Random Telegraph Noise Time Constant Extraction Method
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current technologies face challenges in accurately determining and characterizing random telegraph noise (RTN) time constants and magnitudes in semiconductor devices, which are crucial for modeling and simulating device performance as technology nodes shrink.
Innovation Solution
A method and system that measure parameters at various sampling frequencies to identify stable states, determine time constants, and compute final values by analyzing the relationship between time constants and sampling frequencies, allowing for the extraction of RTN time constants and magnitudes from semiconductor devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple sampling frequencies are used to measure device parameters, then measurement precision of RTN time constant is improved, but measurement time and complexity increase
Solution Approach 1:
The method performs preliminary measurements at multiple sampling frequencies to establish the relationship between time constants and sampling frequencies before determining the final RTN parameters. This preliminary action at different frequencies allows the system to pre-characterize the device behavior, enabling accurate final parameter extraction without requiring excessively long measurement times at a single frequency.
Solution Approach 2:
The measurement approach dynamically adjusts the sampling frequency based on the device's RTN characteristics. By varying the sampling frequency and observing how the extracted time constants change, the method adapts to capture the true RTN behavior. This dynamic measurement strategy allows the system to identify stable time constant values that are independent of sampling frequency, improving precision without linearly increasing total measurement time.
2Measurement precision
If multiple sampling frequencies are used to measure device parameters, then measurement precision of RTN time constant is improved, but device complexity of the measurement system increases
Solution Approach 1:
The measurement system is designed with multi-functionality to handle multiple sampling frequencies using the same core measurement infrastructure. The analyzer and processing system can operate across different frequency regimes without requiring separate dedicated hardware for each frequency, reducing overall system complexity while maintaining the ability to perform precise RTN characterization.
Solution Approach 2:
The method changes the sampling frequency parameter systematically to extract RTN characteristics. By varying this single measurement parameter and analyzing how the extracted time constants respond, the system achieves high measurement precision without needing to add complex hardware. The approach relies on parameter variation and computational analysis rather than additional measurement equipment.
3Measurement precision
If derivative computation and threshold comparison are used to determine final time constant, then accuracy of RTN parameter extraction is improved, but computational complexity increases
Solution Approach 1:
The method uses feedback through derivative computation and threshold comparison to iteratively refine the RTN parameter extraction. By computing the derivative of time constants with respect to sampling frequency and comparing against a threshold, the system receives feedback on whether the extracted parameters have converged to stable values. This feedback mechanism ensures high extraction accuracy by automatically identifying when the true RTN parameters have been found, without requiring overly complex computational algorithms.
Data Source
AI summary
A method includes, for a device and at each of a plurality of sampling frequencies, measuring a parameter of the device to generate a plurality of signals and each signal represents a time series of values of the parameter. Further, the method includes for each of the plurality of signals, identifying at least two stable states in the time series for each signal. Still further, the method includes determining a time constant associated with each stable state of each signal, determining a relationship between the time constants of each stable state and sampling frequency, and computing a final time constant value for each steady state by computing a derivative of the time constants as a function of sampling frequency relationship and comparing the derivative to a threshold.