Double-sided structured mask for sub-micron resolution proximity i-line mask-aligner lithography

“Is it possible to use a standard mask aligner in proximity mode to manufacture gratings with sub-200nm (half-pitch) resolution?” If you pose this question to any lithography expert, the answer will be simple and short: “Not possible! You need very expensive e-beam or stepper lithography”. Not a surprise. The resolution of a mask aligner in proximity mode is typically some 3μm to 5μm. Working in contact mode with a mask aligner you can achieve below 1μm resolution – but sub-200nm? An alternative approach might be Talbot lithography at 193nm (Excimer laser). But, for sub-200nm (half-pitch) gratings, the depth-of-focus will be tight. This will make it difficult to print on larger substrates or on wafers. The only possible solution for mask aligner lithography might be nano-imprint… which is not well suited for larger series of grating manufacturing.

Researchers of Uwe Zeitner’s group at Fraunhofer IOF/University of Jena, Germany, now found a very clever approach to overcome these limitations of mask aligner lithography. They demonstrated proximity lithography at 10μm gap with an extended depth-of-focus for 350nm period gratings in a standard mask aligner at i-line (365nm) wavelength.

The key idea behind this approach is to use a double-sided photomask, a combination of a wire-grid polarizer and a phase grating providing 0th order cancellation. In their approach, they placed a wire-grid polarizer (WGP) on the top side of the mask to linearly polarize the light from the mask aligner. The wire-grid polarizer was based on 22nm Iridium ridges and 100nm pitch manufactured by e-beam writing and double-patterning. In a next step, a binary phase grating was manufactured on the bottom side. For this phase grating, the IOF researchers used a clever trick proposed by Olivier Parriaux’s group of the University of Saint-Etienne, France, some years ago. For some special combinations of grating ridge width and grating depth, it is possible to cancel out the 0th order of a phase grating. Thus, only the ±1st diffraction orders propagate behind the grating.

The IOF researchers used a grating with 700nm period, 210nm ridge width and 430nm grating depth. From Rigorous Coupled Wave Analysis (RCWA), about 43% of the light should be diffracted in each of the ±1st diffraction orders and only 0.004% should remain in the 0th order. In practice, they still observed some 1% of the light in the 0th order. Not perfect, but good enough to suppress the 0th order significantly. A SUSS MicroTec MA8Gen3 mask aligner, equipped with a bandpath filter (i-line, 365nm) and MO Exposure Optics (6mm wide slit aperture) for illumination with ±2.9° angular spectrum, was used for the experimental work.

As shown in Fig. 1 of this article, the diffracted light propagating behind the phase mask generates a 350nm periodic intensity distribution with almost infinite extension in the direction of light propagation. Printing 350nm periodic gratings at a proximity gap of 10μm or more in a standard mask aligner is well possible. Sure, it is not trivial to manufacture such double-sided photomasks. Especially the 22nm features of the Iridium wire-grid polarizer and the high aspect ratio of the phase mask must have been a challenge. However, the resulting phase mask would allow mass production of sub-200nm (half-pitch) gratings at very low costs. The even more exciting idea is to change the polarization locally by using different wire-grid polarizer configurations on the top side of the mask. I am sure that we will read more about double-sided mask lithography in mask aligners very soon.

---

---

 Add Comment

You must log in to add comments.