Lithography Basics: Extreme Ultraviolet Lithography (EUVL) Technology

In previous issues of the Nikon Review, we discussed different ways to expand the resolution capability of Deep Ultraviolet (DUV) scanners, including phase-shift masks, off-axis illumination, custom illuminator patterns, polarized illumination, and double patterning. All these techniques can be referred to as "k₁ enhancement," where k₁ refers to the coefficient in the equation describing resolution:

The Rayleigh Criterion:

Rayleigh Equation

Of course, Rayleigh’s equation also shows that resolution can be improved by reducing illumination wavelength (λ) or increasing lens NA. However, further progress is limited as essentially no glass is available for use at shorter wavelengths, and in the case of lens NA, we are reaching the fundamental limits of water immersion with no promising alternative materials on the horizon. In addition, improving k₁ beyond its current state of about 0.3 will be extremely difficult (Figure 1). As we reach the fundamental limits of DUV lithography, it is necessary to transition to Next-Generation Lithography or NGL. Currently, the most promising NGL candidate is Extreme Ultraviolet (EUV) lithography. In this article we’ll look at the advantages and challenges associated with this solution.

k₁ Reduction Over Time

Figure 1. Improvement in k1 over time.

Figure 1. Improvement in k₁ over time.

The EUV light source has a wavelength fourteen times smaller than that of ArF DUV tools. This has such a profound effect on resolution that the lens NA can be much smaller than with DUV, and efforts to improve k₁ are practically unnecessary. Figure 2 shows the simulated aerial image of a small 45 nm line-end feature. The top graphic shows that the aerial image from the 1.30 NA ArF immersion tool is barely there, and would require a lot of k₁ enhancing "tricks" to get a well printed image from it. In the case of the EUV tool, even with a much smaller NA (0.25), printing the image is simple. The EUV tool's inherent capacity is much better from the start.

Line-end Feature Aerial Image Comparison

Figure 2. DUV and EUV aerial images for the same 45 nm feature.

With EUV hardware, we are essentially forming images with X-rays, and this has several important ramifications:

DUV uses a laser. EUV uses an incoherent “light source,” powered by an electrical discharge (DPP) or laser radiation (LPP), radiating at 13.5 nm.
DUV scanners use either all-glass optics or mostly glass with a few mirrors. EUV optics are composed entirely of mirrors requiring highly specialized, multilayer coatings.
DUV projection lenses have an NA from 0.92 to 1.30. EUV projection optics have an NA of 0.25 or so.
DUV scanners operate in the air using sealed optics purged with nitrogen gas. EUV scanners are completely encased in a high-vacuum system, meaning that wafers and reticles must pass in and out of airlocks.

A "generic" EUV tool is shown in Figure 3. The basic components are similar to those used on DUV systems, but modified for the all-reflecting in-vacuum design.

Generic EUV System Design

Figure 3. A generic EUV system.

An overview of the Nikon EUV1 system is shown in Figure 4. Multiple EUV1 systems with the following basic characteristics have been manufactured:

A Xe-discharge EUV source
An illumination system that allows users to set the partial coherence in the reticle illumination
A six mirror ring-field projection optic, with 0.25 lens NA
Aberration levels of less than 1 nm RMS and flare of 10% or less
Full field imaging with a scanned slit yielding 26 X 33 mm dies on the wafer

Modules of the Nikon EUV1 System

Figure 4. The Nikon EUV1 system.

Nikon predicts that the EUV1 system will be used in process development for features of 32 nm and less, and that it will show an eminently usable depth of focus and process window for those small features. Over the next few years, it is expected that DUV will remain the primary tool for leading-edge lithography. To meet this need, Nikon is continuing to enable chipmakers to stretch the k₁ factor with immersion ArF systems. Looking forward, EUVL is a likely solution for the 22 nm node and beyond, and aggressive development efforts on EUV1 and next-generation EUV systems are underway (Figure 5).

Nikon EUVL Development Roadmap

Figure 5. The S610C already meets the aggressive CDU requirements for double patterning applications.

Figure 5. The Nikon EUVL development roadmap.

Back to Front Page