Transmission electron microscopy (TEM) forms an image by passing a focused beam of electrons through an ultrathin specimen rather than bouncing electrons or light off a surface. Because electrons have a much shorter wavelength than visible light, TEM can in principle resolve structures far below the ~200 nm diffraction limit of light microscopy — down to individual atomic columns in a well-aligned, aberration-corrected instrument. In practice, almost none of a TEM project’s difficulty is in the physics of the beam. It is in getting a specimen thin enough for electrons to pass through it (electron-transparent), preserved close enough to its real structure to be worth imaging, and stained or contrasted enough to see. This guide walks through the standard chemical-fixation sample preparation workflow step by step, the alternative routes you’d use instead (negative staining, cryo-EM, FIB milling), what resolution actually depends on, and the failure modes that produce a ruined grid.
What TEM Resolution Actually Depends On
Two different numbers get called “TEM resolution,” and mixing them up is the single most common source of unrealistic expectations. Instrument (point) resolution is a property of the microscope itself — the accelerating voltage, the electron source, and (on modern high-end instruments) whether it has aberration correction. State-of-the-art aberration-corrected TEMs can reach sub-0.1 nm resolution on hard, crystalline, radiation-tolerant materials. Achievable specimen resolution is what you actually get once specimen thickness, staining, radiation sensitivity and contrast are factored in — and for a routine chemically-fixed, resin-embedded biological specimen imaged on a mid-range instrument, that number is closer to 1–2 nm, not the instrument’s theoretical best. The specimen, not the microscope, is usually the limiting factor.
| Regime | Typical practical resolution | What limits it |
|---|---|---|
| Aberration-corrected TEM/STEM, hard crystalline materials | ~0.05–0.1 nm (sub-Ångström to atomic-column) | Instrument optics and lens aberrations, not the specimen |
| Cryo-EM single-particle reconstruction, well-behaved macromolecules | Sub-nanometer to near-atomic after image averaging | Particle number, conformational heterogeneity, radiation dose budget — a computational result, not a single micrograph’s raw resolution |
| Routine chemically-fixed, resin-embedded biological ultrathin section | ~1–2 nm | Section thickness, staining contrast, fixation-induced structural loss |
| Negative-stained particles/viruses | ~2 nm | Stain grain size and specimen flattening on the grid |
Because the achievable number depends this heavily on specimen type, prep quality and instrument class, treat any single quoted “TEM resolution is X nm” figure as a starting range to verify against your own instrument’s spec sheet and your sample class — not a universal constant.
The Standard Sample Preparation Workflow (Chemical Fixation Route)
This is the default workflow for biological tissue, cells and most soft-matter specimens headed for conventional (room-temperature, resin-embedded) TEM. Each step exists to solve a specific problem the next step would otherwise fail on.
- Primary fixation. The specimen is immersed in a buffered aldehyde fixative (commonly glutaraldehyde, sometimes combined with paraformaldehyde) to cross-link proteins and halt cellular activity before ultrastructure degrades. Fixative concentration and buffer osmolarity both need to be matched to the tissue — too weak and structures are poorly preserved; too strong, wrong pH, or wrong osmolarity and you get shrinkage, swelling or extraction artifacts (see the troubleshooting table below).
- Secondary (post-)fixation. A wash step removes unbound aldehyde, then the specimen is post-fixed in osmium tetroxide. Osmium cross-links and stabilizes lipids (which aldehydes alone don’t fix well) and adds heavy-metal mass that later provides image contrast — it is functioning as a fixative and an early contrast agent at the same time.
- Dehydration. Water is removed through a graded series of increasing ethanol or acetone concentrations (e.g., 30% → 50% → 70% → 90% → 100%), because the resins used for embedding are not water-miscible. Moving too fast through the series is a common cause of specimen distortion.
- Infiltration and embedding. The dehydrated specimen is infiltrated with a resin (commonly an epoxy such as Epon- or Spurr-type formulations), often via an intermediate transitional solvent, then oriented and cast into a mold with fresh resin.
- Polymerization (curing). The resin block is cured, typically in an oven at around 60°C for roughly 24–48 hours, producing a hard block that can be trimmed and cut.
- Ultramicrotomy (sectioning). The cured block is trimmed to a small trapezoid around the region of interest, then cut into ultrathin sections — commonly in the 50–90 nm range, with a usable window of roughly 30–150 nm depending on what’s being imaged and at what resolution — on an ultramicrotome using a diamond (or, for less critical work, glass) knife. Sections float off onto a water-filled boat and are picked up on a support grid.
- Grid mounting. Sections are collected on a metal mesh grid (commonly copper), usually pre-coated with a thin support film (e.g., formvar or carbon) so unsupported sections don’t tear or fall through the mesh openings.
- Section staining. Mounted sections are stained with heavy-metal contrast agents — typically uranyl acetate followed by lead citrate — which bind differentially to cellular structures and generate the mass-thickness contrast the image actually depends on. Without this step, a resin section is nearly featureless under the beam.
End to end, this workflow routinely runs from same-day (fixation) to several days (fixation, dehydration, infiltration and a multi-day cure), before sectioning and staining add another session — budget for it as a multi-day protocol, not a same-morning turnaround.
Specimen Thickness, Electron Transparency and Resolution
Electron transparency is the real constraint the whole workflow above is built around: electrons must be able to pass through the specimen without being absorbed, and thinner sections both transmit more electrons and preserve more structural detail.
| Specimen thickness | Electron transparency | Practical consequence |
|---|---|---|
| <50 nm | High | Preferred for high-resolution imaging; less overlap of structures along the beam path |
| 50–100 nm | Still electron-transparent | The typical working range for routine ultrathin sections; resolution degrades gradually as thickness increases |
| >150–200 nm | Increasingly opaque / degraded contrast | Structures overlap along the beam path (projection effect); commonly used deliberately for STEM tomography or thicker semi-thin survey sections, not high-resolution imaging |
This is also why materials-science specimens (semiconductors, metals, ceramics) that can’t be ultramicrotomed usually go through mechanical thinning followed by ion milling or, for site-specific work, focused ion beam (FIB) lift-out — the goal is the same electron-transparent thickness, reached by a different route because resin embedding and diamond-knife sectioning don’t work on hard inorganic materials.
Alternative Preparation Routes
The chemical-fixation/resin-embedding workflow above is the default for tissue and cell biology, but it is not the only route, and choosing the wrong one for your specimen class is a common source of wasted grids.
| Route | Best suited to | What it trades off |
|---|---|---|
| Negative staining | Purified viruses, macromolecular complexes, isolated particles | Fast (minutes, no sectioning) and simple, but limited to ~2 nm resolution and can flatten or distort delicate particles |
| Cryo-EM (plunge-freeze vitrification) | Macromolecules, viruses and complexes in a near-native, fully hydrated state | Avoids staining and dehydration artifacts and can reach near-atomic resolution via single-particle averaging, but requires cryo-holders/cryo-stages, careful ice-thickness control, and substantially more instrument time and image-processing expertise |
| Chemical fixation + resin embedding + ultramicrotomy | Tissue, whole cells, tissue ultrastructure | Well-established, moderate cost, but every fixation/dehydration step risks introducing an artifact and the sample is no longer in a native, hydrated state |
| Mechanical thinning + ion milling / FIB lift-out | Semiconductors, metals, ceramics and other hard inorganic materials | Reaches electron transparency in materials that can’t be cut with a diamond knife, but is slow, requires dedicated equipment, and can introduce beam/milling damage at the very surface being studied |
Choosing an Accelerating Voltage
Accelerating voltage affects both resolution and how much dose the specimen tolerates, so the right voltage depends on what you’re imaging, not just what the instrument’s ceiling is.
| Voltage range | Common use |
|---|---|
| ~80–120 kV | Routine biological ultrathin sections and negative-stained specimens; lower dose helps limit beam damage to soft, radiation-sensitive material |
| ~200 kV | A widely used middle ground for both structural biology (including cryo-EM instruments such as 200 kV cryo-TEMs) and general materials imaging |
| ~300 kV | High-resolution structural biology (the majority of sub-3 Å cryo-EM structures deposited to the PDB have used 300 kV instruments) and demanding materials-science imaging |
Voltage selection is always a trade-off against your specific instrument’s configuration and your specimen’s radiation sensitivity — treat the ranges above as a starting point for discussion with your facility, not a fixed rule for every sample.
Contrast Mechanisms: Reading What You See
A TEM image is not a photograph of the specimen’s surface; it’s a projection built from how the beam interacted with material along its whole path through the section. Three mechanisms dominate, and knowing which one is producing your contrast changes how you interpret the image:
- Mass-thickness contrast. Denser or thicker regions scatter more electrons out of the imaging aperture and appear darker. This is the dominant mechanism in stained biological sections — it’s why heavy-metal stains (uranyl acetate, lead citrate, osmium) are used: they add local mass differences that wouldn’t otherwise exist in low-atomic-number biological material.
- Diffraction contrast. In crystalline specimens, regions oriented to strongly diffract the beam appear darker than regions that don’t — this is how grain boundaries, dislocations and crystal defects become visible in materials-science TEM.
- Phase contrast. At high resolution, interference between scattered and unscattered electron waves produces the fringe-like contrast used to resolve atomic columns and lattice spacings — this is the mechanism behind atomic-resolution imaging, and it is highly sensitive to focus and specimen thickness.
Troubleshooting: Common TEM Sample Artifacts
| Artifact | Likely cause | Fix |
|---|---|---|
| Chatter or repeating knife marks across sections | Dull knife edge, cutting speed too fast, or vibration reaching the microtome | Advance to a fresh diamond knife edge, slow the cutting speed, isolate the microtome from building vibration |
| Compressed or distorted sections | Dull knife or incorrect clearance angle | Use a fresh knife edge; adjust clearance angle; consider chloroform vapor to relax compression before mounting |
| Wrinkled or folded sections on the grid | Sections not fully flattened before or during pickup | Use chloroform vapor or a gentle heat pen over the water boat to flatten sections before grid pickup |
| Weak or uneven contrast | Under-staining, degraded stain solution, or too-thin sections for the stain to bind enough mass | Refresh uranyl acetate/lead citrate solutions (both degrade over time and with light/CO² exposure), extend stain time, confirm section thickness is in the working range |
| “Holes” or extraction of expected structures (e.g., apparent loss of cytoplasmic detail) | Inadequate or mismatched-osmolarity fixation, or excessive dehydration extracting lipids/soluble components | Re-optimize fixative concentration and buffer osmolarity for the specific tissue; avoid rushing dehydration steps |
| Charging (bright flares or image drift near uncoated/thick regions) | Insufficient conductive coating or poor grid grounding, especially on less-conductive specimens | Ensure adequate carbon coating/support film continuity and confirm the grid and holder are properly grounded |
| Progressive loss of detail during imaging (beam/radiation damage) | Cumulative electron dose exceeding what the specimen (especially organic or polymer material) can tolerate | Minimize exposure time per field, use lower-dose/low-magnification screening before high-dose high-resolution capture, consider a lower accelerating voltage where appropriate |
| Astigmatism or persistent drift | Contaminated or misaligned aperture, unstable stage, or insufficient settling time after stage movement | Clean or replace the objective aperture, allow drift-settling time before capture, correct stigmators against a known reference |
TEM vs Other Microscopy and Structural Techniques
TEM is one of several tools that answer overlapping but distinct structural questions, and choosing between them (or combining them) usually matters more than optimizing any single one further. Confocal microscopy stays in the light-microscopy resolution regime but keeps samples alive and fluorescently labeled — the right choice when you need live, functional or multi-color fluorescence data rather than ultrastructural detail. Atomic force microscopy images a surface topographically by physically probing it rather than transmitting a beam through it, and can work in liquid on unfixed samples, at the cost of only imaging a surface rather than internal structure. Dynamic light scattering gives a fast, population-averaged particle-size estimate in solution without any imaging at all — useful as a quick QC check before committing a nanoparticle or protein-complex prep to TEM grids. X-ray crystallography reaches true atomic resolution but requires a well-ordered crystal, whereas cryo-EM (a TEM variant) has become the practical alternative for macromolecules that won’t crystallize. If the specimen in question is a bacterial culture rather than a purified particle prep, a simple Gram stain answers a coarse morphology/classification question at a fraction of the cost and time TEM requires, and is often the right first step before committing to electron microscopy at all. For samples originating from cell culture, the same fixation-quality issues that affect TEM (osmolarity, timing) also affect downstream techniques like western blot and agarose gel electrophoresis sample handling, even though those are protein/nucleic-acid rather than imaging techniques. Separations methods such as thin-layer chromatography and gas chromatography answer a completely different question (composition, not structure) but often sit in the same analytical workflow upstream or downstream of an imaging step.
Frequently Asked Questions
What is the actual resolution of a transmission electron microscope?
It depends heavily on the specimen and instrument class. Aberration-corrected instruments on hard, crystalline, radiation-tolerant materials can resolve individual atomic columns (roughly 0.05–0.1 nm). A routine chemically-fixed, resin-embedded biological section on a standard instrument realistically resolves around 1–2 nm — the specimen and its preparation, not the microscope’s theoretical ceiling, is usually what sets the achievable number.
How thin does a TEM sample need to be?
Thin enough to be electron-transparent — commonly under 100 nm, with routine ultrathin sections in the 50–90 nm range and a broader usable window of roughly 30–150 nm depending on the imaging goal. Thicker sections transmit fewer electrons and layer structures on top of each other along the beam path, degrading resolution.
How long does TEM sample preparation take?
For the standard chemical-fixation/resin-embedding route, budget multiple days: fixation and post-fixation can be same-day, but dehydration, infiltration and a resin cure (commonly ~24–48 hours at around 60°C) add real time before sectioning and staining. Negative staining is far faster (minutes, no embedding or curing), while cryo-EM grid preparation is also comparatively quick but demands more downstream imaging and processing time.
What’s the practical difference between TEM and SEM sample preparation?
TEM requires the specimen to be thin enough for electrons to pass through it (electron transparency is the whole point of the workflow above). Scanning electron microscopy (SEM) images a surface by detecting electrons reflected or emitted from it, so specimens don’t need to be sectioned to a specific transparent thickness — the preparation goals for SEM center on surface conductivity and preserving surface topography rather than transmission thickness.
Can TEM be used on living or fully hydrated samples?
Not in the standard sense — TEM operates under high vacuum, which is incompatible with a living, liquid-hydrated specimen. Cryo-EM is the closest practical answer: vitrifying (rapidly freezing without ice-crystal formation) a hydrated specimen preserves it in a near-native structural state, even though the imaging itself still happens on a frozen, non-living grid under vacuum.
Why is osmium tetroxide used in TEM sample preparation?
It serves two roles at once during post-fixation: it cross-links and stabilizes lipid membranes (which aldehyde fixatives alone preserve poorly), and its heavy-metal mass contributes early contrast that carries through to the final stained, imaged section.
