Upload
others
View
5
Download
0
Embed Size (px)
Citation preview
EUV Mask and Optics Contamination
Greg Denbeaux
College of Nanoscale Science and EngineeringUniversity at Albany
July 15, 2009
Outline
• Carbon contamination• Resist Outgassing• Out-of-band radiation• Mask contamination and topography
College of Nanoscale Science & EngineeringNanoFab Complex at the University at Albany
ASML EUV ADT
EUV MET
EUV ROX (outgassing )EUV MiMICS (contamination)
SEMATECH EUV MBDC
Carbon Contamination
Carbon contamination occurs in EUV exposure tools Dominant species deposited is carbon – but inorganic contamination may be larger problem if not cleanableDirect photodissociation and/or secondary electron emission may be mechanism of contamination growth
AdsorbedMolecules
EUVVacuum Chamber
Multilayer
J. Hollenshead and L. Klebanoff, JVST B 24(1), Jan/Feb 2006
A. Wüest, IEUVI optics contamination TWG 2007
Carbon contamination
e- e-e-
Dissociated Molecules
Resist Outgassing and eXposure (ROX) Tool
Photodiode
Energetiq Source
Collector
300 mm waferLoad Lock
Mass spectrometer
Zr Filter
Sample resist outgassing result
28 CO
39
41
56
91 119
Tert-butyl-benzene (C10H14)
Isobutene (C4H8)
(Resist courtesy of Prof. Robert Brainard)AMU
Mol
ecul
es/c
m2
Contamination results due to outgassed species
No significant reflectivity loss for these species at these pressures and doses
Chamber Conditions
Chamber Pressure (Torr)
Exposure time (hours)
Total Dose (J/cm2)
Number of pulses (millions)
Reflectivity drop (ΔR/R%)
Clean(background)
2.5 x 10-8 8 29 36 0.35
Benzene 1 x 10-6 8 29 36 0.35
Tert-Butanol 3 x 10-6 8 11.5 36 -0.09
Diphenyl Sulfide 1 x 10-6 4.2 15 19 0.1
Diphenyl Sulfide 1 x 10-6 3.6 13 16 -0.23
Diphenyl Sulfide 1 x 10-6 2.9 42 13 0.1
EUV MiMICS (mask contamination tool)
EUV Source
Mask
Multilayer Mirror
SiZr Filter
Designed aperture
Carbon containing gas
XYZ stage to hold 6” mask and covers full travel range, with height adjustment
Automated load-lock for mask loading
Best pressure of 1*10-7 Torr
G. Denbeaux, et al., “Accelerated contamination testing of EUV masks.” Proc. SPIE 2008Y.J Fan, et al., “Effect of carbon contamination of EUV masks on imaging.” EUVL Symposium 2008
Removing the filter caused rapid contamination!
EUV Source
Multilayer Mask
Multilayer mirror
Filter (Si, SiZr, MgF2, Open)
Reflectivity of Mo/Si mirror
Rapid contamination from exposure without a filterPhoto of carbon contamination on mask 1 cm
The problem with out-of-band (OOB) radiation• There is vacuum ultraviolet radiation from the EUV plasma sources (Sn,
Xe, Li, …)• Mo/Si multilayers reflect the radiation
• The absorption depth of ~ 60 nm radiation is shorter than for EUV radiation in relevant materials– This leads to higher secondary electron emission for OOB radiation – and
may contribute to increased rate of optics contamination– This leads to higher direct photoabsorption in the carbon containing
molecules – and may contribute to increased rate of optics contamination
Photoabsorption and Secondary Electron Yield
The wavelength band from 40 nm to 80 nm yields:• higher photoabsorption cross-section for C, H, and O than for EUV• higher secondary electron yield from carbon or Ru layer than for EUV
C9H20
Absorption data from www.cxro.lbl.gov
EUV Source
4.7˚
22.74˚
slit
Glass
Filter
Mask Plane
Bellow
Sample for EUV Exp.
25
22
Collector
Bellow
Bellows100mm
Redesigned system with custom flat field spectrometer for 22‐124 nm
Spectrometer has been installed to measure out-of-band contamination
Mask Contamination and Topography
Reticle SEM was used to inspect the mask before and after the contamination
Observed larger CD after contamination on the mask
Contaminated with designed aperture
1mm
~ 20 nm Carbon contamination
Before contamination
CD=152.6±1.3 nm
100nm
CD=176.6±1.7 nm
100nm
After contamination
Contamination adds to surface roughness
For this mask, multilayer RMS roughness measured with AFM was 0.37 nmAfter 20 nm of carbon contamination, roughness was .60 nm
225 nm lines with 900 nm spaces(45 nm lines at wafer plane)
Printing Results of contaminated mask
Larger dose increase due to contamination than would be predicted by simple carbon absorption
10% of 40nm target CD
Dose (mJ/cm^2)
CD
(nm
) ~ 20 nm of carbon
P. Naulleau et al., “Status of EUV micro-exposure capabilities at the ALS using the 0.3-NA MET optic,” Proc. SPIE 5374, pp. 881–891, 2004
Printing results from Berkeley MET
Contamination topography affects required dose
Panoramic software simulation compared to the experimental data
Conformal deposition requires more dose to achieve target CD than direct deposition
CD compensation failure due to carbon contamination
Can compensate for shadowed features on clean mask
However, after enough contamination, the dose for different orientations diverges
This might be a bigger problem than the simple loss of throughput due to contamination
Conclusions• Photoresists outgas many molecules that can be detected
• Will the detected molecules contaminate optics when exposed for thousands of hours
• Are we missing any molecules (too large AMU), or fragments that could contaminate the optics?
• Out-of-band radiation contaminates more rapidly than EUV
• Spectral filtering can solve this problem
• Mask contamination topography is important
• Dose required to print is larger than simple carbon absorption
• CD compensation may fail due to carbon contamination
Acknowledgments• College of Nanoscale Science and Engineering
– Alin Antohe– Yu-Jen Fan– Rashi Garg– Chimaobi Mbanaso– Petros Thomas– Leonid Yankulin– Robert Brainard
• SEMATECH– Frank Goodwin– Vibhu Jindal– Andrea Wüest– Warren Montgomery