Si/Ge multilayer structures formed by the deposition of relatively small amounts of Ge layers (less then the critical thickness for 3D islands formation) are obtained by molecular beam epitaxy. Their structural and optical properties are investigated in detail. Appropriate growth parameter of the stacked island structures lead to significant increasing of the luminescence efficiency even at room temperature.

Scaling of devices requires not only shallow junctions but also high levels of dopant activation. For boron as the main p-type dopant, the latter requirement is especially problematic since small clusters of boron atoms and self-interstitials, known also as boron-interstitial clusters (BICs), were found to deactivate and immobilize large fractions of the implanted atoms during post-implantation annealing. In this article, the properties of BICs are reviewed and their influence on semiconductor processes are highlighted.

Fluorine and boron co-implantation within amorphous silicon has been studied in order to meet the process challenges regarding p+ ultra-shallow junction formation. Previous experiments have shown that fluorine can reduce boron TED (Transient Enhanced Diffusion), enhance boron solubility and reduce sheet resistance. In this study, boron diffusion characteristics prior to solid phase epitaxial regrowth (SPER) of the amorphous layer in the presence of fluorine are addressed. Samples were pre-amorphized with Si+ at a dose of 1x1015 ions/cm2 and energy of 70 keV, leading to a deep continuous amorphous surface of approximately 1500 Å. After pre-amorphization, B+ was implanted at a dose of 1x1015 ions/cm2 and energy of 500 eV, while F+ was implanted at a dose of 2x1015 ions/cm2 and energies ranging from 3 keV to 9 keV. Subsequent furnace anneals for the F+ implant energy of 6 keV were conducted at 550°C, for times ranging from 5 minutes to 260 minutes. During annealing, the boron in samples co-implanted with fluorine exhibited significant enhanced diffusion within amorphous silicon. After recrystallization, the boron diffusivity was dramatically reduced. Boron in samples with no fluorine did not diffuse during SPER. Prior to annealing, SIMS profiles demonstrated that boron concentration tails broadened with increasing fluorine implant energy. Enhanced dopant motion in as-implanted samples is presumably attributed to implant knock-on or recoil effects.

Traditionally, the choice of dopant has been limited to those species with the highest solid-solubility, however, Laser Thermal Processing (LTP) is not fundamentally limited by solid-solubility. Therefore, it is of interest to evaluate alternate dopants that have previously been excluded due to low solid-solubility. To this end, alternate dopants of 14N, 121Sb (n-type), 27Al, 70Ga, and 115In (p-type) are compared to conventional dopants of As and B respectively, after LTP and post-LTP thermal processing. Dopants were implanted into <100> silicon wafers of opposite background doping type that had previously been amorphized to a depth of approximately 300 angstroms by a 15 keV 28Si+ implant of 1x1015/cm2 dose. An implant energy of 5 keV was sufficiently low to confine the implanted ions to the amorphous layer, with the exception of B, which required an energy of 2 keV. All species were implanted at doses of 1x1014, 5x1014 and 1.5x1015/cm2. Samples were LTP utilizing a 308 nm, 18 ns laser pulse with a fluence of 0.680 J/cm2. Post-LTP thermal processing of the samples consisted of a 900 °C rapid thermal anneal (RTA) in a nitrogen ambient for a duration ranging from spike to 300 seconds. Measurements of the sheet resistance, mobility and carrier concentration were taken after both LTP, and the post-LTP thermal processing. Experimental results show that LTP of alternate dopants increases the electrically active carrier concentration of Ga, Al and Sb above solid-solubility. Additionally, the amount of deactivation upon post- LTP thermal processing depends on the alternate dopant species.

The work investigates simple transient enhanced diffusion (TED) behavior, which is a reflection of the interstitial behavior in the system. The analysis shows that TED depends mainly on two factors: the intial demage profiles and the DICI product. We find that, based on these two inputs, the extent of TED can be accurately diffusion capacity (DICI) which is compared to values previously extracted from diffcusion and silicon self-diffusion experiments.

Scanning Spreading Resistance Microscopy (SSRM) is now widely used for two-dimensional doping profiling with high spatial resolution. The need for a high force between the tip and the sample in order to obtain a good electrical contact, leads to a fast degradation of the tip (and the sample) while scanning. Tip damage is mainly due to the shear force occurring while scanning in contact mode at high forces leading to breakage (cleavage) of sharp tips or a rapid increase of tip radius (wear). The latter adversely affects the accuracy of the electrical measurements, as the contact radius is a determining parameter for quantification. The strong abrasive force also necessitates the use of tips composed of very hard material such as doped diamond, which has however a limited resistivity, and so far prevented the use of metallic probes. In addition the high force also prevents the simultaneous acquisition of high quality topography data. The solution to these problems is obtained by implementing the Modulated Force Principle (MFP). The latter consists of applying a variable (for instance pulsed) force while scanning, reducing the force during the lateral movement of the tip and synchronizing the electrical measurements with the high force periods. The latter results in lower lateral forces and introduces a quasi multi point contact mode. MFP also allows to obtain a better topography image by synchronizing the topography measurement with the low force part of the force cycle. The MFP leads to a drastic reduction of the surface and probe damage while maintaining high quality electrical data. The implementation of multiplexed detectors within the force cycle further enables the simultaneous acquisition of spreading resistance and topography during one scan, and/or the combination with multiple linear current detectors, capacitance sensors or tunneling current measurements.

A radiochemical neutron activation analysis (RNAA) procedure has been developed, critically evaluated, and shown to have the necessary sensitivity, chemical specificity, and matrix independence to certify phosphorus at ion implantation levels. 32P, produced by neutron capture of 31P, is chemically separated from the sample matrix, and measured using a beta proportional counter. Analysis of 12 pieces of a silicon wafer, implanted at a nominal dose of 8.5 x 1014 atoms·cm-2, yielded a phosphorus concentration of (8.30 ± 0.16) x 1014 atoms·cm-2 (expanded uncertainty), in agreement with a consensus value of (8.23 ± 0.33) x 1014 atoms·cm-2 (1s uncertainty) determined from a round robin SIMS investigation. The relative standard deviation of < 2 % is suitable for certification of a Standard Reference Material.

Carrier Illumination™ (CI) is an optical technique for non-destructive in-line monitoring of post-anneal junction depth and pre-anneal PAI depth and dose with wafer mapping capabilities. This work intends to extend the use of the CI-measurements from a range-specific quantitative measurement towards a more universal quantitative analysis of junction depth, profile abruptness and implant dose. For that purpose this paper presents a systematic study of the CI response to a wide variety of post anneal implant processes, varying parameters including implant species, dose and energy, annealing condition, and surface preparation.

Samples containing B, BF2 and As-implants with and without Ge PAI layers, with junction depths between 10-120 nm, were measured. In addition near-ideal box-like profiles (as obtained with CVD-growth) were fabricated and measured. For the abrupt CVD profiles, CI measures the junction position with sub-nm resolution independent of the CI-analysis conditions. For more graded profiles resulting from annealed implants, the correlation to the SIMS junction depth becomes a function of generation laser current (which is proportional to the applied power). As the concentration level, at which the correlation is made, can be adjusted over a concentration range of approximately 2x1018 to 2x1019/cm3 by changing the laser current, a route towards correlating the CI measurement with profile abruptness becomes feasible.

In this work, variations in electron potential are incorporated into a Kinetic Lattice Monte Carlo (KLMC) simulator and applied to dopant diffusion in silicon. To account for the effect of dopants, the charge redistribution induced by an external point charge immersed in an electron (hole) sea is solved numerically using the quantum perturbation method. The local carrier concentrations are then determined by summing contributions from all ionized dopant atoms and charged point defects, from which the Fermi level of the system is derived by the Boltzmann equation. KLMC simulations with incorporated Fermi level effects are demonstrated for charged point defect concentration as a function of Fermi level, coupled diffusion phenomenon and field effect on doping fluctuations.

The transient enhanced diffusion (TED) of high-dose implanted P is simulated taking into account Ostwald ripening of end-of-range (EOR) defects. First, we integrated a basic diffusion model based on the simulation of in-diffusion, where no implanted damages are involved. Second, from low-dose implantation, we developed a model for TED due to {311} self-interstitial (I) clusters involving Ostwald ripening and the dissolution of {311} clusters. Third, from medium-dose implantation, we showed that P-I clusters should be taken into account, and during the diffusion, the clusters are dissolved to emit self-interstitials that also contribute to TED. Finally, from high-dose implantation, EOR defects are modeled and we derived a formula to describe the time-dependence for Ostwald ripening of EOR defects, which is more significant at higher temperatures and longer annealing times. The simulation satisfactorily predicts the TED for annealing conditions, where the calculations overestimate the diffusion without taking Ostwald ripening into account.

Implanted fluorine has been observed to behave unusually in silicon, manifesting apparent uphill diflusion [1]. We are further motivated to understand the behavior of implanted fluorine in silicon by experiments which suggest that fluorine reduces boron diflusion [2, 3, 4, 5] and enhances boron activation in shallow junctions [2, 3]. In order to investigate fluorine behavior, we calculated the energy of fluorine defect structures in the framework of density functional theory (DFT). Besides identifying the ground-state con.guration of a single fluorine atom in silicon, a set of energetically favorable fluorine defect structures were found. The latter strongly suggests a distinct fluorine diflusion mechanism, which was implemented in a continuum diflusion simulation and compared to experimental data.

High-resolution x-ray diffraction (HRXRD) was used to monitor silicon-on-insulator (SOI) processing steps. The use of HRXRD is attractive since it is non-destructive and can be applied directly to product wafers. We show the usefulness of this technique for the characterization of amorphizing implants for shallow junctions, solid phase re-crystallization of implanted junctions, cobalt-silicide formation, and oxidation; all are critical processes for CMOS fabrication on SOI.

Single-shot Excimer Laser Annealing (ELA) was performed onto Si surface that was previously B+ implanted with or without Ge+ pre-amorphization. As a result, p+ type USJ (Ultra-Shallow Junction) has been formed. In process analysis, using Infrared Spectroscopic Ellipsometry (IR-SE) has been performed and compared with conventional 4-point probe method. Also, the corresponding crystallinity for the USJ of Si surface has been studied using Ultraviolet-Visible (UV-Vis) Spectroscopic Ellipsometry. In the case of pre-amorphization by Ge+ implantation, the laser energy density threshold required for melting the surface, and therefore for electrical activation, decreased drastically because of the difference in the thermodynamic properties of the amorphized Si. Estimation of the junction depth shows a shallower junction when using UV-SE, as compared to IR-SE. This can be explained by the fact that, in the UV range, the crystallinity of the top layer is predominant while IR-SE is more sensitive to dopant activation. This efficient single-shot ELA is a candidate for the USJ formation for sub-0.1 νm CMOS transistors. The effective method for investigating the activation state related to the crystallinity by using UV-SE and IR-SE is expected to apply as a non-contact analytical tool for USJ formation.

CVD-grown lightly C-doped superlattices with peak C concentrations of 2.1018/cm2 and 2.1019/cm2 were annealed in NH3, N2/H2, N2, and O2 ambient gases to investigate the influence of a range of point-defect conditions on C diffusion at the nanometer scale. C profiles were measured by secondary-ion mass spectroscopy. The profiles exhibit exponential-like diffusion consistent with a ‘long hop’ diffusion process with a characteristic migration length λ (=19 ± 3 nm at 850°C). Within experimental errors the value of ë is the same for all the ambient gases used, whereas the migration frequency g increases by two orders of magnitude as the ambient gas is changed from NH3 ambient (interstitial undersaturation) to O2 ambient (interstitial supersaturation), and decreases as a function of C concentration in the as-grown superlattice. The results confirm that C diffuses predominantly by a kick out mechanism under nearequilibrium diffusion conditions. Initial results support the chemical-pump model for suppression of diffusion in C-doped silicon.

In traditional models of P diffusion in Si, vacancy assisted diffusion mechanism has been assumed. More recently, experiments have determined that for intrinsic P diffusion in Si, the interstitial assisted diffusion mechanism dominates. We have performed ab-initio pseudopotential calculations to study P diffusion in Si. Special care is taken with regard to structure minimization, charge state effects and corrections. We calculated the defect formation energies and migration barriers for the various competing P-interstitial complex diffusion mechanisms for low concentration P region, as well as the energetics of different charge states P-vacancy complex diffusion. For interstitial mediated diffusion of P-Sii pair in Si, we find the overall diffusion activation energies calculated are 3.1 eV for neutral case, and 3.4 eV for +1 charge case. This is in agreement with experimental observation that the interstitial mechanism dominates for intrinsic P diffusion in Si. For vacancy mediated diffusion, our calculations are in agreement with previous calculations result in the neutral case. We obtained the lower bounds for diffusion activation energy of 3.8 eV for (PV)0 and 3.4 eV for (PV). A further evaluation of the numbers would require a proper treatment of the energy states in the band-gap due to Jahn-Teller relaxations.

Suppression of short channel effect (SCE) by utilizing the technology of formation of ultrashallow junctions is one of the important issues. The annealing process of implantation-damage that induces transient enhanced diffusion during a subsequent thermal process, such as low-pressure chemical vapor deposition (LPCVD) for gate sidewall spacer, should be optimized. To pursuit high performance of MOSFETs, parasitic resistance must be reduced with scaling. On the other hand, it is difficult to decrease the parasitic resistance in the region of contact junction, which is a function of physical constant such as Schottky barrier height of silicide materials and solid solubility of dopant. The elevated source/drain structure reduces parasitic resistance of contact junction due to reduction of resistance of diffusion beneath salicide materials. Cobalt salicide is widely used till 100nm node. However, cobalt salicide has disadvantage in the thermal budget for shallow junction and quantity of silicon consumption during silicidation. Nickel salicide is one of the candidates for successor of cobalt salicide to 70 nm node or above, because of its characteristics of low temperature formation, low silicon consumption and low contact resistivity on p+ junctions. In this paper, S/D engineering will be discussed from the viewpoint of the process integration of sub-100 nm physical gate length complementary metal-oxide-semiconductor field-effect transistor (CMOSFET) device.

Transient enhanced diffusion (TED) is a challenge that the semi-conductor industry has been faced with for more than two decades. Numerous investigations have been conducted to better understand the mechanisms that govern this phenomenon, so that scale down can be acheived. {311} type defects and dislocation loops are known interstitial sources that drive TED and dopants such as B utilize these interstitials to diffuse throughout the Si lattice. It has been reported that a two-step anneal on Ge preamorphized Si with ultra-low energy B implants has resulted in shallower junction depths. This study examines whether the pre-anneal step has a measurable effect on the end of range defects. Si wafers were preamorphized with Ge at 10, 12, 15, 20 and 30keV at a dose of 1x1015cm-2 and subsequently implanted with 1x1015cm-2 1keV B. Furnace anneals were performed at 450, 550, 650 and 750°C; the samples were then subjected to a spike RTA at 950°C. The implant damage was analyzed using Quantitative Transmission Electron Microscopy (QTEM). At the low energy Ge preamorphization, little damage is observed. However at the higher energies the microstructure is populated with extended defects. The defects evolve into elongated loops as the preanneal temperature increases. Both the extended defect density and the trapped interstitial concentration peak at a preanneal temperature of 550°C, suggesting that this may be an optimal condition for trapping interstitials.

During spike annealing of ultra-shallow junctions, large fractions of the dopants form a partially active pile-up at the interface between silicon and the screening oxide layer. In this paper, we show results of sheet resistance, SIMS and high resolution Elastic Recoil Detection measurements to investigate the physical and electrical behaviour of B and As dopant atoms at the interface.

Our results show that the fraction of dopants segregated at the interface is as high as 30-50% for B, but is dependent on dose and the type of screening oxide. Concentrations of up to 3e20cm-3 and more of active dopants are found on the Si side of the interface. The presence of nitrogen in the oxide at the interface causes a higher and sharper pile-up. Results indicate that a similar peak is expected for As, with active concentrations above 6e20cm-3. In an HF dip, the pile-up is removed together with the oxide or deactivated during native oxide regrowth.

Further experiments show that immediately after removing the screening oxide in an HF dip the sheet resistance for B decreases sharply due to carrier accumulation, then raises to about 6-9% above the initial level depending on the oxide and dopant species. The sharp decrease in resistance is not observed for As.

In this contribution we present a model for transient enhanced diffusion of boron in silicon. This model is based on the usual pair diffusion mechanism including non-equilibrium reactions between the dopant and the free point defects, taking into account their various charge states. In addition to, and fully coupled with the dopant diffusion we model the growth and dissolution of the interstitials and boron interstitials clusters associated with the anneal of the self-interstitial supersaturation created by the implantation step. It is thus possible to simulate a rather large set of experimental conditions, from conventional predeposition steps, to RTA after low energy implantation.

One of the important challenges in semiconductor industry is to sustain high concentration of dopant atoms electrically active in very small areas. In investigating the optimum post implantation treatment methods that will help to attain these conditions, the local structural information around the dopant atom is crucial. In this study, we have used secondary ion mass spectroscopy (SIMS) and x-ray absorption fine structure spectroscopy (XAFS) to obtain the concentration depth profiles and the local structural information around the Sb atom in laser thermal annealed (LTA) Sb implants in Si wafers. The Sb implant doses used in this work are 6.4x1015/cm2 and 2.0x1016/cm2. The XAFS results for the 6.4x1015/cm2 Sb dose sample do not exhibit any rhombohedral-Sb precipitation as the Fourier Transformed (FT) data can be fit successfully using only substitutional-Sb in the Si lattice sites. However, a multi-shell analysis of the Fourier Transformed (FT) data for the 2.0x1016/cm2 Sb dose sample clearly indicates there is a substantial contribution from the Sb-Sb scattering, which is a signature of precipitated form of Sb.