Advertisement

Atomic Force Microscopy in the Life Sciences

  • Matthias W. AmreinEmail author
  • Dimitar Stamov
Chapter
Part of the Springer Handbooks book series (SHB)

Abstract

The last decade have established AFM as a technique in life sciences applications ranging from single molecules to living cells and tissues. AFM still remains one of the few microscopy tools to offer premium resolution on bio-macromolecules at near physiological/native sample conditions. The demand for correlative immunochemical and ultrastructural characterization of macromolecular complexes and cells has made the combination of AFM and advanced optical microscopy techniques almost ubiquitous in every life sciences lab. This chapter gives an overview of the instrumentation and most common imaging modes used in AFM nowadays, as well as different preparation protocols for single molecule and cell applications. We finish with application examples that feature some recent developments in state-of-the-art AFMs as tools to study molecular and cellular dynamics with high spatial and temporal resolution.

Keywords

atomic force microscopy (AFM) high-resolution imaging correlative microscopy feedback loop tip-sample interaction sample preparation nanomechanical mapping force spectroscopy 

Cells are highly organized in compartments and functional units down to the macromolecular level. The frameworks of these functional and structural units are either protein complexes or complexes of proteins with nucleic acids or with lipids. The map of the human genome and the systematic mapping of the proteins expressed in tissues and cells (proteomics) are part of a concerted effort to rapidly advance the understanding of the functions of macromolecular units and the cell. However, proteomics reveals only the basic inventory of a cell, and the inventory is insufficient to explain the function of each element and the orchestration of the components. As with Einstein's image of the closed watch, understanding life is inconceivable without observing the structures behind function. Microscopy plays an important role by placing the molecular elements into a structural context.

Because the pace of discovery of these elements has been increased substantially by proteomics, the need for more sophisticated microscopy has also substantially increased in recent years. Atomic force microscopy ( ) plays a specific role in life science microscopy by allowing imaging to be combined with locally probing functions of macromolecular elements. Most microscopes depend on radiation, emitted and recorded at a distance from the sample, using lenses. The resolution power of these microscopes in the life sciences is limited by diffraction and/or damage to the sample by the illuminating beam. Super resolution microscopies have pushed these boundaries. In contrast, the scanning tunneling microscope ( ) was the first microscope to rely on an effect only present in the immediate vicinity of a physical probe and the sample. In an STM, an electrically conductive needle approaches the sample until current flows. The current is based on a quantum-mechanical tunneling effect and flows before the tip and sample physically touch. During scanning, the tip–sample distance remains constant by keeping the tunneling current constant. The movement of the tip relative to the sample reflects the sample topography and is recorded. Not only has the atomic surface lattice structure been revealed for many crystalline samples, but also single atom defects have been imaged directly. The STM immediately created great interest among biologists because of its outstanding resolution power and the absence of radiation damage [31.1, 31.10, 31.2, 31.3, 31.4, 31.5, 31.6, 31.7, 31.8, 31.9]. However, the application of STM in the life sciences suffered from the poor electrical conductivity of most biological matter, and reproducible images were obtained only after metal-coating of the specimens [31.11, 31.12, 31.13, 31.14, 31.15, 31.2]. Nevertheless, early applications demonstrated the resolution power and high signal-to-noise ratio at the macromolecular level offered by the new microscopy. Although unintentional at first, it also became clear that the probe could be used to manipulate macromolecular structures individually (Fig. 31.1a,b).

Fig. 31.1a,b

STM topographies of Rec–DNA complex (a). (From [31.11]. Reprinted with permission from AAAS) and an aberrant bacteriophage capsid (b). (Reprinted from [31.12] Copyright 1989, with permission from Elsevier). In the upper right corner, a hole has been created by applying a voltage pulse between the tip and the sample. Both images immediately reveal the molecular arrangement of the respective complex structure

The basic principles of the STM were soon extended to form a suite of new devices, the scanning probe microscopes (SPM ). For all SPM, a physical probe is scanned over the sample in nanometer distance at most. Sample properties are mapped from a very small interaction volume in the near field of the sample and the probe. The meaning of the term of near field differs for the various types of SPM. In an AFM, short-range forces between a tip and the sample take the role of the tunneling current in the STM, thus making the new paradigm applicable to electrically nonconductive samples. This is how AFM gained ground in the life sciences.

The direct measurement of the position of the surface means that the achievable resolution is not limited by the environment—the AFM is well suited for imaging molecular and cellular specimens under physiological conditions, in buffer. In many cases, the sample preparation is very straightforward and at a basic level requires only the immobilization of the sample to be imaged. In the case of adherent cells, for instance, the sample can be imaged directly without any special preparation. The local measurement at the sample surface also allows direct quantitative measurement of dimensions and forces. In addition to the topographic image that is built up by scanning the tip over the surface, AFM can simultaneously probe the mechanical properties of the surface, by dynamically oscillating the cantilever that supports the tip. Tip–sample adhesion can also be probed, and maps of surface viscoelasticity or adhesion built up. One application is to coat the tip with appropriate molecules, such as antibodies, other proteins, or sugars to generate maps of specific ligand–receptor binding over the surface. The AFM can also be combined with other techniques, which means that the combination with optical microscopy is often particularly worthwhile for life sciences research. Other techniques, such as electrochemistry or electrophysiology, can also be used simultaneously, and in-situ probing of cell response to electrical, chemical, or mechanical stimulation is possible. The foundation of AFM in the life sciences has now been laid as a result of the profound understanding of the interactions between the tip and the sample and the possibility to subtly control the interactions by choosing an appropriate sensor and by adjusting the imaging environment [31.16, 31.17, 31.18].

To appreciate the contribution of AFM to the life sciences, it is necessary to consider it in relation to other microscopy techniques currently in use. In the realm of single molecules, the motivation for using AFM comes mainly from the high resolution attainable and the direct measurement of samples without coating. Most of the knowledge so far gained by AFM in the life sciences comes from studying single macromolecules or macromolecular complex structures. Biological membranes, for example, can be imaged in their native state at a lateral resolution of \(0.5{-}1\,{\mathrm{nm}}\) and a vertical resolution of \(0.1{-}0.2\,{\mathrm{nm}}\). Conformational changes that are related to functions can be resolved to a similar resolution. In the study of macromolecular structure, AFM competes with transmission electron microscopy. In many cases, the outcome is comparable and either microscope can be used with equal success. For example, measuring the contour length of a plasmid DNA or determining the binding site of a protein to a DNA may be performed by either microscope. However, solving the three-dimensional structure of macromolecules or macromolecular complexes that cannot be solved by x-ray crystallography or by nuclear magnetic resonance ( ) is better accomplished by cryo-electron microscopy, because AFM reveals only a topographical view of the structures. On the other hand, when a molecular assembly is not highly defined structurally, and individual units differ from each other substantially, AFM may allow a specific question to be answered in a straightforward manner, because of its uniquely high signal-to-noise ratio at molecular dimensions.

The strength of AFM does not lie in imaging alone but in the possibility of combining microscopy with an experiment at the molecular level. In a number of cases, molecular images have been obtained with sufficient resolution to individually recognize single macromolecules and then address the molecules individually with the stylus of the AFM (Fig. 31.2).

Fig. 31.2

The adenosine triphosphate (ATP ) synthase is a proton-driven molecular motor with a stator (seen here) and a rotor. Simply counting the number of individual proteins inside the ring structure was not possible with electron micrographs. From [31.19]

Most prominently, single-molecule force spectroscopy combined with single-molecule imaging has provided unprecedented possibilities for analyzing intramolecular and intermolecular forces, including the study of the folding pathway of proteins. Probing the self-assembly of macromolecular complexes and measuring the mechanical properties of macromolecular springs are other examples in which AFM has made substantial contributions to the life sciences.

Imaging living cells with AFM is also performed in conjunction with local probing of the sample. The images by themselves are usually used only to obtain proper experimental control. The extreme flexibility of the cell membrane means that the images often show a combination of the cell topography with the mechanical stiffness of cytoskeletal fibers and vesicles below the cell surface. An important example of a meaningful application of AFM with living cells is the measurement of the constrained diffusion of elements of the plasma membrane. These measurements have substantially contributed to establishing the nature of lipid rafts. In other applications, the cellular response to defined mechanical stress has been studied in conjunction with hearing or with the myogenic effect (Fig. 31.3).

Fig. 31.3

Neuronal cell, cultured on an electronic chip. The chip is designed such as to pick up an action potential of the cell. The image demonstrates proper tracing of the cell surface. In a future application, an appropriately designed stylus might be used as an additional electrode to excite or record an action potential at any location of the cell body or a process of the neuronal cell. From [31.20]

Over the last decade, a number of AFM developments have been introduced, which has made the technique almost ubiquitous in any life sciences lab. Performing local spectroscopy or manipulation of samples at the nanoscopic scale in addition to imaging is no longer a challenge. As a matter of fact, nowadays force spectroscopy is almost an order of magnitude faster, compared to what it used to be \(10{-}15\) years ago, which allows the mapping of the properties of the studied samples at almost imaging speed [31.21]. In addition, the AFM acquisition rates of near 1 frame per second (fps) on living cells, as well as video rates of over \({\mathrm{10}}\,{\mathrm{fps}}\) on single molecules, have made it possible to look at dynamic processes, and not just simple snapshots, as it used to be in the past. The improved cantilever sensitivity, models, and immobilization protocols have enabled us to further study the spatial relationships and interactions between macromolecules, cells, and tissues. Immunohistochemistry and immunocytochemistry, which used to be etalons for establishing the structural relationships in cells and tissues, are continuously being replaced by recognition AFM studies that allow reproducible identification of molecules, cells, and species. High-resolution cell images now give a clear picture of the apical cell ultrastructure, sometimes down to the molecular level. AFM is currently almost routinely applied to study folding/unfolding and, thus, reconstruct the energy landscape of individual multidomain molecules. Combined with the application of ultra-small and sub-pN-sensitive cantilevers [31.22, 31.23], it is almost possible to identify the structural transition events of individual molecules.

In the following section, AFM and its elements are described. The basics for preparing macromolecular and cellular samples are then described. Finally, a few selected examples highlighting AFM experiments are given, in which a combination of imaging with sample manipulation has been used to understand macromolecular or cellular function. A comprehensive review of all AFM applications in the life sciences is beyond the scope of this chapter.

31.1 Instrumentation and Imaging

In AFM, the topology of the sample is traced by a sharp stylus that is scanned line by line over the sample. For most setups, the stylus is located at the free end of a cantilever spring. Every elevation on the sample causes the stylus to move up and bend the cantilever upward, and every depression makes the lever move down. Stylus and cantilever are usually microfabricated from silicon or silicon nitride. The cantilever is typically a fraction of a millimeter long and a few micrometers thick. The softer the specimen, the softer the cantilever spring should be in order for it to trace the sample surface rather than deform it. The shape of the stylus is crucial too. It may be tetrahedral or extended, with a high aspect ratio, depending on the requirement. The radius of curvature at the apex of the stylus may be as small as \({\mathrm{2}}\,{\mathrm{nm}}\) (the apex of the stylus is also referred to as the tip) (Fig. 31.4).

Fig. 31.4

The basic working principle of AFM . Reprinted from [31.24]

Accurate measurement of the deflection of the cantilever is the basis for accurate measurement of the sample topology. It also allows proper control of the loading force of the stylus onto the sample. It turns out that this latter aspect is particularly important in life sciences applications of AFM (see below). The first AFMs used an STM behind the cantilever to measure the deflection. Deflection has also been measured by the change in electrical capacitance between the cantilever and a reference electrode or by means of a piezoresistor integrated with the cantilever [31.25]. Most AFMs now use an optical pointer to measure cantilever deflection. This detection system is fully adequate, as it poses no limitation to AFM resolution.

In the optical pointer detection, a laser beam is focused onto the back of the free end of the cantilever (Fig. 31.4). The laser beam is then reflected off the cantilever onto a four-segment photodiode. Prior to imaging, the four-segmented laser diode is aligned until all four segments are equally illuminated. For imaging, the stylus is then loaded onto the sample. This causes the free end of the cantilever to bend upward, and the laser beam now illuminates the two upper segments more strongly. The signals from the two upper segments of the diode are compared to the two lower segments \(\left[\left(A+B\right)-\left(C+D\right)\right]\) to derive the amount of deflection of the lever in the \(z\)-direction. The load is preset by the user, depending on the application, and is related to the deflection of the cantilever
$$F=k\Updelta S\;,$$
where \(\Updelta S\) is the deflection in the \(z\)-direction and \(k\) is the spring constant of the cantilever. The force is typically selected within the range of less than \({\mathrm{100}}\,{\mathrm{pN}}\) to a few nN, depending on the application.

In operation, the cantilever is deflected from the preset value by the sample topology, and the reflected laser beam is moved up or down. The original deflection is then restored via a feedback loop by a motion of the scanner perpendicular to the sample plane (referred to as the \(z\)-direction). The position of the scanner with respect to the tip is recorded and used as the topographical AFM image.

Torsion of the cantilever may also occur during scanning, when the tip experiences friction with the sample. When the cantilever becomes twisted, the laser beam is moved sideways. The amount of torsion and, hence, friction, is then derived from comparison of the signals from the two right-hand and the two left-hand segments of the photodiode \(\left[\left(A+C\right)-\left(B+D\right)\right]\). Maps of local friction are used to reveal materials contrast in addition to the topographical image.

In dynamic AFM modes , the cantilever is oscillated near its resonance frequency. An interaction with the sample is detected from the reduction of the amplitude or a phase shift with respect to the freely-oscillating cantilever.

An AFM does not necessarily need to be based on a cantilever at all. In an instrument combining AFM topographical imaging with near- field optical imaging ( or ), a tapered optical fiber is used as the stylus in most current instruments. It is oscillated parallel to the sample. Dampening of this oscillation is used as the feedback signal. In another alternative setup to the cantilever-based AFM, the sample is mounted on the membrane of an electret microphone, and the output of this microphone is used for feedback [31.26, 31.27]. This setup performs equally as well as the more traditional cantilever setup (Fig. 31.5a,b).

Fig. 31.5a,b

Actin filaments (a) (reprinted from [31.27] Copyright 1996, with permission from Elsevier) and an archebacterial S-layer (HPI-layer) (b) imaged with an AFM based on an electret sensor, rather than a cantilever sensor

In addition to a highly sensitive probe, AFM depends on a precise scanner. The scanner is attached either to the probe or the sample. It allows the sample to be scanned with respect to the stylus in the plane of the sample (referred to as the \(x,y\) plane) and adjusting the relative height of the sample and the probe (referred to as the \(z\)-direction) with subatomic accuracy [31.28]. AFM scanners are made of voltage-driven piezoceramic elements.

AFM imaging is a mechanical process, usually operated in a closed feedback loop. Even an apparently crisp, high-resolution topographical image need not necessarily reflect the true sample topography. The reliability of the topographical images depends on the properties of the feedback loop, the tip, and sample geometries, and on how much the sample deforms upon imaging. Eventually, the accuracy of the topographical images is limited by noise (Fig. 31.6). Each of these aspects is discussed below. Also discussed is how the development of AFM technology might lead to improved instrument performance.

Fig. 31.6

The reliability of an AFM topographical image depends on the sample and tip geometries, the sample deformation, the quality of the mechanical feedback loop of the instrument, and noise. From [31.20]

31.1.1 Geometry of the Stylus

When the stylus moves over a sample, the effective point of contact of the tip with the sample also changes. The topographical image is, colloquially speaking, convoluted with the tip geometry. The early days of AFM (and STM) were plagued by ill-characterized tips. Multiple whiskers created images that contained the same object multiple times (Fig. 31.7a,b).

Fig. 31.7a,b

A single active filament, imaged with a single tip (a) and after the tip has become a triple tip (b). From [31.20]

For a blunt stylus, each prominent object of the sample resulted in a local image of the tip itself rather than the local sample topology, and the overall appearance of such images was cloudy. Because both the tip geometry and the sample topology matter, high-resolution images were sometimes obtained with an apparent blunt tip for very flat samples. This is because even the bluntest of tips has a rough surface, covered with fine asperities.

Commercially available cantilevers now come with a well-characterized stylus, and the apex may have a very small radius of curvature (typically down to \({\mathrm{2}}\,{\mathrm{nm}}\)). The shape of the stylus needs to be selected while keeping the sample in mind. Biological membranes or arrays of proteins with little overall height variation, for example, are well imaged by a pyramid-shaped stylus that ends in an apex of a small radius of curvature, whereas a sample with a prominent topology with steep flanks needs to be scanned by an elongated, needle-like stylus of high aspect ratio. In cases where the hydrodynamic drag on the samples needs to be minimized, special cantilevers with extended beak-like indenters have been further developed, and so on. During imaging, even a well-characterized, sharp stylus may become mechanically damaged or pick up a contaminant that renders it blunt. In these cases, the probe needs to be cleaned (e. g., by washing in ultrapure water, containing a detergent). If this is not successful, it has to be replaced.

When selecting an appropriate stylus, not only the expected sample topology must be considered. The sample compliance is an equally important aspect, as a sharp tip may strongly deform a soft sample. This aspect is described below.

31.1.2 Tip–Sample Interaction

The tip–sample interactions need to be considered carefully, because they critically influence the success of the experiments. In AFM, most of the time the stylus is loaded onto the sample either intermittently (referred to as intermittent contact mode or tapping mode) or constantly (contact mode; see below). This is achieved by approaching the probe and sample and bending the cantilever until the desired loading force is achieved.

In addition to the loading force, exerted by the cantilever spring, there are additional forces \(F_{\text{StSa}}\) that act between the sample and the stylus (see below). They may be repulsive or attractive. An attractive force makes the effective load of the tip onto the sample greater than what would be assumed from the bending of the cantilever. A repulsive force that acts prior to physical contact delocalizes the load of the tip onto the sample. The total loading force of the tip onto the sample after contact becomes
$$F=k\Updelta S+F_{\text{StSa}}\;.$$

Evaluating these interactions may be pursued by acquiring force-versus-distance curves (referred to also as force spectroscopy ). Thereby, the tip is approached to the sample and the deflection of the cantilever recorded as a function of the vertical position of the scanner with respect to the sample. The deflection of the cantilever can then be converted into a force using the spring constant of the lever (Fig. 31.8a,b).

Fig. 31.8a,b

Force-versus-distance curve. For the example shown here, the tip first experiences a long-range repulsive force upon approaching the sample, even before the tip and sample are in physical contact. Close to the sample, the tip becomes strongly attracted by the van der Waals force . In this moment, the attractive force gradient becomes greater than the force gradient by the cantilever spring. This causes the tip to snap into physical contact with the sample (the perpendicular part of the approach curve). Once physical contact has been made, the cantilever is deflected linearly by the approaching scanner. On the way back, the tip may stick to the sample by adhesion until the pull by the cantilever forces it out of contact. After [31.20]

31.1.3 Sample Deformation

When physical contact between the tip and the sample is established, the sample deforms until the contact area has sufficiently increased such that the load is accommodated (Fig. 31.9).

Fig. 31.9

Parameter plot of the loading force \(F\) versus the penetration depth \(\sigma\) for three different radii of curvature of the apex \(R\) and for two types of samples with differing stiffness. A Young's modulus \(E\) of \({\mathrm{0.1}}\,{\mathrm{GPa}}\) might reflect the cytoskeletal components of a living cell, whereas \({\mathrm{1}}\,{\mathrm{GPa}}\) could be ascribed to hierarchical protein structures. The plot shows that even at a very low force, a sharp stylus dives right into the softer sample. It is, therefore, crucial to select the right kind of an apex radius for each application. After [31.20]

The deformation strongly determines the resolution and trustworthiness of AFM imaging and must, therefore, be considered carefully. Models relate the deformation of a solid body to the loading force of a stylus. According to Sneddon [31.17], for example, the repulsive force \(F\) for a stylus being loaded onto a solid, homogeneous body is
$$F=\frac{E_{\text{S}}}{2\left(1-v_{\text{S}}^{2}\right)}\left[\left(\eta^{2}-R^{2}\right)\ln\left(\frac{R+\eta}{R-\eta}\right)-2\eta R\right],$$
where \(E_{\text{s}}\) is Young's modulus, \(\nu_{\text{S}}\) is Poisson's ratio of the sample, and \(R\) is the apex radius of the stylus (the deformation of the stylus is neglected). With increasing loading force, the radius \(\eta\) of the contact area between the tip and the sample increases. The penetration depth \(\sigma\) of the tip and the radius of the contact area \(\eta\) are related [31.17]
$$\sigma=\frac{1}{2}\eta\ln\left(\frac{R+\eta}{R-\eta}\right).$$
Hence, for highest resolution, macromolecular samples need to be imaged at minimal load. Under optimal conditions, subnanometer scale resolution has been obtained on protein samples [31.29]; for a review [31.30]. Müller et al found on a two-dimensional regular array of the protein bacteriorhodopsin that at a load exceeding about \({\mathrm{100}}\,{\mathrm{pN}}\), the resolution dropped and the molecules became deformed both vertically and laterally [31.31]. Because the molecular lattice of bacteriorhodopsin is known, the change in topography could be assigned to a distinct conformational change of the protein upon the higher load (Fig. 31.10).
Fig. 31.10

Force-dependent surface topography of bacteriorhodopsin demonstrating the effect of force variations on the topography of the cytoplasmic purple membrane surface. The initial force of \({\mathrm{300}}\,{\mathrm{pN}}\) (bottom) was decreased during the scan to \({\mathrm{100}}\,{\mathrm{pN}}\) (top). A conformational change is distinct: donut-shaped bacteriorhodopsin trimers transform into units with three pronounced protrusions at their periphery. Inset: noise reduced image at higher magnification. Reprinted from [31.31], Copyright 1995, with permission from Elsevier

A convenient way to detect sample deformation upon too high a load is the comparison of images from trace and retrace [31.29, 31.32, 31.33]. Trace and retrace refer to the line-by-line motion of the scanner. All commercial AFMs allow two individual images to be acquired, one using the line traces when the stylus moves from left to right, and the other one using the traces on the way back.

The actual load leading to high-resolution images might be even smaller than expected from the force setting of the microscope. Yang and co-workers [31.34] proposed that short-range interactions with a local asperity give rise to high-resolution contrast, while longer-range interactions with blunter parts of the tip help support the load of the tip in contact with the sample. They have argued that in solution, the long-range force required for repulsion of the body of the tip is electrostatic. They adjust the supporting electrolyte so that the asperity just touches the sample lightly.

To obtain details of living cells and tissues by AFM, a meaningful image is often obtained only after a loading force of a few nanonewtons has been applied. Such high loads lead to deformation of the cell of up to several hundred nanometers [31.35]. The cell membrane is then pressed onto intracellular structures such as the nucleus, cytoskeletal elements, and vesicles, which in consequence become visible [31.35, 31.36, 31.37]. In these cases, the AFM images reflect the local plasticity of the living cells more than their true surface topography. The fast scanning developments discussed further below offer an interesting alternative for resolving the topographical features of individual cells by effectively reducing the time in which the cantilever tip intermittently interacts with the cell surface.

31.1.4 Forces Between the Apex of the Stylus and the Sample

A van der Waals force \(F_{\text{vdW}}\) is always present between the tip and the sample. The main contribution to \(F_{\text{vdW}}\) is the dispersion force, caused by the dipole-induced dipole interaction and is present between all kinds of materials. Lifshitz theory allows the forces between two geometrically shaped surfaces to be calculated. In the case of a flat surface (representing the sample) and a sphere (being used as an approximation for the apex of the stylus), \(F_{\text{vdW}}\) is [31.38]
$$F_{\text{vdW}}=\frac{-H_{a}R}{6d^{2}}\;,$$
where \(R\) is the radius of the tip (radius of the tip apex) and \(d\) is the distance between apex and sample; \(H_{\text{a}}\) represents the Hamaker constant , which characterizes the interaction of the two surfaces (media) across a third medium. For example, for two mica surfaces in water, \(H_{\text{a}}\) is \({\mathrm{2.2\times 10^{-20}}}\,{\mathrm{J}}\) and for two silicon oxide surfaces in water, \(H_{\text{a}}\) is \({\mathrm{8.3\times 10^{-21}}}\,{\mathrm{J}}\) [31.38]. For hydrocarbons in water, \(H_{\text{a}}\) lies between \(0.2{-}1.0E-20\,{\mathrm{J}}\) [31.39]. The van der Waals force between particles is always attractive in air and attractive for most situations in an aqueous solution.

When imaging in aqueous solution, additional interactions between the apex of the stylus and the sample need to be considered. Many of the commonly used supports and probes, as well as most biological samples are charged in an aqueous environment. This is because they usually carry weak acidic and basic functional groups. They dissociate in an aqueous solution, according to their equilibrium constants. The net charge density of a surface in water depends on the density of the functional groups, their p\(K\) values, and the pH of the buffer solution.

The Derjaguin, Landau, Verwey, Overbeek ( ) theory quantitatively describes the total force between charged interfaces in aqueous solution. It considers the electrostatic double-layer interaction caused by surface charges and the van der Waals forces and neglects entropic or steric contributions. Unlike the van der Waals interaction, the electrical double-layer repulsion depends on the sign and magnitude of the surface charge density, the ion concentration, and pH. A charged surface attracts counterions in water. At the solid–liquid interface, a charge cloud on the order of molecular dimensions is created as a transition region. In this so-called electrical double layer (EDL ), the counterions balance the charge of the surface. The density of the charges surrounding the surface falls off exponentially with distance \(z\) from the surface (Debye–Hückel approximation)
$$\psi=\psi_{0}\mathrm{e}^{-z/\lambda_{\text{D}}}\;.$$
\(\psi_{0}\) represents the potential at the surface. The Debye length \(\lambda_{\text{D}}\) is the thickness of the EDL
$$\lambda_{\text{D}}=\sqrt{\frac{\varepsilon_{0}\varepsilon_{\text{e}}k_{\text{B}}T}{e^{2}\sum_{i}c_{i}q_{i}^{2}}}\;,$$
where \(\varepsilon_{0}\) represents the vacuum permittivity, \(\varepsilon_{\text{e}}\) the dielectric permittivity of the electrolyte, \(k_{\text{B}}\) the Boltzmann constant, \(T\) the absolute temperature, \(e\) the unit charge, \(c_{i}\) the concentration, and \(q_{i}\) the ionic charge of the \(i\)-th component of the liquid. Note the strong dependence of the double layer thickness on the valence of the ions.
If the tip and the sample approach each other, the electrical double layers of the two interfaces become perturbed when they begin to overlap (Fig. 31.11a,b). This results in a force that is known as the double-layer force \(F_{\text{el}}\); \(F_{\text{el}}\) decreases exponentially with distance \(d\) between the two surfaces. For a stylus with an apex radius of curvature of \(R\) and a planar sample (at a surface potential \(<{\mathrm{50}}\,{\mathrm{mV}}\)) [31.39])
$$F_{\text{el}}=\frac{4\uppi R\sigma_{1}\sigma_{2}\lambda_{\text{D}}}{\varepsilon_{\text{e}}\varepsilon_{0}}\mathrm{e}^{-d/\lambda_{\text{D}}}\;,$$
where \(\sigma_{1}\) and \(\sigma_{2}\) represent the surface charge density of the stylus and the specimen, respectively, \(\varepsilon_{0}\) is the vacuum permittivity, \(\varepsilon_{\text{e}}\) the dielectric permittivity of the electrolyte, and \(d\) the distance between the two surfaces. This equation is a simplification of any real situation. At a distance much below \(\lambda_{\text{D}}\), it is necessary to resort to numerical solutions for which there are no simple expressions.
Fig. 31.11a,b

Specific interaction of ions at solid–liquid interfaces. The surfaces are negatively charged. (a) The ions produce an interfacial region of excess solute concentration, the electrical double layer, which consists of the Debye layer and the counterions bound at the surface (Helmholtz layer). (b) Two negatively charged surfaces at very small separation. The ion clouds of the electrical double layer overlap and cause a repulsive force. Reprinted from [31.40] Copyright 1997, with permission from Elsevier

In addition, when the two charged surfaces approach each other, the local ion concentration also changes. This means a shift in the equilibrium conditions of the charged groups with the ions in solution. Hence, the ionizable functional groups of the surface may become neutralized, according to the new equilibrium conditions, and the surface charge density is decreased. This phenomenon is called charge regulation and causes a less strong repulsive force for surfaces charged with a similar sign than would occur without charge regulation; \(F_{\text{el}}\) can even become attractive at very small distances [31.38].

To evaluate the overall force that is relevant for the interaction between the stylus and the sample, the electrical double-layer force and the van der Waals force are added together to get the total DLVO force (\(F_{\text{DLVO}}\)) (Fig. 31.12).

Fig. 31.12

Force versus distance curves showing the DVLO force between an alumina tip and a mica surface under varying pH conditions and at fixed ion concentration. Mica is naturally negatively charged at neutral pH, less charged at low pH, and increasingly charged at higher pH. At low surface charge, the van der Waals attraction dominates and causes the tip to snap onto the sample. At higher charge, the van der Waals attraction becomes increasingly screened by the repulsive electrical double-layer repulsion. Reprinted from [31.41], Copyright 1991, with permission from Elsevier

$$F_{\text{DLVO}}=\frac{4\uppi R\sigma_{1}\sigma_{2}\lambda_{\text{D}}}{\varepsilon\varepsilon_{0}}\mathrm{e}^{-d/\lambda_{\text{D}}}-\frac{H_{\text{a}}R}{6d^{2}}\;.$$

For high-resolution imaging, the electrical double-layer repulsion may have to be reduced such that the stylus can effectively come into physical contact with the sample rather than riding on the electrical double layer. This can be achieved by increasing the (bivalent) ion concentration [31.16, 31.39, 31.41, 31.42].

Additional interactions need to be considered to the DLVO force, including steric, hydrophobic, and hydrophilic interactions. When the microscopy is performed in air, often a water bridge occurs between the stylus and sample. This is because under ambient conditions, most surfaces are covered by a thin water layer. The resulting meniscus force may be strong (e. g., in the order of \({\mathrm{10^{-7}}}\,{\mathrm{N}}\)) and pull the tip effectively onto the sample. This will result in poor resolution and may even cause damage to the sample and stylus. It is notable that the van der Waals interaction in air is usually about an order of magnitude stronger than in an aqueous environment. The detrimental effects occurring in air through these forces may partially be overcome by choosing an appropriate imaging mode (intermittent contact mode, see below). The properties of the stylus may also be changed to avoid formation of a water bridge between sample and tip. Knapp et al have established a method for rendering the \(\mathrm{Si_{3}N_{4}}\) probes hydrophobic [31.43]. The probes were coated with a Teflon-like polymer by glow discharge in a hexafluoropropene (HFP, Hoechst AG, Frankfurt, Germany) atmosphere [31.43, 31.44].

31.1.5 The Feedback Loop of the AFM

The quality and speed of the imaging process of the AFM depend on the elements that take part in tracking a sample surface, united in the feedback loop. The elements of the feedback are the regulator, the cantilever spring, the scanner element that moves the sample up and down, as well as the interaction between the sample and the apex of the stylus and the sample topology itself (Fig. 31.13).

Fig. 31.13

The feedback loop of the AFM. The boxes symbolize the frequency response of the elements of the loop. After [31.20]

During scanning, the sample topology causes the stylus to move up or down, and the cantilever is deflected accordingly. The deflection is measured by means of the optical pointer and compared to the setpoint in the regulator. The setpoint represents the loading force chosen by the user. Deviation from the setpoint is relayed to the driver of the scanner, and the height of the sample with respect to the stylus is readjusted, etc. Hence, the scanning process translates the sample topology into a time-dependent signal. The signal may be decomposed (by the Fourier transform) into a spectrum of sine waves. Each sine wave of the spectrum is distinct based on its amplitude and frequency. The smaller the features are in the plane of the sample and the faster the probe is scanned over the sample, the higher the frequencies. The taller the features of the sample, the higher the amplitude of the respective frequencies. Each frequency is also distinct by its phase. The phase describes how much one sine wave is shifted with respect to all other sine waves. The quality of the feedback loop refers to how accurate phase and amplitude is relayed through feedback loop as a function of the frequency.

The maximum possible imaging rate depends on the feedback control speed (bandwidth), as well as sample fragility [31.45]. Assuming a sample with a spatial feature frequency of \(1/\lambda\) means that the \(z\)-feedback scan frequency required to track the topography should be \(f=v_{\text{s}}/\lambda\) (where \(v_{\text{s}}\) is the scanning velocity in the fast \(x\)-direction). To compensate for tracking of the rapidly changing topography signal, the feedback bandwidth \(f_{\text{b}}\) needs to be equal to or greater than \(f\). Therefore, for a sample with an area of \(W\times W\) and \(N\) lines in the slower \(y\)-direction, the imaging time \(t\) can be calculated by using the relationship \(t\geq 2WN/\left(\lambda f_{\text{b}}\right)\).

The \(f_{\text{b}}\) is limited by a number of time delays in the feedback loop, such as the time to measure the cantilever oscillation amplitude \(\tau_{\text{a}}\), cantilever response time \(\tau_{\text{c}}\), \(z\)-scanner reaction time \(\tau_{\text{s}}\), and the controller error integration time \(\tau_{i}\). Theoretical and experimental formulations for all of these are obtained [31.45, 31.46], given as
$$\tau_{\text{a}}=\frac{1}{2f_{\text{c}}}\;,\quad\tau_{\text{c}}=\frac{Q_{\text{c}}}{\uppi f_{\text{c}}}\;,\quad\tau_{\text{s}}=\frac{Q_{\text{s}}}{\uppi f_{\text{s}}}\;,$$
where \(f_{\text{c}}/f_{\text{s}}\) and \(Q_{\text{c}}/Q_{\text{s}}\) are the first resonant frequency and quality factor of the cantilever\(/z\)-scanner, respectively. Thus, the final form of \(f_{\text{b}}\) can be expressed as
$$f_{\text{b}}=\frac{\alpha f_{\text{c}}/8}{\left(1+\displaystyle\frac{2Q_{\text{c}}}{\uppi}+\displaystyle\frac{2Q_{\text{s}}f_{\text{c}}}{\uppi f_{\text{s}}}+2f_{\text{c}}(\tau_{p}+\tau_{i}+\delta)\right)}$$
including \(\delta\) as the sum of time delays other than those mentioned above.

The Scanner

The scanner is one of the elements of the feedback loop that can be limiting in terms of tracking speed and precision. Subatomic-precision scanners are the key to the success of STM, AFM, and nanotechnology in general. Such scanners are based on piezoceramic elements. They deflect in an electrical field. Within a certain range, the deflection is approximately linear to the applied voltage. A scanner is characterized by its scan range (\(x,y\)), the \(z\)-range, and the frequency response.

Over time, the scan range available for commercial AFMs has increased, particularly the \(z\)-range or distance the tip is able to move up and down over the surface. For cell imaging, a \(z\)-range of at least \(10{-}15\,{\mathrm{{\upmu}m}}\) is required, otherwise the tip will be unable to move high enough to scan over the cell nucleus. Larger \(x\) and \(y\) scan ranges of \({\mathrm{100}}\,{\mathrm{{\upmu}m}}\) or more are also generally required for cell imaging, and sometimes also for imaging much smaller samples, which may be distributed inhomogeneously over the surface. The piezoceramic material used for all AFM scanners suffers from various problems of nonlinearity, hysteresis, and creep [31.47]. Although it is possible to move the tip very precisely, this is against a large background of position changes due to longer-term effects in the piezo material, as it continues moving slowly for a long time after a voltage jump is applied (creep) or moves variable distances depending on its history (hysteresis). These problems have been addressed by adding position sensors (such as capacitive or strain-based sensors) along the movement axes, so that the tip movement is no longer set merely by converting the desired position into a simple voltage. The current position of the tip is constantly read by the position sensors and the nonlinearity of the piezo material, and its changes over time can be constantly corrected, using another feedback loop. These linearized piezo systems are also becoming available in commercial AFM systems and are likely to become standard.

There are two possibilities generally used for \(xy\) scanning: AFMs with tip-scanning or sample-scanning design. For general AFM, these are equivalent, and both have their advantages and disadvantages. In terms of optimizing an AFM for biological samples, tip scanning generally has clear advantages, as mentioned previously. On a basic practical level, much of AFM imaging for life sciences research has to take place in aqueous solutions, usually containing a reasonably high concentration of salt, and there is always a greater risk of expensive damage when high-voltage piezoelectronics is placed under the sample. Even when a tip scanner is used, the AFM must be carefully designed so that the electronics and piezo elements above the sample are protected from accidental spillage or water vapor. This is particularly important if the sample is to be heated to \({\mathrm{37}}\,{\mathrm{{}^{\circ}\mathrm{C}}}\), when the evaporation is significant, and condensed water vapor could easily collect in an unsealed AFM head. There is also a more fundamental problem with sample scanning, however, as simultaneous AFM and optical imaging cannot generally be performed. When the sample is moved in the \(z\)-direction, which usually lies along the optical axis, the objective lens can be tracked with the sample by using a piezo-actuated lens holder, but this is not possible for lateral scanning movements across the optical axis.

Large scan ranges negatively affect the frequency response; large scanners are slow to respond. The bandwidth (i. e., the usable frequency range) of the scanner is limited by its first mechanical resonance frequency (also natural frequency) in the \(z\)-direction, and this depends on the size and shape of the element as well as the material properties. Accurate sample tracking is achieved only when the actuator is used well below its resonant frequency. Close to the resonance, the sensitivity rises sharply and, even more importantly, the deflection becomes increasingly delayed with respect to the driving voltage (i. e., there is a negative phase shift between the driving voltage and the deflection). As discussed above, too much of a delay is not acceptable, because the probe can then no longer trace the relief of a specimen in real time.

The Probe and the Sample

The dynamic response of the probe is more difficult to understand than the frequency response of the scanner. Prior to establishing contact with the sample, the probe (cantilever) may be treated as a beam with one end fixed and one end free. Establishing the frequency response is straightforward, as described, for example, in [31.48]. The smaller and the stiffer the cantilever, the higher its first resonant frequency. Note that the frequency response of a free cantilever is always established prior to imaging in the intermittent contact mode. However, after establishing contact, the beam is in an intermediate state between a beam with both ends fixed and one end fixed. Moreover, contact to the sample is complicated. The cantilever and the sample are coupled via the force interaction between the tip and the sample and the compliant sample itself. Mechanically, the situation resembles a series of three dampened springs, the cantilever, the tip–sample interaction, and the viscoelastic sample. There is no analytical solution for this situation [31.49], and the frequency response is best determined experimentally. Smaller and stiffer cantilevers will still lead to a broader bandwidth of this element of the feedback loop.

Noise

The instrument's noise lowers the accuracy with which the topology is traced. In a well-designed instrument, out of all elements of the feedback, the thermal motion of the cantilever is the major source of noise. The cantilever produces one \(k_{\text{B}}T\) per resonance interval (\(k_{\text{B}}\) is Boltzmann's constant and \(T\) is the absolute temperature). This will typically cause the cantilever to oscillate by a few angstroms. When the stylus rests in firm contact with the sample, the thermal motion of the cantilever is strongly reduced and will not affect resolution directly. However, the thermal energy now creates a fluctuation of the loading force. In the case of a soft, susceptible sample, this in turn will create a marked fluctuation of the sample deformation. The thermal force noise within the context of the harmonic oscillator model is given by [31.50]
$$\Updelta F=23\sqrt{\frac{4k_{\text{B}}Bk}{\omega_{0}Q}}\;,$$
where \(\Updelta F\) is the RMS force noise , \(B\) the bandwidth at which the microscope operates, \(T\) the absolute temperature, \(k\) the spring constant of the cantilever, \(\omega_{0}\) the radial resonant frequency, and \(Q\) the mechanical quality factor; \(Q\) parameterizes the sharpness of the resonance, being the ratio of the resonant frequency to the full width at half height of the peak. In air or vacuum, very high \(Q\) (on the order of thousands) can be achieved. In liquid, \(Q\) is naturally much lower due to the hydrodynamic drag. For example, for a typical cantilever used to scan a molecular sample in an aqueous solution (\(k={\mathrm{0.1}}\,{\mathrm{N/m}}\), \(\omega_{0}={\mathrm{20}}\,{\mathrm{kHz}}\), \(Q=1\), \(B={\mathrm{10}}\,{\mathrm{kHz}}\)) the thermal force noise is \(5{-}20\,{\mathrm{pN}}\). From the equation above it follows that an ideal cantilever for low-noise operation should have a low spring constant, but at the same time a high resonant frequency. For a given size of a cantilever, however, the stiffer the lever, the higher the resonant frequency. This dilemma can be overcome by employing smaller (shorter) cantilevers [31.51]. The effect of the thermal noise on the image quality will also be reduced by operating the microscope at a lower bandwidth. However, this comes at the cost of reduced scan speed.

It is notable that noise is not evenly spread over the entire frequency spectrum. For large \(Q\), there is a large reduction of the thermal noise away from the resonant peak. In water, the thermal noise is more evenly distributed over the whole frequency range. At low frequencies, other sources of mechanical motion become progressively more important. This is referred to as \(1/f\) noise and is notable foremost in contact mode imaging (see below).

Improvements to the Speed and Accuracy of the AFM

Fast scanning and high-speed AFM has been made possible as a result of a range of developments that improve the feedback control and increase the feedback bandwidth to reduce the delays discussed in the previous sections at the controller level. Further improvements feature high-frequency \(z\)-scanners, as well as significantly smaller\(/\)shorter cantilevers.

While interacting with the sample, the cantilever reacts to the exerted force with a response time \(\tau_{\text{c}}\) (see above), and the time needed for measuring its oscillation amplitude is \(\tau_{\text{a}}\) [31.52]. In order to decrease both delays and apply this in a scenario with sensitive samples, it is necessary to minimize the cantilever force constant \(k_{\text{c}}\) and maximize the cantilever resonance frequency \(f_{\text{c}}\), which comes from reducing the cantilever dimensions
$$k_{\text{c}}=\frac{Ew}{4}\left(\frac{t}{l}\right)^{3},\quad f_{\text{c}}=\frac{(1.8751)^{2}}{2\uppi}\frac{t}{l^{2}}\sqrt{\frac{E}{12\rho}}\;,$$
where \(E\) is the Young's modulus of the cantilever material, \(w/t/l\) are the width/thickness/length of the cantilever, and \(\rho\) is the density. A direct result of the reduced size is that the thermal noise of short cantilevers \([({k_{\text{B}}}T)/{k_{\text{c}}}]^{1/2}\), in agreement with the equipartition theorem, is much lower. Thermal vibrations over larger \(f_{\text{c}}\) range (typically \(0\)\(2f_{\text{c}}\)) also result in a significantly lower thermal noise density, making such cantilevers \(5{-}10\)-fold quieter on average.
Another advantage of short cantilevers is their contribution to the overall increased angle deflection detection sensitivity \(\Updelta\varphi\) read out in optical beam deflection ( ) detectors, which for a given \(z\)-displacement at the free cantilever and (\(\Updelta z\)) is given by
$$\Updelta\varphi=\frac{3\Updelta z}{2l}\;.$$
Higher displacement sensitivity improves the tip–sample force control and also minimizes the invasiveness, particularly on sensitive and very dynamic samples [31.53, 31.54].

31.1.6 Imaging Modes

Various schemes/principles currently exist and are being used to classify the AFM imaging modes.

Contact Mode (Also Called DC Mode or Static Mode)

In contact mode, the tip is in permanent contact with the sample, while the vertical \(z\)-deflection of the cantilever is kept constant by the feedback loop. For relatively stiff samples, the \(z\)-position of the cantilever is well defined, only affected by the force noise. In biology, the method is typically applicable only to tightly-packed, relatively flat macromolecular structures, including two-dimensional protein crystals and membrane patches [31.53, 31.55]. These structures can withstand the high lateral force exerted by the tip.

Lateral forces will cause torsion of the cantilever and can be recorded as a second imaging channel to show sample inhomogeneity with respect to friction. In this way, it is possible to distinguish between regions of different chemical composition, density, or structure.

Dynamic AFM Modes

In dynamic AFM, the cantilever is oscillated with a certain amplitude over the sample, rather than being maintained at a constant vertical deflection. Different excitation approaches used include acoustic shaking of the cantilever by applying alternating current (AC ) voltage [31.56, 31.57], application of a magnetic field to ferromagnetically coated cantilevers [31.58, 31.59], or photothermal (laser) actuation that causes the bending of the cantilever [31.60]. In acoustic drive, an oscillator supplies a voltage drive to a piezoelectric actuator that generates sound waves in the cantilever holder. The frequencies are typically from tens of kilohertz to megahertz. When the driving frequency is near a bending-mode resonance of the cantilever, the cantilever is driven into an oscillation. A frequency scan with acoustic drive results in many peaks, only some of which yield a usable signal [31.51, 31.57]. This is because the transmission of the acoustic wave is not uniform over the frequency range. Furthermore, an advantage of the acoustic excitation is that it is also independent of the cantilever selection. In magnetic drive, the cantilever is coated with a magnetic film, or a magnetic particle is glued onto the end of the cantilever. The oscillating magnetic field of a solenoid in the vicinity of the cantilever causes the cantilever to oscillate. In the case of photothermal drive, an intensity-modulated laser periodically heats up the cantilever base and, thus, thermally induces stress in the cantilever, which leads to bending. The direct energy transfer in that case, avoids any spurious resonances, although a minimal heating of the sample due to the applied laser power of a few mW should be considered. In that case, the modulation laser wavelength is naturally in a spectrum range that is different than the detection laser.

In contrast to contact mode, which senses a force, dynamic AFM is a technique that senses an interfacial stiffness. As frequency is increased, dynamic AFM also becomes sensitive to increased interfacial damping. One way to define the tip motion for small amplitudes is to consider the cantilever-tip ensemble as a point-mass spring, where the motion can be approximated by a nonlinear second-order differential equation [31.61]
$$m\frac{\partial^{2}z(t)}{\partial t^{2}}+kz\left(t\right)+\frac{m\omega_{0}}{Q}\frac{\partial z\left(t\right)}{\partial t}=F_{\text{ts}}+F_{0}\cos\omega t\;,$$
where \(F_{0}\) and \(\omega\) are the amplitude and angular frequency of the driving force; \(Q\), \(\omega_{0}\) and \(k\) are the quality factor, angular resonance frequency, and force constant of the free cantilever, respectively. In the absence of tip–sample forces (\(F_{\text{ts}}\)), the equation above describes the motion of a forced harmonic oscillator with damping.
If a sinusoidal excitation of the cantilever base at external drive frequency \(\omega\) with an amplitude \(A\) is applied, then as a result, the cantilever tip will oscillate in the steady state around its equilibrium position [31.62] with
$$z\left(t\right)=Z(\omega)\sin\left[\omega t+\varphi(\omega)\right].$$

In dynamic AFM, the cantilever amplitude, its resonant frequency, and the phase shift of the oscillation, link the dynamics of the vibrating cantilever to the tip–surface interactions [31.61]. Each of them changes with respect to the gradient of attractive or repulsive forces acting during the oscillation cycles and can be used as a feedback parameter to measure the sample surface while the tip is in vicinity to the sample. For small amplitudes, especially in air, the attractive forces might prevent a stable oscillation of the cantilever and higher amplitudes are used, which is usually referred to as intermittent contact mode. During its oscillation, the tip drives through the whole potential, i. e., repulsive and attractive forces act on the tip, but far away from the sample there is nearly no interaction.

In an oversimplified form, the forces are attributed to either conservative forces associated with a shift in resonance frequency (i. e., van der Waals, polarization, electrostatic, London dispersion, Pauli repulsion), or dissipative interactions leading to change in the amplitude, or \(Q\) (i. e., plastic deformation, viscosity) [31.63]. A range of different experimental and analytical solutions has been suggested to separate the conservative and dissipative interactions during these nonlinear shifts, which are beyond the scope of this chapter [31.64, 31.65, 31.66, 31.67, 31.68, 31.69, 31.70, 31.71].

Amplitude modulation ( ) AFM is the most commonly used imaging mode in life sciences applications, in which the cantilever is driven with a constant amplitude close to its resonance frequency. The damping of the cantilever oscillation amplitude is used as a feedback mode to reconstruct the sample topography. The phase signal is used to get a compositional contrast of the sample. The frequency is kept fixed during imaging.

Phase modulation (PM ) AFM uses the phase shift between driving and oscillation signal as a constant feedback for imaging. In addition, it is possible to use an additional amplitude gain control (AGC ) loop to keep a constant amplitude signal during imaging. The additional output amplification applied to the \(z\)-piezo drive to compensate for the cantilever excitation amplitude damping is used as a parameter to evaluate the energy dissipation due to the tip–sample interaction forces [31.72].

Frequency modulation ( ) AFM uses the frequency shift, arising from the conservative tip–sample interactions, as a feedback to reconstruct the sample surface. The oscillation amplitude is kept constant, typically not more than \({\mathrm{1}}\,{\mathrm{nm}}\), at the selected cantilever resonance frequency. The excitation amplitude in that case provides details on the energy dissipation between tip and sample. The method allows the separation of long and short-range cantilever interactions. Real FM-AFM is typically only done in ultrahigh vacuum, although recent developments have also enabled its application to biological samples in liquid [31.73, 31.74].

Drive amplitude modulation ( ) AFM is a recent development, that builds on the feedback scheme with two nested loops. The first maintains a constant amplitude by regulating the driving force, and a second is used to reconstruct the topography based on the driving force. A parallel phase-locked-loop allows to separate between conservative and dissipative interactions. The feedback in DAM-AFM used to reconstruct topography is the driving force, unlike FM-AFM, which uses the frequency. It is free of the well-known feedback instability associated with the noncontact-to-contact transition that occurs in FM-AFM due to imminent tip contamination occurring on biological samples [31.65, 31.75].

Force Modulation Mode

Force modulation is a form of dynamic contact mode imaging. The tip contacts the surface, and the feedback loop maintains constant vertical deflection to construct the topographical image, just as for regular contact mode imaging. In addition, a low off-resonant oscillation is induced either to the tip or to the sample. The amplitude damping of the cantilever modulation scales with the relative viscoelastic properties of the sample.

Force-Controlled Mapping (Quantitative Imaging)

Conventional AFM modes, in particular static imaging, have well-known drawbacks for biological samples, which are typically high, exhibit steep edges, sticky, soft, and loosely attached. These drawbacks originate from the relatively high lateral forces, a problem partly overcome in dynamic AFM, but with the trade-off of sometimes having lower resolution. This is eliminated by the use of a force curve-based mapping of the sample, where the force between the preset force applied to the sample is accurately controlled. Recording the detailed vertical movement of the cantilever at every pixel with linear velocity, opens a range of possibilities for force curve analysis and extracting detailed topographical and mechanical information from the sample. The imaging speed of conventional force mapping is typically limited by the bandwidth of the z-scanner and the hydrodynamic drag exerted on the cantilever. Recent developments, including short/small cantilevers, and the use of high-resonance frequency scanners have enabled the simultaneous acquisition of topographic and quantitative mechanical information [31.21, 31.76].

High-Speed Imaging

The application of conventional AFM imaging for imaging of soft samples (particularly in liquid) is fundamentally limited to the use of very low imaging rates, due to the rather slow feedback of the available setups. Similarly, a range of dynamic processes, taking place on the second and even millisecond scale is impossible to capture. Technological developments in the last two decades have paved the way towards the application of small cantilevers with high resonant frequencies, piezoactuator-based sample scanners, and improved OBD detectors to meet the demand for faster \(xy\)-movement of the AFM cantilevers and improved tip–sample interaction force control [31.77, 31.78]. The higher oscillation frequencies of the cantilever, higher feedback bandwidth, and enhanced \(xy\)-movement allow us to increase the speed of the cantilever without decreasing the resolution of the measurement or damaging very sensitive samples.

31.1.7 Optimizing AFM Design for Life Sciences Applications

The range of applications for AFM is very large, from atomic resolution imaging on crystal surfaces to imaging or manipulation of whole cells. An AFM designed to work under high vacuum and to measure defects in atomic lattices has different basic requirements from one that is designed predominantly for samples in liquid, with full optical microscope integration being a priority. On biological samples, it is realistic to talk about submolecular (not atomic resolution); for samples of macromolecules and larger structures, the highest resolution is mainly defined by issues such as the tip–sample interaction and sample deformation, and is limited down to \({\mathrm{1}}\,{\mathrm{nm}}\). A number of challenges at that end are conveniently overcome by bringing samples into their native environment, which for biological molecules and cells is liquid. This furthermore, as previously discussed, reduces the number of conformational changes induced in the majority of samples. Then, technical issues such as integrated optics, flexibility of handling, and sample environment carry more weight.

Combining AFM and Optical Microscopy

Combining AFM with optical information has become a standard when working with biological samples above Abbe's resolution limit , as well as with single molecules carrying immunolabels. By combining both techniques, higher resolution structural information can be generated with AFM. This is subsequently correlated and combined with optical imaging to provide information about the composition, and subsequently the function, of the structures/molecules being studied [31.79]. The convenience of combining high-resolution AFM data with optical information in an inverted optical microscope ( ) has been previously demonstrated [31.80]. The choice of tip scanner design is crucial for applications where the optical image of the specimen needs to remain in focus during AFM measurements [31.81]. Accomplishing this in the case of a sample-scanning setup is more complicated, as the specimen is constantly moving in the optical image, therefore true simultaneous imaging is not really possible, and the features cannot be imaged without smearing. For a well-designed AFM, it is possible to simultaneously obtain high-contrast and high-resolution optical images, for a range of conventional and advanced widefield and confocal-based microscopy techniques. These include, but are not limited to, laser scanning microscopy ( ) [31.82, 31.83], total internal reflection fluorescent ( ) microscopy [31.84, 31.85], fluorescence correlation spectroscopy ( ) [31.86, 31.87], fluorescence resonance energy transfer ( ) [31.88]. Furthermore, AFM setups are now more and more commonly combined, with super-resolution fluorescence microscopy, as demonstrated for stimulated emission depletion ( ) [31.89] microscopy, structured illumination microscopy (SIM ), and stochastic optical reconstruction microscopy ( ) [31.24, 31.90]. The example given below shows a correlation between a confocal, STED, and an AFM map of microtubules in Cos7 cells measured in liquid (Fig. 31.14a-f).

Fig. 31.14a-f

Cos7 cells labeled with Atto 647 N. (a) Confocal image, (b) STED image (both raw data), (c)  rendered view of AFM measured height extracted from AFM force curves (45\({\times}\)45 pixelated scan), (d) confocal image and (e) STED image (both linear deconvoluted), (f) elasticity map calculated from AFM force curves. Reprinted from [31.89], Copyright 2012, with permission from Elsevier

To make sure that structural and optical information can be truly correlated, it is important to remove any distortions arising from lens aberrations and nonlinear alignment of mirrors in the optics system. Optical artifacts such as nonlinear stretching, rotating, and off-setting of light microscope images are present in nearly all types of optical setups. By taking advantage of the accuracy of already existing commercial closed-loop AFM systems, it is possible to correct for any lens imperfections and transfer the optical image to the calibrated AFM coordinate system [31.24]. In short, the cantilever is moved to a set of 25 points in real space, using the \(x,y\)-piezo scanners. At each point an optical image is acquired and, subsequently, the tip location within the optical image is automatically determined. A transform function is then calculated using both sets of 25 points, and this transform is applied to the optical image as it is imported into the SPM software [31.91]. This allows the immediate possibility to carry out all measurements directly by choosing locations from within the optical image.

In one sense, it is relatively straightforward to build a combined optical and atomic force microscope, since the AFM can be built around a standard inverted optical microscope. This allows the use of the optics through the objective lens under the sample (if the sample is transparent, which is often the case), while at the same time the AFM can approach from above. In practice, care must be taken that the entire system has the required mechanical stability, that sources of drift or vibration are minimized, and that the open nature required of such an AFM does not compromise its stability and reduce the resolution achievable. There are also two potential problems that must be considered in the instrument design that relate directly to the optical imaging: the optical quality of the support for the sample, through which the optical image must be formed, and the light path to illuminate the sample from above.

For high-quality optical images, the objective lenses are optimized for thin glass coverslips, which are mechanically unstable sample supports for AFM imaging. The working distance of high-magnification objectives is very small, so the objective lens must be positioned close to and beneath the sample, often with an oil or water droplet between the lens and the coverslip. These considerations must be addressed by the appropriate design of sample holders (which must generally also function as a liquid cell) that are able to provide enough mechanical support for the sample without interfering with access for the microscope objective beneath [31.92]. In addition, these are typically combined with \(130{-}170\,{\mathrm{{\upmu}m}}\) thick commercially available glass supports/surfaces (e. g., coverslips, or glass-bottom Petri dishes), which have various surface modifications (hydrophilic/hydrophobic), coating protocols (extracellular matrix proteins), and special designs (i. e., \({\mathrm{40}}\,{\mathrm{{\upmu}m}}\) elastic polymer films), which allow tailoring the sample preparation and imaging conditions to particular cellular types, biomolecules, etc.

For common biological samples such as cells or vesicles, there is very low optical contrast in brightfield illumination, so some enhanced optical contrast mechanism is required. The most common choices are phase contrast (which provides contrast based on the sample refractive index) or differential interference contrast ( , which is sensitive to local differences in the refractive index from point to point). In both cases, the path length and optical quality of any components in the illumination path are critical. To use the optical microscope in its standard configuration, and hence have the optimal optical system, the microscope requires illumination from above, using the condenser optics, long enough to accommodate the AFM head [31.93]. This means that the illumination light must pass through the AFM in some way. The ideal optical solution is to be able to use the condenser optics provided by the optical microscope manufacturer and build the AFM on an open frame around the optical path that does not interfere with the sample illumination.

Another issue for experiments involving simultaneous AFM and optical microscopy is the laser illumination used to measure the cantilever deflection. Although the majority of this is reflected by the cantilever, a significant proportion spills over the cantilever edges and can pass through the sample into the optical microscope image. For brightfield, phase contrast, or DIC this can be removed using a simple filter in the optical path to cut the required wavelength, but for fluorescence this can severely limit the wavelength regions available and, hence, the dyes that can be used for labeling. This is a particular problem when the sample should be labeled with more than one dye to identify different components, or in more advanced optical techniques such as FRET, where two fluorescent dyes are used simultaneously. Historically most AFMs used a detection laser in the red region of the optical spectrum, because of the availability of small, low-power red laser diodes. The standard for AFMs in life sciences has shifted to infrared laser illumination, which allows the full visible spectrum to be used for optical imaging. Using an infrared detection laser also results in better performance of the AFM itself, in that it removes artifacts caused by optical interference between the light reflected from the cantilever and from the sample in the path to the photodiode detector.

Sample Environment

In addition to the considerations already discussed about the need for using coverslips as a sample support for many optical imaging applications, there are other practical issues raised by life sciences samples. Temperature control of the sample is often important for studying molecular reactions or whole cells under physiological conditions, and, for example, many lipid bilayers used as model cell membranes undergo phase transitions over the range from around 10 to \({\mathrm{37}}\,{\mathrm{{}^{\circ}\mathrm{C}}}\) and above. Perfusion is also important, particularly for in-situ experiments and when introducing molecules for reactions, blocking, or changing the properties of the solution. For live cell work, it may also be necessary to allow a gas exchange to equilibrate the cell medium with \({\mathrm{5}}\%\) \(\mathrm{CO_{2}}\).

The priorities for the design of the cantilever and sample holders should also include the easy disassembly and cleaning of all the components that are in contact with the sample, so that, for example, ultrasonication or autoclaving is possible. When working in liquid, all parts of the instrument that come in contact with the fluid are sources of possible sample contamination. The issues vary somewhat depending on the applications. For instance, if the aim is to image or stimulate living cells, then molecular-sized contamination is not so important for AFM imaging itself, but the sample must remain sterile, and the cells must not be exposed to any chemical contaminants that will elicit a biological response. For single-molecule imaging, the molecules may not be affected by traces of certain ions leaching from metal surfaces, but every macromolecular contaminant that could adhere to the surface is a problem for AFM imaging.

31.2 Sample Preparation

To study their native structures and probe their functions, most cellular or macromolecular samples need to be kept in an aqueous environment. Hydrophilic and hydrophobic interactions promote correct folding of the polypeptide chains into a protein and are responsible for the formation of micelles, bilayers, and membranes from lipids and proteins. The conformation of membrane proteins is determined by hydrophobic interactions with lipid tails and hydrophilic interactions with their heads and the surrounding water [31.94, 31.95, 31.96]. The pH, electrolyte type, and its concentration and temperature also influence the structure and function. The function of macromolecular structures depends not only on their native conformation but often requires even more exacting environmental conditions with respect to pH and temperature and may depend on the presence of coenzymes or ATP, for example. Living cells are not only sensitive to pH, ionic strength, temperature, and \(\mathrm{CO_{2}}\) levels; they usually need a specific and often highly complex medium in which to grow. When biological structures are allowed to air dry, they are subjected to a high force caused by the change in surface tension as the water evaporates. The energy involved is considerable. As a result, even macromolecules become severely flattened and collapsed [31.100, 31.97, 31.98, 31.99]. For these reasons, most sample preparation techniques described below are for samples in buffer. Most of the time, immobilizing cells and macromolecular structures is all that is needed for AFM analysis. AFM analysis of macromolecular samples is demanding with respect to the necessary cleanliness of the support and the purity of the buffer. All surfaces immediately become covered with hydrocarbons when exposed to ambient air. Even double-distilled water can be a source of organic contaminants. A layer of these hydrocarbons on the sample or the probe could cause disturbances with AFM measurements.

As a result, the sample supports should be prepared or activated immediately before use. Ultrapure water (fresh Milli-Q water; \(\leq{\mathrm{18}}\,{\mathrm{M{\Upomega}{\,}cm^{-1}}}\)) should be used to prepare all buffer and rinsing solutions, because it contains fewer hydrocarbons and macroscopic contaminants than conventional bidistilled water.

In the following sections, we describe suitable supports and immobilization techniques for both macromolecular and cellular samples.

31.2.1 Macromolecular Samples

Immobilizing macromolecular structures aims at attaining a homogeneous distribution of the specimens in a close-to-native conformation. Tight binding of the biological specimens to the support surface will prevent them from clustering. They may also better withstand the forces that arise between the probe and the sample for most AFM. On the other hand, the structure can be substantially distorted and proteins may even denature by strong binding. This is well known from transmission electron microscopy ( ) as well as from STM of biological specimens [31.100, 31.101]. The specimens may adsorb with preferential orientations depending on the binding conditions [31.102, 31.103, 31.104, 31.33]. In addition to suitable binding properties, the support should be as smooth as possible, so that it does not interfere with the structure of the biological specimen in the final image. Furthermore, it should be relatively chemically inert to prevent contamination resulting from the solution or nonspecific reactions with the biological system.

Specimen Supports

Glass coverslips are widely used as an amorphous specimen support and can be used either unmodified or altered to change their physisorption or chemisorption properties. The surface can be almost featureless on the scale of macromolecular specimens. They are best suited for all experiments in which visible light is transmitted across the sample, as in SNOM or in the combined light and AFM setups. Before use, organic contaminants, dust, and other particles are removed by washing once with concentrated HCl\(/\)\(\mathrm{HNO_{3}}\) (3 : 1) and five times for \({\mathrm{1}}\,{\mathrm{min}}\) with Millipore water in an ultrasonic bath (\({\mathrm{50}}\,{\mathrm{kHz}}\)). This process makes the coverslips clean and smooth (RMS roughness \(\approx{\mathrm{0.5}}\,{\mathrm{nm}}\)).

The most commonly used support for imaging biological specimens in the AFM is mica. Mica minerals are characterized by their layered crystal structure and can be readily cleaved for a clean, atomically flat surface. Muscovite mica is the most commonly used form. The average surface charge density of muscovite mica in water is \(\sigma_{\text{m}}=-{\mathrm{0.0025}}\,{\mathrm{C/m^{2}}}\) (0.015 electron per surface unit cell).

Gold surfaces can be easily prepared by vapor deposition. They are chemically inert against \(\mathrm{O_{2}}\) and stable against radicals. They bind organic thiols or bifunctional disulfides with high affinity, which can be used to covalently attach biological macromolecules (see below). Hegner et al [31.105] developed a relatively simple and reliable method for preparing ultraflat gold [31.106] surfaces, consisting of atomically flat micron-sized terraces. Thin carbon films commonly used in TEM are smooth on molecular dimensions and adsorb macromolecules well when freshly prepared. High-vacuum carbon evaporators are common in electron microscopy laboratories.

Physisorption: DLVO Force

The most common technique for immobilizing biomolecules onto a support is by physisorption. Usually, the objects are immobilized out of an aqueous solution. They become attached to a support when there is an overall attractive force that pulls the surfaces into contact. The relevant force for adsorption is the DLVO force. Hydrophobic and hydrophilic interactions may also play a role. The DLVO force for two planar surfaces [31.38] per unit area is
$$\begin{aligned}\displaystyle F_{\text{DLVO}}&\displaystyle=F_{\text{el}}\left(z\right)+F_{\text{vdW}}\left(z\right)\\ \displaystyle&\displaystyle=\frac{2\sigma_{\text{su}}\sigma_{\text{sp}}}{\varepsilon_{o}\varepsilon_{\text{e}}}\mathrm{e}^{-z/\lambda_{\text{D}}}+\frac{-H_{\text{a}}}{6\uppi z^{3}}\;.\end{aligned}$$

The DLVO force between charged surfaces is highly susceptible to ion concentration, and conditions can thus be adjusted to achieve good adsorption [31.40] (Fig. 31.15).

Fig. 31.15

Dependence of the DLVO force on ion concentration (1\(:\)1 monovalent electrolyte) and distance between a macromolecular sample and a mica support. Whereas the attractive van der Waals force is mainly unaffected by the electrolyte, the double-layer repulsion decreases with increasing salt concentration. The surface charge densities were \(-{\mathrm{0.0025}}\,{\mathrm{C/m^{2}}}\) for mica [31.38] and \(-{\mathrm{0.05}}\,{\mathrm{C/m^{2}}}\) for purple membrane [31.39], respectively. The Hamaker constant was \({\mathrm{3\times 10^{-19}}}\,{\mathrm{J}}\). Reprinted from [31.40], Copyright 1997, with permission from Elsevier

When the electrical double-layer repulsion between the two surfaces has vanished, they rapidly coalesce to minimize the interaction energy [31.38]. For the example shown in Fig. 31.15, at electrolyte concentrations above \({\mathrm{50}}\,{\mathrm{mM}}\) KCl, the amount of adsorbed membranes rapidly increased and reached its maximum at about \({\mathrm{150}}\,{\mathrm{mM}}\) KCl. Note that adsorption occurred even though both mica and the sample carried a net negative charge.

Physisorption: Hydrophobic and Hydrophilic Interactions

There is an attractive interaction between hydrophobic surfaces in water. The attractive interaction potential is larger than the van der Waals potential and can be very long range. The nature of these long-range forces has not yet been fully elucidated. Hydrophilic molecules, on the other hand, tend to disorder the surrounding water molecules and prefer contact with water molecules. Hence, the molecules repel each other. These repulsive, hydrophilic forces are also referred to as hydration, structural, or solvation forces. They may cause the DLVO theory to fail at small distances between two hydrophilic surfaces. With respect to adsorption, hydrophobic molecules do not attach to a hydrophilic surface and vice versa. For example, hydrophilic purple membranes did not adsorb to highly hydrophobic supports such as derivatized glass. The hydrophilic and hydrophobic interactions can cause an oriented adsorption of molecular structures.

Physisorption: Preparation of the Support

With mica, an active surface is conveniently obtained by cleaving the layered mica crystals prior to specimen adsorption. For most other supports, the active surface cannot be produced so easily. These supports are usually covered by hydrocarbon contaminants and behave more or less hydrophobically. Glass, silicon wafers, and many thin films can be rendered hydrophilic by exposure to glow discharge (for example, in a plasma cleaner, \({\mathrm{1}}\,{\mathrm{min}}\), \(p={\mathrm{0.1}}\,{\mathrm{mbar}}\), with air as the residual gas) right before use. Thin carbon films become negatively charged. For those specimens that adsorb more efficiently to hydrophobic surfaces, a suitable silanization protocol might be better suited [31.107].

Coating is another way to improve physisorption on many specimen supports, and it has been used for a long time by electron microscopists [31.108, 31.109]. For example, poly-l-lysine can be used for coating glass and mica and to render the coated surfaces positively charged. This allows cells, tissues, and plasma membranes that are usually negatively charged to be readily adsorbed. Objects that carry charge in an uneven distribution can be adsorbed in a defined orientation on a poly-l-lysine-coated surface. For example, purple membrane mainly consists of a light-driven proton pump that builds up an electrochemical potential across the membrane. Illuminated by light, purple membranes show an asymmetric charge distribution and adsorb to polylysine-coated surfaces in an oriented fashion [31.102, 31.103]. More than \({\mathrm{90}}\%\) of the membranes attach with their cytoplasmic surface towards the poly-l-lysine under specific conditions (pH \(\approx 9\)). At pH below 4, most membranes (\({\mathrm{94}}\%\)) were directed with their extracellular surface toward the coated surface.

Chemical Bonding

Covalent bonding can be a very reliable technique to allow firm binding of biological specimens to a support. Some of the first high-resolution AFM topographs of protein structures in buffer solution were obtained using this technique [31.104]. It appears that covalent binding does not interfere with the macromolecular structure anymore than physisorption. Bonding of the macromolecular specimens can be accomplished using chemically modified supports. Karrasch et al [31.104] developed a protocol to crosslink biological systems to a silanized glass coverslip. The silane (APTES, Fluka Chemie AG, Buchs, Switzerland) contained a free amino group that allowed it to react with the succinimide ester group of the photo-crosslinker ANB-NOS (Fluka Chemie; \(\lambda={\mathrm{312}}\,{\mathrm{nm}}\)). Proteins were then bound to the interface by activating the photo-crosslinker with UV radiation. This method resulted in the first high-resolution images of protein structures by AFM in buffer (Fig. 31.16).

Fig. 31.16

Hexagonally packed intermediate layer. Reprinted from [31.104], Copyright 1993, with permission from Elsevier

Epitaxial gold surfaces can be effectively functionalized by alkanethiols. They form ordered, self-assembled monolayers that are tightly bound to the gold surface via chemisorption of the sulfur atoms. The monolayers are further stabilized by the lateral hydrophobic interactions of the alkyl chains [31.105, 31.110, 31.111]. The latter can carry head groups at the free end that allow oriented covalent anchoring of macromolecular structures [31.105, 31.112, 31.113]. Wagner et al bound protein structures via their amino groups with an N-hydroxysuccinimide-terminated monolayer on gold [31.114, 31.115].

Langmuir–Blodgett Films

There are amphiphilic substances that naturally form insoluble monomolecular films on an air–water interface. They exhibit a water-soluble polar or charged head group and a highly apolar tail. This causes them to attach to an air–water interface with the head group immersed in the water and the tail toward the air. The most prominent example is the pulmonary surfactant that forms at the interface of the respiratory gas lumen and the solvation layer that covers the alveolar epithelium of lungs. Surfactant layers can be formed ex vivo in a Langmuir trough to study their biophysical properties under defined conditions or for the purpose of microscopic examination. Langmuir films of lipids have also been used to mimic biological membranes [31.116], or they have served as a substrate to bind and crystallize proteins in two dimensions for TEM and AFM investigations [31.117]. AFM proved to be outstandingly well suited for studying the structure and mechanical properties of these thin layers.

To prepare films for microscopy, the amphiphilic substances are spread at the air–water interface of a Langmuir trough. They are then compressed by a movable barrier by a specified amount. To perform AFM on the air side of the film, the monolayers may be transferred from the air–water interface onto a solid support by slowly pulling a hydrophilic support out of the aqueous phase across the interface (Langmuir–Blodgett transfer; [31.118]). The film is deposited as the support is moved vertically across the air–water interface. It is then inspected by AFM in air. To do microscopy on the aqueous side of the film, the monolayer can also be deposited by dipping a hydrophobic substrate from the air side across the interface into the water. If a first lipid layer is deposited on the upstroke onto a hydrophilic substrate and then another layer added on the down stroke of the sample, a complete lipid bilayer has been formed. This bilayer may contain membrane proteins. It is interesting to note that deposition of a bilayer onto a mica substrate arrests the lipids of the first lipid layer. These lipids are no longer free to diffuse in the plane of the membrane. If the support is glass, both the lipids bound to the support and those within the second layer facing the aqueous phase are free to diffuse.

Finally, Langmuir–Blodgett transfer may not be necessary, and films of pulmonary surfactant have been studied directly at the air–water interface [31.119].

31.2.2 Cells

Successful immobilization of living cells largely depends on the cell type. There are cells with adherent growth (for example, epithelial cells, fibroblasts, or glial cells), and cells that grow in suspension without contact with a substrate (for example, bacterial cells or erythrocytes). Adhesive cells can be more readily imaged with the AFM, whereas cells that grow in suspension have to be immobilized to be imaged. It is notable that cells may change their shape, physiology, and even their life cycle once bound to a substrate. A variety of techniques have been developed to immobilize living cells. Cells are best imaged with an AFM that is combined with a light microscope.

Adsorbing Cells on Glass Coverslips

Cells that naturally adhere to a substrate can either be cultured on an appropriate support and subsequently imaged, or plated on the support and monitored shortly after they have established cell–substrate contact. For both procedures, the glass coverslip must be thoroughly cleaned. If cleaned with water, the glass support has to be dried in air or a stream of \(\mathrm{N_{2}}\) to prevent plated cells from possible osmotic shock. Coverslips have been coated with poly-l-lysine, collagen [31.120], proteoglycans, laminin, or fibronectin to improve adhesion.

For imaging individual, adherent cells with AFM, the density of the cell suspension has to be chosen such that enough space remains for the cells to spread out. The time required for the cells to attach and spread depends on the cell type. Before imaging, the samples have to be rinsed with buffer solution to remove cells that are not firmly attached and, if feasible, examined by conventional light microscopy. Specific cells that are cultured on a solid support spread out to a thickness of up to \(0.5{-}1.0\,{\mathrm{{\upmu}m}}\) over large areas in the periphery. In these thin regions it is possible to monitor the re-organization of the intracellular cytoskeleton with fast scanning. Alternatively, if temporal resolution is not a limiting factor, force mapping of larger cell areas can be easily accomplished over the full \(z\)-range of the conventional AFM scanners (up to \({\mathrm{15}}\,{\mathrm{{\upmu}m}}\)).

Immobilizing Nonadherent Cells

A stable immobilization of cells that grow in suspension and do not establish substrate interactions in their natural environment is difficult to obtain. Hörber et al [31.121] established a method to trap single cells by a micropipette and image the exposed part with the AFM. The setup makes it possible to use the advantages of the micropipette technique and to enhance the inner pressure of the cell. This is an advantage, because the spring constant of a cell surface may be very low (for example, \(\approx{\mathrm{0.002}}\,{\mathrm{N/m}}\)) [31.35]. Hence, the cell is extensively deformed by any reasonable interaction force with an AFM probe.

Permeable supports provide the possibility of measuring additional properties of cells (for example, permeability, diffusion, and voltage characteristics of the plasma membrane) while they are being imaged by AFM. The cells may attach onto substrates with a much smaller pore size than the average diameter of the cell, or individual cells may be trapped in pores that are only slightly smaller than the average cell diameter. Kasas and Ikai [31.122] used Millipore filters (Millipore PCF, Millipore Corp., Bedford, MA) with pore sizes similar to that of the cell diameter for trapping yeast cells. Hoh and Schoenenberger [31.35] cultured Madin–Darby Canine Kidney (MDCK) epithelial cells (average lateral diameter \(\approx{\mathrm{10}}\,{\mathrm{{\upmu}m}}\)) on polycarbonate filter supports (Millipore PCF, \({\mathrm{12}}\,{\mathrm{mm}}\) diameter) with a much smaller pore size (\({\mathrm{0.4}}\,{\mathrm{{\upmu}m}}\)) than the average cell diameter.

Attachment of Living Cells for Single-Cell Force Spectroscopy

In particular cases, rather than imaging, the topic of interest is the direct measurement of cell–cell [31.123, 31.124], or cell–substrate interactions [31.125]. In that case, a suitable immobilization strategy has to be introduced either directly on a tipless cantilever and/or on the substrate. The common immbolization molecules/chemistries feature either nonspecific binding to a group of polyphenols and lectins, such as concanavalin A, poly-l-dopamine, Cell-Tak, or functionalization with extracellular matrix molecules, specific for particular cell types, such as laminin, fibronectin, collagens, etc. In rare cases, such as work with T-cells, it is even necessary to coat the surfaces with antibodies in order to trigger cell attachment. The protocol for isolating individual cells, and their attachment to cantilevers to the state in which they are ready to be measured against other cells, is described in [31.123]. Briefly, after trypsin-induced detachment of cells from their cell culture plates, they are introduced to an interface with both adhesive (discussed above) and nonadhesive coating (e. g., bovine serum albumin (BSA )). Furthermore the cells are captured via the adhesive tipless cantilever, while sitting on the nonadhesive part of the substrate, and carried over the section of the substrate with immobilized cells to perform single-cell force spectroscopy ( ) measurements.

Recent developments, including hollow microchanneled cantilevers even allow the direct aspiration of individual cells and their use in continuous SCFS measurements [31.126, 31.127].

31.3 Imaging and Locally Probing Macromolecular and Cellular Samples: Examples

The fast scanning developments in the latest generation of AFM setups with the already mentioned technological and software improvements opens up a range of possibilities. The following examples show, but are not limited to, applications ranging from single molecules to living cells.

31.3.1 High-Resolution Imaging of Single Molecules

The application of AM-AFM in combination with fast scanning now conveniently allows users to achieve structural resolution that is comparable with the details coming from crystallographic data, as previously demonstrated for bacteriorhodopsin (BR ) and described in detail in [31.24]. The sample is arranged in a 2-D protein crystal of trimeric polypeptide molecules consisting of seven transmembrane alpha helices each and functions as a light-driven proton pump in the Halobacterium species. The vast majority of earlier high-resolution studies with this molecule reported in the literature, were performed exclusively with contact mode imaging [31.128].

Whereas, imaging of 2-D protein crystals in membrane patches is still something that might be possible with static AFM methods, this is not the case for much softer molecules, such as DNA . The double helical structure of DNA which, as reported in literature, is typically revealed by applying forces in the range of about \(50{-}100\,{\mathrm{pN}}\) [31.129]. Although such forces can be maintained with classical contact AFM, the lateral stress applied on the samples hinders the substructural DNA features. The above example with plasmid DNA in liquid was recorded with fast AM-AFM and allows us to resolve the sub-\({\mathrm{3.4}}\,{\mathrm{nm}}\) helical pitch over the entire double-stranded structure, containing a total of ten base pairs (Fig. 31.17).

Fig. 31.17

Topography images of soft double-helical pUC19 plasmid DNA recorded with AM-AFM in liquid on a polycathionic protein-coated mica

31.3.2 Studying Biomolecular and Cellular Dynamics

Living cells and biomolecular systems are highly dynamic, and most of the processes involved easily take place on the second, and even millisecond scale. Such dynamic processes are commonly addressed by optical techniques that offer sufficiently high temporal resolution, but cannot address the spatial resolution that is necessary to elucidate the structural features of individual molecules. At the same time, premium resolution techniques, such as electron microscopy, are not applicable to biological samples in their native or near-physiological environment. It is also necessary to emphasize that understanding individual processes requires both the compartmentalization of individual molecules and interaction partners, to enable high-resolution imaging, as well as a substantial decrease in the time necessary to a acquire a single image without compromising the resolution. Fast and high-speed AFM, with some sample prep considerations, is currently being directly applied to study protein aggregation kinetics, single-molecule dynamics, morphological changes in living cells and membranes, and many more, many of which have been excellently reviewed in [31.52, 31.53].

As representatives of one of the most abundant and extracellular matrix proteins, fibrillar collagens such as collagen type I have received a lot of attention over the last five decades, due to their large interactome, and hierarchical structural, and mechanical stability. The self-assembly process, though, remains incompletely understood. As previously demonstrated, amplitude-modulation fast AFM imaging has been successfully applied to noninvasively study and modify the kinetics of collagen type I fibrillar nanomatrix [31.54]. As shown in Fig. 31.18a-i below, by modifying the buffer compositions and pH value used, the fibrillogenesis can be tuned to enable optimal analysis.

Fig. 31.18a-i

Collagen kinetics observed at physiological pH. Collagen type I fibrillogenesis in PBS buffer (standard, with \({\mathrm{2}}\,{\mathrm{mM}}\) KCl) with pH 7.4 and monomer concentration of \({\mathrm{30}}\,{\mathrm{{\upmu}g/mL}}\) was imaged at the mica interface with a USC cantilever. The timestamps are relative to the beginning of the imaging in (a) (\(t_{0}\)). (b) Formation of initial \(300{-}600\,{\mathrm{nm}}\) long oligomeric intermediates of collagen type I with no D-banding, formed within \({\mathrm{41}}\,{\mathrm{s}}\) after the start of the imaging. (di) Representative frames of a consecutive set recorded at a line rate of \({\mathrm{15}}\,{\mathrm{Hz}}\) (resolution of 256\(\times\)256 pixels) over the original image location in (a) and (b) (before cropping) to reach a higher spatial resolution of the dynamic process. (c) A Boltzmann-sigmoidal fit of the dynamic coverage of the imaged area with collagen fibrils within the first 12 frames, representing the two characteristic phases of collagen I kinetics: nucleation and exponential phase. The different area size in (a) and (b) was normalized and plotted together with the relative covered area in the next 10 frames; \(x\)-scan size in (a,b) and (di) is 2 and \({\mathrm{1}}\,{\mathrm{{\upmu}m}}\), respectively, with a \(z\)-height for all images of \({\mathrm{3}}\,{\mathrm{nm}}\). Reprinted from [31.54], Copyright 2015, with permission from Elsevier

The lack of D-banding in Fig. 31.18a-i was previously associated with the importance of higher [\(\mathrm{K^{+}}\)] for the formation of the characteristic for collagen type I \({\mathrm{67}}\,{\mathrm{nm}}\) periodicity [31.130]. In fibrillar collagens, this periodicity is defined by the alignment and lateral staggering of monomers, typically driven by hydrophobic and electrostatic interactions and governed by the amino acid composition. In native structures, it shows an axial polarity, which is a result of asymmetric amino acid sequences and parallel staggering within the fibrils [31.131]. Typically, the substructure of the D-band in collagen I is visible after staining with heavy metals and sample preparation for high-resolution TEM studies. Interestingly, due to the mechanical clustering nature of the amino acid regions along the collagen molecules, such information is also possible to resolve by application of fast scanning and high resolution imaging as well (Fig. 31.19a,b).

Fig. 31.19a,b

Sub-D-periodic structure of collagen type I revealed by fast AM-AFM. The sub-banding of collagen type I reconstituted in PBS buffer (\({+}\)\({\mathrm{200}}\,{\mathrm{mM}}\) KCl) with pH 7.3 was recorded at \({\mathrm{15}}\,{\mathrm{Hz}}\) and resolution of 8192\(\times\)1532 pixels. The time for recording each of the images was \({\mathrm{102}}\,{\mathrm{s}}\); \(z\)-scale in the height (a) and phase (b) channel is \({\mathrm{2}}\,{\mathrm{nm}}\) and 3 degrees respectively. Reprinted from [31.54], Copyright 2015, with permission from Elsevier

In essence, this implicates fast AFM as a high-throughput data acquisition tool, as shown by the \(8192{\times}{\mathrm{1532}}\,{\mathrm{pixels}}\) image recorded over about \({\mathrm{2}}\,{\mathrm{min}}\). Since the sub-D-periodic structure of collagen exhibits a number of smaller bands, which could be arranged differently in different polymorphic or mutated forms of collagen types, this opens up a new possibility of using fast AFM as a high-resolution molecular fingerprinting technique as well.

Whereas, imaging of single molecules or flat macromolecular arrays is something that is typically accomplished on a height scale of a few nanometers, cells exhibit a strong heterogeneity in their surface roughness. Typically fast AFM in such cases is applicable to lamellipodia regions where the stronger cell–substrate contacts can be accommodated by the \(z\)-range of the instrument or directly over areas of the cell membrane. Conventional dynamic and static imaging modes are not very suitable, due to the comparatively long acquisition times and relatively slow feedback being unable to cope with the rather soft and topographically inhomogeneous samples. A characteristic system setup allows a simultaneous acquisition in cell culture medium, at controlled temperature, and optional perfusion of fluids or \(\mathrm{CO_{2}}\) for the case where such saturation is essential for the cells. Due to the transparent design of petri dishes, available also with glass bottom thickness suitable for high NA optical microscopy, it is possible to simultaneously acquire optical and AFM data. As can be seen from Fig. 31.20a,b, the cell periphery can be continuously imaged without affecting the end protrusions and contractile processes taking place under the cell membrane.

Fig. 31.20a,b

Realtime dynamics of living Chinese hamster ovary (CHO ) cells. (a) shows two consecutive frames, taken nearly 37 s apart, that show the multitude of cytoskeletal reorganization events taking place at the cell edge. The cells were kept at \({\mathrm{37}}\,{\mathrm{{}^{\circ}\mathrm{C}}}\) in DMEM/F-12 cell culture medium. The images in (b) depict a membrane ruffling event recorded with an acquisition rate of close to \({\mathrm{1}}\,{\mathrm{fps}}\). The line rate in that case was \({\mathrm{120}}\,{\mathrm{Hz}}\) for an image of \(2\times{\mathrm{1}}\,{\mathrm{{\upmu}m}}\) image (\(256\times 128\) pixels)

Magnifying over the surface of the cells allows us to further increase the acquisition speed during fast AM-AFM. As can be seen from Fig. 31.20a,b, it is possible to follow/track the direction of the plausible membrane ruffling processes, with a temporal resolution of \({\mathrm{1.1}}\,{\mathrm{s/frame}}\).

31.3.3 Macro and Nanomechanical Mapping of Single Molecules, Cells, and Tissues

Conventional force–distance spectroscopy has been routinely applied for nearly two decades in life sciences research to study the mechanics of single molecules, cells, and tissues. Depending on the heterogeneity of the substrate, the shape of the indenter can widely differ, but the most common types are pyramidal, conical, spherical, cylindrical, etc. [31.132]. In turn, there is a variation of contact mechanics models applied for quantifying the deformation of the materials, as reviewed in [31.133].

In the context of imaging though, one of the greatest advantages of the force–distance based approach, as previously discussed, is the theoretical elimination of lateral forces. During force-controlled mapping, the movement algorithm records a complete force curve at every pixel while only performing lateral movement between pixels. In the past, conventional force maps used to take hours, however as mentioned previously, recent technological developments have sped up the acquisition rates by close to an order of magnitude. This effectively makes it possible to acquire a complete map with a sufficiently high spatial resolution recorded at an imaging speed of a few minutes using a linear velocity movement of the cantilever within the entire force curve. The numerous options for proper force curve analysis and extraction of various material properties still use either standard or custom fit algorithms, namely work of adhesion, maximum adhesion, contact point determination, Young's modulus, etc. The real advantage for imaging is that the full force curve behind every pixel enables the recreation of the so-called zero-force image (having no indentation), or simply the topography and indentation information from the sample at different forces within the range of the applied set point (Fig. 31.21a-c), as shown for living cells.

Fig. 31.21a-c

Force-controlled of living Cos7 cells, and bacteriorhodopsin. The complete force curve behind every pixel was used to reconstruct the topography (a), as well as the sample indentation (b) at the full force range applied to the cells, imaged in cell medium at \({\mathrm{37}}\,{\mathrm{{}^{\circ}\mathrm{C}}}\). The nanodeformation of the subtrimeric structure of bacteriorhodopsin at different forces is shown in (c); \(z\)-ranges in (ac) are \(0{-}800\,{\mathrm{nm}}\), \(400{-}0\,{\mathrm{nm}}\), and \(0{-}600\,{\mathrm{pm}}\), respectively. Reprinted from [31.24]

This further enables the 3-D tomographic reconstruction of the sample structure and the correlation of the information coming from different channels. This is very interesting, as on multistructured/multilayered materials this would enable the determination of various hierarchical levels. In the context of living cells, if applied over the entire cell, these could be show an initial indentation of the cell membrane, further deformation of the cytoskeleton components, and eventually compression of the underlying organelles or nucleus.

The high sensitivity and versatility of this method becomes even more obvious when applied to single molecules. This is illustrated with the resolution of the subtrimeric structure of the purple membrane protein bacteriorhodopsin (Fig. 31.21a-c).

31.3.4 Spectroscopy Applications for Single Molecules and Cells

A particular strength of the AFM is that it is able to combine imaging modes sensitive to different properties of the sample with direct measurements of forces and interactions. These measurements can be carried out at particular points (selected, for instance, from a sample that has just been imaged), or built up over a grid to map the surface properties.

Force spectroscopy in AFM refers to the measurement of tip–sample interaction forces at a point as the height of the cantilever base is varied. This allows measurement of forces either pushing into the sample (elasticity, rheology, etc.), or pulling away from the sample (adhesion, including recognition of specific molecular binding, unfolding, or stretching of molecules bound between the tip and the sample surface). This has been reviewed in [31.134].

Single-Molecule Force Spectroscopy (SMFS)

is commonly applied for either detecting the adhesion forces between receptor-ligand pairs [31.135, 31.136], enzymatic catalysis, or unknown molecular interaction candidates. Pioneered more than two decades ago [31.137, 31.138], the technique has undergone a fundamental evolutionary leap involving different functionalization protocols for attachment of molecules. This involves attachment of molecules to the AFM tip, either directly or via linkers/spacers that negate the conformational changes induced by the direct binding of biomolecules to the AFM tip or cantilever surface. The investigated specimens are either immobilized in the form of arrays or are spatially confined in the form of 2-D crystals or patches.

Furthermore, in recent years, SMFS has also emerged as a tool to evaluate the mechanical stability and folding pathways of individual multidomain polyproteins. The basic concept here is that by applying a negative load on the individual segments, it is possible to evaluate the sawtooth pattern transitions for individual molecules between their folded (native) and unfolded (stressed) state [31.139]. By evaluating the folding/unfolding trajectories, it is possible to reconstruct all of the folding pathways of a molecule into a funnel-like free energy landscape with a number of energy minima. For example, this is a very common tool for studying the differences between wild type versus proteins with point-like mutations, the roles of intrinsically disordered proteins, and to test putative treatment agents in certain neurodegenerative disorders, such as Alzheimer's, Parkinson's, or Huntington's disease [31.140]).

Unfolding proteins and DNA and separating molecular binding partners have become important and active offspring of AFM, referred to as force spectroscopy. It provides insight into protein and DNA folding and conformation of the nature of binding pockets. Müller and his associates used this technique in conjunction with high-resolution imaging to gain insight into the interaction forces between the individual protomers of a regular bacterial surface layer, e. g., the hexagonally packed intermediate ( ) layer of Deinococcus radiodurans. After imaging the HPI layer, the AFM stylus was attached to individual protomers by enforced stylus–sample contact to enable force spectroscopy experiments. Imaging of the HPI layer after recording force-extension curves allowed unfolding forces to be correlated with the structures being unfolded. By using this approach, individual protomers of the HPI layer were found to be removed at pulling forces of \(\approx{\mathrm{300}}\,{\mathrm{pN}}\). Furthermore, it was possible to sequentially unzip entire bacterial pores formed by six HPI protomers (Fig. 31.22a-c).

Fig. 31.22a-c

The six protomers of an individual pore can be sequentially pulled out of the HPI layer. AFM topograph of the inner surface of the HPI layer prior to (a) and after (c) the pulling experiment. Note that one entire pore is missing. (b) The force-extension curve shows a sawtooth pattern with six force peaks of about \({\mathrm{300}}\,{\mathrm{pN}}\) each corresponding to the extraction of one protomer. The height of the force peaks corresponds to the binding force of a protomer to its neighbors; the stretching distance between protomer disruption events (\(7.3{\pm}{\mathrm{1.6}}\,{\mathrm{nm}}\)) corresponds to the length of the molecular linker connecting a protomer to its neighbors. From [31.30]

Single-Cell Force Spectroscopy (SCFS)

Single-cell force spectroscopy ( ) is a technique that measures the interactions between a cell, attached to a cantilever, and another interface of interest, which can be a substrate patterned with individual molecules [31.125] or another cell. The immobilization protocols for individual cells were described in Sect. 31.2.2, Attachment of Living Cells for Single-Cell Force Spectroscopy. SCFS has also been successfully applied to the study of cell–cell interactions. In such cases, following the attachment of a living cell to the cantilever, it is typically brought into contact with another cell for a short period of time, and the cantilever is retracted from the sample. This allows the measurement of either the adhesion work that is necessary to separate the cells, maximum adhesion, or even the study of individual receptor–ligand unbinding events. In a typical top-view setup on an inverted optical microscope, the cantilever optically blocks the apical side of the cell–cell contact (Fig. 31.23a-d).

Fig. 31.23a-d

Mirror-based side view optical setup applied in SCFS measurements with Xenopus laevis CNC cells. (a) Conventional AFM setup incorporating an inverted light microscope. For indentation force measurements, a microbead is immobilized on a tipless AFM cantilever and positioned above a single target cell. (b) Bright field image of a cantilever carrying a microbead above the target cell in standard view. (c) Mirror-based side view setup using lateral illumination. A pedestal is necessary to prevent the overhanging side view mirror from contacting the sample support during cantilever approach. (d) The microbead indenter immobilized on a V-shaped cantilever in contact with the target cell single cell in side view. Reproduced from [31.141] with permission from the Royal Society of Chemistry

This is particularly important if cells display blebbing behavior. In order to correlate mechanical, morphological, and adhesive information gained during AFM measurements, a mirror-based side view optical setup allows a more accurate assessment of the exact shape of the cell membrane contact area during indentation. This was recently demonstrated as shown in Fig. 31.24a-e.

Fig. 31.24a-e

Bleb formation promotes cell–cell detachment. (a) Side view time lapse image series of two cells separating at a \({\mathrm{0.2}}\,{\mathrm{{\upmu}m/s}}\) retraction speed (left) and the corresponding force–distance curve (right). The cell–cell contact area remains bleb free through the separation phase. (b) Side view time lapse image series of a cell pair detaching prematurely after bleb movement through the contact zone and the corresponding force–distance curve (right). (c) Magnified view of images from the image series at increased frame rate (\({\mathrm{10}}\,{\mathrm{s}}\)) immediately before cell–cell separation. Arrows indicate antiparallel bleb movement through the cell contact zone. Blebs are highlighted in red. (d) Mean cell–cell detachment forces and (e) mean cell–cell detachment times of bleb and bleb-free groups (mean \(\pm\) SD). Numbers above the bars indicate the number of cell pairs tested in each experiment. Reproduced from [31.141] with permission from the Royal Society of Chemistry

The results showed that the nonblebbing membranes (270\({\pm}\)\({\mathrm{140}}\,{\mathrm{Pa}}\)) are stiffer than blebbing membranes (170\({\pm}\)\({\mathrm{120}}\,{\mathrm{Pa}}\)). The stiffness of blebbing membranes is comparable in blebbistatin-treated cells (140\({\pm}\)\({\mathrm{120}}\,{\mathrm{Pa}}\)). This is consistent with the absence of a functional actin cytoskeleton in bleb protrusions. The single-cell force spectroscopy was used to quantitate 15-fold lower adhesion in blebbing versus nonblebbing cells. This supports the model that the softer and motile membrane blebs formed in the cell–cell contact area either promote cell–cell detachment or prevent adhesion reinforcement [31.141].

31.3.5 Substrate Manipulation: Pulmonary Surfactant

Alternatively, manipulation of the surface such as dissection or alignment is possible using the tip to apply forces and modify the sample. The ability to locally probe and manipulate a sample in addition to imaging is a unique strength of AFM. The potential experiments are as diverse as the research questions and address both isolated macromolecular structures and cells. This is demonstrated here for a few examples.

A mixed lipid–protein film of pulmonary surfactant covers the hydrated lung epithelia to the air. This highly cohesive and mechanically stable film reduces the otherwise high surface tension of the air–aqueous interface to almost zero, as is required for the structural stability of the alveolar lung and ease of breathing. The mechanical stability of the film depends on the formation of patterned monolayer domains of de-saturated phospholipids, intercepted by multilayer areas, containing the unsaturated lipids also present in surfactant. This molecular architecture is conveyed to the film by the surfactant-specific proteins SP-B and/or SP-C. The proteins crosslink the multilayer areas to the monolayer. This molecular arrangement was discovered by AFM [31.142]. First, a topographical image was obtained (Fig. 31.25a-c).

Fig. 31.25a-c

Functional pulmonary surfactant forms molecular films of molecular monolayer regions and areas of monolayers with stacks of lipid bilayers crosslinked to it (a). Crosslinking is revealed by physically removing the bilayers by an intentionally high load of the AFM probe (b,c). From [31.20]

Then, the loading force was increased so as to physically remove the lipid stacks on top of the molecular monolayer. After removal of the multilayers, the monolayer showed crevices on the circumference of the earlier multilayer patch and holes, evenly distributed over the area previously covered. This was indicative of a crosslinking structure that had also been removed by the stylus. Mere lipid layers, formed on top of a lipid monolayer in the absence of surfactant proteins, left no traces behind after being removed by the stylus.

Multilayers crosslinked to the monolayer, as opposed to multilayers merely adsorbed to the monolayer, appear to be crucial for surfactant function. At low surface tension, the interfacial film is subjected to high lateral pressure. Crosslinked multilayered areas take up the lateral load together with the monolayer. Otherwise, the load resides on the monolayer alone. However, lipid monolayer films containing unsaturated lipids are not able to withstand a high film pressure without collapsing. They, therefore, cannot reduce tension by a degree required for proper function of the lung. This condition is likely to be responsible for lung failure after acute lung injury where the surfactant is distinct by an elevated level of cholesterol [31.143]. Excess cholesterol prevents the formation of crosslinked multilayer stacks, and surface tension is high [31.144].

31.4 Outlook and Perspective

AFM has come a long way since the invention of the scanning probe technique in 1986. A number of milestones, such as the first image in liquid in 1987 [31.145], the first observation of a biomolecular process in 1989 [31.146], the first article on high-speed AFM in 1991 [31.147], the first high-res protein scan in liquid in 1994 [31.148], the first single protein unfolding paper in 1997 [31.139], the first cell–cell adhesion paper measured with AFM in 2000 [31.149], etc., are a testament to the continuous interest shown in the technique, which is continuously undergoing new developments. The possibility of combining imaging, applying forces, and studying dynamics with high-spatiotemporal resolution, at nearly physiological conditions, without explicit sample modifications, have arguably made the technique the best suited tool for studying biomolecules and cells at work [31.30]. The exponentially increased publication trend with AFM in biology [31.150] proves that, as Paul Hansma envisioned, the technique is a worthy successor to the already established and instrumental tools, such as electron and light microscopes. Even more, SPM instruments today offer full integration with inverted light optical microscopes, giving researchers the flexibility of utilizing the benefits of both techniques with the aim of achieving a unified correlative microscopy approach. Setting up crossdisciplinary seamless measuring systems, as shown by the integration of super-resolution microscopy with AFM, pushes the boundaries even further.

The future of AFM in the life sciences appears promising. The constant ongoing developments open up a range of possibilities for applying specialized imaging modes that help to unravel structural problems, study biomechanical phenomena, and dynamic processes on timescales that were not addressable by the AFMs of a decade ago. Perhaps most promising of all is the transfer of contemporary AFM instruments from predominantly physics/biophysics labs to molecular and cell biology labs in both academia and industry in the last \(10{-}15\) years. The technology transfer ensures that both basic scientific questions and clear and real biomedical needs can be addressed with ample speed and knowledge. The firm footing of AFM in the life sciences is hopefully going to mean that we are not far from the day when it can be directly applied to heal or cure diseases.

Notes

Acknowledgements

The authors would like to thank Oilibhe Pabsch and Detlef Knebel at JPK Instruments for inspiring discussions and critical reviews during the preparation of the manuscript.

References

  1. M. Amrein, R. Durr, A. Stasiak, H. Gross, G. Travaglini: Scanning tunneling microscopy of uncoated RecA–DNA complexes, Science 243(4899), 1708–1711 (1989)CrossRefGoogle Scholar
  2. M. Amrein, H. Gross, R. Guckenberger: STM of proteins and membranes. In: STM and SFM in Biology, ed. by O. Marti, M. Amrein (Academic Press, Cambridge 1993) pp. 127–175CrossRefGoogle Scholar
  3. P.G. Arscott, V.A. Bloomfield: STM of DNA and RNA. In: STM and SFM in Biology, ed. by O. Marti, M. Amrein (Academic Press, Cambridge 1993) pp. 259–272CrossRefGoogle Scholar
  4. R. Guckenberger, W. Wiegräbe, A. Hillebrand, T. Hartmann, Z. Wang, W. Baumeister: Scanning tunneling microscopy of a hydrated bacterial surface protein, Ultramicroscopy 31(3), 327–331 (1989)CrossRefGoogle Scholar
  5. S.M. Lindsay, N.J. Tao: Potentiostatic deposition of molecules for SXM. In: STM and SFM in Biology, ed. by O. Marti, M. Amrein (Academic Press, Cambridge 1993) pp. 229–257CrossRefGoogle Scholar
  6. M.J. Miles, T.J. McMaster: Protein assemblies and single molecules imaged by STM. In: STM and SFM in Biology, ed. by O. Marti, M. Amrein (Academic Press, Cambridge 1993) pp. 177–204CrossRefGoogle Scholar
  7. M.J. Miles, T. McMaster, H.J. Carr, A.S. Tatham, P.R. Shewry, J.M. Field, P.S. Belton, D. Jeenes, B. Hanley, M. Whittam, P. Cairns, V.J. Morris, N. Lambert: Scanning tunneling microscopy of biomolecules, J. Vac. Sci. Technol. A 8(1), 698–702 (1990)CrossRefGoogle Scholar
  8. A. Stemmer, A. Hefti, U. Aebi, A. Engel: Scanning tunneling and transmission electron microscopy on identical areas of biological specimens, Ultramicroscopy 30(3), 263–280 (1989)CrossRefGoogle Scholar
  9. M.A. Voelker, S.R. Hameroff, J.D. He, E.L. Dereniak, R.S. McCuskey, C.W. Schneiker, T.A. Chvapil, L.S. Bell, L.B. Weiss: STM imaging of molecular collagen and phospholipid membranes, J. Microsc. 152(2), 557–566 (1988)CrossRefGoogle Scholar
  10. M.E. Welland, M.J. Miles, N. Lambert, V.J. Morris, J.H. Coombs, J.B. Pethica: Structure of the globular protein vicilin revealed by scanning tunnelling microscopy, Int. J. Biol. Macromol. 11(1), 29–32 (1989)CrossRefGoogle Scholar
  11. M. Amrein, A. Stasiak, H. Gross, E. Stoll, G. Travaglini: Scanning tunneling microscopy of RecA-DNA complexes coated with a conducting film, Science 240(4851), 514–516 (1988)CrossRefGoogle Scholar
  12. M. Amrein, R. Dürr, H. Winkler, G. Travaglini, R. Wepf, H. Gross: STM of freeze-dried and Pt-Ir-C-coated bacteriophage T4 polyheads, J. Ultrastruct. Mol. Struct. Res. 102(2), 170–177 (1989)CrossRefGoogle Scholar
  13. B.L. Blackford: A metallic replica/anchoring technique for scanning tunneling microscope or atomic force microscope imaging of large biological structures, J. Vac. Sci. Technol. B 9(2), 1253 (1991)CrossRefGoogle Scholar
  14. G. Travaglini, H. Rohrer, M. Amrein, H. Gross: Scanning tunneling microscopy on biological matter, Surf. Sci. 181(1/2), 380–390 (1987)CrossRefGoogle Scholar
  15. J. Zasadzinski, J. Schneir, J. Gurley, V. Elings, P. Hansma: Scanning tunneling microscopy of freeze-fracture replicas of biomembranes, Science 239(4843), 1013–1015 (1988)CrossRefGoogle Scholar
  16. H.-J. Butt, M. Jaschke, W. Ducker: Measuring surface forces in aqueous electrolyte solution with the atomic force microscope, Bioelectrochem. Bioenerg. 38(1), 191–201 (1995)CrossRefGoogle Scholar
  17. M. Heuberger: Elastic deformations of tip and sample during atomic force microscope measurements, J. Vac. Sci. Technol. B 14(2), 1250 (1996)CrossRefGoogle Scholar
  18. A.L. Weisenhorn, P. Maivald, H.-J. Butt, P.K. Hansma: Measuring adhesion, attraction, and repulsion between surfaces in liquids with an atomic-force microscope, Phys. Rev. B 45(19), 11226–11232 (1992)CrossRefGoogle Scholar
  19. H. Seelert, A. Poetsch, N.A. Dencher, A. Engel, H. Stahlberg, D.J. Müller: Structural biology: proton-powered turbine of a plant motor, Nature 405(6785), 418–419 (2000)CrossRefGoogle Scholar
  20. M. Amrein: Atomic force microscopy in the life sciences. In: Science of Microscopy, ed. by P.W. Hawkes, J.C.H. Spence (Springer, New York 2007) pp. 1025–1069CrossRefGoogle Scholar
  21. C. Braunsmann, J. Seifert, J. Rheinlaender, T.E. Schäffer: High-speed force mapping on living cells with a small cantilever atomic force microscope, Rev. Sci. Instrum. 85(7), 073703 (2014)CrossRefGoogle Scholar
  22. D.T. Edwards, J.K. Faulk, A.W. Sanders, M.S. Bull, R. Walder, M.-A. LeBlanc, M.C. Sousa, T.T. Perkins: Optimizing 1-\({\upmu}\)s-resolution single-molecule force spectroscopy on a commercial atomic force microscope, Nano Lett. 15(10), 7091–7098 (2015)CrossRefGoogle Scholar
  23. D.T. Edwards, T.T. Perkins: Optimizing force spectroscopy by modifying commercial cantilevers: improved stability, precision, and temporal resolution, J. Struct. Biol. 197(1), 13–25 (2017)CrossRefGoogle Scholar
  24. D.R. Stamov, S.B. Kaemmer, A. Hermsdörfer, J. Barner, T. Jähnke, H. Haschke: BioScience AFM—capturing dynamics from single molecules to living cells, Microsc. Today 23(06), 18–25 (2015)CrossRefGoogle Scholar
  25. M. Tortonese, R.C. Barrett, C.F. Quate: Atomic resolution with an atomic force microscope using piezoresistive detection, Appl. Phys. Lett. 62(8), 834–836 (1993)CrossRefGoogle Scholar
  26. M. Schenk, M. Amrein, P. Dietz, R. Reichelt: Simultaneous STM/ AFM imaging of biological macromolecular assemblies: experimental set-up and first results. In: 13th Int. Congr. Electron Microsc. Ed. Phys. Ulis Paris (Editions de Physique, Les Ulis 2999) pp. 755–756Google Scholar
  27. M. Schenk, M. Amrein, R. Reichelt: An electret microphone as a force sensor for combined scanning probe microscopy, Ultramicroscopy 65(1/2), 109–118 (1996)CrossRefGoogle Scholar
  28. M. Amrein, A. von Nahmen, M. Sieber: A scanning force- and fluorescence light microscopy study of the structure and function of a model pulmonary surfactant, Eur. Biophys. J. 26(5), 349–357 (1997)CrossRefGoogle Scholar
  29. F.A. Schabert, A. Engel: Reproducible acquisition of Escherichia coli porin surface topographs by atomic force microscopy, Biophys. J. 67(6), 2394–2403 (1994)CrossRefGoogle Scholar
  30. A. Engel, D.J. Müller: Observing single biomolecules at work with the atomic force microscope, Nat. Struct. Biol. 7(9), 715–718 (2000)CrossRefGoogle Scholar
  31. D.J. Müller, G. Büldt, A. Engel: Force-induced conformational change of bacteriorhodopsin, J. Mol. Biol. 249(2), 239–243 (1995)CrossRefGoogle Scholar
  32. J.H. Hoh, G.E. Sosinsky, J.P. Revel, P.K. Hansma: Structure of the extracellular surface of the gap junction by atomic force microscopy, Biophys. J. 65(1), 149–163 (1993)CrossRefGoogle Scholar
  33. D.J. Müller, W. Baumeister, A. Engel: Conformational change of the hexagonally packed intermediate layer of Deinococcus radiodurans monitored by atomic force microscopy, J. Bacteriol. 178(11), 3025–3030 (1996)CrossRefGoogle Scholar
  34. J. Yang, J. Mou, J.-Y. Yuan, Z. Shao: The effect of deformation on the lateral resolution of atomic force microscopy, J. Microsc. 182(2), 106–113 (1996)CrossRefGoogle Scholar
  35. J.H. Hoh, C.A. Schoenenberger: Surface morphology and mechanical properties of MDCK monolayers by atomic force microscopy, J. Cell Sci. 107(Pt 5), 1105–1114 (1994)Google Scholar
  36. L. Chang, T. Kious, M. Yorgancioglu, D. Keller, J. Pfeiffer: Cytoskeleton of living, unstained cells imaged by scanning force microscopy, Biophys. J. 64(4), 1282–1286 (1993)CrossRefGoogle Scholar
  37. M. Fritz, M. Radmacher, H.E. Gaub: Granula motion and membrane spreading during activation of human platelets imaged by atomic force microscopy, Biophys. J. 66(5), 1328–1334 (1994)CrossRefGoogle Scholar
  38. J.N. Israelachvili: Intermolecular and Surface Forces, 12th edn. (Academic Press, London 2007)Google Scholar
  39. H.-J. Butt: Measuring local surface charge densities in electrolyte solutions with a scanning force microscope, Biophys. J. 63(2), 578–582 (1992)CrossRefGoogle Scholar
  40. D.J. Müller, M. Amrein, A. Engel: Adsorption of biological molecules to a solid support for scanning probe microscopy, J. Struct. Biol. 119(2), 172–188 (1997)CrossRefGoogle Scholar
  41. H.-J. Butt: Measuring electrostatic, van der Waals, and hydration forces in electrolyte solutions with an atomic force microscope, Biophys. J. 60(6), 1438–1444 (1991)CrossRefGoogle Scholar
  42. W.A. Ducker, T.J. Senden, R.M. Pashley: Direct measurement of colloidal forces using an atomic force microscope, Nature 353(6341), 239–241 (1991)CrossRefGoogle Scholar
  43. H.F. Knapp, W. Wiegräbe, M. Heim, R. Eschrich, R. Guckenberger: Atomic force microscope measurements and manipulation of Langmuir–Blodgett films with modified tips, Biophys. J. 69(2), 708–715 (1995)CrossRefGoogle Scholar
  44. R. Guckenberger, M. Heim, G. Cevc, H. Knapp, W. Wiegrabe, A. Hillebrand: Scanning tunneling microscopy of insulators and biological specimens based on lateral conductivity of ultrathin water films, Science 266(5190), 1538–1540 (1994)CrossRefGoogle Scholar
  45. T. Ando, T. Uchihashi, N. Kodera, M. Shibata, D. Yamamoto, H. Yamashita: High-speed AFM for observing dynamic processes in liquid. In: Atomic Force Microscopy in Liquid: Biological Applications, ed. by A.M. Baró, R.G. Reifenberger (Wiley-VCH, Weinheim 2012) pp. 189–209CrossRefGoogle Scholar
  46. T. Ando, T. Uchihashi, T. Fukuma: High-speed atomic force microscopy for nano-visualization of dynamic biomolecular processes, Prog. Surf. Sci. 83(7–9), 337–437 (2008)CrossRefGoogle Scholar
  47. P. Rahe, J. Schütte, W. Schniederberend, M. Reichling, M. Abe, Y. Sugimoto, A. Kühnle: Flexible drift-compensation system for precise 3D force mapping in severe drift environments, Rev. Sci. Instrum. 82(6), 063704 (2011)CrossRefGoogle Scholar
  48. C.A. Van Eysden, J.E. Sader: Resonant frequencies of a rectangular cantilever beam immersed in a fluid, J. Appl. Phys. 100(11), 114916 (2006)CrossRefGoogle Scholar
  49. Á. Budó, C. Haltenberger: Theoretische Mechanik, 12th edn. (Deutscher Verlag der Wissenschaften, Berlin 1990)Google Scholar
  50. D. Sarid: Scanning Force Microscopy: with Applications to Electric, Magnetic, and Atomic Forces (Oxford Univ. Press, New York 1994)Google Scholar
  51. T.E. Schäffer, J.P. Cleveland, F. Ohnesorge, D.A. Walters, P.K. Hansma: Studies of vibrating atomic force microscope cantilevers in liquid, J. Appl. Phys. 80(7), 3622–3627 (1996)CrossRefGoogle Scholar
  52. T. Ando, T. Uchihashi, N. Kodera: High-speed AFM and applications to biomolecular systems, Annu. Rev. Biophys. 42(1), 393–414 (2013)CrossRefGoogle Scholar
  53. T. Ando, T. Uchihashi, S. Scheuring: Filming biomolecular processes by high-speed atomic force microscopy, Chem. Rev. 114(6), 3120–3188 (2014)CrossRefGoogle Scholar
  54. D.R. Stamov, E. Stock, C.M. Franz, T. Jähnke, H. Haschke: Imaging collagen type I fibrillogenesis with high spatiotemporal resolution, Ultramicroscopy 149, 86–94 (2015)CrossRefGoogle Scholar
  55. D.J. Müller, D. Fotiadis, S. Scheuring, S.A. Müller, A. Engel: Electrostatically balanced subnanometer imaging of biological specimens by atomic force microscope, Biophys. J. 76(2), 1101–1111 (1999)CrossRefGoogle Scholar
  56. P.K. Hansma, J.P. Cleveland, M. Radmacher, D.A. Walters, P.E. Hillner, M. Bezanilla, M. Fritz, D. Vie, H.G. Hansma, C.B. Prater, J. Massie, L. Fukunaga, J. Gurley, V. Elings: Tapping mode atomic force microscopy in liquids, Appl. Phys. Lett. 64(13), 1738–1740 (1994)CrossRefGoogle Scholar
  57. C.A.J. Putman, K.O. Van der Werf, B.G. De Grooth, N.F. Van Hulst, J. Greve: Tapping mode atomic force microscopy in liquid, Appl. Phys. Lett. 64(18), 2454–2456 (1994)CrossRefGoogle Scholar
  58. M.A. Lantz, S.J. O'Shea, M.E. Welland: Force microscopy imaging in liquids using ac techniques, Appl. Phys. Lett. 65(4), 409–411 (1994)CrossRefGoogle Scholar
  59. W. Han, S.M. Lindsay, T. Jing: A magnetically driven oscillating probe microscope for operation in liquids, Appl. Phys. Lett. 69(26), 4111–4113 (1996)CrossRefGoogle Scholar
  60. D. Ramos, J. Tamayo, J. Mertens, M. Calleja: Photothermal excitation of microcantilevers in liquids, J. Appl. Phys. 99(12), 124904 (2006)CrossRefGoogle Scholar
  61. R. García: Dynamic atomic force microscopy methods, Surf. Sci. Rep. 47(6–8), 197–301 (2002)CrossRefGoogle Scholar
  62. B. Voigtländer: Amplitude modulation (AM) mode in dynamic atomic force microscopy. In: Scanning Probe Microscopy, ed. by B. Voigtländer (Springer, Berlin, Heidelberg 2015) pp. 187–204CrossRefGoogle Scholar
  63. U. Dürig: Conservative and dissipative interactions in dynamic force microscopy, Surf. Interface Anal. 27(5/6), 467–473 (1999)CrossRefGoogle Scholar
  64. F.J. Giessibl: A direct method to calculate tip–sample forces from frequency shifts in frequency-modulation atomic force microscopy, Appl. Phys. Lett. 78(1), 123–125 (2001)CrossRefGoogle Scholar
  65. M. Jaafar, D. Martínez-Martín, M. Cuenca, J. Melcher, A. Raman, J. Gómez-Herrero: Drive-amplitude-modulation atomic force microscopy: from vacuum to liquids, Beilstein J. Nanotechnol. 3, 336–344 (2012)CrossRefGoogle Scholar
  66. T. Fukuma, J.I. Kilpatrick, S.P. Jarvis: Phase modulation atomic force microscope with true atomic resolution, Rev. Sci. Instrum. 77(12), 123703 (2006)CrossRefGoogle Scholar
  67. B. Gotsmann, C. Seidel, B. Anczykowski, H. Fuchs: Conservative and dissipative tip–sample interaction forces probed with dynamic AFM, Phys. Rev. B 60(15), 11051–11061 (1999)CrossRefGoogle Scholar
  68. B. Gotsmann, H. Fuchs: Dynamic AFM using the FM technique with constant excitation amplitude, Appl. Surf. Sci. 188(3/4), 355–362 (2002)CrossRefGoogle Scholar
  69. K. Schröter, A. Petzold, T. Henze, T. Thurn-Albrecht: Quantitative analysis of scanning force microscopy data using harmonic models, Macromolecules 42(4), 1114–1124 (2009)CrossRefGoogle Scholar
  70. S.D. Solares, G. Chawla: Frequency response of higher cantilever eigenmodes in bimodal and trimodal tapping mode atomic force microscopy, Meas. Sci. Technol. 21(12), 125502 (2010)CrossRefGoogle Scholar
  71. G. Chawla, S.D. Solares: Mapping of conservative and dissipative interactions in bimodal atomic force microscopy using open-loop and phase-locked-loop control of the higher eigenmode, Appl. Phys. Lett. 99(7), 074103 (2011)CrossRefGoogle Scholar
  72. Y.J. Li, N. Kobayashi, Y. Naitoh, M. Kageshima, Y. Sugawara: Phase modulation atomic force microscopy in constant excitation mode capable of simultaneous imaging of topography and energy dissipation, Appl. Phys. Lett. 92(12), 121903 (2008)CrossRefGoogle Scholar
  73. H. Yamada, K. Kobayashi, T. Fukuma, Y. Hirata, T. Kajita, K. Matsushige: Molecular resolution imaging of protein molecules in liquid using frequency modulation atomic force microscopy, Appl. Phys. Express 2(9), 095007 (2009)CrossRefGoogle Scholar
  74. T. Fukuma: Advanced instrumentation of frequency modulation AFM for subnanometer-scale 2D/3D measurements at solid–liquid interfaces. In: Noncontact Atomic Force Microscopy, ed. by S. Morita, F.J. Giessibl, E. Meyer, R. Wiesendanger (Springer, Cham 2015) pp. 435–460CrossRefGoogle Scholar
  75. J. Melcher, D. Martínez-Martín, M. Jaafar, J. Gómez-Herrero, A. Raman: High-resolution dynamic atomic force microscopy in liquids with different feedback architectures, Beilstein J. Nanotechnol. 4, 153–163 (2013)CrossRefGoogle Scholar
  76. L. Chopinet, C. Formosa, M.P. Rols, R.E. Duval, E. Dague: Imaging living cells surface and quantifying its properties at high resolution using AFM in QI\({}^{\mathrm{TM}}\) mode, Micron 48, 26–33 (2013)CrossRefGoogle Scholar
  77. T. Ando, N. Kodera, E. Takai, D. Maruyama, K. Saito, A. Toda: A high-speed atomic force microscope for studying biological macromolecules, Proc. Natl. Acad. Sci. 98(22), 12468–12472 (2001)CrossRefGoogle Scholar
  78. L.M. Picco, L. Bozec, A. Ulcinas, D.J. Engledew, M. Antognozzi, M.A. Horton, M.J. Miles: Breaking the speed limit with atomic force microscopy, Nanotechnology 18(4), 044030 (2007)CrossRefGoogle Scholar
  79. A. Monserrate, S. Casado, C. Flors: Correlative atomic force microscopy and localization-based super-resolution microscopy: revealing labelling and image reconstruction artefacts, ChemPhysChem 15(4), 647–650 (2014)CrossRefGoogle Scholar
  80. M. Cascione, V. de Matteis, R. Rinaldi, S. Leporatti: Atomic force microscopy combined with optical microscopy for cells investigation: AFM combined with optical microscopy, Microsc. Res. Tech. 80(1), 109–123 (2017)CrossRefGoogle Scholar
  81. C. Löbbe, K. Poole: Two microscopes, one software: combining atomic force and optical microscopy, GIT Imaging Microsc. 8(4), 43–44 (2006)CrossRefGoogle Scholar
  82. B. Städler, T.M. Blättler, A. Franco-Obregón: Time-lapse imaging of in vitro myogenesis using atomic force microscopy, J. Microsc. 237(1), 63–69 (2010)CrossRefGoogle Scholar
  83. A. Mescola, S. Vella, M. Scotto, P. Gavazzo, C. Canale, A. Diaspro, A. Pagano, M. Vassalli: Probing cytoskeleton organisation of neuroblastoma cells with single-cell force spectroscopy, J. Mol. Recognit. 25(5), 270–277 (2012)CrossRefGoogle Scholar
  84. W. Christenson, I. Yermolenko, B. Plochberger, F. Camacho-Alanis, A. Ros, T.P. Ugarova, R. Ros: Combined single cell AFM manipulation and TIRFM for probing the molecular stability of multilayer fibrinogen matrices, Ultramicroscopy 136, 211–215 (2014)CrossRefGoogle Scholar
  85. J. te Riet, J. Helenius, N. Strohmeyer, A. Cambi, C.G. Figdor, D.J. Müller: Dynamic coupling of ALCAM to the actin cortex strengthens cell adhesion to CD6, J. Cell Sci. 127(7), 1595–1606 (2014)CrossRefGoogle Scholar
  86. M. Hilsch, B. Goldenbogen, C. Sieben, C.T. Höfer, J.P. Rabe, E. Klipp, A. Herrmann, S. Chiantia: Influenza A matrix protein M1 multimerizes upon binding to lipid membranes, Biophys. J. 107(4), 912–923 (2014)CrossRefGoogle Scholar
  87. L.L.G. Schwenen, R. Hubrich, D. Milovanovic, B. Geil, J. Yang, A. Kros, R. Jahn, C. Steinem: Resolving single membrane fusion events on planar pore-spanning membranes, Sci. Rep. 5(1), 1–15 (2015)CrossRefGoogle Scholar
  88. T. Kihara, C. Nakamura, M. Suzuki, S.-W. Han, K. Fukazawa, K. Ishihara, J. Miyake: Development of a method to evaluate caspase-3 activity in a single cell using a nanoneedle and a fluorescent probe, Biosens. Bioelectron. 25(1), 22–27 (2009)CrossRefGoogle Scholar
  89. B. Harke, J.V. Chacko, H. Haschke, C. Canale, A. Diaspro: A novel nanoscopic tool by combining AFM with STED microscopy, Opt. Nanoscopy 1(1), 3 (2012)CrossRefGoogle Scholar
  90. J.V. Chacko, F.C. Zanacchi, A. Diaspro: Probing cytoskeletal structures by coupling optical superresolution and AFM techniques for a correlative approach, Cytoskeleton 70(11), 729–740 (2013)CrossRefGoogle Scholar
  91. T. Jähnke, M. Haggerty: Method for the operation of a measurement system with a scanning probe microscope and a measurement system,  US Patent 8368017 B2 (2013)Google Scholar
  92. O. Sünwoldt, D. Knebel: Device for receiving a test sample, US Patent 8506909 B2 (2013)Google Scholar
  93. D. Knebel, T. Jähnke, O. Sünwoldt: Apparatus and method for a scanning probe microscope, US Patent 7473894 B2 (2009)Google Scholar
  94. A. Engel, A. Hoenger, A. Hefti, C. Henn, R.C. Ford, J. Kistler, M. Zulauf: Assembly of 2-D membrane protein crystals: Dynamics, crystal order, and fidelity of structure analysis by electron microscopy, J. Struct. Biol. 109(3), 219–234 (1992)CrossRefGoogle Scholar
  95. B.K. Jap, M. Zulauf, T. Scheybani, A. Hefti, W. Baumeister, U. Aebi, A. Engel: 2D crystallization: From art to science, Ultramicroscopy 46(1), 45–84 (1992)CrossRefGoogle Scholar
  96. T. Haltia, E. Freire: Forces and factors that contribute to the structural stability of membrane proteins, Biochim. Biophys. Acta Bioenerg. 1228(1), 1–27 (1995)CrossRefGoogle Scholar
  97. E. Kellenberger, J. Kistler: The physics of specimen preparation. In: Unconventional Electron Microscopy for Molecular Structure Determination, ed. by W. Hoppe, R. Mason (Vieweg+Teubner, Wiesbaden 1979) pp. 49–79CrossRefGoogle Scholar
  98. E. Kellenberger, M. Häner, M. Wurtz: The wrapping phenomenon in air-dried and negatively stained preparations, Ultramicroscopy 9(1/2), 139–150 (1982)CrossRefGoogle Scholar
  99. I. Wildhaber, H. Gross, A. Engel, W. Baumeister: The effects of air-drying and freeze-drying on the structure of a regular protein layer, Ultramicroscopy 16(3/4), 411–422 (1985)CrossRefGoogle Scholar
  100. W. Baumeister, M. Barth, R. Hegerl, R. Guckenberger, M. Hahn, W.O. Saxton: Three-dimensional structure of the regular surface layer (HPI layer) of Deinococcus radiodurans, J. Mol. Biol. 187(2), 241–250 (1986)CrossRefGoogle Scholar
  101. Z.H. Wang, T. Hartmann, W. Baumeister, R. Guckenberger: Thickness determination of biological samples with a zeta-calibrated scanning tunneling microscope, Proc. Natl. Acad. Sci. 87(23), 9343–9347 (1990)CrossRefGoogle Scholar
  102. K.A. Fisher: Oriented adsorption of purple membrane to cationic surfaces, J. Cell Biol. 77(2), 611–621 (1978)CrossRefGoogle Scholar
  103. S.B. Hayward, D.A. Grano, R.M. Glaeser, K.A. Fisher: Molecular orientation of bacteriorhodopsin within the purple membrane of Halobacterium halobium, Proc. Natl. Acad. Sci. 75(9), 4320–4324 (1978)CrossRefGoogle Scholar
  104. S. Karrasch, M. Dolder, F. Schabert, J. Ramsden, A. Engel: Covalent binding of biological samples to solid supports for scanning probe microscopy in buffer solution, Biophys. J. 65(6), 2437–2446 (1993)CrossRefGoogle Scholar
  105. M. Hegner, P. Wagner, G. Semenza: Immobilizing DNA on gold via thiol modification for atomic force microscopy imaging in buffer solutions, FEBS Letters 336(3), 452–456 (1993)CrossRefGoogle Scholar
  106. D. Knebel, M. Amrein, K. Voigt, R. Reichelt: A fast and versatile scan unit for scanning probe microscopy, Scanning 19(4), 264–268 (1997)CrossRefGoogle Scholar
  107. L.S. Shlyakhtenko, A.A. Gall, A. Filonov, Z. Cerovac, A. Lushnikov, Y.L. Lyubchenko: Silatrane-based surface chemistry for immobilization of DNA, protein–DNA complexes and other biological materials, Ultramicroscopy 97(1–4), 279–287 (2003)CrossRefGoogle Scholar
  108. D. Mazia: Adhesion of cells to surfaces coated with polylysine. Applications to electron microscopy, J. Cell Biol. 66(1), 198–200 (1975)CrossRefGoogle Scholar
  109. B. Jacobson, D. Branton: Plasma membrane: Rapid isolation and exposure of the cytoplasmic surface by use of positively charged beads, Science 195(4275), 302–304 (1977)CrossRefGoogle Scholar
  110. H. Sellers, A. Ulman, Y. Shnidman, J.E. Eilers: Structure and binding of alkanethiolates on gold and silver surfaces: Implications for self-assembled monolayers, J. Am. Chem. Soc. 115(21), 9389–9401 (1993)CrossRefGoogle Scholar
  111. H. Wolf, H. Ringsdorf, E. Delamarche, T. Takami, H. Kang, B. Michel, C. Gerber, M. Jaschke, H.-J. Butt, E. Bamberg: End-group-dominated molecular order in self-assembled monolayers, J. Phys. Chem. 99(18), 7102–7107 (1995)CrossRefGoogle Scholar
  112. D.P. Allison, T. Thundat, K.B. Jacobson, L.A. Bottomley, R.J. Warmack: Imaging entire genetically functional DNA molecules with the scanning tunneling microscope, J. Vac. Sci. Technol. A 11(4), 816–819 (1993)CrossRefGoogle Scholar
  113. M. Hegner: Modified DNA immobilized on bioreactive self-assembled monolayer on gold for dynamic force microscopy imaging in aqueous buffer solution, J. Vac. Sci. Technol. B 14(2), 1418 (1996)CrossRefGoogle Scholar
  114. P. Wagner, M. Hegner, H.-J. Guentherodt, G. Semenza: Formation and in situ modification of monolayers chemisorbed on ultraflat template-stripped gold surfaces, Langmuir 11(10), 3867–3875 (1995)CrossRefGoogle Scholar
  115. P. Wagner, M. Hegner, P. Kernen, F. Zaugg, G. Semenza: Covalent immobilization of native biomolecules onto Au(111) via N-hydroxysuccinimide ester functionalized self-assembled monolayers for scanning probe microscopy, Biophys. J. 70(5), 2052–2066 (1996)CrossRefGoogle Scholar
  116. H. Bader, K. Dorn, B. Hupfer, H. Ringsdorf: Polymeric monolayers and liposomes as models for biomembranes. In: Polymer Membranes, ed. by M. Gordon (Springer, Berlin, Heidelberg 1985) pp. 1–62Google Scholar
  117. A. Brisson, A. Olofsson, P. Ringler: Two-dimensional crystallization of proteins on planar lipid films and structure determination by electron crystallography, Biol. Cell 80, 221–228 (1994)Google Scholar
  118. K.B. Blodgett, I. Langmuir: Built-up films of barium stearate and their optical properties, Phys. Rev. 51(11), 964–982 (1937)CrossRefGoogle Scholar
  119. D. Knebel, M. Sieber, R. Reichelt, H.-J. Galla, M. Amrein: Scanning force microscopy at the air–water interface of an air bubble coated with pulmonary surfactant, Biophys. J. 82(1), 474–480 (2002)CrossRefGoogle Scholar
  120. E. Henderson, P. Haydon, D. Sakaguchi: Actin filament dynamics in living glial cells imaged by atomic force microscopy, Science 257(5078), 1944–1946 (1992)CrossRefGoogle Scholar
  121. J.K. Hörber, W. Häberle, F. Ohnesorge, G. Binnig, H.G. Liebich, C.P. Czerny, H. Mahnel, A. Mayr: Investigation of living cells in the nanometer regime with the scanning force microscope, Scanning Microsc. 6(4), 919–929 (1992), discussion 929–930Google Scholar
  122. S. Kasas, A. Ikai: A method for anchoring round shaped cells for atomic force microscope imaging, Biophys. J. 68(5), 1678–1680 (1995)CrossRefGoogle Scholar
  123. C.M. Franz, A. Taubenberger, P.-H. Puech, D.J. Muller: Studying integrin-mediated cell adhesion at the single-molecule level using AFM force spectroscopy, Science's STKE 2007(406), pl5 (2007)CrossRefGoogle Scholar
  124. J. Kashef, C.M. Franz: Quantitative methods for analyzing cell–cell adhesion in development, Dev. Biol. 401(1), 165–174 (2015)CrossRefGoogle Scholar
  125. L. Dao, U. Weiland, M. Hauser, I. Nazarenko, H. Kalt, M. Bastmeyer, C.M. Franz: Revealing non-genetic adhesive variations in clonal populations by comparative single-cell force spectroscopy, Exp. Cell Res. 318(17), 2155–2167 (2012)CrossRefGoogle Scholar
  126. E. Potthoff, O. Guillaume-Gentil, D. Ossola, J. Polesel-Maris, S. LeibundGut-Landmann, T. Zambelli, J.A. Vorholt: Rapid and serial quantification of adhesion forces of yeast and mammalian cells, PLoS ONE 7(12), e52712 (2012)CrossRefGoogle Scholar
  127. A. Sancho, I. Vandersmissen, S. Craps, A. Luttun, J. Groll: A new strategy to measure intercellular adhesion forces in mature cell–cell contacts, Sci. Rep. 7, 46152 (2017)CrossRefGoogle Scholar
  128. D.J. Müller, F.A. Schabert, G. Büldt, A. Engel: Imaging purple membranes in aqueous solutions at sub-nanometer resolution by atomic force microscopy, Biophys. J. 68(5), 1681–1686 (1995)CrossRefGoogle Scholar
  129. A. Pyne, R. Thompson, C. Leung, D. Roy, B.W. Hoogenboom: Single-molecule reconstruction of oligonucleotide secondary structure by atomic force microscopy, Small 10(16), 3257–3261 (2014)CrossRefGoogle Scholar
  130. F. Jiang, H. Hörber, J. Howard, D.J. Müller: Assembly of collagen into microribbons: effects of pH and electrolytes, J. Struct. Biol. 148(3), 268–278 (2004)CrossRefGoogle Scholar
  131. D.J.S. Hulmes, A. Miller, D.A.D. Parry, K.A. Piez, J. Woodhead-Galloway: Analysis of the primary structure of collagen for the origins of molecular packing, J. Mol. Biol. 79(1), 137–148 (1973)CrossRefGoogle Scholar
  132. I.N. Sneddon: The relation between load and penetration in the axisymmetric Boussinesq problem for a punch of arbitrary profile, Int. J. Eng. Sci. 3(1), 47–57 (1965)CrossRefGoogle Scholar
  133. J.J. Roa, G. Oncins, J. Diaz, F. Sanz, M. Segarra: Calculation of Young's modulus value by means of AFM, Recent Pat. Nanotechnol. 5(1), 27–36 (2011)CrossRefGoogle Scholar
  134. D. Alsteens, H.E. Gaub, R. Newton, M. Pfreundschuh, C. Gerber, D.J. Müller: Atomic force microscopy-based characterization and design of biointerfaces, Nat. Rev. Mater 2, 17008 (2017)CrossRefGoogle Scholar
  135. P.-H. Puech, D. Nevoltris, P. Robert, L. Limozin, C. Boyer, P. Bongrand: Force measurements of TCR/pMHC recognition at T cell surface, PLoS ONE 6(7), e22344 (2011)CrossRefGoogle Scholar
  136. Y. Lee, Y. Kim, D. Lee, D. Roy, J.W. Park: Quantification of fewer than ten copies of a DNA biomarker without amplification or labeling, J. Am. Chem. Soc. 138(22), 7075–7081 (2016)CrossRefGoogle Scholar
  137. G.U. Lee, D.A. Kidwell, R.J. Colton: Sensing discrete streptavidin–biotin interactions with atomic force microscopy, Langmuir 10(2), 354–357 (1994)CrossRefGoogle Scholar
  138. V. Moy, E. Florin, H. Gaub: Intermolecular forces and energies between ligands and receptors, Science 266(5183), 257–259 (1994)CrossRefGoogle Scholar
  139. M. Rief: Reversible unfolding of individual titin immunoglobulin domains by AFM, Science 276(5315), 1109–1112 (1997)CrossRefGoogle Scholar
  140. J. Yu, J. Warnke, Y.L. Lyubchenko: Nanoprobing of $${\upalpha}$$-synuclein misfolding and aggregation with atomic force microscopy, Nanomed. Nanotechnol. Biol. Med. 7(2), 146–152 (2011)CrossRefGoogle Scholar
  141. C. Gonnermann, C. Huang, S.F. Becker, D.R. Stamov, D. Wedlich, J. Kashef, C.M. Franz: Quantitating membrane bleb stiffness using AFM force spectroscopy and an optical sideview setup, Integr. Biol. 7(3), 356–363 (2015)CrossRefGoogle Scholar
  142. A. von Nahmen, M. Schenk, M. Sieber, M. Amrein: The structure of a model pulmonary surfactant as revealed by scanning force microscopy, Biophys. J. 72(1), 463–469 (1997)CrossRefGoogle Scholar
  143. G. Karagiorga, G. Nakos, E. Galiatsou, M.E. Lekka: Biochemical parameters of bronchoalveolar lavage fluid in fat embolism, Intensive Care Med. 32(1), 116–123 (2006)CrossRefGoogle Scholar
  144. L. Gunasekara, S. Schürch, W.M. Schoel, K. Nag, Z. Leonenko, M. Haufs, M. Amrein: Pulmonary surfactant function is abolished by an elevated proportion of cholesterol, Biochim. Biophys. Acta Mol. Cell Biol. Lipids 1737(1), 27–35 (2005)CrossRefGoogle Scholar
  145. O. Marti, B. Drake, P.K. Hansma: Atomic force microscopy of liquid-covered surfaces: Atomic resolution images, Appl. Phys. Lett. 51(7), 484–486 (1987)CrossRefGoogle Scholar
  146. B. Drake, C. Prater, A. Weisenhorn, S. Gould, T. Albrecht, C. Quate, D. Cannell, H. Hansma, P. Hansma: Imaging crystals, polymers, and processes in water with the atomic force microscope, Science 243(4898), 1586–1589 (1989)CrossRefGoogle Scholar
  147. R.C. Barrett: High-speed, large-scale imaging with the atomic force microscope, J. Vac. Sci. Technol. B 9(2), 302 (1991)CrossRefGoogle Scholar
  148. S. Karrasch, R. Hegerl, J.H. Hoh, W. Baumeister, A. Engel: Atomic force microscopy produces faithful high-resolution images of protein surfaces in an aqueous environment, Proc. Natl. Acad. Sci. 91(3), 836–838 (1994)CrossRefGoogle Scholar
  149. M. Benoit, D. Gabriel, G. Gerisch, H.E. Gaub: Discrete interactions in cell adhesion measured by single-molecule force spectroscopy, Nat. Cell Biol. 2(6), 313–317 (2000)CrossRefGoogle Scholar
  150. P. Parot, Y.F. Dufrêne, P. Hinterdorfer, C. Le Grimellec, D. Navajas, J.-L. Pellequer, S. Scheuring: Past, present and future of atomic force microscopy in life sciences and medicine, J. Mol. Recognit. 20(6), 418–431 (2007)CrossRefGoogle Scholar

Copyright information

© Springer Nature Switzerland AG 2019

Authors and Affiliations

  1. 1.Dept. of Cell Biology & AnatomyUniversity of CalgaryCalgaryCanada
  2. 2.Bruker Nano GmbHBerlinGermany

Personalised recommendations