Introduction

Scanning Probe Microscopy (SPM) is a branch of microscopy which produces images of surfaces using a physical probe that scans the object. An image of the surface is obtained by mechanically moving the probe in a raster scan of the object, line by line, and recording the probe-surface interaction as a function of position. One of the types of SPM is the atomic force microscopy (AFM), which invention, in 1981, earned its inventors, Gerd Binnig and Heinrich Rohrer, the Nobel Prize in Physics (Binnig et al. 1986). The essential property of AFM is the interaction force between a sharp tip (<10 nm) and the sample surface at a very short distance (0.2–10 nm probe-sample separation). The tip is attached to the free end of a flexible V-shaped or rectangular cantilever and is brought very close to a surface. It is surprising that a sharp tip adapted to a cantilever with a spring constant weaker than the equivalent between atoms can image both conductive and non-conductive samples (Rugar and Hansma 1990). AFM enables manipulations of nano-sized objects relying on high resolution imaging from the nano to mesoscale structures able to resolve slight differences in surface height and to image transparent structures. In contrast with Scanning Electron Microscopy (SEM), it has a great advantage that samples do not need to be coated, thus lowering the cost of the operation. It can also be operated in different environments (air, liquid, vacuum) and it is applied to conductive and non-conductive materials.

Since the invention of AFM (Binnig et al. 1986), great progress has been made regarding imaging biological samples. It is a useful tool, for example, for the analysis of biological membranes without damaging them. It is also used in other fields of biosciences (Engel and Gaub 2008), for instance, imaging of DNA and soluble proteins. From bacteria to chloroplasts and to animal cells, a variety of biological samples can be imaged by AFM. It enabled advances in live-cell characterization (Weyn et al. 1998; Higgins et al. 2000; Crawford et al. 2001; Arce et al. 2004). So far, AFM is unique in its capacity of imaging biological samples in their natural environment at sub-nanometer resolution (e.g. Drake et al. 1989; Engel and Gaub 2008).

Diatoms, unicellular microalgae occurring in aquatic (fresh or salt water) or terrestrial habitat, represent a good model of naturally evolved micro to nano-sized biological structures. Their siliceous cell wall, called a frustule, is the most characteristic feature of these microorganisms and is generated by silica biomineralization. The process of diatom biosilica formation is thought to be controlled by silica deposition and nanopatterning (Sumper and Kröger 2004). Biosilica is mainly constituted by amorphous hydrated silica (SiO2·2H2O), containing a small proportion of organic macromolecules. The diatom frustule consists of two valves (epivalve and hypovalve) resembling a petri dish, joined together by a series of silica bands (girdle bands or copulae) connected along the margins (Round et al. 1990) and involving the organic parts of the cell. The epivalve always overlaps the hypovalve. The uppermost part of each valve is called valve face while the downturned side of the valve surrounding the valve face, which is visible only in girdle view, is called mantle. The first girdle band directly associated with the valve is called valvocopula. The full series of all linking siliceous girdle bands form a structure called cingulum (epicingulum or hypocingulum). All these silicious diatom units are illustrated on Fig. 1. The first study using AFM microscopy on cleaned diatoms from mud samples imaged in air was done by Linder et al. (1992), where it was shown that AFM could compete with SEM, at medium magnification, imaging the same features, but with a low cost operation because there is no need to coat the sample. However, the first AFM studies on the characterization of diatom silica nanostructure, were done by Almqvist et al. (2001) on dried (in ethanol) pennate diatom Navicula pelliculosa Kützing (Hilse) and by Crawford et al. (2001) on the living pennates Pinnularia viridis (Nitzsch) Ehrenberg and Hantzschia amphioxys (Ehrenberg) Grunow, proving it was possible to image diatoms alive, in natural conditions. The number of publications applying AFM to diatom studies has increased becoming a separate research discipline. There are many diatoms studies, not only about the silica nanostructured morphology (e.g., Crawford et al. 2001; Gebeshuber et al. 2003, 2005; Losic et al. 2006a, b, 2007a; Tesson and Hildebrand 2010; Scheffel et al. 2011), but also about properties of mucilage layers (Higgins et al. 2002, 2003a; Arce et al. 2004; Pletikapić et al. 2011) and, partially understudied, physical and mechanical properties (e.g., Almqvist et al. 2001; Francius et al. 2008).

Fig. 1
figure 1

SEM microphotographs showing the organization of the silicious diatom units

Full size image

The most complete review with respect to AFM techniques used in the study of diatoms is from Svetličić et al. (2011). This study only focused on the AFM imaging of extracellular polysaccharides and the quantification of physical forces, based on the detection of the supramolecular structures of polysaccharide fibrils produced by marine diatoms, either attached to the diatom cell wall or released into the surrounding seawater. However, with our review, our intention is to go deeper in the knowledge of diatoms, exploring the basic structures that construct the frustule, understanding the mechanical properties of such structure, the composition of the mucilage layer that encases the frustule, and complementing the study of Svetličić et al. (2011), regarding his review about diatom extracellular polysaccharides. From our knowledge, it is the first general review about the use of AFM microscopy to study diatom cells as a whole.

Thus, the main objectives of this review are to summarize elementary scientific literature and papers on all fields related to diatom imaging in AFM and present particular methods, principles, and outcomes of AFM application in diatom studies. The review focuses on the following subjects: (1) AFM method and its modes, (2) methods of sample preparation, (3) advantages and limitations of using AFM, (4) application of AFM to the study of biomineral formation and ultrastructure in diatoms, (5) mucilage layer characteristics and functions, (6) micromechanical properties of the frustule, (7) AFM illustrations for taxonomy, and (8) future perspectives and conclusions.

AFM method and its modes

In all force microscopy types, five components are required: (1) a sharp tip mounted on a soft cantilever spring, (2) a way of sensing the cantilever’s deflection, (3) a feedback system to monitor and control the deflection, (4) a mechanical scanning system (usually piezoelectric) that moves the sample with respect to the tip in a raster pattern, and (5) a display system that converts the measured data into an image (Rugar and Hansma 1990).

Atomic force microscopy (AFM) has three operation modes: (a) non-contact mode, (b) contact mode, (c) and tapping mode.

  1. (a)

    Non-contact mode. The probe in non-contact mode operates in the attractive force region and the tip-sample interaction is minimized. Non-contact mode uses a stiff cantilever, oscillating in the attractive regime, allowing the tip to be quite close to the sample, but not touching it. The detection scheme is based on measuring changes to the resonant frequency or amplitude of the cantilever. The tips mainly used for this mode are silicon probes. The advantages of this mode type are very low force exerted on the sample (10−12 N), extended probe lifetime. On the contrary, the disadvantages are the following: lower lateral resolution and slower scan speed to avoid contact with fluid layer, usually only applicable in extremely hydrophobic samples with a minimal fluid layer usually need ultra-high vacuum (UHV) to have best imaging. No works on diatoms were performed with this mode.

  2. (b)

    Contact mode is the first and foremost operation mode widely used for diatom studies. The tip either scans at a constant small height above the surface or under the conditions of a constant force touching the surface. The cantilever must be soft enough to be deflected by very small forces and must have a high resonant frequency not to be susceptible to vibrational instabilities. Silicon nitride tips are often used in this mode. Contact mode allows fast scanning and it is good for rough samples used in friction analysis. It is preferred by some authors (Crawford et al. 2001; Francius et al. 2008; Pletikapić et al. 2012) because contact mode is able to remove soft upper layers revealing the underlying harder silica structures of diatoms. However, extensive smearing effects have been regularly observed when imaging soft biological samples, such as diatom trails/biofilms, living animal cells and polymer layers (Weyn et al. 1998; Higgins et al. 2000; Crawford et al. 2001), and lateral forces can distort the image and forces and damage/deform soft samples.

  3. (c)

    Tapping mode or intermittent contact mode or more generally dynamic force mode (DFM). In this mode, the cantilever oscillates close to its resonance frequency. An electronic feedback loop ensures that the oscillation amplitude remains constant, so that a constant tip-sample interaction is maintained during scanning. Forces that act between the sample and the tip do not only cause a change in the oscillation amplitude, but also a change in the resonant frequency and phase of the cantilever. The amplitude is used for the feedback, and the vertical adjustments of the piezoscanner are recorded as a height image. Simultaneously, the phase changes are presented in the phase image (topography). Silicon probes are used primarily for tapping mode applications. A stiff cantilever is oscillated closer to the sample than in non-contact-mode. In this mode, large vibrations of the probe are applied for imaging.

The following are advantages of the tapping mode:

  • Measurements can be made at ambient conditions.

  • When the oscillating probe hits the sample, its short-time interactions with minimal shear, are less destructive even with stiffer probes than the tip-sample forces in contact mode.

  • Can be used on both dry and wet sample surfaces.

  • Different components of the sample which exhibit difference adhesive and mechanical properties will show a phase contrast and therefore even allowing a compositional analysis.

  • Allows high resolution of samples that are easily damaged and/or loosely held to a surface.

  • Higher lateral resolution (1 nm to 5 nm).

  • Lower forces and less damage to soft samples in air, almost no lateral forces.

The following are the disadvantages of the tapping mode:

  • It images in a slower scan speed than in contact mode thus it is more challenging to image in liquids because slower scan speeds are needed.

  • The tip touches periodically the surface; therefore, manipulation of the sample as well as contamination of the tip is possible.

It seems that tapping mode is the most appropriate technique for diatom studies (e.g., Heredia et al. 2008; Scheffel et al. 2011), because, for example, it enables analysis of the topography of the soft mucilage layers (Pletikapić et al. 2011; Svetličić et al. 2011, 2013), allows the adhesive trails of the gliding diatoms to be imaged in the natural hydrated state (e.g. Higgins et al. 2000; Gebeshuber et al. 2002), and, for instance, contact mode often damages soft surfaces (soft biological samples, polymers, etc.). This type was also elected by several authors (Almqvist et al. 2001; Higgins et al. 2002; Gebeshuber et al. 2003; Hildebrand et al. 2008; Pletikapić et al. 2011; Svetličić et al. 2011).

However, a combination of contact and tapping modes was used by the previous authors with the exception of Heredia et al. (2008) and Scheffel et al. (2011). The reasons for this election will be further discussed in detail, in the following sections.

Methods of sample preparation

There are several factors to bear in mind for a good sample preparation to AFM imaging, always depending on the properties/structure of diatoms that significantly influence the results analysis (e.g., image quality, prolong time of analysis).

For AFM analysis, either live material or treated-diatom samples can be used. To treat samples prior to AFM operations in order to eliminate organic content of diatom cells using acids, e.g., sulphuric acid (Losic et al. 2006a, b, 2007a, b, 2008; Kröger et al. 2002; Hildebrand et al. 2008) can be used; however, as some micro- and nanostructures and less silicified frustules may dissolve in strong acids, then rinsing in hydrogen peroxide can be used instead (De Stefano et al. 2011). A treated-diatom suspension is transferred then onto suitable substrate (e.g., glass slides, cleaved muscovite mica, or silica wafer), which is the most important factor to be taken into account because it enables attachment of the sample to a holder. A good attachment is needed to allow stable scanning and for not contaminating the AFM tip. While Linder et al. (1992), for example, fixed dead diatoms directly to their respective sample holders when studying silica structures, whereas cleaved muscovite mica was used by Pletikapić et al. (2011, 2012) and Svetličić et al. (2011) to study the extracellular polysaccharides production. Almqvist et al. (2001) used diatoms grown on mica surface; diatoms were transferred in a drop that was spread on the mica surface to study micromechanical properties. For immobilization of diatom cells, polylysine was the mounting medium most frequently used to both treated (Higgins et al. 2002, 2003a, b; Losic et al. 2006a, b, 2007a, b, 2008; Tesson and Hildebrand 2010) or live diatoms (Pickett-Heaps et al. 1990; Higgins et al. 2002). Polylysine-coated slides (e.g., Tesson and Hildebrand 2010) or polylysine-coated silicon wafer (Losic 2006a, b, 2007a, b, 2008; Scheffel et al. 2011) can be used for diatom attachment in silica structure studies. For example, Gebeshuber et al. (2003) used a “naturally selected” approach to study adhesion strength involving freshwater snails feeding on algae, from a sample of numerous benthic freshwater diatom species growing on glass slides. With this approach, three species that obviously produce outstanding natural adhesives were obtained, e.g., Eunotia sudetica O.Müller, Navicula seminulum Grunow and an unidentified species. AFM was also used to compare adhesion forces between extracellular polymers associated with surfaces of live cells of marine fouling diatoms of the Navicula genus and two surfaces of different physicochemical properties, namely mica and Intersleek, in a simulated marine environment (Arce et al. 2004). For investigations of the hydrated organic layers in live diatoms, samples can be mounted onto glass coverslips coated with polyethylenimine (Crawford et al. 2001). Authors such as Pletikapić et al. (2011) undertook all experiments in fluid with diatom media or K-medium and a saline solution of 10 g L−1 NaCl was used by Francius et al. (2008), both to image marine live diatoms.

Related to working conditions, samples can be imaged in vacuum or in air both for polysaccharides (Arce et al. 2004) or silica structure (Almqvist et al. 2001; Heredia et al. 2008) studies. For silica structures, samples can have an additional treatment as lyophilization and then be embedded in a low viscosity epoxy resin (e.g., Embed-812) (Heredia et al. 2008). However, for polysaccharides imaging and force measurements, the information is contradictory; imaging in air (Arce et al. 2004) avoids an inconvenient motion of polysaccharides; however, the drop deposition method (Balnois and Wilkinson 2002) by adsorption is preferred (Pletikapić et al. 2011, 2012) because the contact and rinse time can be controlled and optimized and few aggregates were observed. Thus, for systems involving polysaccharides on living cells, liquids mimic natural cellular conditions.

Experiments and manipulations were always performed at room temperature.

Advantages and limitations of using AFM

Effortless sample preparation, able to image conductive and non-conductive materials and transparent structures, low cost operation related to SEM and no need of coatings, and the possibility of imaging biological samples in their natural environment at a subnanometer resolution are the major advantages of AFM. In addition, AFM has the ability to determine three-dimensional surface topology of the hard material, since it is especially able to resolve slight differences in surface height, which is useful in imaging nanoscale features resulting from silica polymerization in diatoms (Hildebrand et al. 2008). It also enables high resolution imaging from the nano to mesoscale (Hildebrand et al. 2009). The girdle bands which due to their small size and thinness are relatively transparent to electrons are resolved by AFM demonstrating the advantages of AFM to image small differences in surface morphology, providing new insights, and confirming evidence for models of diatom silica structure formation (Almqvist et al. 2001; Hildebrand et al. 2009). High image resolution, for example, AFM, is able to elucidate some details of the surface topography of the cells (e.g. setae). In particular, the ability of AFM to operate in different environments, as for example, to operate in fluids (i.e., diatom growth media), enabled to interact directly with the adhesive strands and determine their local nanoscale properties and do not damage biological samples. Taxonomical studies can be improved since small structures like pores can clog with the coatings used in SEM. AFM enables chemical mapping through the use of derivatized tips and allows investigation of the specific interaction forces between the diatom adhesive molecules (single strands in living cells and/or purified molecules) and cantilever tips that have been chemically modified (i.e., antifouling probes). AFM is able to determine micromechanical properties and the possibility of mapping mineral or organic domains at nanometer resolution (Hildebrand et al. 2008). Interestingly, AFM has also allowed the direct visualization of hydrated extracellular polymeric substances (EPS) adsorbed on solid surfaces to a resolution that was not previously accessible by any other microscopy technique (Higgins et al. 2000). Furthermore, AFM allows working directly on living cells in order to explore and manipulate cell surface properties at nanometer resolution and to analyze cell wall proteins at the single molecule level (Francois et al. 2013). Currently, AFM is the only technique that can image live microbial cells in real time and in a nanometer resolution (Francius et al. 2008).

Only few limitations of this technique have been determined so far. For example, it can be difficult to image large changes in height due to the limits posed by the movement of the Z piezo scanning element, that in most AFM’s is limited to about 10 μm. Furthermore, imaging larger diatoms has limitation in the XY scan range of about 100 μm. Another limitation of AFM is that the microscopes are operating through piezoelectric devices, which show a limited operation range in the spatial directions. This characteristic may force one to change the head of the AFM from sample to sample. Moreover, those piezoelectrics show a nonlinear behavior which has to be calibrated and corrected.

Application of AFM to the study of biomineral formation and ultrastructure in diatoms

Silica is the second most abundant mineral biogenically produced, slightly under calcium carbonate (Lowenstam and Epstein 1957). Biosilica formation is dominant in aquatic organisms, among which, diatoms are the most abundant and widespread. Biosilica features are formed in silica deposition vesicles (SDVs), while the precursor for biosilica formation inside SDVs is unknown. Most likely, it could be the monosilicic acid. Biosilica determines nano to mesoscale structures. Nanoscale and mesoscale morphology are independent of microscale structures and size. Different nano and mesoscale structures can be found within the same organism, and different mesoscale arrangements of nanoscale silica can be observed, which include random and ordered structures with the latter most commonly observed in a linear arrangement (Hildebrand et al. 2009). At the nanoscale, the structures vary greatly even on the same side of the same feature. Diatom valve formation can progress in different stages. For example, in pennate diatoms, different features are formed in certain order (e.g., first the raphe is formed and then the ribs) (Pickett-Heaps et al. 1990). However, there is small variation in the size/range of the silica particles within a particular frustule component, valve, or girdle band (Crawford et al. 2001). Thus, diatoms show different and well-organized patterns from nano to microscale morphologies. Different cell wall silica structures, including valves, girdle bands, and setae (extension of the diatom valve which are more elongate than spines; they join frustules together to form chains) (Fig. 2), can be studied (e.g., Hildebrand et al. 2008) to characterize surface topology, showing particle size differences between valve or girdle bands within the frustule and most significant differences between species. Crawford et al. (2001) was one of the first authors who analyzed morphological differences in frustules of different species via imaging diatom silica structures using AFM. In his study, girdle bands and valves of Pinnularia viridis and Hantzschia amphioxys were shown in cross-section, revealing silica composed of a conglomerate of different colloidal spheres with granular texture. The same arrangement was found on Coscinodiscus sp. by Losic et al. (2006b). Furthermore, AFM revealed the nanoscale of girdle band features, which consisted of particulate silica even in the relatively featureless portion and ribs that were composed of organized agglomerates of particulate silica (Tesson and Hildebrand 2010). Cross sectioning and AFM revealed that all structures in Cyclotella cryptica Reimann, J.C.Lewin & Guillard are formed from nanoparticulate silica, and these are assembled into larger particles which are organized on an even higher level to form mesoscale structure. A distinguished nanostructured silica granular surface of diatoms cell wall was also identified by Noll et al. (2002).

Fig. 2
figure 2

SEM microphotograph of Bacteriastrum sp. (by Carlos E. Wetzel) showing its elongated setae

Full size image

Frustule and central spine could be resolved at small levels of organization with great detail at both micro- and nanoscale in a study of Heredia et al. (2008) with Ditylum brightwellii (T.West) Grunow in Van Heurck. It was shown by AFM that there is a fundamental difference between these two structures consisting in a solid smooth surface in the case of the cell walls and a surface with nanospheres in the case of the central spine as well as in the central hollow. Other morphological differences included the presence of pores between 0.5–1.0 μm in diameter in the cell wall, contrasting to the case of the central spine (Heredia et al. 2008). AFM allowed us to describe the growth of fultoportulae. After its initial formation in the x/y plane, it grew in the z-axis direction distally, and rib silica deposition has moved past the location of the portula (Tesson and Hildebrand 2010). Morphological differences in frustules can incorporate number of layers, density, shape, size, and geometric configuration of pores. Three porous layers are identified in the frustule of Coscinodiscus sp. The outer porous layer, known as cribellum, whose surface consists of about 2 μm diameter porous domes hexagonally packed across the frustules. Cribellum is followed by a second porous layer, called cribrum, a thin covering of silica having fine pores, which is connected to the internal (third) porous silica plate containing the so called foramen holes. (Losic et al. 2006a). Since AFM-closed loop imaging is able to obtain very accurate sample topography, measurements of pores could be done as well. In Cocconeis placentula Ehrenberg, the depth of AFM tip penetration into the pores ranged between 2 and 11 nm, while in the case of wider pores of Cocconeis placentula var. euglypta (Ehrenberg) Grunow, the tip penetration ranged between 7 and 42 nm (Strzelecki et al. 2012). Losic et al. (2006b) analyzed detailed nanotopography of porous surface of Thalassiosira eccentrica (Ehrenberg) Cleve emend Fryxell & Hasle and Coscinodiscus sp. The external frustule structure of Thalassiosira eccentrica is very analogous to the internal frustule structure of Coscinodiscus sp. in terms of the porosity and pore patterns. Both species showed a consensus lower size limit for pores in the frustule (about 40 nm) (Losic et al. 2006b). Thalassiosira eccentrica produces a frustule with one silica structural layer featuring two different porous surfaces. Two layers (not three as Coscinodiscus sp.) with the outer surface with large (800 nm diameter) holes (foramen) while the inner surface contains a porous wall with pores comparable in size to the Coscinodiscus sp. cribellum. The inner and outer surfaces of the frustule wall of both diatoms are hence in reverse order. In contrast to previous SEM studies, high resolution AFM images clearly reinforce that the diatom frustule is built from silica nanoparticles. The size of these silica structures (nodules) varied (from 20 to 70 nm) on different layers. Both species show smaller silica nodules and lower surface roughness on the inner layer than on the outer layer (Losic et al. 2007a).

Height information from AFM was helpful in imaging nanoscale features of spherical silica particles on the ridges of the outer Thalassiosira pseudonana Hasle & Heimdal valve surface, the textured surface of Ditylum brightwellii valve ribs, and the different silica morphologies in the proximal and distal faces of girdle bands (Hildebrand et al. 2009).

AFM imaging of setae allowed seeing these elongated substructures in a detail not previously registered, as for example the regularly repeating structure in Chaetoceros gracilis F.Schütt setae (Hildebrand et al. 2008). Also, the ribs in the base layer of the Thalassiosira species were clearly identified in AFM by Hildebrand et al. (2009).

A fustule is not composed of a unique block unit of silica despite the different morphologies common in the same species. However, there are identical construction features in valve-linear and girdle bands (Hildebrand et al. 2008). Structural intermediates of the mesoscale can be enriched through the application of synchronized culture approaches (Hildebrand et al. 2007), allowing more detailed examination of the stages and processes involved in structure formation (Hildebrand et al. 2006).

Biosilica morphology of the porous structures that transverse the valve wall of Dytilum brightwellii showed macropores of about 100 nm in diameter (Heredia et al. 2008). The central spine, in topographic images, showed a more complex layered structure, while the valve showed simpler double-layered silica structure with thickness of 100 nm each. The same pore size was found by Gebeshuber et al. (2003) and Francius et al. (2008). In addition, the central spine was studied in more detail. The elementary structures of the central spine were its cavity, the nanospheres, and the array in silica layers. There is a fundamental difference between the central spine and the frustule (valves): the contrast of the solid and smooth surface of the frustule with the nanospheres and the hole of the central spine. No pores are present in the central spine, so no connections in between the central hollow of the central spine and the cytoplasm of the cell. The opposite was seen in the frustule surface with pores of about 0.5–1.0 μm in diameter. The spheres of the central spine are of 20 nm in diameter (Chiappino and Volcani 1977; Almqvist et al. 2001; Hildebrand et al. 2008).

AFM was also successfully applied in studying the surface features of silica structure, which seemed quite smooth, but the high resolution imaging revealed some small-scale granularities, and other larger-scale features were superimposed on the basic structure (Crawford et al. 2001). Silica on the valve surface of the large centric diatom Coscinodiscus granii L.F.Gough was examined by AFM and, in contrast to Pinnularia viridis and Hantzschia amphioxys, did not have a mild texture, but granular, consisting of amalgamated particles of silica with 100–200 nm in diameter (Hildebrand et al. 2008).

Specifically in thin structures, AFM also allows interpretation of how the underlying organics involved in structure formation are organized (Hildebrand et al. 2009). Pletikapić et al. (2012) proposed a cell wall model based on AFM, presenting individual silica nanoparticles incorporated in an organic matrix. The 15 nm sized silica spheres in the valve region connecting raphe with the girdle bands were identified after acid treatment. The silica spheres were neither fused together nor forming a nanopattern. The organization of girdle band and valve regions allows the high flexibility needed for movement and adjustment to distinctive environments while maintaining the integrity of the cell (Pletikapić et al. 2012).

Formation of opalescent films by self-assembly of silica nanoparticles produced in the growth medium of marine algae (diatoms) was demonstrated for the first time in 2008 by Losic et al. AFM images of opalescent films show ordered arrays of silica nanoparticles with different diameters depending on the colors observed by light microscopy. The film forming silica nanoparticles were either released by the diatoms during reproduction or after cell death. The opalescent film was formed in Coscinodiscus sp. culture medium, but not in culture of Thalassiosira eccentrica. Alternatively or in addition, silica nanoparticles could be formed through an action of biomolecular catalysts such as silaffins or long chain polyamines released into the growth media (Losic et al. 2008).

Manipulation of diatom frustules using external stimuli was demonstrated by Clarson et al. (2009). Structural features of the biosilica of Aulacoseira granulata (Ehrenberg) Simonsen cultured under quiescent environments were studied in AFM, which showed that the pores were typically ~530 nm wide, while the depth of silica “mesh” was ~120 nm. This analysis revealed the hierarchical nature of the biosilica structure.

Tesson and Hildebrand (2013) made the connection, for the first time, between structure, composition, and localization of insoluble organic matrices associated with the silica formation. The calcofluor-stained material from Amphora salina W.Smith studied by AFM indicated that it consists of two halves connected by fibers. It was discovered as a novel glucose polymer with a role during silica structure formation.

Mucilage layer characteristics and functions

Understanding how the organic envelope (essentially composed of polysaccharides and proteins), formed after the silica structure formation, relates to the silica structure, is important. For example, previous studies have sought to determine the chemical composition of mucilaginous coatings from diatom species after extracting mucilaginous layer from the cell and subsequent analysis (Lewin 1955; Lewin et al. 1958; Tokuda 1969). However, Crawford et al. (2001) and other previous authors (Higgins et al. unpublished data) were the first to investigate the hydrated organic coatings of living diatom cells. AFM in contact mode showed that the mucilage layer could be removed aside to reveal the hard underlying surface of the frustule cell wall (Crawford et al. 2001). Thus, the examination of Pinnularia viridis by Crawford et al. (2001) made possible, for the first time, the study of the hydrated organic layer covering the silica in living diatom cells and also the silica itself in natural hydrated conditions. One particularity revealed by contact mode was an outer mucilage layer of thick mucilaginous material interrupted only in the proximity of the raphe fissure (narrowly open and unclogged by mucilage) covering the cell walls (Crawford et al. 2001). Also the adhesive mucilages secreted by Craspedostauros australis E.J.Cox showed cells encased in a cohesive, amorphous outer mucilage layer, continuous over the entire frustule except at the raphe openings. A second adhesive mucilage consisted of strands secreted at the raphe, a distinct slit in the silica wall, involved in cell-substratum attachment and motility (Higgins et al. 2002).

Polysaccharide characterization was also possible, combining carbohydrate analysis with AFM imaging to study the pennate diatom, Pinnularia viridis. Chemical and AFM data for the WW (warm water extract) and HB (hot bicarbonate/EDTA extract) fractions indicated that compositional differences were associated with important changes in the morphology and properties of the cell surface mucilage. Soluble polymers relatively enriched in fucose granted a degree of softness and compressibility to the mucilage, whereas most of the mucilage comprised firmer more gelatinous polymers comparatively enriched in rhamnose (Chiovitti et al. 2003). Using a modification of the fractionation procedure of Ford and Percival (1965), the polysaccharides extracted with hot alkali that dissolved the silica, were characterized by constituent sugar and linkage analyses. In conclusion, the value of combining sequential extraction and chemical analysis with AFM of the diatom cell surface was demonstrated. Enrichment of certain carbohydrates in different fractions was related to observations of changes to the surface covering, providing information of the structure and the in situ architecture and properties of the extracted macromolecules. However, mannose was the dominant sugar in the polysaccharides from all four species analyzed (Pinnularia viridis, Craspedostauros australis, Thalassiosira pseudonana, and Nitzschia navis-varingica Lundholm & Moestrup) (Chiovitti et al. 2005). There is also a high capacity of AFM to investigate the diatom extracellular polysaccharides at a subnanometer resolution. For example, from a ubiquitous marine diatom Cylindrotheca closterium (Ehrenberg) Reimann & J.C.Lewin, isolated from the northern Adriatic Sea, with its EPS visualized at a single cell level (Pletikapić et al. 2011). A simple procedure for AFM imaging of diatom cells on mica under ambient conditions (in air) to achieve visualization of their EPS with molecular resolution was applied. The EPS represents a web of polysaccharide fibrils with two types of cross-linking: fibrils association forming junction zones and fibril–globule interconnections with globules connecting two or more fibrils. The fibril heights were 0.4–2.6 nm while globules height was in the range of 3–12 nm (Pletikapić et al. 2011).

Visualizing a biofilm as a gel network (Pletikapić et al. 2011), seeing the arrangement of the fibrillar network, could be obtained with high resolution only when imaged in air. AFM images of the marine gel network and the network formed by polysaccharides extracted from the C. closterium culture medium share the same features regarding the fibril heights, pore openings, and the mode of fibril association, indicating that the gel phase in the Northern Adriatic macroaggregates can be formed directly by a self-assembly of diatom released biopolymers (Pletikapić et al. 2011).

AFM also offers possibility of characterizing the stickiness and elasticity of high molecular weight AOM (algal derived organic material) (e.g., biopolymers). The stickiness of AOM in terms of its adhesive and cohesive strength varies between algal sources. Cohesion between AOM was found generally stronger than adhesion of AOM on AOM-free surfaces (Villacorte et al. 2015). Overall, the composition as well as the physico-chemical characteristics of AOM will likely dictate the severity of fouling in membrane systems during algal blooms.

Another study of Higgins et al. (2003a), with the marine species Craspedostauros australis (mentioned above), Nitzschia navis-varingica and the freshwater species Pinnularia viridis, was done using AFM to investigate the topography and material properties of the mucilage layer of live cells. Contrary to previous studies, this study showed that surface mucilage layers display distinct nanostructural features. Tapping mode images of C. australis revealed a continuous and soft mucilage layer encasing the silica cell wall, consisting of a smooth flat surface, interrupted by regions with groove-like indentations and force measurements showed no adhesion. On the other hand, P. viridis presented a mucilage layer with a densely sphere-like aspect. Nitzschia navis-varingica showed an outer surface with absence of a mucilage layer in the girdle region, in contrast to the other two species. It was found that the nanostructure and adhesive properties of the mucilage layers differ between live cells of C. australis and P. viridis. Surprisingly, N. navis-varingica appears to lack an outer mucilage layer, which raises fundamental questions about the overall function of this layer (Higgins et al. 2003a).

Hydrated trails left behind by gliding Pinnularia viridis demonstrated that AFM in fluid tapping mode was the most accurate mode to image the topology of soft hydrated mucilage, due to the minimal physical interference caused by the action of the cantilever tip (Higgins et al. 2000). Force mode AFM demonstrated the non-adhesive, compressible, and soft nature of this mucilage layer (Higgins et al. 2002).

Atomic force microscopy (AFM) also allows visualizing polymer networks produced by a marine diatom in culture, Cylindrotheca closterium, isolated from the northern Adriatic Sea (Svetličić et al. 2011). Applying high resolution imaging by AFM demonstrated the capacity of the diatom in producing polysaccharides to form polymer gel networks by self-assembly. Isolated polysaccharide fractions from C. closterium culture can form gel network by self-assembly that resembles the gel network of macroscopic gel phase occurring in the northern Adriatic Sea. Macromolecular properties of the isolated polysaccharide fraction of marine gel confirmed marked polyelectrolytic behavior allowing a major solubility of polysaccharides in salt solutions favoring chain aggregation and gel formation (Svetličić et al. 2011, 2013).

A novel strategy of colony formation was detected using AFM. It was possible to visualize and characterize the cell jacket structure of Bacteriastrum jadranum (Bosak et al. 2012). This species forms loose but regular chains with distinct heterovalvate terminal cells. The colonial cells and their siliceous projections, the setae, were not in direct contact; instead, they were enclosed within the optically transparent organic matrix. At nanoscale resolution, the cell jacket appears as a cross-linked fibrillar network organized into a recognizable structure. Using the high resolution imaging technique of AFM, it was possible to see the presence of an EPS, which was the most obvious and probable matrix constituent. The discovered polysaccharide fibrillar network was highly organized and delicately structured with a monomolecular fibril height of 0.6 nm. The Bacteriastrum polysaccharide jacket represents a fundamental part of the cell, as the conjunction of the polymer network with the frustule appearing very tight, and such specific and unequaled patterns have never been found in self-built polysaccharide gel networks, usually found in the marine environment (Bosak et al. 2012).

There are several AFM studies that have explored the nature of mucilage layers of biofilms as adhesives (Higgins et al. 2002; Stal and de Brouwer 2003; Arce et al. 2004; Dugdale et al. 2005), most man-made adhesives fail in wet conditions, owing to chemical modification of the adhesive or its substrate. Diatoms produce adhesives that are extremely strong and robust both in fresh- and in seawater environments.

Craspedostauros australis and Pinnularia viridis were also investigated to understand how the adhesive mucilage and mechanism of cell-substratum adhesion work. AFM force measurements revealed that the adhesive strands originating from the non-driving raphe of live C. australis and P. viridis are highly extensible and accumulate to form tethers (resultant of elongation/stretching of composite adhesive mucilage strands secreted from raphes during the onset of cell adhesion and reorientation). In most cases, AFM force measurements recorded an interaction with a number of adhesive strands that are secreted from the raphe. The force curves of C. australis and P. viridis displayed a sawtooth pattern, suggesting the successive unbinding of modular domains when the adhesive strands were placed under stress (Higgins et al. 2003b). In conclusion, the adhesive properties of the C. australis and P. viridis strands were similar and, in most cases, the mechanical response of their adhesives under stress gave rise to a sawtooth pattern in the force curves that has been commonly observed for natural adhesives (Smith et al. 1999). The performance of “fly-fishing mode” experiments (Rief et al. 1997) enabled the attachment of single adhesive strands useful to the cantilever for their characterization (Higgins et al. 2003b). This study provided valuable information about protist adhesives used during biofouling and may suggest possible mechanisms to prevent it (Higgins et al. 2003b).

Adhesion is an important property of mucilage layers when antifouling properties of different materials are studied. AFM observation of a living marine diatom Navicula sp. helped to evaluate the adhesive properties of EPS associated with individual cells on two materials dissimilar in physicochemical characteristics, such as mica and the silicone elastomer, Intersleek in a study of Arce et al. (2004). The work strength of detachment of a single diatom from these surfaces, determined from force versus distance curves, using either the same or different diatoms in different growth phases, strongly depended on the individual Navicula cell and not on their growth stage. The values and shapes of force versus distance curves supported the argument that the adhesion of Navicula sp. to surfaces with different physicochemical properties is governed by macromolecular specificity of diatom EPS. Thus, biological AFM can be successfully applied to carry out in situ characterization of cell adhesion to different surfaces. This is a promising method for the rapid assessment of diatom attachment to antifouling materials. Force versus distance curves generated during these experiments revealed comparable cell adhesion strengths on Intersleek and mica, indicating that Navicula diatoms secrete extracellular polymeric substances with hydrophobic and hydrophilic properties (Arce et al. 2004). Natural adhesives might uncover promising and valuable information to future engineering approaches. The first in vivo study of diatoms using atomic force microscopy (AFM) was performed on three chain-forming benthic freshwater species, Eunotia sudetica, Navicula seminulum and a yet unidentified species by Gebeshuber et al. (2002). These species were directly imaged while growing on glass slides. The thickness of the organic case enveloping the siliceous skeleton of the cell (of 10 nm) was determined with focus on the natural adhesives produced by these unicellular organisms to adhere to other cells or the substratum. The phase-imaging and force-pulling experiments revealed some specific characteristics of these natural adhesives that might be of use in designing man-made analogues that function in wet environments. Engineering stable-underwater adhesives currently poses a major technical challenge (Gebeshuber et al. 2003).

Micro- and nanotribology considered as a mechanical interaction of moving bodies is the science of friction, adhesion, lubrication and wear on the scale of micrometers to nanometers, and the force scale of millinewtons to nanonewtons (Gebeshuber et al. 2005). Some sort of lubrication of interactive surfaces is present in some diatom species because several diatom species actively move. For example, Bacillaria paxillifer (O.F.Müller) T.Marsson forms colonies, in which the single cells slide against each other, or the cell colonies of Ellerbeckia arenaria (Moore ex Ralfs) R.M.Crawford in Schmidt & Crawford, which can be elongated by as much as a major fraction of their original length colonies. Thus, it might also be of tribological interest, to grow diatoms due to their highly efficient self-lubrication while girdle bands telescope, as the cells grow. Also, an unknown benthic freshwater diatom species was investigated in vivo at nanoscale with an atomic force microscope and bead-like features on the edges of certain girdle bands were found, which might well act as lubricant either by means of ball bearings or as solid lubricant or following a lubrication strategy which is still completely unknown to engineers (Gebeshuber et al. 2005).

Micromechanical properties of the frustule

Mechanical properties of the frustule have been studied to a lesser extent, due to the challenges associated with force application and detection at a relevant scale. Improving knowledge on these properties is essential for better understanding of attributes of the frustule. The most frequent micromechanical properties analyzed by AFM are elasticity, hardness, and adhesion, mostly in the frustule and in the mucilage layer but in a lower extent.

From an evolutionary perspective, structural and mechanical properties of the frustule must reflect trade-offs between defense and growth. Other siliceous structures have also evolved in some chain-forming species as means of holding neighboring cells together. Characterization and quantification of mechanical properties of these structures are crucial for the understanding of the relationship between form and function in diatoms. Only a handful of studies have addressed this issue, like AFM micro-indentation experiments conducted in the marine, chain-forming species Lithodesmium undulatum Ehrenberg (Karp-Boss et al. 2014). In this study, elastic moduli of cells and their linking structures were studied and examined to identify whether the cell division results in variations of local elasticity (Karp-Boss et al. 2014).

The first AFM experiments measuring nanoindentation and deflection in Navicula pelliculosa (Almqvist et al. 2001), found that hardness and elasticity were similar to other silicas; however, high resolution mapping revealed important differences in these two properties in different locations of the silica in the valve. For example, Losic et al. (2007b) mapped the mechanical properties of Coscinodiscus sp. using AFM-based nanoindentation and found that the lowest values of hardness and elastic modulus on the cribellum and on the girdle band and the observed variation in the mechanical properties could be attributed to differences in pore size, pore distances, and porosity, but also to a different biomineralization process (Losic et al. 2007b). On the other hand, Hildebrand et al. (2008) showed that hardness and elasticity generally decreased towards the edges of the valve.

Using a novel atomic force microscopy imaging technique with Peak Force Tapping Mode, Pletikapić et al. (2012) characterized elasticity and deformation of a weakly silicified marine diatom Cylindrotheca closterium. Elasticity and deformation were measured over the entire cell surface in seawater at a resolution that was not previously achieved. The fibulae were the stiffest (200 MPa) and the least deformable (only 1 nm). Girdle band region appeared as a series of parallel stripes. The valve region was complex with average values of Young’s modulus (29.8 MPa) and deformation (10.2 nm) with high standard deviations (Pletikapić et al. 2012). An AFM-calibrated glass needle determined elasticity of silica structures. Elasticity tests showed that ringent girdle bands were more flexible than complete ones (Coscinodiscus sp.) and that both long chain clusters of Nitzschia palea (Kützing) W.Smith and central spines of Ditylum brightwellii have certain elasticities (Wang et al. 2012).

A rigorous nanoindentation study of the species Coscinodiscus sp., monitored hardness and elasticity in four distinct structural features of the frustule. Hardness varied from 0.059–0.53 GPa and elastic modulus from 1.7–6.51 GPa, depending on the feature and its location. The lowest values were found on the cribellum and the girdle band (Loisic et al. 2007b).

Using in situ AFM, Francius et al. (2008) found that fusiform, triradiate, and ovoid morphotypes of Phaeodactylum tricornutum Bohlin showed major differences in surface ultrastructure and mechanical properties. While fusiform and triradiate forms show organic, smooth, non-structured surfaces, the oval type may have a silica valve typical of diatoms (Borowitzka and Volcani 1978) in one valve whereas the other valve is organic. The ovoid silica surface is coated with extracellular polymers, which were possibly involved in adhesion and gliding motility. The cell wall of the silicified ovoid form is about five times stiffer (Young modulus of ~500 kPa) than that of the two non-silicified forms (~100 kPa). The cell wall elasticity shows major variations depending on the location: the girdle region of both fusiform and ovoid forms is five times softer than the valve, the arms of the triradiate form are softer than the core region, and the mucilaginous footprints of the ovoid form show only moderate stiffness, compared with the valve (Francius et al. 2008).

Harper and Harper (1967) were one of the first to use the fundamental principles of AFM (i.e., deflection of a lever) to measure adhesion strength of several raphid diatoms. By pushing a glass fiber down onto the diatoms, presumably with part of it sufficiently adhered to their non-driving raphe, these researchers were able to measure the amount of fiber deflection required to detach the diatoms from the substratum.

Micromechanical properties of the adhesive strands secreted from the diatom raphe can also be studied as shown by Higgins et al. (2002). Mechanical force was exerted on the attached strands by retracting the cantilever from the cell surface, thus allowing an investigation of the mechanical behavior of the adhesive(s) under stress, as well as the binding forces required to detach the strands from the tip. The first measurements in the study of the adhesive forces were hampered by the continual bridging of the adhesive strands between the cantilever tip and cell surface during successive extension–retraction cycles (Higgins et al. 2002).

AFM illustrations for taxonomy

Amphipleura pellucida Kützing has long been applied in microscopic studies as one of the standard objects used to determine the resolving power of light microscopy. The surface components of A. pellucida could be resolved by light microscopy if the used optical systems were of high numerical aperture (>1.40). Just with the first electronic microscopy studies (Kolbe & Golz 1943), features of A. pellucida were better resolved. From a taxonomical point of view, new data about A. pellucida is presented here, showing topographical images of the frustule with some of the valve’s typical features as ribs, raphe, and pores (Fig. 3), putting in evidence the small size of the pores which are found to be at the limit of the light microscope resolution. These AFM images also enable the comparison of pore size and height, as well as pore-to-pore distances.

Fig. 3
figure 3

Amphipleura pellucida. AFM microphotographs, showing the frustule topography. a, b Internal view of the central part of the valve showing the ribs fused in the middle. c External view of the middle of the valve showing the small pores (at the limit of the light microscope’s resolution); d, e Ribs bifurcating in the distal parts and enclosing a shortened raphe

Full size image

Amphipleura pellucida samples were treated with hot hydrogen peroxide method (using H2O2, 35%, 120 vol) in order to eliminate the organic matter and to obtain clean diatom suspensions. Finally, the oxidized samples were rinsed three times with deionized water by decantation. AFM analyses were performed under ambient condition in tapping mode with a PicoSPM® LE with large scanner controlled by Pico Scan software (Agilent, USA). Diatom valves were analyzed using a tip with a radius curvature of 10 nm, a frequency range of 140–390 kHz and a force constant range of 3.1–37.6. The raw topographic image was tilted manually with SPIP software (Image Metrology). This procedure was done according to the methodology used in Hlúbiková et al. (2012).

Future perspectives and conclusions

Diatoms nanotechnology, a new interdisciplinary area, attracted attention of scientists since 2005, emphasizing recent advances in diatom biomineralization; thus, AFM can help in detecting small differences in surface morphology ultrastructure and biomineral formation, showing a diversity of nano and mesoscale silica morphologies. Effortless sample preparation, imaging of conductive/non-conductive materials and transparent structures, ability to resolve slight differences in surface height, low cost operation related to SEM, and no need of coating, are the major advantages of AFM. On the other hand, AFM enables study of the properties of mucilage layers and EPS revealing the topology of the soft hydrated mucilage layer. Polymer networks and related micromechanical properties, as for example adhesion, can provide new approaches for the creation of antifouling materials, providing new insights into general diatom study and future applied technologies.

To sum up, AFM applicability in diatom studies relates a wide range of scientific disciplines and covers and extensive range of possible applications. It represents a rich information source and important basis for many research fields dealing with diatoms and should certainly be in attention of scientists.