Graduate Thesis Or Dissertation


Resolution Enhancement In Photolithography Via STED Inspired Process and Novel Mask Development Public Deposited

Downloadable Content

Download PDF
  • Photolithography is a process that transfers a two dimensional pattern onto the surface of a substrate via optical exposure, and is a key enabling technology for IC fabrication. Being an optical patterning process, the resolution is diffraction limited. A constant push for decreasing transistor size requires corresponding improvements in resolution. Typically, resolution improvements have come from reduced exposure wavelength, improved optics, improved photoresist chemistry, multiple patterning, and resolution enhancement techniques.

    This thesis develops a resolution enhancement technique, originally inspired by a technique used for super-resolution microscopy, capable of extending the critical dimension resolution of lithographic tools beyond the diffraction limit. Due to the non-linear response of positive tone photoresist to optical dose, it is possible to super-localize features, much like in the case of STED microscopy. Unlike other lithographic super-resolution techniques, this is accomplished with a single wavelength, over a broad area in the far-field, with i-line (near UV) illumination, and is compatible with commercial i-line steppers.

    By saturating the response of photoresist, features can be localized to the center of deep intensity nulls. These nulls are produced via optical interference. Models are developed to explore the optical requirements and limits of the resolution enhancement. Resist processing is optimized and these models are validated using laser interference lithography. While periodic patterns can be produced through traditional multi-beam interference, arbitrary two dimensional patterns cannot. To enable arbitrary patterning, novel lithographic masks, pixelated polarization phase shifting masks (P3SM), with both local polarization and phase control are developed. Simulations are written to explore the mask design space, and P3SMs are designed and fabricated. These masks iv are demonstrated by projection lithography using a lab-built i-line stepper.

    Modeling shows that the ultimate limit to sub-diffraction feature size is set by both optical contrast and resist contrast. Additionally, the scaling of size with dose is contrast dependent. When optical contrast is perfect, size scales as dose1/2, as is the case for STED. Experiments validated these limits and scaling rates, and demonstrated features as small as 50 nm written at i-line. Deterministic linewidth control is demonstrated, as is low line edge roughness. Phase masks that incorporate polarization control solve key issues in traditional phase masks. Not only is interference contrast maximized in multiple simultaneous orientations, but also the added degree of freedom enables line end termination and resolution of phase conflicts. A strategy for polarization mask design is proposed based on a solution to the Laplace equation. Initial mask experiments demonstrate the line end termination enabled by polarization control.

    These dimensions are similar to what may be achieved using scanning near-field, DUV, or e-beam lithography, yet achieved with far-field near UV exposures over a large area. Deterministic linewidth control and low LER make this process viable for fabrication at length scales well below those typically achieved with i-line tools. Finally, the addition of pixelated polarization control to photomasks solves the key issues preventing broader application of phase masks in lithography.

Academic Affiliation
Degree Grantor
Commencement Year
Last Modified
  • 2021-02-24
Resource Type
Rights Statement