Focussed ion beam preparation and in situ nanoscopic study of Precambrian acritarchs

Precambrian Research 140 (2005) 36–54

Focussed ion beam preparation and in situ nanoscopic studyof Precambrian acritarchs

Andre Kempea,1, Richard Wirthb, Wladyslaw Altermannc,∗,Robert W. Starka, J. William Schopfd, Wolfgang M. Heckla

a Department fur Geo- und Umweltwissenschaften, Sektion Kristallographie, Ludwig-Maximilians-Universitat Munchen,Theresienstr. 41, D-80333 Munchen, Germany

b GeoForschungsZentrum, Sektion 4.1, Telegrafenberg, D-14473 Potsdam, Germanyc Centre Biophysique Moleculaire (CBM), CNRS (Centre National de la Recherche Scientifique),

Rue Charles-Sadron, 45071 Orleans Cedex 2, Franced Department of Earth and Space Sciences, Institute of Geophysics and Planetary Physics,

University of California, Los Angeles, CA 90095-1567, USA

Received 10 December 2004; received in revised form 4 July 2005; accepted 14 July 2005

Abstract

The taphonomic nanostructure of acritarch cell walls from the c. 650 million years old Chichkan Formation was studiedwith optical microscopy (OM), Raman spectroscopy, scanning electron microscopy (SEM), atomic force microscopy (AFM),and transmission electron microscopy (TEM). The integration of high-resolution methods and classical optical microscopy

sil. Par-celln of sec-, AFMkerogen

anome-eneitiestion, bute cell.y in theinated

allows for the assessment of the relationship of the fossil to the embedding chert and of the authenticity of the fostial etching of paleontological (150�m) and petrographic (30�m) thin sections rather than maceration of the semi-stablestructures served as the preparation method suitable for AFM and SEM studies. Focussed Ion Beam (FIB) preparatiotions normal to cell walls, yielded stable thin foils of the same thin sections and microfossils as investigated by OMand SEM. Unicells that appeared excellently preserved by optical microscopy standards, consisted of disconnectedparticles, dispersed in the cryptocrystalline quartz matrix and arranged in stacks of variable spacing on micro- to nter scale. The density of carbon particles was found to be correlative to the stability of cell walls and to inhomogin the chert. SEM and AFM images of cell cross-sections are directly comparable at the same scale of magnificaAFM offers higher resolution possibilities and 3-D information on the arrangement of particulate carbon within thWhereas the microscopic appearance of cells was highly variable within the same rock unit, from the same localitChichkan Formation, the nanoscopic structure of kerogen was found to be similar in all cells, consisting of multi-lam

∗ Corresponding author. Present address: Geology, Geo- and Environmental Sciences, Luisenstr. 37, D-80333 Munich, Germany.Tel.: +49 89 2180 6552; fax: +49 89 2180 6514.

E-mail addresses: [email protected] (A. Kempe), [email protected] (W. Altermann).1 Fax: +49 89 2180 4334.

0301-9268/$ – see front matter © 2005 Elsevier B.V. All rights reserved.doi:10.1016/j.precamres.2005.07.002

A. Kempe et al. / Precambrian Research 140 (2005) 36–54 37

(sheeted) amorphous carbon films, built up of layers with measured thickness between 10 and 20 nm, as revealed by TEM andAFM.© 2005 Elsevier B.V. All rights reserved.

Keywords: Atomic force microscopy (AFM); Transmission electron microscopy (TEM); Focussed Ion Beam (FIB); Scanning electronmicroscopy (SEM); Acritarchs; Precambrian microfossils; Petrifaction; Microfossil micro-structure

1. Introduction

In paleontology, the micro-structure of fossil tissueshas frequently been used for the characterization oforganisms, transmission electron microscopy (TEM)being the prominent tool for viewing objects in cross-section on the nanometer scale. For the preparation ofan ultra thin section of a cell, the carbonaceous materialis usually isolated by chemical dissolution of the rockmatrix, and extraction of the organic residue (Hemsleyand Glasspool, 1999; Jones and Rowe, 1999a;Loeblich, 1970; Loeblich and Drugg, 1968; Loeblichand Tappan, 1969, 1971; Talyzina, 2000; Talyzina andMoczydlowska, 2000; Talyzina et al., 2000; Wellmanand Axe, 1999; Yang et al., 1998). Large fossils, likeplant leaves, can directly be grasped and treated as awhole (Jones and Rowe, 1999b; Osborn et al., 2000;Taylor, 1999), whereas small objects like microbes orpollen and spores have to be embedded in a solid resinbefore thinning (Jones and Rowe, 1999c). For example,in carbonates and cherts the palynomorphs and unicells

eter scale (Kempe, 2003). Cell walls are often dam-aged beyond recognition by diagenetic processes. Thismakes very old fossils difficult to recognize unam-biguously and has led to controversy on the authen-ticity of the world’s oldest fossils (Brasier et al., 2002;Kazmierczak and Kremer, 2002; Schopf, 1993, 2004;Schopf et al., 2002a). Because of the discontinuouspreservation, Precambrian microfossils tend to fallapart when the supporting quartz matrix is dissolved.Many objects that are well preserved by Precambrianstandards were thus excluded from close examinationby TEM in the past. A new integrative preparation musttherefore create a stable section of the organic fossilwithin its mineralic host, in order to allow detailedinvestigation of the kerogenous remnants and their rela-tionship to the embedding rock matrix.

Yet another constraint for a new preparation method,besides integrativeness and stability, is the possibilityof applying several different methods of morphologicalanalysis to the same specimen at all scales of obser-vation. This is absolute prerequisite for a nanoscale

ec-yd

eva-sed

are usually isolated by maceration of the rock, thensieved or concentrated by centrifugation (Williamsonet al., 1999) and thus, removed from their original siteof preservation and separated from the large portionof organic residue. This separation, due to technicalnecessities in preparation for TEM, causes a significant

study of poorly preserved microbes by scanning eltron microscopy (SEM) or atomic force microscop(AFM) (e.g. House et al., 2000; Kazmierczak anAltermann, 2002; Kempe et al., 2002; Kudryavtset al., 2000), because the examination of carbonceous fragments at high resolution needs to be ba

loss of information and is a potential source of con-tamination. In palynology, it would be most desirableto study pollen in the context of the associated plant( yo fer-a theh aleo-b

thep cro-f howr notp om-

on a reliable identification of the biological struc-tures. In Precambrian micropaleontology, fossil recog-nition is mainly achieved by optical microscopy ands ion,b thea ion-s 1T icalc oidsa ,1 dert for-m d on

Guignard et al., 1998; Harley, 1997) and in the studf microbial life single celled organisms are prebly analyzed integratively within the context ofosting stromatolite structure and the preserved piocoenosis.

Another critical obstacle in fossil preparation isoor preservation, especially of Precambrian mi

ossils. Even cells that appear well preserved and sich detail under the optical microscope usually doossess a contiguous wall structure on the micr

tatistically representative morphological descripteing the most important method that allows forssertion of the fossil and embedding rock relathips (Buick, 1991; Schopf, 1992a; Altermann, 200).hree informal classes equivalent to the biologlasses of prokaryotic filaments, prokaryotic coccnd eukaryotic cells have been established (Schopf992a,b), based chiefly on inspection of cells un

he optical microscope. At least 279 taxonomic inal species of the Precambrian are known, base

38 A. Kempe et al. / Precambrian Research 140 (2005) 36–54

close similarities to extant unicellular organisms. Qual-ity of preservation ranges from very thick, dark brownand black, contiguous carbonaceous fossil walls thatare often multi-layered and show internal structure, tofaint shades of carbonaceous material that represent the‘ghosts’ of destroyed and mineral-replaced fossil walls.

In this paper, we present new methods for theanalysis of fossil unicells on the nanometer scale,within their original embedding rock context, suitablefor a combined study with optical microscopy (OM),Raman spectroscopy, SEM, AFM and TEM. We haveapplied Focussed Ion Beam (FIB) preparation to chert-permineralized Precambrian acritarchs, producing a∼8�m× 20�m× 100 nm rock foil containing a sec-tion of the fossil wall for TEM imaging, directly froman optical thin section. After the extraction of the foil,the fossil unicell is still left in the original embeddingrock in the thin section. A similar approach has beenused bySchopf and Oehler (1976)andOehler (1977),but seems to have been discarded. The method, allowsfor direct comparison of the results of different tech-niques, applied to the very same individual microfossil.

2. Materials and methods

2.1. Specimens

The fossil acritarchs (leiospheres) studied hereare permineralized in carbonaceous cherty stro-mo auM RGS lh calmt lublec eds ga thers

2

2mm-

d ace( on-

ics, Jersey City, NJ, USA);∼1 cm× 1 cm rectanglesenclosing the circled areas were cut from the sections;the cementing medium underlying these rectangles wasdissolved in acetone; and the fossil-bearing rock sliceswere re-cemented onto microscope slides. Individualmicroscopic fossils, situated close to the uppermost sur-faces of paleontological thin sections,∼300�m thick,were located and photographed in transmitted whitelight with a digital video camera connected to a ZeissAxiovert inverted microscope, then marked on the backside of the glass carrier with a diamond scribe. In a care-ful step-by-step process, controlled under an opticalmicroscope, the section surface was then hand-groundwith corundum powder and water on a glass plate,eroding roughly half the fossil and exposing the car-bonaceous material of the cell walls on the surface ofthe section (Fig. 1A). Surfaces were hand polished withdiamond powder and ethanol lubricant on a swivel disc,in successive steps with decreasing grain sizes of 6, 3, 1and 0.5�m, in order to reduce irregularities in the rocksurface to a minimum. Thus, later chemical preparationof the section surface was dependent only on the mate-rial contrast of the quartz in the rock and the kerogen incell walls. The sections were submerged for 20–50 minin fluid HF 5% with the surfaces upside-down so as tolet detached particles sink to the bottom of the acidbath. An optimum exposure time of 30 min could beverified for most fossils. Cells were photographed inreflected white light with the above-mentioned opticalsystem, acquiring search-maps and reference imagesf ur-f

2rst

s veda thins inga ionb ionb 0 kV.P bys ber.F d att pre-c allw iredA he

atolites (Conophyton gaubiza) of the ∼650 Mald Chichkan Formation from the Maly Karatountains, Ayusakan, of southern Kazakhstan (PPample #1473;Moore and Schopf, 1992). Severaundred individual fossils were located by optiicroscope in paleontological thin sections (∼150�m

hick), cemented onto glass slides by acetone-soement. Intact, relatively thick-walled, well-preservpheroidal acritarchs >30�m in diameter, representinbout 2% of detected fossils, were selected for furtudy.

.2. Preparation

.2.1. Preparation for AFMFossil-containing areas were demarcated by 3-

iameter circles drilled into the thin section surfSonatron sonic disintegrator drill, Kenyon Electr

or AFM data. Subsequently, AFM studies of the sace structure were conducted.

.2.2. Preparation for SEM and TEMAfter AFM examination, specimens were fi

putter-coated with a very thin gold layer that sers a trace line of the etched surface structure. Theection was then coated with carbon for SEM imagnd to avoid charging of the specimen under theeam. TEM foil preparation occurred in a focusedeam device by a Ga-ion beam accelerated to 3reparation of ultra-thin foils by FIB is controlledecondary electron imaging in the preparation chamor this application, the fossil wall must be expose

he surface of the thin section and marked. Theise location for cutting a TEM foil across the cell was determined by comparing the previously acquFM image with the SE image obtained in the FIB. T


Fig. 1. Preparation of cell walls for AFM and TEM. (A) Sketch of a fossil cell with exposed surface after grinding and etching; the red part ofthe fossil wall represents the location of a TEM foil that was cut normal to cell wall. (B) SE image of a FIB prepared TEM foil typical for cell3, before extraction from the thin section. The double wall structure appears as an etched prominence on the section surface (box) and as rowsof bright pores along the rim of the cell visible in the foil. (C) Magnified SEM image of the boxed area in B. (D) Optical micrograph of cell 2before FIB preparation. (E) Optical micrograph of cell 2 after FIB preparation; positions of removed cell wall (black) and TEM foil (white) areindicated graphically.

oil-free vacuum system of the FIB guarantees that thespecimen will not be contaminated by carbon duringthe preparation process. The surface of the foil needsto be protected from being sputtered by the ion beamby depositing a thin Pt-layer on top of the surface. Pt-deposition occurs by decomposing an organic Pt-gasin the ion beam. Details of the FIB milling processare given inWirth (2004)and references therein. The

TEM ready foil measures 15–20�m by up to 10�mand is about 150 nm thick. The foil is cut free at itsbase and on both sides. It is lifted out of the excava-tion site using an optical microscope. A glass fiber isattached to the foil as a special manipulator. Adhesiveforces will keep the foil at the fiber’s tip, so that it canbe lifted and placed onto a TEM grid with a perforatedcarbon film as support for the foil. No further carbon


coating of the specimen is required. A final control ofthe cutting position and location of the TEM foil wasachieved by optical microscopy in transmitted light.Cell 2 is shown before FIB preparation inFig. 1D andafter the TEM foil has been extracted inFig. 1E, wherethe former location of the cell wall and the position ofthe TEM foil are highlighted.

The perforated polymer film on the TEM coppergrid is electron transparent but gives rise to a weakmass absorption contrast, which is superimposed onthe contrast of the object under investigation. However,the supporting film is perforated, which always allowsto find regions where the sample is not overlapped bythe film.

2.2.3. Preparation for OMBecause thick paleontological sections limit pet-

rographic microscopy, some selected specimens wereprepared for petrographic investigations, subsequentlyto AFM and SEM investigations. In order to turn thethick sections into petrographic (30�m) thin sections,the sections were glued with the rock surface on anotherglass carrier, with acetone resistant glue. To avoidsticking of the old glass carriers to the new glass car-riers, the old ones were covered with a thin coat ofVaseline.

After 24 h drying time, the rock sections, now sand-wiched between the old and new glass carriers, weredipped in acetone to remove the old glass carriers, aftera few minutes. Thereafter, the sections were groundd ump f thet ark-i e thei ot ber tingt ingg ondm .

2

byRe -R eD a)0 cro-

Raman, and line-scan Raman imaging capabilities. Allthree gratings of the system (each having a groove den-sity of 1200 grooves mm−1) were holographic, and thespectrograph was typically set in a subtractive config-uration. Due to the confocal capability of the system,use of a 100× objective (having an extended workingdistance of 3.4 mm, a numerical aperture of 0.8, andnot requiring immersion oil) provided a planar reso-lution of <1�m and, by use of a confocal hole sizeof 150–200�m, a vertical resolution of 1–3�m. Acoherent krypton ion laser equipped with appropriateoptics provided laser wavelengths ranging from blueto infrared, of which 476 nm was used here to acquirespectra centered at 1400 cm−1 and extending from 800to 1960 cm−1. The typical laser power was <8 mWover a∼1�m spot, an instrumental configuration wellbelow the threshold resulting in radiation damage tofossil specimens such as those analyzed (Schopf et al.,2002b). Spectra of the carbonaceous walls of acritarchswere acquired from chert-embedded specimens situ-ated <10�m beneath the surface of a thin section andcentered in the path of the laser beam projected throughan Olympus BX40 microscope. To enhance the opticalimage of specimens in unpolished thin sections, thearea analyzed was covered by a thin veneer (∼1�mthick) of Type B non-drying microscopy immersionoil (R.P. Cargille Laboratories Inc.), the presence ofwhich has been shown to have no appreciable effecton the Raman spectra acquired (Schopf et al., 2002b).All spectra shown (Fig. 3) are normalized to the corre-s . Noe d int

2

ea-s o,M 3-p e of1 am nt . Thei 300p it-t andS atel,S odes

own manually on a glass plate, with 800 corundowder and under constant microscopic control o

hickness. In sections processed in this way, mng by pen was avoided, because under acetonnk penetrates and soaks into the rock and cannemoved completely from the rock surface, obliterahe microfossils. To facilitate the observation durrinding, the fossil location was marked by a diamark on the reverse side of the new glass carriers

.3. Raman spectroscopy

Prior to study by AFM, all fossils were analyzedaman spectroscopy using procedures ofKudryavtsevt al. (2000)and Schopf et al. (2002b). The laseraman spectra presented inFig. 3were obtained by thilor XY (formerly, Instruments S.A.; now JY Horib.8 m triple-stage system with macro-Raman, mi

ponding spectral response function of the systemvidence of the effects of polarization was detectehe spectra of any of the specimens analyzed.

.4. Atomic force microscopy

The surface structure of etched fossils was mured by use of a Topometrix ‘Explorer’ AFM (Veecannheim, Germany) which is equipped with aiezo scanner having a maximum planar rang30�m in each of the planar (x, y) directions andaximum vertical (z) range of 10.5�m, mounted o

he video camera-connected inverse microscopemages were acquired at a digital resolution ofixels per line in each image, applying the interm

ent contact mode (using NCHRW-reflex-coatedSS-NCH cantilevers from Nanosensors, Neuchwitzerland), selected among various scanning m


tested as resulting in the least disruption of the scannedsurface.

The surface structure of Cell 2 (Fig. 6F) was imagedwith a ‘Dimension’ (Veeco) closed loop AFM intappingTM mode, at a digital resolution of 512 linesper image and 1024 pixels per line, using Nanosensors’SSS-NCH cantilevers.

2.5. Transmission electron microscopy

TEM was performed in a Philips ‘CM200’ trans-mission electron microscope operating at 200 kV andequipped with a LaB6 electron source. The TEM isequipped with a Gatan Imagin Filter GIFTM and allowsenergy filtered imaging. Plasmon imaging using elas-tically scattered electrons applying a 20 eV windowto the plasmon peak in the electron energy loss spec-trum (EELS) allows discrimination between holes inthe specimen and areas with very small sample thick-ness. In both cases in TEM bright field images, avery bright contrast would be visible. However, inplasmon images holes appear black, because in vac-uum no electrons from the primary beam are scatteredinelastically.

The carbon concentration in the TEM foils was ana-lyzed by analytical electron microscopy (AEM) usingthe C–K X-ray fluorescence. The same signal can beused for carbon element mapping. AEM was performedwith an EDAX X-ray analyzer. The spot size was about4 nm and the dwell-time for each position was 45 ms.F itha s theX areas ap-p eptf sityw on-s eakb hichd butt en-t pec-t mpr

cen-t thins odeF an

Oxford Instruments EDX spectrometer, with a spatialresolution of∼1�m3.

3. Results

3.1. Preservation

Optical microscopy revealed that all cells selectedfor investigation were generally well preserved by opti-cal standards, with the organic carbon in the cell bodiesof dark yellow to brown color and in most cases appear-ing as continuous structures (Fig. 2). Nevertheless, cellsdisplayed various degrees of detail-preservation: Someof the acritarchs are slightly deflated and have thick, butdiffuse walls. Others are perfectly round. An exception-ally large and well-preserved example of an acritarch,with a thick wall is shown in petrographic thin sectionin Fig. 2. It has a clear, dark brown cell wall with brownhalo inside and a narrower radiance zone outside thecell. The cell center is very clear. The cell is locatedat the edge of a very fine chert nest, rich in carbon.Within the cell, however, quartz crystallites are ratherlarge, 10–20�m across and of flame chalcedony type,with sharp boundaries (Fig. 2B).

Details of the cell wall are visible in a digital blow-up in Fig. 2C and D. The cell wall has indistinct,slightly diffuse boundaries. Many small “pearl-chain”-like, regular, carbonaceous structures are located onthe inner cell wall and protrude into the cell or area thso entso owsi d inp el es.C pt areb out5 g-u

of av nets saictl llerp fossil

or the acquisition of an elemental map, a window wcertain energy range is selected, which include-ray fluorescence line of interest. The integratedelected by the window defines the signal used for ming. This signifies that, even if there is no signal exc

or the background intensity, this background intenill create a weak signal during the mapping. Cequently, the elemental maps always exhibit a wackground distribution of the selected element, woes not reflect the real distribution of the element

he intensity from the background. Additional elemal maps were acquired by electron energy loss sroscopy (EELS) (three window method and/or juatio imaging) using the C–K edge (Fig. 7B).

SEM and point spectroscopy of the carbon conration on the etched surface on the petrographicections was performed with a Hitachi Cold Cathield Emission microscope that is equipped with

rranged parallel with the cell wall. At different depf focus the spheroidal nature of the single elemf the chains, arranged in clusters, is visible (arr

n Fig. 2E). These structures were only observeetrographic thin sections of 30�m thickness, wher

ight refraction is low enough for good optical imagarbonaceous chains are∼1�m in diameter and u

o about 10�m long and the spheroidal elementselow 1�m in diameter, arranged in clusters of ab�m, and coating the inner wall of the cell with a relar, honeycomb-like layer.

Some cells studied are preserved at the edgeery finely crystalline chert, within a transition zoo a somewhat coarser chert (c. 1�m). Most of theections are of equant quartz (equally large, moextured recrystallized chert) of about 10�m crystal-ite size, with sharp crystal boundaries. Only smaatches are preserved as fine chert, mostly in the


Fig. 2. Acritarch cell 4 in a petrographic, 30�m thin section. (A) Cross-section through the cell displaying a relatively sharp, distinct cell wall(depending on the depth of focus) and a halo of kerogen inside and outside the cell wall. (B) The same cell under crossed Nichols, showingextinction properties of the cell-permineralizing quartz. In the lower right corner the chert is very finely crystalline. The other parts of the imagedisplay a coarser, recrystallized chert and a network of fine crystallites along the coarse quartz grain boundaries, where kerogen is concentratedas well. The cell wall itself is permineralized by finely crystalline chert, but coarser chert crystals penetrate it from the outside and chalcedonicquartz fills the cell interior, protruding into the inner kerogenous halo. (C–E) digital blow-ups of the cell wall parts as indicated by rectangles in(A), but taken at varying depths of focus. In (C) and (D) (arrows), the pearl-like chains of kerogenous, spheroidal structures are clearly visible.In (E), strings of such spheroids (arrows) and honeycomb-like clusters (in the background) are clearly discernible. Such preservational detailswithin the embedding rock are only visible by optical microscopy as SEM and AFM measure only the surface of the sample, and TEM has notenough penetration potential and transmits only nanometer-thick samples.


vicinity and in areas where kerogen is concentrated.Within most cells, like in the example described above(Fig. 2), the crystal size was found to be markedlylarger, recrystallized to above 1–5�m, than within andoutside the cell walls, where it is below 1�m, if notrecrystallized. In some cells, the center is filled withequant, recrystallized flame chalcedony. Carbon is notclearly discernable in the walls by optical microscopy,because its particles are below the resolution ability ofthe lens. Cells often also appear to contain Fe oxidesor hydroxides.

TEM investigations in the direct vicinity of the wallstructures and enclosing sheath confirmed the differ-ence in grain size of quartz grains inside and outsidethe cells, being especially distinct where the carbona-ceous fossil wall was more intact. It also showed thatfine grains are typically linked to a high porosity of thechert, which was found to be especially dominant ina ∼5�m wide margin around the cell wall. A typicalexample of this type of differential mineralization ispresented inFig. 5A, where the SiO2-filling of cell 1consists of larger polycrystalline quartz with diffrac-tion contrasts visible (dark bands in the grains). Thechert outside the carbonaceous cell wall is composed ofnano- to microcrystalline quartz grains that are riddledwith cavities and organic carbon. The same differen-tiation was less pronounced in cells with a less intactwall structure (e.g. cell 2,Fig. 6A) and totally absentin poorly preserved fossils, as in cell 3 where a persis-tent nano- to microcrystalline quartz texture permeatest e-n entd theg

s ap owt low0 aleb om-b enti ostsc rals,n onsw inq clu-s tincth bout5 on-

stitute the cell wall. Where recrystallization is moreadvanced, this kerogenous brown “fog” is pushed awayalong the crystallization front and the quartz is clear,while the kerogen is concentrated along, and coats orwraps around quartz crystal boundaries. The chert isclearly finest where richest in kerogen.

3.2. Stability/integrity of cell walls

For 32 analyzed specimens of acritarchs from thesame rock unit and locality in the Chichkan Forma-tion, three different grades of preservation rangingfrom intact to semi-intact, to disrupted, were defined.Best preservation was represented by cells with intactcarbonaceous cellular structures and high chemicalresistance to hydrofluoric acid, poorest preservation bycellular structures which were non-contiguous on themicrometer scale and very unstable upon HF expo-sure. Three typical unicells with an organic doublewall structure of declining grade of stability upon HF-exposure are presented inFigs. 5–7; best preservationbeing represented by cell 1 (Fig. 5) which shows adistinct complete cell wall, poor preservation beingrepresented by cell 3 (Fig. 7) displaying a very thininterrupted wall structure, and intermediate preserva-tion in cell 2 (Fig. 6), where the semi-contiguous wallshowed only small gaps, filled by quartz crystals. TEMfoils transecting the cell walls of all qualities of preser-vation could be acquired without exception by the FIBtechnique.

3

acho ert-e ectraa ci-m tp ssilsaa m( rac-t tedR ons,“ l.,2 al.,2 d, at∼ ing

he thin wall. Fig. 7 displays this type of homogeous permineralization with quartz grains of differiffraction contrast, due to random orientation ofrains.

The kerogen throughout the sections displayatchy distribution and is very fine, mostly far bel

he resolution ability of the optical microscope (be.1�m), giving the sections a dark brown and prown diffuse appearance. Small, rare ghosts of a rhic mineral (dolomite?) replaced by silica are pres

n parts of some sections. Rare cubic mineral ghould represent former diagenetic evaporite mineow thoroughly silicified. Several cells have inclusiithin the cell interior. Some are fluid inclusionsuartz (pale and transparent) and some mineral inions, Fe-rich and dark. Some cells display a disalo of kerogen around them, which can be up to a0% of the width of the cell, but clearly does not c

.3. Carbonaceous structure on micrometer scale

Raman spectra of the kerogenous cell walls of ef the 32 acritarchs studied were acquired from chmbedded specimens analyzed in situ. Typical spre shown inFig. 3, acquired from each of the speens illustrated here (Figs. 1, 2, 5–7). The two mosrominent features of the Raman spectra of the fonalyzed are a band of high intensity at∼1600 cm−1

nd a group of bands centered at 1300–1350 c−1

Fig. 3). Both of these spectral features are chaeristic of the carefully studied and well-documenaman spectra of polycyclic aromatic hydrocarb

PAH’s” ( Mapelli et al., 1999a, 1999b; Rigolio et a001; Castiglioni et al., 2001a, 2001b; Negri et002). In such spectra, the most intense Raman ban1600 cm−1, is ascribed to synchronous aromatic r


Fig. 3. Raman spectra taken from the walls of individual cells 1–4.

stretching whereas the bands in the 1200–1400 cm−1

range are attributed to modes of aromatic ring defor-mation and totally symmetric breathing (Mapelli etal., 1999b). The spectra acquired from the Chichkanacritarchs (Fig. 3) are thus consistent with a large bodyof data establishing that fossil kerogens are composedpredominantly of more or less regularly stacked arraysof interlinked PAH’s (Durand et al., 1977; Durand,1980; Faulon et al., 1989; van Krevelen, 1993;Vandebroucke, 2003). More detailed Raman analysesshow the geochemically moderately mature Chichkankerogen to be exceptionally little altered in comparisonwith the kerogenous components of more than a scoreof other fossiliferous Precambrian deposits (Schopfet al., 2005).

The carbonaceous wall structures exposed on therock surface by HF-etching were imaged with AFM,later combined with SEM/EDX analysis. On the petro-graphic thin sections, elemental mapping for C, Si andFe was performed. Fe mapping revealed very low-Fepresence, not reflecting the shape of the cell, but onlyan irregular, patchy Fe distribution. Also the attempt tomap for Al, Mg, Ca and K did not reveal any elementdistribution patterns related to cell morphology. Obvi-ously carbonates, phyllosilicates and feldspar do notplay any role in the fossilization process of these cells.It appears, however, that Ca and Mg slightly reflectsome recrystallization shapes of quartz.

In all cases of C-mapping, the area maps showed ad vedc l.

The same can be seen, but less clearly, in Si distribu-tion maps, where C substitutes for Si. From the pointspectra taken on elevated structures within the cell walland on elevated particles in the direct outer vicinityof the cell wall (comp.Fig. 6G), it can be seen thatthe C-peak is relatively low in comparison to the Si-peak (and O-peak). However, with increased resolutionand magnification, in spot measurements, the C-peakrises significantly. Spot mapping thus confirmed thatthe parts of the cell wall that form elevated structuresafter HF preparation consist of carbon.

Although all of the 32 specimens possessed anorganic wall, that was optically coherent over largestretches, only 9 showed fossil walls that were stableat acid exposure. In very well preserved cells (darkunder optical microscope and of dense carbon distribu-tion in TEM section) HF-etching produced walls rising200 nm above the rock surface that could be imagedwith the AFM already after 20 min of acid expo-sure. These cell walls grew as high as∼1000 nm after30–40 min of exposure (e.g.Fig. 4K and L). Longeretching times of up to 70 min resulted in 2000 nmhigh cell walls and large cavities in the centers ofthe cells, where pure quartz, recrystallised after flamechalcedony, was present. Large holes proved to bevery inconvenient for AFM imaging, thus defining theupper limit of etching time. Optically thin-walled fos-sils showed less differential etching contrast, producinga flat relief that was best detectable by AFM when thepolished surface was only slightly eroded by etching( g,t d sot

ingw s of

F ea-s catedb

istribution correlating with the shape of the preserell or with the SiO2 crystallite along the cell wal

e.g.Fig. 7G and H). After more than 30 min etchinhe whole structure became increasingly degradehat material difference became unrecognizable.

The different grades of stability versus HF-etchere documented by AFM in the three example

ig. 4. Topography of fossil walls after 20 min HF-exposure, mured by AFM. Positions of the corresponding transects are indiy blue lines inFigs. 5L, 6F and 7H.


cells 1–3.Fig. 4 displays profiles through the etchedwalls of cells 1–3 (along the blue trajectories inFigs. 5L, 6F and 7H), after 20 min of etching, showingthat good quality of preservation corresponds to tallwall structures in cell 1, with lower topography in cells2 and 3.

Along with height, the compactness of the fossilstructures increased with preservation quality. In theAFM images of cell 1 (Fig. 5K and L), a very soliduninterrupted fossil wall prevails in about half the cir-cumference of the whole cell, the inner wall beingmuch more pronounced than the outer structure. Cell

F of a p whitea unifor representsc ht part , amorphousc DX), c along thebficteagss(oomw

ig. 5. Structural analysis of cell 1. (A) TEM bright field imagereas represent holes in the foil or very thin carbon layer. Theoarse-grained quartz. The patchy-appearing contrast in the rigarbon (light grey) and holes (white). (B) C-elemental map (E
oundary of the coarse-grained quartz and the fine-grained quartz. (C
ossil boundary shown in (A) and (B). Coarse SiO2 in dark grey and fine Sin black. (D) TEM bright field image of the area indicated by a box inarbon, and adhering to the overhead quartz grain at label f and spo (D). (F) Energy filtered plasmon image from the area labeled f inntirely consisting of amorphous carbon. (G) Plasmon image from tmorphous carbon material. (H) Electron energy loss spectrum reprrey; background fit in dark grey; spectrum after subtraction of backghowing high concentration of carbon in the center and along the oucale sketch of the fossil wall structure, superimposing the dark browK), represented by the dashed line. (K) AFM image of the leiospherf high carbon concentration in the double wall and center. Image takf the boxed area in (K). The blue line indicates the transect througeasured. (M) Magnified AFM image, corresponding to the box in (L)all structure.

art of the TEM foil cut normal to the cell wall. Bright to nearlym dark grey contrasting material in the left part of the imageof the image is due to fine-grained quartz (darker grey or black)orresponding to (A). It shows an increased amount of carbon

) Sketch of the main components in the carbonaceous structure along theO2 in light grey, Corg. in red, Pt-cover on surface from FIB preparation(A), showing a wavy, string-like dark grey film, built up of amorphous

anning over the vacuum at label g. (E) EDX, Carbon map corresponding(D); this section of the film shows a substructure, with regular spacing,he area labeled g in (D). The image shows a three-layer structure in theesentative for all carbonaceous structures analyzed; original spectrum inround as black line. (I) Optical photomicrograph of a well-preserved cell,ter double wall structure, as indicated by the dark brown color. (J) True

n areas in (I), represented by the continuous line, with the etched relief ine after 20 min etch with 5% HF, showing elevated structures at localitiesen after erosion of∼1�m of material from the surface. (L) AFM image

h the cell wall along which the topographic profile shown inFig. 4 was, showing a 3-D view of stacks of carbonaceous components in the inner


Fig. 5. (Continued ).

2 displayed a less homogeneous carbonaceous wall,with spots of high elevation above the etched rocksurface, e.g. the structure in the center ofFig. 6F,but less pronounced wall for most of its expansion.EDX spot analysis (Fig. 6G) demonstrated the car-bonaceous nature of the elevated structures within thewall (green squares), as well as of those elevated struc-tures dispersed in the embedding chert outside thecell, which envelop the inner cell wall within a 5�mthick spherical margin. Cell 3 had the thinnest etcheddouble wall structure of all cells, which appears as astippled half circle in the upper portion ofFig. 7G.Higher magnified topographical maps (Fig. 7H andI) showed the non-continuous nature of this fossilwall.

TEM bright field imaging (Figs. 5A, 6A and 7A)displayed cavities and the carbonaceous content of thecells in light grey to white contrast, contrary to the darkgrey contrast, typical for the micro- to nanocrystallinequartz matrix of the embedding rock. In approximation,all bright parts in the TEM images can thus be regarded

as portions of the fossil. Cell walls are well visibleas round, approximately 1–3�m thick structures. Ele-mental maps (Si–K, C–K) of the fossil walls showedthat quartz components were interfusing the cell wallsand comprised the major part of the whole fossil struc-tures, so that carbonaceous cell walls were discontinu-ous in most cases. The walls of the acritarchs describedhere can thus be regarded as a quartz–kerogen com-posite or as kerogen disrupted by quartz crystallites.The carbon constituent of the composite showed gen-erally sharp boundaries towards the inside of the celland flame-chalcedony. Carbon constituent that reached3–10�m towards the outside of the cell, formed a lesswell-defined outer boundary. In the better-conservedcells, the innermost part of the fossil wall consistedof larger intact pieces of carbonaceous film, whereasthe outward part hosted dissected fragments of sucha carbon film. The same characteristic distribution ofcarbon has been described for eukaryotic unicells fromthe Bitter Springs Formation, also applying TEM stud-ies of whole rock specimens (Oehler, 1977; Schopf and


Oehler, 1976), being interpreted as cell walls and cellsheaths.

Out of the three examples of TEM analysis shownhere, only cell 1 has an intact inner cell wall, theouter wall and surrounding area comprise of discon-tinuous carbonaceous particles (Fig. 5A). The lesswell-preserved cell 2 displayed on a large scale, analmost coherent inner wall structure (Fig. 6A and B)and an exterior carbonaceous wall and carbon halo thatis interfused by larger volumes of quartz. The singleportions of the cell wall, viewed on a smaller scale

consist of flat carbon fragments, 1–3�m in diame-ter, very closely spaced, aligned and interrupted bysmall volumes of quartz crystals. The cell wall car-bon was thus non-coherent in cell 2.Fig. 6C–E depictsingle fragments of carbon, each surrounded and over-lapping with crystalline quartz. In cell 3, all carbona-ceous parts are separated by large volumes of quartzand the inner- and outer cell wall are indistinguishable(Fig. 7A).

The structure of the carbon particles comprising thecell walls was documented in detail in elemental maps

F f a par e areasr grey c (B) SketchoPfaaoiso

ig. 6. Structural analysis of cell 2. (A) TEM bright field image oepresent holes in the foil or very thin material. The patchy dark
f areas with accumulated carbon along the fossil boundary shown it-coat in black. (C) Plasmon image from the area labeled c in (A), s
rom the area labeled d in (A), showing stacks of carbonaceous sheerea labeled e in (A), showing stacks of carbonaceous sheets in planfter 20 min etch, showing revealed carbonaceous structures, where rf ∼500 nm of material from the surface. The blue line indicates the t

n Fig. 4was measured. (G) SEM image of that portion of the fossil wtructures shown in (F) (e.g. arrows); this spectrum was taken at the sf one carbonaceous platelet outside the cell, labeled h in (F).

t of the TEM foil cut normal to the cell wall. Bright or nearly whitontrasting material in the image represents crystalline quartz.

n (A); SiO2 unmarked, Corg. in red, Au-coat in yellow, C-coat in orange,howing stacks of carbonaceous sheets in side view. (D) Plasmon imagets in plan view, embedded in quartz grains. (E) Plasmon image from theview, embedded between quartz grains. (F) AFM image of the leiosphereemoval of quartz matrix has created depressions. Image taken after erosionransect through the cell wall along which the topographic profile shownall labeled g in (F); EDX point spectrum, representative of all excavatedpot labeled with a green circle in (G) and (F). (H) AFM image in 3-D view



obtained by EDX and EELS. In TEM, wall structureswere best traced by carbon elemental mapping (EDX)in very well-conserved fossils, like cell 1. The EDXimage inFig. 5B displays a contiguous carbon struc-ture within the quartz frame filling cell 1 and carbonenrichment in the proximal cavities outside the cell.A more precise assignation of carbon to the cellularstructure with EDX was possible at higher magnifica-tion, as shown in the close-up mapping of one sectionof the inner cell wall, inFig. 5D–E. Visualized at thenanometer scale, the carbonaceous wall appears as anultra-thin film, adhering to quartz crystals of the cell-filling (top) and the embedding quartz matrix (bottom)in some places, but spans empty cavities in other parts(Fig. 5D). The carbon signal shown inFig. 5E clearlymaps the laminar structure. Elemental point spectrafrom the carbonaceous film and crystals were taken onmultiple spots. In less well-preserved fossils, the car-bon signal of smaller cellular fragments was dominated

by the background noise in large survey scans, but wasvisualized for smaller portions of cell wall at highermagnification.Figs. 7C–D shows fragments of car-bonaceous film embedded in microcrystalline quartzframework.

As an overall result, carbon mapping confirms thatcarbon films are partly filling inter-granular cavitiesin the crystalline SiO2-matrix. EELS spectra acquiredwith a spot size of 55 nm show a C–K edge, charac-teristic for amorphous carbon. All carbon structuresfound were amorphous by identification with the estab-lished fingerprint technique (Garvie et al., 1984), whichallows for discrimination between the phases of amor-phous carbon, graphite and diamond. No crystallinecarbon was found.Fig. 5H depicts a typical spectrumof such amorphous carbon. The light grey shaded spec-trum is the raw spectrum. The dark grey-shaded spec-trum is the modeled background and the black curverepresents the background-subtracted spectrum.


3.4. Wall structure on nanometer scale

Whereas the structure of the entire cells was foundto be of variable stability, completeness and coherenceat all grades of preservation, the carbonaceous com-ponent of the fossil walls consisted of similar films ofamorphous multi-sheeted carbon. These films are pre-sented at nanometer resolution in TEM (EELS:Fig. 5Fand G,Fig. 6C–E and as a plasmon image inFig. 7B),where each additional carbon sheet increases the massadsorption contrast in the image. Single fragments ofC-film in cavities between the quartz crystals, in cells2 and 3, showed various shapes and sizes, typical foramorphous sheeted carbon material.

Fig. 6C displays planar stacked carbon sheets of veryregular, rectangular shape at the surface of the polished

and etched thin section.Fig. 6D shows a planar sheetedcomplex at a depth of∼0.5�m below the section sur-face, with amoebae-like forms that vary greatly in shapefor each lamina and which overlap the neighboringquartz grains.Fig. 6E depicts a planar sheeted parti-cle of compact, irregular shape at a depth of∼2�m,also overlapping the surrounding quartz crystals.

In EELS images, the carbonaceous substance is dif-ficult to discriminate from the very tiny quartz, andvery thin carbon sheets are difficult to distinguish fromvacuum, by the optical impression. Therefore, carbonwas mapped in all EELS images and superimposed onthe corresponding bright field images. The outlines ofthe carbonaceous material have been indicated by stip-pled lines in the bright field images inFig. 6D and E.A clear distinction of quartz and carbon was made by

F om a fo tructuresi portion g a 20 eVw aceous lline quartzi ge is th tifying thec corre filled witha map (E regularlyp outer True scales n area line). (G)A e elev the doublewTi

ig. 7. Structural analysis of cell 3. (A) TEM bright field image frn quartz grains together with carbon-rich material. (B) The leftindow to the plasmon peak in the EELS spectrum; a carbon

n dark grey and vacuum in black. The right portion of the imaentral structure as carbonaceous. (C) TEM bright field imagemorphous carbon between quartz crystallites. (D) Elementalreserved cell, showing high concentration of carbon along theketch of the fossil wall structure, superimposing the dark browFM image of the leiosphere after 20 min etch, showing very fin
all structure. Image taken after erosion of∼200 nm of material from thehe blue line indicates the transect through the cell wall along which t
mage, corresponding to the box in (H), showing a 3-D view of the car

il cut normal to the cell wall. The bright areas indicate porous sof the image is an energy filtered TEM plasmon image applyinmulti-layered film is represented by light grey scales, crystae corresponding carbon elemental map (EELS), clearly iden

sponding to the box labeled c in (A), exhibiting “pore”-spaceDX) corresponding to (C). (E) Optical photomicrograph of the

double wall structure, as indicated by the dark brown color. (F)s in (E) (continuous line) with the etched relief in (G) (dashed

ated structures at the localities of high carbon concentration in
surface. (H) AFM image, corresponding to the box labeled h in (G).
he topographic profile shown inFig. 4was measured. (I) Magnified AFMbonaceous components of the outer wall structure.



carbon–K edge mapping presented inFig. 7B, wherethe left portion of the figure is a plasmon mode image(energy filtered), in which crystalline quartz appears indark grey, carbon in bright grey and vacuum in black.The right portion of the image is the corresponding car-bon elemental map with the C–K edge intensity shownin red color. The images display a multi-layered pieceof carbon film that wraps around quartz crystals, form-ing a bifurcation in its lower part, altogether appearingwrapped and folded.

The intact wall of cell 1 was imaged at nanometerresolution, displaying a carbonaceous film in cross-section (Fig. 5F and G). Also here, the C-film is sheeted,each lamina having a thickness of∼10 nm, which isan approximation since the structure is folded. Someinternal structures with larger mass absorption contrastthan in the surrounding film were found in one lamina(Fig. 5F). These structures are arranged perpendicularto the film surface, forming some sort of cross-channelsor supports. The channels have a regular spacing of

∼10 nm. The total width of the film in a section at itsthinnest point is 30 nm.

For carbonaceous flakes that have been imaged intop view, like inFig. 6E, the thickness of carbon sheetswas calculated using the zero loss peak intensity (I0)and the total spectrum (It). From total intensity of thespectrum (It) and the intensity of the zero loss peak(I0) in the equationt/λ = ln (It/I0), the ratio of specimenthickness (t) divided by the electron mean free path (λ)can be calculated (Malis et al., 1988; Egerton, 1989).The calculated electron mean free path,λ, for carbonat an acceleration voltage of 200 kV is about 150 nm.The resulting thickness for the multi-layered particleshown inFig. 6E is 100 nm.

AFM analysis of the surface structure of fossil wallsat nanometer resolution revealed sheeted carbon par-ticles that were found in fossils of all qualities ofpreservation, but did not generally cover the whole wallstructure.Fig. 5M displays a section of the wall of cell1, where an array of stacked carbonaceous particles,


ranging from 100 nm× 400 nm to 400 nm× 1000 nm,prevails. Such an arrangement was also found in cell3 (Fig. 7I). The size of these particles is comparableto the dimensions of the carbonaceous flakes visible inTEM section and has been evaluated statistically for allfossils (Kempe, 2003), showing distributions in thick-ness with peaks at∼20 and∼300 nm. The regularity inspacing, the size distribution and the three dimensionalorientation patterns are striking. In the particular exam-ples of cells 1 and 3, carbon particles seem to standupright and parallel to the cell radius. However, thesection through cell 2 showed carbonaceous particlesin the wall and directly outside the wall, that were notregularly oriented. Most particles are lying flat in thesubstrate in this particular section of the cell, like theparticle displayed inFig. 6H, located within the sheath-ing margin of the cell. At this orientation, the particlesare similar in shape when viewed in AFM and TEM.

The multi-sheeted carbon structure seen in transmis-sion EM has been reported before (Kempe et al., 2002)and referred to as “platelet structure”.Fig. 6H shows athree dimensional view of a single carbonaceous parti-cle (or platelet), imaged at a digital spatial resolution of1 nm and an approximate physical spatial resolution of3–5 nm. In the topographic profile, the same amoebae-like flat shape and the multi-layered composition thatprevails in the TEM images is visualized. The step sizeof single carbonaceous layers was measured, and var-ied between 15 and 30 nm, as visualized in the heightprofile, corresponding to the pale blue line inFig. 6H.

4

m-b m-b edf reed t then re off itum -ut itelyd ils.

ellst er theo able

resistance to hydrofluoric acid. All other cells tended tofall apart upon preparation, during mild and selectiveetching with HF. These 10 stable cells were imaged insection by atomic force microscopy. After etching, eachfossil was protruding to a different degree from the sur-rounding rock surface and the cells that produced verystable high walls had a very solid, compact wall struc-ture. Those cells with low walls consisted of looselydispersed carbon particles embedded in and disruptedby quartz, crystallized during diagenesis and at laterstages of the rock’s history.

Focussed ion beam preparation enabled us to exam-ine microfossils of varying preservation quality, at thenanoscale under TEM and AFM. The method is partic-ularly interesting for the study of Precambrian fossils,where poor preservation can cause a major difficultyin preparation of macerates and where taphonomic andpreservation details are important for the judgement ofthe authenticity of the objects investigated. It is remark-able that the extraction of foils by the FIB techniquewas achieved without losses for very solid cell walls, aswell as for the more fragile fossil walls. The enframingchert showed ideal sputtering characteristics so that flat,unwrapped foils with maximum sizes of 8�m× 20�mand a thickness of∼100 nm were produced for eachcell, whereas etching of the petrographic slides pro-duced walls that fell apart at a maximal wall heightof ∼2�m. FIB thus proved to be an excellent tool forprobing fossil unicells at precisely defined locations.

TEM examination of the thin foils showed thatt icleso e of1 k att ur-r undn . Thes etro-g shedf liza-t apsa n <1a is-i tz asu silso tal-l afterm hortfi like

. Summary and conclusions

AFM imaging of fossils in petrographic slides, coined with SEM and optical microscopy and the coination with TEM imaging of ultra thin foils, prepar

rom the same slides, provide information on the thimensional structure and composition of fossils aanometer to micrometer scale and on the natu

ossilization of the cells. In combination with in sicro-isotopic analysis (House et al., 2000) and moleclar analysis of fossil remnants (Brocks et al., 1999),

he methods introduced herein can help to definecide on the authenticity of the alleged microfoss

In a study of 32 Precambrian spheroidal unichat appeared as excellently preserved fossils undptical microscope, only 10 cells showed consider

he carbonaceous cell structure consists of partf amorphous carbon (kerogen) in the size rang–3�m. Particles are dispersed in the chert roc

he original location of the cell walls and in their soundings, and the carbon apparently wraps aroanometer to micrometer scaled quartz crystallitesame can be observed by optical microscopy of praphic thin sections, where the carbon has been pu

orward along the crystal boundaries, by the crystalion force of the growing quartz crystals, and wrround such crystals which have a size betweend 5�m. This behaviour bears a high risk for m

nterpretation of such carbon wraps around quarnicells, when cherts are investigated for microfosnly by SEM. Such carbon wrapped quartz crys

ites and carbon films embedding quartz, appearild etching as coccoidal, cucamber-shaped or s

lamentous single cell-like structures and colony-


assemblages, in a two dimensional surface investigatedunder the SEM. Various shapes mimicking dividingcells and cell colonies can be produced by etching ofkerogenous chert or by simple breaking along weak-ness fractures, and easily be misinterpreted for micro-fossils, as discussed byAltermann (2001), especially asEDAX carbon mapping will show they carbon contentwithout revealing they quartz crystallite nature. Thebest method to avoid such misinterpretation is previ-ous thorough investigation of thin sections by opticalmicroscopy and identification of microfossils withintheir rock content.

Whereas the structural constitution of cells as awhole was greatly variable within the individual stro-matolites sampled, the nanoscale structure of kerogenwas remarkably homogeneous. By analysis of Ramanspectra, the fossil kerogen was found to show molecularsimilarity to interstellar dust as well as to quenched car-bonaceous composites (QCC) and hydrogenated amor-phous carbon (HAC), which all consist of islands ofaromatic (sp2) bonded C atoms joined together witha variety of peripheral sp2- and sp3-bonded hydrocar-bons (Kwok, 2004). As in these three classes of car-bonaceous material, electron micrography of the fossilkerogen showed a structure, varying from amorphousto electron diffractive.

In all fossil walls as well as in the intact wall in cell 1,the kerogen consisted of multi-layered, amorphous car-bon film, or small fragments of such a film. The outershape of particles varied within one fossil and fromc ed,c wasa telyr eetsc foilp ughc thet thee flati ick-n heetsh essb gree-me letste g-i of

single components, of controlled (SEM, TEM, AFM)locations within the particular cell.

The three dimensional arrangement of kerogen par-ticles within the cell walls was apparent in AFM topo-graphical mapping. Here, flat particles or platelets,arranged in stacked arrays, in the manner of tiles,constitute solid cell walls. Stacks of carbon plateletsappeared dense in those cases, where the spacing inTEM section was also found to be dense, and vice versa.However, the described stackings of carbon plateletswere not continuous over entire cell walls and theywere not found in all cells. These irregularities may bedue to local variances in degradation processes (celllysis) and to (post-) diagenetic alteration.

The spacing of carbon particles (platelets) wasapproximately regular within one fossil but variedgreatly from one unicell to the other. Not surprisingly,those cells with the densest distributions of carbon inthe fossil wall were also most resistant to acid expo-sure. Optically, we consider those fossils with dense,dark carbon to be well preserved. In this sense, cells1–4 were ideally preserved, because they displayeda largely uninterrupted carbonaceous wall. However,TEM and AFM revealed the discontinuous nature ofall but one fossil (all but cell 1), which corresponded tothe varying grades of stability of the cell walls upon HFetching. This structural observation obviously must beused to redefine the term “preservation quality”.

The biological and taphonomic significance of thehere described preservation is yet enigmatic. The mor-p truc-t inga rlyst seso ia-g oft riali slyt yed.M red,r st nara or-p veryw za-t eta-m the

ell to cell, being angular or lobate, flat or wrappompact or spliced, but the multi-layered structurelways found to be laminated with an approximaegular lamina thickness. The thickness of the should be measured directly in the cases where thereparation had produced a normal section throarbon sheets. Another possibility of calculatinghickness of single laminae is the evaluation ofnergy loss spectrum on kerogen particles lying

n the focal plain. Both methods yielded average thess between 10 and 20 nm per carbon sheet. Save also been imaged by AFM, yielding a thicknetween 10 and 30 nm. These results are in good aent with structural data published before (Kempet al., 2002). The carbonaceous nature of the plate

hen deduced by laser Raman spectroscopy (Kempet al., 2002) has now been confirmed with TEM ima

ng, EDAX element mapping and spot analyses

hology and size of the platelets and the internal sure of the carbon films, with sheeted lamina hav

thickness of∼10 nm and with supports regulapaced perpendicularly to the laminae, at∼10 nm dis-ance, is certainly a function of combined procesf post mortem bacterial cell degradation and denesis, but mainly of the metamorphic history

he rock. Reorganization of the genuine cell matento polycyclic aromatic hydrocarbons has obviouaken place, the original biochemistry being destroost of the cell (wall) structures have been alte

esulting in the uniform multi-layered carbon filmhat consist of planar domains of interlinked plaromatic makro-molecules, which are electron amhous. The detected internal ultra structures inell-preserved fossils reflect most likely recrystalli

ion effects during late diagenesis and weak morphosis. However, the detailed contribution of


involved processes remains yet to be experimentallyevaluated.

In summary, we conclude that AFM and TEMtechniques combined with the Focussed Ion Beampreparation method and Laser Raman spectroscopysignificantly support the investigations of Precambrianmicrofossils, adding new, up to now inaccessibleinformation on the fossilization processes and on thenanoscopic structure of the fossils. However, opticalmicroscopy remains an important and inevitable basicmethod in Precambrian micropaleontology, providingcrucial information on the fossils and their relationshipto the hosting rock.

Acknowledgements

Financial support through DAAD (Kempe, Heckl),NSF (Schopf) and Le STUDIUM (Altermann) isgratefully appreciated. Annie Richard, Department ofPhysics, University of Orleans, France helped withSEM and EDAX analyses; A.B. Kudryavtsev, Univer-sity of Birmingham, Alabama, USA carried out Ramanspectroscopic measurements. Pat Eriksson UniversityPretoria, critically read the manuscript.

References

A isal

B Kra-002.416,

B hean285,

B anhert-451.

C ofd ondro-

C cti-den69–

D ude:

Advances in Organic Geochemistry 1975. ENADIMSA, Madrid,pp. 753–779.

Durand, B. (Ed.), 1980. Kerogen, Insoluble Organic Matter fromSedimentary Rocks.Editions Technip, Paris, 519 pp.

Egerton, R.F., 1989. Electron Energy-Loss Spectroscopy in the Elec-tron Microscope. Plenum Press, New York.

Faulon, J.L., Vandenbrouke, M., Drappier, J.M., Behar, F., Romero,M., 1989. 3D Chemical model for geological macromolecules.Org. Geochem. 16, 981–993.

Garvie, L.A., Craven, A.J., Brydson, R., 1984. Use of electron-energyloss near-edge fine structure in the study of minerals. Am. Min-eralogist 79, 411–425.

Guignard, G., Thevenard, F., van Konijnenburg-van Cittert, J.H.A.,1998. Cuticle ultra structure of the cheirolepidiaceous coniferHirmeriella muensteri (Schenk) Jung. Rev. Palaeobot. Palynol.104 (2), 115–141.

Harley, M.M., 1997. Ultra structure of pollen from some Eocene palmflowers (Messel, Germany). Mededelingen Nederlands Instituutvoor Toegepaste Geowetenschappen TNO 58, 193–209.

Hemsley, A.R., Glasspool, I.J., 1999. Megaspore ultra structure. In:Jones, T.P., Rowe, N.P. (Eds.), Fossil Plants and Spores: Mod-ern Techniques. Geological Society of London, United Kingdom(GBR), London.

House, C.H., Schopf, J.W., McKeegan, K.D., Coath, C.D., Harri-son, T.M., Stetter, K.O., 2000. Carbon isotopic composition ofindividual Precambrian microfossils. Geology 28, 707–710.

Jones, T.P., Rowe, N.P. (Eds.), 1999a. Fossil Plants and Spores: Mod-ern Techniques. Geological Society of London, United Kingdom(GBR), London.

Jones, T.P., Rowe, N.P., 1999b. Plant cell walls. In: Jones, T.P., Rowe,N.P. (Eds.), Fossil Plants and Spores: Modern Techniques. Geo-logical Society of London, United Kingdom (GBR), London.

Jones, T.P., Rowe, N.P., 1999c. Embedding techniques; adhesivesand resins. In: Jones, T.P., Rowe, N.P. (Eds.), Fossil Plantsand Spores: Modern Techniques. Geological Society of London,

K liza-

K rth’s

K -Bio-

K ckl,ro-120.

K 00.ssils.

K unds

L ibu-ze,li 70

L arlyma,

ltermann, W., 2001. The oldest fossils of Africa—a brief reappraof reports from the Archean. J. Afr. Earth Sci. 33, 427–436.

rasier, M.D., Green, O.R., Jephcoat, A.P., Kleppe, A.K., Vannendonk, M.J., Lindsay, J.F., Steele, A., Grassineau, N.V., 2Questioning the evidence for Earth’s oldest fossils. Nature76–81.

rocks, J.J., Logan, G.A., Buick, R., Summons, R.E., 1999. Arcmolecular fossils and the early rise of eukaryotes. Science1033–1036.

uick, R., 1991. Microfossil recognition in Archean rocks:appraisal of spheroids and filaments from a 3500 My old cbarite unit at North Pole, Western Australia. Palaios 5, 441–

astiglioni, C., Mapelli, C., Negri, F., Zerbi, G., 2001a. Originthe D line in the Raman spectrum of graphite: a study baseRaman frequencies and intensities of polycyclic aromatic hycarbon molecules. J. Chem. Phys. 114, 963–974.

astiglioni, C., Negri, F., Rigolio, M., Zerbi, G., 2001b. Raman avation in disordered graphites of the A1’ symmetry forbidk0 phonon: the origin of the D line. J. Chem. Phys. 115, 373778.

urand, B., Marchand, A., Amiell, J., Combaz, A., 1977. Etde Kerogenes par resonance paramagnetiqueelectronique. In

United Kingdom (GBR), London.azmierczak, J., Altermann, W., 2002. Neoarchean biominera

tion by benthic coccoid cyanobacteria. Science 298, 2351.azmierczak, J., Kremer, B., 2002. Thermal alteration of the Ea

oldest fossils. Nature 420, 477–478.empe, A., 2003. Nanostructures in Precambrian Fossils. Geo

Nano, 1. Utz Verlag, Munich, Germany, 155 pp.empe, A., Schopf, J.W., Altermann, W., Kudryavtsev, A.B., He

W.M., 2002. Atomic force microscopy of Precambrian micscopic fossils. Proc. Natl. Acad. Sci. U.S.A. 99 (14), 9117–9

udryavtsev, A.B., Schopf, J.W., Agresti, D.G., Wdowiak, T.J., 20In situ laser-Raman imagery of Precambrian microscopic foProc. Natl. Acad. Sci. U.S.A. 98 (3), 823–826.

wok, S., 2004. The synthesis of organic and inorganic compoin evolved stars. Nature 430, 985–991.

oeblich Jr., A.R., 1970. Morphology, ultra structure and distrtion of Paleozoic acritarchs. Atti - Istituto Veneto di ScienLettere ed Arti. Classe di Scienze Matematiche e Natura(G), 705–788.

oeblich Jr., A.R., Drugg, W.S., 1968. New acritarchs from the EDevonian (Late Gedinnian) Haragan Formation of OklahoUSA. Tulane Studies Geol. 6 (4), 129–137.


Loeblich Jr., A.R., Tappan, H., 1969. Acritarch excystment and sur-face ultra structure with descriptions of some Ordovician taxa.Revista Espanola de Micropaleontologia 1 (1), 45–57.

Loeblich Jr., A.R., Tappan, H., 1971. New observations of the ultrastructure of Asketopalla, an Ordovician acritarch. J. Paleontol.45 (5), 899–901.

Malis, T., Cheng, S.C., Egerton, R.F., 1988. EELS log-ratio techniquefor specimen thickness measurement in the TEM. J. ElectronMicrosc. Technol. 8, 193–200.

Mapelli, C., Castiglioni, C., Meroni, E., Zerbi, G., 1999a. Graphiteand graphitic compounds: vibrational spectra from oligomers toreal materials. J. Mol. Struct. 480–481, 615–620.

Mapelli, C., Castiglioni, C., Zerbi, G., Mullen, K., 1999b. Commonforce field for graphite and polycyclic aromatic hydrocarbons.Phys. Rev. B 60, 12710–12725.

Moore, T.B., Schopf, J.W., 1992. Geographic and geologic data forPPRG rock samples. In: Schopf, J.W., Klein, C. (Eds.), TheProterozoic Biosphere—A Multidisciplinary Study. CambridgeUniversity Press, Cambridge.

Negri, F., Castiglioni, C., Tommasini, M., Zerbi, G., 2002. A compu-tational study of the Raman spectra of large polycyclic aromatichydrocarbons: toward molecularly defined subunits of graphite.J. Phys. Chem. A 106, 3306–3317.

Oehler, D.Z., 1977. Pyrenoid-like structures in Late Precambrianalgae from the Bitter Springs Formation. J. Paleontol. 51,885–901.

Osborn, J.M., Phipps, C.J., Taylor, T.N., Taylor, E.L., 2000. Struc-turally preserved sphenophytes from the Triassic of Antarctica;reproductive remains of Spaciinodum. R. Palaeobot. Palynol. 111(3–4), 225–235.

Rigolio, M., Castiglioni, C., Zerbi, G., Negri, F., 2001. Density func-tional theory prediction of the vibrational spectra of polycyclicaromatic hydrocarbons: effect of molecular symmetry and sizeon Raman intensities. J. Mol. Struct. 563–564, 79–87.

Schopf, J.W., 1992a. Informal revised classification of proterozoiczoicity

S logic, C.dy.

S ert:6.

S hunt.W.,

and, pp.

Schopf, J.W., Czaja, A.D., Kudryavtsev, A.B., Agresti, D.G.,Wdowiak, T.J., Kempe, A., Altermann, W., Heckl, W.M., 2002a.The oldest evidence of life, ISSOL meeting June 30–July 5, Oax-aca, Mexico.

Schopf, J.W., Kudryavtsev, A.B., Agresti, D.G., Wdowiak, T.J.,Czaja, A.D., 2002b. Laser-Raman imagery of Earth’s earliest fos-sils. Nature 416 (7), 73–76.

Schopf, J.W., Kudryavtsev, A.B., Agresti, D.G., Czaja, A.D.,Wdowiak, T.J., 2005. Raman imagery: a new approach to assessthe geochemical maturity and biogenicity of permineralized Pre-cambrian fossils. Astrobiology 5, 333–371.

Schopf, J.W., Oehler, D.Z., 1976. How old are the eukaryotes? Sci-ence 193, 47–49.

Talyzina, N.M., 2000. Ultra structure and morphology of Chuaria cir-cularis (Walcott, 1899) Vidal and Ford (1985) from the Neopro-terozoic Visingso Group, Sweden. Precambrian Res. 102 (1/2),123–134.

Talyzina, N.M., Moczydlowska, M., 2000. Morphological and ultra-structural studies of some acritarchs from the Lower CambrianLukati Formation, Estonia. R. Palaeobot. Palynol. 112 (1–3),1–21.

Talyzina, N.M., Moldowan, J.M., Johannisson, A., Fago, F.J.,2000. Affinities of Early Cambrian acritarchs studied by usingmicroscopy, fluorescence flow cytometry and biomarkers. Rev.Palaeobot. Palynol. 108 (1/2), 37–53.

Taylor, T.N., 1999. The ultra structure of fossil cuticle. In: Jones,T.P., Rowe, N.P. (Eds.), Fossil Plants and Spores: Modern Tech-niques. Geological Society of London, United Kingdom (GBR),London.

Vandebroucke, M., 2003. Kerogen: from types to models of chem-ical structure. Oil & Gas Science and Techology - Rev. IFP 58,243–269.

van Krevelen, D.W., 1993. Coal: Typology–Chemistry–Physics–Constitution, 3rd ed. Elsevier, Amsterdam, 979 pp.

Wellman, C.H., Axe, L., 1999. In: Jones, T.P., Rowe, N.P. (Eds.),mac-R),

W lgae;ossilty of

W forences

Y theum23–

microfossils. In: Schopf, J.W., Klein, C. (Eds.), The ProteroBiosphere—A Multidisciplinary Study. Cambridge UniversPress, Cambridge, pp. 1119–1168.

chopf, J.W., 1992b. Proterozoic prokaryotes: affinities, geodistribution, and evolutionary trends. In: Schopf, J.W., Klein(Eds.), The Proterozoic Biosphere—A Multidisciplinary StuCambridge University Press, Cambridge, pp. 195–218.

chopf, J.W., 1993. Microfossils of the Early Archean Apex chnew evidence for the antiquity of life. Science 260, 640–64

chopf, J.W., 2004. Earth’s earliest biosphere: status of theIn: Eriksson, P.G., Altermann, W., Nelson, D.R., Mueller,Catuneanu, O. (Eds.), The Precambrian Earth: TemposEvents. Developments in Precambrian Geology. Elsevier516–539.

Extracting plant mesofossils and megafossils by bulk acideration. Geological Society of London, United Kingdom (GBFossil plants and spores; modern techniques.

illiamson, B.J., Jones, T.P., Berube, K.A., 1999. Calcareous aanalytical techniques. In: Jones, T.P., Rowe, N.P. (Eds.), FPlants and Spores: Modern Techniques. Geological SocieLondon, United Kingdom (GBR), London.

irth, R., 2004. Focused Ion Beam (FIB): a novel technologyadvanced application of micro- and nanoanalysis in geosciand applied mineralogy. Eur. J. Miner. 16, 863–876.

ang, W., Zhang, H., Xu, F., 1998. A tentative study onultra structure of the Middle Permian acritarch Micrhystridifrom western Yunnan, China. Chin. Sci. Bull. 43 (21), 181826.

Documents

Focussed ion beam preparation and in situ nanoscopic study of Precambrian acritarchs