Photomask

Photomask (English reticle) are projection templates whose main application is the photolithographic patterning in the manufacture of microelectronic circuits and microsystems. They usually consist of high-purity quartz glass or calcium fluoride ( lithography with laser light of wavelength 248 nm and 193 nm) and are for example provided on one side with a thin structured chromium layer.

• 3.1 EUV mask

Background and application

Photomasks ( resist german) in the photolithographic patterning of the photoresist used. Simplified one can describe this process as follows. The mask is irradiated with light. The transparent and opaque regions of the mask form a shadow on the resist layer and light in the exposed areas brings about a chemical reaction in the photoresist. After further steps (see Photolithography ( Semiconductor Technology ) ) arises as a structured photoresist layer, that is, on the wafer are areas where the photoresist layer is still present or has been removed. This layer will be used in subsequent manufacturing steps to deposit patterned layers of other materials or to be generated by etching.

Photomasks must be completely error- free, because an error would be reflected in the exposure in each chip or the (single chip masks) on each wafer. Therefore the highest requirements in terms of transmission, planarity, material purity and temperature stability are placed on the material. This requirement in conjunction with the required precision ( widths and positional accuracy of a few nanometers ) require extremely complex and expensive production (laser or electron beam writer ) in which even variations of the geomagnetic field must be balanced to produce flawless masks. Because of these high demands, a photomask costs depending on the specification up to 250,000 EUR. Photomasks are generally the most expensive "materials", which is required for the production of integrated circuits. While cost of raw wafers in the range of a few hundred or thousand euros, a complete mask set can ( about 20 to 60 pieces are required for the various process steps ) cost several million euros.

One of intra-industry survey, in 2011 over 80 % photomasks were fabricated on 6-inch substrates. In addition, 5 -inch substrates (11 %) were also prevalent. The major portion (over 50 %) of photomasks were still made ​​for products in the 250 nm technology node and beyond; These are usually masks for installations that use a radiation source, the i- line of a mercury vapor lamp or KrF excimer laser. Photomasks for modern integrated circuits ( below the 130 - nm technology node ), however, accounted for only around 20%. In this area, the proportion of masks by a good 50 % increase for the 65 nm technology node and below, so for today's critical levels of the top products of Globalfoundries, Intel, Qualcomm, Renesas, Samsung, TSMC, etc., to 9.2%. It should be noted here is that logic products, although still the largest share of manufactured masks form (56%), masks for memory and microprocessor circuits but only 10.6 % and 1.5 %, respectively.

Photomasks for the " classic " photolithography

To increase the resolution of the wafer exposure process, more elaborate versions of classic chrome masks were developed. They can be classified as follows:

• Binary mask ( engl. binary mask interface, BIM) Chrome mask (german chrome on glass, COG ): As an industry standard usually used synonymously with binary mask
• OMOG Mask (german opaque MoSi on glass, OMOG, dt opaque MoSi on glass).

• Chrome-free phase mask (English chromeless phase-shift mask, CPM)
• Alternating phase mask (English alternating phase-shift mask, AltPSM also Levenson PSM called ) Rim - phase masks (german rim phase-shift mask, Rim -PSM )

Binary or chromium mask

A binary mask is a photomask in which the pattern to be imaged in the form of non-permeable, that is absorbent, and "open" ( transparent ) places is included. In general, this is accomplished through a strongly absorbing thin film on a transparent substrate. The chrome mask is the "classic " photomask. It consists of a glass substrate onto which a chromium layer is applied structured lichtunduchlässige. This type of photomask is the version most often used because it is the cheapest and fastest produced. For many less critical process steps in semiconductor development their resolution is sufficient. Binary masks are also available with different absorber materials available to call here is the OMOG Mask (german opaque MoSi on glass ), which uses instead of chromium a sufficiently thick layer of so-called molybdenum silicon ( MoSi ). In order to optimize the optical properties of such masks to suppress, for example, the reflection of rückgestahltem from the wafer light, the layers have often with the thickness changing composition of molybdenum and silicon with oxygen, nitrogen and carbon, which is why the term molybdenum silicon is not entirely correct. Another variant is the "thin OMOG mask " with a layer stack comprising a thinner molybdenum silicon ( MoSi ), and a thin overlying layer of chromium.

Chrome-free phase mask

In chromeless phase mask (English chromeless phase-shift mask, CPM) is omitted entirely, a coating of the substrate. Contrast the structure is obtained exclusively by the phase shift of light through the glass substrate respectively in etched trenches.

Alternating phase mask

In alternating phase mask is a combination of chromium and chromium-free mask mask. In addition to the states " opaque " ( chrombeschichted ) and "see-through " ( chrome-free ) of a pure binary mask there are clear points where was etched into the glass substrate. They are thus "deeper" than the regular transparent places, the depth is set so that there is a to the normal opaque Make a 180 ° phase difference. The areas are alternately (hence " alternating " ) next to one another, so that there is a pattern in which the two variants are always separated by transparent bodies by an opaque area ( " opaque " / " transparent " / " opaque " / " transparent 180 ° phase- shifted ", etc.). In this way, the contrast of the image can be increased.

Due to the complicated calculation of the distribution of etched glass trenches (possibly conflicting requirements when colliding structure edges ), the data processing is extremely costly.

A slightly modified version is called the Rim phase mask (of English. Rim = hem ). With this, the transparent and opaque regions of a skirting, i.e. separated saumartigen transparent area, which causes a phase shift of 180 °.

Halbtonphasenmaske

A Halbtonphasenmaske (English attenuated phase-shift mask, AttPSM ) is very similar to the COG or binary mask under construction. The crucial difference between these types of masks, the thickness of the absorber material. Unlike the binary masks coated with absorber material areas are not 100% opaque but weak partially transparent (hence " semitone " ) and effected by the absorber, a phase shift of 180 ° experiences. The transmittance (in terms of the incident light on the mask ), this is between 4 and 20%, but usually at 6 %. The layer thickness is chosen so that the radiation used for lithographic imaging on passing through a phase shift of 180 ° with respect to radiation which penetrates only the glass substrate undergoes. The contrast of the imaged structure and thus the resolving power to take by.

The absorber material can usually all typical materials such as chromium oxide (CrO ), chromium oxynitride ( CrOxNy ) Molybdänsiliziumoxid ( MoSiOx ) or Molybdänsiliziumoxinitrid ( MoSiOxNy ), are used. However, the layer thickness, and thus the above-mentioned optical properties are adapted to the illumination wavelength must be used. Was for the i-line lithography ( 365 nm ) and are inter alia Chromoxicarbonnitrid ( CrOxCyNz ) used. Halbtonphasenmasken for KrF and ArF excimer laser lithography equipment ( " DUV lithography " with a wavelength of 248 nm and 193 nm) are largely only masks with a structure layer of silicon nitride ( Si3N4), with about 5% molybdenum ( mo ) is doped, also referred to as MoSi.

Halbtonphasenmasken were introduced in the early 1990s and are used to today.

Tritone phase mask

The Tritone phase mask ( " three-tone phase mask " ) can be thought of as a combination of binary and Halbtonphasenmaske. Fundamental to the function is a layer stack of a partially transparent (often MoSi ) and an overlying opaque absorber layer (usually chromium). The layers are structured differently so that undichsichtige ( Cr MoSi ), partially transparent ( MoSi ), and transparent ( uncoated) regions are created. In this way, through the chromium layer light components that are not needed for the projection or interfere would be removed and the contrast can be increased in the photoresist.

Photomasks for the next-generation lithography

It is foreseeable that the successor technology of the 193- nm immersion lithography ( art 2012) will use this new significantly different functional mechanisms. Some of these techniques, which are often summarized under the name of next-generation lithography, further based on the reduction of the wavelength of electromagnetic radiation, such as EUV and X-ray lithography. However, since they lie in a wavelength range in which the properties of the materials differ significantly from that in the optical and nachenultravioletten area, the implementation of these techniques requires new mask shapes and funktionsmechnismen.

EUV mask

EUV masks are designed for a light in the extreme ultraviolet wavelength of 13.5 nm. They are used in EUV lithography. Due to the high absorption of useful substrate materials in this spectral EUV masks can not be used in transmission. However, the reflectance of common coating materials is also very low. One manages multi- layer systems (English multilayer, ML ), which lie directly below the patterned area and function as bragg shearing interference levels. The coating consists of 40-50 Bragg pairs that are typically formed from a molybdenum and a silicon layer. The layer thicknesses are designed for an angle of incidence of 6 °, and to reach reflectivities of about 65%. For the structuring tantalum -based absorbers are used ( tantalum dioxide, tantalum ).

Production

The production of the desired structures is usually done by direct writing laser and electron beam lithography, the laser lithography with 60.3 % constitutes the largest share due to the high proportion of masks for 250 - nm technology node and beyond. Masks with smaller structures are produced by electron beam lithography. The proportion of plants with variable shaped beam (vector shaped e-beam ) with good 37.5% ( 24.0% taking advantage of energies greater than 50 keV).

The actual manufacturing of the masks is similar to photolithography. In the case of conventional chrome masks to a blank ( the blank) with a thin layer of chromium coated (often by sputter deposition ). The desired structures are created on a finished coated substrate by unneeded chrome is removed.

Depending on which photolithographic processes ( contact exposure, projection exposure, etc.), the mask is designed, the structures are the same as the structures of the subsequent photoresist layer on the wafer (scale 1:1, market share approximately 12.6% ) or in the scale 4:1 (about 43.6 %) or 5:1 (about 41.4%) increased as the later target structures on the wafer.

Manufacturer of photomasks include the Japanese company Dai Nippon Printing (DNP ) and Hoya, the American company Toppan photomasks (USA) - participated in the Advanced Mask Technology Center ( AMTC, Germany ) in a joint venture with Global Foundries - and Photronics (USA) and Taiwan Mask Corporation ( TMC, Taiwan) and Compugraphics (part of OM Group, UK / USA).

Possible defects

Already in the preparation may cause various defects. For example, particles or other obstructions cause already during deposition of the chromium layer does not arise confined areas. Also, it can lead to imaging errors on the mask in the exposure by the electron beam, this can be both open sectors and non- addition areas be opened. In addition, provide electrostatic discharges a potential source of defects dar. Here, an electric voltage between two unconnected areas chromium discharges through an arc ( sub-micron ). Similar to the arc evaporation, this leads to Sputtereffekten on the chromium layer and thus to defects. Due to the high cost per mask minor defects are costly to individual work corrected instead of the mask to make again. For the correction of holes, an additional layer of chromium is applied locally and removed excess material. For this purpose, in recent years, various methods based on Laserstrahlverdampfen and deposition, focused-ion -beam sputter etching, ion beam assisted deposition, developed by micromanipulation AFM techniques and electron beam assisted methods.

The main " defects" where use of photomasks are contamination by particles from the air or by abrasion of the Hantierungssysteme. However the latter by appropriate choice of materials and minimizing the contact area of ​​the plants with the mask largely manageable. To protect the very expensive mask against dirt from the air, a pellicle is often used in addition. It is a transparent film ( for example, nitrocellulose ), which is stretched on a frame made of plastic. The frame is mounted on the side of the patterned layer. This is to prevent particles from getting directly on the patterned layer of the mask and thus lead to imaging errors. Contamination of the pellicle, however, are significantly by the distance to the structure layer (usually 3 or 5 mm) outside the focus of the exposure system and disturb the picture or only to a much lesser extent thus. In addition, accreted particles as on the unstructured side of the mask can be easily with nitrogen Blow off without damaging the mask. Furthermore, reducing the risk of mechanical damage and replacement of the pellicle is much cheaper than cleaning or repairing the mask itself

A further increasingly winning important defect is the formation of crystals on the mask surface, the so-called haze (English for, extractor ',' opacity '). This is, for example, ammonium sulfate crystals that form, for example, residues of ammonium hydroxide and sulfuric acid ( as sulfur source). Both materials are used for the cleaning of the screens. In haze that slowly grows on the photomask during operation, the sulfur originates mostly from the ambient air sulfur dioxide present.