알림
뒤로
mobile background

IRIS sensor

WINTEL'S PRODUCTS


IRIS (Intra-chamber Radio-frequency Inspection Sensor) Definition
A sensor that measures the assembly state of semiconductor process chambers

NEEDS · Equipment Supplier


• Pre-assembly tolerance management per individual component (traditional method)


• Post-assembly error management — cannot fully represent complex structures


• No chamber-to-chamber assembly tolerance management (cannot evaluate chamber-level quality equivalence)


• Harder to establish management criteria as chamber structures become more complex


NEEDS · Mass Production Fab.


• No way to verify identical reassembly state during equipment setup


• No way to confirm identical recovery state before/after PM


• Limited ability to observe long-term drift of specific parts during PM cycle


SOLUTION · IRIS Sensor 


• Measures spectral data representing chamber assembly state


• Expresses assembly tolerance via similarity index → enables evaluation of installation and recovery state at the chamber level


• Analyzes long-term change by spectral band → enables real-time long-term change evaluation of specific parts during operation


• Expresses identity qualitatively and quantitatively via spectrum comparison


Chamber assembly
Chamber assembly
Spectral information on internal space dimensions
Spectral information on internal space dimensions
Real-time long-term change analysis and tracking of space (= component dimensions)
Real-time long-term change analysis and tracking of space (= component dimensions)
IRIS Sensor Functions and Benefits
A sensor that evaluates consistency, repeatability, and recoverability of semiconductor process chambers
✅ Examples of Actual Applied Functions

📡 

Monitoring (FDC)


- Tracks long-term change of contamination during PM cycle

- Tracks long-term change of consumable parts during PM cycle

- Verifies repeatability of process gap

- Verifies repeatability of per-wafer process

- Verifies wafer-less sub-process repeatability (base check, burn-in check)

- Verifies chamber state before/after remote plasma cleaning


🔬

Diagnostics


- Real-time measurement of process gap (scaled gap)

- Real-time measurement of TES height (scaled height)

- Measurement of stabilization time after PM

- Measurement of recovery rate after PM

- Provides Tool to Tool Matching comparison data




🎯

Prediction


- Predicts maintenance timing 






✅ Expected Benefits 

Chip Manufacturing Fab.


• Improves productivity and reduces maintenance costs by tracking long-term changes

• Component-level management tool via chamber physical state diagnostics

• Tool to Tool Matching evaluation across PM cycles, Process Modules, and Systems


Equipment Supplier


• Verification of Tool to Tool Matching before and after equipment delivery

• Component-level improvement and development tool for mass production environments


IRIS Sensor Measurement Principle and Signal Processing
Analyzes electromagnetic resonance characteristics formed inside the chamber
Measurement


- Scans 3–14 GHz frequency band

- Records S₁₁ (reflectance) signal

- Stores spectral data periodically (~2 sec intervals)

- Active measurement — measures during non-process periods (~1mW)

- Structural changes reflected as resonance peak changes in spectrum

Signal Processing


- Treats each S₁₁ spectrum as a state data unit

- Defines initial state data

- Calculates similarity between initial and current state data

- Observes time-series data of similarity




Analysis


- Observes similarity changes in response bands linked to specific components

- Classifies synchronized state data as needed

- Observes long-term changes (PM cycle unit analysis)

- Observes short-term changes (wafer to wafer, Lot to Lot)


✅ 1D Resonance vs. 3D Resonance

The 3D chamber interior spectrum expresses complex resonance characteristics as components interlink and change

→ This is why specialized analysis algorithms and programs are required

1D Resonance
1D Resonance
3D Resonance
3D Resonance

Comparison: OES handles 1D optical resonance spectra at the atomic/molecular level → Peak interpretation is intuitive

IRIS Sensor Measurement Principle and Signal Processing – Resonance Inside the Chamber
f = 1.882GHz, simulated by CNU
f = 1.882GHz, simulated by CNU

▪ Scans frequency while measuring the chamber's reflectance spectrum

▪ S₁₁ (reflectance) is the ratio of reflected power to incident forward power

▪ Injects wavelengths (= frequencies) at the scale of the internal chamber structures



𝑆11=10 log10 PrΤPf Pr : Reflected Power, Pf : Forward power

✅ Examples of Actual Signals
Same model, same process: PM1 vs. PM2
Same model, same process: PM1 vs. PM2
Different model, different process: PM1
Different model, different process: PM1
IRIS Sensor Measurement Principle and Signal Processing – Time-Series State Data & Similarity

Similarity

A metric for evaluating the degree of identity between different state data

Baseline data 

• Initial state data for calculating long-term and short-term time-series similarity

• e.g.) Stabilized state data after PM

• e.g.) Stabilized idle state data from one of PM1, PM2, PM3

Current data

• Current state data

• fingerprint of current chamber state

• e.g.) Idle state, run state, or state data observed during PM

Similarity Time-Series Graph vs. Time Axis (blue line)

• Time-series data summarizing chamber state into a single similarity value

• Calculated using full or partial frequency range state data as needed

Sync.

• Points with identical chamber H/W state

• e.g.) Chuck Home Position

Response band

• Frequency range reflecting changes of a specific component (parts, process variables) 

Reference data (Library)

• State data for a specific variable combination

• e.g.) State data at Home position, process position

• e.g.) State data at a specific step of the process recipe

3D time-series representation of state data recorded over 4 days before and after PM work
3D time-series representation of state data recorded over 4 days before and after PM work
IRIS Sensor Measurement Principle and Signal Processing – Time-Series Similarity
✅ Time-Series Similarity = Summarized History of Chamber State
Model A, Chamber 1: 30-day time-series similarity after PM
Model A, Chamber 1: 30-day time-series similarity after PM
Model A, Chamber 2: 30-day time-series similarity after PM
Model A, Chamber 2: 30-day time-series similarity after PM
Model B: 30-day time-series similarity after PM
Model B: 30-day time-series similarity after PM
By analyzing the trend of time-series patterns, it is possible to evaluate the consistency, repeatability, and recoverability of the chamber
Early post-PM, idle
Early post-PM, idle
Early post-PM, run
Early post-PM, run
Late post-PM, run
Late post-PM, run
Before and after PM
Before and after PM
IRIS Sensor Measurement Principle and Signal Processing – Analysis Graphs
Possible to isolate, observe, and manage factors that impair chamber consistency, repeatability, and recoverability
✅ Individual Analysis Graph = Statistical value of time-series similarity filtered from a specific perspective


🔶 BASIC : Synced monitoring 

(stand-by state: median)

run time, hrs
run time, hrs

🔶 BASIC : Synced monitoring 

(stand-by state: range)

run time, hrs
run time, hrs

🔶 BASIC : Long-term prediction 

(To PM point: stand-by state trend)

run time, hrs
run time, hrs

🔶 CUSTOMIZED : Synced diagnostics 

(TES height: response band)

run time, hrs
run time, hrs
IRIS Sensor Measurement Principle and Signal Processing – Installation Reproducibility & Sensor Precision
IRIS Installation Reproducibility vs. Viewport Installation Reproducibility vs. Reproducibility After Maintenance
✅ IRIS Installation Reproducibility vs. Viewport Installation Reproducibility vs. Reproducibility After Maintenance


Viewport adaptor
Viewport adaptor
Only IRIS re-installations
Only IRIS re-installations
After viewport adaptor re-installations
After viewport adaptor re-installations
After multiple maintenances
(2 process modules, wet PM/regular PM)
After multiple maintenances
(2 process modules, wet PM/regular PM)
✅ IRIS Precision, Measurement Error
Discriminability for 5μm step movement in process equipment where process gap precision is critical
Discriminability for 5μm step movement in process equipment where process gap precision is critical
Continuous measurement error for the same process equipment
Continuous measurement error for the same process equipment
IRIS Sensor Configuration, Specifications & Installation Environment

Sensor Unit


Components

  • Adaptor flange / Antenna / Network Analyzer / Case
  • L-type, I-type form factor options

Functions

  • Generation / radiation / reception of electromagnetic signals
  • Operating temperature control

Specifications

  • Output: ~1mW
  • Frequency range: 3–14 GHz
  • Min. bandwidth: 3 MHz
  • Max. resolution: 11,000 pts/full scan
  • Operating temp.: <50°C (sensor internal)

Control Console


Components

  • PC / Monitor / Keyboard / Mouse / Stand / Cables

Functions

  • Runs measurement & analysis programs
  • Data storage
  • Sensor power supply
  • Sensor data communication

Specifications

  • Max. sensor connections: 3 EA
  • Sensor power: 5V, <3A

Measurement Software


License

  • 3 sensors/copy

Functions

  • Measurement variable control
  • Real-time measurement data storage
  • Basic similarity time-series display
  • Stat. data storage

Specifications

  • Min. scan interval: ~2.5 sec/full scan
  • Max. connected sensors: 3 sensors/copy

Analysis Software


Auto Analysis (for Mass Production Fab.)

  • 3 sensors/copy
  • Synchronized monitoring, diagnostics & predictive analysis
  • Response frequency band search
  • State classification (idle, run, PM)
  • Analysis result (stat. data) storage
  • FDC-linked data generation
  • Real-time analysis linked with measurement software
  • Post-analysis of stored large-scale data

Manual Analysis (for Research)

  • 1 sensor/copy
  • Basic time-series similarity display
  • Selective file-unit analysis
  • Post-analysis of stored data

INSTALLATION ENVIRONMENT (RECOMMENDED)


• Metal-shielded structure forming an enclosed space  

• Chamber equipped with a viewport that allows electromagnetic waves to pass through 

• Chamber where mesh, holes, slits, etc. do not excessively interfere with electromagnetic wave transmission

IRIS Auto Analysis Program (IRIS Data Analysis ver. 2)
✅ MP-CSB Integrated Module


Main view
Main view
Analysis view
Analysis view
T3M view
T3M view


Standard Features

  • • Equipment stand-by state monitoring (degrade-recovery trend) 
  • • PM timing prediction 
  • • Stabilization time and recovery rate calculation after PM 
  • • Auto-recognition of PM work period and auto-reset for new PM cycle 
  • • Tool to Tool Matching comparison data 
  • • Real-time operating rate information

Customizable Features

  • • Build observation and diagnostic functions for physical chamber components with significant impact on equipment stability (Process gap, TES, Edge ring etch amount, etc.)

    • Build sub-recipe observation functions for equipment stability management (base pressure check, auto burn-in check, etc.)

    • Perform solution optimization as mini-projects for other customer pain points

Use Scenario 1: Basic Single-Chamber Type for Mass Production Fab.
✅ When comprehensive analysis — Monitoring, Diagnostics, Prediction — within the PM (Preventive Maintenance) cycle is important


Use Scenario 2: Multi-Chamber Linked Type for Mass Production Fab.
✅ When PM is frequent and quality control before/after PM is important — when Tool to Tool Matching is critical (using T3M table)


PM timing differs per sensor. Time-series history within PM cycle also differs.
PM timing differs per sensor. Time-series history within PM cycle also differs.

A: Similarity difference values between long-term baseline data (example)

- Uses baseline data calculated after each PM as the representative value for the PM cycle

- When the reference point is changed, relative difference values with the remaining points are displayed

- Can quantitatively verify how much the recovered chamber state differs per PM cycle


B: Operating rate, stabilization time, recovery rate history