Compressive Scanning Electron Microscopy for Reduced Beam Damage

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current scanning transmission electron microscopy (STEM) techniques face inefficiencies in data acquisition time and electron beam dose due to redundant information collection and limitations in signal-to-noise ratios, leading to suboptimal throughput and radiation damage to samples.

Innovation Solution

The implementation of compressive sensing with rapid modulation of the electron beam intensity or scan speed and path, using a system with a fast beam-current modulator and mathematical reconstruction techniques to minimize data acquisition time and electron exposure, while utilizing arbitrary illumination patterns to reduce radiation damage.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional STEM diffraction scanning is performed to obtain high-resolution structural information, then measurement precision is improved, but loss of time increases significantly as each acquisition may take a significant fraction of a second or even multiple seconds

Engineering Contradiction:
Improvehigh-resolution structural informationVSAvoidacquisition time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent applies compressive sensing theory to acquire data in advance through undersampled measurements, capturing only the most informative projections during the scan. This preliminary data collection is then mathematically reconstructed to produce high-resolution images, eliminating the need for exhaustive scanning of all pixel locations and thereby dramatically reducing acquisition time while maintaining measurement precision

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent extracts and retains only the essential information from the data acquisition process by using compressive sensing to identify and measure only the most informative projections. This extraction approach discards redundant measurements while preserving the critical structural information needed for high-resolution imaging, thus reducing total acquisition time without sacrificing precision

Inventive Principle:
Principle #2Taking out (Extraction)

2Loss of information

If conventional STEM diffraction scanning is performed to obtain comprehensive data, then information return is improved, but loss of time increases as total scan may take hours

Engineering Contradiction:
Improveinformation returnVSAvoidtotal scan time
Core Design Contradiction:
Loss of informationVSLoss of time

Solution Approach 1:

The patent performs preliminary data acquisition using compressive sensing that captures the essential information content through undersampled measurements. The mathematical reconstruction process then recovers the complete information from these partial measurements, enabling comprehensive data retrieval in a fraction of the time required for conventional exhaustive scanning

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent changes the acquisition parameters by using compressive sensing theory to determine optimal undersampling rates and projection angles. This parameter optimization allows the system to acquire sufficient information with far fewer measurements than conventional methods, dramatically reducing total scan time while maintaining complete information return through advanced reconstruction algorithms

Inventive Principle:
Principle #35Parameter changes

3Productivity

If HAADF-STEM scanning is performed to achieve fast acquisition, then productivity is improved, but measurement precision deteriorates as vastly less information is retrieved from each electron

Engineering Contradiction:
Improvescan speedVSAvoidinformation per electron
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The patent combines the fast scanning capability of HAADF-STEM with compressive sensing to acquire data preliminarily at high speed through undersampled measurements. The mathematical reconstruction then recovers the detailed information that would otherwise require slower, more comprehensive scanning, thereby maintaining high productivity while improving the information extracted from each electron

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent optimizes the acquisition parameters by using compressive sensing to determine the optimal balance between scan speed and information content. This allows the system to operate at high speeds like HAADF-STEM while recovering complete information through reconstruction, effectively resolving the trade-off between productivity and measurement precision

Inventive Principle:
Principle #35Parameter changes

4Ease of operation

If conventional STEM scanning is performed with regular 2D array pattern, then ease of operation is improved, but productivity deteriorates as much of the information obtained is redundant

Engineering Contradiction:
Improveautomated scanningVSAvoiddata throughput
Core Design Contradiction:
Ease of operationVSProductivity

Solution Approach 1:

The patent maintains the ease of automated operation by implementing compressive sensing within the existing scanning framework. The system automatically selects optimal undersampled projections and performs mathematical reconstruction without requiring manual intervention, thereby preserving ease of operation while dramatically improving productivity by eliminating redundant information collection

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent changes the scanning parameters from a regular 2D array pattern to an optimized undersampled projection pattern determined by compressive sensing theory. This parameter change maintains automated operation simplicity while maximizing data throughput by acquiring only the most informative measurements and using mathematical reconstruction to recover complete information

Inventive Principle:
Principle #35Parameter changes

5Measurement precision

If electron beam dose is increased to improve signal-to-noise ratio, then measurement precision is improved, but object-affected harmful factors increase as beam damage to sample becomes the resolution limitation

Engineering Contradiction:
Improvesignal-to-noise ratioVSAvoidbeam damage to sample
Core Design Contradiction:
Measurement precisionVSObject-affected harmful factors

Solution Approach 1:

The patent applies compressive sensing to acquire data preliminarily through undersampled measurements that require lower electron beam doses. The mathematical reconstruction process then recovers high-quality images from these low-dose measurements, achieving good signal-to-noise ratios without subjecting the sample to high beam doses that would cause damage and limit resolution

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent extracts and retains only the essential information through compressive sensing measurements that are optimized to minimize the number of electrons required. This extraction approach achieves adequate signal-to-noise ratios with reduced beam doses, thereby improving measurement precision while reducing beam damage to the sample

Inventive Principle:
Principle #2Taking out (Extraction)

Data Source

PatentEP3061118B1System and method for compressive scanning electron microscopy
Publication Date: 2021.09.29 LAWRENCE LIVERMORE NAT SECURITY LLC
  • EP3061118B1 patent drawingFigure 1
  • EP3061118B1 patent drawingFigure 2

AI summary

A scanning transmission electron microscopy (STEM) system is disclosed. The system may make use of an electron beam scanning system configured to generate a plurality of electron beam scans over substantially an entire sample, with each scan varying in electron-illumination intensity over a course of the scan. A signal acquisition system may be used for obtaining at least one of an image, a diffraction pattern, or a spectrum from the scans, the image, diffraction pattern, or spectrum representing only information from at least one of a select subplurality or linear combination of all pixel locations comprising the image. A dataset may be produced from the information. A subsystem may be used for mathematically analyzing the dataset to predict actual information that would have been produced by each pixel location of the image.