Why Do We Care About CaF2?
Ruth DeJule, Associate Editor -- Semiconductor International, 1/1/1999
As lithography tool wavelengths transition from 365 and 248 nm to 193 and 157 nm, requirements concerning optical transmission and radiation resistance become more stringent. For projection and illumination optics, monocrystalline calcium fluoride (CaF2) is expected to meet required specifications.
The low transmission of optical glasses developed for i-line (365 nm) steppers prohibit their use at 248 nm. Instead, synthetic fused silica displaying a very high homogeneity, low stress birefringence, high transmission and sufficient resistance to the high-energy DUV radiation was developed. However, for 193 nm tools, fused silica is inappropriate because of the considerable decrease in optical transmission below 200 nm. In addition, interaction between UV radiation and quartz glass results in color center formation and a change in the refractive index (compaction).
The material of choice for 193 nm optics, CaF2, shows no compaction and has notably improved transmission and resistance against radiation-induced absorption over fused silica. Further requirements for optical lithography include homogeneity >1 ppm and stress birefringence <1 nm/cm in <111> single crystals containing no striae, bubbles or inclusions for blanks up to 300 mm in diameter and 100 mm in thickness. While i-line glasses and fused silica meet these requirements, meeting specifications in the production and supply of CaF2 crystals is non-trivial.
 |
|
Fig. 1. Qualification of stress birefringence is performed at a CaF2 crystal growth factory in Eisenberg, Germany. (Source: SCHOTT) |
Progress, however, is being made by companies such as Bicron (Solon, Ohio), Canon's Optron Inc. (Ibaraki, Japan) and SCHOTT to improve yields and control the growth process. Using its existing crystal growth furnaces, SCHOTT has successfully developed and produced material with qualities required for first-generation prototype 193 nm steppers such as ASML's 5500/900 193 nm tool. Typically, CaF2 crystals that meet the high quality requirements for use in projection optics are available only in single pieces, though high-temperature vacuum furnaces are currently being designed and built for the production of large crystals.
Compounding the uncertainties surrounding timing and demand for the material, stepper manufacturers are reducing the amount of CaF2 required in next-generation 193 nm steppers to cope with limited supplies. Difficult as it may be, what may be needed to push supply forward is a firm commitment. For while investment in new furnaces by CaF2 producers is limited, a large number of new furnaces is all that is necessary to accelerate development and availability of improved crystals, Moersen said. To meet peak demand for CaF2, he mentioned supporting a number of furnaces that will be operated for only one year, though questions of economic feasibility are apparent. Installation of new furnaces may require a one year lead time and three to six months ramp-up time.
Thus far, technological feasibility has been demonstrated, but the quality
and yield of CaF2 crystals must be improved. What may be needed is
more development programs and greater cooperation among suppliers and end users.
(Next month: CaF2 manufacturing issues)
