Semiconductor Plasma Ignition Control for Reflected RF Power

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Variations in processing chamber conditions cause unpredictable reflections of Radio Frequency (RF) power during plasma ignition in semiconductor manufacturing, leading to defects, longer ignition times, and potential damage to the RF source.

Innovation Solution

A trained model is used to predict and optimize the amount of RF power reflected back to the RF source by adjusting input parameters, such as RF source settings and tuning capacitors, to minimize and stabilize the reflected power.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If standard RF power parameters are used for plasma ignition, then plasma can be ignited, but the reflected RF power varies unpredictably causing defects and potential RF source damage

Engineering Contradiction:
Improveplasma ignition consistencyVSAvoidreflected RF power damage
Core Design Contradiction:
ReliabilityVSObject-affected harmful factors

Solution Approach 1:

The system performs preliminary measurements of reflected RF power during the plasma ignition process before the plasma is fully established. By measuring the reflected power trajectory in advance, the system can predict potential damage conditions and adjust parameters proactively to prevent RF source damage and substrate defects.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system continuously monitors reflected RF power during plasma ignition and uses this feedback to dynamically adjust RF power parameters. The control system modifies the RF power waveform in real-time based on the measured reflected power, creating a closed-loop control system that maintains plasma ignition consistency while preventing excessive reflected power from damaging the RF source.

Inventive Principle:
Principle #23Feedback

2Reliability

If RF power is increased to ensure plasma ignition, then plasma can be ignited more reliably, but the reflected RF power increases causing more damage and defects

Engineering Contradiction:
Improveplasma ignition reliabilityVSAvoidreflected RF power magnitude
Core Design Contradiction:
ReliabilityVSObject-generated harmful factors

Solution Approach 1:

The system dynamically adjusts RF power parameters during the plasma ignition process rather than using fixed parameters. By making the RF power waveform adaptive and time-dependent, the system can provide sufficient power for reliable ignition while minimizing reflected power at each moment of the ignition sequence, thus preventing damage without sacrificing ignition reliability.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system changes multiple RF power parameters including voltage, current, frequency, and waveform shape to optimize the ignition process. By adjusting these parameters based on real-time reflected power measurements, the system achieves reliable plasma ignition while maintaining reflected power within safe limits, preventing both damage and defects.

Inventive Principle:
Principle #35Parameter changes

3Manufacturing precision

If RF power parameters are tightly controlled for uniform energy distribution, then energy is provided uniformly to the plasma, but variations in chamber conditions cause unpredictable plasma ignition timing

Engineering Contradiction:
Improveenergy distribution uniformityVSAvoidplasma ignition adaptability
Core Design Contradiction:
Manufacturing precisionVSAdaptability or versatility

Solution Approach 1:

The system performs preliminary measurements of chamber conditions and reflected power characteristics before initiating plasma ignition. By characterizing the chamber state in advance, the system can predict the optimal ignition timing and parameter adjustments needed to achieve uniform energy distribution while adapting to variations in chamber conditions between different substrate processes.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system uses real-time feedback from reflected power measurements to adjust RF power parameters during ignition. This feedback mechanism allows the system to maintain uniform energy distribution to the plasma while adapting to varying chamber conditions, ensuring consistent ignition timing across different substrate processes despite tight control on energy distribution parameters.

Inventive Principle:
Principle #23Feedback

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

The solution reduces the magnitude and variability of reflected RF power, improving plasma ignition consistency and reducing defect rates in semiconductor substrates while protecting the RF source.

Implementation Method 1

Plasmas used in the processing chamber may be ignited by providing radio frequency (RF) energy to the chamber from an RF source

Methodology Applied
Scientific EffectRadio Frequency (RF) energy: Electromagnetic Induction

Implementation Method 2

electro-magnetic energy is applied to at least one precursor gas or precursor vapor to transform the precursor into a reactive plasma

Methodology Applied
Scientific EffectPlasma generation: Plasma

Implementation Method 3

Plasma-enhanced chemical vapor deposition (PECVD) process is a chemical process wherein electro-magnetic energy is applied to at least one precursor gas or precursor vapor to transform the precursor into a reactive plasma

Methodology Applied
Scientific EffectPlasma-enhanced chemical vapor deposition: Plasma Enhanced Chemical Vapour Deposition

Data Source

PatentUS11894217B2Plasma ignition optimization in semiconductor processing chambers
Publication Date: 2024.02.06 APPLIED MATERIALS INC
  • US11894217B2 patent drawing
  • US11894217B2 patent drawing
  • US11894217B2 patent drawing

AI summary

A method of reducing reflected Radio Frequency (RF) power in substrate processing chambers may include accessing input parameters for a processing chamber that are derived from a recipe to perform a process on a substrate. The input parameters may be provided to a model that has been trained using previous input parameters and corresponding sensor measurements for the chamber. A predicted amount of reflected RF power may be received from the model and it may be determined whether the predicted reflected RF power is optimized. The input parameters may be repeatedly adjusted and processed by the model until input parameter values are found that optimize the reflected RF power. Optimized input parameters may then be provided to the chamber to process the substrate.