Optipedia • SPIE Press books opened for your reference.
Step-and-repeat and step-and-scan modes of wafer patterning require masks with no killer defects in order to achieve good yields. While masks are made without any defects that result in nonfunctional die, preventing particles from depositing on masks during extended mask usage is a challenge, even in state-of-the-art cleanrooms. To avoid new printable defects, pellicles are attached to photomasks.1 Pellicles are thin (~1 µm) polymer films stretched across a frame that is attached to the mask (Fig. 7.16). Typical frame heights are 5–10 mm, with 6.35 mm a common value, the same as the thickness of the photomask blank. Particles deposited on the pelliclized photomask fall onto the pellicle or glass backside of the mask, and are therefore several millimeters away from the chrome features that are being imaged. With small depths-of-field, these particles are not going to be in focus.
With typical depths-of-focus less than 1 µm, millimeter stand offs are expected to blur the image of the particles significantly. The requirements of pellicles go beyond blurring the images of particulate defects. Pellicle stand-offs must be large enough to prevent defects from reducing the light intensity of the desired mask patterns significantly. Early theoretical studies2 showed that image intensities are not affected by amounts greater than 10% as long as the pellicle stand-off is at least as large as
where M is the lens reduction, NA is the numerical aperture of the lens, and d is the diameter of the particle on the pellicle. More detailed theoretical investigations have shown that imaging is protected for particles only about onehalf that given by Eq. (7.3). For example, process windows are not reduced significantly with pellicle stand-off distances of 6.3 mm and particles < 90 µm.3 Usually a pellicle needs to be attached only to the chrome side of the mask, since the glass blank itself serves the same purpose as the pellicle with respect to particles.
Pellicle films are usually polymers, with nitrocellulose and forms of Teflon being common. These materials must be mechanically strong when cast as thin films, be transparent, and resistant to radiation damage. Good light transmission through the pellicles is a combination of transparency and optimization of the thin-film optics.4 Transmission through a nonabsorbing film as a function of film thickness is shown in Fig. 7.17. As can be seen, the pellicle transmission is maximized at particular thicknesses, and pellicles are fabricated at thicknesses corresponding to such maxima. In some instances, antireflection coatings are applied to the pellicle films. Transparency and resistance to radiation damage become more of a challenge as the wavelength is shortened. Transitions to shorter wavelengths have always required some level of pellicle re-engineering. It has been possible to design and fabricate pellicles that have high transmission at multiple wavelengths; for example, there are suitable pellicles for both i-line and KrF lithography.
Pellicle frames are usually made of anodized aluminum. Small holes are commonly drilled into the frame so the pressure of the air in the space enclosed by the mask and pellicle remains equal to the ambient air
pressure.5 These holes are very small, or involve circuitous routes, to prevent these air paths from becoming routes for particles that would defeat the purpose of the pellicle. If the surface of the frame attached to the mask is not very flat, applying the pellicle to the mask can cause the mask to distort, resulting in registration and focus errors. Induced registration errors on the order of 100 nm have been measured.6 These distortions are often unrecognized, because reticle registration measurements are made prior to pellicle application in many mask shops. The magnitude of the distortion depends upon the compliance of the adhesive used to bond the pellicle frame to the mask. Less mask distortion occurs with flatter pellicle frames and more flexible adhesives.
- V. Shea and W. J. Wojcik, “Pellicle cover for projection printing system,” U.S. Patent # 4,131,363 (1978).
- A. Flamholz, “An analysis of pellicle parameters for step-and-repeat projection,” Proc. SPIE 470, pp. 138–146 (1984).
- P-Y. Yang, M. Yeung, and H. Gaw, “Printability of pellicle defects in DUV 0.5-µm lithography,” Proc. SPIE 1604, pp. 106–117 (1991).
- W.N. Partlo and W.G. Oldham, “Transmission measurements of pellicles for deep-UV lithography,” IEEE Trans. Semicond. Manufact. 4(12), pp. 128–133 (1991).
- R.W. Murphy and R. Boyd, “The effect of pressure differentials on pelliclized photomasks," Proc. SPIE 2322, pp. 187–210, (1994).
- W. Chen, J. Carroll, G. Storm, R. Ivancich, J. Maloney, O. Maurin, and E. Soudeillet, “Pellicle-induced reticle distortion: an experimental investigation,” Proc. SPIE 3546, pp. 167–172 (1998).