Quantifying reaction spread and x-ray exposure sensitivity in hydrogen silsesquioxane latent resist patterns with x-ray spectromicroscopy

Abstract

Direct-write soft x-ray lithography with an ϳ50 nm diameter beam is used to pattern features in hydrogen silsesquioxane ͑HSQ͒ thin films. Scanning transmission x-ray microscopy of the undeveloped patterns ͑latent patterns͒ at the oxygen K-edge reveals a two-stage cross-linking mechanism. Oxygen and silicon near edge x-ray absorption fine structure spectra of latent patterns show an increase in oxygen content and no change in silicon content within exposed regions. A dose and thickness dependent spatial spread of the cross-linking reaction beyond the exposure boundaries is observed and quantified in detail. Strong area-dependent exposure sensitivity ͑attributed to cross-linking beyond the exposed region͒ is observed in latent patterns. A lateral spread in the cross-linking of Ͼ70 nm ͑full width at half maximum͒ is observed on both sides of the lines created with 580 eV x-rays ͑ = 2.14 nm͒ in 330Ϯ 50 nm thick HSQ films at low dose ͑0.6Ϯ 0.3 MGy, 27Ϯ 12 mJ/ cm 2 ͒ ͑1 MGy= 10 6 J / kg absorbed energy͒. At a higher dose ͑111Ϯ 29 MGy, 5143Ϯ 1027 mJ/ cm 2 ͒, this spread increased to 150 nm. Preliminary results indicate that latent line widths increased with increasing delay between film spin-coating and exposure. Sharper lines are observed after room temperature development of the latent HSQ patterns in NaOH/NaCl solution ͑onset dose of 3.9Ϯ 1.0 MGy, 181Ϯ 36 mJ/ cm 2 ͒ due to the removal of material below a critical degree of cross-linking. Given the short range of low energy secondary electrons in condensed media ͑Ͻ10 nm at Յ580 eV͒, the observed spread is likely due to the propagation of reactive ions or radicals beyond the exposed regions.

Zone plate focused soft X-ray lithography

“…Surprisingly, material was removed more than 300 nm away from the expected exposure area. It is this type of exposure-spread phenomenon that has limited the minimum achievable feature sizes for direct write X-ray patterning in the past [23][24][25][26]. Understanding the origin, which in turn could lead to strategies to minimize the spread, was a major goal of this research.…”

“…The features of this exposure spreading mechanism are quantifiable -the lateral extent of the spread depends on the polymer used and dose [24,25]. Other direct write X-ray patterning studies have noted that this takes place not only in organic resists, but also inorganic resists [23,26], and self assembled monolayers [36]. The radiation based lithography literature contains numerous examples of effects which can cause the actual exposure area to be larger than intended (i.e.…”

“…Thermal treatment methods have also led to reduced minimum features sizes in PMMA [47]. Other resists such as HSQ [26] could offer higher resolution than PMMA. One can also envisage a resist and developer system designed specifically for focused soft X-ray lithography.…”

“…Larciprete et al [23] used focused 640 eV soft X-rays at ELETTRA to pattern LiF films. Wang et al [24,25] However, in all reported direct write X-ray patterning experiments to date the best minimum feature size of either undeveloped or developed patterns is only 90 ± 14 nm in hydrogen silsesquioxane (HSQ) [26], despite writing with focused X-ray sources of much finer spot size. The resolution limitation is due to an exposure spreading phenomenon whose origin has not been understood until this work.…”

Direct-write X-ray lithography using a hard X-ray Fresnel zone plate

“…Recently, there has been a considerable amount of effort devoted to developing direct-write X-ray lithography (X-ray writers) using synchrotron light sources (Caster et al, 2010;Leontowich & Hitchcock, 2011;Leontowich et al, 2013). The concept of X-ray writers, i.e.…”