Nanowaves

The process of transferring a pattern to the surface of a wafer is commonly known as photolithography. It was first used in the printing industry, but since the 1950s it has become the dominant pattern delineation process in microelectronics. Over the last several decades, one of the key enabling and driving forces behind the semiconductor industry has been photolithographic technology. In fact, when measured as cost per function, roughly half of the improvements in integrated circuit (IC) technology can be attributed to improvements in photolithography.

Photolithography can be viewed as a three step process. First, a photosensitive material (also called photoresist) is coated onto the surface of a wafer. This layer is then exposed to ultraviolet (UV) light through a patterned mask consisting of clear and dark areas. The UV light interacts with the photoresist rendering the exposed (or unexposed) areas soluble in a developing solution. After immersion or spraying the surface of the wafer with this developing solution, the desired pattern appears on the wafer surface. Etching is then used to remove the material that is no longer protected by photoresist after development. Hence the resist pattern is transferred to the underlying layers. There are two classes of etching: wet etching in a chemical solution, and dry etching in a plasma or vapor phase etchant. Wet etching is generally isotropic, which means the material is etched in all directions at the same rate. Anisotropy can be achieved with dry etching thanks to energetic ion bombardment in a plasma. Plasma etching is used in the IC industry to produce small features where the exact dimensions are critical. However this technology involves much more expensive and complicated hardware than wet etching.

Moore's law states that the number of devices on a chip doubles every 18 months. Hence the requirements of fine line photolithography grow more demanding every day. Many technologies have been proposed and developed such as x-ray and electron beam lithography to improve the performance of photolithography limited by diffraction, but so far none have succeeded in replacing the UV technique, due to such limitations as the complexity of mask fabrication and low throughput. It has always been more cost effective to push for an incremental improvement of photolithography technology than move toward adopting an entirely new technology. One such example of this evolution is the use of immersion lithography, which many manufacturers have now embraced. Here, instead of the air gap, light passes through a liquid medium that has a refractive index greater than one. Consequently the numeric aperture increases and the resolution is enhanced. The depth of focus is also increased by at least a factor of the refractive index. For water, the resolution improvement can be over 40 % compared with the “dry” technique.

Dense, fine resist patterns often collapse as they come into contact with each other at their tips. Resist collapse depends on the aspect ratio (height:width) of the resist pattern, and the maximum height of a resist pattern is lower for fine structures. SU-8, an epoxy based photoresist, is the most commonly used thick resist for electroforming masks, but removing the cross-linked SU-8 is challenging. KMPR complements SU-8 for electroforming molds. The major advantage of a KPMR negative resist compared to SU-8 is that it can be removed using commercially available chemical removers. Structures with aspect ratios of 18:1 can be made in a repeatable fashion using KMPR resists. It demonstrates superb adhesion as well as good mechanical properties, vertical sidewalls, and an excellent dry etching resistance.

This month's cover image is a scanning electron microscope (SEM) image of KMPR negative photoresist wires on a cleaved silicon substrate. The wires are approximately 350 nm wide and 1 μm thick. Here, KMPR was used as a mask for plasma etching. These wires, which were supposed to be straight, were deformed after development in an unexpected but interesting manner as quasi-periodic undulations formed. This wiggling phenomenon was only observed for high aspect ratio wires but was quite repeatable from sample to sample. Similar undulations have been observed for narrow porous dielectric wires using a TiN mask. Darnonet al. explained that the release of residual stresses of TiN deformed the dielectric material1. Although these distortions are very esthetic, they could be catastrophic for the fabrication of well-defined micro- and nanostructures. In the future it would be interesting to analyze the mechanical stresses responsible for this effect, and to compare with other photoresists at the same dimensions.

The unique CrossBeam® (FIB-SEM) Workstations from Carl Zeiss combine the unequalled imaging power of the proprietary GEMINI® field emission technology with a sophisticated, high performance focused ion beam column into one extraordinarily powerful system. The superior eucentric stage, together with a sophisticated, compact multi-channel gas injection system makes the CrossBeam® Workstation the ultimate analysis, inspection, and modification tool.

For more information about Carl Zeiss NTS and our comprehensive particle-beam product portfolio including SEM, TEM, FIB-SEM and the unique Helium-Ion microscopes, go to www.zeiss.com/nts.

Further Reading
[1] Darnon et al. Appl Phys Lett, 91 (2007), p. 194103