial structures as active regions of deep submicron (< 0.25 gtm)metal oxide silicon field effect transistors (MOSFETs).The structures were formed by ultra high vacuum chemical vapor deposition (UHV-RTCVD) using Si2H6 and B2H6 as the source gases and H2 as the carrier gas at 750°C - 800°C and at a total pressure of 80 mTorr. With the high growth rates provided by Si2H6, thermal budgets were kept below the limit of boron diffusion in Si resulting in extremely abrupt doping transitions pushing the depth resolution limits of secondary ion mass spectroscopy. The structures consist of three epitaxial layers with thicknesses ranging from 100 A to 625 T. The top layer on which the gate oxide is formed is lightly doped (lxl016 cm-3) to minimize vertical electrical field and ionized impurity scattering for higher MOSFET channel mobility. The second layer is doped to lx1018 cm-3 for suppression of punchthrough short channel effects and finally the third layer is doped to lx 1017 cm-3 to decrease the parasitic source/drain junction capacitance that will result from the relatively high doping density of the intermediate layer. To minimize dopant diffusion in Si, low temperature (or low thermal budget) processes were employed for gate oxidation and polysilicon implant activation. A typical source/drain activation anneal was also included in sample preparation in order to simulate complete MOSFET fabrication assuming remaining steps could be carried out at lower temperatures with little contribution to dopant diffusion. Our results indicate that after all process steps a lightly doped region can still be obtained under the gate oxide with sufficient thickness to contain the MOSFET inversion layer. In these structures, the threshold voltage is determined by the doping density and thickness of the top two layers and can be easily tailored to the desired value by optimizing these parameters. With the range of parameters used in this study, our measurements show threshold voltages within the range desired for 0.1 µm MOSFETs.

Measurement and modeling of boron diffusion in Si and strained Si1−xGex epitaxial layers during rapid thermal annealing

Boron concentration profiles in rapid thermally annealed Si and strained Si1−xGex in situ doped, epitaxial layers were measured using secondary-ion-mass spectroscopy. Comparison of the Si1−xGex samples to the Si samples after rapid thermal annealing revealed a retarded B diffusivity inside the strained Si1−xGex layers. A simple empirical expression for the B retardation, which depended linearly on the Ge concentration, was developed and incorporated into a diffusion model for dopants in heterostructures. This model accurately simulated the measured B concentration profiles over a wide range of Ge fractions (0%–10%), B peak concentrations (2×1018–3×1019cm−3), and rapid thermal annealing conditions (900–1025 °C for 20–30 s).

Comparison of boron diffusion in Si and strained Si1−xGex epitaxial layers

We have investigated boron diffusion in Si and strained Si1−xGex, in situ doped, epitaxial layers. During inert ambient annealing at 860 °C, boron diffusion is observed to be slower in Si0.83Ge0.17 than in Si for boron concentration levels between 5×1016 and 2.5×1019 cm−3. Computer simulations of the measured boron profiles for annealed samples indicate that the effective boron diffusivity Deff in Si0.83Ge0.17 is approximately an order of magnitude lower than that in Si. This disparity is found to increase with increasing boron concentration.

Suppression of anomalous diffusion of ion-implanted boron in silicon by laser processing

Anomalous diffusion of ion-implanted boron in silicon has been suppressed by using the laser annealing (LA) technique. For the rapid thermal annealing process, the high-dosage boron implant significantly enhanced the anomalous diffusion of boron in Si largely due to increased density of interstitial clusters. The mechanism that electrically active dopants contribute to diffusion is confirmed. The dopant activation is primarily determined from the heating process rather than the holding time interval. Hence, the optimum annealing regime, attaining high-performance shallow p+-n junctions, is to increase the dopant activation efficiency during the temperature-ramping process as well as to shorten the holding interval. Having an ultrahigh heating rate, the LA technique serves as an excellent annealing scheme to significantly suppress the anomalous diffusion and considerably promote the dopant activation.

Influence of implant condition on the transient-enhanced diffusion of ion-implanted boron in silicon

Under high-dosage boron implant, the implant condition was found to be important for reducing the transient-enhanced diffusion of boron in Si and forming shallow p+n junction with good rectifying characteristics as well as high dopant concentration in the diffused region. The BF2+-implantation resulted in not only an excellent dopant activation but also a reduced anomalous diffusion due to the formation of amorphous layer and the scarce defects underneath the amorphous/crystalline (a/c) interface. The B+-implanted crystalline samples manifested a poor activation efficiency, and a largely anomalous diffusion at the high temperature with prolonged-time annealing ascribed to much damage induced by the high-dose implant. The B+-implanted pre-Si+-amorphized samples also displayed severely transient-enhanced diffusion in spite of the good dopant activation.

The role of extended defects on transient boron diffusion in ion-implanted silicon

Transient tail diffusion of boron in Si(100) has been investigated as a function of boron implant condition, dose, energy, time, temperature and silicon or germanium post-implantations at doses above and below the amorphization threshold. Boron was implanted at energies in the range 5–380 keV and at doses in the range 2× 10 12 to 5 × 10 15 cm −2 either along [100] or in a random direction. Significantly longer transient boron tail diffusion is observed for implants along [100]. These results reflect the differences between the random and channeling implants in the position of the damage distribution relative to the boron profile. Post-implantation with 1 MeV 29Si ions below the amorphization threshold can reduce boron tail diffusion if the silicon dose is high enough (5 × 10 13 cm −2 or greater) to form extended defects during the anneal. Lower silicon doses do not influence boron diffusion. Annealing of extended defects also results in anomalous diffusion. These results demonstrate that silicon interstitials cause the enhanced boron diffusion. Amorphization of boron-implanted silicon with 29Si or 73Ge ions prevents transient boron tail diffusion during subsequent annealing if the boron profile is completely incorporated in the amorphized region.

Effect of Fluorine on the Dopant Diffusion of Through-Oxtoe Implanted Boron in Si - a Correlation with Microstructural Defects

ABSTRACTIn case of boron through-oxide implant, it has been shown that the knocked-in oxygen atoms segregate at initially nucleated dislocation sites during the incubation and no significant junction movement is detected. The trapping of oxygen proceeds up to a certain time at which oxygen-precipitation occurs and this leads to an ejection of excess Si interstitials and further enhancing boron diffusion. However, with fluorine addition we believe that fluorine incorporation in SiO2 and/or SiO2/Si interface not only releases the strain gradient but also suppresses the silicon interstitials ejection and by this means suppresses the oxidation-enhanced boron diffusion. Correlated results of TEM microdefect structures and spreading resistance profiles are used to further support our postulation.