Semiconductor analysis at the atomic scale

In semiconductor manufacturing, advanced challenges demand more than just high-resolution imaging. (S)TEM instruments offer a suite of powerful techniques that enable precise characterization at the atomic scale. These tools are now essential in addressing process control, material analysis, and device optimization in leading-edge semiconductor labs.


Precise metrology of semiconductor devices

S/TEM metrology has become a necessary element of all leading-edge wafer fabrication workflows, since highly specific measurements are needed to generate statistically relevant data with sub-Angstrom accuracy. This data allows manufacturers to calibrate toolsets, diagnose failure mechanisms, and optimize overall process yield.

 

We offer metrology workflows that are flexible enough to enable manual, semi-automated, or even fully automated solutions, depending on your needs. The Thermo Scientific Metrios 6 (S)TEM is designed to provide the fast, precise measurements that semiconductor manufacturers need to develop and control their wafer fabrication processes.  

The Metrios 6 (S)TEM provides automated TEM imaging and metrology on gate-all-around structures.

Energy-dispersive X-ray spectroscopy (EDS) for qualitative or quantitative composition analysis

Energy dispersive Xray spectroscopy (EDS) integrated into Thermo Scientific (S)TEMs adds essential compositional insight to high-resolution imaging. The electron beam excites core-shell electrons in the specimen, producing characteristic X-rays that are detected and converted into energy spectra. These spectra yield rapid, quantitative elemental maps that help semiconductor developers assess material composition, track contamination, and validate process uniformity with high spatial precision.

EDS map of gate-all-around transistor

Electron energy-loss spectroscopy (EELS) for compositional and chemical information

In semiconductor development, electron energy-loss spectroscopy (EELS) is a powerful tool for atomic-scale chemical and structural analysis. EELS enables detailed mapping of elemental composition, oxidation states, and electronic structure, critical for understanding complex device architectures and material interfaces. With high-energy resolution and sub-Angstrom spatial precision, EELS helps engineers detect nanoscale variations in dielectric layers, analyze bandgap properties, and monitor changes from processing or aging. These insights support failure analysis, material selection, and performance optimization in the next generation of semiconductor nodes.

Nitrogen mapping in a GaN device with EELS (left to right, GaN, SiN, TiN, all).

4D STEM

4D STEM works by scanning an electron probe across a sample and recording a diffraction pattern at each pixel, resulting in a 4D dataset, two dimensions for scanning and a two-dimensional diffraction pattern. This technique can be used for mapping crystal grain orientation and strain distribution by measuring local lattice distortions. Electric and magnetic fields indicated by shifts in the transmitted electron beam's phase at each scan point can give insights to dopant regions when combining 4D STEM and DPC capabilities.

Diffraction patterns are used in generating 4D STEM data.

Differential phase contrast (DPC)

Our DPC technology enables high-resolution mapping of electric and magnetic fields by detecting subtle deflections of the electron beam within a sample. This allows for quantitative visualization of charge distribution, electric potential gradients, and interface fields, critical for analyzing semiconductor devices. With low-dose imaging options like iDPC, it also supports high-contrast analysis of beam-sensitive materials and ultra-thin structures common in advanced semiconductor nodes.

Arrows highlight the transition between the n-type and p-type doped regions of the p-n junction.

Strain analysis for semiconductor devices

Strain engineering is critical to improving carrier mobility, performance, and reliability in advanced semiconductor devices. Thermo Scientific TEM solutions provide precise, nanoscale strain analysis to support process development and device optimization.

Strain distribution shown along the (110) direction.

Geometric phase analysis (GPA)—Extracts quantitative strain from high-resolution TEM or S/TEM images, revealing local lattice distortions in transistors, gate stacks, and heterostructures.

Nanobeam and convergent beam electron diffraction (NBED/CBED)—Quantifies strain with nanometer precision, ideal for assessing epitaxial layers and channel regions.

Precession electron diffraction (PED)—Reduces dynamical effects for accurate, large-area strain mapping in complex 3D architectures.

4D STEM—Captures diffraction patterns at every probe position to map strain and rotation fields across complete device cross-sections.


Process automation for high-throughput semiconductor TEM analysis

In semiconductor metrology and failure analysis, process automation enhances the efficiency, consistency, and throughput of TEM workflows. Automated routines handle tasks such as feature detection, navigation, focusing, and data acquisition, minimizing user variation and accelerating time-to-result. Integrated automation enables seamless switching between imaging and analytical modes like S/TEM, 4D STEM, EELS, and EDS, ensuring reproducible data across multiple lamellae and devices. This results in faster, more reliable nanoscale characterization that supports yield improvement and process optimization in advanced semiconductor manufacturing.

Metrios Smart Automation is used to identify features for automated metrology.

Automated navigation and feature recognition—AI-driven algorithms identify and target specific device features such as GAA nanowires, FinFETs, and 3D NAND structures without manual intervention.

Consistent image and data acquisition—Automated focusing, alignment, and drift correction ensure reproducible imaging and analytical results across multiple lamellae and sessions.

Integrated multi-modal operation—Seamlessly combines S/TEM, TEM, EELS, and EDS in unified workflows, enabling correlated structural and compositional analysis.

Recipe-based and recipe-free modes—Allows both standardized process monitoring and flexible, adaptive exploration depending on analysis needs.

Throughput and reliability—Reduces operator variability, shortens time-to-data, and supports statistically meaningful process control and failure analysis.

For Research Use Only. Not for use in diagnostic procedures.