Computer modeling of contamination and cleaning of EUV source optics
D.Astakhov1,2, Y. Mankelevich3, D. Lisitsin1, I. Popov2, V. Krivtsun1,3, V.Medvedev1,3, A. Yakunin5, M. van Kampen5, P. van Zwol5 , D.Labestki5
1 RnD-ISAN/EUV Labs, Sirenevy Bulevard Str. 1, Troitsk, Moscow, 108840, Russia
2 ISTEQ BV, High-Tech Campus 9, Eindhoven, The Netherlands
3 Institute for Spectroscopy RAS (ISAN), Fizicheskaya str. 5, Troitsk, Moscow 108840, Russia
4 Skobel’tsyn Institute of Nuclear Physics, Lomonosov Moscow State University, Leninskie gory, Moscow 119991, Russia
5 ASML Netherlands B.V., De Run 6501, 5504DR Veldhoven, The Netherlands
Motivation for flow model
● Collector mirror need to be
protected from tin debris
● Two main approaches:
− Gas flow
− Gas flow + B-field
● Goal: fast transient 3D model
for flow + plasma conditions
Test configuration of LPP EUV source chamber
Region for test model ●
●
Transient simulation for test configuration
Tin vapour reach collector. This simulation resulted in net removal of tin for the most part of the collector. (see slide 16)
Included physics
● Gas heating due to ions stopping
● Spectrally resolved WUV radiation absorption in gas
● Reflection from collector, EUV to IF
● H2 dissociation, recombination
● Multi component diffusion
● Particulate debris tracking
● Set of surface & volume chemical reactions
● Tin deposition to surfaces
● Tin cleaning by atomic hydrogen
● Disturbance of the tin droplets trajectory by flow field
Relevant time & space scales hierarchy
(1) Laser pulse & hot tin plasma Not resolved,
● t < 500 ns, h < 1mm
input data
(2) Ions stopping & radiation absorption Partially resolved,
t ~ 1 μs, h ~ chamber size
●
time scale
integrated out
(3) pulse-to-pulse repetition time
● t ~ 10 μs
(4) Transport processes in chamber
● t > 10 ms, h ~ chamber size
Transient CFD
(5) Tin cleaning and deposition on the collector mirror
● t > 1 hour
Model approach: CFD + plasma source
EUV plasma characteristics
● Ions: angular - energy distribution + charge states
● Radiation: angular resolved spectrum
Plasma energy-mass-momentum source for CFD
● Particle tracing for ions
● Ray tracing for radiation
Firing pattern
Computational fluid dynamics for multi-component gas
•Mesh accounts for complex vessel geometry
Tin cleaning model
● Tin cleaning agent
○ Atomic hydrogen vs ions
○ Ratio of ion and radical fluxes to the collector
○ Energy distribution of ion flux
● Tin cleaning product and redeposition
○ SnH4 vs SnHx
Tin cleaning agent
Tin etching by H radicals and by H2 plasma was observed in many
experiments:
● D. J. W. Klunder et al., SPIE 5715 (2005)
….
● O.V. Braginsky et al., J. Appl. Phys. 111 (2012)
○ Sn is etched by atoms, probability is → 0.5e-5 SnH4 per At.H
○ Sn may be etched also by ions, but with low overall contribution 13 years
○ Ion energy below 20eV.
● D. Ugur et al., Chemical Physics Letters 552, 122 (2012).
○ Sn is etched by atoms, probability is → 1e-5 SnH4 per At.H
….
● D. T. Elg et al Plasma Chem Plasma Process 38, 223 (2018).
○ Sn is effectively etched by H ions, limiting factor is energy
Tin cleaning agent
Tin etching by H radicals and by H2 plasma was observed in many
experiments: O. V. Braginsky et al Journal of Applied Physics 111, 093304 (2012).
● D. J. W. Klunder et al., SPIE 5715 (2005)
….
● O.V. Braginsky et al., J. Appl. Phys. 111 (2012)
○ Sn is etched by atoms, probability is ~ 0.5e-5 per At.H
○ Sn may be etched also by ions, but with low overall contribution
○ Ion energy below 20eV.
● D. Ugur et al., Chemical Physics Letters 552, 122 (2012).
○ Sn is etched by atoms, probability is ~ 1e-5 per At.H
….
Sn cleaning by ions can be 10 -- 100x more
● D. T. Elg et al Plasma Chem Plasma Process 38, 223 (2018). efficient then by atoms, but still have small
○ Sn is effectively etched by H ions, limiting factor is energy contribution in these experiments.
In order to translate to the EUV source chamber model:
● Expected ratio of ions to radicals fluxes
● Expected energy distribution function of the ion flux on the
collector mirror
Estimation of ions to radical ratio for fluxes on the collector
Recombination of ions:
• H2+, H3+ lost due to two-body dissociative recombination in volume
• Sn+n to Sn++ lost charge due to charge exchange with H2 flux ratio Hions/(At.H + Hions)
• Sn++ , Sn+ and H+ recombine via collisional radiative process
Recombination of H radicals:
• Atomic hydrogen recombination in volume is a very slow
• Recombination on the surface is a main sync
Estimation energy distribution function (EDF) of the ion flux to the collector mirror
● EUV radiation forms plasma in front of the collector
mirror
● Escape of fast photo electrons to the walls facilitate
of plasma sheath formation
● Ions are accelerated by the sheath potential
○ Once plasma near the mirror cools down
→ ion energy drops to near gas temperature
level
Way to estimate -- Particle-in-Cell model for the EUV
induced plasma applied locally near the collector mirror.
● Te cools down quickly to almost room temperature
due to collisions with gas
● Most ions impact surface with very low energy. IEDF over 1us, after 1us T_e is about 1000K
Most ions are cold → At.H provides most of tin cleaning
Tin cleaning products and redeposition
Common assumptions are:
●Tin etching is Sn + 4H → SnH4
●Redeposition is due to SnH4
○ SnH decomposition is EUV or plasma induced
4
○ Thermal decomposition rate is too low (Tamaru 1956)
●Direct measurements of SnH4 decomposition yields no effect
For other systems it is known that intermediate products can be much more
reactive, e.g. CH3 vs CH4; SiH3 vs SiH4 etc.
● Proposed set of reaction:
○ Sn + xH → SnHx
○ SnHx + surface → redeposition
○ SnHx + H + M → SnH4
Tin cleaning model key experiment → filament + lateral flow
experiment by Piter van Zwolle and Maarten van Kampen
H2 flow tin covered sample
● Tin is pre-deposited on the sample
○ initial profile is uniform
hot filament
● Spatial distribution of tin over sample
after exposure show measurable
redeposition
○ The amount of redeposition is
inconsistent with rate for SnH4
decomposition
1D analytical example: effect of tin redeposition
Z
● For H radicals and Sn the surface is a “sync term”
SnHx
At.H, Sn ○ Sn deposits with g ~ 1
○ H radicals recombine with g ~ 1e-4 .. 0.1
L ● For SnHx the surface with tin is a source
○ SnHx is produced due to etching
○ Need to be transported any by diffusion and flow
○ Efficiently redeposits back on Sn
Balance between contamination and cleaning
● In 1D approximation, with diffusive transport only:
Assumption:
● Tin covers only part of the surface,
PAC = S_covered/S_total
● fluxes of H and Sn are constant
● For PAC → 0 redeposition is not important
● For PAC ~ 1, redeposition is important, but the transport of
SnHx limit cleaning rate
● Pressure effect → increase of pressure limits the transport,
thus reduce cleaning rate
Modelling of collector contamination/cleaning
● Full model needs to define H, Sn, SnHx fluxes and
test simulation results:
deposited tin on all surfaces in the source chamber
● In order to provide quantitative results the amount of
tin on the surface should be an output
○ Dirty surface → increase of H recombination.
Steady model frequently results in all dirty
collector.
● Model approach: transient fully coupled model for
flow + source terms + species transport. Iteratively
solved to prescribed tolerance.
○ If flow is steady → solve species transport only
○ If flow is unsteady, e.g. start of pulses → can
resolve with small time step.
● Internal scheme is unconditionally stable:
○ dt ~ 1e-6s to resolve pulses
○ dt ~ 1e-4s to resolve flow restructure
○ dt ~ 100s to resolve cleaning/deposition
Net cleaning of the collector can be realised
Conclusions
• We have developed 3D transient model that couples energy and
momentum input from tin plasma to the flow in the EUV source
chamber
• The model takes into account tin deposition and cleaning from surfaces.
Main etch product is assumed to be chemically active SnHx
• The model have ability to smoothly vary time step from pulse-to-pulse (~1e-6s)
resolution to characteristic times (~100s) of cleaning processes
• The model can be used to optimize the chamber geometry, flow
structure etc. for regime during source operation
