TEM Sample Thinning With ROI-Guided FIB Curtaining Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The preparation of ultra-thin TEM samples for semiconductor devices is time-consuming and prone to defects due to thickness variations and 'curtaining' issues, which limits the quality of structural characterization and analysis, especially for samples below 30 nm thickness.
Innovation Solution
A method involving cleaning the sample to remove redeposition layers, identifying the region of interest, and using a focused ion beam to remove material in specific orientations to achieve uniform thinning, including a 'clip lamella roof' step to create a flat surface perpendicular to the ion beam, ensuring precise thickness and minimizing curtaining artifacts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If traditional manual lift-out procedures are used to prepare TEM samples, then sample quality can be maintained, but the preparation time becomes excessively long (45-90 minutes per sample)
Solution Approach 1:
The preparation process is divided into distinct automated stages: initial milling to create a thick lamella, extraction of the lamella from the substrate, thinning to final thickness, and cleaning. Each stage is performed automatically by the FIB system with programmed parameters, eliminating manual intervention while maintaining quality standards.
Solution Approach 2:
Manual mechanical manipulation of the probe and sample is replaced with automated computer-controlled positioning and manipulation systems. The FIB system automatically controls probe movement, milling parameters, and sample positioning, substituting human skill with programmable automation.
2Manufacturing precision
If lamellae are thinned below 100 nm thickness to achieve better TEM imaging, then imaging quality improves, but the process becomes increasingly difficult to control and automate
Solution Approach 1:
A thick lamella (several micrometers) is first created as a preliminary structure that is easier to handle and extract. This thick intermediate structure provides mechanical stability during extraction, and only after successful extraction is it thinned to the final ultra-thin dimensions, separating the difficult extraction task from the precision thinning task.
Solution Approach 2:
The system uses real-time feedback from imaging modes and automated thickness monitoring to control the thinning process. The FIB system continuously adjusts milling parameters based on feedback signals, maintaining precise control over thickness reduction to the target range of 30-100 nm.
3Productivity
If automated FIB milling is used to increase throughput, then preparation speed improves, but curtaining artifacts and thickness non-uniformity increase
Solution Approach 1:
The automated thinning process uses periodic pulsing of the ion beam rather than continuous milling. This allows brief intervals for heat dissipation and reduces cumulative thermal effects that cause warping and non-uniform thinning, while maintaining high overall throughput through automated sequencing.
Solution Approach 2:
The system dynamically changes milling parameters during the process, including ion beam current, pulse duration, and scanning patterns. Parameters are adjusted based on real-time feedback and the specific stage of thinning, allowing optimized control over material removal rates to maintain uniformity while achieving high speed.
4Adaptability or versatility
If multiple samples need to be analyzed for complete semiconductor wafer characterization, then analysis completeness improves, but total processing time becomes prohibitive (hours to days)
Solution Approach 1:
The FIB system is designed to perform multiple functions in sequence without requiring sample removal or repositioning: it can mill multiple lamellae from different locations on the same substrate, extract each lamella, thin it to the required thickness, and clean it. This multi-functionality allows complete wafer characterization to be performed on a single substrate in one automated session.
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 significantly reduces the time required for sample preparation, enhances the quality of ultra-thin TEM samples by preventing defects and ensuring accurate thickness, thereby improving the robustness and throughput of the process.
Implementation Method 1
milling away material with an ion beam to create trenches on either side of the region of interest
Implementation Method 2
cleaning the sample to remove a redeposition layer
Data Source
AI summary
A method of preparing a sample for transmission electron microscopy (TEM) analysis is provided. The method comprises cleaning the sample to remove a redeposition layer, imaging the cleaned sample and identifying a location of a region of interest within the sample, and removing material from the sample, based on the identified location of the region of interest within the sample. Advantageously, the sample thinning step is performed based on a detected location of a region of interest. This thinning step involves removal of uneven surfaces (the “lamella roof”) and thinning the remaining bulk substrate to remove redundant material, so that the silicon substrate volume between the surface of the sample and the region of interest has a defined thickness.