Michael E. Clarke, Annie Xia, Joseph Smith, Bipin Parekh, Entegris, Inc.
Liquid immersion lithography (LIL) has catapulted to the forefront of all lithography discussions recently. With several production-worthy toolsets delivered in 2004-2006, LIL is no longer a curiosity, but a mainstream technology.
The semiconductor industry has consistently delivered improved chip performance by doubling the number of transistors designed on a chip every two to three years. This has been achieved primarily by shrinking linewidth (feature size). A key technology enabling this achievement has been optical lithography, which has relied on ever-decreasing wavelength light sources to print finer lines. Currently, dry lithography using 193 nm ArF Excimer laser is targeted to print down to 65 nm feature sizes. Table 1 summarizes the various lithography technologies introduced over the last 20 years.
 Table 1: Key dates for various lithography technologies |
To print finer feature sizes beyond 65 nm, the industry is preparing to extend “193 nm dry lithography” to “193 nm liquid immersion lithography” using ultrapure water (UPW) (see Fig. 1). Immersion lithography has the effect of reducing the exposure wavelength (193 nm immersion can be considered as 134 nm lithography).
 Figure 1: Schematic of immersion lithography concept. |
In the immersion lithography process, a higher refractive index liquid (e.g., deionized water (DI), index = 1.44) is placed between the final lens and the wafer (replacing the lower-index air, index = 1). The higher refractive index of water increases the Numerical Aperture (NA) creating a higher imaging resolution and an increased depth-of-focus of up to 50 percent for printing finer circuit lines onto wafers than a “dry” lens system will allow.
Current development plans show that the DI immersion process can potentially extend conventional 193 nm optical lithography to at least the 45 nm node.1, 2
With the potential benefit of this technology also comes uncertainty and risk. Lithography suppliers are not used to dealing with liquid in their tools and are seeking cooperation from traditional liquid handling companies. In DI immersion lithography, the optical lens is in intimate contact with the water. This poses a risk of lens contamination by impurities in the water. The water quality must be maintained at the highest level of clarity (low absorbance) and purity (parts-per-trillion levels of contaminants) to ensure transmission of imaging radiation through the water. For example, the 193 nm optical absorbance in high-purity water is typically 0.01/cm and varies strongly with any trace amounts of absorbing extrinsic impurities3-6. It is critical that all impurities in water be removed to trace levels at the point-of-use to achieve high process effectiveness (see Table 2).
 Table 2. Water quality for immersion lithography |
null
UPW parameters and effects on immersion process
Key considerations in designing an ultrapure water system for immersion lithography are selection of high purity, inert construction materials, integration of purification techniques in an optimal sequence, control of trace levels of ions, dissolved gases and organics extractables, and precise temperature and flow control.
A POU UPW water system must purify and condition the high-purity fab water and deliver it to the immersion lithography tool lens. The contaminants to be removed from UPW are dissolved gases and bubbles, particles, total oxidizable carbon (TOC) and non-volatile residue (NVR), and ionic and organic extractables that are added from the process materials and piping components. Temperature control is also very important to ensure constant optical properties (refractive index) and to ensure constant physical properties (e.g. density, surface tension and gas solubility). Protection of the optical lens is another important consideration as deposition of haze producing contaminants from water onto the lens and the etching adversely affect lens transmission and durability.
Degassing
The International Technology Roadmap for Semiconductors (ITRS), published in late 2005, identifies bubbles in UPW as a key contaminant in the immersion lithography process.7 The liquid must be free from bubbles that may be caused by the scanning process, exposure, or the fluid delivery, recovery, and recirculation system. Degassing (using vacuum) reduces dissolved gas concentration from the UPW feed and gasses/bubbles generated by the ultraviolet (UV) oxidation source. It is critical to remove the dissolved gasses to ppb levels using clean devices that have low extractables and particle shedding. Conventional degassers are efficient at typical flow rates (>75 percent efficient) but a high level of extractables (TOC and ionic) from these units limit their use to upstream of the UV source, as roughing degassers. Teflon® degassers are more expensive due to lower efficiency (>40 percent efficient), but their cleaner design make them ideal for use after the UV source. The degasser ionic extraction data in Table 3 shows the benefit of low metallic extractables from the all Teflon® (pHasor®) degasser compared to a conventional degasser.
 Table 3. Cleanliness comparison of degassers |
null
Particle control
Since particles in UPW can deposit on the wafer or cast a shadow during the wafer exposure causing defects, it is imperative to remove particles down to 0.03 μm size using POU filtration. One solution is to use a 0.03 μm-rated all-Teflon® filter (QuickChange®) in a disposable format to minimize the handling contamination. The filter uses non-dewetting technology with high particle retention: LRV > 2.5 of 0.03 μm (> 99.7 percent removal), and extremely low extractables.8 Data in Figure 2 show there is no particle shedding by the filter (feed and filtrate particle levels are the same for steady and pulsed flow of clean water though the filter).
 Figure 2: Filter cleanliness. |
Alternatively, a 0.02 μm polyvinylidene fluoride (PVDF) based filter can be used for efficient particle removal from UPW.
TOC reduction
Published research work indicates that organics in UPW can absorb deep ultraviolet (DUV) energy from the stepper and cause defects. Also, organics can deposit on lens causing haze and lens performance impairment. Thus organics need to be reduced from the incoming parts-per-billion (ppb) levels down to the parts-per-trillion (ppt) levels. Entegris has developed a UV oxidation-ion exchange process to reduce total oxidizable carbon (TOC) to ppt levels by breaking down most organic molecules into CO2 and H2O. Other organic molecules might only be ionized and removed. TOC reduction is a direct function of flow rate through both the UV source and purifier. A key to delivering a low TOC water is to use cleaner system components, with reduced leachables. Figure 3 shows the TOC reduction achieved in the immersion system using the oxidation-ion exchange process. The flow rate dependence of TOC removal is shown in Figure 4.
 Figure 3: TOC reduction in an immersion water system. |
null
 Figure 4: TOC removal depends on system water flow rate. |
null
Ions purification
As per the 2005 ITRS guidelines, ion removal from UPW to ppt level is critical. A mixed-bed ion exchanger polisher can be used to deionize the fab water to produce ppt level ions in the system. Table 4 summarizes UPW quality delivered by the immersion system. It also shows that the system components are clean and do not add ionic impurities.
 Table 4. Ionic water quality from th eimmersion system |
null
Temperature control
Temperature changes can have large effects in immersion lithography. A stable water temperature prevents imaging defects by eliminating refractive index changes. Inlet fab water temperature swings are not well known, nor controlled to the degree required. The required accuracy of the point-of-dispense temperature measurement is 0.001°C or 1 mK. Entegris has developed a recirculating water bath method for an “off-tool” water purification system. The output of the recirculating bath would flow through Teflon® heat exchangers and water-jacketed tubing. For a final “on-tool” temperature control, a second stage system is proposed, which would be a “polisher” that controls the final temperature close to the wafer. Plots in Figure 5 show the thermal control for the system to achieve the target temperature is very good. The system is stable to < 0.01°C for target temperature at 20.5°C. Temperature measurement details show accuracy is + 0.0012°C and resolution is + 0.0001°C.
 Figure 5: Stable and accurate temperature control for UPW system. |
null
Liquid flow control
In the immersion process, it is important to deliver a stable DI flow precisely/repeatedly to the illuminated area to prevent any bubble attachment on the wafer or the lens during the filling process. The water-filling rate, over the wafer topography, should remove the resist reaction products, water-soluble resist components and the heat generated during the exposure. The current flow rate control required is in the range from 0.4 to 3 LPM at steady state with slower flow rate at initial fill, to ensure complete filling under the lens, followed by a faster rate during scanning to ensure byproduct removal and meniscus integrity during stage movement. An Entegris flow control module is used in the immersion lithography UPW system to maintain a highly repeatable and stable flow rate through the illuminated area.
null
Michael Clarke is a principal engineer and research and development manager in liquid systems for Entegris. He earned his B.S. degree from the Massachusetts Institute of Technology and an M.S. degree from Rensselaer Polytechnic Institute, both in Mechanical Engineering. Mr. Clarke can be reached at (978) 436-6533 and via e-mail at [email protected].
Bipin Parekh is senior consulting engineer for dispense and subsystems for Entegris. He has a Ph.D. in chemical engineering from State University of New York at Buffalo. Mr. Parekh can be reached at (978) 436-6636 and via e-mail at [email protected].
Annie Xia is currently an application engineer in liquid systems for Entegris. She obtained her B.S. and M.S. degrees in chemical engineering from the Massachusetts Institute of Technology. Ms. Xia can be reached at (978) 436-6651 and via e-mail at [email protected].
Joseph Smith is the product manager for Entegris liquid systems. He earned his B.S. in mechanical engineering from the New Jersey Institute of Technology, and his M.S. degree from Boston University. Mr. Smith can be reached at (978) 436-6663 and via e-mail at [email protected].
References
1. Maloney, Lawrence D. “Tomorrow’s Lithography: It’s All Wet,” Design News, June 2005.
2. Mulkens, Jan; Bob Streefkerk; Martin Hoogendorp; Richard Moerman; Martijn Leenders; Fred de Jong; Marco Stavenga; Herman Boom. Optical Microlithography XVIII, edited by Bruce W. Smith. Proceedings of SPIE, Vol. 5754, Bellingham, WA, 2005, pp. 710-724.
3. Robinson, Chris; Dan Corliss; John Barns; Kevin Cummings; Hans Jansen; Brian Lee; John Woodbeck; Frank Goodwin; Yayi Wei; Peter Benson; Richard Housley. “Immersion Lithography Water Quality at Albany Nanotech,” 2005 International Symposium for Immersion Lithography, Brugges, Belgium, September 2005.
4. Gil, D.; T. Bailey; D. Corliss; M.J. Brodsky; P. Lawson; M. Rutten; Z. Chen; N. Lustig; T. Nigussie; K. Petrillo; C. Robinson. “First Microprocessors with Immersion Lithography,” Optical Microlithography XVIII, edited by Bruce W. Smith. Proceedings of SPIE, Vol. 5754, Bellingham, WA, 2005, pp. 119-128.
5. Liberman, V.; S.T. Palmacci; D. E. Hardy; M. Rothschild; A. Grenville. “Controlled Contamination Studies in 193 nm Immersion Lithography,” Optical Microlithography XVIII, edited by Bruce W. Smith. Proceedings of SPIE, Vol. 5754, Bellingham, WA, 2005, pp. 148-153.
6. Peng, Sheng; Roger H. French; Weiming Qiu; Robert C. Wheland; Min Yang; Michael F. Lemon; Michael K. Crawford. “Second generation Fluids for 193 nm Immersion Lithography,” Optical Microlithography XVIII, edited by Bruce W. Smith. Proceedings of SPIE, Vol. 5754, Bellingham, WA, 2005, pp. 427-434.
7. 2005 ITRS
8. Funahashi, Isamu; Takuya Nagafuchi; Bipin Parekh. “Wafer environment Nanoparticle Contamination Control and Defect Reduction in Front-End-Of-Line (FEOL) Cleaning Processes,” CleanRooms, July 2005, pp. 21-23.