Radiation imagery chemistry: process – composition – or product th – Imaging affecting physical property of radiation sensitive... – Forming nonplanar surface
Reexamination Certificate
1999-03-08
2001-08-14
Duda, Kathleen (Department: 1756)
Radiation imagery chemistry: process, composition, or product th
Imaging affecting physical property of radiation sensitive...
Forming nonplanar surface
C430S325000, C430S950000
Reexamination Certificate
active
06274295
ABSTRACT:
BACKGROUND OF THE INVENTION
One of the problems encountered in practical semiconductor device manufacture is the dependence of the imaging dose of the photoresist on the photoresist thickness. If photoresist films below about 4 &mgr;m (micrometers) thickness are used, as they are in the most common applications for high resolution devices, thin film interference effects lead to periodic behavior of the photoresist sensitivity, the “swing curve”. Photoresist sensitivity can vary dramatically over a small change in film thickness. On a substrate with topography, this limits the accuracy to which photoresist features can be imaged, and it also reduces the apparent depth of focus. Moving from g-line to i-line and eventually DUV radiation brings with it a large increase of the substrate reflectivity, leading to an increase in the magnitude of the problem.
The classic solution to the above problem has been the use of dyed photoresist materials. Dyed photoresists add a non-bleachable absorption to the photoresist (i.e., they increase the Dill B parameter). This reduces the swing curve by causing increased attenuation of the light in the photoresist, even when the photoresist is near to full exposure where, without the dye, it would be quite transparent. Both swing curve and standing waves are reduced by the addition of dye.
However, use of dyed photoresists carries a considerable price: photoresists with all but the very lightest dye loading always exhibit large increases in exposure doses as well as reduced wall angles and final resolution. Both optical and chemical reasons contribute to this deterioration; while the chemical effects (which are due to the changes in the photoresist development parameters caused by the dye addition) may be minimized by judicious choice of the dye's chemical structure, the optical effects (larger Dill B parameter, i.e., increased unbleachable absorption) are inherent in the concept of dye addition, and cannot be avoided. Still, the ease of use of dyed photoresists has made them the most popular solution to reflectivity and swing curve woes.
In the last few years, organic antireflective coatings have entered the market place, and in many cases have proven to be a crucial technology enabling users to safely run processes which previously had low or no latitude. The ease of use of water-soluble antireflective topcoats, which offer only minimal increases in process complexity, has made them a logical choice for users trying to enlarge their process window.
Light falling on a substrate through a thin film undergoes an infinite series of reflections at the boundaries between the thin film and the air as well as at the film/substrate boundary, as shown in FIG.
1
. The incoming and outgoing waves interfere in the film, causing constructive interference if their phase difference is an even multiple of p, and destructive interference if their phase difference is an odd multiple of p. Physicists call this film stack a Fabry-Perot etalon, the theory of which is well established. From these considerations, an approximate formula can be derived (cf., e.g., T. Brunner, Proc. SPIE 1466, 297(1991)) which relates the swing curve ratio of a photoresist to the reflectivity R
t
at the photoresist-air interface, and the reflectivity R
b
at the photoresist/substrate interface:
S=4{square root over (R
t
+L R
b
+L )}·e
−a
r
·d
, (1)
where d is the photoresist thickness.
Essentially, the idea behind the top antireflective layer is to change the phase of light passed through the layer in such as way that the first and the sum of all subsequent reflection amplitudes in
FIG. 1
are out-of-phase by 180 degrees. Incoming and outgoing wave amplitudes interfere destructively, and the swing curve vanishes. Manipulation of the appropriate equations gives two necessary conditions for this phase change:
1. The first and the sum of all subsequent reflection amplitudes are opposite in phase when the film thickness t of the antireflective layer is an odd multiple of t=&lgr;/(4n
t
), where n
t
is the real part of the refractive index of the top layer; and
2. The first and the sum of all subsequent reflection amplitudes are equal in intensity if n
t
={square root over (n
r
+L )}, where n
r
is the refractive index of the photoresist.
It should be realized that the above equation (1) provides a simple first-order approximation to the swing curve ratio. Employing a more exact treatment, one finds that the swing curve ratio for exposure on a diazonaphthoquinone (DNQ) photoresist will not be zero even if the above conditions are exactly met. Using the Prolith/2 lithography simulator, sold by Finle Technology of Austin, Tex., one finds that for a perfect match it is possible to reduce the swing curve to about 5% of the original value (Example 3). If the simulation data are interpolated linearly, one finds that the predicted remaining swing ratio is less than 1%. In practice, the optimum situation will be somewhat worse since the photoresist changes its absorption and its refractive index during exposure, so that a top layer with an unchanging refractive index cannot be matched to both the beginning and end conditions at the same time. This effect is not taken into account by the Prolith™ simulator.
If one looks for chemical materials to make a top antireflective layer, it turns out that while the first condition, t=&lgr;(4n
t
), can be met fairly easily, fulfilling the second one, n
t
={square root over (n
r
+L )}, meets with some fairly formidable obstacles. The refractive index of a photoresist at i-line (365 nanometer) is typically about 1.72 to 1.75, which calls for an antireflective coating with a refractive index of n
t
=1.31 to 1.32. For easier measurement, the literature mostly references the refractive indices at the typical metrology wavelength of 633 nm (nanometers), making the somewhat incorrect assumption that the dispersion behavior of both materials will be identical. In this case, the refractive indices become n
r
=1.64 and n
t
=1.28. Unfortunately, it is difficult to find organic materials with refractive indices in the range of 1.28 to 1.31. Known examples include heavily fluorinated, Teflon®-like polymers which require exotic or environmentally unacceptable solvents such as chlorofluorocarbons as spincasting solvents. Moreover, early examples of these layers had to be removed in a second solvent treatment step prior to development.
The TAR concept gained practical importance only with the advent of water-based, developer-soluble antireflective coatings. These coatings, which were originally developed by researchers at IBM Corp. and which have been described in EP 522990, have a dramatically simplified mode of application. They can be spun on from water directly onto the photoresist in two different ways:
1. in a “coat-bake-coat” process in which the photoresist is first coated and prebaked, and the top coating is then spun on top of the dry photoresist; it becomes non-tacky without the need for a second bake step; or
2. in a “coat-coat-bake” process in which the top coating is spun on top of the wet photoresist film, and both are baked together.
For most photoresist materials, it is even possible to use a single spin coater, since the edge bead removal typically can be set up to clean the coater bowl of all top coat residues. For puddle or spray development, no separate step is needed to remove the top coat after exposure; it is fully developer-soluble and will be washed away in the first few seconds of the development process. For very sensitive processes or for immersion development, it is possible to remove the top coat by a short water rinse.
In order to insure water solubility of the top coat, some compromises have to be made with respect to its refractive index. Consequently, water-based topcoats do not achieve the optimum refractive index of 1.28; one commercial material, the AZ® Aquatar coating, available from AZ Electronic Materials, Somerville, N.J., has a refractive index of 1.41. Sinc
Dammel Ralph R.
Norwood Robert A.
Clariant Finance (BVI) Limited
Duda Kathleen
Jain Sangya
LandOfFree
Light-absorbing antireflective layers with improved... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Light-absorbing antireflective layers with improved..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Light-absorbing antireflective layers with improved... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2501638