IMEC Calls EUV Performance ‘Impressive’
After only five months' experience with a new source on its Alpha Demo Tool (ADT), Kurt Ronse, director of the Advanced Lithography Program at IMEC (Leuven, Belgium), said he is impressed with the extreme ultraviolet (EUV) tool's stability and performance to date. IMEC has demonstrated the ability to resolve 35 nm flash patterns with the scanner.
Laura Peters, Editor-in-Chief -- Semiconductor International, 10/23/2008
After only five months’ experience with a new 16 W source on its Alpha Demo Tool (ADT), Kurt Ronse, director of the Advanced Lithography Program at IMEC (Leuven, Belgium), said he is impressed with the extreme ultraviolet (EUV) tool’s stability and performance to date.
“This is the first data from the full-field scanners, and we are impressed with the stability of the system, which shows in the average CD and CD variation over time,” he said. With this scanner, IMEC has demonstrated the ability to resolve 35 nm flash patterns with acceptable CD uniformity (1.5 nm 3σ).
|
| Measurement of full-wafer vertical (V) and horizontal (H) line CD and CD variation over time is stable. (Source: IMEC) |
Ronse emphasized, however, that much progress still needs to be made in EUV. The top three priority items are now source power, defect-free masks, and production-worthy photoresists. In recent months, the resists and masks swapped places in the priority ranking because of the progress made on the part of materials suppliers in providing improved resists.
The ADT target is resolution &32 nm with a line-edge roughness (LER) &4 nm and dose &15 mJ/cm2 (NA=0.25, σ=0.5). Today, at that dose and resolution, LER is 4-6 nm. Ronse said he is also encouraged because the mask error enhancement factor (MEEF) has come down from ~1.43 for early resists to ~1.2 for the more mature resists.
Source power, which is directly tied to throughput of the scanner, must be increased from 120 W today to ~170 W. “It is very important that the source can maintain these high power levels of output day in and day out — not just for short periods of time.” Ronse stressed that the practical use of a scanner necessitates use at a high power level while the wafer is being exposed, then it is not used again until the next wafer is exposed. “When a source is tested today, it operates for eight hours continuously, but that is not how it will be used in manufacturing; you don’t know how it will actually perform until you interface it with the scanner,” he said.
The throughput of the ADT today is 4 wph; the original goal using this 120 W source, which was brought on in August, was 5-6 wph, according to Ronse. Sometime around the end of 2008, IMEC expects to receive an upgraded, 170 W source. IMEC estimates that to achieve 100 wph throughput on an EUV scanner, an immediate focus (IF) power level of 200-250 W will eventually be needed.
|
| Tracking scanner throughput; production goal is 100 wph. (Source: IMEC) |
|
| The source must provide stable output at immediate focus (IF), over several operations/day, multiple weeks and months. (Source: ASML) |
Finally, progress in the mask area requires better inspection tools and a method for keeping the mask defect-free when it is not inside the scanner. Ronse said that inspection tools today do not have the required resolution to find the defects of interest. He expects EUV masks to require frequent cleans, so fabs likely will need to have in-house procedures for cleaning the masks, unlike with optical mask cleaning, which is done at the mask supplier’s facility. Keeping the masks free of contamination when they are stored in pods may require a nitrogen or vacuum environment; this has yet to be determined.
-
EUV throughput looks worse than quadruple...
who put the eggs in one basket? - 2008-13-11 03:22:00 -
It doesn't look like EUV will be ready in time for 22 nm or 16 nm folks. EUV throughput orders...
LDI - 2008-13-11 03:03:00 -
Things will change since ASML provide new souce. EUV is the only hope for the...
imal - 2008-8-11 15:10:00 -
EUV throughput is a...
you gotta be kidding - 2008-25-10 03:02:00 -
CD mean and 3s drifting...
statistician - 2008-24-10 17:26:00
EUV Requires Enhanced Metrology
04/01/2009Double Patterning Leads Race for 32 nm
10/18/2007Score One for EUV
10/19/2007IMEC’s Ronse Worth a Listen
11/27/2007IMEC Expecting Two EUV Source Upgrades
11/27/2007
