Lithography

A typical integrated circuit consists of various patterned thin films of metals, dielectrics and semiconductors on various substrates such as silicon, gallium arsenide, or germanium. Such a device is fabricated by a technology known as lithography, in which radiation sensitive polymeric materials called resists are used to produce circuit patterns in the substrates. Figure 1 depicts the lithographic process sequence.

Figure 1: Schematic representation of the lithographic process.

The resist material is applied as a thin coating, typically by spin coating over the substrate (wafer) and then heated to remove the casting solvent (post-apply bake, pre-exposure bake, or pre-bake). The resist film is subsequently exposed in an image-wise fashion through a mask (in photo- and X-ray lithography) or directly with finely focused electron beams. The exposed resist film is then developed typically by immersion in a developer solvent to generate three-dimensional relief images. The exposure may render the resist film more soluble in the developer, thereby producing a positive-tone image of the mask. Conversely, it may become less soluble upon exposure, resulting in generation of a negative-tone image. When the resist image is transferred into the substrate by etching and related processes, the resist film that remains after the development functions as a protective mask. The resist film must “resist” the etchant and protect the underlying substrate while the bared areas are being etched. The remaining resist film is finally stripped, leaving an image of the desired circuit in the substrate. The process is repeated many times to fabricate complex semiconductor devices.

For a resist material to be useful in device fabrication:

• it must be capable of spin casting from solution into a thin and uniform film that adheres to various substrates such as metals, semiconductors, and insulators
• possess high radiation sensitivity
• possess high resolution capability, dictated by solubility/insolubility characteristics
• withstand extremely harsh environments, for example, high temperature, strong corrosive acids, and plasmas such as used in subsequent etching, doping and sputtering operations.

In earlier generations of resists (often based on novolok phenolformaldehyde polymers), each absorbed quantum of radiation induced on average less than one chemical reaction within the material. “Photosolubilizable compositions” for photographic and photoresist applications with overall quantum yields greater than one were described by Smith as early as 1973. The idea was limited in practice, however, until thermostable yet photosensitive onium salts of strong acids were developed by Crivello as PAGs for cationic photopolymerization. The first CA resist for microlithography was reported by the university industry team of Fréchet (University of Ottawa) and Willson and Ito (IBM Inc.), who combined onium salts with the acid-deprotectable poly(4-[t-butyloxycarbonyloxy]styrene) (poly-TBOCST), and coined the term chemical amplification. Over 200 articles and several reviews have appeared on this topic between January 1992 and June 1994 alone.

In chemical amplification (CA) resists, the primary photochemical event produces a mobile catalyst that, typically during later postexposure baking (PEB), goes on to induce a cascade of material transforming secondary catalytic events within a 5-25 nm radius. Such chemical amplification thus makes possible an overall quantum yield (the number of material reactions divided by number of absorbed photons) of up to several hundred. Thus, a CA resist must contain:

1. a small amount (ca. 1-5 wt%) of radiation-sensitive catalyst precursor, generally a photoacid generator (PAG);
2. many chemical groups that can react by elimination, addition, or rearrangement only in the presence of catalyst;
3. a polymer matrix able to disperse all other components in a smooth clear film; and
4. optional additives to improve performance or processability; e.g., surfactants, photosensitizers, and etch resistors.

Although in most reported compositions the catalyst-sensitive groups are polymer-bound and the catalyst precursors are free (i.e., i + ii~iii), components i-iii or iv can in principle be interconnected in any combination, as small molecules, homopolymers, copolymers or blends (e.g., i + ii + iii, i~ii + iii, i~ii~iii). Along with their higher sensitivity and contrast in forming images, CA are also better than earlier resists in being more flexible in design and formulation, versatile in radiation source (electromagnetic or particle beams), and compatible with dry (plasma), multilayer, and other advanced pattern transfer techniques. In general, resist systems can be classified on the basis of their design, namely, one-, two-, or multi-component systems. One-component resist systems consist of pure polymers that must combine all the necessary attributes such as substrate protection, radiation sensitivity, and film-forming characteristics. The most popular resist designs in modern lithography are based on two-component systems in which resist functions are provided by two separate components.

Resist systems can also be divided into three groups on the basis of the radiation source: UV or photoresists, electron-beam resists, and X-ray resists. Photolithography that utilizes UV light has been the predominant technology in semiconductor manufacture and will continue to be so in the foreseeable future. X-ray lithography is capable of producing high-resolution, high-aspect-ratio (height-width) images and is considered to be the technology of the future, whereas elctron-beam lithography is used in photomask fabrication. Photolithography can be further subdivided into near-UV (350-450 nm), mid-UV (300-350 nm), and deep-UV (< 300 nm) technologies, depending on the wavelength of the exposure. The resolution is proportional to the exposing wavelength and inversely proportional to the numerical aperture (NA) of the lens. Thus the i-line (365 nm) with a high-NA lens shifting from the g-line (436 nm), has been regarded as the dominant technology in the manufacture of 16-megabit (Mbit) dynamic random access memory (DRAM) devices with a minimum feature size of 0.5 mm. Krypton fluoride (KrF, 248 nm), argon fluoride (ArF, 193 nm), and F2 (157nm) excimer laser technologies are emerging as the minimum feature size continues to shrink far below 0.5 mm down to 80 nm.

The breakthrough that ultimately led to the adoption of 248 nm lithography as the technology of choice for advanced device fabrication was the development of CA resists. The consensus candidate for the next generation of optical lithography tools has been photolithograpy using 193 nm light. At this wavelength, the opacity of traditional aromatic-based materials precludes their use. Alternate resist materials based on aliphatic polymers and dissolution inhibitors have been examined. The introduction of monomers bearing cyclic side groups such as adamantyl or norbornyl significantly improves the etch resistance over that of simple acrylic polymers. Alternately alicyclic structures have been incorporated directly into the polymer backbone offering a second route to 193 nm single layer resists.

The next generation of optical lithography at 157 nm has necessitated new developments in resist technology. Due to their high optical absorbance at 157 nm, single-layer 248 nm and 193 nm photoresists are not a viable alternative. The incorporation of fluorine into polymers have led the way to suitable resist materials.

One of the ultimate tasks in current nanolithography is the ability to fabricate arrays of structures with controlled size and shape, precisely positioned on a suitable surface. The race for shrinking the feature size pushes the limits of conventional lithography requiring fast, low cost, reliable and well-controlled processes of which stencilling is a promising one. The most common applications of stenciling have been patterning of printed circuit board features and interconnect technologies. Potential applications of the technique include (1) deposition on non-conventional and unstable substrate materials (i.e. bio-chemical, hydrophobic), (2) deposition of heterostructures (epitaxial, magnetic, complex oxides, piezoelectric materials) and (3) deposition of nanodevices on CMOS. Identifying, predicting and overcoming issues accompanying the nanostencil lithography is critical to the successful and timely development of this technique with wide range of potential applications.

The deposition through stencils suffers, however, from drawbacks such as the clogging of apertures and the membrane deformation due to the deposition induced stresses. Clogging occurs when the evaporated material accumulates on the membrane itself and inside the apertures. This phenomenon changes the shape of the aperture during the deposition process and leads to a distortion of the deposited pattern, eventually resulting in the complete closure of the aperture. Accumulation of material on the membrane and the subsequent deformation due to the deposition induced stress results in an increased substrate-stencil gap which is vital for pattern definition.

One of the modelling tasks within the NaPa project is to reduce the undesirable effects of these phenomena to the minimum by optimising the geometry of the stencils. It was proposed to stabilise the stencils mechanically by introducing corrugation structures/rims. Tyndall researchers are working on predicting the deposition stress induced deformation of stencils and subsequently establishing optimal corrugation geometries for various stencil designs defined by partners from EPFL, IBM and CNM. The stencils are intended for deposition of nanoresonators, for evaporating nanowires and for fabrication of three levels of interconnections to realise electrical measurement on a single molecule or nanowire. Modelling the clogging effect and subsequent prediction of stencil lifetime is also within our NaPa tasks.

Above Left: Simulated deformation of fragment of nanoresonator stencil.
Above Right: Simulated deformation of the same stencil incorporating corrugation rims. A 67% reduction in deformation is achieved.

Without a clean interactive “tool” for control of global temperatures, we are relying upon the unstable nature of our environment to continue to mend itself, regardless of our abusive actions. In addition to our own actions, history has shown that nature often has global extinction events. There are many natural generators of greenhouse gases.

If the methane permafrost melts, billions of people may die from reduced global agricultural harv Continue reading »

Rotating magnetic field as a sum of magnetic vectors from 3 phase coils.

An electric motor converts electrical energy into kinetic energy. The reverse task, that of converting kinetic energy into electrical energy, is accomplished by a generator or dynamo. In many cases the two devices differ only in their application and minor construction Continue reading »

The next 10 years promises to be an exciting period in the history of computers and networks as nanotechnology takes off to redefine a new level in the way computers are manufactured. It’s not entirely radical as the Lithographic principles behind the manufacturing process can be adopted for nanotech processes. What is revolutionary are the minute molecular-level sizes at which those circuit boards can now be made. This is the core of nanotech Continue reading »