James P. DeYoung

Synthesis of Fluoropolymers in Liquid and Supercritical Carbon Dioxide Solvent Systems

Unlike a large portion of the many billions of pounds of organic and halogenated solvents used in industry every year, CO2 is inexpensive, of low toxicity, and environmentally and chemically benign. The prime factors inhibiting the widespread use of this attractive solvent replacement have been the disappointingly low solubility of most materials in CO2 in both liquid and supercritical states, and a less than complete understanding of its physical properties. Discoveries made over the past decade identifying the solubility of siloxanes and many amorphous fluoropolymers in CO2 have facilitated a greater understanding of the solvency properties of this uniquely “tunable” solvent. The subsequent development of amphiphilic surfactants incorporating lyophilic fluorocarbon or siloxane segments with lipophilic or hydrophilic moieties opens the door for the synthesis of a broad range of hydrocarbons. Carbon dioxide technology is of particular significance with respect to the manufacture of many fluoropolymers that require nonaqueous synthesis. Reduction of unwanted side reactions such as β-scission and opportunities for unique separation and purification processes demonstrate that CO2 is not only an adequate replacement for chlorofluorocarbons, but in many cases a superior one.

Mechanism of Extreme Ultraviolet Photoresist Development with a Supercritical CO2 Compatible Salt

The mechanism of developing an extreme ultraviolet (EUV) commercial photoresist with supercritical carbon dioxide (scCO2) and a CO2 compatible salt (CCS) solution was studied. The cloud point of CCS in CO2 and the pressure at which the photoresist dissolves in CCS/scCO2 were determined for temperatures between 35 and 50 degrees C. For this temperature range, it was found that the CCS cloud point ranges between 11.2 and 16.1 MPa, while the photoresist dissolution point ranges from 15.5 to 21.3 MPa. The kinetics of the CCS/scCO2 development was modeled using a simplified rate equation, where the rate-limiting steps were photoresist dissolution and mass transfer. The effects of temperature, mass transfer, pressure, and CCS concentration on photoresist removal rate were further explored experimentally using a high-pressure quartz crystal microbalance (QCM). Increasing temperature (35-50 degrees C) at a constant fluid density of 0.896 g/mL was found to increase the removal rate following an Arrhenius behavior with a photoresist dissolution energy of activation, Ea, equal to 79.0 kJ/mol. The removal was zero order in CCS concentration, signifying photoresist phase transfer, photoresist mass transfer, or both were rate limiting. Mass transfer studies showed that circulation enhanced the photoresist removal rate, but that the mass transfer coefficient was independent of temperature from 35 degrees C to 50 degrees C. In pressure studies, increasing pressure (27.6-34.5 MPa) at a constant temperature of 40 degrees C increased the removal rate by enhancing the fluid density, but at 50 degrees C increasing pressure had little effect on the removal rate. When the total CCS concentration was in large global excess over the number of Bronsted acid groups in the polymer (2400:1 at 5 mM CCS concentration), the mass of photoresist removed varied linearly with time. At lower CCS concentrations but still in global excess of the number of Bronsted acid groups, the photoresist removal slowed (0.5 mm CCS, approximately 240:1) or was prevented (0.03 Mm CCS, approximately 15:1) due to partitioning of the CCS between the CO(2)-rich phase and the film. The CCS partitioning into the resist was found to decrease with increasing temperature, revealing an enthalpy-driven CCS absorption.

Line-width roughness analysis of EUV resists after development in homogenous CO 2 solutions using CO 2 compatible salts (CCS) by a three-parameter model

Line Width Roughness (LWR) of resists constitutes one of the main obstacles in the race of further shrinking the feature dimensions of fabricated devices. Thus, the reduction and control of LWR is one of the biggest challenges of next generation lithographies. In this paper, the LWR output of a new development process of EUV resists which uses homogeneous carbon dioxide (CO2) solutions containing CO2 compatible salts (CCS) has been examined. The measurement and characterization of LWR has been made through the analysis of CD-SEM images and the application of a three-parameter model. The three parameters involved in this model (sigma value σLWR, correlation length ξ, roughness exponent α) determine both the spatial aspects (spectrum) of LWR as well as the interplay between LWR and local CD variations. It is found that wafers developed with CCS process gives substantially lower LWR parameters (σLWR,ξ) than comparable TMAH developed samples. Also, the impact of the preparation of resist wafer (exposure time, PAG and quencher level) and the development conditions (temperature, CCS concentration) on LWR parameters is examined so that we are able to identify trends to lead toward optimized LWR performance.

Development of EUV resists in homogenous CO 2 solutions using CO 2 compatible salts (CCS): a kinetic view of dissolution of conventional resists in supercritical CO 2

A novel method for the development of standard EUV photoresists in CO2 using CO2 compatible salts (CCS) is described and examined using a quartz crystal microbalance (QCM) technique in CO2. The fundamental steps of this development process are proposed to be 1) photoresist modification via CCS interaction with Bronsted acid groups in the resist; and 2) dissolution of the modified resist into CO2. Removal rates of non-exposed photoresist from the quartz crystal can be studied in real time under a variety of conditions to help elucidate the kinetics and mechanism of the CCS development process. A series of runs from 35 to 50 C at the same CO2 density showed a strong effect of temperature on the development rate. In the presence of a large excess of CCS, the kinetics of dissolution appear to be zero order. An Arrhenius plot generated by this data gave an activation energy of 79.0 kJ/mol. The zero order kinetics of dissolution was confirmed with a series of runs at various CCS concentration that showed insignificant rate effects. Dynamic flow of the CCS solution in the QCM cell gave ~50% increase in rate of resist removal relative to the standard conditions using static CCS solution. The indication from the three sets of QCM experiments described here is that dissolution of the CCS modified polymer is the slow step in CCS development.

Development of EUV resists in supercritical CO2 solutions using CO2 compatible salts (CCS) : Results from a two level full factorial design of experiments (DOE)

Direct development of EUV resists in homogeneous carbon dioxide (CO2) solutions containing CO2 compatible salts (CCS) has been demonstrated. These CCS complexes have been designed and prepared such that the anion and/or the cation of the salt contains at least one CO2-soluble portion. In the described method, standard positive tone EUV resists are processed in supercritical CO2 containing less than 20 mM CCS, at pressures ranging from 3500 to 5500 psi, 35 to 65 C with cycle times as short as 1 minute to give reverse image development. Substantial reduction in image collapse and LER/LWR has been observed; large aspect ratios approaching 10 have been measured in dense line/space features, and dense lines with 3 sigma LWR values that are 30% smaller than comparable TMAH developed samples have been observed. This paper will describe results from a two level full factorial design of experiments (DOE) using four factors. Two experimental samples were processed at each point, along with four center point runs to give a total of 36 experimental samples. An additional seven supplemental runs were also prepared. Responses measured for the DOE include half height resist thickness, resist height, contrast, iso/dense bias, aspect ratio of collapse, line edge roughness, and line profile. Details of how the responses were measured will be presented in the paper. Effects of individual factors on each response, along with the significant interactions between factors will be reported. An optimized parameter space will be selected and additional follow-up experiments will be described.