AI summary by TwinMind

China’s “New” EUV Light Source

Asianometry

15m 58s

Huawei EUV Announcement Background

  • Social media reports Huawei testing China domestically developed EUV machine
  • Uses laser induced discharge plasma (LDP) EUV light source
  • Contrasts with ASML's laser produced plasma (LPP) method
  • Claims LDP is more efficient, smaller, simpler, better energy efficiency than LPP
  • Limited information available on machine specifications

EUV Light Source Fundamentals

  • Target: 13.5 nanometer light production
  • ASML's LPP process:
    • CO2 laser hits tin droplet twice
    • First pulse flattens droplet to disc shape
    • Second powerful pulse creates tin plasma
    • Plasma emits 13.5 NM light
  • Optical system challenges:
    • 6-10 mirror reflections required
    • Each mirror reflects <70% of EUV light
    • Requires high photon generation (measured in Watts)
    • Initial goal: 250 Watts minimum for economic viability

Historical EUV Light Source Development

  • Early 2000s: LPP had serious problems
    • Unstable CO2 laser hitting microscopic tin target from meters away
    • Community considered this extremely difficult
  • Alternative explored: Discharge Produced Plasma (DPP)
    • Strong electrical current discharge between electrodes
    • Creates "lightning bolt" effect
    • Vaporized fuel pre-ionized between electrodes
    • Powerful capacitor bank provides discharge
    • Current flows along z-axis creating magnetic field
    • Field constricts particles creating plasma
    • Temperatures up to 200,000°C
    • Called "Z pinch" - same tech investigated for nuclear fusion in 1950s

Early DPP Research Organizations

  • Extreme Ultraviolet Lithography System Development Association (EUVLA) - Japan
    • Large joint organization with Japanese companies/academics
  • Philips Extreme UV - joint venture
    • Philips + Fraunhofer Institute for Laser Technology
  • Fuel options explored: Xenon, tin, lithium
    • EUVLA studied Xenon DPP
    • Philips championed tin DPP
    • Lithium rejected due to diffusion/poisoning issues

DPP Performance Characteristics

  • Xenon fuel DPP sources used in early prototypes
    • Micro exposure tool (2004)
    • Intel and Sematech R&D labs
  • Tin emerged as best candidate:
    • More ions produce light at 13.5 nanometers
    • Less light absorption by own ions
    • 2% wall plug power conversion vs 0.5-0.9% for Xenon
  • DPP advantages: Simple method, early shipping (2003)
  • Major limitation: Repetition rate scaling
    • Thousands of discharges/second needed for 120 wafers/hour
    • Thermal damage concerns for electrodes
    • Tin vapor condensation on cold surfaces

Evolution to LDP Technology

  • Mid-2000s: Philips Extreme UV evolved DPP → LDP
  • LDP setup features:
    • Two rotating tin supply discs as electrodes
    • Connected to large energy storage capacitors
    • Discs rotate through temperature-controlled liquid tin baths
    • Pulsed CO2 or neodymium YAG laser fires at tin film
    • Creates tin cloud "pre-plasma" between electrodes
    • Electric discharge generates Z pinch for hot dense plasma
  • LDP advantages over LPP:
    • Guaranteed laser-tin contact (vs missing droplets in LPP)
    • Heat mitigation through liquid tin bath cooling
    • Foil trap for debris mitigation

Commercial LDP Development

  • 2005: Japan's USHIO bought 50% of Philips Extreme UV
  • 2008: USHIO purchased remaining 50%, renamed to Extreme Technologies
  • 2010: Revealed tin DPP Source Collector Module (SCM)
    • 20,000 amp discharges
    • Lightning strikes lasting few hundred nanoseconds
    • 10,000-100,000 discharges per second
    • Adjustable pulse repetition rate and energy
  • ASML initially used LDP in first Alpha demo tool
  • ASML monitored both methodologies:
    • LDP through Extreme Technologies
    • LPP through Cymer and Gigaphoton startups

LDP Performance Limitations

  • 2008: Extreme announced 500 watts EUV power at plasma
  • Power measurements:
    • "At plasma": power at light source stage
    • "At intermediate focus": power reaching wafer after reflections
  • SCM performance with standard collector: 14 watts at intermediate focus = 7-8 wafers/hour
  • With improved collector: 34 watts (still below competitors)
  • Competitor performance:
    • Cymer: 250 watts at intermediate focus
    • Gigaphoton: 104 watts at intermediate focus
  • To achieve 500 watts at intermediate focus:
    • Would need 4,000 watts at plasma
    • Requires 50,000 pulses/second at 80mJ per pulse
    • Possibly too much heat generation to manage
  • Physical limitations observed:
    • Higher power input created larger (not brighter) plasma
    • Larger plasma exceeded collector mirror capture capability
    • Energy waste similar to Xenon plasma studies

ASML's Technology Choice

  • ASML switched from LDP to LPP
  • 2011: USHIO achieved 30 watts at intermediate focus (up from 14 watts)
  • By then ASML had committed to LPP path
  • Community continued debating choice through 2014
  • USHIO continued LDP for mask blank inspection tools
  • 2022: USHIO tools showed ~250 watts at plasma (likely local maximum)
  • LPP scaling success:
    • ASML San Diego team hit 740 watts EUV power
    • Expected productization soon

Current Chinese EUV Development

  • China pursuing multiple EUV light source methods:
    • LPP, free electron lasers, laser assisted discharge plasma
  • 2023: Research from Harbin Institute of Technology
    • School of Aeronautics and Astronautics
    • Small research cluster working on DPP/LDP since early 2010s
    • Multiple Chinese organizations investigating LDP plasma dynamics
  • Published research appears unimpressive:
    • Lower conversion efficiencies than 2005-2006 setups
    • Likely hiding actual capabilities

Technology Assessment and Predictions

  • China will eventually achieve EUV capability
    • Massive funding and talent allocation
    • Benefit of knowing technology works
    • Access to published research on "golden path"
    • Conceptual knowledge represents ~60% of development
  • Initial Chinese EUV machines likely:
    • Not economically competitive with ASML initially
    • <100 wafers/hour throughput
    • Worse resolution than ASML
    • May be ridiculed as "boondoggle" or "toy"
  • Potential innovations:
    • Clever twists to overcome brightness limitations
    • Novel two-mirror optic systems
    • ASML should study and consider integrating innovations
  • Chinese/Asian advantage: Tolerance for years of unprofitable operations
  • ASML's current gross margins: ~51%
  • Margins may need reduction to compete with Chinese challenge

Alternative EUV Research

  • Russian academics exploring 11.2 nanometer EUV light
  • Using ruthenium-beryllium mirrors
  • Unlikely to scale but demonstrates continued innovation
  • Acknowledgment that early researchers didn't have all answers

TwinMind

Get summaries like this for your own meetings

TwinMind takes notes in the background during meetings, lectures, and calls, then writes the summary for you.

Free plan to get started

No bot joins your calls

Phone, desktop & Chrome apps

Your next meeting, summarized like this

TwinMind listens in the background and turns meetings, lectures, and calls into searchable notes you can chat with. Start free.

Available on phone, desktop & Chrome

Your next meeting, summarized like this

TwinMind listens in the background and turns meetings, lectures, and calls into searchable notes you can chat with. Start free.

Available on phone, desktop & Chrome

Your next meeting, summarized like this

TwinMind listens in the background and turns meetings, lectures, and calls into searchable notes you can chat with. Start free.

Available on phone, desktop & Chrome