ELSEVIER Microelectronic Engineering 46 (1999) 255-258 Comparative Study of AZPF514 and UVIII Chemically Amplified Resists for Electron Beam Nanolithography Zheng Cui and Phil Prewett Central Microstructure Facility, Rutherford Appleton Laboratory, Chilton Didcot, Oxon OX11 0QX, UK Two commercial positive tone chemically amplified resists, AZPF514 and UVIII, were investigated and compared for their performance in e-beam direct write of wafers at sub-150nm resolutions. The resist sensitivity, process latitude and post exposure delay effect were studied. Both resists are of high sensitivity. They are insensitive to variation of post exposure bake conditions. The post exposure vacuum delay effect is much more pronounced in AZPF514. AZPF514 also suffers severely the "T-topping" effect. Feature dimension below 150nm cannot be achieved with AZPF514, while sub-50nm lines have been obtained with UVIII. 1. INTRODUCTION The IC industry is on a cross road in choosing microlithography tools for 130nm and below feature dimensions. Among a number of non-optical tools, electron beam direct write or mix-and-mmch lithography is an attractive choice because of the potential of using the tool for several generations of ICs. In the mix-and-match mode, e-beam lithography is most likely to be used to define contact holes and poly lines, where both positive and negative tone resists are used. The conventional high resolution positive tone e-beam resist for contact hole definition has been the PMMA, which is unsuitable for dry etch pattern transfer and its very low sensitivity prohibits high throughput direct write of wafers. Chemically amplified resists (CARs) have demonstrated their high sensitivity in deep UV optical lithography and high resistance to dry etch. Although there have been reports of using CARs in high resolution e-beam lithography [1-2], to employ them in an industrial environment for future generation IC production questions remain to be answered as to the exposure and process latitude, post exposure bake, post exposure delay effect and their high resolution capabilities. In this paper, results are presented on the comparative study of AZPF514 and UVIII positive chemically amplified resists for e-beam lithography. Process conditions were optimised for both resists at target resolutions of 150nm and below. The effects of exposure and development condition on resist sensitivity and contrast were investigated. The 0167-9317/99/$- see front matter PII: S0167-9317(99)00075-1 important issues related to CARs, such as the post exposure bake latitude and the post exposure delay effect, have been quantified. Sub-50nm features have been achieved using UVIII resist. 2. EXPERIMENT The AZPF514 and UVIII chemically amplified resists are commercially available from Clariant and from Shipley. Electron beam exposures were carried out using Leica EBMF10.5, EBML-300 and VB6- HR vector beam systems at 50keV and 100keV, with resist thickness ranging from 0.5~tm to 1.0gin. Silicon wafers were used as the substrate. Post applied bake (PAB) or softbake and post exposure bake (PEB) were carried out on a hot plate. For AZPF514, the optimum PAB condition is 120°C for 2 minutes and the optimum PEB condition is 60°C for 1 minute. For UVIII, the PAB is at 150°C for 1 minute and the PEB is at 140°C for 1 minute. PEB effect for both resists was investigated by varying the temperature and time. Wafers were normally baked immediately after the e-beamexposure to avoid any delay effects. For the investigation of post exposure delay (PED) effect, the wafers were left in a vacuum chamber after the exposure for a controlled period before the baking. The standard developers, AZ518MIF for AZPF514 and CD26 for UVIII, were used with varied development time. Wafers were developed in immersion mode. The e- beam lithography results were quantified by CD measurements of resist feature cross-sections using the on-line metrology tool of a Hitachi-S4000 SEM. © 1999 Elsevier Science B.V. All rights reserved.