Advanced analytical techniques for the study of themorphology and chemistry of Proterozoic microfossilsWacey, D., Battison, L., Garwood, . R. J., Hickman-Lewis, K., & Brasier, M. D. (2017). Advanced analyticaltechniques for the study of the morphology and chemistry of Proterozoic microfossils. Geological SocietySpecial Publication, 448, 81-104. DOI: 10.1144/SP448.4

Published in:Geological Society Special Publication

DOI:10.1144/SP448.4

Document VersionPeer reviewed version

Link to publication in the UWA Research Repository

General rightsCopyright owners retain the copyright for their material stored in the UWA Research Repository. The University grants no end-userrights beyond those which are provided by the Australian Copyright Act 1968. Users may make use of the material in the Repositoryproviding due attribution is given and the use is in accordance with the Copyright Act 1968.

Take down policyIf you believe this document infringes copyright, raise a complaint by contacting repository-lib@uwa.edu.au. The document will beimmediately withdrawn from public access while the complaint is being investigated.

Download date: 20. Jul. 2018

12345678910111213141516171819202122232425262728293031323334353637383940414243444546474849505152535455565758

Advanced analytical techniques for studying the morphology

and chemistry of Proterozoic microfossilsQ1

DAVID WACEY1,2*, LEILA BATTISON3, RUSSELL J. GARWOOD4,

KEYRON HICKMAN-LEWISQ2 3,5,6 & MARTIN D. BRASIER3†

1Centre for Microscopy Characterisation and Analysis, and Australian Research

Council Centre of Excellence for Core to Crust Fluid Systems, The University of

Western Australia, 35 Stirling Highway, Perth, WA 6009, AustraliaQ32School of Earth Sciences, University of Bristol, Life Sciences Building,

24 Tyndall Avenue, Bristol BS8 1TQ, UK3Department of Earth Sciences, University of Oxford, South Parks Road,

Oxford OX1 3AN, UK4School of Earth, Atmospheric and Environmental Sciences, University of

Manchester, Manchester M13 9PL, UK5St Edmund Hall, Queens Lane, Oxford OX1 4AR, UK

6Present address: Homestead, 19 Sunnybank Road, Blackwood, Gwent, NP12 1HT, UK

*Correspondence: David.Wacey@uwa.edu.au

Abstract: This paper outlines the suite of advanced multi-scalar techniques currently availablein the toolkit of the modern Proterozoic palaeobiologist. These include non-intrusive and non-destructive optical, laser and X-ray techniques, plus more destructive ion beam and electronbeam methods. Together, these provide morphological, mineralogical and biochemical data atflexible spatial scales from that of an individual atom to the largest Proterozoic microfossils. Anoverview is given of each technique and a case study from the exceptionally well-preservedTorridonian biota of NW Scotland is presented. This microfossil assemblage was first recognizedover a century ago, but its great diversity and evolutionary importance has only recently come tolight, due in no small part to the research efforts of Martin Brasier.

Modern palaeobiology primarily exists to dis-cover, describe and decode the ancient biosphere,and to understand the course of global evolutionarychange. Stemming from its roots in Victorian natu-ral history, palaeobiology has made good use oftechnological advances to shed light on new dis-coveries (see Sutton et al. 2014; Wacey 2014 andreferences cited therein) and to reveal previouslyunimagined details in historical material (Brasieret al. 2015). As with any modern field of science,palaeobiological research must continually look for-wards to the next potential discovery, utilizing allthe available tools and techniques.

Historically, major discoveries have predomi-nantly dated from the Phanerozoic as a result ofthe relatively well-preserved and easily recoverable

fossils of the macroscopic organisms alive duringthis time. In the search for life’s origins and earlyrecord, attention has inevitably turned to the morepoorly understood Proterozoic and Archaean fossilrecords. The evolutionary history of these expansesof time is much less well established because thereis a shortage of exposed rock of the appropriateage, a relative paucity of fossil material and limita-tions in extracting the relevant information. Fossilsfrom these times are typically microscopic, enig-matic and poorly preserved, although a numberof exceptionally preserved deposits have come tocharacterize the Proterozoic fossil record (e.g. theTorridonian biota, Strother et al. 2011; the Doush-antuo biota, Yin & Li 1978). In both ‘traditional’and ‘exceptional’ examples of preservation, our

†Deceased 16 December 2014.

From: Brasier, A. T., McIlroy, D. & McLoughlin, N. (eds) Earth System Evolution and Early Life:a Celebration of the Work of Martin Brasier. Geological Society, London, Special Publications, 448,http://doi.org/10.1144/SP448.4# 2016 The Author(s). Published by The Geological Society of London. All rights reserved.For permissions: http://www.geolsoc.org.uk/permissions. Publishing disclaimer: www.geolsoc.org.uk/pub_ethics

5960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116

understanding is still limited by the observationaland analytical techniques used to characterize theseimportant specimens.

The approaches traditionally used to studyearly fossil material are essentially borrowed andadapted from the methods used in the study of Palae-ozoic fossils and are best suited to hard-bodiedmacroscopic fossils or compressed organic materialextracted by acid maceration. However, as ourunderstanding of Precambrian environments hasfundamentally improved, it has become clear thatentirely different preservational styles are possible,some of which require novel analytical approaches.Although many Proterozoic carbonaceous fossilscan still be found compressed within shales (Javauxet al. 2004; Agic et al. 2015) and can be extractedfor study by palynological acid maceration tech-niques, microfossil material can also be hosted ina variety of other media, including chert (e.g. Bar-ghoorn & Tyler 1965), pyrite (e.g. Rasmussen2000), authigenic aluminosilicates (e.g. Waceyet al. 2014) and cryptocrystalline phosphate (e.g.Strother et al. 2011). These alternative preservatio-nal styles originate from the biogeochemical con-ditions that prevailed in specific environmentsor across specific periods of time. They are ableto exceptionally preserve microfossils of a widerange of affinities in their original spatial context,often in three dimensions, reflecting a broad spec-trum of taphonomic decay. In these cellular Lager-statte, challenges are posed by the small scale,enigmatic nature and relative scarcity of Proterozoicfossils, as well as by their complex taphonomic andmetamorphic histories. Thus a thorough understand-ing of Proterozoic and Archaean life necessarilycalls for state of the art, high spatial resolution andholistic imaging and analysis techniques.

An increasing number of researchers arenow making use of such techniques to study bothProterozoic and Archaean material, revealingunprecedented levels of detail and allowing thereconstruction of the complex Precambrian bio-sphere. It is still common, however, for these differ-ent approaches to be attempted separately, often bydifferent individual research groups, which can par-tially preclude the synthesis of information and anoverall understanding of local, regional or evenglobal palaeoecologies. Here we present a holisticmethodology for studying Proterozoic fossil depos-its with a consideration of their unique preservatio-nal styles and histories. A set of complementarymicroanalysis techniques has already been pre-sented with respect to Archaean material (Wacey2014). However, with the expansion of the bio-sphere (Knoll 1994), the evolution of eukaryoticcells (Knoll et al. 2006) and the advent of variousmetabolic pathways and trophic tiering (Knoll2015), the Proterozoic fossil record is more complex

and – as a result of its younger age (c. 2500–540 Ma) – arguably better preserved. Thus a greaterpotential wealth of information might be gleanedfrom such deposits, necessitating their study on avariety of spatial scales, as well as assessing boththeir morphology and chemistry.

The following sections detail, in a logical orderfor practical investigation, multiple approaches toexamining a Proterozoic microfossil assemblage,including the following: ‘traditional’ field studyand optical microscopy; X-ray based techniques,including X-ray computed tomography (CT) andX-ray spectroscopy; laser-based techniques, includ-ing Raman spectroscopy and confocal laser scan-ning microscopy (CLSM); infrared spectroscopy;electron-based techniques, including scanning elec-tron microscopy (SEM) and transmission elec-tron microscopy (TEM); and ion-based techniques,including focused ion beam (FIB) milling andsecondary ion mass spectrometry (SIMS). A combi-nation of several of these techniques when investi-gating a single fossil deposit provides the bestopportunity to fully reveal the palaeocology of theProterozoic biosphere. An example of their applica-tion to the microfossiliferous rocks of the 1200–1000 Ma Torridonian Supergroup of NW Scotlandis presented as a demonstrative case study.

Standard palaeobiological techniques Q4

Field study and optical microscopy

A crucial starting point for any palaeobiologicalinvestigation remains a comprehensive field studyand the preparation of candidate material for opticalmicroscopy. As a preliminary investigation, this canprovide an important palaeoenvironmental contextand enable the quantification of the richness, mor-phology and spatial distribution of fossils, plus thedepositional setting and taphonomic history of thefossil deposit.

A detailed sedimentological and stratigraphicstudy should initially be made of the fossiliferousrocks, and the rocks associated with them, to allowaccurate palaeoenvironmental, metamorphic andtectonic interpretations. Such a study will provideregional, local and fine-scale information pertainingto the location, type and energy of the environmentof deposition, as well as any subsequent chemical orstructural changes that may have taken place sincelithification. Fine-scale field observations will alsoallow the identification of candidate fossiliferousmaterial. This may be related to specific preservatio-nal mineralogies, such as cherts (e.g. the c. 1900 MaGunflint Formation; Barghoorn & Tyler 1965) orphosphates (e.g. the c. 600 Ma Doushantuo Forma-tion; She et al. 2013) or be found in association withmacroscopic fossil structures, including siliceous

D. WACEY ET AL.

117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174

and phosphatic stromatolites (e.g. the c. 1900 MaBelcher Supergroup; Hofmann 1976) and micro-bially induced sedimentary structures (e.g. the c.1000 Ma Diabaig Formation; Callow et al. 2011).The collection and documentation of candidatematerial should be methodical and include globalpositioning system localities, orientation data andspecific relationships with larger scale structures.

Polished, uncovered (which are more usefulthan covered for subsequent techniques) petro-graphic thin sections can be prepared from thecollected samples for analysis using optical micro-scopy. Ideally, thin sections should be preparedboth perpendicular and parallel to the bedding direc-tion to capture the full spatial distribution of micro-scopic fossils. Although sections 30 mm thick arerequired for mineral identification using cross-polarized light, the detection of fossil material maybe facilitated by the use of sections up to c. 150 mmthick, provided the encasing medium is sufficientlylight-coloured and free of dark impurities. Thisincreases the chances of capturing entire cellularmaterial and the in situ relationships between differ-ent fossil taxa.

The primary purpose of optical microscopy isto locate and identify fossil material and to docu-ment its spatial distribution and relationship withnon-biological minerals. For the majority of Prote-rozoic carbonaceous fossil deposits, examinationand imaging at all magnifications up to 1000× areneeded to provide a complete context. This canallow the observation of fine structural details upto c. 0.2 mm across, but note that oil immersion isrequired at the highest resolutions to increase clar-ity, which may be detrimental to some subsequenttechniques. The position of fossil material can beidentified and recorded for future reference usingstandard graticules, When fossil material is pre-served with some degree of three dimensionality,focusing through the thickness of the slide canreveal its shape, organization and extent. A rangeof different photomicrography suites is now avail-able for capturing images of such samples (e.g.Synchroscopy Auto-Montage, as demonstrated byBrasier et al. 2005). Many packages contain algo-rithms for stacking focused images from differentdepths within a section to produce a single, focusedimage, or for stitching together images of adjacentfields of view to produce a high-resolution ‘map’of a thin section.

Using a variety of optical micrographic tools,the preliminary identification and quantification offossil material may be carried out, larger scalespatial relationships determined and candidate fos-sils selected for further analysis. This work is vitalfor the initial study of a fossil deposit, but the intrin-sic limitations of this approach preclude its usefor finer scale analyses. Certain media may be

unsuitable for investigation by optical microscopy.Dark-coloured material, or enclosing media withmany impurities, for example, may mask fossildetails and reduce their visibility, especially throughthick sections. Larger microfossils may be cross-cutby the sectioning process, limiting interpretation.Another limitation is that the identification of thechemical constituents of samples is limited to thatwhich can be determined by standard petrographicmethods and may not be sufficient for fine-grainedor finely crystalline material. As carbonaceous fos-sils are often dark coloured, optical analyses willonly be able to resolve their surface shape andstructure, with the fossils themselves masking anyunderlying ultrastructure or interior features. Thusmore versatile high spatial resolution techniquesare required for a better understanding of both fossilmaterial and its preservational medium.

Non-destructive moderate to high spatial

resolution techniques

Non-destructive techniques are classified here astechniques that can be applied to a standard geolog-ical thin section, rock chip or rock hand sample withminimal sample preparation and that do not con-sume or alter the specimen of interest during theanalysis. Hence they can be applied to type speci-mens (including holotypes on loan from museums)and can be utilized as a precursor to more destruc-tive techniques on newly discovered material.

X-ray computed tomography. X-ray CT maps theX-ray attenuation within a rotating sample. Dataare captured as a series of projections that can bereconstructed as two-dimensional (2D) slices andthree-dimensional (3D) visualizations (see Kak &Slaney 2001; Cnudde & Boone 2013 for overviews).The X-ray attenuation is dictated by factors such asthe elemental composition and density, hence X-rayCT can often detect variations in the style of fossilpreservation and mineralization, as well as buildingup 3D models of entire specimens (Conroy & Van-nier 1984; Haubitz et al. 1988; Sutton et al. 2001).The high-resolution form of X-ray CT used for fos-sils is known as X-ray microtomography (mCT) andhas been utilized in palaeobiology for almost twodecades (Rowe et al. 2001; Sutton 2008). It is rou-tinely applied to Phanerozoic vertebrate and inverte-brate fossils, ranging from echinoderms (Rahman &Zamora 2009) to dinosaurs (Brasier et al. 2016) andfrom plants (Spencer et al. 2013) to arthropods(Garwood & Sutton 2010). The study of microfos-sils using CT has become viable in recent yearswith the use of synchrotron-based systems inwhich more intense, monochromatic X-rays resultin improved contrast and greater spatial resolution(Donoghue et al. 2006; Huldtgren et al. 2011).

ANALYTICAL TECHNIQUES FOR PROTEROZOIC MICROFOSSILS

175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232

Recent years have also seen improvements in thespatial resolution of laboratory-based mCT andnano-CT systems where sub-micrometre resolutionsare now possible (Hagadorn et al. 2006; Schiffbaueret al. 2012; Sutton et al. 2014).

Despite these technological advances, configur-ing the correct instrumental parameters for mCTscanning a given microfossil specimen is challeng-ing and some specimens will not be suited to mCTtechniques due to a lack of X-ray attenuation con-trast between the specimen and the matrix and/orthe presence of X-ray opaque minerals. In general,mCT is applied to small rock chips. It is not suitedto geological thin sections because of their highlyanisotropic nature, although thin sections can becut down to a more isotropic shape if allowed bythe owner, or the fossils can be liberated using amicro-corer. Elsewhere in this volume, Hickman-Lewis et al. (2016) report several case studies ofthe mCT scanning of Precambrian microfossil-bearing rocks using two laboratory-based CT scan-ners with spatial resolutions (minimum voxel sizes)of about 5 and 0.5 mm, respectively. They showthat mCT can be a valuable tool to decode the 3Dpetrographic context of such biological material –for example, by highlighting potential organicgrains and laminations, fractures within the matrix,assemblages of detrital heavy minerals and thereplacement of silica by carbonate rhombs (whichare known to reduce the quality of microfossil pres-ervation). Detecting individual microfossils usinglaboratory-based CT remains challenging unlessthe preservation window is particularly favourable(e.g. pyritized microfossils in a silica matrix; seeHickman-Lewis et al. 2016). The use of a synchro-tron-based CT (or laboratory-based nano-CT) sys-tem can improve results by providing more intenseX-rays and improved spatial resolution, but thisrequires more specialist sample preparation (e.g.micro-coring) to obtain sub-millimetre pieces offossiliferous rock, meaning that it can no longer berealistically classified as a non-destructive tech-nique and can seldom be applied to holotypematerial.

X-ray spectroscopy. A logical extension to examin-ing the morphology of microfossils using X-raymicrotomography is to investigate their chemistryusing X-ray spectroscopy. A range of X-ray tech-niques is available to characterize fossiliferousrocks, most performed on a synchrotron beamline(for overviews, see Fenter et al. 2002; Templeton& Knowles 2009) and utilizing both hard X-rays(more penetrating with wavelengths of 1–20 A andphoton energies .5–10 keV) and soft X-rays (lesspenetrating with wavelengths of 20–200 A andphoton energies ,5 keV). X-ray fluorescence map-ping provides semi-quantitative element-specific

maps over flexible spatial scales (micrometres tomillimetres, e.g. Edwards et al. 2014). Near-edgeX-ray absorption fine structure and X-ray absorptionnear-edge structure spectrometry are techniquesthat use soft (low-energy) and hard (high-energy)X-rays, respectively, to excite the core electrons inan element (Templeton & Knowles 2009). Theresulting spectra provide information on both thecoordination chemistry and valence of the elementof interest. Scanning transmission X-ray micros-copy uses soft X-rays to obtain both spectral dataand images of these spectral data (e.g. maps of thespatial distribution of specific elements, valencestates or functional groups) at the nanometre scale,created by rastering samples through an X-raybeam at stepwise-increasing incident X-ray energiesto cover the absorption edges of the elementsof interest (e.g. Lawrence et al. 2003). Althoughthese types of analyses do not destroy the specimen,specialist sample preparation (e.g. micro-cored rockchips; doubly polished thin sections no more thanc. 100 mm thick) means that permission for holotypespecimens to be analysed in this way is unlikely tobe granted. Beam damage can also affect subse-quent chemical analyses.

In terms of Proterozoic microfossils, much ofthe interest in X-ray spectroscopy surrounds thechemical bonding of carbon. The energy resolutionof X-ray absorption fine structure/X-ray absorptionnear-edge structure is excellent (c. 0.1 eV), soclosely spaced peaks can be resolved. Hence carbonbound in aromatic groups, aliphatic groups, ketones,peptides, carbonyls, carboxyls and carbonate can allbe distinguished from one another (Bernard et al.2007). Such spectra may help to characterize cellu-lar v. extracellular organic components and theinterfering signals from carbonate minerals can besubtracted. De Gregorio et al. (2009) applied thismethodology to powders of organic material fromthe 1878 Ma Gunflint Formation and showed thatpolyaromatic carbon, carboxyl and phenol groupshad all been preserved in this ancient kerogen. Sim-ilarly, the bonding characteristics of other elementscommon in organic material (e.g. S, N, P, O) mayhelp to determine whether they are present as orga-nic or inorganic forms in ancient fossiliferous rocks.For example, Lemelle et al. (2008) used X-ray fluo-rescence to quantify the amounts of sulphur withinthe cell walls of coccoid microfossils from the c.750 Ma Draken Formation, Svalbard before usingX-ray absorption near-edge structure techniques todetermine the speciation of sulphur. They showedthat the sulphur was a reduced organic form and wasmost likely present as a thiophene-like compound.

Confocal laser scanning microscopy. The tech-nique of CLSM provides high spatial resolutionmorphological data (,100 nm is possible) allowing

D. WACEY ET AL.

233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290

the visualization of microfossils in three dimensions(for overviews, see Halbhuber & Konig 2003;Sutton et al. 2014). Under ideal conditions data col-lection from standard polished or unpolished geo-logical thin sections is rapid and CLSM is able toresolve tiny morphological features that may beunclear or hidden when viewed under light micro-scopy, as well as giving a true 3D perspective tothe distribution of microfossils (Schopf et al.2006; Cavalazzi et al. 2011). However, natural sam-ples are rarely ideal for the application of this tech-nique. CLSM relies on the fact that organic materialauto-fluoresces when excited by a laser of a specificwavelength. The system can accurately focus andscan at different depths within a microfossil speci-men and can exclude fluorescence outside theplane of focus; 3D images are then built up combin-ing the data acquired from successive planes offocus (see Amos & White 2003). Hence anythingthat interferes with the transmission or detectionof this signal severely degrades the quality of thefinal images obtained. For example, specimens situ-ated a long way below the surface of a thin section orwith thick opaque walls will not provide sharpCLSM images. Similarly, a specimen surroundedby plentiful fluorescing organic detritus, or onethat is embedded in a mineral that internally reflectsthe fluorescence signal, may be problematic. Thematurity of the organic material also affects thequality of the data, with the auto-fluorescence signaldissipating as the organic material becomes moregeochemically mature and loses more of its hetero-atoms (i.e. evolves towards the structure of graph-ite). Hence CLSM is of greatest use when appliedto thin-walled organic microfossils preserved insilica (and, to a lesser extent, phosphate) and housedin rocks of low metamorphic grade. In these casessignificant insights into the 3D morphology andtaphonomic preservation of Proterozoic micro-fossils may be obtained. For example, in the Neo-proterozoic Buxa Formation, CLSM was able todemonstrate the 3D organization of groups of fila-mentous microfossils (Schopf et al. 2008). In the850 Ma Bitter Springs Formation and the 650 MaChichkan Formation, notches, tears, grooves andsurface ornamentation were all detected in micro-fossils using CLSM (Schopf et al. 2006), whereasin the c. 580 Ma Doushantuo Formation CLSMrevealed parts of fibrous tissues and cell walls withinfossil alga that were not visible by any other means(Chi et al. 2006).

Laser Raman microspectroscopy and imagery.Raman spectroscopy is a versatile, non-intrusiveand non-destructive in situ technique. It can beused to identify the mineralogy of microfossils andtheir host rocks and is particularly sensitive to themolecular structure and geochemical maturity of

carbonaceous phases such as kerogen – the primeconstituent of organic-walled microfossils (fordetails, see Beyssac et al. 2002; Fries & Steele2011). In addition, when used in the confocal imag-ing mode, Raman spectroscopy can provide 2D and3D chemical and structural maps of microfossils atmoderate spatial resolution (potentially ,1 mm).Raman spectroscopy can be applied to rock chipsand standard uncovered geological thin sections.Data are acquired via laser excitation of the chemi-cal bonds within the sample. This excitation pro-duces characteristic spectra depending on theminerals and compounds present. Maps can be con-structed of the spatial distribution of various spectralparameters, including the intensity of a given peak(also sometimes referred to as a band) or the ratiosof two given peaks.

For the field of Proterozoic palaeobiology, thepeaks of interest are often associated with car-bon. In perfectly crystalline graphite, a single first-order peak occurs at 1582 cm21, attributed tostretching of the C–C bonds in basal graphite planes(known as the G or graphite peak) (Jehlicka et al.2003). Second-order peaks occur at c. 2695 and2735 cm21. Imperfectly crystallized graphitic car-bons, including kerogens, have additional peaksat c. 1355 cm21 (known as the D1 or disorderedpeak) and c. 1620 cm21 (D2, occurring as a shoulderto the G peak) and a single broad second-order peakat c. 2700 cm21. The specific position, width andrelative intensities of these peaks vary dependingon the degree of ordering of the carbon and theseparameters have been characterized in carbon ofvarying metamorphic grade in an attempt to useRaman spectroscopy as an indicator of the antiquityof carbon in ancient rocks (Tice et al. 2004). Thisis by no means an exact science because the start-ing composition of organic material in differentmetamorphic terrains, both geographically andtemporally, may differ. Putative carbonaceousmicrofossils should, however, exhibit very similarRaman spectral features to other carbonaceousmaterial in the same rock specimen because bothshould have undergone the same maturation pro-cesses. Raman spectra cannot be used to unequivo-cally determine the biogenicity of an ancientcarbonaceous object because similar spectra tothose of biogenic kerogens are seen in laboratory-synthesized abiological disordered carbonaceousmaterial (Pasteris & Wopenka 2003). However,the co-occurrence of a kerogenous compositionwith features that optically resemble cellular mate-rial provides promising preliminary data regardingbiogenicity that can be further tested using tech-niques with a higher spatial resolution.

As with CLSM, the highest quality data areobtained from specimens close to the surface of athin section and it has been suggested that for viable

ANALYTICAL TECHNIQUES FOR PROTEROZOIC MICROFOSSILS

291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348

3D maps of kerogen to be produced, the entire speci-men of interest should be no more than 6–8 mmbelow the surface (Marshall & Olcott Marshall2013). The best data will come from specimenslying under translucent minerals such as quartzc. 1–5 mm below the surface of a thin section;microfossils associated with phases that fluorescestrongly under the laser excitation beam may notprovide usable spectra. Care must also be takennot to confuse the carbon signature of interest withthat produced by (1) the polymer used to attachthe thin section to the glass slide, (2) any coatingthat may have been applied to the section duringprevious analyses and (3) overlapping peaks in thevicinity of carbon peaks – of particular note hereis the 1320 cm21 peak of hematite (Marshall et al.2011). The carbon spectrum can also be artificiallymodified by using too high a laser power or by ana-lysing right at the surface of a thin section that hasbeen polished (Fries & Steele 2011). Both of theseconditions should always be avoided. Raman spec-troscopy can also be used to elucidate some struc-tural information from the minerals that hostputative microfossils. Several minerals produceRaman spectral peaks that vary in intensity depend-ing on their crystallographic orientation relative tothe incoming laser. This feature can be used, forexample, to image the distribution of the crystallo-graphic axes of quartz to see whether putativemicrofossil material occurs between grain boundar-ies, is enclosed by entire grains or occurs in cracks(Fries & Steele 2011).

Examples of Raman spectroscopy applied toProterozoic microfossils include a study by Fries& Steele (2011), who mapped the carbon D tocarbon G peak intensity ratio (an indicator ofgraphite domain size) to show micron-sized varia-tions in the structure of kerogen within and aroundexamples of Huroniospora from the 1878 MaGunflint Formation. This potentially reflects initialheterogeneities in the biological material. Alsowithin the Gunflint Formation, Wacey et al. (2013)used Raman spectroscopy to demonstrate that Gun-flintia microfossils were dominantly carbonaceousin composition, but were preserved as pyrite inmicroenvironments where anoxia had allowed theformation of pyrite via the metabolic activity ofsulphate-reducing bacteria. Raman spectroscopyhas been used extensively by Schopf and coworkersto characterize Proterozoic microfossils (Schopfet al. 2005, 2008; Schopf & Kudryavtsev 2005,2009), culminating in the Raman index of preserva-tion. This correlates the geochemical maturity ofthe kerogen, the fidelity of microfossil preservation,the H : C and N : C ratios of organic material andthe metamorphic grade of the rocks. Examples havebeen reported from 22 chert units ranging in agefrom 2100 to 400 Ma (Schopf et al. 2005).

Micro-Fourier transform infrared spectroscopy.Micro-Fourier transform infrared (FTIR) spectro-scopy is a vibrational technique that providescomplementary information to that obtained fromorganic material using Raman spectroscopy. In par-ticular, it provides data pertaining to the functionalgroups attached to carbon chains and their bondingenvironment within organic material (Mayo et al.2004; Dutta et al. 2013; Chen et al. 2015). Differentpeaks in an IR spectrum arise due to different vibra-tional behaviours in the bonds of groups such asCH2, CH3, CZN, CvO and others. FTIR spectro-scopy can be applied non-destructively, but requiresdoubly polished thin sections. The main drawbackis currently the limited spatial resolution that canbe obtained, with recent studies reporting only ac. 15 mm2 spot size in the transmission mode (Quet al. 2015). This is sufficient to characterize largerProterozoic acritarchs in palynological extracts(Arouri et al. 1999; Marshall et al. 2005) and groupsof smaller filamentous and coccoid microfossils(Igisu et al. 2009), but is insufficient to determinethe difference between the wall chemistry and inter-nal chemistry of most Proterozoic organisms. Thespatial resolution problem may be circumventedby using a micro-FTIR system attached to a syn-chrotron beamline, where spot sizes of ,5 mmhave been achieved for some parts of the spectra(Bambery 2016). However, this may require morespecialist, often extremely difficult, sample prepara-tion (e.g. ,20 mm thickness, unglued slice).

Of particular interest are data from extantmicroorganisms, which suggest that FTIR may pro-vide domain-specific information, whereby specificcomponents (e.g. lipids) of different domains of life(i.e. prokaryote, eukaryote and archaea) may havecharacteristic ratios of CH2 and CH3 groups intheir IR spectra (Igisu et al. 2009, 2012). This hasled to FTIR being used in Proterozoic assemblagesin an attempt to decode the phylogenetic affinityof microfossils (Igisu et al. 2009, 2014 Q5). Igisuet al. (2009) analysed microfossils in their mineralmatrix and thus concentrated on the CHx (2500–3100 cm21) region of the spectrum. This type ofresearch is very much in its infancy and a betterunderstanding, both of the changes in CH2/CH3

during post-mortem alteration processes and of thespectral parameters of differentiated cells in multi-cellular organisms, is required for these data tobecome a robust domain-level signature. Insuffi-cient data currently exist for comparisons of organicmaterial from different terranes and of differentmetamorphic grades using this technique. Neverthe-less, FTIR analyses from the 850 Ma Bitter SpringsFormation, Australia and the 1878 Ma Gunflint For-mation, Canada suggest that organisms in these fos-sil assemblages belong to Bacteria rather thanArchaea or Eukarya (Igisu et al. 2009). Likewise,

D. WACEY ET AL.

349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406

combined FTIR and Raman data from the 1485 MaWumishan Formation, China (Qu et al. 2015) sug-gested that the organic material was derived fromprokaryote cyanobacteria and was characterizedby a homogenous and low CH3/CH2 ratio. FTIRdata from acritarchs from the c. 575 Ma Tanana For-mation, Australia suggest that Tanarium are proba-bly eukaryotic micro-algae, but Leiosphaeridia maybe Bacteria (Igisu et al. 2009, based on data pre-sented in Marshall et al. 2005).

Destructive high spatial resolution techniques

Focused ion beam milling and scanning electronmicroscopy. The technique of SEM has traditionallybeen of limited use in characterizing Proterozoicmicrofossils in geological thin sections becausethe majority of microfossils are embedded withinthe thin section and below the reach of this surface-based technique. SEM has, however, provided highspatial resolution morphological data from the sur-faces of individual microfossils in acid-etched rocksor those extracted from their host rock using acidmaceration. This has revealed, for example, delicatewall ultrastructures that could not be resolved underthe light microscope (Javaux et al. 2004; Moczy-dlowska & Willman 2009; Agic et al. 2015).

The use of SEM in Precambrian palaeobiologyhas been reinvigorated by a new generation of dual-beam instruments where the user has access to bothan FIB and an electron beam (for overview, seeYoung & Moore 2005). A highly focused beam ofheavy ions (usually Ga+) can be used to sputterions from the sample surface, essentially cuttinginto the sample with very high (nano-scale) preci-sion (for details, see Wirth 2009). The electronbeam can be used to image the results. Additionaldetectors can be inserted to image backscatteredelectrons as well as secondary electrons, allow ele-mental analysis (using an energy-dispersive X-rayspectroscopy (EDS) detector), or even phase detec-tion and crystallographic mapping (using an elec-tron backscatter diffraction detector). FIB millingcan be used to cut into, or through, specific featuresin a thin section or rock chip, allowing the structureperpendicular to the surface to be better visualized(Westall et al. 2006). A number of sequential slicescan be milled through an object, with images orother data acquired after each slice has been milled.The latter is termed FIB-SEM nano-tomographyand allows the 3D reconstruction and visualizationof microfossils at very high spatial resolutions (fordetails, see Wacey et al. 2012). The resolutionattainable is essentially dictated by the 3D size ofthe object to be analysed, plus the available time,although instrumental resolution limits may comeinto play for very small objects. Slice thicknessesare set by the user and can be ,50 nm; however,

for practical reasons 100–200 nm slices have com-monly been used. Proterozoic microfossils havebeen visualized using FIB-SEM nano-tomographyfrom the 1878 Ma Gunflint Formation (Waceyet al. 2012, 2013), the c. 1700 Ma Ruyang Group(Schiffbauer & Xiao 2009; Pang et al. 2013) andthe c. 1000 Ma Torridon Group. In the former, FIB-SEM data were key in revealing heterotrophic bac-teria attached to, and fossilized in the act of decom-posing, larger organisms (Wacey et al. 2013). Thedrawbacks of FIB-SEM nano-tomography includeits destructive nature – the analysed specimen iscompletely consumed and only a digital record ofits existence will remain – plus the restrictive time-scales involved both in analysing objects c..30 mm in diameter (24 hours or more beamtimerequired) and in processing and reconstructing thedata. A number of options exist for processing andvisualizing such data (and data from other 3Dtechniques such as X-ray CT), ranging from free-ware products – such as the serial palaeontologicalimage editing and rendering systems SPIERS (Sut-ton et al. 2012), Drishti (Limaye 2012) and Blender(Garwood & Dunlop 2014) – to more advanced (butexpensive) products such as AVIZO (http://www.vsg3d.com). The choice of software will dependon the budget, time constraints, the quality of theraw data and whether there is an interest in produc-ing just images, or images plus movies (for an over-view of the options, see Sutton et al. 2014).

Transmission electron microscopy. The techniqueof TEM covers a number of separate sub-techniquesthat can all be performed in a transmission elec-tron microscope. At its most simple, TEM is avery high spatial resolution imaging technique,capable of resolving objects separated by as littleas c. 0.1 nm. A standard TEM image results fromvariable electron scattering as a beam of electronsis accelerated at high voltage through an ultrathin(ideally ≤100 nm) sample; a true high-resolutionimage is a phase-contrast image with atomic-scaleresolution, allowing the visualization of the arrange-ment of atoms within a sample (Williams & Carter2009). This provides information about the crystal-linity of a sample, its lattice structure and anydefects it may have.

Sample preparation is the key to obtaining high-quality data and in this regard FIB has revolution-ized the use of TEM in Precambrian palaeobiology.Before the advent of FIB, sample preparation forTEM involved either grinding up a rock, extractingorganic material by acid maceration, or using ionpolishing, meaning that the context of the putativemicrofossils was often lost. It was very difficult toobtain samples of uniform (and ultrathin) thicknessand contamination was widespread. FIB millingnow allows individual microfossils, or even specific

ANALYTICAL TECHNIQUES FOR PROTEROZOIC MICROFOSSILS

407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464

parts of individual microfossils, to be targeted withgreat accuracy in their host thin section. Ultrathinwafers (typically about 15 mm × 10 mm × 100 nm)can then be extracted from below the surface ofthe thin section (hence eliminating the possibilityof contamination) and mounted on a TEM grid(for an overview, see Wacey et al. 2012).

In addition to morphology, a number of otherparameters can also be analysed by TEM, includingelemental composition, bonding and oxidation state,crystal structure (leading to mineral identification)and crystal orientation. The elemental compositionof a sample can be determined at the nano-scaleusing either EDS or by isolating and mapping spe-cific energy windows from an electron energy lossspectrum. The fine structure of peaks within an elec-tron energy loss spectrum can also be used to shedlight on the bonding and oxidation state of the ele-ment of interest – for example, distinguishing disor-dered carbon from graphite (Buseck et al. 1988) andFe2+ from Fe3+ (Calvert et al. 2005). For advancedcrystallography and mineral identification, selectedarea electron diffraction provides quantitative infor-mation on the distances between atomic planes incrystalline materials and allows the orientation ofseveral grains of the same mineral to be comparedwith one another.

The technique of TEM has been used in Protero-zoic palaeobiology for several decades, with earlyimages of microfossils extracted from their hostrock in the c. 850 Ma Bitter Springs Formation,Australia reported by Oehler (1977). A number ofstudies have investigated the wall architecture ofProterozoic acritarchs in an attempt to decodetheir taxonomic affinities because TEM can detectvariations in the electron density and texture of dif-ferent layers within cell walls at nanometre-scaleresolution. These include studies from the c.575 Ma Tanana Formation, Australia (Arouri et al.1999; Moczydlowska & Willman 2009), where therecognition of a trilaminar sheath structure waspart of a suite of evidence suggesting that the micro-fossils were chlorophyte algae. TEM helped to elu-cidate the nanostructure of carbon particles makingup the cell wall in the 650 Ma Chichkan Formation,Kazakhstan (Kempe et al. 2005). In the c. 1450 MaRoper and Ruyang groups of Australia and China,respectively (Javaux et al. 2004), at least four differ-ent types of wall ultrastructure suggested a greaterdiversity of eukaryote clades in these deposits thancould have been recognized by standard opticaltechniques. TEM has also been used to investigatethe interplay of microfossil walls with the mineralsin which they have been preserved, with studiesfrom the 1878 Ma Gunflint Formation showinghow nano-grains of silica disrupt the carbonaceouswalls of bacteria as they are fossilized (Moreau &Sharp 2004; Wacey et al. 2012). Data from the

c. 750 Ma Draken Formation, Svalbard showedboth the cell membrane and cytoplasm of the coc-coid microfossil Myxococcoides embedded withinnano-grains of silica (Foucher & Westall 2013).TEM data from the c. 580 Ma Doushantuo Forma-tion, China helped to decode the relationshipsbetween preserved microfossils and the phosphategranules in which they were contained and sug-gested that phosphate precipitation was likely tohave been microbially mediated (She et al. 2013).

Secondary ion mass spectrometry. As applied to thefield of Proterozoic palaeobiology, SIMS is a sur-face analysis technique, whereby the elemental orisotopic composition of a sample can be determinedat moderate to high spatial resolution and with greatsensitivity (i.e. many elements can be detected evenwhen present at only the parts per billion level). Thesurface of a sample is sputtered with an ion beamand the secondary ions ejected from the sampleare collected and analysed using a mass spectrome-ter (for details, see Ireland 1995). Two differenttypes of SIMS instruments are commonly used inpalaeobiological investigations.

(1) The large radius secondary ion mass spectro-meter is used to accurately determine the sta-ble isotope ratios of key biogenic elements(e.g. carbon, sulphur), plus the ratios of radio-genic isotopes, in order to date rock formationscontaining microfossils (see, for exampleStern et al. 2009; Farquhar et al. 2013; Willi-ford et al. 2013). Such instruments cananalyse objects as small as c. 10–20 mm indiameter and the isotopic data can have a pre-cision better than 0.5 parts per thousand (‰).

(2) In NanoSIMS, the mass spectrometer has adifferent geometry and is thus capable of ele-ment (ion) mapping with a lateral resolutiondown to c. 50 nm (see Kilburn & Wacey2015 for details). The NanoSIMS instrumentcan also give accurate isotopic measurementsfrom objects ,5 mm, albeit with poorer preci-sion (generally .1‰) than the large radiussecondary ion mass spectrometer.

Both forms of SIMS can be applied to surface fea-tures in standard geological thin sections and rockchips, although some specialist sample preparationis needed so that the sample and appropriatestandards can be correctly mounted together withinthe instrument. This generally involves mountingpieces of thin sections or rock chip alongside analyt-ical standards in resin discs. SIMS is partiallydestructive in that layers of the surface material(as deep as c. 200 nm during isotope analysis withlarge radius SIMS) are consumed during the analy-sis. Small specimens may be entirely consumed bythe analysis, whereas larger specimens can be repol-ished after analysis to look like new.

D. WACEY ET AL.

465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522

A number of Proterozoic microfossils have beenanalysed by SIMS in the last 15 years. House et al.(2000) were the first to determine the carbon iso-tope composition of individual microfossils usingmaterial from the c. 850 Ma Bitter Springs and1878 Ma Gunflint formations, finding d13C signa-tures (221 to 245‰) consistent within specificmetabolic pathways (namely the Calvin cycle andacetyl-CoA). This work was refined by Willifordet al. (2013), who analysed microfossils from fourProterozoic assemblages (Gunflint, Bitter Springs,plus the c. 650 Ma Chichkan and c. 740 Ma Min’yarformations) with greater precision and reproducibil-ity. They were able to show considerable variabilityof d13C within individual assemblages that mayreflect the preservation of the original metabolic dif-ferences between different components of eachbiota and also potential heterogeneities in molecularpreservation in single microfossils. It must be notedat this stage that non-biological reactions are able toproduce similar d13C fractionations (McCollom &Seewald 2006), so a d13C value must be supportedby a definitive biological morphology to prove thebiogenicity of ancient carbonaceous objects.

SIMS has also been used to investigate meta-bolic pathways involving sulphur in Proterozoicorganisms. Wacey et al. (2013) determined the d34Scomposition of pyritized microfossils from the1878 Ma Gunflint Formation, finding sulphur frac-tionations (d34S ¼ +7 to +22‰) consistent withpyrite formation via the activity of sulphate-reduc-ing bacteria in sulphate-starved sediment porewaters. In the same study, Wacey et al. (2013)used NanoSIMS to map the residual carbon andnitrogen associated with the pyritized microfossilsand found reproducible differences in the preserva-tion of organic material between two different typesof organism (Huroniospora v. Gunflintia). Gunflin-tia was poorly preserved, which suggests that itwas more prone to decay by heterotrophic bacteria(that also mediated pyrite formation) than Huronio-spora. NanoSIMS mapping of organic microfossilsin the c. 850 Ma Bitter Springs Formation has shownthe co-occurrence of carbon, nitrogen and sulphurin such microstructures (Oehler et al. 2006) andattempts have been made to quantify the ratios ofnitrogen to carbon to distinguish different compo-nents of microbial communities, or to distinguishbiological from co-occurring abiotic organic mate-rial (Oehler et al. 2009; Thomen et al. 2014),although the SIMS community has yet to agree onthe robustness of these methods.

A Proterozoic case study: the 1200–1000 Ma

Torridonian lakes

The effectiveness of combining multiple high spa-tial resolution in situ techniques is demonstrated

here using a case study of microfossils from the1200–1000 Ma Torridonian Supergroup of NWScotland. Not all the described techniques wereapplied to the Torridonian material to avoid theduplication of data and to keep costs and processingtimes to reasonable levels. For example, we felt inthis case that higher quality 3D morphologicaldata could be acquired using FIB-SEM rather thanCLSM, and that the detailed chemistry could bebetter (and more cheaply) determined using TEMrather than X-ray spectroscopy. We present dataobtained from light microscopy, SEM, mCT, laserRaman, NanoSIMS, TEM and FIB-SEM nano-tomography, which together provide a detailedcharacterization of a number of components of theTorridonian biota.

Methods

Optical microscopy. Polished and uncoveredpetrographic thin sections of 30 and 100 mm thick-ness were examined under Nikon Optiphot-Poland Nikon Optiphot-2 microscopes with 4×, 10×,20×, 40× and 100× (oil immersion) lenses at theDepartment of Earth Sciences, University of Oxfordand with a Leica DM2500M microscope with 4×,10×, 20× and 50× lenses at the Centre for Micro-scopy Characterisation and Analysis (CMCA),The University of Western Australia. Images werecaptured using Synchroscopy imaging software(Acquis and Auto-montage) at Oxford and usingToupview imaging software at CMCA. Post-pro-cessing, for example the colouring of cells inFigures 2 and 3, was carried out in Adobe Photoshop(GIMP is an open source alternative).

Scanning electron microscopy of palynologicalspecimens. Palynological samples were preparedat the Department of Animal and Plant Sciences,University of Sheffield using conventional acidmaceration techniques (Grey 1999). FollowingHCl–HF–HCl acid maceration, the residues weresieved using a 10 mm mesh. They were then treatedto a heavy liquid separation using zinc chloride, fol-lowed by further sieving at 10 mm. The organic res-idues were mounted directly onto glass slides usingepoxy resin. SEM imaging was carried out using aJEOL JSM-840A scanning electron microscopeat the Department of Earth Sciences, Universityof Oxford.

X-ray micro-computed tomography. Computedtomography scans were performed at the Manches-ter X-ray Imaging Facility using a Nikon Metris225/320 kV X-ray CT system in a customized bay(tungsten reflection target; current/voltage of130 mA/80 kV; no filtration; 3142 projections of708 ms exposure collected with a 2000 × 2000

ANALYTICAL TECHNIQUES FOR PROTEROZOIC MICROFOSSILS

523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580

detector; reconstructed dataset 5.1 mm voxels) and aZeiss Xradia Versa 520 system (standard transmis-sion target; current/voltage of 62 mA/160 kV; stan-dard in-built, high-energy 2 Zeiss filter; 4× opticalmagnification; 501–1001 projections of exposuresbetween 0.5 and 2 s collected with 4× binningusing a 2000 × 2000 detector; reconstructed data-sets with 1–2 mm voxel size). Additional propaga-tion-based phase-contrast scans were performed atthe TOMCAT beamline of the Swiss Light Source(Paul Scherrer Institut, Villigen, Switzerland; 1001projections of 700 ms exposure; 37 keV monochro-matic beam; 4× objective; a LAG:Ce 100 mm scin-tillator; reconstructions based on both attenuationand phase used to create datasets with 1.625 mmvoxels). Datasets were reconstructed using the SPI-ERS software suite (Sutton et al. 2012), followingthe methods of Garwood et al. (2012), and Drishti(Limaye 2012), following the methods of Strenget al. (2016).

Laser Raman spectroscopy. Laser Raman analyseswere carried out at the University of Bergen usinga Horiba LabRAM HR800 integrated confocalRaman system and LabSpec5 acquisition and analy-sis software. Samples were standard uncovered geo-logical thin sections, which allowed optical andchemical maps to be superimposed. All analyseswere carried out using a 514.5 nm laser, 100 mmconfocal hole, 1800 grating and 50× objectivelens. The laser was focused at least 1 mm belowthe surface of the thin sections to avoid surface pol-ishing effects. For mineral identification fromRaman spectra, dual acquisitions were taken fromeach analysis point, each with an acquisition timeof 4 s. Raman maps were acquired with a 1.5 mmspatial resolution.

Transmission electron microscopy of focused ionbeam milled wafers. The TEM wafers were preparedusing two dual-beam FIB systems (FEI NovaNanoLab) at the Electron Microscopy Unit of theUniversity of New South Wales and at AdelaideMicroscopy at the University of Adelaide. Electronbeam imaging was used to identify the microfossilsof interest in standard polished thin sections coatedwith c. 30 nm of gold, allowing site-specific TEMsamples to be prepared. The TEM sections were pre-pared by a series of steps involving different Ga+

ion beam energies and currents (see Wacey et al.2012), resulting in ultrathin wafers of c. 100 nmthickness. These TEM wafers were either attachedto Omniprobe copper TEM holders or depositedon continuous-carbon copper TEM grids. TEMdata were obtained using an FEI Titan G2 80-200TEM/STEM system with ChemiSTEM Technol-ogy operating at 200 kV, plus a JEOL 2100 LaB6

transmission electron microscope operating at

200 kV equipped with a Gatan Orius charge-coupled device camera and Tridiem energy filter.Both instruments were located at CMCA.

Nano secondary ion mass spectrometry. Ion map-ping was performed using a CAMECA NanoSIMS50 system at CMCA, with instrument parametersoptimized as described in Wacey et al. (2011). Anal-ysis areas were between 12 × 12 mm and 25 ×25 mm with a resolution of 256 × 256 pixels (soeach pixel measured between 47 and 98 nm), witha dwell time of 5–15 ms per pixel and a primarybeam current of c. 2.5 pA. The secondary ionsmapped were 24C2

2, 12C14N2, 28Si2, 32S2 and56Fe16O2; charge compensation was achieved usingthe electron flood gun.

Focused ion beam scanning electron microscopynano-tomography. Sequential FIB milling andSEM imaging was carried out on a Zeiss AurigaCrossbeam instrument at the Electron MicroscopyUnit of the University of New South Wales usingthe method of Wacey et al. (2012, 2014). Keyparameters were adjusted to suit the specific size andnature of each sample of interest. Initial trencheswere milled using a 9 nA beam current and theimaged face was cleaned using a 2 nA beam current;the ion beam current for slice milling was 2 nA, theelectron beam voltage for imaging varied betweenabout 800 V and 5 kV, the step sizes between sliceswere between 75 and 200 nm and the image capturetimes were around 30 s per frame. In some samples,dedicated trenches were milled to obtain elemental(EDS) maps of microfossils that were not subse-quently milled for 3D analysis.

To visualize the data, FIB-SEM images werestacked, aligned and cropped using SPIERSalign(Sutton et al. 2012). The resultant stacks wereimported into SPIERSedit (Sutton et al. 2012),where a number of masks were added to segmentindividual components (e.g. cell walls, cell con-tents) of the microfossil assemblage. The resultingfiles were exported and loaded into SPIERSview(Sutton et al. 2012) to generate the 3D surfacerenderings.

Results

Multiple seasons of fieldwork had been completedto gain a firm understanding of the geological con-text of the host rocks before the Torridonian micro-fossils were subjected to the high spatial resolution,in situ microanalysis described here. In addition,over 100 thin sections and hand samples had beenstudied to understand the depositional context andpost-depositional history of the rocks and to isolateonly the very best and most promising samples forfurther study. A large amount of optical microscopy

D. WACEY ET AL.

581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638

work had also been completed to form an estimateof the morphological diversity of the biota. Thiswork has all been peer reviewed and published (Cal-low et al. 2011; Strother et al. 2011; Battison &Brasier 2012; Strother & Wellman 2015), thus giv-ing a firm platform on which to build this high-resolution work. A summary of some of the mostcommon components of the Torridonian biota asobserved by optical microscopy is given in Figure 1.

Scanning electron microscopy data. As may beexpected, the range of morphologies visible inSEM analysis (Fig. 2) was broadly comparablewith that observed within thin sections of the phos-phate (Fig. 1, plus Battison & Brasier 2012). Manysimple vesicles and tubular morphotypes wereobserved, with SEM imaging affording enhancedresolution of their shape and wall structure. In par-ticular, differences in the physical responses of

structures to compression hint at differences inthe cell wall architecture. Two principal wallresponses were observed. Thicker walled (wall atleast 1 mm thick) specimens accommodate flatten-ing with broad, rounded, velvet-like folds or largecreases (Fig. 2a). By contrast, thin-walled vesicles(,0.5 mm) accommodate compression with finewrinkles irregularly distributed across the surfaceand are apparently more prone to small tears (Fig.2b). The flattening of these walls during preparationdoes not allow the resolution of any ultrastructurallamination, but a synthesis of the taphonomic res-ponse and wall thickness may be used to enhancethe interpretation of microfossils studied by opticalmicroscopy.

A number of unique forms of microfossils werealso observed by SEM. This is probably due to theprocessing of larger quantities of material duringpreparation by acid maceration, as well as the

Fig. 1.

Colour

online/

colour

hardcopy

Optical microscopy of Torridonian microfossils, demonstrating the common morphotypes present in theassemblage. (a) Highly degraded dark-walled vesicle. (b) Pristine dark-walled vesicle. (c) Light-walled vesicle,potentially possessing a double wall. (d) Cluster of light-walled spheroidal unicells, most with a dark spot indicatingthe potential preservation of cell contents. (e) Cluster of light-walled cells with mutually adpressed cell walls. (f) Pairof spheroidal unicells with very prominent dark inner sphere. (g) Partially decomposed filamentous sheath.(h) Filamentous sheath with bulbous termination housing potential spheroidal cell. (i) Colony of light-walled ellipticalcells comparable to Eohalothece lacustrina described by Strother & Wellman (2015). (j) Pair of cells that may havedivided shortly before fossilization, each containing a dark spot. Scale bars 20 mm for (a–i) and 10 mm for ( j).

ANALYTICAL TECHNIQUES FOR PROTEROZOIC MICROFOSSILS

639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696

enhanced resolution afforded by SEM imaging.Of note were two morphotypes, the first (Fig. 2c)consisting of a vesicle c. 50 mm in diameter, orna-mented with regular pits c. 10 mm across, witheach pit possessing a raised ‘collar’ c. 2 mm wideand 2 mm high. This form bears some resemblanceto the basal vesicle of Cheilofilum hysteriopsis But-terfield (see Butterfield 2005, figs 8 and 10) or thefreshwater green microalga Botryococcus braunii(see Vandenbroucke & Largeau 2007, pl. e) in itspossession of flanged openings. The second form(Fig. 2d) is a spherical hollow vesicle c. 20 mm indiameter with a spongy textured wall and irregularlydistributed, rounded or sub-circular holes c. 1–3 mm across. This morphotype is particularly nota-ble for its retention of 3D structure following mac-eration, indicating significant rigidity of the wall.In addition, non-vesicular membranous organicmatter with an irregularly pustulate and pitted tex-ture and an amorphous architecture was distributedabundantly among the structurally distinguishable

vesicles and sheaths. The size and nature of thismaterial was similar to the amorphous extra-poly-meric substances secreted by mat-forming organ-isms in modern microbial ecosystems (cf. Pactonet al. 2007), but could also be amorphous kerogen.This material was occasionally seen containedwithin thin sections as a light-walled membrane,but its texture and extent was clearer under SEManalysis.

Of particular note among the vesicles, sheathsand putative extra-polymeric substances weresmall coccoid or baccilate forms seen to be coloniz-ing, to varying degrees, some of the larger fossilstructures. These were associated with pits withinthose larger structures and were apparently embed-ded within a membrane that linked them to thehost fossil (Fig. 3). We interpreted these forms asfossils of heterotrophic bacteria preserved feedingon the larger Torridonian microbial flora and thisinterpretation reinforced observations made pre-viously using light microscopy (see Battison &

Fig. 2.

Colour

online/

colour

hardcopy

Torridonian microfossils imaged and analysed by SEM, coloured for easier interpretation. (a) Large,thick-walled vesicle showing velvet-like folds. (b) Smaller, thin-walled vesicles with a crinkled surface and finelyirregular outline. (c) Vesicle with large hemispherical pits bounded by raised rims or ‘collars’. (d) Subspherical rigidvesicle retaining a 3D structure and bearing many irregular rounded holes. Pink coccoid structures attached to thevesicles (a, b) are potential fossil heterotrophs (see also Fig. 3). Sample CAI-7, macerated from phosphate from theCailleach Head Formation. All scale bars 10 mm.

D. WACEY ET AL.

697698699700701702703704705706707708709710711712713714715716717718719720721722723724725726727728729730731732733734735736737738739740741742743744745746747748749750751752753754

Brasier 2012, fig. 9, where evidence for heterotro-phy included roughly circular holes in large micro-fossil vesicles and inferred clumps of heterotrophicbacteria pseudomorphing decayed vesicles).

X-ray micro-computed tomography data. Micro-CTwas explored as a method to investigate the petro-graphic context of cellular material and was alsotested to determine whether individual microfossils

could be detected and their 3D morphology charac-terized. Scans of rock chips from the CailleachHead Formation using a Nikon Metris 225/320 kVX-ray CT system with 5.1 mm voxels revealed phos-phate nodules as a slightly denser phase that couldbe distinguished from the surrounding matrix sedi-ment (Fig. 4a, purple colour). It also suggestedthat phosphate was present in small quantitiesclose to, but exterior to, the main nodule. Rounded

Fig. 3.

Colour

online/

colour

hardcopy

Evidence of bacterial heterotrophy in SEM images. (a, b) Rounded pits and occasional holes, irregularlydistributed on the surface of the walls of larger vesicles; (b) is an enlargement of boxed area in (a). (c, d) Collapsedcoccoid or baccilate cells c. 5 mm across, occupying pits in the walls of larger vesicles, occasionally with a thinraised lip; (d) is an enlargement of boxed area in (c) with heterotrophs false-coloured pink. (e) Densely packedcolony of coccoid and baccilate cells (pink) continuous with amorphous degraded vesicular or extra-polymericsubstances material (grey–green). (f) Higher magnification of colony in boxed area of (e) showing collapsedcoccoid and baccilate structures arranged randomly with possible supporting and sheathing membrane. SampleCAI-7, macerated from phosphate from Cailleach Head Formation. Scale bars 20 mm for (a, c, e) and 10 mm (b, d, f).

ANALYTICAL TECHNIQUES FOR PROTEROZOIC MICROFOSSILS

755756757758759760761762763764765766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804805806807808809810811812

concentrations of a very dense phase, most likely tobe an iron-rich mineral such as pyrite or iron oxide,were shown to be present both within and outsidethe nodule (Fig. 4a, gold colour). Hence CT couldbe used in future investigations as a pre-screen ofrock fragments to determine the best position withinthe rock from which to cut thin sections. The NikonCT scans detected phases of lower density withinthe phosphate nodules that may be organic micro-fossils. However, the spatial resolution of thisinstrument was insufficient to determine whetherthese lower density objects were indeed microfos-sils or simply lower density sediment grains (e.g.quartz) scattered through the phosphate nodules.

Higher resolution scans of a different rock chip(with 1.625 mm voxels) conducted at the SwissLight Source demonstrated a complex sedimentarytexture – here both phosphate and other densephases were present in the form of evenly spacedrounded to angular fragments within the scannedrock chips (Fig. 4b), with no evidence of well-formed nodules of phosphate. The lack of evidencefor nodules suggested that this rock chip would notbe a promising target for the further investigationof microfossils.

The CT scans of a sub-portion of the sampleexamined in the Nikon instrument, performedusing a Zeiss Xradia Versa 520 with voxels of

Fig. 4.

Colour

online/

colour

hardcopy

X-ray microtomography analysis of Torridonian rock chips. (a) Reconstruction of a CT scan of a rock chipusing the Nikon instrument (voxels c. 5 mm), highlighting part of a phosphate nodule (purple) within a quartz-richsediment (grey), plus a number of higher density grains that are probably pyrite or iron oxide (gold coloured).(b) Reconstruction of an X-ray scan of a second rock chip using the Swiss Light Source Synchrotron (voxels1.625 mm). This shows a mixture of phosphate and other denser phases rather evenly distributed through the rockchip with no distinct phosphate nodule. (c, d and f, g) Reconstruction of two putative vesicles identified in a higherresolution CT scan using the Zeiss Xradia Versa instrument (voxels c. 1.5 mm). The light micrograph images (e andh) show specimens observed in thin sections that may be analogous to those identified using CT. Scale bars 2 mmfor (a), 500 mm for (b) and 20 mm for (c–h).

D. WACEY ET AL.

813814815816817818819820821822823824825826827828829830831832833834835836837838839840841842843844845846847848849850851852853854855856857858859860861862863864865866867868869870

c. 1.5 mm, detected a small number of low-densityobjects that strongly resembled the microfossilsobserved in thin sections (Fig. 4c, d, f and g).These objects were analogous to some of the largestand darkest walled vesicles seen in thin section(Fig. 4e, h) and CT allowed them to be viewedfrom multiple orientations in 3D space. These puta-tive fossils were also frequently found close to thevery high density phases (presumably iron oxideor pyrite). The combined evidence suggested thatmCT at this resolution was only capable of detect-ing the largest and thickest walled components ofthe Torridonian biota. We also suggest that theincreased density contrast when such fossils occurin close proximity to iron oxide or pyrite aids detec-tion by CT. The remaining components of the biota(e.g. the examples shown in Fig. 1) are essentiallyinvisible on X-ray CT scans conducted at theseresolutions. The biggest challenge for future workwill be identifying workflows to isolate knownmicrofossils for future scanning.

Raman spectroscopy data. Raman data inform aboutthe dominant mineralogy of the Torridonian micro-fossils and their surrounding matrix, plus the struc-ture and thermal history of any organic carbon

present. Raman maps from the Cailleach HeadFormation (Fig. 5a–c) demonstrated that the micro-fossils were indeed carbonaceous (Fig. 5b) and thatthe dominant fossilizing phase was apatite (Fig. 5c).Raman spectroscopy also showed that the intra-cellular inclusions (Fig. 5a arrows), common inmany of the spheroidal fossils from this formation,were also carbonaceous in composition. Hencethese inclusions probably represent plasmolysed(shrunken) cell contents or, in some cases, couldrepresent a fossilized cell nucleus. The Raman spec-tra in the first-order region of carbon showed the twomain bands (D1 at about 1350 cm21 and G at about1600 cm21) characteristic of disordered carbona-ceous material. The D1 band was very broad (fullwidth at half peak maximum of c. 120 cm21) witha shoulder on its low wavenumber side. This shoul-der was caused by a small band at c. 1150 cm21,which was only observed in very disordered carbo-naceous material (Marshall et al. 2005). The G bandappeared to have been shifted considerably fromits value in crystalline graphite (1582 cm21) to avalue of c. 1610 cm21. This reflected an overlap ofthe G band with a well-developed disorder band(D2) at c. 1620 cm21. The spectrum indicatedthat the carbonaceous material had a very weak

Fig. 5.

Colour

online/

colour

hardcopy

Raman analysis of microfossils from the Torridonian Supergroup. (a) Optical photomicrograph of twococcoid microfossils from the Cailleach Head Formation, each containing dark interior spheroids (arrows). (b)Raman map of the carbon G c. 1600 cm21 peak showing that the microfossils have carbonaceous walls and the darkinterior spheroids are also carbonaceous. This suggests that they are clumps of degraded cellular material orremnants of a cell nucleus. (c) Raman map of the major calcium phosphate (apatite) c. 960 cm21 peak showing thata large proportion of the mineralizing phase is apatite. The patchy appearance of the apatite suggests the presence offurther mineral phases, interpreted to be clay minerals as detected in higher resolution SEM and TEM analyses (seeFigs 7 & 8). (d) Optical photomicrograph of a microfossil from the Stoer Group. Raman maps of (e) the carbon G c.1600 cm21 peak, (f) the pyrite c. 380 cm21 peak, (g) the calcite c. 1090 cm21 peak and (h) the albite c. 510 cm21

peak demonstrating that the microfossil is partially pyritized, but some carbonaceous composition remains and thatthe sediment is dominantly calcite and feldspar. Scale bars 10 mm.

ANALYTICAL TECHNIQUES FOR PROTEROZOIC MICROFOSSILS

871872873874875876877878879880881882883884885886887888889890891892893894895896897898899900901902903904905906907908909910911912913914915916917918919920921922923924925926927928

structural organization, had experienced little orno metamorphism (cf. Wopenka & Pasteris 1993)and was consistent with the previously suggestedmaximum heating of only c. 1008C (Stewart & Par-ker 1979).

Not all microfossils are preserved purely as car-bon. In the Stoer Group, Raman spectroscopyrevealed that significant portions of the microfossilwalls have been pyritized, although some carbona-ceous signal remains (Fig. 5d–f). The matrix miner-alogy was also different here, with typical phasesincluding calcite and albitic feldspar (Fig. 5g, h).These data indicated that different suites of lakeswithin the Torridonian had different chemistries,with those of the Stoer Group being sulphate-richand phosphate-poor compared with those of theCailleach Head and Diabaig formations (for furtherdetails on contrasting fossil preservation in theselakes, see Wacey et al. 2016).

NanoSIMS data. The NanoSIMS technique wasused as an additional tool to determine whetherthe microfossils were composed of carbonaceousmaterial and then to determine whether any addi-tional elements of biological interest were preser-ved within their cell walls or intracellular space.NanoSIMS uniquely revealed significant (but notquantifiable) amounts of nitrogen and sulphurwithin cellular material from the Diabaig Formation(Fig. 6). These data were collected from FIB-milledwafers and so the nitrogen and sulphur came fromcell walls located below the surface of a thin section.This negated the possibility that these biologicalsignals came from surface contamination and pro-vided an improvement on previous NanoSIMSmethodology where ion mapping was performedon surface features (e.g. Oehler et al. 2006, 2009).The co-occurrence of C, N and S in microstructuresthat have a cellular morphology is strong evidence

Fig. 6.

Colour

online/

colour

hardcopy

NanoSIMS analysis of a microfossil from the Diabaig Formation. (a) Optical photomicrograph of alight-walled spheroidal cell with ruby red intracellular particles. (b) Overview of a FIB-milled wafer prepared forNanoSIMS from the region indicated by the yellow line in (a). Note the contrast between the large dark grey grains,which equate to the ruby red grains in (a), and the remainder of the wafer, plus holes in the wafer probably inducedby excessive FIB milling. (c) NanoSIMS ion map of nitrogen measured as CN2. (d) NanoSIMS ion map of sulphurmeasured as S2. (e) NanoSIMS ion map of iron oxide measured as FeO2. (f) Three-colour overlay of nitrogen(blue), iron oxide (red) and silicon (green) showing that the large dark grains are iron oxides and they are locatedjust inside the cell wall (intracellular). The other mineral phases are dominantly clays and quartz. Scale bar 20 mmin (a) and 5 mm for (b–f). Note scale bar in (c) also applies to (d–f).

D. WACEY ET AL.

929930931932933934935936937938939940941942943944945946947948949950951952953954955956957958959960961962963964965966967968969970971972973974975976977978979980981982983984985986

of the biogenicity of such structures. Although thisis less relevant to the Torridonian material, the bio-genicity of which is well accepted, it is a very usefultool for the investigation of older and/or more con-troversial fossil material. Building up a databaseof the C, N and S concentrations of different typesof organic material may also be useful in helpingto determine whether different components ofcells (i.e. the cell wall, membrane, nucleus and cyto-plasm) can be preserved in exceptional circum-stances. NanoSIMS also revealed the nature ofsome non-carbonaceous intracellular inclusionswithin the Diabaig Formation. These inclusionsare ruby red in colour in optical microscopy (Fig.6a) and NanoSIMS showed that they were ironoxides (Fig. 6e, f ) and that at least some occurredin direct contact with the inner cell wall. Theseinclusions were rare, found in ,1% of Torridonianmicrofossils, but may indicate a unique intracellularchemistry in this small proportion of specimens.

Transmission electron microscopy data. The TEMdata revealed the chemistry of the fossilizingmineral phases and the ultrastructure of the micro-fossils at a spatial scale (nanometres) unattainableby any other technique. For example, ChemiSTEM(STEM-EDS) elemental mapping combined withselected area electron diffraction has shown thatphosphate is not necessarily the dominant mineralresponsible for the exceptional microfossil preser-vation in the Cailleach Head and Diabaig formations(cf. Raman and optical data). In fact, the mineralsimmediately adjacent to most vesicle walls are iron-rich clay minerals of the chlorite group or potas-sium-rich clay minerals similar to illite (Fig. 7;see Wacey et al. 2014 for details on clay mineralidentification). Phosphate only dominates at some

distance (tens to hundreds of nanometres) awayfrom the cellular material. The interiors of manymicrofossils were also filled with potassium-richclay minerals (Fig. 7), although phosphate grainswere also common in many cell interiors (e.g.Wacey et al. 2014, fig. 8). STEM-EDS in the trans-mission electron microscope detected small C and Fpeaks in the phosphate spectra, confirming that thephosphate was francolite (carbonate fluorapatite),the common low-temperature form often associatedwith fossils.

The TEM imaging revealed sub-componentsof microfossil walls that were not previously recog-nized. In many cases a presumed single, thickvesicle wall was shown to consist of multiple com-ponents. These included a thicker inner wall sittingwithin a thinner outer wall, perhaps suggesting acyst housed within a vegetative cell, or even morecomplex arrangements of up to four distinct layerswithin a ‘wall zone’ (Fig. 7). Such arrangements aretoo complex for simple prokaryote cells. Hence thisstrongly suggested a eukaryotic component to thebiota. These complex layered walls were also pre-served in clay minerals. Hence the combined datasuggested that the fidelity of microfossil preser-vation may be enhanced by the early precipitationof clay minerals and that microfossil preservationin clay minerals may be of even higher qualitythan in phosphate.

Focused ion bean scanning electron microscopydata. Two types of data were acquired usingFIB-SEM: chemical and 3D morphological. Chem-ical data were acquired by simply slicing into amicrofossil using an FIB and then analysing thechemistry of a cross-section through the microfossilusing SEM-EDS. This provided similar data to

Fig. 7.

Colour

online/

colour

hardcopy

TEM analysis of a FIB milled wafer extracted from a Torridonian microfossil. (a) Optical photomicrographof a dark-walled spheroidal microfossil from the Cailleach Head Formation. (b) Overview of the FIB-milled waferextracted from the region marked by the yellow line in (a) showing a complex wall structure and different mineralphases (indicated by different levels of grey within the image) inside and outside of the microfossil (from Waceyet al. 2014). (c) Three-colour overlay of ChemiSTEM elemental maps of carbon (blue), aluminium (orange) andcalcium (pink) from the region indicated by the dashed box in (b). Carbon represents the organic material of themicrofossil walls and at least four separate walls (or wall layers) can be seen. Calcium represents apatite, thedominant mineral phase outside the microfossil. Aluminium represents clay minerals that infill the microfossil,occur between the walls of the microfossil and occur in minor amounts outside the microfossil. Black areas areholes in the TEM wafer. Scale bar 10 mm in (a), 2 mm in (b) and 1 mm in (c).

ANALYTICAL TECHNIQUES FOR PROTEROZOIC MICROFOSSILS

987988989990991992993994995996997998999100010011002100310041005100610071008100910101011101210131014101510161017101810191020102110221023102410251026102710281029103010311032103310341035103610371038103910401041104210431044

STEM-EDS in TEM, but at a more flexible spatialscale (i.e. it could be applied to larger fossils, albeitat a lower spatial resolution). These data reinforcedthose acquired using TEM, showing that, in fossilswith complex walls (interpreted as eukaryotes),clay minerals occurred in direct contact with micro-fossil walls, in between multiple walls and in micro-fossil interiors, whereas calcium phosphate tendedto occur exterior to the fossil (Fig. 8). The pattern

was less defined in simpler prokaryote fossils, withphosphate mixed with clay minerals typically occur-ring both exterior and interior to the cell (Fig. 9a, b).

Morphological data in three dimensions wereacquired using FIB-SEM nano-tomography, where-by sequential FIB slicing was followed by imagingusing SEM. This provided an excellent visualizationof the cellular material located below the surfaceof the thin section (Fig. 9b) that would otherwise

Fig. 8.

Colour

online/

colour

hardcopy

FIB-SEM-EDS analysis of a microfossil from the Cailleach Head Formation. (a) Optical photomicrograph ofa dark-walled spheroidal vesicle showing the location of the FIB-milled area and direction of view for the otherpanels in the figure (from Wacey et al. 2014). (b) Secondary electron image showing the FIB-milled face below thesurface of the thin section. The EDS elemental maps of the FIB-milled face shown in (b) are given below. Carbon(light blue) represents the organic microfossil walls, highlighting a thick inner cyst wall and thinner outer vegetativecell wall. Phosphorus (red), calcium (pink) and moderate levels of oxygen (green) represent apatite, the dominantfossilizing mineral outside the microfossil. Iron (blue), plus moderate amounts of silicon (turquoise), aluminium(orange) and oxygen, represents iron-rich clay, occurring between the two microfossil walls, replacing parts of theouter wall and continuing for 1–2 mm outside the outer wall. Potassium (yellow), plus silicon, aluminium andoxygen represents potassium-rich clay restricted to the interior of the vesicle. Scale bars 5 mm.

D. WACEY ET AL.

1045104610471048104910501051105210531054105510561057105810591060106110621063106410651066106710681069107010711072107310741075107610771078107910801081108210831084108510861087108810891090109110921093109410951096109710981099110011011102

have been hidden by the overlying fossil material(Fig. 9a). In addition, individual cells and cell con-tents could be visualized from multiple orientationsin 3D space (Fig. 9c–f). This is particularly usefulfor accurately locating the position of organic intra-cellular inclusions (Fig. 9c–f). In the example pre-sented here, these inclusions were most likelyshrunken remnants of the cytoplasm of simple pro-karyote cells, but in future it may be possible todetect the preserved remnants of eukaryotic nucleior organelles using such methods.

Conclusions

We have provided an overview of the types ofhigh-resolution techniques currently available tothose researchers interested in characterizing Prote-rozoic microfossils and their associated mineralsand fabrics. The techniques have been classifiedeither as non-destructive, hence applicable to allmaterial including holotypes, or destructive, appli-cable where the conservation of the specimen is nota requirement. Non-destructive techniques includelaser Raman spectroscopy, CLSM, SEM, infraredspectroscopy, X-ray CT and X-ray spectroscopy,

although specialized (and partly destructive) samplepreparation is required to obtain the highest spatialresolution data using the latter two methods.Destructive techniques include SIMS, where thesurface layers of a microfossil are sputtered awayduring analysis, TEM, where an ultrathin slicemust be extracted from the microfossil, and FIB-SEM nano-tomography, which consumes the entirespecimen during analysis.

Maximum information is gained by the consil-ience of multiple approaches to a microfossil assem-blage, but in reality there will be some trade-offbetween time and budget constraints, efforts to con-serve the best specimens and the spatial resolutionrequired. The destructive techniques of TEM andFIB-SEM provide the greatest spatial resolution,whereas SIMS uniquely provides isotopic data. Asensible workflow would involve an analysis ofthe petrographic context and a significant numberof representative specimens using non-destructiveavenues, followed by the focused analysis of a fewspecimens by destructive techniques.

A case study from the Torridonian of NW Scot-land, a microfossil assemblage whose importancehas been highlighted by work led by Martin Brasier,demonstrated the additional insights that these

Fig. 9.

Colour

online/

colour

hardcopy

Three-dimensional FIB-SEM nano-tomography of a Torridon microfossil. (a) Optical photomicrograph of acluster of light-walled spheroidal cells from the Cailleach Head Formation (from Wacey 2014). (b) Example of anFIB-milled slice through the cluster of microfossils in the region indicated by the dashed line in (a). Note thatportions of at least eight cells can be seen in this image, some of which are hidden from view below other cells inthe optical photomicrograph. Note also dark material inside the upper central cell (dashed arrow). (c–f) 3D model ofthe cell indicated by the solid arrow in (b) viewed from four different orientations, showing the location ofpreserved cell contents (blue) with respect to the cell wall (yellow). Note that part of the cell wall in (f) has beenremoved to better visualize the cell contents. Scale bar 10 mm in (a) and 5 mm for (b–f). Note scale bar in (c) alsoapplies to (d–f).

ANALYTICAL TECHNIQUES FOR PROTEROZOIC MICROFOSSILS

1103110411051106110711081109111011111112111311141115111611171118111911201121112211231124112511261127112811291130113111321133113411351136113711381139114011411142114311441145114611471148114911501151115211531154115511561157115811591160

high-resolution techniques can offer. Micro-CT pro-vided a rapid way to determine the locality of thephosphate nodules that house microfossils andother petrological details. SEM revealed a numberof new morphotyes not previously recognized inoptical work and hinted at different taphonomicresponses by different types of cell and vesiclewalls. TEM revealed the fine-scale distribution ofmineral phases in and around the cellular materialand showed that clay minerals played an importantpart in the exceptional preservation of this biota.Raman spectroscopy, together with NanoSIMS,revealed details of the organic material making upthe cells, including its thermal maturity and bio-chemistry in terms of the C, N and S contents. FIB-SEM nano-tomography provided a detailed 3D viewof a number of fossilized cells, including the loca-tion of the remains of organic cell contents.

We acknowledge the facilities, scientific and technicalassistance of the Australian Microscopy and MicroanalysisResearch Facility at the Centre for Microscopy Character-isation and Analysis, The University of Western Australiaand the Electron Microscopy Unit, The University of NewSouth Wales. These facilities are funded by the universi-ties, state and commonwealth governments. We acknowl-edge the following for technical assistance: Matt Kilburnand Paul Guagliardo for assistance with NanoSIMS; Mar-tin Saunders for assistance with TEM; Charlie Kong forassistance with FIB; Phil Withers and the ManchesterX-ray Imaging Facility, which was funded in part by theEPSRC (grants EP/F007906/1, EP/F001452/1 and EP/I02249X/1); Nicola McLoughlin for access to the Univer-sity of Bergen laser Raman facility; Owen Green for thinsection preparation and access to the Oxford Universityscanning electron microscopy; Imran Rahman, Alan Spen-cer, Jonny Waters and Nidia Alvarez for assistance onbeamtime and the Paul Scherrer Institut, Villigen, Switzer-land for the provision of synchrotron radiation beamtimeon the TOMCAT beamline at the Swiss Light Source;and Charles Wellman for the acid maceration of microfos-sils. DW was funded by the European Commission and theAustralian Research Council (FT140100321). RG is a Sci-entific Associate at the Natural History Museum, Londonand a member of the Interdisciplinary Centre for AncientLife (UMRI). This is CCFS paper XXXQ6 . We thank KevinLepot and James D. Schiffbauer for their review commentsthat improved the quality of this paper.

References

Agic, H., Mocztdlowska, M. & Yin, L.-M. 2015.Affinity, life cycle, and intracellular complexity oforganic-walled microfossils from the Mesoproterozoicof Shanxi, China. Journal of Paleontology, 89, 28–50.

Amos, W.B. & White, J.G. 2003. How the confocal laserscanning microscope entered biological research. Biol-ogy of the Cell, 95, 335–342.

Arouri, K., Greenwood, P.F. & Walter, M.R.1999. A possible chlorophycean affinity of some

Neoproterozoic acritarchs. Organic Geochemistry,30, 1323–1337.

Bambery, K.R. 2016. Synchrotron infrared micro-spectroscopy of single cells at the Australian Synchro-tron. Abstract presented at the 24th Australian Confer-ence on Microscopy and Microanalysis, 31 January–4February 2016, Melbourne, Australia.

Barghoorn, E.S. & Tyler, S.A. 1965. Microorganismsfrom the Gunflint Chert. Science, 147, 563–577.

Battison, L. & Brasier, M.D. 2012. Remarkably pre-served prokaryote and eukaryote microfossils within1 Ga-old lake phosphates of the Torridon Group,NW Scotland. Precambrian Research, 196–197,204–217.

Bernard, S., Benzerara, K. & Beyssac, O. 2007.Exceptional preservation of fossil plant spores in high-pressure metamorphic rocks. Earth and Planetary Sci-ence Letters, 262, 257–272.

Beyssac, O., Goffe, B., Chopin, C. & Rouzard, J.N.2002. Raman spectra of carbonaceous material inmetasediments: a new geothermometer. Journal ofMetamorphic Geology, 20, 859–871.

Brasier, M.D., Green, O.R., Lindsay, J.F., McLough-

lin, N., Steele, A. & Stoakes, C. 2005. Critical test-ing of Earth’s oldest putative fossil assemblage fromthe �3.5 Ga Apex chert, Chinaman Creek, WesternAustralia. Precambrian Research, 140, 55–102.

Brasier, M.D., Antcliffe, J., Saunders, M. & Wacey,D. 2015. Changing the picture of Earth’s earliest fos-sils (3.5–1.9 Ga) with new approaches and new dis-coveries. Proceedings of the National Academy ofSciences of the United States of America, 112,4859–4864.

Brasier, M.D., Norman, D.B. et al. 2016. Remarkablepreservation of brain tissues in an Early Cretaceousiguanodontian dinosaur. In: Brasier, A.T., McIlroy,D. & McLoughlin, N. (eds) Earth System Evolutionand Early Life: a Celebration of the Work of MartinBrasier. Geological Society, London, Special Publica-tions, 448, http://doi.org/10.1144/SP448.3

Buseck, P.R., Bo-Jun, H. & Miner, B. 1988. Structuralorder and disorder in Precambrian kerogens. OrganicGeochemistry, 12, 221–234.

Butterfield, N.J. 2005. Reconstructing a complex earlyNeoproterozoic eukaryote, Wynniatt Formation, Arc-tic Canada. Lethaia, 38, 155–159.

Callow, R., Battison, L. & Brasier, M.D. 2011.Diverse microbially induced sedimentary structuresfrom 1 Ga lakes of the Diabaig Formation, TorridonGroup, northwest Scotland. Sedimentary Geology,239, 117–128.

Calvert, C.C., Brown, A. & Brydson, R. 2005. Deter-mination of the local chemistry of iron in inorganicand organic materials. Journal of Electron Spectro-scopy and Related Phenomena, 143, 173–187.

Cavalazzi, B., Westall, F., Cady, S.L., Barbieri, R. &Foucher, F. 2011. Potential fossil endoliths in vesicu-lar pillow basalt, Coral Patch Seamount, eastern NorthAtlantic Ocean. Astrobiology, 11, 619–632.

Chen, Y., Zou, C., Mastalerz, M., Hu, S., Gasaway, C.& Tao, X. 2015. Applications of micro-Fourier trans-form infrared spectroscopy (FTIR) in the geologicalsciences – a review. International Journal of Molecu-lar Sciences, 16, 30223–30250.

D. WACEY ET AL.

1161116211631164116511661167116811691170117111721173117411751176117711781179118011811182118311841185118611871188118911901191119211931194119511961197119811991200120112021203120412051206120712081209121012111212121312141215121612171218

Chi, H., Xiao, Z., Fu, D. & Lu, Z. 2006. Analysis of fluo-rescence from algae fossils of the NeoproterozoicDoushantuo Formation of China by confocal laserscanning microscope. Microscopy Research and Tech-nique, 69, 253–259.

Cnudde, V. & Boone, M.N. 2013. High-resolution X-raycomputed tomography in geosciences: a review of thecurrent technology and applications. Earth-ScienceReviews, 123, 1–17.

Conroy, G.C. & Vannier, M.W. 1984. Noninvasivethree-dimensional computer imaging of matrix-filledfossil skulls by high-resolution computed tomography.Science, 226, 456–458.

De Gregorio, B.T., Sharp, T.G., Flynn, G.J., Wirick, S.& Hervig, R.L. 2009. Biogenic origin for Earth’s old-est putative microfossils. Geology, 37, 631–634.

Donoghue, P.C.J., Bengtson, S. et al. 2006. Synchro-tron X-ray tomographic microscopy of fossil embryos.Nature, 442, 680–683.

Dutta, S., Hartkopf-Froder, C., Witte, K., Brocke, R.& Mann, U. 2013. Molecular characterization of fossilpalynomorphs by transmission micro-FTIR spectro-scopy: implications for hydrocarbon source evaluation.International Journal of Coal Geology, 115, 13–23.

Edwards, N.P., Manning, P.L. et al. 2014. Leaf metal-lome preserved over 50 million years. Metallomics, 6,774–782.

Farquhar, J., Cliff, J. et al. 2013. Pathways for Neo-archean pyrite formation constrained by mass-indepen-dent sulfur isotopes. Proceedings of the NationalAcademy of Sciences of the United States of America,110, 17638–17643.

Fenter, P., Rivers, M., Sturchio, N. & Sutton, S. (eds)2002. Applications of Synchrotron Radiation in Low-Temperature Geochemistry and Environmental Sci-ences. Reviews in Mineralogy and Geochemistry, 49.Mineralogy Society of America, Washington DC.

Foucher, F. & Westall, F. 2013. Raman imaging ofmetastable opal in carbonaceous microfossils of the700–800Ma Old Draken Formation. Astrobiology,13, 57–67.

Fries, M. & Steele, A. 2011. Raman spectroscopy andconfocal Raman imaging in mineralogy and petrogra-phy. In: Dieing, T., Hollricher, O. & Toporski, J.(eds) Confocal Raman Microscopy. Springer Seriesin Optical Sciences, 158, 111–135.Q7

Garwood, R.J.& Dunlop, J.2014.The walkingdead:Blen-der as a tool for paleontologists with a case study onextinct arachnids.Journal of Paleontology, 88, 735–746.

Garwood, R.J. & Sutton, M.D. 2010. X-ray micro-tomography of Carboniferous stem-Dictyoptera: newinsights into early insects. Biology Letters, 6, 699–702.

Garwood, R.J., Ross, A., Sotty, D., Chabard, D.,Charbonnier, S., Sutton, M. & Withers, P.J.2012. Tomographic reconstruction of neopterous Car-boniferous insect nymphs. PLoS One, 7, e45779.

Grey, K. 1999. A Modified Palynological PreparationTechnique for the Extraction of Large NeoproterozoicAcanthomorph Acritarchs and other Acid InsolubleMicrofossils. Geological Survey of Western Australia,Record 1999/10.

Hagadorn, J.W., Xiao, S. et al. 2006. Cellular and sub-cellular structure of Neoproterozoic animal embryos.Science, 314, 291–294.

Halbhuber, K.-J. & Konig, K. 2003. Modern laser scan-ning microscopy in biology, biotechnology and medi-cine. Annals of Anatomy, 185, 1–20.

Haubitz, B., Prokop, M., Doehring, W., Ostrom, J.H.& Wellnhofer, P. 1988. Computed tomography ofArchaeopteryx. Palaeobiology, 14, 206–213.

Hickman-Lewis, K., Garwood, R., Withers, P. &Wacey, D. 2016. X-ray microtomography as a toolfor investigating the petrological context of Precam-brian cellular remains. In: Brasier, A.T., McIlroy,D. & McLoughlin, N. (eds) Earth System Evolutionand Early Life: a Celebration of the Work of MartinBrasier. Geological Society, London, Special Publica-tions, 448, http://doi.org/10.1144/SP448.11

Hofmann, H.J. 1976. Precambrian microflora, BelcherIslands, Canada: significance and systematics. Journalof Palaeontology, 50, 1040–1073.

House, C.H., Schopf, J.W., McKeegan, K.D., Coath,C.D., Harrison, T.M. & Stetter, K.O. 2000. Carbonisotopic composition of individual Precambrian micro-fossils. Geology, 28, 707–710.

Huldtgren, T., Cunningham, J.A., Yin, C., Stampa-

noni, M., Marone, F., Donoghue, P.C.J. & Bengt-

son, S. 2011. Fossilized nuclei and germinationstructures identify Ediacaran ‘animal embryos’ asencysting protists. Science, 334, 1696–1699.

Igisu, M., Ueno, Y., Shimojima, M., Nakashima, S.,Awramik, S.M., Ohta, H. & Maruyama, S. 2009.Micro-FTIR spectroscopic signatures of bacteriallipids in Proterozoic microfossils. PrecambrianResearch, 173, 19–26.

Igisu, M., Takai, K. et al. 2012. Domain-level identifica-tion and quantification of relative prokaryote cell abun-dance in microbial communities by micro-FTIRspectroscopy. Environmental Microbiology Reports,4, 42–49.

Ireland, T.R. 1995. Ion microprobe mass spectrometry:techniques and applications in cosmochemistry, geo-chemistry, and geochronology. In: Hyman, M. &Rowe, M. (eds) Advances in Analytical Geochemistry.JAI Press, Greenwich, CT, 1–118.

Javaux, E.J., Knoll, A.H. & Walter, M.R. 2004. TEMevidence for eukaryotic diversity in mid-Proterozoicoceans. Geobiology, 2, 121–132.

Jehlicka, J., Urban, O. & Pokorny, J. 2003. Ramanspectroscopy of carbon and solid bitumens in sedimen-tary and metamorphic rocks. Spectrochimica Acta A,59, 2341–2352.

Kak, A.C. & Slaney, M. 2001. Principles of Computer-ized Tomographic Imaging. Society of Industrial andApplied Mathematics, Philadelphia.

Kempe, A., Wirth, R., Altermann, W., Stark, R.W.,Schopf, J.W.& Heckl, W.M. 2005.Focussed ion beampreparation and in situ nanoscopic study of Precam-brian acritarchs. Precambrian Research, 140, 36–54.

Kilburn, M.R. & Wacey, D. 2015. NanoSIMS as an ana-lytical tool in the geosciences. In: Grice, K. (ed.) Prin-ciples and Practice of Analytical Techniques inGeosciences. Royal Society of Chemistry, Cambridge,1–34.

Knoll, A.H. 1994. Proterozoic and early Cambrian pro-tists: evidence for accelerating evolutionary tempo.Proceedings of the National Academy of Sciences ofthe United States of America, 91, 6743–6750.

ANALYTICAL TECHNIQUES FOR PROTEROZOIC MICROFOSSILS

1219122012211222122312241225122612271228122912301231123212331234123512361237123812391240124112421243124412451246124712481249125012511252125312541255125612571258125912601261126212631264126512661267126812691270127112721273127412751276

Knoll, A.H. 2015. Life on a Young Planet: the First ThreeBillion Years of Evolution on Earth. 2nd edn. PrincetonUniversity Press, Princeton, NJ.

Knoll, A.H., Javaux, E.J., Hewitt, D. & Cohen, P.2006. Eukaryotic organisms in Proterozoic oceans.Philosophical Transactions of the Royal Society B,361, 1023–1038.

Lawrence, J.R., Swerhone, G.D.W., Leppard, G.G.,Araki, T., Zhang, X., West, M.M. & Hitchcock,A.P. 2003. Scanning transmission X-ray, laser scan-ning, and transmission electron microscopy mappingof the exoploymeric matrix of microbial biofilms.Applied and Environmental Microbiology, 69,5543–5554.

Lemelle, L., Labrot, P., Salome, M., Simionovici, A.,Viso, M. & Westall, F. 2008. In situ imaging oforganic sulfur in 700–800 My-old Neoproterozoicmicrofossils using X-ray spectromicroscopy at the SK-edge. Organic Geochemistry, 39, 188–202.

Limaye, A. 2012. Drishti: a volume exploration andpresentation tool. In: SPIE Optical Engineering+Applications. International Society for Optics and Pho-tonics, 85060X–85060X.Q8

Marshall, C.P. & Olcott Marshall, A. 2013. Ramanhyperspectral imaging of microfossils: potential pit-falls. Astrobiology, 13, 920–931.

Marshall, C.P., Javaux, E.J., Knoll, A.H. & Walter,M.R. 2005. Combined micro-Fourier transform infra-red (FTIR) spectroscopy and micro-Raman spectro-scopy of Proterozoic acritarchs: a new approach topalaeobiology. Precambrian Research, 138, 208–224.

Marshall, C.P., Emry, J.R. & Olcott Marshall, A.2011. Haematite pseudomicrofossils present in the3.5-billion-year-old Apex Chert. Nature Geoscience,4, 240–243.

Mayo, D.W., Miller, F.A. & Hannah, R.W. (eds)2004. Course Notes on the Interpretation of Infra-red and Raman Spectra. Wiley Online Library, Hobo-ken, NJ.

McCollom, T.M. & Seewald, J.S. 2006. Carbon isotopecomposition of organic compounds produced by abi-otic synthesis under hydrothermal conditions. Earthand Planetary Science Letters, 243, 64–84.

Moczydlowska, M. & Willman, S. 2009. Ultrastructureof cell walls in ancient microfossils as a proxy to theirbiological affinities. Precambrian Research, 173,27–38.

Moreau, J.W. & Sharp, T.G. 2004. A transmission elec-tron microscopy study of silica and kerogen biosigna-tures in �1.9 Ga Gunflint microfossils. Astrobiology,4, 196–210.

Oehler, D.Z. 1977. Pyrenoid-like structures in Late Pre-cambrian algae from the Bitter Springs Formation.Journal of Paleontology, 51, 885–901.

Oehler, D.Z., Robert, F., Mostefaoui, S., Meibom, A.,Selo, M. & McKay, D.S. 2006. Chemical mapping ofProterozoic organic matter at submicron spatial resolu-tion. Astrobiology, 6, 838–850.

Oehler, D.Z., Robert, F. et al. 2009. NanoSIMS:insights to biogenicity and syngeneity of Archaean car-bonaceous structures. Precambrian Research, 173,70–78.

Pacton, M., Fiet, N. & Gorin, G. 2007. Bacterial activityand preservation of sedimentary organic matter: the

role of exopolymeric substances. GeomicrobiologyJournal, 24, 571–581.

Pang, K., Tang, Q. et al. 2013. The nature and origin ofnucleus-like intracellular inclusions in Paleoprotero-zoic eukaryote microfossils. Geobiology, 11, 499–510.

Pasteris, J.D. & Wopenka, B. 2003. Necessary, butnot sufficient: Raman identification of disordered car-bon as a signature of ancient life. Astrobiology, 3,727–738.

Qu, Y., Engdahl, A., Zhu, S., Vajda, V. & McLough-

lin, N. 2015. Ultrastructural heterogeneity of carbona-ceous material in ancient cherts: investigatingbiosignature origin and preservation. Astrobiology,15, 825–842.

Rahman, I.A. & Zamora, S. 2009. The oldest cinctan car-poid (stem-group Echinodermata), and the evolution ofthe water vascular system. Zoological Journal of theLinnean Society, 157, 420–432.

Rasmussen, B. 2000. Filamentous microfossils in a3,235-million-year-old volcanogenic massive sulphidedeposit. Nature, 405, 676–679.

Rasmussen, B., Fletcher, I.R., Brocks, J.J. & Kilburn,M.R. 2008. Reassessing the first appearance of eukary-otes and cyanobacteria. Nature, 455, 1101–1104. Q9

Rasmussen, B., Blake, T.S., Fletcher, I.R. & Kilburn,M.R. 2009. Evidence for microbial life in synsedimen-tary cavities from 2.75 Ga terrestrial environments.Geology, 37, 423–426. Q10

Rowe, T.B., Colbert, M., Ketcham, R.A., Maisano, J.& Owen, P. 2001. High-resolution X-ray computedtomography in vertebrate morphology. Journal ofMorphology, 248, 277–278.

Schiffbauer, J.D. & Xiao, S. 2009. Novel application offocused ion beam electron microscopy (FIB-EM) inpreparation and analysis of microfossil ultrastructures:a new view of complexity in early eukaryotic organ-isms. Palaios, 24, 616–626.

Schiffbauer, J.D., Xiao, S., Sharma, K.S. & Wang, G.2012. The origin of intracellular structures in Edia-caran metazoan embryos. Geology, 40, 223–226.

Schopf, J.W. & Kudryavtsev, A.B. 2005. Three-dimensional Raman imagery of Precambrian micro-scopic organisms. Geobiology, 3, 1–12.

Schopf, J.W. & Kudryavtsev, A.B. 2009. Confocal laserscanning microscopy and Raman imaging of ancientmicroscopic fossils. Precambrian Research, 173,39–49.

Schopf, J.W., Kudryavtsev, A.B., Agresti, D.G.,Czaja, A.D. & Wdowiak, T.J. 2005. Raman imagery:a new approach to assess the geochemical maturity andbiogenicity of permineralized Precambrian fossils.Astrobiology, 5, 333–371.

Schopf, J.W., Tripathi, A.B. & Kudryavtsev, A.B.2006. Three-dimensional confocal optical imagery ofPrecambrian microscopic organisms. Astrobiology, 6,1–16.

Schopf, J.W., Tewari, V.C. & Kudryavtsev, A.B. 2008.Discovery of a new chert-permineralized microbiota inthe Proterozoic Buxa Formation of the Ranjit Window,Sikkim, northeast India, and its astrobiological impli-cations. Astrobiology, 8, 735–746.

She, Z., Strother, P. et al. 2013. Terminal Proterozoiccyanobacterial blooms and phosphogenesis docu-mented by the Doushantuo granular phosphorites

D. WACEY ET AL.

1277127812791280128112821283128412851286128712881289129012911292129312941295129612971298129913001301130213031304130513061307130813091310131113121313131413151316131713181319132013211322132313241325132613271328132913301331133213331334

I: in situ micro-analysis of textures and composition.Precambrian Research, 235, 20–35.

Spencer, A.R.T., Hilton, J. & Sutton, M.D. 2013.Combined methodologies for three-dimensional recon-struction of fossil plants preserved in siderite nodules:Stephanospermum braidwoodensis nov sp (Medullo-sales) from the Mazon Creek lagerstatte. Review ofPalaeobotany and Palynology, 188, 1–17.

Stern, R.A., Bodorkos, S., Kamo, S.L., Hickman, A.H.& Corfu, F. 2009. Measurement of SIMS instrumen-tal mass fractionation of Pb isotopes during zircondating. Geostandards and Geoanalytical Research,33, 145–168.

Stewart, A.D. & Parker, A. 1979. Palaeosalinity andenvironmental interpretation of red beds from theLate Precambrian (‘Torridonian’) of Scotland. Sedi-mentary Geology, 22, 229–241.

Streng, M., Butler, A.D., Peel, J.S., Garwood, R.J. &Caron, J. 2016. A new family of Cambrian rhyncho-nelliformean brachiopods (Order Naukatida) with anaberrant coral-like morphology. Palaeontology, 59,269–293.Q11

Strother, P.K. & Wellman, C.H. 2015. Palaeoecologyof a billion-year-old non-marine cyanobacteriumfrom the Torridon Group and Nonsuch Formation.Palaeontology, http://doi.org/10.1111/pala.12212

Strother, P.K., Battison, L., Brasier, M.D. & Well-

man, C.H. 2011. Earth’s earliest non-marine eukary-otes. Nature, 473, 505–509.

Sutton, M.D. 2008. Tomographic techniques for thestudy of exceptionally preserved fossils. Proceedingsof the Royal Society B, 275, 1587–1593.

Sutton, M.D., Briggs, D.E.G., Siveter, D.J. & Siveter,D.J. 2001. An exceptionally preserved vermiformmollusk from the Silurian of England. Nature, 410,461–463.

Sutton, M.D., Garwood, R.J., Siveter, D.J. & Siveter,D.J. 2012. SPIERS and VAXML: a software toolkit fortomographic visualization and a format for virtualspecimen interchange. Palaeontologica Electronica,15, 5T.

Sutton, M.D., Rahman, I.A. & Garwood, R.J.2014. Techniques for Virtual Palaeontology. Wiley-Blackwell, Oxford.

Templeton, A. & Knowles, E. 2009. Microbial trans-formations of minerals and metals: recent advancesin geomicrobiology derived from synchrotron-basedX-ray spectroscopy and X-ray microscopy. AnnualReviews in Earth and Planetary Science, 37,367–391.

Thomen, A., Robert, F. & Remusat, L. 2014. Determina-tion of the nitrogen abundance in organic materials byNanoSIMS quantitative imaging. Journal of AnalyticalAtomic Spectrometry, 29, 512–519.

Tice, M.M., Bostick, B.C. & Lowe, D.R. 2004. Thermalhistory of the 3.5–3.2 Ga Onverwacht and Fig TreeGroups, Barberton greenstone belt, South Africa,inferred by Raman microspectroscopy of carbonaceousmaterial. Geology, 32, 37–40.

Vandenbroucke, M. & Largeau, C. 2007. Kerogen ori-gin, evolution and structure. Organic Geochemistry,38, 719–833.

Wacey, D. 2014. In situ morphologic, elemental andisotopic analysis of Archean life. In: Dilek, Y. &

Furnes, H. (eds) Evolution of Archean Crust andEarly Life. Modern Approaches in Solid Earth Sci-ences, 7. Springer, Berlin, 351–365.

Wacey, D., Kilburn, M.R., Saunders, M., Cliff, J. &Brasier, M.D. 2011. Microfossils of sulfur metaboliz-ing cells in �3.4 billion year old rocks of Western Aus-tralia. Nature Geoscience, 4, 698–702.

Wacey, D., Menon, S. et al. 2012. Taphonomy of veryancient microfossils from the �3400 Ma StrelleyPool Formation and �1900 Ma Gunflint Formation:new insights using focused ion beam. PrecambrianResearch, 220–221, 234–250.

Wacey, D., McLoughlin, N. et al. 2013. Nano-scaleanalysis reveals differential heterotrophic consumptionin the �1.9 Ga Gunflint Chert. Proceedings of theNational Academy of Sciences of the United States ofAmerica, 110, 8020–8024.

Wacey, D., Saunders, M. et al. 2014. Enhanced cellularpreservation by clay minerals in 1 billion-year-oldlakes. Nature Scientific Reports, 4, 5841, http://doi.org/10.1038/srep05841

Wacey, D., Brasier, M. et al. 2016. Contrasting micro-fossil preservation and lake chemistries within the1200-1000 Ma Torridonian Supergroup of NW Scot-land. In: Brasier, A.T., McIlroy, D. & McLough-

lin, N. (eds) Earth System Evolution and Early Life:a Celebration of the Work of Martin Brasier. Geolog-ical Society, London, Special Publications, 448,http://doi.org/10.1144/SP448.6 Q12

Westall, F., de Ronde, C.E.J., Southam, G., Grassi-

neau, N., Colas, M., Cockell, C. & Lammer, H.2006. Implications of a 3.472–3.333 Gyr-old subaerialmicrobial mat from the Barberton greenstone belt,South Africa for the UV environmental conditions onthe early Earth. Philosophical Transactions of theRoyal Society B, 361, 1857–1875.

Williams, D.B. & Carter, C.B. 2009. Transmission Elec-tron Microscopy: a Textbook for Materials Science.2nd edn. Springer Science and Business Media,New York.

Williford, K.H., Ushikubo, T., Schopf, J.W., Lepot,K., Kitajima, K. & Valley, J.W. 2013. Preservationand detection of microstructural and taxonomiccorrelations in the carbon isotopic compositions ofindividual Precambrian microfossils. Geochimica etCosmochimica Acta, 104, 165–182.

Wirth, R. 2009. Focused ion beam (FIB) combined withSEM and TEM: advanced analytical tools for studiesof chemical composition, microstructure and crystalstructure in geomaterials on a nanometre scale. Chem-ical Geology, 261, 217–229.

Wopenka, B. & Pasteris, J.D. 1993. Structural character-ization of kerogens to granulite-facies graphite: appli-cability of Raman microprobe spectroscopy. AmericanMineralogist, 78, 533–557.

Yin, L. & Li, Z. 1978. Precambrian microfloras of south-west China with reference to their stratigraphic signifi-cance. Memoir Nanjing Institute of Geology andPalaeontology, Academia Sinica, 10, 41–108.

Young, R.J. & Moore, M.V. 2005. Dual-beam (FIB-SEM) systems. In: Gianuzzi, L.A. & Stevie, F.A.(eds) Introduction to Focused Ion Beams. Instrumenta-tion, Theory, Techniques and Practice. Springer Sci-ence and Business Media, New York, 246–268.

ANALYTICAL TECHNIQUES FOR PROTEROZOIC MICROFOSSILS