{
  "id": "transmission-electron-microscope",
  "title": "Transmission Electron Microscope",
  "lang": "en",
  "source_lang": "en",
  "summary": "A transmission electron microscope uses electrons passing through a thin specimen to reveal internal structure, crystallography, and chemical composition at nanometre to atomic scales.",
  "category": "science",
  "subcategory": "physics",
  "type": "concept",
  "aliases": [
    "TEM",
    "transmission electron microscopy"
  ],
  "generated_at": "2026-11-03",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/transmission-electron-microscope/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/transmission-electron-microscope/",
      "json": "https://aiwiki.page/zh/transmission-electron-microscope.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "A transmission electron microscope (TEM) is an instrument that forms images and diffraction patterns using a beam of electrons transmitted through a sufficiently thin specimen. Electromagnetic lenses control the beam and magnify the information carried by the transmitted electrons. Unlike a light microscope, a TEM can resolve features substantially smaller than the wavelength of visible light, including individual atomic columns under suitable conditions.\n\nTransmission electron microscopy combines imaging with measurements of crystal structure, elemental composition, chemical bonding, and electronic properties. It is central to research on metals, ceramics, semiconductors, nanoparticles, and biological structures. Its capabilities depend not only on electron optics but also on specimen preparation, detector performance, computational analysis, and the specimen’s resistance to irradiation. Conventional TEM, scanning transmission operation, and cryogenic methods use related physical principles but emphasize different signals and experimental conditions."
    },
    {
      "heading": "Electron waves and interaction with matter",
      "text": "The resolving power of a TEM begins with the wave properties of electrons, described by quantum mechanics. An electron accelerated through an electrical potential difference acquires a momentum and an associated de Broglie wavelength. Including the relativistic correction, the wavelength is\n\n\\[\\lambda=\\frac{h}{\\sqrt{2m_e eV\\left(1+\\frac{eV}{2m_e c^2}\\right)}},\\]\n\nwhere \\(h\\) is the Planck constant, \\(m_e\\) is the electron rest mass, \\(e\\) is the magnitude of its charge, \\(V\\) is the accelerating voltage, and \\(c\\) is the speed of light. At 100 kilovolts the wavelength is approximately 3.70 picometres; at 200 kilovolts it is approximately 2.51 picometres. These wavelengths are much smaller than typical interatomic distances, but they do not by themselves determine the achievable image resolution.\n\nElectrons interact strongly with the electrostatic potential of matter. In elastic scattering, an electron changes direction without appreciable energy transfer to the specimen. Coherent elastic scattering produces diffraction and the interference responsible for much high-resolution image contrast. In inelastic scattering, an electron transfers energy to excitations such as collective electron oscillations, electronic transitions, or lattice vibrations. These processes supply analytical information but can also introduce background, chromatic blur, and specimen damage.\n\nBecause scattering is strong, specimens must usually be thin. Many routine specimens are tens to roughly one hundred nanometres thick, although suitable thickness varies greatly with composition, accelerating voltage, and the information required. Atomic-resolution imaging often requires thinner regions, especially for materials containing heavy elements. A specimen that is too thick produces substantial multiple scattering, overlaps structures along the beam direction, and reduces the interpretability of both images and spectra.\n\nA TEM image is therefore not simply a magnified shadow. It records selected consequences of the interaction between an incident electron wave and the specimen, followed by propagation through an imperfect optical system. Changes in thickness, orientation, composition, focus, or lens settings can alter contrast even when the underlying structure remains unchanged. Higher accelerating voltages generally improve penetration and shorten the wavelength, but they can increase some forms of radiation damage. Lower voltages may protect certain materials against atomic displacement while increasing scattering and making chromatic aberration more important. Operating voltage is consequently an experimental compromise rather than a universal measure of microscope quality."
    },
    {
      "heading": "Electron sources, lenses, and the microscope column",
      "text": "A TEM column contains an electron source, illumination optics, a specimen stage, an objective lens, intermediate and projector lenses, and recording devices. The column operates under high vacuum because collisions with gas molecules would scatter the beam and contamination would impair both the source and the specimen. Different parts of the instrument can use different vacuum levels, with particularly stringent requirements around field-emission sources.\n\nElectron guns commonly use thermionic or field-emission sources. Thermionic guns heat a tungsten filament or a lanthanum hexaboride crystal until electrons escape from its surface. Field-emission guns extract electrons from a sharp tip using a strong electric field. Cold field-emission sources operate without the high tip temperature characteristic of thermionic emission; Schottky sources combine heating with field-assisted emission. Their brightness, energy spread, stability, and operating requirements differ. High brightness is especially valuable when forming a small, intense probe for scanning or spectroscopy.\n\nMagnetic lenses use the Lorentz force to bend electron trajectories. In a conventional lens, a current-carrying coil and shaped ferromagnetic pole pieces establish an approximately rotationally symmetric magnetic field. Condenser lenses control beam convergence and illuminated area. The objective lens forms the first image and diffraction pattern of the specimen; intermediate and projector lenses select and magnify these distributions. Switching lens excitations allows an operator to view either an image plane or a diffraction plane without fundamentally changing the electron–specimen interaction.\n\nApertures select electrons by position or scattering angle. A condenser aperture controls illumination conditions, an objective aperture selects portions of the diffraction pattern for image formation, and a selected-area aperture restricts the specimen region contributing to a diffraction pattern. Stigmators compensate for astigmatism, in which the lens has different focal strengths in different transverse directions. Beam deflectors provide alignment, scanning, and image shifting. Specimen holders may add tilt, cooling, heating, electrical contacts, mechanical loading, or other functions.\n\nImages were historically viewed on fluorescent screens and recorded on photographic film. Digital systems include scintillator-coupled cameras and detectors that register electrons directly. Detector properties influence sensitivity, spatial sampling, frame rate, and the amount of useful information obtained per incident electron. The entire column also depends on environmental control: mechanical vibration, stray magnetic fields, temperature changes, acoustic noise, and unstable electrical supplies can blur an image. At atomic resolution, the microscope, its supporting room, and its electronic controls effectively form a single measurement system."
    },
    {
      "heading": "Image formation and contrast mechanisms",
      "text": "Conventional TEM commonly uses broad illumination to form a magnified image of a specimen area simultaneously. Several contrast mechanisms can operate together, and interpreting an image requires identifying which dominate. The displayed brightness is not generally a direct map of height, elemental concentration, or atomic number; it depends on the electrons admitted to the imaging system and on their phase relationships.\n\nIn bright-field imaging, the image is formed mainly from the transmitted beam, often with an objective aperture excluding strongly scattered electrons. Regions that scatter more electrons outside the aperture can appear darker. This produces mass–thickness contrast in many noncrystalline specimens: thicker regions or regions containing more strongly scattering elements often transmit fewer electrons into the selected angular range. Biological sections stained with heavy elements exploit this mechanism. However, in crystalline specimens, changes in orientation can produce strong diffraction contrast that outweighs simple thickness effects.\n\nIn dark-field imaging, the central transmitted beam is excluded and one or more scattered beams form the image. A crystal grain or precipitate that contributes to a selected reflection appears bright against a darker background. This is useful for mapping phases, identifying orientation relationships, and locating defects. Dislocations and stacking faults modify local diffraction conditions through their displacement and strain fields. Their contrast can change or disappear when a different reflection is selected, allowing crystallographic analysis rather than merely visual identification.\n\nHigh-resolution transmission electron microscopy often relies on phase contrast produced by interference between transmitted and diffracted electron waves. The objective lens converts phase differences into measurable intensity variations. Defocus and lens aberrations determine how different spatial frequencies are transferred into the image. As a result, atomic columns can appear bright, dark, or weakly visible under different conditions. Lattice fringes show periodic information, but their appearance alone does not establish the positions or chemical identities of individual atoms.\n\nThe contrast-transfer function describes this frequency-dependent imaging response under specified approximations. At frequencies where transfer is weak or changes sign, structural information may be suppressed or contrast may reverse. Focal-series reconstruction, electron holography, and comparison with calculated images can recover or constrain information not apparent in a single exposure. For thicker crystals, simulations often use the multislice method, which models successive scattering and propagation through thin computational slices. Reliable interpretation therefore combines microscope settings, specimen geometry, diffraction information, and image simulation rather than assuming that the most visually distinct feature directly represents an atom."
    },
    {
      "heading": "Preparing specimens for transmission",
      "text": "Specimen preparation determines which structures survive, which become visible, and which artifacts may be mistaken for real features. A specimen must be sufficiently thin, mechanically supported, compatible with the microscope environment, and appropriate for the intended measurement. A thin specimen is not necessarily representative: selecting a small transparent region can favor a particular grain, surface condition, particle size, or damaged layer.\n\nFor conventional biological TEM, tissue is commonly chemically fixed, dehydrated, embedded in resin, and cut into ultrathin sections with an ultramicrotome. Aldehyde fixatives stabilize many cellular components through chemical cross-linking, while osmium tetroxide is widely used in membrane preservation and contrast enhancement. Stains containing uranium or lead increase electron scattering from selected structures. Sections often lie in the range of roughly 50–100 nanometres, though the appropriate thickness depends on the specimen and purpose. Fixation, dehydration, staining, and sectioning can each alter dimensions or remove material, so the image represents a prepared specimen rather than an untouched living cell.\n\nNegative staining offers a different route for small particles, including viruses and purified molecular assemblies. A heavy-element stain surrounds and outlines the particles, providing strong contrast at relatively modest electron doses. Drying and interaction with the support can flatten or distort structures. In cryo-electron microscopy, rapid cooling instead immobilizes material in vitreous ice, avoiding the large ice crystals produced by ordinary freezing. Thin aqueous films are commonly vitrified by plunging into a cryogen such as liquid ethane; thicker biological specimens may require high-pressure freezing followed by thinning.\n\nMaterials specimens can be prepared by mechanical polishing, dimpling, electropolishing, or ion milling. Focused ion beam instruments are especially important for extracting site-specific lamellae from semiconductor devices, interfaces, and individual microstructural features. They can select a region first located by other imaging methods and thin it to electron transparency. Ion exposure can nevertheless introduce implantation, amorphization, redeposition, or preferential removal of particular components. Lower-energy final polishing can reduce some of these effects without necessarily eliminating them.\n\nPowders and nanoparticles are often dispersed onto support films carried by metal grids, commonly a few millimetres in diameter. Carbon films, perforated supports, and thin inorganic membranes provide different combinations of stability and background scattering. Preparation must also account for the intended analysis: a copper grid can complicate identification of copper in a specimen, and a carbon support can contribute to a carbon signal. Contamination from solvents, handling, protective coatings, and residual hydrocarbons can become disproportionately important when the region under examination contains only a small number of atoms."
    },
    {
      "heading": "Diffraction and crystallographic information",
      "text": "Electron diffraction reveals periodic structure by recording the angular distribution of scattered electrons. For a crystalline specimen, diffracted beams reflect the geometry of its reciprocal lattice. The relation between wavelength, plane spacing, and scattering angle is commonly expressed through Bragg’s law,\n\n\\[2d\\sin\\theta=n\\lambda,\\]\n\nwhere \\(d\\) is the interplanar spacing, \\(\\theta\\) is the Bragg angle, and \\(n\\) is an integer. Electron wavelengths are so short that many useful diffraction angles are small. A single crystal near a suitable orientation often produces an array of spots, whereas many randomly oriented crystallites produce rings. Noncrystalline materials generally produce broad, diffuse features.\n\nSelected-area electron diffraction uses an aperture in an image plane to restrict the contributing specimen region. It can identify phases, determine crystal orientation, measure lattice spacings, and establish relationships between adjacent grains or between a precipitate and its matrix. In a simplified small-angle treatment, the distance \\(R\\) of a reflection from the direct beam is approximately related to spacing by \\(Rd=L\\lambda\\), where \\(L\\) is the effective camera length. Accurate measurements require calibration because lens settings, detector geometry, specimen height, and distortion affect the scale.\n\nConvergent-beam electron diffraction uses a converging illumination cone, producing discs rather than nearly point-like reflections. Intensity variations within and between these discs contain information about symmetry, orientation, thickness, and, under suitable conditions, local strain. Nanobeam diffraction restricts illumination to a small region, allowing structures smaller than those conveniently isolated by a selected-area aperture to be examined. Precession electron diffraction tilts the incident beam around an axis during acquisition, sampling more reciprocal space and reducing some orientation-sensitive intensity variations.\n\nElectron diffraction differs importantly from X-ray diffraction because electrons interact much more strongly with matter. A transmitted electron can undergo several scattering events, giving rise to dynamical diffraction. Reflection intensities may therefore depend strongly on specimen thickness and orientation rather than only on the squared magnitude of a structure factor. Even the appearance of reflections absent in a simple kinematic calculation can require consideration of multiple scattering. Quantitative structure determination must account for these effects or use collection and analysis strategies that limit their consequences.\n\nThree-dimensional electron diffraction collects patterns while a small crystal is tilted through a range of orientations. This allows structural investigation of crystals too small for many conventional single-crystal X-ray experiments, including inorganic compounds and molecular crystals. Diffraction mapping also supports studies of phase transformations, deformation, and nanoscale strain. Its spatial resolution, angular precision, and dose requirements differ from those of direct imaging, so diffraction and imaging provide complementary constraints rather than interchangeable pictures of the same property."
    },
    {
      "heading": "Scanning transmission and chemical analysis",
      "text": "In scanning transmission electron microscopy (STEM), a finely focused probe scans across the specimen and detectors record signals at each position. Many modern TEM instruments can operate in both conventional TEM and STEM modes. The distinction is primarily one of illumination and signal collection: conventional TEM forms a broad-field image through imaging lenses, whereas STEM constructs an image sequentially from position-resolved detector measurements.\n\nAn annular dark-field detector collects electrons scattered outside the central beam. At sufficiently high collection angles, much of the signal can be understood in terms of relatively incoherent scattering. Under suitable thin-specimen and imaging conditions, high-angle annular dark-field intensity increases strongly with atomic number, making heavy-element columns conspicuous among lighter elements. The dependence is not a universal fixed power law, because thickness, electron channeling, detector geometry, and composition also matter. Annular bright-field and other detector arrangements can reveal complementary information, including light-element columns in favorable specimens.\n\nEnergy-dispersive X-ray spectroscopy analyzes characteristic X-rays emitted after incident electrons remove inner-shell electrons and higher-energy electrons fill the vacancies. The resulting photon energies identify elements, while appropriately corrected intensities can estimate composition. STEM can map these signals across interfaces, nanoparticles, and device structures. Practical limits include low counting statistics, overlapping spectral lines, absorption, detector geometry, and additional X-rays generated in supports or nearby material.\n\nElectron energy-loss spectroscopy measures the energy lost by transmitted electrons. Low-loss spectra contain information about collective electronic excitations and, in appropriate experiments, band gaps and dielectric response. Core-loss edges identify elements, while fine structure near an edge can reveal aspects of bonding, coordination, and oxidation state. Interpretation requires attention to energy resolution, background subtraction, plural scattering, and the comparison between measured spectra and reference data or calculations. Chemical changes inferred from a spectrum must also be separated from changes induced during irradiation.\n\nThese methods make the instrument valuable for studying dopant distributions, oxide interfaces, catalyst particles, segregation at grain boundaries, and layered materials. Atomic-scale chemical maps are possible in suitable stable specimens, but nominal probe size alone does not establish analytical resolution: scattering within the specimen and the spatial extent of the excitation also contribute.\n\nIn four-dimensional STEM, a two-dimensional diffraction pattern is recorded at each point of a two-dimensional scan. This produces datasets that can be analyzed for orientation, strain, and other properties. Electron ptychography uses overlapping probe positions and computational reconstruction to recover specimen information from diffraction intensities. Its resolution and dose efficiency depend on acquisition geometry, detector performance, specimen behavior, and the validity of the reconstruction model."
    },
    {
      "heading": "Biological imaging and three-dimensional reconstruction",
      "text": "Biological TEM addresses a fundamental contrast problem: many important biological structures consist mainly of relatively light elements and tolerate only limited electron exposure. Conventional stained sections nevertheless reveal cellular organization on scales inaccessible to ordinary light microscopy. Membranes, organelles, cytoskeletal structures, and viral particles can be distinguished through their morphology and staining patterns. Electron microscopy contributed decisively to the characterization of cellular ultrastructure, including the internal membranes of the mitochondrion and the organization of the endoplasmic reticulum.\n\nThe biological meaning of an image depends on preparation and labeling. A dark membrane profile in a stained section reflects the interaction of fixatives, stains, and section geometry with the original structure. Immunogold labeling uses antibodies associated with electron-dense gold particles to locate selected molecular targets. Such labeling can connect ultrastructure with molecular identity, although accessibility, specificity, labeling efficiency, and the displacement between label and target constrain interpretation.\n\nSingle-particle cryo-EM analyzes many projections of nominally identical particles embedded in vitreous ice. Computational alignment and classification estimate particle orientations and separate structural states, after which a three-dimensional density map is reconstructed. The method has produced high-resolution structures of large protein complexes, membrane proteins, and assemblies such as the ribosome. Its success depends on particle quality, orientation distribution, ice thickness, image contrast, and the control of motion and radiation damage. The reconstructed map is an estimate derived from many observations rather than a direct three-dimensional photograph of one molecule.\n\nElectron tomography instead records a tilt series of the same specimen region. Reconstruction combines the projections to estimate a three-dimensional volume. Cryo-electron tomography can preserve macromolecular assemblies within cellular environments, while tomography of resin sections is widely used for cellular architecture. Because holders and specimen geometry usually prevent a full range of tilt angles, the reconstruction has incomplete angular information, often described as a missing wedge. This causes direction-dependent resolution and can elongate or obscure structures.\n\nRepeated structures within tomograms can be extracted and averaged through subtomogram analysis, improving signal and sometimes enabling detailed structural interpretation in situ. Thick cells present a penetration problem, so cryogenic focused ion beam milling can prepare thin lamellae while maintaining vitrified conditions. These approaches connect molecular structure with cellular context, but their resolution is constrained by dose sharing across the tilt series, specimen thickness, alignment accuracy, and structural variability. Distinguishing a reproducible molecular feature from preparation effects or reconstruction artifacts requires independent checks and careful treatment of heterogeneous populations."
    },
    {
      "heading": "Resolution, radiation damage, and measurement reliability",
      "text": "TEM performance is limited by more than electron wavelength. Conventional round lenses suffer from spherical aberration: electrons passing through different parts of the lens do not converge to precisely the same point. Chromatic aberration arises because electrons of different energies are focused differently. Source energy spread, inelastic scattering, lens instability, specimen drift, vibration, and detector response all influence the recorded information. The attainable resolution therefore depends on the whole experiment, not simply on magnification or accelerating voltage.\n\nAberration-corrected electron microscopy uses multipole elements to compensate for selected lens aberrations. Correcting spherical aberration permits smaller probes in STEM and improved transfer of high spatial frequencies in TEM. Monochromators can narrow the energy distribution, while specialized systems also correct chromatic aberration. Sub-ångström information is attainable with modern instruments under suitable conditions. Nevertheless, point resolution, information limit, probe diameter, and precision in locating a fitted atomic position describe different quantities and should not be treated as synonyms.\n\nElectron irradiation can change the object being measured. Knock-on damage occurs when momentum transfer displaces an atom from its site. Radiolysis results from electronic excitation and ionization that break or rearrange chemical bonds, and is especially important in biological and many insulating materials. Beam-induced heating, charging, mass loss, and chemical reactions can also occur. Hydrocarbons on a specimen or in the column may decompose and form carbonaceous deposits under irradiation, gradually obscuring a region or altering its analytical signal.\n\nDose is commonly expressed as incident electrons per unit area, often electrons per square ångström. If other factors remain unchanged, increasing the number of detected electrons improves shot-noise-limited precision approximately with the square root of that number. Beam-sensitive specimens, however, may be damaged before sufficient counts can be collected for a desired measurement. Cooling can reduce some damage processes and preserve vitrified structures, but it does not prevent all radiation chemistry. Dose rate, elapsed time, and acquisition sequence can matter as well as total dose.\n\nQuantitative work requires explicit treatment of measurement uncertainty. Image scale, diffraction camera length, energy dispersion, detector sensitivity, and specimen thickness may each require calibration or independent estimation. Drift can deform a scanned image, while scan nonlinearity can mimic strain. Image processing and machine-learning methods can improve visibility or automate analysis, but they may also suppress unexpected features or generate plausible-looking structure unsupported by the raw data. Reliability rests on retaining the original measurements, validating processing steps, comparing independent observations, and separating instrumental resolution from the strength of a particular structural or chemical inference."
    },
    {
      "heading": "Historical development and observation under controlled conditions",
      "text": "Transmission electron microscopy emerged from the realization that electron beams could be focused in ways analogous to light. Hans Busch’s 1926 work established the focusing action of magnetic fields on electron beams. In 1931, Max Knoll and Ernst Ruska built an early electron microscope using magnetic lenses. Ruska’s subsequent work produced an instrument in 1933 with resolving power exceeding that of light microscopes. Siemens introduced a commercial transmission electron microscope in 1939. These developments transformed electron optics from an experimental demonstration into a practical means of examining fine structure.\n\nBiological applications expanded as preparation methods improved. Ultrathin sectioning, suitable embedding resins, heavy-element staining, and improved fixation made cellular interiors accessible. Materials research developed diffraction contrast and high-resolution imaging into tools for studying dislocations, precipitates, crystal defects, and interfaces. Scanning transmission methods, associated especially with the work of Albert Crewe and colleagues, brought high-brightness sources and localized analytical signals into increasing use. Ruska received half of the 1986 Nobel Prize in Physics for his fundamental work in electron optics and the design of the first electron microscope.\n\nLate twentieth-century advances increasingly addressed limitations imposed by lenses and specimens rather than wavelength. Experimental aberration correction became practical in the 1990s, with major demonstrations in STEM in 1997 and TEM in 1998. Improvements in mechanical stability, computing, and detectors expanded the range of quantitative experiments. Direct electron detection was particularly important to high-resolution cryo-EM because rapid, dose-efficient recording allowed motion correction and better use of weak signals. The 2017 Nobel Prize in Chemistry recognized Jacques Dubochet, Joachim Frank, and Richard Henderson for developing cryo-electron microscopy for high-resolution structure determination of biomolecules in solution.\n\nModern instruments also examine changes rather than only static structures. Heating holders reveal processes such as particle coarsening, phase transformations, and defect motion. Electrical-biasing holders investigate devices and electrically driven changes, while mechanical holders can deform small specimens during observation. These measurements relate structural evolution to controlled external stimuli, although the thin specimen geometry and contact arrangement may differ substantially from bulk conditions.\n\nEnvironmental TEM introduces gas around the specimen while differential pumping protects other parts of the column. Sealed gas cells and liquid cells use thin electron-transparent windows to contain environments incompatible with the ordinary column vacuum. Such arrangements enable observations of catalysts, electrochemical processes, and particle growth. Their interpretation must account for scattering by windows and fluid, restricted geometry, and electron-induced chemistry. In liquid water, for example, irradiation generates reactive products that can alter the reaction being observed; distinguishing externally driven behavior from beam-driven behavior is part of the measurement itself."
    }
  ],
  "links": [
    "electron",
    "microscope",
    "quantum-mechanics",
    "planck-constant",
    "speed-of-light",
    "scattering",
    "diffraction",
    "electric-field",
    "lorentz-force",
    "magnetic-field",
    "interference",
    "cell",
    "virus",
    "cryo-electron-microscopy",
    "reciprocal-lattice",
    "braggs-law",
    "calibration",
    "structure-factor",
    "atomic-number",
    "energy-dispersive-x-ray-spectroscopy",
    "x-ray",
    "electron-energy-loss-spectroscopy",
    "oxidation-state",
    "mitochondrion",
    "protein",
    "ribosome",
    "measurement-uncertainty",
    "machine-learning",
    "nobel-prize-in-physics",
    "nobel-prize-in-chemistry"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}