Files in this item



application/pdfPANICI-THESIS-2019.pdf (4MB)
(no description provided)PDF


Title:Investigation of plasma diffusion in extreme ultraviolet lithography sources
Author(s):Panici, Gianluca
Advisor(s):Ruzic, David N.
Contributor(s):Curreli, Davide
Department / Program:Nuclear, Plasma, & Rad Engr
Discipline:Nuclear, Plasma, Radiolgc Engr
Degree Granting Institution:University of Illinois at Urbana-Champaign
Abstract:Advancement of technology in the semiconductor industry is dependent on cost-effective manufacturing of integrated circuits. Novel methods are currently being investigated for the limiting step, lithography, to cut costs and drive innovation. The current standard lithography technique utilizes a 193 nm excimer laser and complicated techniques to reduce this wavelength including multiple patterning and immersion lithography. The leading technique for next generation lithography is extreme ultraviolet (EUV) light at the 13.5 nm wavelength. Shifting to this wavelength from the current industry standard of 193 nm allows an increase in resolution without the reduction in depth of focus present in immersion lithography. It would also cut back on the number of steps needed, increasing its cost-effectiveness compared to multiple patterning. EUV light is created by a laser-produced plasma as opposed to an excimer laser. This change in photon generation brings a new problem of energetic ions and neutrals that can damage the optics in the system. Understanding the transport of this debris is fundamental to creating an effective debris mitigation technique. This work investigates the plasma transport both theoretically and experimentally. An ambipolar diffusion model was developed and used to predict the expected peak diffusion times at various pressures and distances from the source plasma. The momentum cross-sections used in this model were derived using the method proposed by Ruzic [1]. The interaction of each plasma component with the background gas is hypothesized and secondary plasma development is predicted. The photons travel out first but are not expected to contribute to any secondary plasma via photo-ionization or photo-electric effect. The electrons, both fast and bulk, are not expected to contribute to any secondary plasma in this work but may do so at industrial conditions. The ions are expected to slow significantly due to collisional processes. They also will transfer charge readily to the background, creating low energy ions and high energy, forward scattering neutrals. The predictions are then evaluated by comparing to triple langmuir probe data. The XTS 13-35 EUV source is used as the plasma source. Two different pinch gases, argon and nitrogen, are utilized to examine the trend with projectile ion mass. Pressure is ramped to examine the trend with background collisions. As expected, at higher pressures the diffusion time elongates, accounting for increased collision frequency. Increasing the pressure from 10 mtorr to 1 torr decreases the diffusion coefficient and mobilities of the plasma by two orders of magnitude, which in turn causes the diffusion time to increase by a factor of 15. A similar trend is apparent with increasing distance from the pinch, as an increased number of collisions is expected. With plasma lifetimes exceeding a millisecond at industrial pressures, it is expected that in a commercial EUV system there is a constant background of plasma from previous pulses. This constant background showcases the need for the characterization undertaken in this work.
Issue Date:2019-12-11
Rights Information:Copyright 2019 Gianluca Panici
Date Available in IDEALS:2020-03-02
Date Deposited:2019-12

This item appears in the following Collection(s)

Item Statistics