In situ phosphorus‐doped epitaxial silicon films have attracted significant attention as source and drain materials because low specific contact resistivities have been achieved on such films by increasing the active carrier concentration using millisecond laser annealing. However, the active phosphorus concentration that can be achieved using millisecond laser annealing is much less than the incorporated concentration. To increase the activation efficiency, nanosecond laser annealing with a dwell time ≈104 times shorter than that of millisecond laser annealing is investigated and the diffusion, strain, microstructure, and electrical properties of single‐ and multipulse nanosecond laser‐annealed samples are examined. The melting depth simulation classifies the energy density regions and explains the limited diffusion in nanosecond laser annealing. After multipulse nanosecond laser annealing, more phosphorus is activated without diffusion than by millisecond laser annealing. Moreover, almost all the incorporated phosphorus atoms are activated by the nanosecond laser, which melts in situ phosphorus‐doped epitaxial silicon films without major strain loss. The increased active carrier concentration presents an opportunity to achieve low contact resistivity characteristics.

As transistor sizes reduce, the effect of contact resistivity on power consumption increases. To reduce contact resistivity, heavily phosphorus‐doped Si grown via in situ phosphorus‐doped (ISPD) Si epitaxial growth is studied actively. Laser spike annealing to the heavily phosphorus implanted (IMP) layers is demonstrated to replace ISPD Si epitaxial growth process and phosphorus profiles and strain characteristics are evaluated. Regardless of the doping method, the phosphorus concentrations of both samples and their tensile strains are equivalent. After laser annealing, the metal‐silicidation is conducted to measure contact resistivity. The Ni‐silicide formed on IMP sample has 3D clusters inducing greater morphological degradation than ISPD samples. The contact resistivity of IMP sample measured using the circular transmission line model (CTLM) (1.2–8.3 × 10−8 Ω cm2) is similar to that of the ISPD sample (1.1–5.5 × 10−8 Ω cm2) after Ni‐silicidation of the ISPD layer. This study performs strain engineering by achieving low contact resistance at lower cost while applying strain using the IMP process rather than the epitaxial process.

We propose a method to reduce the contact resistivity of titanium silicide (TiSi2) on phosphorus-doped epitaxial silicon by introducing a thin Se layer at the Ti/Si interface. The Se interlayer delays transition from the C49 to the C54 phase and changes the dominant diffusing species of TiSi2 formation from Si to Ti. In addition, the Se interlayer worsens the interface roughness between TiSi2/Si. The contact resistivity of the sample with the Se interlayer improves by one order of magnitude, which is significant. This improvement is attributed to the suppressed diffusion of P and low Schottky barrier height.

An amorphous Si (a-Si) solar cell with a back reflector composed of zinc oxide (ZnO) and silver (Ag) is potentially the most plausible and flexible solar cell if a graphite sheet is used as the substrate. Graphite supplies lightness, conductivity and flexibility to devices. When a graphite sheet is used as the substrate, carbon can diffuse into the Ag layer in the subsequent p-i-n process at 200–400 °C. To prevent this, we added an oxide layer as a carbon diffusion barrier between the carbon substrate and the back reflector. For the carbon diffusion barrier, silicon oxide (SiO2) or tin oxide (SnOx) was used. We evaluated the thermal stability of the back reflector of a carbon substrate using secondary-ion mass spectrometry (SIMS) to analyze the carbon diffusion barrier material. We confirmed the deposition characteristics, reflectance and prevention of carbon diffusion with and without the barrier. Finally, the structures were incorporated into the solar cell and their performances compared. The results showed that the back reflectors that were connected to a carbon diffusion barrier presented better performance, and the reflector with an SnOx layer presented the best performance.