Semiconductor device manufacturing: process – Coating with electrically or thermally conductive material – To form ohmic contact to semiconductive material
Reexamination Certificate
2002-12-04
2004-04-13
Elms, Richard (Department: 2824)
Semiconductor device manufacturing: process
Coating with electrically or thermally conductive material
To form ohmic contact to semiconductive material
C438S624000, C438S636000, C438S637000
Reexamination Certificate
active
06720256
ABSTRACT:
FIELD OF THE INVENTION
The invention relates to the field of fabricating integrated circuits and other electronic devices and in particular to an improved method of photoresist patterning that provides a wider process latitude and higher yield during the formation of dual damascene structures.
BACKGROUND OF THE INVENTION
The manufacture of integrated circuits involves a lithographic process to form a pattern in a photoresist layer and one or more etching steps to transfer each pattern into a substrate. While most photoresists are optimized for front end of the line (FEOL) applications that generally have smaller feature sizes or critical dimensions (CD) than back end of the line (BEOL) applications, BEOL processes have unique challenges that require special solutions in some cases. For example, when forming a damascene structure, the lithography process usually must contend with considerable topography and has to be compatible with dielectric materials that can poison the photoresist.
There is an ever present demand for smaller circuits in semiconductor devices in order to achieve higher performance. From a photoresist standpoint, a smaller CD in a feature that will become a circuit is more easily printed when the film thickness is reduced or when the exposure wavelength is decreased according to the equation R=k&lgr;/NA. R is the minimum CD that can be resolved while k is a process constant, &lgr; is the exposure wavelength, and NA is the numerical aperture of the exposure tool. Thinner films help to lower k but in the example of a damascene process, the uneven surface of the substrate prevents a thin photoresist film from becoming planar. Film thickness non-uniformity leads to problems with controlling reflectivity off the substrate during exposure and as a result there is a wide variation in printed CD for each feature in the photoresist pattern. In addition, if photoresist fills the via part of the damascene structure, it is difficult to remove during the patterning step and residue or scum is often left behind that can increase cost due to rework and lower yield
A bilayer concept has been introduced to overcome the difficulties associated with imaging a thin single layer of photoresist. Typically, the top layer in a bilayer scheme is a thin film of photoresist containing a small percentage of an element like silicon that can easily form an etch resistant oxide in an oxygen plasma. The bottom layer is thicker so that it can form a planar surface over topography and often contains highly absorbing material that minimizes reflectivity to improve top layer patterning. The bottom layer is usually not photosensitive but may be comprised of a crosslinkable component that reacts with a thermally generated acid. The resulting crosslinked layer tends to be more unreactive toward the solvent or other components of the top layer to avoid mixing of the two layers. In theory, the thin photosensitive layer on a planar underlayer should provide a path to forming small feature sizes with a good process latitude during patterning of the top layer. The silicon containing film provides a high selectivity for transferring a pattern through a thick underlayer during an oxygen etch.
However, a lack of maturity in silicon containing photoresist technology has prevented widespread acceptance of the bilayer concept in the industry. When the composition of the silicon containing photoresist is optimized to provide more silicon content for better etch selectivity, the imaging performance usually suffers. CD non-uniformity across the wafer is one particular problem that is observed when imaging immature bilayer photoresists.
In order to achieve finer resolution in patterns, the exposing wavelength (&lgr;) has been steadily shifting lower in recent technology generations or nodes. For larger feature sizes from about 300 nm to about 1 micron, i-line (365 nm) or g-line (436 nm) exposure tools are more popular because of a lower cost of ownership. For the 130 nm to 250 nm nodes, Deep UV (248 nm) exposure tools are considered state of the art. Meanwhile, 193 nm exposure tools are thought to be the best solution for reaching the 100 nm node and a 157 nm technology is being developed for the 70 nm node.
With the shift to 248 nm and 193 nm exposure wavelengths, a new lithography concept was introduced in which photoresists operate by a chemical amplification (CA) mechanism whereby one molecule of strong acid is capable of causing hundreds of chemical reactions in an exposed film. A strong acid is generated by exposing a photosensitive component and the acid reacts with acid labile groups on a polymer in positive tone photoresist. In negative tone photoresist, the acid initiates a crosslinking reaction. Unfortunately, this new approach is quite sensitive to traces of base compounds such as airborne amines or amines that diffuse into the chemically amplified photoresist from an adjacent layer. Sometimes, amine concentrations as low as parts per billion (ppb) can inhibit or “poison” the CA reaction enough to prevent a pattern from being formed.
Another concern that has recently appeared for 193 nm and 248 nm applications is outgassing from silicon containing resists. High energy exposures are capable of breaking C—Si or O—Si bonds that can release volatile fragments of silicon containing material which can deposit on the lens component of the exposure system. Eventually, the deposit is converted to SiO
2
that forms a permanent coating and irreversibly damages the expensive optics.
An option in the bilayer approach is to expose a silicon free photoresist layer over an underlayer and then selectively introduce a silicon reagent into either the exposed or unexposed regions. This method of treating a photosensitive film with a silicon compound that reacts to become incorporated into the film is called silylation. In U.S. Pat. No. 5,922,516, the underlayer is a photoresist that has been thermally crosslinked at a temperature between 110° C. and 140° and a silicon compound in a vapor phase reacts with the top photoresist. This method appears to be limited to an underlayer which is a 365 nm or 436 nm sensitive material because Deep UV (248 nm) and 193 nm photoresists are typically high activation energy systems that are normally processed up to 150° C. without crosslinking. A lack of maturity in silylation tools is another concern for this technique.
A related bilayer method described in U.S. Pat. No. 5,286,607 involves silylation of the underlayer which is a photoresist and then patterning a second photoresist layer above the silylated material. Generally, silylation does not provide the same uniformity of silicon content across a film and into a film as when the coated layer already contains silicon. Moreover, an optimized bilayer system has the refractive index (RI) of the underlayer matched to the top layer so that the patterning process produces a vertical profile in the top layer. In other words, the sidewalls on the photoresist image form a straight line and do not have a foot at the interface with the underlayer. A foot is defined as a lateral extension from the base of a photoresist image where some photoresist remains on a substrate. When the (RI) of the two layers is not matched, a foot on the patterned image or photoresist residue on the underlayer tend to form which leads to problems in the subsequent etch transfer step. It is difficult to control the RI with a silylation technique because the composition as a function of depth is not uniform unlike the case with a film formed by spin coating a silicon containing material.
In another prior art example, U.S. Pat. No. 4,882,008 teaches how to perform a silylation and subsequent etch processes in the same chamber. However, the method does not deal with new challenges of photoresist poisoning by amines or with the unique challenges of patterning to form damascene structures.
U.S. Pat. No. 6,120,974 describes a novel polymer that forms an amine after a 193 nm patterned exposure and which forms a sulfonic acid following a second blanket exposure to neu
Chao Li-Chih
Lin Li-Te S.
Wu Tsang-Jiuh
Elms Richard
Taiwan Semiconductor Manufacturing Company
Wilson Christian D.
LandOfFree
Method of dual damascene patterning does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method of dual damascene patterning, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method of dual damascene patterning will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3240844