Micro Raman Spectroscopy of Carbonaceous Material in Microfossils and Meteorites: Improving a Method for Life Detection

Abstract

The identification of biosignatures in Earth's ancient rock record and detection of extraplanetary life is one of the primary goals in astrobiology. Intrinsic to this goal is the improvement of analytical techniques and protocols used to identify an unambiguous signal of life. Micro Raman spectroscopy is a nondestructive method that allows for in situ identification of a wide range of minerals and compounds. The use of D (∼1350 cm⁻¹) and G (∼1580 cm⁻¹) band parameters to infer the biogenicity of carbonaceous materials in fossils has become a commonly used analytical tool, but carbonaceous compounds from different sources often share the same spectroscopic characteristics. Microfossil studies do not always take into consideration a nonbiological source for the carbon in their samples and therefore still rely on morphology as the primary mode of identification. Comprehensive studies that consider all carbon sources are typically done on metasediments, coals, or meteorites, and the results are not clearly applicable to microfossil identification. In this study, microfossils from a suite of sedimentary rock samples of various ages were analyzed with micro Raman spectroscopy to investigate the nature and provenance of carbonaceous material. To further constrain D- and G-band carbon characteristics, micro Raman analyses were also performed on well-characterized meteorite samples as abiological controls. The results appear to show a correlation of precursor carbonaceous material with D-band parameters and thermal history with G-band parameters. This systematic study lays the groundwork for improving the use of the G- and D-band trends as useful indicators of the origin of carbon in microfossils. Before unambiguous biosignatures can be established, further work characterizing the carbonaceous material in microfossils of different ages, thermal histories, and host rock compositions is needed. Key Words: Raman spectroscopy—Microfossils—Biosignatures—Carbonaceous material—Thermal history. Astrobiology 13, 103–113.

1. Introduction

I n studies of ancient rocks on Earth, the majority of techniques used to investigate biogenicity have coupled the observation of morphology such as filaments, coccoids, or laminations with some manner of carbonaceous signature specific to life. As the original organic components break down during progressive metamorphism, hydrogen, oxygen, nitrogen, and sulfur compounds are lost (Eglinton, 1969). The carbon that remains is entrained in the petrology of newly forming minerals, such as phosphates, iron oxides, iron silicates, and sulfides (Beveridge et al., 1983; Konhauser, 1998; Konhauser and Urrutia, 1999). The carbon in these rocks can form as the result of the metamorphic conversion of organic matter, but the abiotic precipitation of graphite from carbon-rich fluids can also occur (Luque et al., 1998; Pasteris, 1999). Therefore, differences between abiogenic and biogenic carbon in ancient rocks are often difficult to discern by their spectral signatures alone.

Confocal Raman spectroscopic imaging is useful in this regard as the spectral signatures of carbonaceous materials can be mapped and correlated to contextual, morphological features (Kudryavstev et al., 2001; Schopf and Kudryavstev 2009; Bower, 2011; Marshall et al., 2011). Raman spectra of carbon are typically composed of disordered and ordered (graphitic) bands (“D” at ∼1350 cm⁻¹ and “G” at ∼1580 cm⁻¹, respectively). The peak characteristics of the D and G bands, such as width and intensity, can vary depending on a range of factors, including the degree of thermal alteration (metamorphic grade), precursor carbonaceous material, and crystallinity of the graphitized carbon (Pasteris and Wopenka, 2003; Busemann et al., 2007; Quirico et al., 2009; Marshall et al., 2010). The G band is a signature of graphitic crystalline order, and the D band is a resonant band indicative of disorder in the crystal lattice. Unlike the G band, the D band's spectral position varies slightly in response to laser excitation wavelength and power (Wang et al., 1990; Sood et al., 2001). The D band is also thought to reflect in-plane defects arising from the presence of hetero atoms (H, N, O) (Beny-Bassez and Rouzaud, 1985). A recent trend in micropaleontology has been to use the characteristics of these spectral features to investigate the origin of graphitized microfossils in ancient rocks (Kudryavstev et al., 2001; Schopf et al., 2005). These spectral features have been used as sensitive indicators of the metamorphic history of carbonaceous matter in several fossil studies (Wopenka and Pasteris, 1993; Allwood et al., 2006; Bernard et al., 2007; Marshall et al., 2010). However, the studies on ancient microfossils still rely primarily on the morphology of the feature investigated, with Raman microscopy confirming the presence of carbonaceous material. The D- and G-band parameters alone have been described as “necessary but not sufficient” for unambiguous life detection (Pasteris and Wopenka, 2003). Adding to the potential misinterpretation of Raman peak parameter signatures in life-detection studies is the lack of standardized comparative studies of biogenic and abiogenic carbonaceous material (Pasteris and Wopenka, 2003; Marshall et al., 2010). Carbon spectral features that are indicative of thermal history in highly metamorphosed rocks such as primitive meteorites or coals are not always applicable to sedimentary rocks that have experienced little metamorphism (Beyssac et al., 2002a; Rahl et al., 2005; Quirico et al., 2009).

Here, we address these issues by comparing the G- and D-band parameters of microfossils in a suite of sedimentary rocks of different ages and metamorphic histories with two well-characterized carbonaceous chondrites of abiological origin. The goal of this study was to constrain G- and D-band parameters to further understand their usefulness in determining the source of carbonaceous materials in geological samples.

2. Materials and Methods

Thin sections of 30 μm thickness were made from samples of fossiliferous rocks from different localities and ages: the 400 Ma Rhynie Formation, 600 Ma Duoshanto Formation, 1.9 Ga Gunflint Formation, 2.7 Ga Tumbiana Formation, and 3.5 Ga Apex and Strelley Pool chert deposits (Table 1). To provide comparisons with materials of known abiogenic origin, sections were also made from two well-characterized carbonaceous chondrite samples: Allende and Murchison (Table 1).

Table 1.

Age, Host Material, Known Peak Metamorphic Temperatures, Mineral Assemblages, and Origin of the Carbonaceous Material of the Fossil and Meteorite Samples Analyzed in This Study

Sample	Age	Host material	Peak metamorphic T	Mineral assemblages	Origin of carbonaceous material
List of Fossil Samples Analyzed
Rhynie chert (Scotland)	400 Ma	Chert	Progressive diagenesis, low-grade metamorphism (<150°C)	Pyrite, anatase, goethite, siderite	Plants, algae, fungi, crustaceans, prokaryotes
Duoshanto phosphorite (South China)	570 Ma	Phosphorite	Diagenesis	Dolomite, barite, apatite, sulfides	Algae, metazoans
Gunflint chert (Ontario, Canada)	∼1.9 Ga	Chert	Low grade (<150°C)	Sulfides, CO3, Fe oxides	Prokaryotes, algae
Tumbiana Formation (Fortescue Group, Australia)	2.7 Ga	Carbonate	Prehnite-pumpellyite (100–300°C)	Pyrite, chlorite, sericite, anatase, silica, aragonite	Prokaryotes
Strelley Pool chert (Pilbara, Australia)	∼3.5 Ga	Chert	Lower to mid greenschist	Dolomite, Fe oxides, alunite	Prokaryotes or C-rich hydrothermal fluids
Apex (Warawoona Group, Western Australia)	3.5 Ga	Chert	Lower to mid greenschist	Barite, Fe oxides, sulfides, chlorite, sericite, leucoxene	Prokaryotes or C-rich hydrothermal fluids
List of Meteorite Samples Analyzed
Murchison	∼4.6 Ga	Carbonaceous chondrite CM2(3)	Low T aqueous alteration	Pyroxene, olivine, Fe oxides/ sulfides, serpentine	Solar nebula (plasma gases): aliphatic and aromatic hydrocarbons, fullerenes, carboxylic acids, amino acids, puric nucleic bases, sugars
Allende	∼4.6 Ga	Carbonaceous chondrite CV3	Aqueous alteration, T>1000°C, followed by rapid cooling	Fe silicates, oxides, no volatiles	Solar nebula (plasma gases)

The Early Devonian (400 Ma) Rhynie chert in Aberdeen, Scotland, contains well-preserved remains of early vascular plants, lichens, algae, cyanobacteria, and insects (Boyce et al., 2003; Krings et al., 2007). The cherts are part of a sedimentary and volcanic succession (interbedded with shales and sandstones) that precipitated from silica-rich fluids in one of the earliest known subaerial hot spring environments associated with andesitic volcanism (Rice et al., 2002; Trewin et al., 2003).

The ∼600 Ma Neoproterozoic Duoshanto Formation in South China contains the fossilized remains of algae, metazoan embryos, and acritarchs—taxonomically unclassified microfossils (Xiao et al., 1998; Xiao and Knoll, 2000). The formation consists of transgressive deposits of carbonate, shale, and phosphatic shale successions and volcanic ash beds. The fossils are found in the upper-sequence phosphorite beds that grade upward into a dolomitic micrite and grainstone (Xiao et al., 2002; Condon et al., 2005). Phosphorites are lithified organically induced phosphate-rich muds that are typically associated with marine carbonates (Sheldon, 1981). These sediments lie above glaciogenic rocks and were only lightly affected by local orogenic events (Xu et al., 1997; Barfod et al., 2002).

The cherts of the ∼1.9 Ga Gunflint Formation, Ontario, Canada, contain well-documented microfossils that include filamentous and coccoidal remains of stromatolite-building and planktonic microorganisms (Cloud, 1965; Licari and Cloud, 1968; Darby, 1974; Knoll et al., 1978; Awramik and Semikhatov, 1979; Lanier, 1989). These sediments were deposited in a shelf environment in the Animike Basin during a transgressive-regressive-transgressive cycle (Schulz and Cannon, 2007). The deposits consist of proximal stromatolites, lagoonal ribbon chert-carbonates, offshore grainstones that were derived from siliceous and iron-rich chemical mud layers, as well as black shales and volcanic ash beds in the upper half (Pufahl and Fralick, 2000; Fralick et al., 2002). The bedded cherts in the Gunflint Formation are considered to be either primary or diagenetic precipitates that replaced the original stromatolitic carbonate (Winter and Knauth, 1992; Maliva et al., 2005). These rocks have experienced subgreenschist facies metamorphism (T<150°C).

The 2.7 Ga Tumbiana Formation is a stromatolitic limestone that contains organic globules (Lepot et al., 2008; Phillipot et al., 2009). It is part of the Fortescue Group of the Hamersley Basin, Australia, which consists of calcareous sandstone, stromatolitic carbonate, and micaceous sandstone, all with interlayers of volcanic siltstone (Thorne and Trendall, 2001). The carbonates were deposited in a shallow subaqueous coastal lacustrine or marine environment (Sakurai et al., 2005). The host rocks experienced only prehnite-pumpellyite facies metamorphism—thermal alteration between 100–300°C (Smith et al., 1982).

The ∼3.5 Ga Strelley Pool and Apex cherts are part of the Warrawoona Group in Western Australia. The Apex cherts contain filamentous and coccoidal features, as well as disseminated carbonaceous material; the Strelley Pool cherts contain conical stromatolitic features, as well as filamentous microstructures (Schopf and Packer, 1987; Schopf, 1993; Brasier et al., 2005). The Warrawoona Group consists of mainly basaltic rocks interbedded with felsic volcanic rocks and metasedimentary units that contain the cherts (Van Kranendonk and Pirajano, 2004). The Apex chert is part of a 10 m thick bedded chert unit interbedded with felsic tuff. The Apex chert deposits were fed by swarms of hydrothermal feeder veins that have since been recrystallized to microcrystalline quartz. Minerals and native metals indicative of high hydrothermal temperatures (250–350°C) are found in the chert veins and associated basaltic host rocks (Van Kranendonk and Pirajano, 2004; Brasier et al., 2005; Pinti et al., 2009). The Strelley Pool chert is part of a silicified carbonate succession that is also transected by silica-rich veins, and the metabasalts below the cherts have been hydrothermally altered (Van Kranendonk et al., 2002; Lindsay et al., 2005). The metamorphic grade of the Warrawoona Group is up to greenschist facies, ∼250°C (Van Kranendonk et al., 2010).

Two well-characterized meteorites were included in the study to provide comparisons with naturally derived carbonaceous material of known abiogenic origin. Sections were made from two samples of Allende (CV3) and Murchison (CM2) carbonaceous chondrites. The Murchison and Allende meteorites are ∼4.5 Ga in age and fell to Earth in 1969 (Sears and Dodd, 1988). Both were formed from the accretion of solar nebular material but differ in terms of their post-accretion parent body processing. Allende likely experienced greater heating following accretion of its parent body than Murchison, and its carbonaceous materials show this in terms of greater thermal maturity (Sears and Dodd, 1988; Krot et al., 2007). The Allende meteorite is a type 3 CV chondrite. The CV chondrites have a higher Ca/Si ratio compared to CM types and tend to have Ca- and Al-rich inclusions (CAIs) along with high-Fe olivine (Sears and Dodd 1988; Krot et al., 2007). We observed that the Allende samples contain two pools of carbon that will be discussed: graphitic carbon found in the CAIs processed at relatively high temperature, and macromolecular carbon (MMC) in the meteorite matrix. (MMC is a reduced carbon species which spans a range of crystalline ordering ranging from amorphous carbon to poorly ordered graphite.) Murchison is a type 2 CM chondrite that experienced temperatures of only ∼130°C during parent-body processing (Clayton and Mayeda, 1984). The CM chondrites reflect this low temperature in their mineralogy, featuring hydrous minerals including phyllosilicates and sulfates (Sears and Dodd, 1988).

Raman spectra of the sample thin sections were collected by using a WITec scanning near-field optical microscope (AlphaSNOM) running Scan Control Spectroscopy Plus software. The Raman scans were conducted with a frequency-doubled YAG laser with 532 nm wavelength. The laser was focused through a 25 μm diameter fiber and a 100× objective lens with 3–6 s dwell time per pixel and 0.16–2.2 mW/μm² laser power density as measured at the focal plane (Fries and Steele, 2011). The scans were done between 5 and 10 μm into the thin sections rather than at the surface to analyze material unaffected by thin section preparation. Images and virtual maps of the spatial elemental and mineral composition of the microfossils and other features of interest were generated with WITec Project 1.9 software. Deconvolution of the collected spectra was achieved by using Gaussian peak fits in ACD Labs 12 software. Peak parameters, such as peak position, peak intensity ratios (R1), and full width at half maximum of the G- and D-bands (Γ_G and Γ_D) were used to elucidate the differences among different carbon types in the samples. The position of the G band is primarily a function of the energy of vibrational modes in the crystalline domains of the material, which in turn is a function of the mean size of those domains. The full width at half maximum is a function of the distribution of crystalline domain sizes sampled by the excitation beam. Both features, in turn, are functions of the thermal alteration and metamorphic history of the carbonaceous material (Tuinstra and Koenig, 1970; Pasteris and Wopenka, 1991; Wopenka and Pasteris, 1993; Ferrari and Robertson, 2000; Guedes et al., 2005). R1 is the ratio of the intensity of the D band and G band (I_D/I_G) and reflects the degree of order in the carbonaceous material (Wopenka and Pasteris, 1993; Rahl et al., 2005).

3. Results

Micro Raman spectroscopic imaging of the fossil samples revealed carbon-bearing features via D- and G-band peak intensity (see Fig. 1A–1C). This is clearly illustrated by spectral maps of microfossils in the 1.9 Ga Gunflint chert samples that show distinct filaments and coccoids made up of carbonaceous material embedded in the host chert (Fig. 1B). Typical spectra for the Gunflint samples are shown in Fig. 1C: the average D-band peak position was 1339 cm⁻¹, and the G-band was centered at 1600 cm⁻¹; the predominant quartz peaks were 208 and 467 cm⁻¹; second-order peaks associated with hydrated silicate inclusions were also visible (Bower et al., 2010). The G- and D-band peak positions of each microfossil and chondrite sample set were also compared (Table 2). The chert microfossil samples all have similar G-band peak positions close to 1600 cm⁻¹. Of the microfossil samples, Tumbiana carbonates had the highest G-band peak position centered around 1606 cm⁻¹. The lowest values are found in graphitic carbon in the Allende CAI samples.

FIG. 1.

Micro Raman imaging shows spatial relationships between carbonaceous materials and host material. (A) Image of scan area of microfossils in the Gunflint chert. (B) Map of G-band peak intensity, where bright areas indicate highest intensity; dark areas show the strongest quartz (chert) intensity. (C) Spectra of scan area with typical quartz peaks at 208 cm⁻¹ and 467 cm⁻¹; first-order carbon peaks at 1339 cm⁻¹ and 1600 cm⁻¹; second-order peaks at 2778 cm⁻¹, 2900 cm⁻¹, and 3217 cm⁻¹. Color images available online at www.liebertonline.com/ast

Table 2.

Average G-Band and D-Band Values with Standard Deviations for G-Band Peak Position, Full Width Half Height of the G Band (Γ_G), D-Band Peak Position, Full Width Half Height of the D Band (Γ_D), and D-Band Intensity/G-Band Intensity (R1) of Carbonaceous Material in All Samples Studied

Sample	G-band position (cm ⁻¹ )	Γ _G	D-band position (cm ⁻¹ )	Γ _D	R1
400 Ma Rhynie chert	1600±1	65±9	1345±5	198±31	0.65±0.08
570 Ma Duoshanto phosphorite	1594±8	71±12	1338±6	197±8	0.82±0.12
1.9 Ga Gunflint chert	1602±2	51±16	1345±3	134±28	0.65±0.06
2.7 Ga Tumbiana carbonate	1606±2	46±2	1343±4	60±3	1.37±0.07
3.5 Ga Strelley Pool chert	1602±2	52±3	1342±2	56±5	2.02±0.21
3.5 Ga Apex chert	1600±3	60±3	1342±2	56±6	1.96±0.19
4.6 Ga Murchison chondrite	1587±5	92±12	1346±9	180±27	0.91±0.18
4.6 Ga Allende chondrite matrix material	1593±3	61±13	1342±3	61±3	1.55±0.17
4.6 Ga Allende chondrite (CAI)	1578±2	34±4	1346±3	41±3	0.43±0.15

The G-band peak position of the carbonaceous material in the Duoshanto microfossils and Murchison chondrite has the highest degree of scatter of all the samples in this study (Fig. 2). This is indicative of carbon that has not been homogenized by metamorphism. While their originating materials are very different, both Duoshanto and Murchison carbon materials are naturally heterogeneous. The Duoshanto samples contain a variety of fossil types with precursor materials that were structurally different from one another; the Murchison samples contain several different complex organic molecules (Xiao and Knoll, 2000; Engel and Macko, 2001). The progressive breakdown and subsequent graphitization of these materials during low-temperature alteration produced zones of varying disorder reflected in the wide range of G-band peak positions in the Duoshanto and Murchison samples (Wopenka and Pasteris, 1993; El Amri et al., 2005; Bernard et al., 2007).

FIG. 2.

Comparison of Γ _G to G-band peak position indicates thermal maturity of carbonaceous materials in geological samples; arrow points in direction of increasing thermal maturity. Microfossil samples mainly cluster together with higher G-band peak positions and lower Γ _G (area in circle); Allende CAIs do not follow the same trend and fall into the zone of crystalline graphite (area in polygon).

Comparisons of the average D-band peak positions between each sample set do not reveal any straightforward trends (Table 2). The graphitic carbon in the Allende CAI samples had the highest peak positions centered around 1346 cm⁻¹. The peak positions for the Duoshanto phosphorite samples were the lowest around 1338 cm⁻¹. The disparity in peak positions of these two particular sample sets might indicate some difference between known biogenic and abiogenic carbon with lower and higher D-band peak positions, respectively. This shift to lower D-band wavenumbers in the Duoshanto samples does not, however, correspond with an overlap of the hematite 1313 cm⁻¹ band (Marshall et al., 2011). Even so, none of the other microfossil samples follow the same trend, so this cannot be used as a proxy to prove biogenicity. One point of interest is in the standard deviations of the D-band peak position of the carbon in a few of the samples (Table 2). For example, the organically complex Murchison chondrite has the biggest degree of scatter of ±9 wavenumbers. This is followed by the multi-species microfossils in the Duoshanto phosphorite (±6) and Rhynie chert (±5).

Comparisons between the average G- and D-band spectra of the carbonaceous matter also reveal some interesting characteristics (Fig. 3). Overall, in the younger fossiliferous rocks, the Rhynie chert, Duoshanto phosphorite, and Gunflint chert, the first-order carbon D bands are wider than the G bands (Fig. 3A). A comparison of the full width half height of the G and D bands (Γ_G, Γ_D, respectively) shows the same trend in a more quantified way (Fig. 3C). The Tumbiana carbonate, Apex chert, Strelley Pool chert samples are similar to each other, with a narrower and more pronounced D band relative to the G band (Fig. 3A, 3C). This indicates that the range of sizes of disordered domains shows less variation than in the young fossiliferous rocks. The same results have been found in previous studies showing a progressive narrowing of the D band with increasing thermal alteration as short-range disorder decreases and graphite crystals increase in domain size (Pasteris and Wopenka, 2003; Schopf and Kudryavstev, 2009). The Allende matrix and CAI samples both feature narrow D and G bands, while the carbon spectra in the Murchison samples look more like those of the fossil samples (Fig. 3B). In comparison with the known metamorphic temperatures for each sample set, the overall trend here indicates a wider D band with lower temperature and a narrower D band with higher temperature.

FIG. 3.

Representative first-order carbon spectra for each sample used in this study. (A) Fossil samples listed from youngest sample and lower metamorphic temperature at the top (1), to oldest samples and highest metamorphic temperatures at the bottom (6); (B) Meteorite samples 7–9. All lower-temperature samples <150°C have wider, lower-intensity D bands relative to the G band; higher-temperature samples have mainly higher intensity, narrow D bands. (C) Graphical comparison of G- and D-band widths, with much wider D bands relative to G bands for the younger samples and Murchison meteorite; numbers correspond to spectra in panels (A) and (B). Color graphics available online at www.liebertonline.com/ast

The trend is similar with the Γ_G in the meteorite samples: the graphitic Allende CAIs have the lowest Γ_G, and the low-temperature Murchison chondrites have the highest (Table 2, Fig. 3C). The values for Γ_G for the carbon in the microfossil samples also exhibit the same characteristics, with the younger, less thermally overprinted samples having a higher Γ_G than the older samples. Previous studies on carbonaceous materials in metamorphic rocks have shown the same correlation of lower Γ_G with a higher metamorphic grade (Wopenka and Pasteris, 1993; Allwood et al., 2006).

The Γ_G can also be compared with the G-band peak position to indicate the thermal maturity of the carbon in a sample (Pasteris and Wopenka, 1991; Spotl et al., 1998; Beyssac et al., 2002a, 2002b; Perraki et al., 2006; Fries and Steele, 2011). In Fig. 2, the data mainly follow a linear trend that reflects the thermal maturity of the samples. Of the fossil samples, the Duoshanto phosphorite falls on the upper end of the line, representative of lower thermal alteration. This is in agreement with the known thermal history in that the Duoshanto also has undergone the least degree of alteration of the fossil samples. The rest of the fossil samples cluster together toward the lower end of the trend line indicative of a higher thermal maturity.

The Murchison chondrite samples exhibit similar trends to the Duoshanto samples, which likely reflects the lower maturity of the Murchison carbon (Fig. 2). The Allende matrix samples, however, fall between the higher-grade fossil samples and thermally less mature Duoshanto and Murchison samples. The Allende matrix carbon has experienced heating close to at least 1000°C, so we would expect the Allende matrix samples to lie well past minimally metamorphosed microfossil samples that only experienced temperatures as high as ∼350°C. The Allende CAI-hosted carbon does not lie on the trend line, indicating that it has crystallized beyond the amorphous carbon–MMC–polycrystalline carbon trendline as defined by the G-band graph and is predominantly single-crystal graphite on the scale of the Raman excitation beam (Fries and Steele, 2008). The meteorite samples follow the same overall trend as the microfossils, but the differences in precursor materials (solid-phase biogenic vs. abiogenic gas-phase precipitate) increase the difficulty in making close comparisons between these two very different sample types (Fries, 2011).

The average D- and G-band intensity ratios (R1) of the samples are graphically shown in descending value order in the inset in Fig. 4. The two similar chert samples, Apex and Strelley Pool, had the highest R1 values, and the graphitic Allende CAI-hosted carbon had the lowest (Table 2). For the fossil samples in this study, the higher R1 values mainly correlate with a higher metamorphic grade. This is in agreement with previous studies on carbons in metamorphic rocks where higher R1 values were associated with a higher degree of structural organization (Jehlicka et al., 2003 ). The values for the Duoshanto samples, however, are the exception. If the trend is higher R1=higher metamorphic grade, Duoshanto should have the lowest values of all the fossil samples. The anomalous R1 value for Duoshanto might be explained by its phosphorite host, which is unique in this data set. While microfossils in phosphorites have been documented in at least one other locality, this rock type is the least studied in comparison to cherts and nonphosphatic carbonates (Brasier and Singh, 1987).

FIG. 4.

Comparison of Γ _D to R1. (A) Cluster of samples where precursor carbonaceous material was compositionally complex—Rhynie chert, Duoshanto phosphorite, Gunflint chert, Murchison chondrite, ∼27°C>T<∼150°C. (B) Cluster of samples where precursor materials may have been less compositionally complex and hydrothermally overprinted—Apex chert, Strelley Pool chert, Tumbiana carbonate, Allende matrix, ∼300°C>T<∼1000°C. (C) Purely crystalline graphitic carbon in Allende CAI, T>1000°C. (Inset) Graph of average R1 values for all samples where the microfossils follow the trend of higher R1 with greater thermal maturity.

When Γ_D is plotted against the D- and G-band intensity ratio (R1), the sample sets cluster into different groups that correlate with the known origins of the precursor carbonaceous material in each sample (Fig. 4). The Rhynie and Duoshanto samples both contain the remains of a wide variety of organisms, including plants, arthropods, and microbes; the Gunflint chert contains fossilized prokaryotes and some primitive eukaryotes. The Murchison chondrite, which contains highly complex organic compounds, also falls into the cluster with the younger microfossil samples (Fig. 4A). In-plane defects from O, H, N heteroatoms that affect D-band characteristics may be the reason for the clustering of the fossil samples and Murchison chondrite (Beyssac et al., 2002a). Similar results have been shown for coals and chondrites, where precursor materials controlled the characteristics of the Raman spectra (Quirico et al., 2009). The Apex and Strelley Pool samples overlap and are clustered with Tumbiana and Allende matrix, and the common denominator in this case may be a hydrothermal origin and siliceous rock matrix (Fig. 4B). As with other parameters considered in this study, the crystalline graphitic Allende CAI samples plot separately from all other samples, because they have crystallized well in excess of the other materials (Fig. 4C).

4. Discussion

As shown in this study and elsewhere, distinctions can be found between G- and D-band spectral characteristics of carbonaceous materials of disparate origins. In high-grade metamorphic rocks, the G-band peak position has been shown to increase with increasing temperatures and pressures (Beyssac et al., 2002a). In the case of the lightly metamorphosed microfossil samples, however, our data agrees more with a study on low-grade metamorphic rocks by Rahl et al. (2005), in which G-band peak positions greater than 1600 cm⁻¹ were associated with metamorphic temperatures only up to 200°C.

In meteorite-hosted carbon, the shift upward in G-band position and downward in Γ_G continues up to a maximum temperature regime as MMC crystallizes toward polycrystalline graphite, but peak position deviates from the thermal trendline in the case of crystalline graphite (Busemann et al., 2007). This is exemplified here by the graphitic carbon in the Allende CAI-hosted samples, which had the lowest G-band peak position in this study centered around 1578 cm⁻¹.

There is a pronounced increase of Γ_D relative to Γ_G that correlates with the complexity of the carbonaceous material in the younger microfossils and Murchison, whereas the older microfossils and high-temperature meteorites do not exhibit this trend (Fig. 3). The evolution of the G-band in our samples indicates differences in the thermal maturity, suggesting that the older microfossil samples experienced much greater temperatures than the younger ones. These data, however, do not definitively differentiate between deposition of graphite from carbon-rich high-temperature fluids and high-temperature thermal modification of pre-existing in situ organic material. The meteorite samples follow similar trends, in that higher alteration temperature correlates with higher G-band peak position and lower Γ_G; however, the alteration temperatures of meteorite samples do not compare directly with those of the fossil samples. Therefore, these correlations between spectral parameters rely on starting material and cannot be applied universally.

Our values and trends for Γ_D are in agreement with previous studies, where the Γ_D was shown to be larger for more amorphous carbon and narrower for more graphitized carbon. This occurs because the range of sizes of disordered material diminishes with annealing (Wopenka and Pasteris, 1993). These trends may actually reveal useful information regarding the nature of the precursor carbonaceous materials in all the samples, regardless of sample provenance. While it is true that the D band is affected by external influences such as laser power and wavelength, this resonant band also reflects the degree of disorder in carbonaceous materials, and the Γ_D reflects the range of amorphous domain sizes (Le Guillou et al., 2012).

Rahl et al. (2005) studied metamorphosed sedimentary rocks from different geographic locations and found a range of R1 values that varied depending on metamorphic grade. In that case, some lower-grade sample (T<200°C) R1 values actually overlap with those of a much higher grade (T ∼ 450°C) with an initial increase in R1 as temperature increases. Wopenka and Pasteris (1993) showed the same trend with a range of higher R1 values up until the chlorite grade metamorphic facies, at which point R1 values decrease with increasing metamorphic grade. The trend for the meteorite samples in this study shows a decreasing R1 with increasing metamorphic grade. Beyssac et al. (2002a) showed that a decrease in R1 values reflects an increase in metamorphic grade for carbonaceous material in metasediments, where the minimum metamorphic temperature affecting their samples was 330°C. Others describe similar results in graphitized materials in low- to high-grade metamorphic rocks (Jehlicka and Beny, 1992; Wopenka and Pasteris, 1993; Kwiecinska et al., 2010).

It may be impractical to use R1 values alone to elucidate the origins of carbonaceous materials in a wide range of sample types. When Γ_D is plotted against the D- and G-band intensity ratio (R1), however, the sample sets cluster into different groups that correlate with the precursor carbonaceous material in each sample (Fig. 4). The carbon in the Rhynie chert, Duoshanto phosphorite, Gunflint chert, and Murchison chondrite samples were all derived from highly complex organic materials, and these samples are clustered together when the Γ_D is compared with R1. In support of this trend, the samples containing carbon derived from less complex carbonaceous material, Apex chert, Strelley Pool chert, Tumbiana carbonate, and Allende matrix, all cluster together when the same parameters are compared. Thus, D-band parameters can be used to differentiate more complex precursor carbonaceous materials from the purer graphitic structures, and this may yet prove to be more useful for assigning the origin of carbonaceous materials as additional samples are analyzed to improve statistical relevance.

5. Conclusions

The improvement of analytical methods for life detection is an ongoing process. Nondestructive techniques such as Raman spectroscopy bring the scientific community closer to the goal of unambiguously identifying signs of life by providing essential information on the origins of materials in geological samples. Most target points in geological samples are distinct carbonaceous features related to biological processes (e.g., morphological features such as filaments, coccoids, flakes, laminae), but there are also geological samples that contain carbonaceous materials not necessarily associated with a specific physical feature. To complicate matters more, many of these features are often contained within a matrix of carbonaceous material that may or may not be related to the structures of interest. Fortunately, the unique mapping capabilities of micro Raman spectroscopy clarify these distinctions by showing the spatial relationships between carbonaceous materials and associated geological features. This tool is essential for microfossil studies where it is necessary to couple morphological and geochemical information in assigning biogenicity.

Despite the inability to unambiguously identify biogenic carbon from abiogenic carbon, the G- and D-band trends described in this study are promising. The continuing improvement of techniques such as micro Raman spectroscopy to understand the thermal processing and origin of carbonaceous material in geological samples is therefore valuable for life detection on Earth and other planets.

Footnotes

Acknowledgments

Support was provided by the NASA Astrobiology Institute at the Carnegie Institution of Washington and the NASA Postdoctoral Program. The authors wish to thank Phil Fralick of Lakehead University and Mark Jirsa of the Minnesota Geological Survey for some of the Gunflint chert samples, Maia Schweizer for the Duoshanto and Rhynie chert samples, and Owen Green for the Apex chert and Strelley Pool chert samples. We would also like to acknowledge the input of Mihaela Glamoclija and Karyn Rogers of the Carnegie Institution of Washington.

Abbreviations

CAIs, Ca- and Al-rich inclusions; MMC, macromolecular carbon.

References

Allwood

A.C.

, Walter

M.R.

, Marshall

C.P.

2006. Raman spectroscopy reveals thermal palaeoenvironments of c. 3.5 billion-year-old organic matter. Vib Spectrosc, 41:190–197.

Awramik

S.M.

, Semikhatov

M.A.

1979. The relationship between morphology, microstructure, and microbiota in three vertically intergrading stromatolites from the Gunflint Iron Formation. Can J Earth Sci, 16:484–495.

Barfod

G.H.

, Albarede

, Knoll

A.H.

, Xiao

, Telouk

, Frei

, Baker

2002. New Lu-Hf and Pb-Pb age constraints on the earliest animal fossils. Earth Planet Sci Lett, 201:203–212.

Beny-Bassez

, Rouzaud

J.N.

1985. Characterization of carbonaceous materials by correlated electron and optical microscopy and Raman microspectroscopy. Scanning Electron Microscopy. O'hare

A.M.F.

SEM Inc.: Chicago, IL, 119–132.

Bernard

, Benzerara

, Beyssac

, Menguy

, Guyot

, Brown

G.E.

, Goffe

2007. Exceptional preservation of fossil plant spores in high-pressure metamorphic rocks. Earth Planet Sci Lett, 262:257–272.

Beveridge

T.J.

, Meloche

J.D.

, Fyfe

W.S.

, Murray

R.G.E.

1983. Diagenesis of metals chemically complexed to bacteria: laboratory formation of metal phosphates, sulfides, and organic condensates in artificial sediments. Appl Environ Microbiol, 45:1094–1108.

Beyssac

, Goffe

, Chopin

, Rouzaud

J.N.

2002a. Raman spectra of carbonaceous material in metasediments: a new geothermometer. Journal of Metamorphic Geology, 20:859–871.

Beyssac

, Rouzaud

J.-N.

, Goffe

, Brunet

, Chopin

2002b. Graphitization in a high-pressure, low-temperature metamorphic gradient: a Raman microspectroscopy and HRTEM study. Contrib Mineral Petrol, 143:19–31.

Bower

D.M.

2011. Micro Raman spectroscopic investigations of mineral assemblages in parallel to bedding laminae in 2.9 Ga sandstones of the Pongola Supergroup, South Africa. J Raman Spectrosc, 42:1626–1633.

10.

Bower

D.M.

, Steele

, Kater

2010. Micro Raman spectroscopic investigations of mineral and carbonaceous materials: redefining parameters for the detection of biosignatures [abstract A112] Goldschmidt 2010 Conference Abstracts, Geochim Cosmochim Acta. 74,11:Supplement 112.

11.

Boyce

C.K.

, Cody

G.D.

, Fogel

M.L.

, Hazen

R.M.

, Alexander

C.M.O.’D.

, Knoll

A.H.

2003. Chemical evidence for cell wall lignification and the evolution of tracheids in early Devonian plants. Int J Plant Sci, 164:691–702.

12.

Brasier

M.D.

, Singh

1987. Microfossils and Precambrian-Cambrian boundary stratigraphy at Maldeota, Lesser Himalaya. Geology Magazine, 124:323–345.

13.

Brasier

M.D.

, Green

O.R.

, Lindsay

J.F.

, McLoughlin

, Steele

, Stoakes

2005. Critical testing of Earth's oldest putative fossil assemblage from the ∼3.5 Ga Apex chert, Chinaman Creek, Western Australia. Precambrian Res, 140:55–102.

14.

Busemann

, Alexander

C.M.O.’D.

, Nittler

L.R.

2007. Characterization of insoluble organic matter in primitive meteorites by microRaman spectroscopy. Meteorit Planet Sci, 42:1387–1416.

15.

Clayton

R.N.

, Mayeda

T.K.

1984. The oxygen isotope record in Murchison and other carbonaceous chondrites. Earth Planet Sci Lett, 67:151–161.

16.

Cloud

P.E.

1965. Significance of the Gunflint (Precambrian) microflora. Science, 148:27–35.

17.

Condon

, Zhu

, Bowring

, Wang

, Yang

, Jin

2005. U-Pb ages from the Neoproterozoic Duoshanto Formation, China. Science, 308:95–98.

18.

Darby

D.G.

1974. Reproductive modes of Huroniospora microreticulata from cherts of the Precambrian Gunflint Iron-Formation. Geol Soc Am Bull, 85:1595–1596.

19.

Eglinton

1969. Organic Geochemistry. Springer-Verlag: New York.

20.

El Amri

, Maurel

M.-C.

, Sagon

, Baron

M.-H.

2005. The micro-distribution of carbonaceous matter in the Murchison meteorite as investigated by Raman imaging. Spectrochim Acta A Mol Biomol Spectrosc, 61:2049–2056.

21.

Engel

M.H.

, Macko

S.A.

2001. The stereochemistry of amino acids in the Murchison meteorite. Precambrian Res, 106:35–45.

22.

Ferrari

A.C.

, Robertson

2000. Interpretation of Raman spectra of disordered and amorphous carbon. Phys Rev B Condens Matter Mater Phys, 61:14095–14107.

23.

Fralick

, Davis

D.W.

, Kissin

S.A.

2002. The age of the Gunflint Formation, Ontario, Canada: single zircon U-Pb age determinations from reworked volcanic ash. Can J Earth Sci, 39:1085–1091.

24.

Fries

2011. Carbon mineralogy—a comprehensive approach to planetary sciences [abstract 5425] In 74th Annual Meteoritical Society Meeting. Lunar and Planetary Institute Houston.

25.

Fries

M.D.

, Steele

2008. Graphite whiskers in CV3 meteorites. Science, 320:91–93.

26.

Fries

M.D.

, Steele

2011. Raman spectroscopy and confocal Raman imaging in mineralogy and petrography. Confocal Raman Microscopy. Dieing

, Hollricher

Toporskis

Springer-Verlag: Berlin, 111–136.

27.

Guedes

, Noronha

, Prieto

A.C.

2005. Characterisation of dispersed organic matter from lower Palaeozoic metasedimentary rocks by organic petrography, X-ray diffraction and micro-Raman spectroscopy analyses. International Journal of Coal Geology, 62:237–249.

28.

Jehlicka

, Beny

1992. Application of Raman microspectrometry in the study of structural changes in Precambrian kerogens during regional metamorphism. Org Geochem, 18:211–213.

29.

Jehlicka

, Urban

, Pokorny

2003. Raman spectroscopy of carbon and solid bitumens in sedimentary and metamorphic rocks. Spectrochim Acta A Mol Biomol Spectrosc, 59:2341–2352.

30.

Knoll

A.H.

, Barghoorn

E.S.

, Awramik

S.M.

1978. New microorganisms from the Aphebian Gunflint Iron Formation, Ontario. J Paleontol, 52:976–992.

31.

Konhauser

K.O.

1998. Diversity of bacterial iron mineralization. Earth-Science Reviews, 43:91–121.

32.

Konhauser

K.O.

, Urrutia

M.M.

1999. Bacterial clay authigenesis: a common biogeochemical process. Chem Geol, 161:399–413.

33.

Krings

, Kerp

, Hass

, Taylor

T.N.

, Dotzler

2007. A filamentous cyanobacterium showing structural colonial growth from the Early Devonian Rhynie chert. Rev Palaeobot Palynol, 146:265–276.

34.

Krot

A.N.

, Yurimoto

, Hutcheon

I.D.

, Libourel

, Chaussidon

, Tissandier

, Petaev

M.I.

, MacPherson

G.J.

, Paque-Heather

, Wark

2007. Type C Ca, Al-rich inclusions from Allende: evidence for multistage formation. Geochim Cosmochim Acta, 71:4342–4364.

35.

Kudryavstev

A.B.

, Schopf

J.W.

, Agresti

D.G.

, Wdowiak

T.J.

2001. In situ laser-Raman imagery of Precambrian microscopic fossils. Proc Natl Acad Sci USA, 98:823–826.

36.

Kwiecinska

, Suarez-Ruiz

, Paluszkiewicz

, Rodriques

2010. Raman spectroscopy of selected carbonaceous samples. International Journal of Coal Geology, 84:206–212.

37.

Lanier

W.P.

1989. Interstitial and peloid microfossils from the 2.0 Ga Gunflint Formation: implications for the paleoecology of the Gunflint stromatolites. Precambrian Res, 45:291–318.

38.

Le Guillou

, Rouzaud

J.N.

, Bonal

, Quirico

, Derenne

, Remusat

2012. High resolution TEM of chondritic carbonaceous matter: metamorphic evolution and heterogeneity. Meteorit Planet Sci, 47:345–362.

39.

Lepot

, Benzerara

, Brown

G.E.

Jr. , Phillipot

2008. Microbially influenced formation of 2,724-milion-year-old stromatolites. Nat Geosci, 1:118–121.

40.

Licari

G.R.

, Cloud

P.E.

1968. Reproductive structures and taxonomic affinities of some nannofossils from the Gunflint Iron Formation. Proc Natl Acad Sci USA, 59:1053–1060.

41.

Lindsay

J.F.

, Brasier

M.D.

, McLoughlin

, Green

O.R.

, Fogel

M.L.

, Steele

, Mertzman

S.A.

2005. The problem of deep carbon—an Archean paradox. Precambrian Res, 143:1–22.

42.

Luque

F.J.

, Pasteris

J.D.

, Wopenka

, Rodas

, Barrenechea

J.F.

1998. Natural fluid-deposited graphite: mineralogical characteristics and mechanisms of formation. Am J Sci, 298:471–498.

43.

Maliva

R.G.

, Knoll

A.H.

, Simonson

B.M.

2005. Secular change in the Precambrian silica cycle: insights from chert petrology. Geol Soc Am Bull, 117:835–845.

44.

Marshall

C.P.

, Edwards

H.G.M.

, Jehlicka

2010. Understanding the application of Raman spectroscopy to the detection of traces of life. Astrobiology, 10:229–243.

45.

Marshall

C.P.

, Emry

J.R.

, Marshall

A.O.

2011. Haematite pseudomicrofossils present in the 3.5-billion-year-old Apex chert. Nat Geosci, 4:240–243.

46.

Pasteris

J.D.

1999. Causes of uniformly high crystallinity of graphite in large epigenetic deposits. Journal of Metamorphic Geology, 17:779–787.

47.

Pasteris

J.D.

, Wopenka

1991. Raman spectra of graphite as indicators of degree of metamorphism. Can Mineral, 29:1–9.

48.

Pasteris

J.D.

, Wopenka

2003. Necessary, but not sufficient: Raman identification of disordered carbon as a signature of ancient life. Astrobiology, 3:727–738.

49.

Perraki

, Proyer

, Mposkos

, Kaindl

, Hoinkes

2006. Raman micro-spectroscopy on diamond, graphite and other carbon polymorphs from the ultrahigh-pressure metamorphic Kimi Complex of the Rhodope Metamorphic Province, NE Greece. Earth Planet Sci Lett, 241:672–685.

50.

Phillipot

, Van Kranendonk

M.J.

, Van Zuilen

M.A.

, Lepot

, Rividi

, Teitler

, Thomazo

, Blanc-Valleron

M.-M.

, Rouchy

J.-M.

, Grosch

, de Wit

2009. Early traces of life investigations in drilling Archean hydrothermal and sedimentary rocks of the Pilbara Craton, Western Australia and Barberton Greenstone Belt, South Africa. C R Palevol, 8:649–663.

51.

Pinti

D.L.

, Mineau

, Clement

2009. Hydrothermal alteration and microfossil artefacts of the 3,465-million-year-old Apex chert. Nat Geosci, 2:640–643.

52.

Pufahl

P.K.

, Fralick

P.W.

2000. Depositional environments of the Paleoproterozoic Gunflint Formation [Field Trip 4] Proceedings of the Institute on Lake Superior Geology, 46. The Institute on Lake Superior Geology: Houghton, MI.

53.

Quirico

, Montagnac

, Rouzaud

J.N.

, Bonal

, Bourot-Denise

, Duber

, Reynard

2009. Precursor and metamorphic condition effects on Raman spectra of poorly ordered carbonaceous matter in chondrites and coals. Earth Planet Sci Lett, 287:185–193.

54.

Rahl

J.M.

, Anderson

K.M.

, Brandon

M.T.

, Fassoulas

2005. Raman spectroscopic carbonaceous material thermometry of low-grade metamorphic rocks: calibration and application to tectonic exhumation in Crete, Greece. Earth Planet Sci Lett, 240:339–354.

55.

Rice

C.M.

, Trewin

N.H.

, Anderson

LI.

2002. Geological setting of the Early Devonian Rhynie cherts, Aberdeenshire, Scotland: an early terrestrial hot spring system. J Geol Soc London, 159:203–214.

56.

Sakurai

, Ito

, Ueno

, Kitajima

, Maruyama

2005. Facies architecture and sequence stratigraphic features of the Tumbiana Formation in the Pilbara Craton, northwestern Australia: implications for depositional environments of oxygenic stromatolites during the late Archean. Precambrian Res, 138:255–273.

57.

Schopf

J.W.

1993. Microfossils of the Early Archean Apex chert: new evidence of the antiquity of life. Science, 260:640–646.

58.

Schopf

J.W.

, Kudryavstev

A.B.

2009. Confocal laser scanning microscopy and Raman imagery of ancient microscopic fossils. Precambrian Res, 173:39–49.

59.

Schopf

J.W.

, Packer

B.M.

1987. Early Archean (3.3-billion to 3.5-billion-year-old) microfossils from Warrawoona Group, Australia. Science, 237:70–73.

60.

Schopf

J.W.

, Kudryavstev

A.B.

, Agresti

D.G.

, Czaja

A.D.

, Wdowiak

T.J.

2005. Raman imagery: a new approach to assess the geochemical maturity and biogenicity of permineralized Precambrian fossils. Astrobiology, 5:333–371.

61.

Schulz

K.J.

, Cannon

W.F.

2007. The Penokean orogeny in the Lake Superior region. Precambrian Res, 157:4–25.

62.

Sears

D.W.G.

, Dodd

R.T.

1988. Overview and classification of meteorites. Meteorites and the Early Solar System. Kerridge

J.F.

, Shapley Matthews

The University of Arizona Press: Tucson, AZ, 3–34.

63.

Sheldon

R.P.

1981. Ancient marine phosphorites. Annu Rev Earth Planet Sci, 9:251–284.

64.

Smith

R.E.

, Perdrix

J.L.

, Parks

T.C.

1982. Burial metamorphism in the Hamersley Basin, Western Australia. Journal of Petrology, 23:75–102.

65.

Sood

A.K.

, Gupta

, Asher

S.A.

2001. Origin of the unusual dependence of Raman D band on excitation wavelength in graphite-like materials. J Appl Phys, 90:4494–4497.

66.

Spotl

, Houseknecht

, Jaques

1998. Kerogen maturation and incipient graphitization of hydrocarbon source rocks in the Arkoma Basin, Oklahoma and Arkansas: a combined petrographic and Raman spectrometric study. Org Geochem, 28:535–542.

67.

Thorne

A.M.

, Trendall

A.F.

2001. Geology of the Fortescue Group, Pilbara Craton, Western Australia. Western Australia Geological Survey Bulletin, 144:249.

68.

Trewin

N.H.

, Fayers

S.R.

, Kelman

2003. Subaqueous silicification of the contents of small ponds in an Early Devonian hot-spring complex, Rhynie, Scotland. Can J Earth Sci, 40:1697–1712.

69.

Tuinstra

, Koenig

J.L.

1970. Raman spectrum of graphite. J Chem Phys, 53:1126–1130.

70.

Van Kranendonk

M.J.

, Pirajano

2004. Geochemistry of metabasalts and hydrothermal alteration zones associated with c. 3.45 Ga chert and barite deposits: implications for the geological setting of the Warrawoona Group, Pilbara Craton, Australia. Geochemistry: Exploration, Environment, Analysis, 4:253–278.

71.

Van Kranendonk

M.J.

, Hickman

A.H.

, Smithies

R.H.

, Nelson

D.R.

2002. Geology and tectonic evolution of the Archean North Pilbara terrain, Pilbara Craton, Western Australia. Econ Geol, 97:695–732.

72.

Van Kranendonk

M.J.

, Smithies

R.H.

, Hickman

A.H.

, Wingate

M.T.D.

, Bodorkos

2010. Evidence for Mesoarchean (∼3.2 Ga) rifting of the Pilbara Craton: the missing link in an early Precambrian Wilson cycle. Precambrian Res, 177:145–161.

73.

Wang

, Alsmeyer

D.C.

, McCreery

R.L.

1990. Raman spectroscopy of carbon materials: structural basis of observed spectra. Chem Mater, 2:557–563.

74.

Winter

B.L.

, Knauth

L.P.

1992. Stable isotope geochemistry of cherts and carbonates from the 2.0 Ga Gunflint Iron Formation: implications for the depositional setting, and the effects of diagenesis and metamorphism. Precambrian Res, 59:283–313.

75.

Wopenka

, Pasteris

J.D.

1993. Structural characterization of kerogens to granulite-facies graphite: applicability of Raman microprobe spectroscopy. Am Mineral, 78:533–557.

76.

Xiao

, Knoll

A.H.

2000. Phosphatized animal embryos from the Neoproterozoic Duoshanto Formation at Weng'an, Guizhou, South China. J Paleontol, 74:767–788.

77.

Xiao

, Zhang

, Knoll

A.H.

1998. Three-dimensional preservation of algae and animal embryos in a Neoproterozoic phosphorite. Nature, 391:553–558.

78.

Xiao

, Yuan

, Steiner

, Knoll

A.H.

2002. Macroscopic carbonaceous compressions in a terminal proterozoic shale: a systematic reassessment of the Miaohe biota, South China. J Paleontol, 76:347–376.

79.

, Rowley

D.B.

, Jiayu

, Jin

, Yuan-Dong

, Ren-Bing

1997. Late Precambrian through Early Paleozoic stratigraphic and tectonic evolution of the Nanling Region, Hunan Province, South China. Int Geol Rev, 39:469–478.