Transmission electron microscopy (TEM) imaging

About the author

Teemu Myllymäki | Materials Scientist, Co-founder & CEO, Measurlabs

Teemu holds an M.Sc. in Organic Chemistry and a Ph.D. in Applied Physics, with doctoral research centered on biomimetic material design using supramolecular chemistry and molecular self-assembly. During his PhD, transmission electron microscopy was a core part of his day-to-day research. Since then, he has continued working hands-on with electron microscopes as part of Measurlabs' analytical services.

Author profile at Measurlabs · LinkedIn

Why we wrote this

Textbooks cover the physics of TEM well. Instrument manuals explain the operational controls. Neither captures the judgement calls, the recurring pitfalls, or the interpretive nuance that comes from running samples day after day.

At Measurlabs, TEM imaging is part of our routine analytical work. We handle a broad range of customer projects: semiconductor failure analysis, battery material characterization, catalyst development, MOF structural studies. Patterns emerge from that work that do not appear in the literature.

This article is written for students entering the workforce with theoretical knowledge but limited hands-on experience. It is also written for practising scientists who understand the technique conceptually but have not yet operated the instrument themselves. We combine graduate-level scientific rigour with practical insight drawn from real customer projects. If this text ends up cited in a master's thesis or a journal article's methods section, we consider that a success.

What is TEM?

Transmission electron microscopy achieves sub-ångström spatial resolution by exploiting the wave nature of electrons. The de Broglie wavelength of accelerated electrons is several orders of magnitude smaller than that of visible light, which bypasses the diffraction limits of optical microscopy and allows direct visualisation of atomic arrangements, crystal lattices, and microstructural defects.

The architecture of a transmission electron microscope is conceptually similar to an inverted optical microscope. A high-energy electron beam replaces the light source, and electromagnetic lenses replace glass optics. The beam is generated by an electron gun at the top of the column, accelerated to kinetic energies of 80 to 300 keV, and focused by a condenser lens system onto an ultra-thin specimen. The transmitted electrons are focused by the objective lens to form a magnified image, then further magnified by projector lenses before reaching the detector. Magnification ranges from 50x to over 1,500,000x.

How TEM differs from SEM and optical microscopy

In optical microscopy, diffraction limits resolution to roughly 200 nm. In SEM, a focused beam scans the surface and generates topographical and compositional signals from an interaction volume extending micrometres into the bulk. In TEM, the beam passes through an ultra-thin specimen, and spatial and phase information encoded in the transmitted electrons reconstructs a projected image of the sample's internal structure at atomic resolution.

TEM is the right tool for resolving internal microstructures, grain boundaries, crystal defects, phase boundaries, and atomic-scale compositional gradients. It cannot directly measure surface topography or analyse bulk specimens without extensive preparation. Choosing between TEM, SEM, and AFM is primarily a question of what information is needed, and this is covered in the comparative section below.

What TEM can and cannot tell you

TEM provides atomic-resolution images of crystal structures and defects, diffraction patterns for phase identification and orientation mapping, elemental maps at sub-nanometre resolution via EDX and EELS, and local strain field measurements. It does not provide quantitative three-dimensional topography, bulk compositional averages, or analysis of specimens that cannot be thinned to electron transparency. Sample preparation is the limiting factor far more often than the instrument itself.

The physics of electron-matter interaction

Elastic and inelastic scattering

When the electron beam strikes the specimen, electrons undergo elastic and inelastic scattering. Elastic scattering deflects electrons without energy loss, preserves phase information, and provides the basis for diffraction contrast and high-resolution phase-contrast imaging. Inelastic scattering transfers energy to the specimen through phonon excitation, plasmon excitation, or inner-shell ionisation. It carries chemical information exploited by EELS and EDX, but also introduces noise and chromatic aberration into the image.

Bright-field and dark-field imaging

In bright-field TEM, the directly transmitted beam forms the image. Thicker regions and heavier elements appear darker. Objective apertures in the back focal plane of the objective lens selectively block high-angle diffracted electrons, enhancing mass-thickness or diffraction contrast. In dark-field TEM, the direct beam is blocked and only diffracted electrons form the image. Crystalline phases and specific orientations appear bright against a dark background.

Diffraction and phase contrast

When the beam encounters a crystalline specimen, interference between electron waves scattered from successive lattice planes produces selected-area electron diffraction (SAED) patterns. These patterns are the primary tool for phase identification, orientation mapping, and lattice parameter measurement. At high resolution, phase contrast imaging is formed by the interference between the direct and multiple diffracted beams, providing direct visualization of atomic columns and lattice spacings. This requires a thin, well-oriented specimen and precise defocus control.

Spherical and chromatic aberration

Electromagnetic lenses are affected by two primary optical defects. Spherical aberration arises because electrons traversing the lens periphery are refracted more than those near the centre. This produces a focal spread rather than a singular focal point and limits point-to-point spatial resolution. Chromatic aberration arises because electrons with slightly different kinetic energies focus at different planes along the optical axis. Inelastic scattering within the specimen worsens the energy spread and amplifies this effect.

Aberration correctors address both defects. These computer-controlled multipole assemblies use hexapole or octupole lenses to introduce negative spherical aberration that cancels the positive spherical aberration of the objective lens. Modern corrected instruments achieve spatial resolutions below 50 pm and energy resolutions below 5 meV.

Instrument components and their functions

The electron gun

The electron gun generates the beam via thermionic emission, using a heated tungsten filament or LaB6 rod, or via cold-field emission (CFEG), which applies an intense electric field to a sharp tungsten tip. CFEG sources offer higher beam brightness, better spatial coherence, and a lower energy spread. For high-resolution work, CFEG is the better choice.

The condenser lens system

The dual condenser lens system, together with condenser apertures, controls the illumination spot size and convergence angle. The convergence angle is a critical parameter. Too large introduces aberrations; too small sacrifices probe intensity.

The objective lens and apertures

The objective lens forms the primary magnified image. Its aberrations dominate the resolution limit of the entire instrument. Objective apertures in the back focal plane control which electrons contribute to the image.

Intermediate and projector lenses

These lenses magnify the image formed by the objective lens before it reaches the detector.

Detectors

The image is recorded by a CCD, CMOS sensor, or direct electron detector. Direct electron detectors, and specifically event-based architectures that perform electron counting in hardware via FPGAs, offer the highest sensitivity, lowest noise, and fastest readout speeds. These are discussed in the detector section below.

Comparative metrology: TEM, SEM, and AFM

The choice of characterization technique depends on the spatial resolution required, the properties of interest, and the constraints of the sample.

SEM scans a focused beam across the surface, generating secondary electrons, backscattered electrons, and characteristic X-rays from an interaction volume extending into the bulk. It provides topographical and compositional maps with lateral resolution typically in the 1 to 10 nm range. It is well suited for screening large areas and imaging samples with significant surface relief. It does not provide quantitative vertical metrology or internal structural information.

AFM uses a sharp cantilever tip to mechanically probe the surface and generates a true three-dimensional topographical map with sub-nanometre vertical resolution. It operates in ambient air, liquid, or controlled gas environments without vacuum or coating requirements. Beyond topography, AFM provides data on mechanical stiffness, friction, and localised electrical or magnetic properties.

SEM is the right choice for rapid imaging of large areas or surface relief. AFM is the right choice for quantitative surface topography and for samples that cannot tolerate vacuum or electron beam exposure. TEM is required for internal structural analysis at atomic resolution. Using TEM when SEM would suffice is a waste of preparation time and instrument time. For more information, read our in depth comparison of TEM and SEM.

Sample preparation

The analytical quality of TEM is bounded by the structural integrity of the sample. The technique requires electron transmission, so the region of interest must be thinned to below 100 nm. A specimen that is too thick will produce data dominated by chromatic aberration and multiple scattering, not by the material's true structure.

Sample preparation is where most TEM projects succeed or fail. A poorly prepared sample produces uninterpretable data regardless of instrument quality. The choice of method depends on the ductility, conductivity, hardness, and crystalline structure of the material.

Electropolishing for metals and conductive alloys

For conductive metals and alloys, twin-jet electropolishing produces electron-transparent foils with no mechanical damage. A 3 mm metallic disc is mechanically ground and dimpled, then mounted in a twin-jet electro-polisher as the anode. Two cathode nozzles direct jets of temperature-controlled electrolyte against the disc while a direct current is applied.

The process requires operating in the plateau region of the current-voltage polarisation curve (Region 2), where a viscous layer forms evenly across the surface and produces smooth, uniform anodic dissolution. Low voltages (Region 1) preferentially attack grain boundaries and create etching artefacts. High voltages (Region 3) cause oxygen evolution, pitting, and surface relief. Electropolishing yields a pristine foil with zero mechanical damage, but uses corrosive chemicals and can alter the surface chemistry of reactive alloys.

Focused ion beam (FIB) milling for site-specific extraction

When a specific grain boundary, catalytic precipitate, or transistor gate oxide must be isolated, focused ion beam (FIB) milling in a dual-beam FIB-SEM system is the standard approach. The workflow proceeds as follows:

1. The region of interest is identified via the SEM column.

2. A protective layer of platinum, tungsten, or carbon is deposited over the site via in situ gas injection.

3. A gallium ion beam mills away the surrounding bulk material in a step-trench pattern, isolating a lamella attached to the bulk by a small bridge.

4. A nanomanipulator probe is welded to the lamella using ion-beam-deposited platinum, the bridge is severed, and the lamella is transferred to a TEM copper grid.

5. The mounted lamella is thinned from both sides using progressively lower-energy ion beams until it reaches electron transparency at 30 nm.

FIB preparation introduces artefacts that must be actively managed. Gallium ion bombardment induces structural amorphisation along the cut walls, drives gallium implantation into the specimen lattice, and causes micro-cracking from thermal or mechanical stress. Low-voltage argon ion milling and plasma cleaning are not optional finishing steps. They are necessary to remove amorphous damage layers and hydrocarbon contamination before imaging.

The method allows cross-sectional TEM imaging.

Mechanical preparation and ultramicrotomy

For hard or brittle materials, tripod polishing grinds samples into precise wedge geometries. Dimpling reduces the centre of a disc to near-transparency while leaving the outer rim intact for handling.

For soft materials and biological tissues, ultramicrotomy uses a diamond knife to section resin-embedded samples into ultra-thin slices. Grinding, crushing, and carbon replication extraction are used for isolated nanoparticles and fine powders.

Imaging modes and techniques

Conventional TEM

Bright-field and dark-field imaging is the starting point for most TEM sessions. It gives a rapid overview of microstructure, phase distribution, and defect density before more advanced modes are deployed.

High-resolution TEM (HRTEM)

HRTEM exploits phase contrast to image individual atomic columns and crystal lattice fringes. Correct interpretation gives direct visualisation of atomic arrangements, dislocation cores, interface structures, and point defects. Misinterpretation is a genuine risk. Phase contrast images are not straightforward structural maps; they are sensitive to defocus, specimen thickness, and orientation, and require careful comparison with simulations to extract reliable structural information.

Scanning TEM (STEM) and HAADF-STEM

In scanning transmission electron microscopy (STEM), the optics form the beam into an atomic-scale probe that scans point-by-point across the specimen. High-angle annular dark-field (HAADF) detectors capture Rutherford-scattered electrons, generating images where contrast scales with atomic number. This gives direct visual discrimination of heavy versus light elements, known as Z-contrast imaging.

Four-dimensional STEM (4D-STEM)

In 4D-STEM, a complete convergent-beam electron diffraction (CBED) pattern is recorded at every scan coordinate, generating a four-dimensional data cube. This enables post-acquisition computation of virtual bright-field, dark-field, or SAED images from any region within the scan. It also enables orientation mapping, electric and magnetic field mapping via Differential Phase Contrast, and strain distribution mapping from sub-pixel lattice distortions. These datasets are often terabytes in size. Open-source platforms such as LiberTEM and py4DSTEM handle the processing, increasingly with AI-assisted feature extraction.

Electron ptychography

Ptychography is a computational phase-retrieval technique built on 4D-STEM data. By recording overlapping CBED patterns as the probe scans, iterative algorithms reconstruct both the phase and amplitude of the electron wave through the specimen. This circumvents the numerical-aperture limitations of physical magnetic lenses and enables resolution in the deep-sub-ångström regime at a fraction of the electron dose required by conventional HAADF-STEM. The low-dose capability makes ptychography the technique of choice for beam-sensitive materials such as zeolites, MOFs, and supramolecular crystals.

Energy-dispersive X-ray spectroscopy (EDX)

When the beam ejects inner-shell electrons from specimen atoms, characteristic X-rays are emitted as higher-energy electrons fill the vacancies. STEM-EDX detectors capture these X-rays to generate elemental maps, allowing visualization of chemical distributions such as the core-shell structure of catalyst nanoparticles.

Electron energy-loss spectroscopy (EELS)

EELS measures the energy lost by transmitted electrons during inelastic scattering. Different elemental bonds and electronic states absorb specific quantized amounts of energy, making EELS a nanoscale spectrometer. Combined with monochromated electron sources, it provides data on local bonding states, oxidation states, and plasmonic responses with energy resolutions below 5 meV. EELS is significantly more sensitive than EDX for light elements such as lithium, carbon, nitrogen, and oxygen. For comprehensive characterisation, both techniques are used together on the same specimen area.

Cryo-TEM

To image hydrated or water dispersed samples, such as liposome dispersions and powders in their natural state, the sample need to be vitrified, i.e. frozen, fast. The vitrification prevents ice crystals from forming during the freezing process. The ice crystals would cause artifacts in the image and could cause damage to the specimen. The vitrification is done by dipping the sample in liquid nitrogen-cooled propane. The reason propane is used is that as liquid nitrogen boils it causes a lot of bubbles disturbing the vitrification. Also the low heat capacity of nitrogen compared to propane causes the sample to freeze slower compared to propane. The operator can benefit both the lack of boiling, higher heat capacity and the good availability of liquid nitrogen by immersing a propane vessel in liquid nitrogen for vitrification. The vitrified sample is then transferred in cryo-TEM for imaging.

Image acquisition, interpretation, and artifacts

Acquisition parameters

Accelerating voltage involves a genuine trade-off. Higher voltages improve resolution and penetration but increase knock-on radiation damage. For beam-sensitive organics and biological samples, 80 keV is the right choice even at some resolution cost.

Common artifacts

Artifact recognition is the most important practical skill in TEM analysis. Artefacts are not rare. They are encountered routinely, and the failure mode is mistaking them for real material features.

Preparation artifacts. FIB-induced amorphization produces a surface layer that is easily mistaken for a genuine amorphous phase. Gallium implantation produces spurious EDX signals. Mechanical grinding introduces dislocation damage. Low-voltage Ar-ion milling and plasma cleaning are the standard remedies.

Contamination. Carbon from airborne hydrocarbons deposits under the beam, forming a dark growing spot that obscures fine detail and distorts EELS signals. Plasma cleaning before loading is necessary, not optional.

Diffraction artifacts. When a crystalline grain rotates into a strong Bragg condition, image contrast changes sharply and can be mistaken for a composition gradient or phase boundary. Systematic tilting and diffraction pattern cross-checking is required to distinguish genuine contrast from orientation effects.

Fresnel fringes. Bright or dark bands appear at interfaces and edges due to phase contrast from abrupt potential changes. Their position shifts with defocus. They are a diagnostic tool, but are easily misread as structural features by those unfamiliar with them.

Beam damage. Radiolytic damage destroys organic bonds. Knock-on displacement damages crystalline lattices. Thermal effects can sinter nanoparticles or induce phase transformations. Structural changes that occur during imaging are real changes to the material, not the pre-existing structure. The sample must be monitored throughout acquisition.

Contrast interpretation pitfalls

HAADF-STEM images are often described as directly interpretable because heavier elements appear brighter. This holds for well-separated atomic columns in simple structures. In thicker specimens, dynamical diffraction effects produce non-monotonic contrast even in HAADF images. Column-by-column composition assignment in multi-component alloys and oxides requires comparison with image simulations.

Data analysis and quantification

The primary software platform for TEM data acquisition and processing is Gatan Microscopy Suite (DigitalMicrograph). ImageJ/Fiji is widely used for lattice spacing measurements, particle size distributions, and layer thickness quantification. For 4D-STEM datasets, LiberTEM and py4DSTEM provide GPU-accelerated processing and orientation mapping.

Standard quantitative outputs include lattice spacings from HRTEM fringe patterns, layer thicknesses at interfaces, nanoparticle size distributions, grain size and orientation statistics from SAED or 4D-STEM, elemental compositions from EDX or EELS, and strain fields from 4D-STEM diffraction disk displacement. Strain mapping from 4D-STEM has become a routine measurement in semiconductor device characterisation.

Applications

Metallurgy and grain boundary engineering

Macroscopic properties such as tensile strength and ductility are governed by nanoscale defect interactions, specifically the interactions between dislocations and grain boundaries as described by the Hall-Petch relationship. Dynamic in situ TEM deformation provides direct observation of these events.

When dislocations impinge on a grain boundary, four outcomes are possible. The dislocation is transferred directly into the adjacent grain, absorbed into the boundary as an extrinsic grain boundary dislocation, accommodated at the boundary while simultaneously emitting a new dislocation into the adjacent grain, or repelled back into the source grain. The outcome is determined by the geometric angle between slip planes across the boundary, the resolved shear stress on the adjoining slip systems, and the magnitude of the residual Burgers vector left at the boundary.

In thermoelectric SnTe, TEM has shown that dislocations initially trapped within grains can be driven to migrate and aggregate into dense arrays at grain boundaries. This microstructural arrangement reduces charge carrier scattering while scattering phonons, improving thermoelectric efficiency. The implication is clear: for thermoelectric materials, grain boundary engineering is not a secondary consideration but a primary design variable.

Semiconductor industry

The electronics industry and nanotechnology laboratories use TEM to examine thin film materials. The method is commonly applied to identify imperfections, failures, and (with EDX or EELS) impurities. TEM is also used for electron diffraction to determine the crystal structure of solid samples and reveal crystallographic orientations with high accuracy.

Gate oxide breakdown is a primary failure mechanism in FinFET logic architectures. Degradation of ultra-thin dielectric layers is caused by Negative Bias Temperature Instability (NBTI), where energetic holes sever hydrogen bonds at the silicon-dielectric interface. As hydrogen diffuses away, trap states accumulate and eventually cause localized leakage currents and oxide rupture.

Isolating a single defective sub-10 nm transistor within a logic array requires Passive Voltage Contrast (PVC), Active Voltage Contrast (AVC), Emission Microscopy (EMMI), or Electron Beam Absorbance Current (EBAC) before FIB-SEM extracts a targeted lamella. TEM analysis of these lamellae reveals defects invisible to other metrology tools: residual metal filaments shorting adjacent lines, contact gouging penetrating the drain-bulk junction, or CMP slurry-induced micro-scratches in Shallow Trench Isolation (STI) regions.

Lithium battery development

The viability of lithium-metal batteries is limited by the instability of the Solid Electrolyte Interphase (SEI) and the nucleation of lithium dendrites. Both metallic lithium and the SEI components are reactive in air and sensitive to electron beam damage, making conventional TEM inadequate for this application.

Cryo-TEM has resolved the role of SEI nanostructure. Two morphologies exist: a mosaic nanostructure with randomly scattered crystalline domains in an amorphous matrix, and a multilayer nanostructure. Lithium dissolution occurs preferentially through the crystalline regions of the mosaic SEI, seeding dendrite nucleation. The ordered multilayer SEI supports uniform lithium stripping and plating, suppresses dendrite formation, and reduces irreversible lithium loss per cycle by a factor of three. The practical implication is that SEI nanostructure, not just chemical composition, determines battery performance. This shifts the design focus towards engineering artificial SEI layers that replicate the multilayer morphology via targeted electrolyte additives.

Catalysis and nanoparticle sintering

Metal nanoparticles dispersed on oxide supports degrade at high operating temperatures via sintering, which reduces catalytically active surface area. In situ gas-phase TEM using windowed membrane flow cells allows observation of nanoparticle evolution at pressures up to 1 bar and above, directly relevant to industrial reaction conditions.

Real-time imaging distinguishes between two sintering mechanisms. In Ostwald Ripening, surface atoms detach, migrate, and absorb into larger particles. In Particle Migration and Coalescence, intact nanoparticles slide across the support and fuse. Identifying which mechanism dominates determines the correct strategy for improving sintering resistance.

Tracking palladium catalysts on MgO supports reveals that Pd transitions from metal nanoparticles to surface-dispersed cations and then to subsurface embedded cations under reaction conditions. Surface-located Pd cations exhibit ethylene hydrogenation activity over 500 times higher than subsurface cations. This result directly identifies the target structural state for catalyst design.

Biology and life sciences

Cryo-TEM is required for biological macromolecules, viruses, and other hydrated specimens. Rapid vitrification in liquid ethane freezes the solvent into an amorphous glass before crystalline ice forms, preserving the sample's native state. Cryogenic temperatures reduce radiolytic and thermal damage during imaging. Cryo-TEM is the primary technique for single-particle analysis, microcrystal electron diffraction (MicroED), and cryo-electron tomography (cryo-ET) in structural biology.

MOFs, 2D materials, and beam-sensitive frameworks

The organic coordination bonds in MOFs and COFs undergo rapid radiolytic damage, and structural collapse occurs before sufficient image contrast is accumulated under conventional STEM conditions. Electron ptychography is the solution: it achieves deep-sub-ångström resolution at a fraction of the dose required by HAADF-STEM. Integrated Differential Phase-Contrast (iDPC-STEM) is effective for imaging light elements such as carbon, oxygen, and nitrogen without destroying the framework.

For transition metal diiodides, which degrade within seconds of air exposure, encapsulation between graphene sheets using an inorganic stamp transfer approach extends the stable sample lifetime from seconds to months. This enabled the first atomic-resolution visualisation of their intrinsic defect dynamics and edge structures.

Limitations and when not to use TEM

Understanding the limitations of TEM is as important as understanding its capabilities. Choosing TEM for the wrong problem wastes significant preparation time, instrument time, and budget.

Sample preparation is destructive. Thinning to below 100 nm is irreversible. For unique or precious samples, the information gained must be weighed against sample consumption before committing to TEM analysis.

TEM provides a 2D projection. Overlapping features produce misleading contrast. True three-dimensional characterisation requires electron tomography, a series of images at many tilt angles followed by computational reconstruction. This is substantially more time-intensive than standard TEM analysis.

The field of view is small. A TEM analysis examines a localised area. Whether that area represents the bulk cannot be answered by TEM alone. For heterogeneous materials, TEM must be complemented by bulk characterisation techniques.

TEM is not suited for bulk compositional analysis. EDX and EELS provide local elemental information, not bulk averages. For overall composition, XRF or ICP-MS are the appropriate techniques.

High-vacuum requirements exclude liquid and biological samples unless cryo-TEM protocols are used. Cryo-TEM adds substantial complexity and cost.

Beam damage constrains the analysis of organic materials, polymers, biological specimens, and MOFs. Low accelerating voltage and low-dose acquisition reduce but do not eliminate this.

Use SEM instead when rapid screening of large areas, surface topography, or compositional mapping across macroscopic features is needed.

Use AFM instead when quantitative three-dimensional surface topography is needed, particularly for samples that cannot tolerate vacuum or electron beam exposure.

In situ and cryogenic platforms

Standard TEM observes a thin foil in high vacuum at room temperature, which does not reflect a material's behaviour under operating conditions.

In situ thermal, electrical, and gaseous platforms

MEMS-based in situ holders apply thermal and electrical stimuli directly to the sample area, covering temperature ranges from below -170 degrees C to above 800 degrees C. For gas-phase experiments, windowed flow cells encapsulate the sample between electron-transparent ceramic membranes, isolating reactive gases from the column vacuum. This allows operation at pressures up to 1 bar and above, directly relevant to industrial catalytic conditions.

Cryogenic electron microscopy

Rapid vitrification preserves samples in their native state and reduces radiation damage, as covered in the biology and battery application sections. Automated cryo-TEM platforms support single-particle analysis, MicroED, and cryo-ET across both biological and materials science applications.

Direct electron detector technology

Event-based direct electron detectors perform electron counting in hardware via FPGAs rather than using software to analyze frames after the fact. The chip identifies, centroiding, and registers individual electron events in real time. This eliminates coincidence loss, the error that occurs when multiple electrons strike the same pixel and are counted as one, and allows operation at much higher dose rates than earlier detector generations.

On-chip Correlated Double Sampling (CDS) eliminates reset noise and maximizes Detective Quantum Efficiency (DQE). The practical result is that beam brightness can be increased substantially while keeping exposure times short. This enables high-throughput continuous rotation tomography, rapid 4D-STEM acquisitions, and improved temporal resolution in dynamic in situ experiments.

Measurlabs TEM imaging services

Measurlabs offers TEM imaging services for materials science, semiconductor characterization, catalysis, energy storage, and life science applications. For a practical example of our service in use, see the following case study: Measurlabs supports Low Noise Factory with routine FIB-STEM imaging of ultra-low-noise microwave amplifiers. For more information or a quote, please contact us using the form.

How this article was produced

This article combines three sources of knowledge. The first is Teemu Myllymäki's scientific training, an M.Sc. in Organic Chemistry and a Ph.D. in Applied Physics, during which TEM was a working instrument rather than a theoretical concept. The second is Measurlabs' experience running customer projects across a wide range of materials, which provides the practical observations embedded in the text. The third is AI-assisted synthesis, used to structure and expand the source material and surface relevant technical literature.

The article was written to bridge the gap between textbook theory and hands-on instrument experience. Technical claims should be verified against the cited primary literature.

Last reviewed: April 15, 2026.