Supramolecular gels are a versatile class of materials, and are of interest in many diverse applications from energy storage to cell culture . The versatile nature of soft matter systems has been of interest to materials scientists for decades . The large number of materials and processing techniques under development will allow these materials to be utilised in increasingly sophisticated applications such as drug delivery, environmental remediation, sensing, optoelectronics, photo-responsive actuators, and healable materials, and a range of biomedical applications among many others . To better understand and guide the development of these chemical systems, robust characterisation techniques and protocols are required. These materials utilise the self-assembly of small molecules into a network of long anisotropic structures which can entangle or cross-link. This network can immobilise the solvent, resulting in a material with both solid-like and liquid-like characteristics. The versatility of these compounds arises from their self-assembly across many length scales (Figure 1), allowing for the control of properties through not only chemical modifications on the monomer, but also changes to the gelation process . Whilst this tunability is advantageous in that it allows for a wide range of functionality and applications, this also gives rise to difficulties in understanding and predicting gel structure and properties . In order to understand the supramolecular structure, a robust characterisation regime is required. A comprehensive work edited by Weiss and Térech discussed in detail the plethora of chemical systems, characterisation techniques, and applications of molecular gels; which still remains deeply insightful despite the rapid advancement of the field . A more recent review by Yu et al. summarised a number of techniques used to characterise supramolecular materials . Two of the most commonly used methods of characterisation of the larger structures present are microscopy and scattering techniques . Across the literature, many investigations rely solely on one of these techniques alone to describe their system. However, both microscopy and scattering have limitations which may prevent a robust characterisation from being achieved. In this review, we aim to highlight and discuss a number of strengths and limitations provided by these techniques and will show how using these characterisation methods to complement each other can provide a much better understanding of self-assembling systems.
Figure 1: Hierarchical assembly occurring across length scales. Molecular interactions result in fibres which can form a network giving rise to the bulk gel.
Review Imaging techniquesImaging is the most widely used and accessible technique for characterising these gel materials. Researchers are looking for the presence of fibrous structures, coils, entanglements, spherulites and other such features that can give them some insight into how the gel is formed in the network via the assembly of interlocking nanostructures and microstructures. Imaging is often inexpensive and accessible both in terms of access and interpretation of the data given. These images are attractive and give an instant visual summary of the network. Depending on the type of imaging used, multiple different length scales can be probed to identify how the individual fibres form and then entangle to form the gel network. In addition to showing the shape of the self-assembled structures, with the aid of software tools such as ImageJ , features such as the fibre width, fibre length, and pore sizes can be quantified . However, measurements are limited to the snapshot captured by the image and require that the fibres are manually measured. This can be error-prone, inconsistent, and time consuming which can both require significant expertise to perform and limit the number of samples that can be feasibly analysed . A recent review by Kubota et al. summarised both the spatiotemporal advantages and disadvantages, as well as the artefacts arising from sample preparation and imaging conditions in the microscopy of molecular self-assemblies . When imaging supramolecular gels, atomic force microscopy (AFM), electron microscopy (EM) techniques including transmission electron microscopy (TEM), scanning electron microscopy (SEM), and scanning transmission electron microscopy (STEM), and confocal laser scanning microscopy (CLSM) are the most common techniques for imaging self-assembled structures. Some of the key features and capabilities of these techniques are summarised in Table 1.
Table 1: Comparison of imaging techniques for organic molecular assemblies.a
CLSM AFM/high-speed AFM TEM/SEM/STEM Cryo-TEM Liquid cell EM spatial resolution 200 nm ≤1–50 nm 1–10 nm 0.2–10 nm ≤3–30 nm acquisition speed (frame−1) 1 s–1 min 100 ms–10 s 30 ms–1 s sample environment and preparation solution, fluorescence labelling dried or solution, sample on substrate dried, sample on grid, (stained with metal) vitrified, sample on grid solution, sample in liquid cell artefacts bleaching, laser toxicity, dye addition tip, force, scanning dehydration, beam damage vitrification, ice crystal formation, beam damage beam damage, diffusional constraint advantage in situ imaging, multicolour imaging, 3 D imaging by Z stack in situ imaging, no staining, imaging with various parameters, high vertical resolution high resolution solves near-atomic structures (single particle analysis), snapshot in situ imaging limitation low spatial resolution, Imaging restricted to fluorescence labels restricted to surface no in situ imaging no in situ imaging, thin film needed lower resolution than other EM techniques, thin liquid sample neededaTable 1 was adapted from , R. Kubota et al., “Microscopic Imaging Techniques for Molecular Assemblies: Electron, Atomic Force, and Confocal Microscopies”, Chem. Rev., © 2021 American Chemical Society. This content is not subject to CC BY 4.0.
CLSM is a fluorescence microscopy technique capable of acquiring high-resolution optical images at selected depths of the sample. This is achieved by exciting compounds with laser light and, by using a pinhole to select focal planes, allows fluorescence at various depths to be focused and imaged. The resulting 2D focal plane images can be stacked to yield a 3D fluorescence image of high resolution, allowing complex morphologies to be visualised (Figure 2) .
Figure 2: Three-dimensional CLSM image of a multicomponent supramolecular structure. The three-dimensional CLSM image is constructed from z-stacked x–y slice images. Figure 2 is from and was adapted with permission from Springer Nature from the journal Nature Chemistry (“In situ real-time imaging of self-sorted supramolecular nanofibers” by S. Onogi; H. Shigemitsu; T. Yoshii; T. Tanida; M. Ikeda; R. Kubota; I. Hamachi), Copyright 2016, Springer Nature. This content is not subject to CC BY 4.0.
While CLSM provides in situ 3D imaging, meaning no drying is required of the material, the material itself needs to be fluorescent (but not quenching upon assembly) or the addition of a dye is required. Such dyes need to be able to stick with the fibres, rather than sitting in the pores. Therefore, the fluorescent material is either chemically attached to the molecules themselves or is interacting with them. Introducing fluorescent dyes to self-assembling systems affects their chemical and mechanical stability as the hydrophobic features can destabilise the self-assembled network . This reliance on fluorescence labels or dyes leading to a change in the self-assembly, meaning a true representation of the supramolecular material may not be investigated which is rarely confirmed when applied to the characterisation of such systems. To identify such changes in structure, other characterisation techniques that probe the structures without the use of fluorescence probes should be used in tandem with CLSM. For example, Raeburn et al. showed that the addition of molecular rotors and fluorescent dyes affects the bulk network of self-assembled materials using rheology. The co-packing was observed to alter rheological properties of the materials arising, for example, from differences in the rate of fibre growth and entanglements. The change in rheological properties when comparing the material and its co-assembly with the probe shows that addition of such probes may lead to measurements which do not truly represent the system of interest . The work by Alakpa et al. highlights how important it is to consider how one assesses the impact of co-assembly on a molecular system . AFM suggested that the addition of the surfactant-like Fmoc-serine (Fmoc-S) to the fibrous Fmoc-diphenylalanine (Fmoc-FF) did not impact the structure of the Fmoc-FF fibres (Figure 3). However, Fourier-transform infrared spectroscopy indicated that the Fmoc-S assembled on the surface of the Fmoc-FF fibres. The addition of carboxylate ions at the fibre surface allows cross-linking to occur, altering the bulk physical properties of the gel .
Figure 3: AFM images of air-dried aqueous Fmoc-FF, Fmoc-S, and 1:1 Fmoc-FF:Fmoc-S solutions. Figure 3 was reprinted from , Chem, vol. 1, by E. V. Alakpa; V. Jayawarna; A. Lampel; K. V. Burgess; C. C. West; S. C.J. Bakker; S. Roy; N. Javid; S. Fleming; D. A. Lamprou; J. Yang; A. Miller; A. J. Urquhart; P. W. J. M. Frederix; N. T. Hunt; B. Péault; R. V. Ulijn “Tunable Supramolecular Hydrogels for Selection of Lineage-Guiding Metabolites in Stem Cell Cultures”, pages 298–319, Copyright (2016), with permission from Elsevier. This content is not subject to CC BY 4.0.
While the proportions of fluorescent probes in the self-assembled material will be much smaller than the example discussed here, it is important to use appropriate methods to show that the addition of any co-assembling component does not impact the nanoscale structure and bulk network. To our knowledge there are very few examples in the literature that undertake analyses to show that the addition of a co-assembling probe does not disrupt the structure of the unmodified material. To adequately show that such adapted systems are relevant to the material of interest, future works should confirm that the modified systems are indeed representative. CLSM also often has poor resolution due to the use of light used for imaging. Other fluorescent imaging techniques such as super resolution fluorescence microscopy (for example, stimulated emission depletion (STED) microscopy or point accumulation for imaging in nanoscale topography (PAINT)) can give amazing details and resolution . Techniques such as STED and PAINT require structures that are specifically designed and stable enough for the longer imaging times, and so they are often not suitable for many systems. Another advantage of these techniques is the possibility of using different probes in a multicomponent network. Onogi et al. reported one of the first examples of direct imaging, in situ, of self-sorted supramolecular nanofibres . TEM was able to show that the mixture of gelators gave rise to nanofibres with a morphology similar to their single-component counterparts. The similarity between these systems made it difficult to assess whether the mixture of gelators resulted in self-sorted or co-assembled fibres using TEM alone. By developing distinct fluorescent probes which could selectively aggregate with, and subsequently stain the supramolecular fibres, CSLM was used to identify the mode of molecular assembly of each component. By comparing the CSLM images arising from the different fluorescent probes, there was shown to be a weak correlation between the resulting images, indicating that the fibres self-sort and separately entangle each other. By obtaining images in three dimensions, the orthogonal assembly could be directly visualised throughout the materials microstructure (Figure 2) . This in situ method of CLSM also allows the kinetics of self-assembly to be elucidated, as demonstrated by Wang et al. . By using fluorescent derivatives to visualise a multicomponent system consisting of neutral and charged gelators (NG and CG, respectively), the self-assembly of neutral and charged fibres (NF and CF, respectively) could be visualised over time. CSLM shows the self-assembly proceeds by the initial self-sorting of NGs to assemble NFs. The self-sorting of NFs decreases the concentration of NGs, increasing the relative concentration of CGs, allowing a higher critical assembly concentration of the CGs (arising from electrostatic repulsions between the CGs) to be reached, leading to the co-assembly of CFs. Not only could CLSM observe the initial kinetic self-sorting of gelators, but it also showed that the homogenous mixture of NFs and CFs can undergo a higher level of self-sorting, resulting in macroscopic phase separation (Figure 4).
Figure 4: (a) 3D CLSM images of macroscopically a self-sorting gel network, where all fibres were stained green, but only CFs incorporated blue stain. (b) 2D CLSM images showing the formation and self-sorting of NFs and CFs over time. (c) Illustration of the multilevel self-sorting process. Figure 4 was adapted from , Y. Wang et al., “Hierarchically Compartmentalized Supramolecular Gels through Multilevel Self-Sorting”, J. Am. Chem. Soc., © 2018 American Chemical Society, distributed under the ACS AuthorChoice/Editors' Choice via CC-BY-NC-ND Usage Agreement, https://pubs.acs.org/page/policy/authorchoice_ccbyncnd_termsofuse.html. This content is not subject to CC BY 4.0.
To gather a three-dimensional surface profile of a material, AFM utilises a highly sensitive cantilever to measure the nanometre to sub-millimetre surface morphology. Attractive or repulsive interactions between the surface and the cantilever induce a bending force in the cantilever providing a mechanical means of probing the surface nanostructure . There are various modes of interaction available to AFM to probe the surface topology. The simplest is constant-height mode, wherein the sample is scanned with the cantilever at a constant height while measuring the deflection. Constant-height mode may result in large deflections and lead to sample damage by inducing lateral shear . To minimise this, a constant-deflection mode can be employed by instead varying the cantilever height with a feedback loop to keep the deflection forces constant. The feedback output is used to generate a height image, however, often an imperfect feedback loop results in an error signal which generates a deflection image. The height image can be used to quantify the height and thickness of surface features, while the deflection image can probe fine surface details . To effectively characterise soft materials, the minimisation of surface forces is essential. This fact made tapping-mode AFM, wherein the cantilever oscillated vertically to reduce the tip–sample contact, a revolutionary method for the characterisation of soft materials . As well as probing the height of the surface, the phase-imaging afforded by the tapping-mode AFM can probe the viscoelastic properties and adhesion forces to qualitatively distinguish different materials in a sample . By relying purely on the interactions between the cantilever and the material surface AFM can overcome the limitations of using fluorescent probes. However, by investigating only the surface of the materials, the bulk gel (which may not be the same as the surface) is inaccessible and cannot be characterised. AFM has the advantage of allowing for the measurement of hydrated samples, and therefore does not suffer from drying artefacts. Since hydrogels are mainly made of water and low concentration hydrogels can be very soft, AFM measurements face the problem of sample drift . Avoiding drift artefacts is difficult so AFM is often performed in a dried state, but imaging of hydrogels in wet state has been achieved on thin films to minimise drift artefacts. Unlike EM techniques, AFM does not require treatments such as metal coating that can alter the structure of the material . AFM was recently used by Sambani et al. to characterise the self-assembly of elastin nanofibres . The high resolution imaging across length scales allows a range of sizes to be probed, identifying the fibres which arise from the formation of fibrils, which are themselves formed from nanofibrils (Figure 5). It is worth noting that difficulties arise in AFM with dealing with such differences in scale as an excessively large step height can result in artefacts that make it hard to detect smaller structures .
Figure 5: (a) 3D AFM topographic image of dried elastin fibre. (b) Indicative height and diameter profile plot of a fibre, with the red line representing the cross-section analysed in plot (c). Height and diameter of the fibre is calculated between the blue and green arrows, respectively. Figure 5 was adapted from (© 2018 K. Sambani et al., published by MDPI, distributed under the terms of the Creative Commons Attribution 4.0 International License, https://creativecommons.org/licenses/by/4.0).
EM techniques, including scanning electron microscopy (SEM) and transmission electron microscopy (TEM), have proven to be some of the most robust and popular supramolecular characterisation techniques. SEM raster scans a fine beam of electrons over a sample surface. As electrons strike the surface, interactions between the beam and sample result in the emission of secondary scattered electrons, back-scattered electrons and X-rays which can be detected to produce an image. In contrast, TEM can measure the bright-field image, among other modes, by blocking the scattered electrons and detecting only the unscattered electrons. This results in a bright-field image where areas that are actively scattering have fewer electrons, resulting in a darker contrast . Together SEM and TEM can enable multiscale quantification of nanostructural and microstructural features (Figure 6) .
Figure 6: The nano-to-micro imaging range of SEM and TEM . Cartoons represent the nanoparticles, pores, nanowires, and composites (from left to right, respectively) which can be probed by electron microscopy.
The versatility of EM can reveal a number of structural features, but the sample preparation required can lead to significant changes in the materials morphology. EM requires many preparative steps to allow structures to be imaged including the application of an ultrathin coating of carbon, gold, or platinum. These are applied to conduct away accumulated surface electrostatic charge . Furthermore, the requirement that TEM samples are prepared as thin films in order to allow electrons to be transmitted through the sample may require mechanical abrasion or ‘blotting’ of the sample to achieve this . Supramolecular materials which are particularly sensitive to shear forces, will be particularly sensitive to morphological changes from the stresses of abrasive thinning. SEM and TEM also may require a fundamental change in the material in order to investigate it, as their high-vacuum environment necessitates that samples are dried. This is problematic for gels in particular, as gels are mostly solvent by composition, it is to be expected that the drying procedures required may lead to a change in morphology due to effectively concentrating the sample, leading to more aggregation, any salts present in the solution could also crystallise and disrupt the network, or change the stability and solubility of any charged structures present. All of these issues could result in statistics which are not representative of the structures in situ . To overcome this, cryogenic techniques are widely used to allow the investigation of a material in its hydrated state.
Cryogenic techniques utilise a vitrification process to maximise the formation of vitreous ice to minimise the formation of ice crystals which can disrupt the material structure. The ability of cryo-EM to probe the structure of soft nanostructured materials has led to a number of reviews discussing its pivotal nature in supramolecular material characterisation . One of the most common methods of vitrification involves plunging the hydrogel into a liquid propane/ethane mixture, transferred under vacuum to a preparation chamber where it is fractured, sublimed, and coated for imaging . While propane/ethane is the most common cryogenic medium, various media exist with different heat transfer capabilities which can lead to differences in the vitrified structure . It is important to consider that the increased thickness of cryo-SEM reduces the surface area-to-volume ratio which reduces the cooling rate reached in the vitrification process. High-pressure freezing (HPF) methods are capable of fixing samples 100 times thicker than plunge freezing. HPF reduces the growth of crystals during the fixation of cryo-SEM samples by using very high pressures of liquid nitrogen, in the order of 210 MPa, and has been reported to improve the preservation of cryo-SEM specimen nanostructure . While allowing thicker samples to be investigated avoids artifacts introduced by preparing thinner samples, cryo-SEM is limited to investigating the surface topology which is not necessarily the same as the bulk morphology.
Cryo-TEM can better probe three-dimensional structures, however, it is important to consider the additional artifacts that arise in cryo-TEM sample preparation. Cryo-TEM requires a thin film sample which can be difficult to prepare with a gel sample . Blotting is a standard technique used to remove excess material from a TEM sample grid, however, even with the advent of robotic instruments it has been shown to be wrought with issues . Not only are the preparations a nontrivial task, but the resulting films are often highly variable, non-uniform, and subject shear stresses which can align, segregate, or transform structures (Figure 7) .
Figure 7: Cartoon of artifacts caused by blotting and thinning. a) Alignment of threadlike micelles (left) and the transformation of threadlike structures into transient lamellar structures (right) . b) Top-down (middle) and cross-section (right) views showing the segregation of structures with small assemblies pushed towards the thin central area and larger ones driven to the thicker areas at the cryo-TEM grid boundary .
Though assemblies at multiple length scales can be captured by cryo-TEM, blotting can cause larger structures to be expelled from the sample, limiting the ability of cryo-TEM to observe the range of hierarchical assemblies . Another significant concern posed by blotting is the brief introduction of an air–water interface which can lead to preferential organisation of particles, as well as unpredictable changes in local concentration on the TEM grid . To overcome such limitations, methods such as those by Arnold et al. have been developed to achieve the preparation of cryo-TEM grids without the need for blotting . A recent review of cryo-TEM preparation techniques covers a number of deposition techniques, their benefits, and limitations was performed by Weissenberger et al. . Even when successfully prepared, the limited thickness can make imaging gel networks difficult . It should also be considered that changing the surface or vessel in which the material is prepared can lead to a change in the microstructure . Therefore, the resulting material on the microscopy grid may not be representative of the typical structures even before drying or vitrification. It is also important to note that any electron microscopy image represents only a small portion of the overall sample and so the resulting characterisation may not account for inhomogeneities throughout the bulk network .
Another significant limitation posed by cryogenic microscopy techniques is their compatibility with organic solvent systems. Firstly, the vitrification procedure poses significant problems for the typical cryogens of liquid nitrogen and liquid ethane. Whilst liquid nitrogen is compatible with organic solvents, the lower cooling rate of ≈7000 K/s may be insufficient to vitrify the materials within organic solvents. Whereas the cooling rate of ethane of up to 100,000 K/s would overcome this, most organic solvents are soluble in liquid ethane making it incompatible with such systems . Furthermore, even when vitrification is achieved, organic solvents are much less stable than aqueous samples under the electron beam. Currently, a typical high resolution cryo-TEM study generally requires a dose of ≥100 e−/Å2, however, vitrified organic solvents have reportedly shown beam damage effects at doses of <50 e−/Å2, making high-resolution cryo-TEM infeasible for many organic solvent systems .
Whilst imaging may have some caveats, it can be a hugely powerful tool when looking at larger areas of materials and looking for uniformity of samples. In materials formed from fibrous nanostructures, the interactions between fibres often dictate the mechanical properties of the material . Therefore, it is crucial to be able to characterise the topology of these structures to better understand how they aggregate. Imaging techniques allow structural elements of a system to be directly observed. This is a significant advantage over scattering which does not provide a direct visualisation, and can be used to better understand the process which gives rise to the observed structure. This advantage was demonstrated by Yuan et al. with their investigation of supramolecular helices . SEM was used to investigate the structures resulting from the self-assembly of a bis-bipyridinium-based compound (1·4Br) resulting in helical supramolecular fibres with different chirality induced depending on the enantiomer of tryptophan present. The direct images obtained from SEM allowed the chirality of the resulting fibres to be easily observed, indicating that the hydrogel matrix induced by racemic tryptophan was comprised of racemic amounts of left-handed (M) and right-handed (P) helices (Figure 8) . Due to the orientational averaging of scattering techniques, directional information is lost. Subsequently, extracting information such as chirality cannot be achieved without additional information such as that obtained from direct imaging . A more comprehensive discussion of the capabilities and limitations of scattering techniques is included later in this review.
Figure 8: (a) Chemical structures of monomer compounds and a schematic of the resulting chiral helical structures. (b) SEM images of the resulting supramolecular helical fibres. (c) SEM images depicting fibre bundling leading to larger fibres with conserved helical chirality. Figure 8 was reproduced from , T. Yuan et al., “Assembly and Chiral Memory Effects of Dynamic Macroscopic Supramolecular Helices”, Chem. Eur. J., with permission from John Wiley and Sons. Copyright © 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim. This content is not subject to CC BY 4.0.
SEM was also able to identify fibre–fibre bundling where smaller fibres further assemble to form thicker fibres while maintaining their helicity. This allowed for the formation and observation of enantiomerically pure fibres of increasing size (Figure 8c). The direct image yielded by microscopy allows the self-assembly process to be easily interpreted, where the rationalisation of such structures may not be achievable with complex indirect imaging techniques. This advantage is epitomised by Jones et al. whose self-assembling urea compounds form helices which can further entangle into a number of significantly more complex braided structures observable by SEM (Figure 9) .
Figure 9: Commonly observed entanglements of urea-based supramolecular helices. (a) Double helix, (b) quadruple helix, (c) Brunnian braid, and (d) nested helices. Figure 9 is from and was adapted by permission from Springer Nature from the journal Nature Chemistry (“Braiding, branching and chiral amplification of nanofibres in supramolecular gels“ by C. D. Jones; H. T. D. Simmons; K. E. Horner; K. Liu; R. L. Thompson; J. W. Steed), Copyright 2019, The Author(s), under exclusive licence to Springer Nature Limited. This content is not subject to CC BY 4.0.
An important feature of self-assembled networks is the defects that arise in the fibre structures as shown in the characterisation of β-hairpin peptide hydrogels by Yucel et al. . The cryo-TEM characterisation indicated that the formation of the hydrogel was directed by self-assembly defects giving rise to branches in the fibrillar structure which was distinct from other reported biopolymeric systems. If the cross-link was a result of entanglement as with similar systems, then the junction would show an increase in contrast due to the two fibril cross-sections increasing the interactions with the electron. The uniform contrast was instead indicative of the branching as a result of valine facial collapse of β-hairpin, giving rise to multifunctional branches which allow cross-linking.
The capabilities of imaging in elucidating the mechanism of aggregation is further exemplified by Jones et al. . It was shown that SEM could directly observe structural defects, resulting in the collapse of helically braided structures into their constituent fibrils which can interact with other fibres (Figure 10). The branching density resulting from these defects could be gauged by SEM since the braid crossings are topologically fixed and act as permanent junctions . Cross-linking caused by such defects can only be observed by localised techniques, since they do not represent the bulk averaged structure. Investigating such local structures can be achieved with imaging techniques while the statistical-averaging of scattering models results in the loss of this information.
Figure 10: (a) SEM image of a single three-stranded braid showing a defect in which the braid separates into separate fibrils. (b) SEM image showing that the separation of branch points in dried gel samples is highly variable. Figure 10 is from and was reprinted by permission from Springer Nature from the journal Nature Chemistry (“Braiding, branching and chiral amplification of nanofibres in supramolecular gels“ by C. D. Jones; H. T. D. Simmons; K. E. Horner; K. Liu; R. L. Thompson; J. W. Steed), Copyright 2019, The Author(s), under exclusive licence to Springer Nature Limited. This content is not subject to CC BY 4.0.
The staggering advancement of the field of electron microscopy continuously provides new opportunities to characterise structures with increasing resolution. Recently, Yip et al. was able to significantly improve upon the state of the art in cryo-TEM to achieve unprecedented structural detail. With a structural resolution of 1.25 Å, individual atoms and hydrogen density in the apoferritin can be visualised (Figure 11) .
Figure 11: Visualization of individual atoms at 1.25 Å resolution. Three apoferritin residues are shown at high density threshold. The true atomic resolution of the map is shown by the clear separation of individual C, N, and O atoms at high thresholds. Figure 11 is from and was adapted by permission from Springer Nature from the journal Nature (“Atomic-resolution protein structure determination by cryo-EM“ by K. Yip; N. Fischer; E. Pakni; A. Chari; H. Stark), Copyright 2020, The Author(s), under exclusive licence to Springer Nature Limited. This content is not subject to CC BY 4.0.
We highlight here two reviews which discuss the fast-paced development of cryo-TEM imaging. Cheng et al. has reviewed the decades of cryo-TEM developments which have given rise to single-atom resolution techniques . Chua et al. discussed a number of factors including advancements in sample preparation, instrumentation, and processing which have improved rapidly . These factors together have allowed electron microscopy to become better, faster, and cheaper; establishing it as a dominant technique for structural characterisation.
Scattering techniquesSmall angle scattering techniques have been used to investigate soft matter systems for decades, with pioneering work such as that by Pierre Térech showing how such techniques can be applied to probe both local and long-range structures of 3D networks . In the decades since there has been a wealth of reviews on soft matter characterisation as interest in such systems increases, and techniques develop at a rapid pace . SAS techniques include small angle X-ray scattering (SAXS), small angle neutron scattering (SANS), and their wide angle (WAXS) and ultra-small angle (USAXS and USANS, respectively) analogues have various advantages and drawbacks compared to one another (Table 2), as well as microscopy techniques in the characterisation of soft matter self-assemblies.
Table 2: Comparison of scattering techniques .
WAXS SAXS USAXS SANS USANS length scales (nm) 0.1–1 1–100 20–3000 2–200 500–2000 sample volume 0.1–500 µL 50–3500 µL radiation sources lab and large facility large facility only typical large facility acquisition time seconds minutes solvent deuteration not required for contrast deuteration often required for contrast artifacts beam damage isotope effects advantages time-resolved analysis, high-throughput contrast matching, magnetic scatteringThese techniques fire a monochromatic beam of radiation through a sample, with the majority of the radiation being transmitted (and blocked by a beamstop) without interacting with any structures in the sample. A small amount of the radiation will interact with either the electron density (SAXS) or the nuclei (SANS) of the structures in the sample resulting in Thompson scattering that can be detected. The resulting scattering is characterised by the scattering vector (q) which is the result of a photon or neutron of wavelength (λ) scattering off the sample at an angle of 2φ (Equation 1 and Figure 12).
Figure 12: Cartoon of a general small-angle scattering setup.
Typically, a 2D detector measures the scattered radiation to determine the scattering intensity (I(q)) and is measured as a function of q. The scattering vector has units of reciprocal space, so the smaller low-q measurements represent larger real-space distances (d) of repeating structures (Equation 2).
The scattering pattern characterises the electron or nuclei density throughout the sample, which is directly affected by the arrangements of molecules and atoms throughout the structure. Therefore, the observed scattering carries bulk-averaged information that is a statistical average of the structures across the macroscopic irradiated volume . Depending on the q-range measured, different structural scales can be investigated. As q increases, smaller structures are probed allowing the topology to be determined. For assembled structures in a solvent, there needs to be an appropriate difference in scattering length density (SLD) of the material compared to the solvent (called the contrast). There are online calculators that can be used to calculate the SLD of molecules and solvents . This is so that the solvent can be suitably removed from the scattering pattern and a model then be used to fit only the material of interest. For X-ray scattering, most solvents have a suitable difference in SLD and so can be used, but should be checked. However, in neutron scattering the protons scatter very strongly, and therefore deuterated solvents are used to essentially hide the solvent to ensure the solvent does not contribute to the scattering pattern and contrast can be achieved this way. Another way to create this contrast is to selectively deuterate the molecule of interest and leave the solvent deuterated. This gives the advantage of contrast matching, where parts of the molecule can be hidden so temporal mapping can be achieved, but all add to the cost and complexity of the experiments. This is discussed in more detail further on. A number of models exist which describe the expected scattering that would arise from a wide range of structural topologies. This fitting allows multiple parameters to be determined. For example, self-assembled materials form worm-like micelles which can be described by various cylindrical models parameters including structure length, Kuhn length, structure radius, and axis ratio . Scattering is representative of the overall system in comparison to the snapshots of individual micelles observed in microscopy techniques . One of the foremost benefits of SAS is the capability to perform in situ measurements, wherein samples can be analysed with minimal change to the materials environment. This allows for structures that are more representative of their typical environment. To achieve this, many specialised cells have been developed to control various environmental parameters including temperature, pressure, irradiation, electromagnetic fields, and flow . Sample environments can also be as simple as a capillary at ambient conditions. Without a need for the sample environment to be under vacuum, as is required for some microscopy techniques as previously discussed, SAS allows for volatile solvent systems to be characterised. In addition, cells that allow for the coupling of SAS with non-X-ray techniques continue to be developed; these include spectroscopy, rheology, and electrochemistry . Ascribing changes in properties of interest directly to changes in the structure further elucidates the interesting properties of self-assembled materials, and better enables rational design. It is important, however, to consider that small changes in the sample environment can lead to significant changes in the self-assembled structures. For example, in aqueous samples, the use of D2O is necessary to produce contrast in SANS analysis and may be assumed to provide an in situ environment. While D2O is chemically identical to H2O, it does exhibit differences in properties including density, viscosity, hydrogen-bond strength, a more pronounced hydrophobic effect. As hydrogen bonding and hydrophobicity are critical to hydrogel characteristics it has been shown that substituting the hydrogel solvent with D2O can lead to structural changes in self-assembled aggregates (Figure 13) . Therefore, by allowing for a better understanding of how self-assembled systems are structured, SAXS and SANS can be a more accurate tool for identifying structural characteristics that give material properties of interest. However, care must be taken to sure a truly representative environment is probed.
Comments (0)