Resist and Exposure Processes for Sub-10-nm Electron and Ion Beam Lithography

The research work described in this thesis deals with studying the ultimate resolution capabilities of electron and ion beam lithography (EBL and IBL respectively) with a focus on resist and exposure processes. The aim of this research was to enlarge knowledge and improve methods on the formation of ultra-high resolution structures. Research sub-projects were defined to focus on specific aspects of ultra-high resolution lithography. The results of these sub-projects form the basis of this thesis. After a short introduction to the principles of lithography, the key points to high resolution technology and the scope of the thesis (Chapter 1), Chapter 2 summarizes the most outstanding and successful attempts performed so far by researchers in the field to improve parameters influencing the lithographic process to achieve formation of ultimately small features at the highest density. The purpose of Chapter 2 is to give reader an understanding of complexity of the lithography process and to give insight into variable ways how to optimize. Therefore the emphasis is given to the experimental results on optimization of each lithography step. First, results on different methods of beam quantification and optimization of crucial parameters in the exposure system are presented. Next, the various tricks to optimise the lithographic process during resist treatments and post-exposure steps as well as during the most important steps – exposure, development and drying are discussed. The last part of this chapter focuses on experimental results of exposure of inorganic resist materials and exposure using ultra thin resist layers as exponents of ultimate resolution. The thesis research started with establishing a quantitative picture of the high resolution electron beam exposure process in ultra thin hydrogen silsesquioxane (HSQ) resist layers. These results are presented in Chapter 3. Three important contributions to quantitative electron beam patterning in the 2-10 nm regime are considered. First of all, electron beam measurements with an advanced knife-edge structure described in this chapter pointed to a minimum e-beam diameter of approximately ~3.9 nm in our lithography system, to be compared with the theoretical estimate of 3 nm. Then, it was demonstrated that adjustment of the beam focus by erroneous beam optimisation due to aging in the marker performance accounts for a broadening of smallest features up to about 3 nm. Additionally the linewidth broadening of about 13 nm in comparing ‘thin’ (5 nm) and ‘thick’ (50 nm) resist layers is attributed to a different density in the secondary electron (SE) exposure of the resist. The combined results provided a quantitative picture of limiting factors for the achievement of ultimate resolution. In chapter 4 the impact of resist thickness on the resolution performance of HSQ electron beam resist is investigated. Thickness of the resist layer was found to have a substantial influence on sensitivity, contrast and surface morphology. Dependence of electron cascade processes on thickness was shown to contribute to the variation of the resist sensitivity and the consequent linewidth broadening with increase of resist thickness. Monte Carlo (MC) simulations were used to get an insight into the exposure process and the energy transfer to the resist. The results of the MC modeling were found to be in good agreement with the experimental results and the proposed mechanism on sensitivity loss and structure linewidth broadening with increase of resist thickness. The molecular cluster size was shown to increase significantly with smaller resist thickness which enhances the surface roughness for ultra thin resist layers. Vacuum drying of resist films was demonstrated to result in smoothening of the resist surface, while it had no strong influence on the ultra high resolution patterning capability of HSQ resist. In Chapter 5, the influence of the exposure temperature on the performance of HSQ resist in ultra high resolution electron beam lithography was investigated. Dependencies of the HSQ contrast and sensitivity with respect to the temperature during exposure were obtained. Besides the increase of sensitivity at elevated temperature up to 90 ºC, a slight degradation of the contrast was observed with exposure temperature rise. An activation energy ~1.6 kcal/mol was obtained from the linear region of the Arrhenius plot of the sensitivity vs. temperature. It was shown that ultra high resolution structures formed at elevated temperatures exhibit improved uniformity in combination with less sensitivity to overdose. The observed effects were attributed to several mechanisms. Amplification of thermally activated processes and their influence on the the resist sensitivity and molecular weight distributions between original and irradiated resists was proposed as the main mechanism which determines contrast and consequent resist resolution with exposure temperature variation. Complementary to the main mechanism, a slight decrease of diffusion range of SE and a decrease of structure’s porosity with a rise of exposure temperature due to enhanced cross linking could take place. In Chapter 6 a method for improving the aspect ratio of ultrahigh-resolution structures in negative electron-beam resist is provided. The idea of this method is formation of a protective “cap” on top of the resist structure by means of electron-beam-induced deposition (EBID) in a self-aligned approach. This way the pattern transfer capabilities maybe enhanced. The process is implemented by a combination of electron-beam lithography and EBID during exposure of the resist material in the presence of a precursor gas. As the result, aspect ratios up to 20% and 14 % higher were obtained for Pt precursor (methylcyclopentadienyl(trimethyl)platinum) and TEOS precursor, respectively. It was shown that the absolute feature-height improvement achieved by means of EBL+EBID compared to conventional EBL is most prominent for thicker resist layers. This effect was related to an enhanced development speed in thicker layers. Tungsten precursor showed no aspect ratio improvement at all. It was demonstrated that erosion of the deposited cap material during development is a serious drawback. It was concluded that the combined EBL+EBID method requires further optimization, whereby the type and flux of precursor gas, the process temperature, and the developer conditions (strength, composition) are the most important parameters. Chapter 7 deals with an investigation of the lithographic process with a scanning sub-nanometer helium ion beam. The lithographic performance of both the positive tone poly(methyl methacrylate) (PMMA) and the negative tone HSQ resist were studied. It was demonstrated that the scanning He+ ion beam has a very high resolution (down to 6±1 nm features in HSQ). Additionally, superior low proximity effect as compared to electron beam exposure was demonstrated. It enabled the formation of extreme high-density features down to pitch of 14 nm in HSQ resist. Furthermore, He+ ion exposure was shown to be several times more effective than electron beam exposure at the same acceleration voltage, whereas the contrast was found to be almost equal. HSQ and PMMA resists exhibited respectively 4 and 17 times higher sensitivities for helium ions than for electrons at the same energy of 30 keV. Dependence of sensitivity on He+ ions energy was found to correlate with the electronic stopping power of ions in resist. Overall, He+ ion beam lithography was shown as a very promising technique for the formation of ultrahigh resolution structures of a high density and having feature sizes far into the sub-10-nm range. In chapter 8, the response of Al2O3 material to bombardment with He+ ions and consequent development was investigated and partly compared to lithographic behavior of WO3. Positive tone and negative tone resist behavior after alkaline development was observed for Al2O3 depending on the exposure dose. Observed lithographic behavior of Al2O3 was ascribed to the role of the material density, structural order and defects influencing the solubility of Al2O3 during development. Exposure to He+ ions does not reduce Al2O3 to metallic state contrary to the heavier atomic weight transition metal oxide WO3. Sub-surface bubbles were not observed after He+ ion beam exposure within investigated dose range. Finally, sub-10-nm patterning of Al2O3 was demonstrated in a negative tone exposure mode. In conclusion, the results presented in this thesis demonstrate several ways of optimizing the lithographic process to ultimate performance. Optimization of the lithographic system, exposure of resists at elevated temperatures, utilization of ultra thin layers, combination of EBL and EBID, exposure of several resist materials with helium ion beam were undertaken. Successful outcome of each step brought additional understanding of the ultra-high resolution lithographic process together with a number of useful findings, which make patterning of sub-10-nm structures at high density more feasible.Kavli Institute of NanoscienceApplied Science