SPIE Membership Get updates from SPIE Newsroom
  • Newsroom Home
  • Astronomy
  • Biomedical Optics & Medical Imaging
  • Defense & Security
  • Electronic Imaging & Signal Processing
  • Illumination & Displays
  • Lasers & Sources
  • Micro/Nano Lithography
  • Nanotechnology
  • Optical Design & Engineering
  • Optoelectronics & Communications
  • Remote Sensing
  • Sensing & Measurement
  • Solar & Alternative Energy
  • Sign up for Newsroom E-Alerts
  • Information for:
SPIE Photonics West 2019 | Call for Papers

2018 SPIE Optics + Photonics | Register Today



Print PageEmail PageView PDF

Micro/Nano Lithography

Under Pressure

Nanoimprint lithography offers an economical, high-resolution alternative to EUV and next-generation lithography.

From oemagazine August 2005
31 August 2005, SPIE Newsroom. DOI: 10.1117/2.5200508.0007

Lithography is the enabling technology for patterning various types of microdevices and structures, including integrated circuits (ICs), micro-electromechanical systems (MEMS), patterned media, and optoelectronic devices. Until now the patterning process has largely taken place through optical lithography, which involves illuminating a photomask and imaging the resultant pattern to expose a photoresist-coated wafer.

As demand for device performance and speed increases, critical linewidth dimensions drop. In IC manufacturing, this trend is described by Moore's Law, but the desire to pattern smaller and smaller critical dimensions exists for many applications. The method for achieving this shrinkage has traditionally been shortening the exposure wavelength, from g-line and i-line illumination to 248- and 193-nm laser light. As each "shrink" has occurred, the need for greater optical precision has grown. This has forced the industry to develop complex lithography tools requiring sophisticated mask technology, super-smooth substrates, and stringent process control. The cost of lithography tools and lithography development has risen dramatically from the early days. The production of extreme-UV (EUV) or next-generation technologies threatens to be more expensive and challenging still.

Press and Pattern

In nanoimprint lithography (NIL), a stamp produced by direct-write electron-beam lithography is used to transfer patterns to a polymer-coated wafer. The stamp substrate typically consists of quartz, nickel, or silicon but may consist of other materials as well. To produce a pattern on a polymer-coated wafer, we apply the stamp to the wafer with pressure and heat to reflow the polymer, leaving an exact image of the stamp (see figure 1). Alternate processes also employ UV light for curing the polymer resist. The polymer is then subject to a gentle etch to remove a miniscule residual layer that is left at the base of the printed features, transferring the pattern into the wafer.

Figure 1. In the nanoimprint lithography process (from the top) a wafer is first metallized (a), then coated with photoresist (b). The imprint stamp impresses the structure on the photoresist layer and UV exposure cures the resist (c). After demolding (d), the residual layer is etched down (e), and reactive ion etching transfers the structure into the metal layer (f).

The technology, although nascent, has already demonstrated the ability to print sub-10-nm structures. NIL also holds the potential to change the cost and complexity dynamics of lithography. Because the method uses UV energy for curing and/or crosslinking the polymer rather than for generating a high-resolution pattern on the photoresist, the control of light energy is no longer a critical concern. This eliminates the performance requirements traditionally imposed by optical lithography on the UV source and optical train.

NIL involves the production of a pattern in the polymer by mechanical deformation rather than chemical reaction. Investigation in the sub-50-nm regime indicates that imprint lithography resolution is limited only by the resolution of the electron beam lithography. Because of the intimate contact between stamp and coated wafer, a release agent must be applied for easy separation of stamp and substrate. These release coatings are typically applied to the stamp after it is patterned.

A clean environment is essential for successful NIL patterning. Particulates potentially contaminate the stamp as well as the substrate; in many cases, however, the particles will stay with the coated substrate rather than with the stamp, which limits the passing-on of printing defects.

Quartz and silicon are the most-used stamp substrates for research applications, but nickel provides better reliability in industrial applications because it has long lifetime and can be used to produce up to 25 inexpensive replicated nickel stamps. It is possible to electrochemically grow a metal sheet on, and produce copies in, nickel metal (a "master" stamp). The master stamp is expensive and can take days to produce, so it is necessary to electrochemically replicate several stamp copies from one master stamp, a process that requires one to two hours. In actual NIL, the stamp copy can impress substrates with its pattern in just 3 to 30 s, depending on polymers.

Several existing imprint lithography methods produce patterns with nanometer-level resolution, such as thermal imprint (Nanonex Inc., Princeton, NJ) and step-and- flash imprint (Molecular Imprints Inc. (MII); Austin, TX; see oemagazine, August 2003, p. 18). Combining the two approaches, as in our simultaneous thermal and UV-imprint NIL, gives better control for downstream processes over a large area. Thermal imprinting forms the pattern in the resist and subsequent UV exposure crosslinks the resist at a constant temperature. The method is fast because it prints over an area as large as a 150-mm wafer all at once, and no temperature ramp is necessary. It also maintains control of the thickness and uniformity of the residual layer (the etch barrier) to as little as 10 nm over the entire wafer.

Full-field NIL printing is still in its early development. The full-field tools available today are characterized by manual-handling and operator-attended alignment systems. Suitable applications are characterized by single-critical-level printing and less aggressive alignment requirements.


With the general trend toward miniaturization of devices, micro-optical elements have become important. So why are micro- and nano-optics becoming available for optical system applications just now? The primary reason has been manufacturability. High-performance and low-cost requirements have traditionally contradicted one another. NIL provides a batch-processing method for optical components with demanding parameters, as with gratings of sub-100-nm dimensions. LG-Electronics Elite (Seoul, Korea) has recently used NIL to produce an optical grating with a 50-nm half pitch, for example. The company plans to use the device in a projection television.

Another application is optical data storage. High-density recording beyond the optical diffraction limit has attracted attention beyond the applications of holographic memory and 3-D optical recording. Researchers have looked to near-field optics to solve this challenge. Today, the minimum structure size of the Blu-Ray disk is about 150 nm, and in the next-generation high-density optical disks, this value will become significantly smaller. Tests are already being conducted with structures as small as 30 nm.

One solution is the super-resolution near-field structure (super-RENS), which can read/write nanometer-scale features. The super-RENS is considered a more feasible way of achieving 60-nm recording, offering a simple recording head design and higher recording speed than today's laser read/write heads. The super-RENS technique is not limited by the wavelength of the light and makes it possible to focus a laser beam in a much smaller spot than with a conventional lens system. It should be noted that all optical disks need a predefined structure, a pattern where it defines the track and pits for data read/write. Nano-imprint lithography is the most promising candidate for production of both super-RENS lens systems and their nano-optical disks.

Figure 2. An SEM image of a master disk for super-RENS applications shows the profound regularity of the data pits. (Image courtesy OBDUCAT AB)

Replication of a nano-optical disk requires fabricating a mold. It has been shown that when the size of the data pits is smaller than 150 nm, the shape of the pits plays an important role in disk player performance. Making the width of the structure greater than the length, for example, improves read/write repeatability and signal-to-noise ratio. An electron beam is the only technique available to make such features on a master disk. For replication, though, NIL provides a solution (see figure 2). We coat a glass substrate with negative-tone photosensitive polymer and using NIL in a simultaneous thermal and UV process, replicate the mold structure in the polymer. The structured polymer is then made into a metal that reflects light from the read head. The polymer remains part of the final disk and no further processing is needed. This is a simple, cost-effective way to produce the next generation of 1 TB optical data storage media.

Infrastructure Process

A challenge to any new technology is the rate and pace of the development infrastructure required for its commercial success. In the case of NIL, the two critical infrastructure pieces are the stamps and the chemistry. Stamps are fundamentally different from photomasks - NIL systems do not involve reduction optics, so the stamp must consist of a 1:1 image of the final device. NIL has demonstrated that any image produced on the stamp at a sub-10-nm level can be replicated through NIL patterning. A number of vendors, including MII, Nanonex, Obducat, Photronics Inc. (Austin, TX), and Quantiscript Inc. (Sherbrooke, QC, Canada) offer stamp-producing services.

NIL chemistry is based on pattern formation through heat and pressure. Imaging is done through the physical stamping of the pattern - a drastic change from the optical stencil and lens system required in optical lithography. MII and Brewer Science (Rolla, MO) have announced an agreement in which Brewer Science produces the MII resist formulations. Nanonex and Micro Resist Technology GmbH (Berlin, Germany) also offer resist variations. Many university research groups are patterning into various mixtures of polymethyl methacrylate, and industrial companies (especially those in the disk media) are developing proprietary formulas.

NIL's promise of simpler tool sets and effective sub-100-nm patterning is being realized today. Independent from the limiting effects of light energy (diffraction, etc.), the ability to replicate sub-100-nm features with a low-cost tool set opens new device-making opportunities, not only in storage media, but in MEMs, biodevices, optical components, and semiconductors. oe

Babak Heidari, Ken Mason

Babak Heidari is CTO of Obducat, Malmo, Sweden; and Ken Mason is manager of Obducat North America, Hampstead, NH.