Recent Trends in Optical Lithography

Recent Trends in Optical Lithography

1. Introduction

As the device size is reduced below 100nm various advanced methods are devised to pattern such small devices. Presently in the industry 193 nm wavelength is used to pattern the devices. Limitations imposed by 193 nm lithography in patterning devices of size less than 100nm has forced the active development of 157 nm and 121 nm lithography. Liquid immersion lithography for the same 193 nm is widely used to pattern low sized device. Practical implementation of short wavelength optical lithography is challenged by the constraints of materials used and optimizing their photochemistry. A transition to the shorter wavelength requires many of the issues related to the material science to be addressed like improving the lens materials and coatings, and the development of transparent and etch resistant photo resists.

2. Lithography at 157 nm

Several changes in the projection systems, photo mask and the photo resists are required for the technology transition from 193 nm to 157 nm. The lens material used for the projection system is crystalline calcium fluoride, optical coatings are made up of fluoride thin films and high purity nitrogen is used for ambient. The quality of the crystal grown for 157 nm should be of higher quality than that of 193 nm. At the new wavelength the absorption coefficient should be less than 0.002cm^-1. This requirement imposes stringent impurity levels in the crystal. The expected residual stress induced birefringence must be less than 1nm/cm and the inhomogenesities in refractive index must be less than 1ppm.

Either the crystal growth or the annealing can't reduce the intrinsic birefringence of calcium fluoride. This effect is more observed in 157 nm than in 193 nm. Lenses made up of two different crystalline orientations can overcome this problem.

Highly transparent antireflective coating is required for optical elements. At 157 nm oxide films used at higher wavelengths are too absorptive. Since fluoride films are used almost universally, their antireflective properties and long term durability to laser radiation exposure should be qualified.

The cleanliness of the purge gas plays vital role at 157 nm. In 193 nm photo induced contamination of optics due to trace contaminants in the purge gas exists. This effect worsens in shorter wavelength side.

Several absorptive materials are under investigation for attenuating phase shift photo mask. Protecting the photo mask from the particle contamination is the important issue related to photo mask in 157 nm. There is a zero tolerance for the particles falling on the photo mask. A very thin membrane, called pellicle, is used to protect the mask. The chosen pellicle materials have shown rapid degradation when pellicle is radiated with a 157 nm laser. Practically 10 % lifetime value of the pellicle must be at least 1KJ/cm2. most of the pellicles tested for 157 nm is found to be having 10% life time of only 75/cm². This failure happened due to the process called as photochemical darkening. [1]

To get good quality photo resists the absorption coefficient of 157 nm photo resists must be reduced to less than or equal to 2um^-1. Photo resists having a thickness of 120 nm have shown resolution of 120 nm. Absorption coefficients of less than 1um^-1 have been reported recently. Incorporation of transparent inorganic nano particles in the polymer of the photo resist accomplishes two goals- increased transparency and enhanced plasma etch resistance (i.e. low etch rate), thus enabling a smaller thickness.

3. Liquid immersion lithography

In this technique higher resolution is achieved by increasing numerical aperture (NA) beyond 1 through the use of immersion liquids. From the equation R=(k1.λ)/NA it is clear that higher resolution can be achieved either by reducing λ and k1 or by increasing the refractive index of the imaging medium. If we employ a liquid of refractive index 1.4 instead of air or nitrogen then the maximum NA achievable approaches to 1.4. This enables the resolution enhancement proportional to 'n'. This can be achieved without changing the any of the established technology base like wavelength, laser, photo masks etc.

Liquid immersion lithography is implemented at deep ultraviolet wavelength of 193 and 157 nm. Availability of transparent liquids is vital to liquid immersion lithography. DI water provides very good absorption coefficient of 0.036cm^-1. This enables the projection system designers to keep sufficient working distance (~1mm) of several millimeters. Presently available transparent liquid at 157 nm has an absorption coefficient of 3cm^-1 requiring a working distance of less than 0.1 mm. For a better working distance reduction in absorption coefficient is necessary.

Photo resist performance is affected by interactions between the immersion liquid and the photo resist. In the absence of laser irradiation thinning of 193 nm photo resist by the liquid is less than 0.5 nm which is well within the acceptable limits. 45 nm device patterning is possible with the 193 nm liquid immersion lithography and 32 nm devices can be manufactured with the advent of 157 nm technology.

4. Lithography at 121 nm

The wavelength of 121.6 nm is also known as Lyman alpha line. This is shortly called as 121 nm. This is the shortest wavelength used in optical lithography. Vacuum based systems are not necessary for 121 nm technology compared to alternatives like Extreme ultra Violet (EUV) at 13 nm or electron beam lithography. Lithographic systems used for 157 nm technology can be engineered to 121 nm technology also because absorption coefficient of most common purge gases and atmospheric constituents are similar.

The main difference of 121 nm from the 157 nm is the use of pulsed discharge sources instead of lasers. The lack of suitable transparent optical materials is imposing difficulties in the development of 121 nm technology. Since high NA over 0.75 is required for the projection systems, the lenses used for optical systems should be of either all refractive (using only lenses) or catadioptric (a combination of lenses and mirrors). A highly transparent lens material like lithium fluoride which has high band gap is used for the manufacture of lenses.

Design of proper photo resist for the new wavelength is always a challenging job. The universally high absorption coefficient of organic polymers necessitates the use of thin layers of photo resist, approximately of thickness 25 to 35 nm. This mentioned thickness of the photo resist is three times less than the thinnest photo resist used at 157 or 193 nm. There can be setback on the development of 121 nm technologies considering the fact that lower thickness of the photo resist may not be compatible with the pattern transfer processing steps such as reactive ion etching.

A survey [2] of leading chip manufacturers finds that for 130 nm node 248 nm lithography is used; for 100 nm node 193 nm lithography is the choice of majority of companies. For the fabrication of 70 nm patterns 157 nm lithography is a workhorse, but several companies have shown interest in Electron Beam Lithography (EBL). For the 50 nm and bellow node there is a great degree of debate in the industry to switch over to either EBL or Extreme Ultra Violet (EUV) lithography.

However the use of optical extension techniques such as phase-shifting masks and off-axis illumination has enabled the industry to find good manufacturability solutions to extend the lifetime of optical lithography. Introduction of 157 nm lithography along with the breakthrough development of fused silica for mask materials has added more nodes to the life time of optical lithography.

5. Conclusions

As the feature size is decreasing to sub micron level, in addition to conventional optical lithography techniques, several new technologies such as x-ray, e-beam direct write (EBDW), extreme UV (EUV), electron beam lithography (EBL) and ion projection lithography (IPL) are emerging in the industry. But breakthrough researches and introduction of liquid immersion technique with new wavelength of 157 nm and 121 nm in the field of optical lithography has pushed the limits. Thus the recent trend in the optical lithography has given more breathing space for the cost sensitive semiconductor industry.

Bibliography

[1] Mordechai Rothschild, Theodore M. Bloomstein, Theodore H. Fedynyshyn, Roderick R. Kunz, Vladimir Liberman, Michael Switkes, Nikolay N. Efremow, Jr., Stephen T. Palmacci, Jan H.C. Sedlacek, Dennis E. Hardy, and Andrew Grenville, Recent Trends in Optical Lithography, Lincoln laboratory journal, volume 14, number 2, 2003

http://www.ll.mit.edu/news/journal/pdf/vol14_no2/14_2recenttrendsop.pdf

[2] Martin McCallum, Lithography Trends: A Review and Outlook, Future Fab Intl. Volume 9, International SEMATECH, (1/7/2000), www.futurefab.com