Nanomaterials (NMs) composed of minerals make up by far the bulk of this planet and are a major part of the critical zone of the Earth. Nanomaterials have been considered one of the principal catalytic components of Earth throughout its history (Hochella et al. 2019), which include for example their roles in the polymerisation and self-assembly of the molecular building blocks of life during early evolution (Ferris and Ertem 1992, Oleson et al. 2010, Xu et al. 2012). Despite their importance through time, NMs are not considered in studies of biogeochemical processes throughout Earth's history, and overall are rarely considered major components of the Earth. Reasons for this may have been a lack of analytical tools and procedure to quantify the presence, size, shape and chemical and mineralogical composition of NMs. Only in the past 25 years, the properties of NMs have been rigorously studied and have been addressed in many review papers (Table 1). The majority of NMs form during weathering and alteration processes of their parent minerals or through heterogeneous nucleation processes involving bacteria, fungi and plants. Hence, associations of NMs with these environmental constituents are of great importance to understand their origin, properties and chemical and mineralogical composition. This review paper addresses the current state of analytical techniques that are used to characterise NMs in associations with minerals, organic matter and organisms.

Nanomaterials are organic, inorganic, or organometallic materials with size- and shape-dependent chemical, physical, and/or electrical properties (Hochella et al. 2019). They may be atomically ordered (crystalline), disordered, or amorphous, and range in size from less than a nanometre up to several tens of nanometres in at least one direction. Their properties depend on their chemical and mineralogical composition and small compositional or structural changes can result in very different chemical and/or physical properties. Contrary to their macroscopic counterparts (with the same mineralogical composition but with sizes larger than tens of nanometres in each direction), NMs have dramatically different chemical and physical properties that affect Earth systems in many important ways. For example, nanometre-size iron oxyhydroxides affect in contrast to their macroscopic counterparts the productivity of phytoplankton in oceans, which control the sequestration of carbon and consequently the carbon cycle and global temperatures (Coale et al. 1996, Blain et al. 2007).

The decrease in size from macroscopic minerals to NMs is associated with a decrease in the total number of atoms and with an increasing fraction of surface atoms (Banfield and Zhang 2001). An increasing fraction of surface atoms leads to a higher degree of surface hydration and protonation, which often results in the weakening of bonds in the surface structure and thus in an increasing destabilisation of NMs (Banfield and Zhang 2001). Depending on the nature of the surrounding environment, this destabilisation leads often but not always to a higher solubility of NMs relative to their larger counterparts (Tang et al. 2004, Dutta et al. 2017, Ramos et al. 2017, Saikia et al. 2018, Sánchez-Peña et al. 2018, Oliveira et al. 2019, Lütke et al. 2020), an important aspect in terms of the bioavailability of toxic elements incorporated into NMs.

The fewer atoms in NMs result in fewer electron energy levels (as observed in micro- to bulk-size materials), and electron energy bands become discrete electron energy states in NMs, thus affecting their physical properties such as conductivity and light absorption (i.e., colour) (Hochella 2008). Higher surface reactivities and different physical properties also alter growth, dissolution, melting, evaporation, and aggregation state of NMs and thus play key roles in their environmental fate. The larger surface area of NMs relative to their larger counterparts also allows them to sorb and transport more effectively solutes, controlling often the fate of hazardous organic compounds and toxic elements in contaminated areas of the critical zone (Hochella and Madden 2005).

The stability of NMs towards dissolution, aggregation and transformation depends among many factors, on their size (which controls their surface energy), surface charge (controlled by pH), the presence of counterions and the presence of polymerised species on the surfaces. All these factors control for example, the stability of Fe-(hydr)oxides (Fe-Ox) (Navrotsky et al. 2008, Guo and Barnard 2013), with adsorbed silica species being the most prominent species that can stabilise FeOx NMs. It should be noted, and this is important for the later discussion, that (a) NMs can be also products of the aggregation of even smaller constituents such as clusters in aqueous solution (considered precursors of NMs); (b) the interaction of NMs with organic matter leads to the stabilisation of the latter and (c) dissolution, aggregation and transformation as well as adsorption processes involving NMs do not only occur in the critical zone but also in the atmosphere, which can result in modified particulate matter (PM) released by for example wildfires. One commonly distinguishes three groups of NMs: (1) Natural NMs which are part of the element cycling on Earth and which form through abiotic and biotic-controlled processes; (2) Incidental NMs which are unintentionally by-products of human activity and have steadily increased since the beginning of the Industrial Revolution; (3) engineered NMs used in many applications due to their designed physical and chemical properties. As the latter group of NMs is less common in the environment, we will focus in this review on natural and incidental NMs (Hochella et al. 2019).

Objectives

Nanomaterials form during abiotic and biotic alterations of minerals or organic matter or nucleate on their surfaces or within their pore spaces. Consequently, NMs are often associated with larger constituents such as mineral grains, organic matter, and organisms such as microbes, fungi and plants. In many cases, structure, stability, chemistry, charge, electronic and magnetic properties and reactivity of NMs are products or controlled by the interaction of NMs and their host materials (Becker et al. 2010). Hence, a thorough understanding of the formation and long-term environmental fate of NMs require the characterisation of associations between NMs and minerals, organic matter and organisms (from here on referred to as NM-associations) as individual entities rather than focusing exclusively on NMs.

The study of NM-associations should include the chemical and mineralogical characterisation of NMs as well as of their hosts, their spatial and textural relation and in the case of their biotic-controlled formation, the identification of the microbial community composition. The analytical characterisation of NM-associations is however challenging as the chemical and physical properties of NMs and associated larger materials and organisms can greatly differ from each other. For example, the chemical and mineralogical characterisation of NMs at the bulk scale is often hindered due to their smaller volume and lower degree of crystallinity relative to their hosts. Furthermore, the textural characterisation of the interface between NMs and their micrometre-size host is not straight forward as this interface is fragile due to differences in hardness and porosity and thus cannot often be preserved during preparation for electron microscopy studies. Here, the range of small versus large volumes of NMs and soft NMs versus hard micrometre-size hosts varies for example between NM-bearing mineral surface coatings in sand- or sandstone-based aquifers to NMs in clay-rich soils to apatite-precursors in association with bioapatite. In sand or sandstone aquifers, NMs are commonly undetected with bulk analytical techniques due to the high proportion of quartz and the heterogeneous and nanocrystalline character of the coatings (Coston et al. 1995, Penn et al. 2001). The phases in the latter coatings are often hydrous and soft in comparison to the hard micrometre-size quartz grains. In clay-rich soils, NMs occur in large volumes as detectable clay minerals, whereas only a small fraction of NM-precursor phases to clay minerals such as amorphous Al-silicates, Fe-Ox and silica remain still undetected with bulk analytical techniques (Tazaki and Fyfe 1987). The latter phases occur commonly on the surface of weathered silicate minerals such as minerals of the feldspar group and are much softer and hydrous relative to their hard micrometre-size silicate minerals. In the case of nano-size precursors to bioapatite such as amorphous calcium phosphate (ACP) and/or octacalcium phosphate (OCP) (Brown and Chow 1976), the difference in their hardness with respect to bioapatite is smaller than between clay minerals and quartz/feldspar and their associations are less difficult to prepare for studies at the micro- and nanometre scale (Witzke et al. 2022). Considering possible differences in volume and physical properties of NMs relative to their macroscopic counterparts, the objectives of this review are: (1) to provide a short overview on different types of NM-associations identified in sediments, soils, ore deposits, tailings and PM, (2) to describe common analytical techniques used for their characterisation, (3) to recommend analytical approaches and to point out advantages and disadvantages of preparation techniques, (4) to suggest new analytical approaches for the study of NM-associations in soils and PM.

There have been numerous review papers on analytical and preparation techniques for the characterisation of NMs in the environment (Table 1), but these reviews focused mainly on NMs rather than on NM-associations. Similarly, reviews exist for each of the below-listed analytical- and preparation techniques (Table 1) and readers interested in detail in one of the techniques should refer to those. This review paper will also not address analytical or computational techniques used to characterise NM-associations in experimental studies (e.g., atomic force and scanning tunnelling microscopy studies of mineral surfaces modified in laboratory experiments). Readers should refer here to the review articles by Becker et al. (2010) and Putnis and Ruiz-Agudo (2021).

Occurrence of nanomaterials in association with minerals, organic matter and microbes

This section addresses the formation of natural and incidental NMs and the environments in which they are associated with minerals, organic matter and microbes. Silicates, oxides, phosphates, and other minerals that are unstable under Earth's surface conditions dissolve or react to form other phases. Intermediate products of this equilibration process are often natural NMs such as amorphous gel-like precursors (Combes et al. 1989, Schindler et al. 2009a, Mantha et al. 2019), which transform into NMs such as amorphous or opaline silica (Jones et al. 1966, Guthrie et al. 1995, Banfield and Barker 1998), hydrous aluminosilicates such as allophane (Henmi and Wada 1976, Wada and Wada 1977), the tubular aluminosilicate imogolite (Gustafsson 2001, Schindler et al. 2019), clay minerals such as kaolinite (Schindler et al. 2019), oxides (e.g., magnetite and haematite; (Banfield and Eggleton 1990, Schindler et al. 2019) and oxyhydroxides (e.g., ferrihydrite and goethite, Schindler et al. (2019), Figure 1a). Note that the minerals allophane and ferrihydrite only form particles in the nano-size range and are termed nanominerals. Other minerals such as 1:1 or 2:1 sheet silicates can undergo further “solid-state conversion” (which is essentially a nano-scale dissolution-reprecipitation process within the sheet structure) to form either interstratified clay minerals (mixtures between different types of sheet silicates) or sheet silicates with a different structure-type such as chlorite (Środoń 1999). During these types of weathering processes, minor elements in the parent material such as Ti, Fe, Cr and Au may be expelled and re-precipitate as anatase (TiO₂), Fe-Ox, chromite and gold NMs (Schindler et al. 2017). Hence, a first generation of NMs can further transform via dissolution-reprecipitation processes leading to the formation of numerous types and generations of NMs.

In the case of an incomplete equilibration process of the mineral (parent phase), newly formed NMs (daughter phases) are often associated with the parent phase, and the spatial and textural relations between the different daughters and parent provide insights into the underlying mechanisms of the equilibration process and in the sequence of formation of the daughters (Figure 1) (Schindler and Hochella 2015, 2016). Similarly, NM-mineral or NM-organic matter associations can form when fluids enter pore spaces of altered rocks, minerals and organic matter in which changes in the environmental conditions or number of nucleation or reactive surface sites leads to higher degrees of supersaturations with respect to the NM (Figure 1; Jadoon and Schindler 2022, Jadoon et al. 2022, Alavijeh et al. 2023). NM-organic matter associations can also form when organic matter sorbs to mineral surfaces in soils (Kleber et al. 2021). In this case, however, the minerals do not have to be nanometric and the formation of mineral-organic matter associations are controlled by the surface charges of the minerals, functional groups of the organic species and charge-neutralising constituents such as extracellular polymeric substances (Guhra et al. 2019).

Contrary to porous mineral- and organic matrices, the formation of NMs within denser high-T minerals (e.g., sulfides, oxides and silicates) occurs when (a) the concentration of an element exceeds the amount that can be structurally incorporated into their hosts (which is temperature dependent) (Reich et al. 2005, 2006); (b) the formation of NMs is energetically more favourable than the structural incorporation of its constituents into the mineral structure (Becker et al. 2010); (c) the growth of the host material is slow and thus provides enough time for the assembly of ions towards NMs (Fougerouse et al. 2016). Well-known examples are the formation and diffusion of Au nanoparticles in pyrite, arsenopyrite, amorphous silica, calcite, rutile and Fe-Ox within epithermal, porphyry and orogenic ore deposits (Reich et al. 2005, Reich et al. 2006, Hough et al. 2011, Fougerouse et al. 2016, Saunders and Burke 2017, Hastie et al. 2021, McLeish et al. 2021).

The formation of NMs such as Fe-Ox, aluminosilicates, Ca-carbonates, and Ca-phosphate can be also induced or controlled by biological processes (Banfield and Zhang 2001, Schindler and Hochella 2015, Sharma et al. 2015). During biologically induced mineralisation, NMs are formed indirectly through redox reactions in the microbial environment related to metabolic processes such as the oxidation of Fe²⁺ to Fe³⁺ or the reduction of sulfate to sulfide (Sharma et al. 2015). In contrast, the formation of NMs during biologically controlled mineralisation is entirely controlled by organisms. Here, NMs form in the microbial cells under certain conditions and are composed of well-defined crystals with narrow particle-size distributions (Sharma et al. 2015). Nanomaterials formed by biologically controlled mineralisation have various functions for the organisms such as the use of magnetite nanoparticles for navigation purposes in magnetotactic bacteria (Schüler and Frankel 1999). Similar to NM-mineral associations, preservation and examination of the spatial and textural associations between NMs and microbes, fungi and plants is of great importance for our understanding of the role of living organisms in alteration processes and thus in the re-cycling of elements within the Earth's crust (Finlay et al. 2019, Wild et al. 2022).

To conclude, natural NMs can form during abiotic- and biotic-induced and controlled processes with weathering processes providing the largest amount of NMs (Hochella et al. 2019). In comparison, incidental NMs are released as NMs during human activity or form during weathering or transformation of by-products of human-activities. Nanomaterials are directly released from human activities during for example combustion, smelting, metallurgical processes and transportation, whereas weathering and transformation of by-products can occur in mine tailings, aquatic bodies, water treatment plants, soils or even shooting ranges (Schindler et al. 2020). Incidental and natural NMs formed during weathering have similar chemical and physical properties whereas incidental NMs directly emitted as NMs can have unique chemical and physical properties rarely observed in the environment. For example, high-T combustion and smelting processes can result in the release of highly reactive NMs such as O-deficient Ti-oxide phases (so-called Magnelli phases) and metallic Pb (Batonneau et al. 2004, Yang et al. 2017), which have not been identified as products of wildfires, ore-forming processes or volcanic activity.

NM-associations involving incidental NMs occur for example in mine tailings. Here, NMs such as ferrihydrite, schwertmannite and jarosite are common constituents and form on the surfaces of primary Fe-sulfides and silicates during their alteration (Bigham and Nordstrom 2000, Blowes et al. 2003). The stability of these NM-associations is of great environmental concern as the corresponding NMs can sequester and mobilise metal(loids) such as As, Pb, Cd, Cu and even transport those for many kilometres within a hydrographic basin (Rodriguez-Iruretagoiena et al. 2016). Incidental NM-associations also occur in mining- and smelter-affected soils. For example, elementary Cu and Ag nanoparticles as well as Cu- and Ag-sulfides and -oxide nanoparticles form in pore spaces of organic matter within the surface layers of organic-rich soils (Figure 1f) (Mantha et al. 2019, Jadoon and Schindler 2022, Jadoon et al. 2022, Alavijeh et al. 2023).

Similar to mine tailings, coal combustion and oil refineries are major anthropogenic activities that generate NM-associations. For example, NMs emitted by oil refineries include fullerenes, graphenes, C nanotubes, and several amorphous phases (Sanchís et al. 2013), whereas those emitted by coal power plants and coal combustion processes are chemically and mineralogically more complex and are composed of fullerenes, C-nanotubes, soot and tar particles as well as amorphous phases, mainly sulfate- and silicate-minerals and complex mixtures of amorphous phases and minerals (Kronbauer et al. 2013, Sanchís et al. 2013, Wilcox et al. 2015, Ward 2016, León-Mejía et al. 2018, Silva et al. 2022, Figure 1c). Carbon-based nanoparticles such as soot are also emitted from wildfires and can interact with other atmospheric components such as sea spray and mineral dust forming micrometre-size NM-associations (Donado et al. 2023). These examples show that incidental NM can also occur in NM-associations, and that those play an important role in the biogeochemical cycling of elements and their impact on environment and health.

As indicated above, characterisations of NM-associations rather than solely NMs are necessary to understand their origin and long-term fate but are challenging to their differences in volume and degree of crystallinity relative to their macrometre-size counterparts. We will start this review paper with analytical techniques that characterise NM-associations at the bulk scale before addressing their characterisation at the micro- and nanometre scale.

Analytical techniques to study NM-associations at the bulk- and micrometre-scale

Formations of NM-associations in rocks, soils and tailings are a direct result of the bulk properties of their hosts namely their mineralogical and chemical composition, physical properties such as porosity and environmental conditions such as water-rock ratio, Eh, pH, T and P. Mineralogical and chemical compositions of environmental samples (e.g., soil, suspended and/or precipitated sediments, organic matter, ashes, rocks and PM) are extremely complex and their characterisation requires often multiple preparation methods, analytical techniques and sometimes even thermodynamic calculations on the potential occurrence of NM-associations in environmental samples that have been affected by pressure and temperature gradients (Silva et al. 2022).

In addition, as a number of structural and electronic properties of NMs and their interfaces towards their host are difficult to determine experimentally, molecular simulations can provide a better understanding of nanoscale phenomena at the molecular- and atomic-scale. These molecular simulations commonly include quantum-mechanical calculations, force-field calculations and molecular dynamics (MD) (Becker et al. 2010).

In terms of the analytical characterisation of NM-associations, one can distinguish between techniques that analyse bulk properties versus those that characterise properties as well as textural and spatial information at the micro- to nanometre scale (Figure 2).

Surface analytical techniques: XPS and ToF-SIMS

Information on the chemical compositions and valences of cations and anions on the surfaces of samples containing NM-associations can be gained with time-of-flight secondary ion mass spectrometry (ToF-SIMS) and X-ray photoelectron spectroscopy (XPS) (van der Heide 2014, Qian et al. 2015, Greczynski and Hultman 2020, Krishna and Philip 2022).

The development of new generations of XPS spectrometers with detectors of higher sensitivities, smaller apertures and charge-compensation systems allow measurements of well-resolved high-resolution XPS spectra from non-conductors samples. Consequently, XPS became one of the most effective techniques for characterising surfaces of NMs (Silva et al. 2013). Chemical and structural information gained with XPS include: (1) the chemical composition of the surface structure (Mantha et al. 2012); (2) structural information on the coordination environment of cations/anions (Schindler et al. 2009b, c); (3) valences of cations and anions on the surfaces (Schindler et al. 2009b, c) and (4) changes in composition with depth using depth profiling (Cerqueira et al. 2011, Mantha et al. 2012). The depth to which this information can be extracted is called information depth and depends on the inelastic mean free path of the respective photoelectrons. It lies for most of the elements in the upper 10 nm (Hochella 1988).

The determination of the chemical composition of a sample with XPS includes low atomic number elements such as N, O, and C (Leung et al. 1999, Haerle et al. 2002, Brydson et al. 2008) as well as the proportion of O, OH and H₂O groups (Schindler et al. 2009c). Structural information gained with XPS include for example an estimation of the polymerisation degree of silicate tetrahedra (important in the case of NM-amorphous silica association; Schindler et al. 2010) or the proportion of strongly versus weakly covalent bonds in NM containing for example hexavalent uranium (i.e., uranyl bonds versus the more ionic U-O equatorial bonds; Schindler et al. 2009b).

ToF-SIMS measures the mass of secondary ions emitted from a surface, which is bombarded by a pulsed primary ion beam of typically 25 keV (van der Heide 2014). The primary beam consists of a low flux ion (Bi, Cs or Ga), which results in the excitation of secondary ions from only the uppermost layers of the surface. The secondary ion beam consists of ionised molecular fragments, which are analysed with respect to their mass to charge ratio. In a time-of-flight (ToF) mass spectrometer, the separation is based on the fact that ions of different mass have different velocities (but the same kinetic energy) and thus different flight times towards the detector (Fearn 2015).

The ToF-SIMS is an important analytical tool for the quantification of chemical elements as well as their isotopes. It creates a complimentary dataset to XPS with respect to the occurrence of major and minor elements on the surface of NM-associations. As both analytical techniques can detect low atomic number elements such as C and as many carbon-bearing NMs have dimensions in the range similar to the depth of information for both instruments, ToF-SIMS and XPS provide valuable information on (a) the occurrence and composition of different carbon-allotropes (e.g., fullerenes) (Cerqueira et al. 2011,  2012, Oliveira et al. 2012) (Figure 3); (b) the chemical composition of nanometre-thick coatings on minerals (Scheidegger et al. 1993); (c) the valence of elements which are exclusively associated with NMs and are absence in their host materials (silica on Fe-hydroxides, (Vempati et al. 1990)).

Preparation of NM-associations for their study at the atomic- and nanometre-scale

The most-often used preparation techniques for the study of NM-associations at the atomic- to nanometre-scale are the dispersion of fine particles on a holder/grid and the more site-specific techniques ultramicrotomy (UM), ion milling (IM) and focused ion beam technique (FIB) (Figure 2) (Heaney et al. 2001, Wirth 2004, Ayache et al. 2010, Tizro et al. 2019).

The simplest and most straightforward preparation technique for the study of NM-associations in colloids or PM is their deposition on grids/holders. Here, solvents such as methanol or mixtures of hexane and acetone can be used to deposit colloids/particles without dissolving water-soluble phases (Donado et al. 2023).

Fundamental principles and advantages and disadvantages of the site-specific preparation techniques UM, IM and FIB have been reviewed in numerous books and papers (Table 1). Here, we briefly describe their functionality and discuss their capabilities to prepare NM-associations as well as their cost efficiency.

In UM, thin slices (30–100 nm) are produced through propagating a fracture induced by the edge of a diamond knife. A pyramidal-shaped sample of either a solid material or a powder embedded into a mould is brought in close contact to the knife and follows a fixed path (so-called D-shaped trajectory) allowing the knife to shave a thin slice at very low speed (Ayache et al. 2010). Advantages of UM are that many slices can be collected in a short period of time (high-cost efficiency) and that a site-specific location of a NM-association can be mounted in the desired orientation within the mould. The major disadvantage is that tearing or crushing of components in mineralogical samples (Figure 1b) during shaving does not allow the preservation of textural and spatial relations in NM-associations, especially those where a soft and porous NM assemblage is associated with a physically hard and homogeneous mineral (e.g., a ferrihydrite coating on a pyrite grain or an amorphous silica coating on a feldspar). Less severe damage is commonly observed when shaving soft organic matter or microbes in association with NMs such as a bacteria-Fe-Ox assemblage in a mine tailings sample.

In the IM technique, a 5–30 μm-thick sample is thinned to electron transparency through abrading the sample with an Ar⁺ ion beam (ionic abrasion and ionic cleaning) (Ayache et al. 2010, Li 2012). The most common procedure for NM-associations includes the mechanically pre-thinning of a mineral fixed to a 3 mm diameter disk (grinding, polishing, dimpling) followed by its ion-mill thinning at various angles of incidence. Ion-mill thinning gradually proceeds from high beam energies at higher incident angle (~ 10°) to lower beam energies at lower incident angles (~ < 3°). The thinning procedure is monitored with an optical microscope or a CCD camera and lasts for several hours. The advantages of this method are that it allows the thinning of specific micrometre-size NM-associations that have been removed from an epoxy-embedded sample with a diamond knife. This micrometre-size sample can be fixated on a Cu-holder and can be subsequently thinned along a feathered loading edge extending laterally over 1000s of micrometres. The disadvantages of the technique are that I. the broad Ar⁺ beam (1–2 mm in diameter) does not allow the site-specific thinning of only a small, isolated area of interest within a NM-association; II. The thinning of mineralogical-heterogeneous samples containing materials of greatly different mill rates will result in a high surface roughness (e.g., nano- to micrometre-size uraninite particles embedded within organic matter). In comparison to the UM method, the IM method is less cost-efficient (normally one sample per day) and is less suitable for the preparation of NM-microbe associations. The advantage of this technique with respect to UM is the better preservation of spatial and textural relations in mineralogically heterogeneous NM-associations.

The FIB technology is the most sophisticated preparation technique for site-specific TEM samples (Heaney et al. 2001, Wirth 2004). Here, a thin lamella of constant thickness is extracted at a precise area of the sample. In the first step, a metal- or carbon-layer is deposited using a gas injection system (GIS) in order to protect the surface area to be extracted (Figure 4a, b). A focused ion beam (commonly Ga⁺ ions) is used to machine two parallel trenches adjacent to the lamella and to cut oft the lamella from the sample prior to its extraction with an internal micromanipulator. Mounting of the lamella on a sample holder and its subsequent thinning completes the preparation of, for example, a TEM sample (Figure 4b). Each step of the preparation process can be monitored with a SEM using a so called a dual-beam system (i.e., focused ion beam and electron beam). In addition to the FIB, SEM and GIS system (Figure 4a), instruments such as the FEI DualBeam™ Helios 600 Nanolab™ are also equipped with a high-resolution field emission gun (FEG); multiple electron detectors for image acquisition, such as through-the-lens detector (TLD), an Everhart-Thornley (ETD), a backscattered electron (BSED) and a EDS detector (Oliveira et al. 2018). Artefacts such as the implantation of Ga⁺ ions, amorphisation, redeposition of material (metals such as native gold and silver are prone to redeposition) or heating of temperature-sensitive minerals can be minimised through tuning down the current of the focused ion beam from several nA during trenching and extraction to several pico-ampere during polishing. Additional cleaning of the extracted lamella with an advanced plasma system can further minimise the amount of implanted Ga⁺ ions or redeposited material (Ribeiro et al. 2010).

FIB lamellae for example TEM can be extracted from almost every interface observed in NM-associations such as NM−bacteria (Obst et al. 2005), NM−mineral−fungal biofilms (Bonneville et al. 2016, Gerrits et al. 2021), NM−mineral (Hellmann et al. 2012), and NM−organic matter (Mantha et al. 2019). However, the successful extraction of a lamella for TEM depends on the occurrence of fractures in the underlying material (not visible when selecting a site for FIB extraction), which often occur along interfaces between NMs and minerals or organic matter. The number of fractures may be minimised through embedding the material in a low viscous resin, but their occurrence cannot be completely avoided as fractures (a) are not always accessible by the resin and (b) form after the exposure of the sample to vacuum or the generation of heat during FIB-trenching.

The advantages of the FIB technology versus UM and IM are its high site-specificity and the preservation of almost all mineralogical, textural and spatial relations of the components in the selected NM-associations. The major disadvantages are (a) the very low cost-efficiency (> 800 $US per sample), (b) the technology is not available at every academic/research institute and (c) that the extraction of a FIB lamella requires highly skilled personal.

Analytical techniques to study nano-scale features

Analytical techniques to study nano-scale features commonly involves TEM but also recently developed techniques such as atom probe tomography (APT), scanning transmission X-ray microscopy (STXM), nano-spectroscopic techniques based on FTIR (synchrotron infrared nano spectroscopy; SINS), Raman (tip-enhanced Raman spectroscopy; TERS), X-ray fluorescence (nano-XRF) and X-ray absorption spectroscopy (nano-XAS) and ptychographic X-ray computer tomography (PXCT). The principles of some of these techniques are shown as schematics in Figure 4c–g.

Atom probe tomography

Atom probe tomography allows the three-dimensional compositional mapping of elements and isotopes in NM-associations (Gault et al. 2012, Lefebvre et al. 2016, Saxey et al. 2018, Reddy et al. 2020). In a process called field evaporation, either ultra-fast voltage or laser pulses are used to erode and convert the atoms on the tip of a needle-shaped specimen into charged ions (Figure 4d). An electric field accelerates the ions towards a position-sensitive detector that registers the time of flight and impact position of individual ion or species (Gault et al. 2012, Lefebvre et al. 2016, Saxey et al. 2018, Reddy et al. 2020). The energy applied to the surface of a specimen is used to calculate the mass-to-charge ratios of the ions based on their travel time, which is almost always sufficient for the identification of an individual atom. Impact positions on the position-sensitive-detector plot the locations of individual atoms (Figure 4d). Locations and the identity of individual atoms can be used to calculate a 3D plot which provide information on the location of NMs in minerals and organic matter (Alavijeh et al. 2023) (Figure 5). Samples for APT are commonly extracted with the FIB technology, where similar to the preparation of a TEM lamella, the selected site is marked and protected with deposited Pt, and then extracted and prepared with a Ga⁺ ions and a micromanipulator (Gu et al. 2020).

Synchrotron-based scanning transmission X-ray microscopy and ptychographic X-ray computer tomography

Scanning transmission x-ray microscopy is performed at synchrotron light sources, which produce intense and tuneable X-ray beams (Henderson et al. 2014). STXM images from X-ray transparent samples are generated with a highly focused X-ray beam with spot sizes as small as 15 nm in size (Figure 4e). The images are obtained at a given photon energy via a raster scan of the sample through a focal point. Here, a sequence of images is collected around the X-ray absorption near edge fine structure (XANES) of an element of interest as a function of photon energy. This sequence of images is commonly referred to as a “stack”. STXM provides data on the valence, chemical and mineralogical composition of NM-associations and similar to EELS, on the occurrence and distribution of organic functional groups across an X-ray image (Rothe et al. 2000, Plaschke et al. 2002). Consequently, STXM is a powerful tool to characterise NM-associations with organic matter and carbon-bearing PM as it allows to visualise (a) the occurrence of inorganic and organic carbon-components and (b) to chemically characterise the distribution of functional groups along the interface of a mineral surface and organic matter (see below). STXM measurements are ideally taken from FIB lamellae or PM and colloids deposited on TEM grids (Moffet et al. 2011, Gu et al. 2020).

Similar to micro-CT tomography, PXCT provides high penetrability and non-destructive 3D visualisation of constituents in NM-associations. A further development of STXM, PXCT is a coherent diffraction imaging technique with the highest spatial resolution currently available (Pfeiffer 2018) (Figure 4f). The technique combines STXM with diffraction imaging and subsequent reconstruction of the images (Dierolf et al. 2010, Pfeiffer 2018). It provides images of extended sample areas with single digit nanometre resolution (with respect to resolved features) and, thus, increases the resolution of a STXM by about one order of magnitude. Although the spatial resolution is ideally only limited by the diffraction of X-rays (i.e., wavelength-limited) rather than by the configuration of the beamline (i.e., X-ray optics), mechanical vibrations of the beamline system such as piezoelectric motors limit the resolution of PXCT to a couple of nanometres.

The basic principle of PXCT is to measure multiple diffraction patterns by scanning a finite X-ray beam over an extended specimen. Illumination of the object multiple times generates an overlap between adjacent illumination positions, which provides subsequently a sufficient over-determination of information on the areas of the object. This over-determination in combination with reconstruction algorithms leads to the generation of electron density plots in 3D space and thus in the visualisation of 3D features in biological and environmental samples (Pfeiffer 2018). Beamlines with PXCT abilities occur for example at the synchrotron facilities Sirius (Brazil; Caterete beamline (Meneau et al. 2021), and Carnauba beamline (Tolentino et al. 2023) Max IV (Lund, Sweden, (Johansson et al. 2021)), ESRF (Grenoble, France, (Martínez-Criado et al. 2016)), the Swiss Light Source at the Paul Scherrer Institute (cSAXS beamline, https://www.psi.ch/en/sls/csaxs) and the ALS at the Lawrence Berkeley National Laboratory (beamline 11.0.2, https://als.lbl.gov/beamlines/11-0-2/).

Advantages of PXCT versus the combination of FIB-TEM-APT are that no sample preparation is needed, and the samples can be characterised without any destruction (note that the volume of the area of interest is limited). Examples of NM-associations characterised with PXCT include for example swelling clay minerals in sandstone (De Boever et al. 2015), magnetite, kerogen and silica associations in micro-fossils (Figure 6, (Maldanis et al. 2020)), magnetosomes in bacteria (Zhu et al. 2016) or nano-size Mn-oxides clusters in rock varnish (Förster et al. 2021).

Challenges and perspectives

The most common challenge when studying NM-associations is that the data obtained at the nanoscale may not be representative for a given environment or medium such as soil, rock, tailings, particulate matter or colloids in a river, lake or sediment. The lack of representative data obtained at the nanoscale is due to the high costs involved in the preparation of FIB lamellae or hours of TEM or APT time to fully characterise a sample. This problem may be solved though creating representative data sets at the bulk and micrometre scale. The datasets obtained at these scales should show for example similar chemical and mineralogical compositions and textural features of NM-associations. A possible way to approach such a multi-scale approach would be a pyramidal analytical approach, where the number of analyses correlates with the width of a pyramid at a given height. In this approach, the analyses at the bulk scale would represent the base of the pyramid, those at the micrometre scale would be at half-height and those at the nanometre-scale at the tip of the pyramid. Considering the time effort to collect and analyse samples at each scale, multi-scale analyses are often conducted over several years and should be carefully planned with respect to a student thesis.

The data collection at the micro- and nanometre-scale can be accelerated using automatism software during SEM and TEM analyses. These are of great importance when examining unconsolidated environmental samples (sediments, soils and tailings material) at the micrometre scale with SEM and colloids and PM with TEM at the nanometre-scale. These analyses are often time consuming and even after many hours of analysis, the obtained data may still be not representative. In recent years, the QEMSCAN (quantitative evaluation of minerals by scanning electron microscopy) system has been used as a tool to search for features at the nanoscale. The QEMSCAN system was developed within the Commonwealth Scientific and Industrial Research Organization (CSIRO) of Australia as a counterpart to the QEM-SEM system. The latter system was based on an automated SEM fitted with four energy dispersive X-ray spectrometers (EDS), which was not capable of quantifying low atomic number elements such as carbon or oxygen in, for example, coal-bearing NM-associations (Ward 2016). Initial research in the characterisation of coal-bearing NM-associations with QEMSCAN commenced in the late 1980s and early 1990s when a proof-of-concept study was initiated, which experimented with various mounting media and automated SEM measurements. During 1997–98, CSIRO's first LEO QEMSCANTM became operational as a fully automated instrument, running a Windows-based PC operating system (Liu et al. 2005). This system has now enhanced low atomic number element capability (carbon and oxygen) due to the use of thin polymer windows and has opened new possibilities for characterising not only minerals in coal but the coal itself (i.e., the macerals), thus providing for the first time, information on the textural setting of minerals in coal-bearing NM-associations.

Conclusions

The study of NM-associations remains an analytical challenge due to their mineralogical and chemical complexities and differences in the physical properties of NMs and associated mineral grains, organic matter, microbes, fungi and plants. Analytical techniques such as the focused ion beam technology, new generations of TEM, automatism software for SEM and TEM analyses and combinations of APT with STXM and PXCT may provide new insights into the structure and properties of NM-associations. The use of the latter technologies offers also new perspectives such as the identification of precursors of NMs in ore deposits and contaminated soils, air and aquatic systems, a better understanding of the role of NM-associations in stabilising carbon in soils and in the transformation of carbon-bearing PM during its interaction with other atmospheric contaminants.