Patterning of porous SiOCH using an organic mask: Comparison with a metallic masking strategy

Abstract

The etching of sub-100-nm porous dielectric trenches has been investigated using an organic mask. The etching process that is performed in an oxide etcher is composed of three steps: a thin dielectric antireflective coating (DARC) layer (silicon containing layer) is etched in the first step, the organic mask [carbon-based layer (CL)] is opened in the second step, and the dielectric layer is etched in the last step. The DARC layer is open in a fluorocarbon-based plasma (CF4∕Ar∕CH2F2) and the main critical dimension issue is the critical dimension control of the trench, which can be adjusted by controlling the amount of polymer generated by the etching chemistry (% of CH2F2). The CL is etched using NH3 based plasmas, leading to straight trench profiles. For dielectric patterning, the etch process results from a delicate trade-off between passivation layer thickness and mask faceting. This is driven by the polymerizing rate of the plasma (% of CH2F2) which controls the trench width. Using an optimized etching process (CF4∕Ar∕2%CH2F2), p-SiOCH trenches can be patterned with straight etch profiles down to 75nm trench width. In this article, the authors have also compared the organic mask and TiN metal hard mask strategies in terms of patterning performances (profile control, porous SiOCH modification, and reactor wall cleaning processes).

Differences in erosion mechanism and selectivity between Ti and TiN in fluorocarbon plasmas for dielectric etch

“…By comparing the BP of TiF 4 and NF 3 , it is obvious that the evaporation probability of NF 3 during plasma processing will be significantly higher than for TiF 4 and Ti removal will be the erosion limiting factor.…”

“…For addition of FC gas, i.e., CF 4 or C 4 F 8 , to the Ar discharge, erosion is taking place by chemical sputtering where removal rates of Ti atoms are increased due to reduced threshold energies of TiF 4 discuss plasma etching of TiN in terms of the boiling point (BP) of volatile species and conclude that in FC plasmas the main etch products are TiF 4 (BP ¼ 284 C) and NF 3 (BP ¼ À128.8 C), which have significantly lower BP than TiF 3 (BP ¼ 1400 C). By comparing the BP of TiF 4 and NF 3 , it is obvious that the evaporation probability of NF 3 during plasma processing will be significantly higher than for TiF 4 and Ti removal will be the erosion limiting factor.…”

“…The highest ES ($15) are achieved for the less polymerizing CF 4 -based plasma conditions where material erosion is an ion-driven process and hardly affected by the thin FC-layer. For more polymerizing C 4 F 8 -based plasmas, material surfaces are covered with a thicker FC film and OSG erosion is limited by both diffusion of F through the FC layer, reduced ion bombardment, and consumption of F by C and H in the OSG (formation of volatile CF 4 and HF etch products), leading to reduced ES. 6 This is in strong contrast to organic masking materials which provide the highest ES in more polymerizing plasma chemistries, e.g., C 4 F 8 /Ar discharges, since organic masks have a higher C and H content and ER are more strongly reduced by a limited supply of F than OSG.…”

“…In Sec. III B, we compare surface composition and ER of Ti and TiN in CF 4 /Ar and C 4 F 8 /Ar plasma discharges for various plasma processing parameters, i.e., self-bias voltage (V SB ), pressure, %CF 4 in CF 4 /Ar, and N 2 and O 2 addition to C 4 F 8 / Ar discharges. In Sec.…”

“…3 So far, research has mainly focused on TiN due to high optical transparency for layers thinner than 20 nm leading to good photolithographic performance in multilayer masking schemes. [3][4][5] Previously, we have studied the erosion of Ti and organosilicate glass (OSG) in CF 4 /Ar and C 4 F 8 /Ar plasma discharges to gain insight into material-and plasma-dependent erosion behavior. 6 For Ti, we have identified three characteristic erosion stages by real-time ellipsometry and post-process x-ray photoelectron spectroscopy (XPS) characterization: (1) removal of a highly oxidized surface, (2) steady state erosion of highly fluorinated Ti covered with a fluorocarbon (FC) layer ($0.9 nm for CF 4 /Ar and $1.5 nm for C 4 F 8 /Ar plasmas), and (3) erosion of SiO 2 after hardmask erosion with some fluorinated Ti remaining on the surface ($2 at.…”