Atomic force microscopy is a technique that creates images of surfaces by physically touching them with an extraordinarily fine needle, building up a picture point by point the way a fingertip reading Braille builds up a word. Unlike optical or electron microscopes, which bounce light or electrons off a sample and collect what comes back, an AFM drags or taps a sharp probe across a surface and records how that probe deflects, achieving resolution down to fractions of a nanometer. The technique works on almost anything, from semiconductor wafers to living cells submerged in fluid, which is a large part of why it has become one of the most versatile instruments in modern science.
How the Probe Builds an Image
The core of every AFM is a tiny cantilever, usually made of silicon or silicon nitride, with a sharp tip at its free end. The tip’s apex can be just a few nanometers across. As this tip is scanned across a surface, forces between the tip and the sample cause the cantilever to bend or shift in frequency. A laser bounced off the back of the cantilever tracks those deflections with extreme sensitivity. A feedback loop continuously adjusts the height of the probe so that the force (or oscillation) stays constant, and the adjustments the system makes at each point become the topographic map you see on screen.
Because the measurement is mechanical rather than optical, AFM sidesteps the diffraction limit that caps the resolution of ordinary light microscopes at roughly half the wavelength of visible light. There is no need for vacuum, either, so samples can sit in air or even in liquid at room temperature. That combination of high resolution and mild operating conditions is what makes AFM appealing for biology, where keeping a sample alive and hydrated matters.
Contact, Tapping, and Non-Contact Modes
The simplest way to run an AFM is contact mode: the tip stays in continuous contact with the surface while it scans. This gives strong, direct signals and works well on hard, flat samples, but it can drag across or damage soft materials. For anything delicate, researchers switch to oscillating modes, where the cantilever vibrates near its resonant frequency and the tip only touches the surface intermittently or not at all.
The two main oscillating approaches are amplitude-modulation (AM, commonly called tapping mode) and frequency-modulation (FM, often called non-contact mode). In tapping mode the cantilever oscillates at a fixed driving frequency and the system tracks changes in oscillation amplitude as the tip encounters surface features. In FM mode the system instead tracks shifts in the cantilever’s resonant frequency caused by tip-sample forces. For imaging in air, the two modes produce comparable resolution, but FM mode has a clear advantage when scanning soft materials in liquid. Experiments comparing the two show that tapping mode exerts a stronger force on soft samples and can deform them, while FM mode applies a much gentler force and captures height values closer to the true surface profile.1Nanotechnology. Imaging of soft matter with tapping-mode atomic force microscopy and non-contact-mode atomic force microscopy If you are imaging something like a protein membrane submerged in buffer solution, FM mode will give you a more accurate picture of its actual shape.
Seeing Individual Chemical Bonds
The most visually striking achievement in AFM has been the direct imaging of chemical bonds within single molecules. By cooling the microscope to cryogenic temperatures, operating in ultra-high vacuum, and attaching a single carbon monoxide molecule to the tip, researchers can resolve the connections between individual atoms in a molecule. The resulting images look almost like textbook structural formulas, showing bond lengths, bond angles, and even differences in bond order between single and double bonds.2Surface Science Reports. Noncontact atomic force microscopy: Bond imaging and beyond
These CO-functionalized tip experiments use a sensor design called the qPlus sensor, which is a quartz tuning fork with a stiff cantilever that oscillates at a well-defined frequency. The carbon monoxide molecule hanging off the end of the tip acts as a kind of molecular-scale probe, flexing slightly as it passes over atoms and bonds and translating those tiny force variations into measurable frequency shifts. The technique has become a go-to method for identifying unknown molecules, confirming reaction products on surfaces, and studying how molecules self-assemble. It is demanding to set up, but it produces images whose clarity is difficult to match with any other method.
Measuring How Things Feel, Not Just How They Look
Beyond imaging topography, AFM can measure mechanical and electrical properties of a sample at every point in a scan. This is where the technique really distinguishes itself from electron microscopy, which excels at morphology but tells you little about how stiff, sticky, or electrically charged a surface is.
One widely used approach is PeakForce quantitative nanomechanical mapping, in which the tip performs a rapid series of tiny indentations as it scans. Each indentation produces a force curve, and from those curves the instrument extracts values for stiffness, adhesion, and deformation at every pixel. Researchers have demonstrated that with a probe of about 240 nanometers and an appropriate contact mechanics model, this mode can map elastic modulus with a spatial resolution of roughly 50 nanometers and a minimum indentation depth of just two to three nanometers.3PubMed. Quantitative mapping of the elastic modulus of soft materials with HarmoniX and PeakForce QNM AFM modes That means you can look at a polymer blend and produce a map showing which component is stiffer, all in a single scan pass.4PubMed Central. PeakForce AFM Analysis Enhanced with Model Reduction Techniques
Other specialized modes extend this philosophy to electrical measurements. Conductive AFM presses a conducting tip against a surface and measures current flow at the nanoscale, which is useful for evaluating thin insulating films, solar cell materials, and semiconductor devices.5PubMed Central. Understanding Current Instabilities in Conductive Atomic Force Microscopy Kelvin probe force microscopy maps variations in surface potential and work function, telling you where charge accumulates on a surface. A recent advance even allows this charge mapping to be performed in water, which opens the door to studying electrochemical processes and charged biomolecules in their native aqueous environment.6ACS Nano. AC Kelvin Probe Force Microscopy Enables Charge Mapping in Water
Probing Living Cells and Single Proteins
AFM has become a standard tool in cell biology. Because the microscope works in liquid at physiological temperature, you can press the tip gently into a living cell and extract a force curve that tells you how stiff or viscoelastic the cell is. Published protocols describe doing exactly this with fruit-fly blood cells and mouse embryonic stem cells, using the shape of the force-distance curve to calculate both the elastic modulus and the time-dependent viscous response of the cell membrane and cytoskeleton.7PubMed Central. Protocol for measuring mechanical properties of live cells using atomic force microscopy Cell stiffness turns out to be a surprisingly informative quantity: cancer cells are often softer than their healthy counterparts, and changes in stiffness can signal shifts in a cell’s differentiation state or response to drugs.
At an even smaller scale, single-molecule force spectroscopy uses AFM to grab one end of a protein and pull it apart, recording the force needed to unfold each structural domain. The resulting force-extension curves reveal how mechanically stable a protein is and what the energy landscape of its folding pathway looks like.8PubMed. The physics of pulling polyproteins: a review of single molecule force spectroscopy using the AFM to study protein unfolding More recent work has pushed into subtler territory, examining how the physical attachment of a protein to the microscope’s cantilever and its flexible polymer linker might alter the folding dynamics being measured, a question that matters for interpreting all such experiments accurately.9PubMed Central. Modulation of a protein-folding landscape revealed by AFM-based force spectroscopy notwithstanding instrumental limitations
A further evolution called FluidFM integrates a microfluidic channel inside the cantilever itself, so the probe can deliver or aspirate tiny volumes of liquid at precisely targeted locations. This makes it possible to inject material into a single cell, extract contents from one, or deposit molecules onto a surface with pinpoint control.10PubMed Central. AFM and FluidFM Technologies: Recent Applications in Molecular and Cellular Biology
High-Speed AFM and Watching Molecules Move
Conventional AFM is slow. Acquiring a single image can take anywhere from one to ten minutes, which is fine for static surfaces but useless for watching a protein change shape or a cell membrane ruffle in real time. High-speed AFM (HS-AFM) solves this by using smaller, lighter cantilevers with higher resonant frequencies, faster scanners, and optimized feedback electronics, bringing frame rates up to several frames per second.
At that speed, you can record movies of individual unlabeled proteins doing their jobs. HS-AFM has captured motor proteins walking along filaments, enzymes opening and closing as they process substrates, and DNA strands being unwound, all at nanometer spatial resolution and subsecond time resolution.11PubMed. Applications of high-speed atomic force microscopy to real-time visualization of dynamic biomolecular processes No other imaging method currently provides that combination of resolution and temporal detail on single molecules without the need for fluorescent labels or stains.
The technique has also been scaled up to image whole living mammalian cells. Researchers have used HS-AFM to watch filopodia extend and retract, membrane ruffles form, endocytic pits open and close, and vesicles being internalized by cells such as HeLa and hippocampal neurons.12PubMed Central. High-speed atomic force microscopy imaging of live mammalian cells Seeing these processes unfold in real time, on live cells, without any chemical fixation or labeling, provides a direct check on what fluorescence microscopy shows and can reveal dynamics that labels might perturb.
Tip Engineering and Why It Matters
The sharpness and durability of the probe tip set a hard ceiling on what an AFM can resolve. A standard silicon tip has an apex radius somewhere around 10 nanometers, which is fine for many applications but becomes a limitation when you need to image narrow grooves, deep trenches, or closely spaced nanostructures. The tip’s finite size means that features smaller than the tip appear broadened in the image, an artifact known as tip convolution.13Nanotechnology. Correction of the tip convolution effects in the imaging of nanostructures studied through scanning force microscopy
One solution is to grow a carbon nanotube directly on the end of a silicon tip. Single-walled nanotube tips achieve apex radii of just two to four nanometers, while multiwalled versions reach three to six nanometers. They also buckle elastically rather than snapping, so they survive accidental crashes that would blunt a bare silicon tip, and they show very low nonspecific adhesion on surfaces in air.14PubMed. Carbon nanotube atomic force microscopy tips: direct growth by chemical vapor deposition and application to high-resolution imaging Another approach coats the tip with nanocrystalline diamond and then etches it into an ultra-sharp nanowire with an apex as small as five nanometers. Diamond tips are extremely hard and resist wear far better than silicon, which is a significant advantage when scanning abrasive materials or running long experiments.15PubMed. Diamond-modified AFM probes: from diamond nanowires to atomic force microscopy-integrated boron-doped diamond electrodes When those diamond probes are made from boron-doped conductive diamond, they double as nanoscale electrodes, combining topographic imaging with local electrochemical measurements in a single pass.
Artifacts and Practical Limitations
AFM’s reliance on a physical probe introduces several categories of error that users have to manage. Tip convolution, already mentioned, broadens features laterally and can make two closely spaced objects appear as one. Software algorithms exist to deconvolve the tip shape from the image, but they require an accurate model of the tip geometry, which itself can change during scanning as the tip wears or picks up debris.
The piezoelectric elements that position the tip and sample introduce their own problems. Piezo actuators exhibit hysteresis (they do not return to exactly the same position when voltage is reversed), creep (they slowly drift after a sudden voltage change), and thermal drift (their dimensions change with temperature). All three distort images, especially at slow scan speeds or over large scan areas.16International Journal of Mechanical Sciences. Role of mechanical and thermal nonlinearities in imaging by Atomic Force Microscope Modern AFMs mitigate these artifacts with closed-loop position sensors that independently measure the actual position of the scanner rather than trusting the voltage command alone, but budget instruments still run open-loop, and users need to be aware of the distortions that introduces.
Speed is another practical limitation. While HS-AFM addresses this for small scan areas and specialized cantilevers, most general-purpose AFMs still acquire images slowly compared to optical or electron microscopes. Scanning a large area at high resolution can take many minutes, during which drift and sample changes accumulate. And unlike electron microscopy, where you can zoom out to survey a large field of view quickly, AFM’s field of view is limited to whatever the piezo scanner can cover, typically on the order of 100 micrometers or less.
How AFM Compares to Electron Microscopy
A common question is why someone would choose AFM over scanning electron microscopy, or vice versa. The two techniques produce superficially similar topographic images, and direct comparisons of the same specimens confirm that their 3D surface reconstructions agree closely, with cross-correlation coefficients above 0.9 for matching surface profiles.17PubMed. A vision-based, 3D reconstruction technique for scanning electron microscopy: direct comparison with atomic force microscopy The differences lie in what else each technique can and cannot do.
SEM requires vacuum (or at least a low-pressure gas environment in environmental SEM), and most biological samples need to be dried, fixed, and coated with a thin conductive layer before imaging. AFM works in ambient air or liquid with no special sample preparation. SEM can survey large areas quickly and easily vary magnification over a huge range, while AFM is restricted to small scan areas and slow acquisition. On the other hand, AFM natively produces quantitative height data with sub-nanometer vertical resolution, whereas extracting true 3D height information from SEM requires stereoscopic reconstruction and calibration. And as discussed above, AFM provides mechanical, electrical, and chemical property maps that SEM simply cannot.
In practice, many labs use both: SEM for rapid survey imaging and AFM for quantitative, property-mapped, high-resolution work on regions of interest identified by SEM or optical microscopy.
Writing with the Tip Instead of Reading
AFM is not limited to passive observation. The same precision that allows the tip to sense a surface also allows it to modify one. In dip-pen nanolithography (DPN), the AFM tip is coated with a molecular “ink” and used to write patterns directly onto a surface, much like a quill pen deposits ink on paper. Capillary transport moves molecules from the tip to the surface through a water meniscus that forms naturally in ambient humidity. Researchers have used DPN to write patterns of gold nanoparticles at nanoscale resolution18PubMed. Direct patterning of gold nanoparticles using dip-pen nanolithography and to deposit biologically active phospholipid membrane stacks with controlled thickness.19PubMed. Comparative height measurements of dip-pen nanolithography-produced lipid membrane stacks with atomic force, fluorescence, and surface-enhanced ellipsometric contrast microscopy
Other AFM-based nanofabrication approaches include local anodic oxidation, where a voltage between tip and sample oxidizes the surface in patterns a few tens of nanometers wide, and nanoshaving, where the tip physically scrapes away a self-assembled monolayer to expose the substrate beneath. These methods are inherently serial (one tip, one line at a time), which limits throughput compared to photolithography, but they offer a level of placement precision and material flexibility that batch techniques cannot match. For research prototyping, biosensor fabrication, and fundamental studies of how molecules behave on surfaces, AFM-based patterning fills a niche that other tools leave open.
Improving Sensitivity Through Higher Resonant Modes
One area of ongoing engineering refinement is cantilever dynamics. The standard approach is to drive a cantilever at its fundamental resonant frequency, but operating at higher resonant modes, where the cantilever vibrates in more complex shapes, can substantially improve sensitivity. Experiments comparing the fundamental mode to the second mode found that the vertical resolution improved from 0.30 nanometers in the fundamental mode to 0.13 nanometers in the second mode.20SAGE Journals (Proceedings of the Institution of Mechanical Engineers, Part N). Flexural sensitivity of high resonant atomic force microscopy cantilever based on optical lever detection The improvement comes from the higher-frequency oscillation being more responsive to small force gradients near the surface. This is one of several ongoing efforts to push the instrument’s detection limits without requiring exotic tips or cryogenic cooling, making finer measurements accessible on more conventional setups.

