Journal of Vacuum Science & Technology B, Vol.25, No.5, 1593-1602, 2007
On the photoresist stripping and damage of ultralow k dielectric materials using remote H-2- and D-2-based discharges
Blanket films of ultralow dielectric constant (ULK) materials and 193 nm photoresist films have been processed downstream from hydrogen and deuterium-based discharges produced using an inductively coupled plasma reactor. Photoresist ashing rates and ULK modifications have been determined as a function of process parameters. The explored ULK materials differed widely in porosity and carbon content. The effect of processing time, substrate temperature (200-300 degrees C), and gas composition on the surface and bulk chemical composition of ULK materials was monitored and quantified by ex situ ellipsometry and time-of-flight secondary ion mass spectrometry (SIMS). The stripping rates of 193 nm photoresist films were found to strongly depend on processing temperature and only weakly on the nature of the H-2/additive gas mixture. The authors found that hydrogen (or deuterium) fully penetrates the high porosity ULK layer, whereas for low porosity materials, such penetration is limited to a 50 nm near-surface region. SIMS measurements also reveal that H-2 (D-2) diffusion into carbon-rich ULK layers can cause substantial carbon depletion throughout the penetration region. ULK damage values increase with temperature and injection of gas additives such as argon, helium, and nitrogen to H-2 or D-2 process gases. For each ULK material, the amount of damage depends on the gas mixing ratio; in general, high percentages of nitrogen in H-2/N-2 (or D-2/N-2) mixtures cause the most damage. Overall, the results demonstrate that ULK ashing damage depends strongly on both ULK material properties and H-2-based plasma process parameters.. In addition, the authors observed in this work a kinetic isotope effect for stripping of 193 um photoresist films in H-2/D-2/N-2-based discharges. For given ashing process conditions, the photoresist ashing rate decreases.by a factor of 1.414 (or square root of 2) in D-2 plasma compared to H-2 plasma. This can be explained by the influence of the H or D mass on the chemical reaction rate through a change in the frequency of nuclear vibrations of the reacting atoms. The presence of the kinetic isotope effect for gas mixtures provides unambiguous evidence of the rate-limiting role of atomic hydrogen in the fundamental etching reaction. Simultaneously processed ULK materials showed minor film thickness changes (< 10 nm) in H-2 or D-2 discharges, and the ULK damage level does not reflect a kinetic isotope effect. Therefore the H/D isotope effect can be used to separate H-2/D-2 associated ashing and etching processes from other chemistries or mechanisms. (C) 2007 American Vacuum Society.