1
Fundamental investigation on CO2 laser-produced Sn plasma
for an EUVL source
Yezheng Tao*, Mark Tillack, Kevin Sequoia, Russel Burdt, Sam Yuspeh, and Farrokh Najmabadi
University of California, San Diego
2008 International Workshop on EUV LithographyJune 10-12th, 2008. Maui, Hawaii, USA.
2
Contents
1. Backgrounds2. CO2 MOPA laser and diagnostics for fundamental EUVL source
research3. Results from CO2 laser-produced Sn plasma
Plasma density profileEffect of wavelength on mass of ablated materialEffect of contamination of the targetEffect of pulse duration
4. Discussions
3
1. Backgrounds - Is fundamental research still necessary for EUVL source?
Issues remaining in CO2 LPP EUVL source 1. Laser – cost of ownership2. CE- higher value is still possible, droplet3. Debris and out-of-band radiation 4. Mass-limited Sn target operation
Fundamental research is necessary
CO2 laser-produced Sn plasma is one of the most promising candidates for HVM EUVL source due to its high efficiency, relatively low cost, and the ability to scale to high power.
4
Oscillator.
LPX-210 i
2. A CO2 MOPA laser for EUV source fundamental research was developed at UCSD
Nd:YAG
laser
Lumonics LaserMark
Pre-Am
pQ
uestek-
Final AMP
DG 535-II3 TEA CO2 lasersLPX-Osc, Questek-Pre-Amp, Lumonics-Final-Amp
1 Nd:YAG laser, trigger air-breakdown Synchronization from DG 535 I
5
Pulse durations from 15 to 110 ns are obtained by an external triggered air breakdown plasma
Temporal shape of CO2 laser pulse obtained using the plasma shutter at various delay times
Characters of the MOPA system
Beam size: φ12 mm Laser pulse durations: 15 to 110 ns Pulse energy: 100 mJ (15 ns), 150 mJ
(25ns), 400 mJ (60 ns), 550 mJ (110ns)
Focal spot: ~ 200 μm (F/10), 100 μm (F/5)
Intensity: 1.5-2×1010 W/cm2, 8 ×1010
W/cm2
6
Comprehensive diagnostics for EUV, debris, and plasma
Laser pulse energy and temporal shape are monitored for each shot.E-Mon, FC, TGS, Interferometer, visible spectrometer and newly developed diagnostics are employed.
ns Nd:Y
AG
laser
CO2 laser
DG 535I
To DG 535 II
7
Temporally resolved in-band 13.5 nm EUV light detector was setup
EUV mirror
EUV PD
EUV mirror: NTT, 13.5 nm, multi-layers Mo/Si EUV PD: IRD AXUV HS5Rising time : 700 psBandwidth: Broadband, 7-17 nm
In-band 4% at 13.5 nm
Transmission of Zr and Zr plus Mo/Si mirror
Diagnostics for temporal shape of in-band 13.5 nm EUV
laser
8
Narrow-band EUV detector provides more accurate information about the dynamics of in- band EUV emission
Broadband light has a slower rising and falling slopes, and has a wider FWHM.This different comes from out-of-band emission.In-band provides more accurate information
9
In-band 13.5 nm EUV imaging
EUV mirror: NTT, 13.5 nm, multi-layers Mo/Si CCD camera: Bandwidth: 4% at 13.5 nmSpatial resolution: 5 μm
CCD
Zr filter Mo/Si mirror
plasma
laser
10
A quarter- static electric energy analyzer was developed
11
3. The MOPA CO2 laser can produce enough intensity to generate efficient 13.5 nm EUV light
In-band CE vs. focus lens position
CE dip located at the best focus and the spectral peak located at 13.5 nm reveals that the laser intensity of the MOPA CO2 laser reaches the optimum intensity for EUV experiment.
Soft x-ray spectra under various intensities
DOF ~ 1.5 mm
12
Accurate time-resolved interferogram is obtained with a fine time synchronization of the MOPA
0 ns 30 ns 50 ns
Data is under analyzing
laser
200 μm
13
CO2 laser ablates much less material as compared with that of Nd:YAG laser
CO2 laser Sn Plasma YAG laser Sn Plasma
For both cases, the probe beam is 532 nm green light.For Nd:YAG laser, the broad black region represents the region with a density above the nc (4×1021 cm-3) for green light.For CO2 laser, the opaque region is very small. The fringe shift gives out the density of the region around marked by the red line, ~1019 cm-3.At least, 100 times less material is ablated by CO2 laser as compared with Nd:YAG laser while CO2 has a higher CE. Efforts to understand the fundamentals are necessary.200 μm
laser
14
CO2 laser is very sensitive to the contaminations
Single peak for 2nd and after shots.Double peaks from fresh target surfaceThe fast peak may come from the contaminations (C,O from pump) on the surface.This double peaks was never observed for Nd:YAG laser.For lower laser intensity, most of the laser energy goes into contaminations.Energetic Sn ions are produced by CO2 laser.
E = 2 keV
15
In-band CE only weakly depends on pulse duration
In-band CE, 2.8 % (2 π), is constantly obtained with CO2 laser pulse with pulse durations from 25 to 110 ns .Long pulse could significantly simplify and reduce the cost of the CO2 laser used in EUVL source.Larger pulse energy accompanying with long pulse makes it easier to realize mass-limited target for droplet targetLong pulse may make it easier for alignment.Further effort is necessary to clarify the plasma physics dominating this effect.
16
The temporal shape of the in-band 13.5 nm EUV light follows that of laser pulse
In-band 13.5 nm EUV light temporally follows laser pulse.For short pulse, EUV light is a little wider than laser, comes.For short pulse, EUV lasts for several ns even after the laser turns offFor long pulse, EUV is shorter than laser. Even the short tail can contribute to a little EUV emission generation.So it is reasonable to get high CE even with a long pulse.
17
Similar soft x-ray spectra are observed for various pulse durations
Spectra with 25 and 50 ns laser pulses are the same.Spectra from 100 ns pulse is similar with those of short pulse, except for a shift of spectral peak. The spectral shift of the peak comes from the temporal average. The low intensity tail of the long pulse contributes a lower Te, resulting a spectral shift towards long wavelength.This confirms the constant CE observed over a wide range of pulse durations.
18
Long pulse and the tails produces extra slow ions
Long pulse produces extra slow ions, arising from the 3 times extra energy. More input energy, more EUV light, more slow ions. GOOD!Short tail produces extra slow ions. OK!Long tail produces a lot of extra slow ions. BAD!
The absorption length of laser light in plasma at various plasma densities for Nd:YAG and CO2 lasers
Absorption length of laser light in plasma(kTe =30 eV, Z=10)
2
2/3)(/1
e
ece
ei
g
e
cabs Zn
kTnnvv
nnl
−∝=
Density Nd:YAG (μm)
CO2 (μm)
nc /20 100 10000nc /10 25 2500
nc /2 1 100
For long pulse Nd:YAG laser, distributed laser absorption is significant.For CO2 laser, most of the laser energy is always locally absorbed around the nc.
The absorption length for 1/e intensity attenuation of laser light in plasma is obtained from D.Attwood, “Soft x-ray and EUV radiation”.
Several previous experiments have shown that CO2 laser-produced Sn plasma has a narrower spectrum than that of Nd:YAG laser, confirming the above estimation.
4. Discussions
20
1. Capabilities to carry out fundamental researches for CO2 laser-produced EUVL source have been develop at UCSD.
2. It was found that in-band CE weakly depends on pulse duration. CE 2.8 %, is obtained over pulse durations from 25 ns to 110 ns.
3. It was noted that CO2 laser is sensitive to the contamination of the Sn target.4. Accurate interferometry was obtained for CO2 Sn plasma, data is under
analyzing.5. Interferogram shows that CO2 laser ablates much less materials as compared
with Nd:YAG laser while CO2 has a higher CE.6. Temporal shape of in-band 13.5 nm EUV emission was observed. It was
shown that the temporal shape of the 13.5 nm EUV light follows that of laser pulse.
7. Long pulse only produces additional slow ions as compared with that of short pulse. Even a short tail may be acceptable.
Summary
Acknowledgements
This work was supported by Cymer Inc. and by the University of California (UC) under the UC Industry-University Cooperative Research Program (ele06-10278).
21