Transistor Threshold Voltage Variation for Random Number Generation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Random number generators face instability due to minimal threshold voltage mismatch between transistors, leading to unpredictable multi-bit outputs, especially under environmental changes.
Innovation Solution
The design method and structure for transistors with exacerbated random dopant fluctuation (RDF) through complementary doping and lateral dopant non-uniformity, widening the threshold voltage variation range, are implemented to enhance the detectability of threshold voltage mismatch between pairs of transistors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If transistors are designed to be essentially identical for random number generation, then the device complexity is reduced and manufacturing is simplified, but the threshold voltage mismatch becomes minimal making detection difficult and output unstable
Solution Approach 1:
The patent applies parameter changes by modifying the doping concentration parameters in the transistor design. Specifically, it introduces complementary doping with different concentration levels in adjacent regions (e.g., n-type and p-type dopants at different concentrations in source/drain regions or channel areas). This parameter modification increases the random dopant fluctuation effect, thereby widening the threshold voltage variation range and making threshold voltage mismatch detectable while maintaining transistor functional equivalence
Solution Approach 2:
The patent implements local quality by creating spatial variations in dopant concentration within the transistor structure. Different regions of the same transistor are doped with different types and concentrations of dopants (e.g., higher concentration in one region, lower in another). This local non-uniformity increases RDF and threshold voltage variation without affecting the overall transistor performance, enabling reliable threshold voltage mismatch detection for random number generation
2Reliability
If complementary doping is used to increase random dopant fluctuation and widen threshold voltage variation range, then threshold voltage mismatch detectability improves, but the manufacturing precision requirements increase
Solution Approach 1:
The patent applies partial or excessive action by intentionally introducing dopant concentrations that exceed standard design specifications. By using higher than normal dopant concentrations and broader concentration variations in complementary doping, the design ensures sufficient random dopant fluctuation effect even with manufacturing tolerances. This excessive doping approach guarantees threshold voltage variation wide enough for reliable mismatch detection while remaining compatible with standard manufacturing capabilities
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
This approach stabilizes the multi-bit output by increasing the likelihood of detecting threshold voltage mismatch, making the output more repeatable and reliable.
Implementation Method 1
exacerbated random dopant fluctuation (RDF)... is achieved through the use of complementary doping in one or more of the transistor components
Implementation Method 2
through lateral dopant non-uniformity within the transistor and, particularly, between the channel region and any halo regions
Data Source
AI summary
Disclosed are a design method and structure for a transistor having a relatively large threshold voltage (Vt) variation range due to exacerbated random dopant fluctuation (RDF). Exacerbated RDF and, thereby a relatively large Vt variation range, is achieved through the use of complementary doping in one or more transistor components and/or through lateral dopant non-uniformity between the channel region and any halo regions. Also disclosed are a design method and structure for a random number generator, which incorporates multiple pairs of essentially identical transistors having such a large Vt variation and which relies on Vt mismatch in pairs of those the transistors to generate a multi-bit output (e.g., a unique identifier for a chip or a secret key). By widening the Vt variation range of the transistors in the random number generator, detecting Vt mismatch between transistors becomes more likely and the resulting multi-bit output will be more stable.


