Diffusion of iron in the silicon dioxide layer of silicon-on-insulator structures

Abstract

The redistribution of iron implanted into the oxide layer of silicon-on-insulator structures has been measured using the secondary ion mass spectroscopy technique after annealing at 900–1050 °C. Iron diffusion has been found to be much faster in the oxide prepared by the separation-byimplantation-of-oxygen (SIMOX) procedure compared to the thermally grown oxide in the bonded and etched-back structures. In the latter case, the Fe diffusivity exhibits a thermal activation with an energy of 2.8 eV, confirming the literature data on silica glass. In the SIMOX oxide, the diffusivity depends only weakly on temperature, indicative of an essentially activation-free diffusion mechanism. Gettering of Fe at below-the-buried-oxide defects in SIMOX wafers has been observed. No iron segregation has been detected at the SiO2–Si interfaces.

Synthesis and characterisation of carbon nanotubes grown on silica fibres by injection CVD

“…Note that the MWCNTs shown in Figure 3c are at the larger end of the size distributions shown in Figure 2a, as they are located at the base of the longest growth experiment, and have been subjected to the greatest ripening; the majority of MWCNTs in the sample are significantly smaller. At high temperature, iron may diffuse into silica, forming iron oxide and an iron-silicon alloy [34,35]. Qu et al [36] has reported similar surface damage caused by the high-temperature pyrolysis of iron phthalocyanine (FePc) on silica-coated carbon fibres.…”

Modeling of Competitive Gettering of Iron in Silicon Integrated Circuit Technology

“…Kononchuk et al 26 fitted iron distribution profiles in SOI wafers obtained by SIMS after heat-treatments. They found that in bonded SOI wafers the diffusivity of iron was 2 ϫ 10 Ϫ14 cm 2 /s at 1050°C, and 10 Ϫ15 cm 2 /s at 900°C.…”

Critical Oxide Thickness for Efficient Single‐Walled Carbon Nanotube Growth on Silicon Using Thin SiO2 Diffusion Barriers

“…Catalyst diffusion through the oxide is determined by the temperature dependent diffusion constant D = D o e −ǫ/kT and the corresponding diffusion length λ ∼ √ Dt, where ǫ is the activation energy for diffusion and t is the diffusion time. Using the bulk diffusion constant for iron in electronic grade silicon dioxide (D o ∼ 10 −4 cm 2 s −1 , ǫ = 2.8 eV ) [40] one would only expect the iron to diffuse ∼ 1.5 nm during the temperature ramp to 900 o C, roughly three times shorter than observed in XPS. The 4 nm diffusion length observed can be explained by a small reduction of ∼ 0.18 eV in the activation energy, yielding a five-fold increase in the effective diffusion…”

“…Indeed, previous studies of metals on ultra-thin silicon oxides have shown that pinholes or defects in the oxide significantly enhances the diffusion process. [40,41,42,43] By determining the critical oxide thickness that prevents silicide formation for iron catalyst, we can extrapolate to determine the critical thickness for cobalt and nickel which are the other common singlewalled nanotube catalysts. For cobalt, the bulk diffusion parameters are…”