FTIR microspectroscopy of carbonaceous matter in ~ 3.5 Ga seafloor hydrothermal deposits in the North Pole area, Western Australia

Abstract

Carbonaceous matter in ~ 3.5 Ga hydrothermal vein deposits from the Dresser Formation, Western Australia, was analyzed using Fourier transform infrared (FTIR) microspectroscopy. Based on the spectroscopy, the carbonaceous matter was mainly composed of disordered aromatic structures, with minor aliphatic C–H functional groups. Spatially resolved analysis supports that the aliphatic C–H signatures are derived from kerogenous macro-molecule and not from free bitumen or other artificial sources. The intensity ratios of the asymmetric aliphatic CH₃ to the asymmetric aliphatic CH₂ (R_3/2 value) in the carbonaceous clots range from 0.22 to 0.51. Thermal alteration may increase or not change the R_3/2 value of organic matter, as suggested by previous experiments, but it is unlikely to be the cause of the substantially lower R_3/2 values when compared with those of primary organic matter. In particular, the low R_3/2 values (< ~ 0.4) suggest that the carbonaceous matter mainly contains aliphatic C–H bonds derived from bacterial cells. The carbonaceous clots may have been possibly produced by abiotic reaction such as Fischer-Tropsch-type (FTT) synthesis. However, the organic matter source only produced by the FTT synthesis is inconsistent with the R_3/2 values for the analyzed carbonaceous clots. The results obtained by combining these spectroscopic features of the carbonaceous clots together with the previously reported isotopic features may possibly suggest that both bacteria and archaea were colonized in the ~ 3.5 Ga Dresser hydrothermal system.

Introduction

Hydrothermal systems have been considered as the origin of life (Holm and Andersson 1998) and the locations of the earliest evolution of life and metabolism (Nisbet 1995). The geological study of ancient hydrothermal systems is crucial to understanding the potential role of hydrothermal systems as the earliest ecosystems. One of Earth’s oldest seafloor hydrothermal deposits occurs in the ~ 3.5 Ga Dresser Formation in Western Australia (Fig. 1). In this formation, prominent silica-barite veins containing carbonaceous matter ubiquitously penetrate pillowed basaltic greenstones (Ueno et al. 2004; Van Kranendonk 2006; Van Kranendonk et al. 2008). The carbonaceous matter in the veins afford an important means of understanding how organic matter was synthesized in the Earth’s early hydrothermal systems, although the biological and abiological processes cannot be fully distinguished. The occurrence of biological activity in the Dresser Formation has been inferred from the observation of putative stromatolites (Walter et al. 1980; Van Kranendonk 2006; Van Kranendonk et al. 2008), morphologically preserved putative microfossils (Ueno et al. 2001), microbially induced sedimentary structures (Noffke et al. 2013), and chemical fossils such as ¹³C-depleted organic matter and methane (Ueno et al. 2001, 2004, 2006) and ³⁴S-depleted pyrites with characteristic ∆³³S (Ueno et al. 2008; Shen et al. 2009). Sulfur isotopic signatures also suggest that microbial sulfate reduction was involved in pyrite formation (Ueno et al. 2008; Shen et al. 2009). In addition, ancient hydrothermal fluid inclusions contain ¹³C-depleted methane, suggesting the presence of methanogenic archaea (Ueno et al. 2006), although it is arguable if those methane were produced by metal-catalyzed abiotic reactions (Sherwood-Lollar and McCollom 2006). The ¹³C-depletion of organic matter indicates the predominance of autotrophic carbon fixation (Ueno et al. 2001, 2004). The organic matter has been characterized by Raman microspectroscopy, FTIR spectroscopy, solid-state ¹³C nuclear magnetic resonance, and pyrolysis analysis (e.g., Ueno et al. 2001; Derenne et al. 2008; Duda et al. 2018). The results of these organic geochemical analyses of isolated insoluble organic matter reveal that the organic matter is dominated by aromatic structures with aliphatic hydrocarbon moieties. In addition, the pyrolysis fragments of the organic matter comprise long-chain aliphatic hydrocarbons and thiophenes, which are likely related to the existence of microbes including sulfate-reducing bacteria (Derenne et al. 2008; Duda et al. 2018).

Fig. 1

Geologic map of the Dresser Formation in the North Pole area. The localities of the silica vein samples 96NP452 and T-58 are shown (modified from Kitajima et al. 2001)

Chemical analyses of Archean organic matter have provided limited but important information about the physiological and phylogenetic characteristics of the source organic compounds (e.g., Hayes 1994; Brocks et al. 1999). Although the observation of morphologically preserved microfossils is a direct evidence of the existence of life, the identification of microfossils based on only morphological analysis is often insufficient (e.g., Brasier et al. 2002; Schopf et al. 2002). The chemical analysis of carbonaceous matter is necessary for understanding its origin and characteristics. However, the bulk analysis of isolated organic matter generally involves the potential risk of post-depositional and experimental contamination of the organic matter (e.g., Brocks et al. 1999; Derenne et al. 2008; Rasmussen et al. 2008; Bourbin et al. 2012). In situ analysis is a more reliable technique that avoids such contamination problems and extracts chemical signatures specific to individual microstructures in petrographic thin sections (e.g., House et al. 2000, 2013; Kudryavtsev et al. 2001; Ueno et al. 2001; Schopf et al. 2002, 2005; Igisu et al. 2006, 2009; van Zuilen et al. 2007; Wacey et al. 2011; Lepot et al. 2013; Williford et al. 2013). FTIR microspectroscopy has been used to analyze Proterozoic organic-walled acritarchs and bacterial fossils, as well as Phanerozoic miospores (e.g., Arouri et al. 1999, 2000; Marshall et al. 2005; Igisu et al. 2006, 2009; Steemans et al. 2010; Qu et al. 2015). These studies demonstrated the usefulness of IR signatures for characterizing organic-walled microfossils, including those with unknown affinity such as acritarchs. However, FTIR microspectroscopy has not yet been applied to Archean carbonaceous matter in petrographic thin sections. FTIR spectroscopy has been performed on isolated Archean carbonaceous matter, and reported the presence of aliphatic C-H stretching absorptions (Derenne et al. 2008). This previous attempt did not employ the in situ microscopic method. It has been suggested that isolated and powdered samples may be affected by contamination from volatile hydrocarbons in laboratory air (e.g., Salisbury and Walter 1989) and/or from other aliphatic C-H bearing components during sample preparation and storage (e.g., Kebukawa et al. 2009). Therefore, the data from powdered samples would need to be interpreted with caution.

In this paper, we first report the FTIR microspectroscopic features of early Archean carbonaceous matter obtained from two ~ 3.5 Ga silica veins in the Dresser Formation in the North Pole area of Western Australia. The IR signatures are then compared with the characteristics of a variety of biologically and abiotically produced organic compounds, and the possible origins of the carbonaceous clots are discussed.

Methods/Experimental

Samples

We analyzed two samples taken from the silica veins (96NP452 and T-58) that intrude on the basaltic greenstones of the Dresser Formation in the North Pole area of Western Australia. The sampling localities of the silica veins are shown in Fig. 1. Detailed descriptions of the localities and samples can be found in the works of Ueno et al. (2001, 2004, 2006). The Dresser Formation consists of pillowed basaltic greenstones and chert beds. In the North Pole area, these rocks have only undergone low-grade metamorphism below their greenschist facies (Buick and Dunlop 1990; Kitajima et al. 2001; Ueno et al. 2004). The lowest chert unit of the formation is intercalated with several barite beds (Buick and Dunlop 1990), and numerous silica veins intrude the greenstones below the chert/barite beds. The silica veins are 0.3–20 m in width and generally > 100 m in length (Ueno et al. 2001). Zircon U-Pb dating yielded two distinct age groups of 3458 ± 2 Ma (Thorpe et al. 1992a) and 3660 ± 52 Ma (Kitajima et al. 2008) of the felsic unit that overlies the Dresser Formation. A model lead age of 3490 Ma was obtained for galena taken from the Dresser Formation (Thorpe et al. 1992b), and this may represent the actual depositional age of the formation.

The first doubly polished petrographic thin section sample was made from sample 96NP452 with a thickness of approximately 60 μm. The preparation process was as described by Igisu et al. (2006). We used alumina paste to polish only one side of the thin section to remove interference fringes. To avoid organic contamination during sample preparation, the sample was attached to a glass slide using only a carbon-based dissoluble adhesive (Aron Alpha). The adhesive was used to remove the doubly polished petrographic thin section from the glass slide by immersing in acetone. Petrographic investigation of the thin section indicated that the silica vein sample was mainly composed of microcrystalline quartz (< ~ 10 μm), carbonaceous matter, sulfides, and carbonates (see for example Fig. 2g, h). The filamentous structures described by Ueno et al. (2001) were not observed in the doubly polished thin section.

Fig. 2

Optical photomicrographs of the analyzed thin sections. a–g Plane-polarized light photomicrographs of the thin section 96NP452. b Composite image of the photomicrographs; the dashed rectangle in b corresponds to the area shown in Fig. 3a. f Enlarged view of the dashed rectangle in e. h Cross-polarized light photomicrograph of the same area in g. i, j Plane-polarized light photomicrographs of the thin section T-58. i Composite image of the photomicrographs. j Enlarged view of the dashed rectangle in i. The locations (50 × 50 μm² dotted squares) of the 12 clots with satisfactory signal/noise ratios and their calculated R_3/2 ratios (1024 scans) are shown in a–d, f, g, j

The second set of doubly polished thin section samples was made from sample T-58 (Ueno et al. 2006) with a thickness of approximately 100 μm. A petrographic examination of the sample revealed three distinctive fabrics, which were denoted by F1–F3 (Fig. 2i). F1 was characterized by coarse-grained quartz (> 1 mm) with abundant fluid inclusions. No carbonaceous matter was observed in F1. F2 comprised microcrystalline quartz chalcedony. Black aggregates in F2 were carbonaceous matter (see Fig. 2i, j) and ion oxides such as hematite. The center of the carbonaceous clot (white part in Fig. 2j) was quartz. F3 consisted mainly of microcrystalline quartz (< ~ 10 μm) and contained carbonaceous matter and ion oxides such as hematite (Fig. 2i). The analyzed carbonaceous clots in the sample were preserved in a microcrystalline quartz matrix or microcrystalline quartz chalcedony, but not in the secondary quartz veins (Fig. 2).

FTIR microspectroscopy

FTIR microspectroscopic analysis was conducted on the two thin section samples using an automated XYZ stage set in an FTIR microspectrometer (JASCO FTIR6200 + IRT7000) with a single-point mercury cadmium telluride detector. The thin section was placed over a hole in the sample holder (Nakashima et al. 1989) and measured as described below.

A reference background spectrum was first measured away from the mounted sample (i.e., in air). A transmission IR spectrum of the sample was then measured. The IR spectrum was described in terms of the IR absorbance, which is a function of the wavenumber (cm⁻¹):

$$ \mathrm{Absorbance}=-{\log}_{10}\left(T/{T}_0\right) $$

(1)

where T and T₀ denote the transmitted intensity of the infrared light at each wavenumber for the sample and air background, respectively. To search for organic signatures, the IR spectra of about 3500 individual carbonaceous clots were first obtained by collecting 64 scans per spot within a spectral range of 4000–1000 cm⁻¹ with a 4 cm⁻¹ spectral resolution. Among the ~ 3500 examined spots, 47 had an IR absorbance of ~ 2925 cm⁻¹, which indicated excess analytical error for the adopted 64-scan acquisition mode. These spots were therefore measured again by collecting 1024 scans per spot within a spectral range of 4000–1000 cm⁻¹ with a 4 cm⁻¹ spectral resolution. This was done to obtain the IR spectra with a better signal-to-noise ratio. Each spectrum was obtained from a spot of 50 × 50 μm² (Figs. 2 and 3a). Duplicate measurements were performed on the 47 spots in the 1024-scan acquisition mode. The analytical errors were determined from the duplicate measurements at the same location as the background spectrum and were found to be smaller than 0.003 absorbance units in the 3000–2800 cm⁻¹ range for the 64-scan acquisition mode and smaller than 0.0007 absorbance units for the 1024-scan acquisition mode. All the spectral data were processed using Spectra Manager (JASCO).

Fig. 3

An example of the optical photomicrographs and transmission IR spectra of the analyzed carbonaceous clots. a Optical photomicrograph. The squares indicate the analyzed areas (50 × 50 μm²). The IR spectra of the two analyzed areas (thick-bordered squares A and B) are shown in b and c. b Transmission IR spectra (64 scans) of the carbonaceous clots with (a) and without (b) aliphatic CH moieties. One scale unit on the vertical axis corresponds to an absorbance of 0.2. The IR spectrum in b is representative of those obtained from the white squares in a (without aliphatic CH bonds). c An enlarged view of the IR spectra in b within 3000–2800 cm⁻¹. The range is indicated by the dotted rectangle in b. The arrow represents the IR band at ~ 2925 cm⁻¹. One scale unit on the vertical axis corresponds to an absorbance of 0.02

Structural characterization of carbonaceous clots

In the aliphatic C–H stretching region (3000–2800 cm⁻¹), the ~ 2925 cm⁻¹ band is due to two asymmetric stretching groups, namely, aliphatic CH₂ and aromatic CH₃ (Jones and Sandorfy 1956; Painter et al. 1981). The band is therefore considered as being the resultant of these two groups. Because the aliphatic CH₂ and aromatic CH₃ groups have symmetric stretching bands with different peak positions at ~ 2850 and ~ 2870 cm^− 1, respectively (Bellamy 1954; Jones and Sandorfy 1956), the contribution of the aliphatic CH₂ groups can be estimated by elucidating the relationship between the intensities at ~ 2925 cm⁻¹ and ~ 2850 cm⁻¹. According to Lambert-Beer’s Law:

$$ \mathrm{Absorbance}=\varepsilon \cdotp c\cdotp d, $$

(2)

where ε is the molar absorption coefficient, c is the concentration, and d is the sample thickness. Based on Eq. (2), the absorbance ratio between the asymmetric CH₂ and symmetric aliphatic CH₂ is converted into the ratio of their molar absorption coefficients as follows:

$$ {\mathrm{Abs}}_{\mathrm{s}}/{\mathrm{Abs}}_a={\varepsilon}_{\mathrm{s}}/{\varepsilon}_a, $$

(3)

where Abs_s and Abs_a are respectively the absorbances (peak heights) of the symmetric aliphatic CH₂ and asymmetric aliphatic CH₂, and ε_s and ε_a are the corresponding molar absorption coefficients. Thus, if the linear correlation between the peak heights at ~ 2925 and ~ 2850 cm⁻¹ is obtained, the ~ 2925 cm⁻¹ band would be observed to be mainly due to the aliphatic CH₂, and less attributable to the aromatic CH₃ group.

The IR signatures of the aliphatic CH₃ and CH₂ ratios generally reflect the degree of branching and the chain length of the aliphatic hydrocarbon moiety (e.g., Lin and Ritz 1993; Coates 2000). To evaluate the relative length and degree of branching of the aliphatic chain in the carbonaceous clots, we used the aliphatic CH₃/CH₂ absorbance ratio (R_3/2):

$$ {R}_{3/2}={\mathrm{Abs}}_{\mathrm{CH}3}/{\mathrm{Abs}}_{\mathrm{CH}2}, $$

(4)

where Abs_CH3 and Abs_CH2 respectively denote the peak heights of the asymmetric stretching bands of the aliphatic CH₃ (end-methyl, ~ 2960 cm⁻¹) and CH₂ (chain-methylene, ~ 2925 cm⁻¹) after linear baseline correction (Igisu et al. 2009). For a more precise determination of R_3/2, we used only the data for the absorbances at ~ 2960 and 2925 cm⁻¹, which were more than twice the analytical error.

Raman microspectroscopy

A laser Raman micro-spectrometer (Nanophoton RAMANtouch) was used to characterize the carbonaceous clots. Each sample was exposed to a green laser (532 nm wavelength) for a period of 5 s at a laser power density of 1 × 10⁵ W/cm². Raman spectra for wavenumbers within ~ 2670–110 cm⁻¹ were acquired with a 600 g/mm grating (spectral resolution of ~ 4 cm⁻¹). A × 100 objective lens (NA = 0.9) was used to achieve Raman analysis spatial resolution of < 1 μm, and the incident laser power was estimated to be ~ 0.4 mW at the sample surface. We only analyzed the carbonaceous matter embedded within the rocks below the surface of the thin section to avoid the effect of polishing, which can induce deformation of carbonaceous matter during sample preparation and thus possibly induce artificial modification of the Raman spectroscopic feature (Pasteris 1989). Prior to sample analyses, wavenumber calibration was determined by comparison with the standard silicon spectra (520 cm⁻¹). All the spectral data were processed using RAMAN viewer (Nanophoton).

Raman spectroscopy has been used as a tool for providing a metamorphic grade indicator of carbonaceous matter in geological samples (e.g., Yui et al. 1996; Beyssac et al. 2002; Rahl et al. 2005; Kouketsu et al., 2014). Several indicators have been proposed to estimate peak metamorphic temperatures by using Raman spectral parameters: peak position, intensity ratio, and full width at half maximum (FWHM) of D band and G band. We chose the D/G band intensity ratio (I₁₃₅₀/I₁₆₀₀) for estimation of the degree of thermal alteration of the carbonaceous matter in order to compare with the previous spectral data of the Proterozoic samples (Qu et al. 2015). The analytical errors for the Raman spectral features were determined by taking duplicate measurements within one area analyzed by FTIR microspectroscopy.

Results and discussion

IR spectra

Figures 3 and 4 show examples of the infrared transmission spectra of the carbonaceous clots in the silica veins. We measured more than 3500 locations in the carbonaceous clots in the thin sections, and only 12 clots had absorbances of ~ 2960, ~ 2925, or ~ 2850 cm⁻¹, which are more than twice the analytical error. The obtained absorbances of ~ 2960, ~ 2925, and ~ 2850 cm⁻¹ are summarized in Table 1. The spectrum shown by line B in Fig. 3 was the most common feature in the analyzed carbonaceous clots. This is quite similar to that of the quartz matrix (black line in Fig. 4a, b1, c1), and such data were not used. Instead, some carbonaceous clots show spectral patterns such as A in Fig. 3 (with a peak at ~ 2925 cm⁻¹). The detailed peak assignments of the spectra are explained below.

Fig. 4

The transmission IR absorbance spectra (1024 scans) of the carbonaceous clots. a Examples of the IR spectra of the carbonaceous clots with aliphatic CH₂ and CH₃ bonds (blue line: spot #6 in Table 1), only aliphatic CH₂ bond (green line: spot #12 in Table 1), and no CH₂ bond (red line: spot #13 in Table 1) together with quartz matrix (black line). IR transmission-reflection spectrum of the carbon-based dissolvable adhesive used in this study on an aluminum foil is shown by the brown line (modified from Igisu et al. 2006). One scale unit on the vertical axis corresponds to an absorbance of 0.2. (b) An enlarged view of the IR spectra within 3000–2800 cm⁻¹. The range for b1 is indicated by the dotted rectangle in a. Spot numbers (#1, #2, etc.) are shown in the upper right in each figure. One scale unit on the vertical axis corresponds to an absorbance of 0.02 in b1 and b4–b12, and 0.1 in b2 and b3. c The IR spectra after linear baseline correction of the spectra in b within 3000–2800 cm⁻¹. Spot numbers (#1, #2, etc.) are shown in the upper right in each figure. One scale unit on the vertical axis corresponds to an absorbance of 0.002 in c1 and c12, and 0.001 in c2–c11. The spectra of the adhesive are re-scaled to a quarter in a, one-tenth in b1, and one-hundredth in c1 for clarity

Table 1 IR and Raman parameters for analyzed carbonaceous clots

The IR spectrum of the adhesive used to prepare the samples is shown in Fig. 4a, b1, c1. The bands at 2988, 2944, 2908, and 2876 cm⁻¹ are due to C–H stretching vibrations. The 1757 cm⁻¹ band is due to ester C=O. The 1450 and 1375 cm⁻¹ bands are due to C–H bending of CH₂ + CH₃ and CH₃ groups, respectively. The bands at 1260 and 1180 cm⁻¹ are due to C–N and C–O, respectively. In the 3000–2800 cm⁻¹ region, the most intense band for the adhesive is observed at 2988 cm⁻¹ (shown by a dotted arrow in Fig. 4b1). Based on the 4 cm⁻¹ wavenumber resolution of the IR spectrometer, the bands of the carbonaceous clots (~ 2960, ~ 2925, and/or ~ 2850 cm⁻¹) are considered to be different from the IR bands of the adhesive. The adhesive used to attach the thin section to the glass has been almost completely washed by acetone. The difference between the absorbances at ~ 2960, ~ 2925, and ~ 2850 cm⁻¹ of individual analyzed spots (Table 1) would reflect the amounts of carbonaceous matter and/or its chemical structure. The saturated bands (1400–1000 cm⁻¹) and seven bands (1995, 1870, 1793, 1684, 1610, 1525, and 1492 cm^− 1) in the figures are due to the Si–O bonds of quartz (Nakashima et al. 1995; Ito and Nakashima 2002; Igisu et al. 2006, 2009; Qu et al. 2015). The IR spectrum around 2300 cm⁻¹ is not shown because this range is affected by absorption by CO₂ in the air. In the region of lower wavenumbers of less than 2500 cm⁻¹, the many intense bands of the quartz matrix mask the small signal from the organic matter (Figs. 3 and 4). Hence, we focused on the IR bands with wavenumbers of > 2500 cm⁻¹.

The contribution of the aliphatic CH₂ group to the intensity in the ~ 2925 cm⁻¹ band was examined by comparison with those in the ~ 2925 and ~ 2850 cm⁻¹ band. Figure 5 shows the relationship between the intensities in the ~ 2925 and ~ 2850 cm⁻¹ bands. The strong correlation between these two bands suggests that they are due to the same component(s) and that the ~ 2925 cm⁻¹ band of the carbonaceous clots is mostly due to the aliphatic CH₂ and is less attributable to the aromatic CH₃ groups. These results indicate that the Dresser carbonaceous clots contain small amounts of aliphatics.

Fig. 5

Relationship between the intensities of the carbonaceous clots at ~ 2925 cm⁻¹ and ~ 2850 cm⁻¹. The ~ 2925 cm⁻¹ band is due to asymmetric stretching vibrations of the aliphatic CH₂ and aromatic CH₃, and the ~ 2850 cm⁻¹ band is due to symmetric stretching vibration of the aliphatic CH₂. The yellow diamonds indicate the data for sample 96NP452, while the green diamond indicates that for sample T-58. The error bars indicate the standard deviations of the duplicate measurements

Raman spectral data

The carbonaceous clots in sample 96NP452 have disordered carbonaceous structures, which were observed as ~ 1600 cm⁻¹ (graphite, G) and ~ 1350 cm⁻¹ (disordered, D) bands in the Raman spectra (Fig. 6). Their Raman spectral features are shown in Table 1 and are similar to each other for this sample. They are also consistent with the previously reported data (Ueno et al. 2001, 2004) and those of carbonaceous matter from lower greenschist facies metasediments (Yui et al. 1996). In addition, it is inconsistent with the carbonaceous matter with different thermal alteration phases observed in ~ 3.46 Ga hydrothermal Apex Chert (Olcott-Marshall et al. 2012; Sforna et al. 2014). These Raman and IR spectral features of the carbonaceous clots suggest that the clots mainly consist of randomly arranged layered polyaromatic domains with minor aliphatic C–H bonds. On the other hand, the carbonaceous clot in sample T-58 does not show clear G or D bands. Instead, it shows a sloping baseline (Fig. 6), possibly suggesting fluorescence produced by hydrogen-rich saturated aliphatic hydrocarbons (Marshall et al. 2005). However, this appears inconsistent with the IR data that the absorbance of aliphatic CH₂ in the carbonaceous clots in sample T-58 falls within the range for those in sample 96NP452 (Fig. 5). The inconsistency between the Raman and IR data of the carbonaceous clots in sample T-58 and sample 96NP452 may be due to difference in the amounts of organic matter. Raman spectral data suggest that the analyzed carbonaceous clots at least in sample 96NP452 are moderately graphitized by metamorphism, supporting indigenous origin in ca. 3.5 Ga rocks rather than contaminations.

Fig. 6

Examples of Raman spectra of the carbonaceous clots. The spectra marked as CH₂, CH₂ only, and no CH₂ were obtained from sample 96NP452 (CH₂: both aliphatic CH₂ and CH₃ bonds were observed in micro-FTIR measurements; CH₂ only: CH₂ bond was observed but CH₃ was not; no CH₂: neither CH₂ nor CH₃ were observed). Spot numbers (#3, #12, and #13) corresponding to Table 1 are also shown. The band around 464 cm⁻¹ is due to the quartz matrix of the silica veins

Spatial distributions of aliphatic C–H groups

The R_3/2 values of the carbonaceous clots range from 0.22 ± 0.03 to 0.51 ± 0.01 (Fig. 7 and Table 1). In sample 96NP452, some of the carbonaceous clots containing aliphatic C–H signatures were located relatively close (< 100 μm) to the secondary quartz veins (Fig. 2b, c), while others were far (> 100 μm) from the veins (Fig. 2e, g). Both low (~ 0.3) and high (~ 0.5) R_3/2 values were observed within one carbonaceous clot (Fig. 2f) and also along one secondary quartz vein (Fig. 2b). In sample T-58, one carbonaceous clot with aliphatic C–H signatures was observed in the quartz chalcedony (F2; Fig. 2i), while the other carbonaceous clots in the quartz chalcedony and microcrystalline quartz did not have aliphatic C–H signatures. These results indicate the absence of an apparent distribution pattern among the carbonaceous clots with and without aliphatic C–H signatures. In addition, there was significant variation among the R_3/2 values within a given sample, and even within a given carbonaceous clot. The aliphatic C–H signatures of the carbonaceous clots are spatially heterogeneous on a micrometer to millimeter scale.

Fig. 7

R_3/2 values of the extant prokaryotes, experimentally maturated cyanobacteria, and the carbonaceous clots. a R_3/2 values of the extant bacterial and archaeal cells and lipids (data from Igisu et al. 2012) together with the standard n-alkane samples (n-C₁₀ and n-C₁₈) (data from Igisu et al. 2009). The R_3/2 values for the representative extant untreated bacterial and archaeal cells range from 0.59 ± 0.01 to 0.70 ± 0.01 and from 0.86 ± 0.04 to 0.95 ± 0.02, and those for bacterial and archaeal lipids range from 0.31 ± 0.09 to 0.41 ± 0.01 and from 0.51 ± 0.05 to 0.67 ± 0.08, respectively. b Results of the heating experiments for the non-embedded and silica-embedded cyanobacterial cells (modified from Igisu et al. 2009, 2018). The R_3/2 values for the heated samples are plotted against the heating temperature. All the experimental run products of the silica-embedded samples were obtained after heating for 600 min except the one indicated by a number (the sample marked was heated for 442 min). Others were obtained after heating for 300 min except the one indicated by a number (the sample marked was heated for 600 min). c R_3/2 values of the carbonaceous clots obtained from the Dresser Formation together with the Proterozoic prokaryotic fossils and carbonaceous matter (CM) in stromatolite. The error bars for the carbonaceous clots indicate the standard deviations of the duplicate measurements. The data for Bitter Springs and Gunflint samples were obtained from Igisu et al. (2009), and that for Wumishan sample was obtained from Qu et al. (2015)

Origin of the carbonaceous matter

Post-depositional thermal alteration of the primary organic matter may have altered the R_3/2 values. The varying R_3/2 values of the carbonaceous clots may also be explained by a mixture of different sources of biological organic compounds and/or abiotically produced compounds. The possible occurrence of anaerobic chemoautotrophic microorganisms was suggested by the carbon isotopic signatures of the carbonaceous matter in the silica veins in the Dresser Formation (Ueno et al. 2004). The carbon isotopic signatures provide the potential chemical evidence of microbial activity in the environment at the time, but do not necessarily offer clues for distinguishing whether the products are derived from the microbial functions of the bacterial and/or the archaeal populations. The existence of sulfate-reducing bacteria has been proposed by the pyrolysis fragments of the organic matter (Derenne et al. 2008). However, the potential risk of post-depositional and experimental contamination has been found in the pyrolysis analysis of the isolated Archean organic matter (e.g., Bourbin et al. 2012). The presence of ancient methanogenic archaea has been inferred from the fluid inclusions in the silica veins containing ¹³C-depleted methane (Ueno et al. 2006), although the abiotic origin of the methane has also been proposed (Sherwood-Lollar and McCollom 2006). We will discuss the origin of the varying R_3/2 values for the carbonaceous clots in the following order: (1) abiological origin, (2) effect of thermal alteration, and (3) biological origin.

Abiological origin

It is known that an FTT synthesis under specific hydrothermal conditions (e.g., in ultramafic rock-associated hydrothermal systems) may produce organic compounds (e.g., Yanagawa and Kobayashi 1992; Holm and Andersson 1998; McCollom et al. 1999; McCollom and Seewald 2006). The organic products of the FTT synthesis are mainly methane and normal alkanes with only a small amount of branched hydrocarbons because the relative abundance of these hydrocarbons decreases with increasing carbon number in the alkyl chains (McCollom and Seewald 2006, 2007). McCollom and Seewald (2006) reported that isobutane was the only branched product of FTT synthesis, but the yield of isobutane (~ 0.02 mmol/kg) was much lower than those of the normal alkanes (e.g., 0.64 mmol/kg for n-butane). Therefore, it is supposed that mixing of the organic matter produced by the synthesis (mainly the short-chain products) increases the R_3/2 value. Although we did not directly determine the carbon number of the aliphatic components in the carbonaceous clots, the R_3/2 values for the normal C₁₀ alkane are estimated to be ~ 0.6 (Fig. 7a; data from Igisu et al. 2009). The relative abundance of the saturated > C₁₀ hydrocarbons produced by the FTT synthesis has been shown to be several orders of magnitude lower than those of the saturated < C₆ hydrocarbons (McCollom and Seewald 2006). Thus, it appears to be difficult to explain the observed R_3/2 values of the carbonaceous clots (0.22–0.51) at this stage based solely on abiotically produced organic matter. Moreover, it is unlikely that the FTT synthesis of organic compounds occurs during the formation of the silica veins. This is because native Fe–Ni minerals and other preferred catalysts for FTT reactions were not observed in the thin sections of the present study.

Effect of thermal alteration

Thermal alteration may increase or not change the R_3/2 value of organic matter. This indicates the preferential loss of the methylene group and/or formation of branched CH₃ in the aliphatic structures during the degradation of the organic matter, as revealed by FTIR measurements of coal (Wang and Griffiths 1985; Ibarra et al. 1996), and heating experiments performed on kerogens (Robin et al. 1977; Huang and Otten 1998) and extant cyanobacteria (Igisu et al. 2009, 2018). Recently, a kinetic approach applied to experimental maturation of cyanobacteria revealed that decreasing rate constants for aliphatic CH₃ were generally similar to or lower than aliphatic CH₂ at 250–450 °C under both non-embedded and silica-embedded conditions. This suggests that aliphatic CH₃ groups degraded more slowly than aliphatic CH₂ groups (Igisu et al. 2018). The metamorphic grade of basaltic greenstones surrounding the silica veins is generally below the greenschist facies (< 350 °C; Kitajima et al. 2001). Thermal alteration may therefore explain the varying R_3/2 values, but is unlikely to be the cause of the substantially lower R_3/2 values when compared with those of primary organic matter.

In addition, the coexistence of carbonaceous matter with different degrees of thermal alteration may cause variation in the R_3/2 value and the presence/absence of aliphatic C–H signatures in the carbonaceous clots. There is the possibility of contribution of some molecules and/or some functional groups, which are very resistant to heat even if they suffered from the same heating processes, to the carbonaceous clots. Assuming that some parts of the molecules and groups were graphitized but some parts were not, there may be differences in the R_3/2 values and Raman spectral features observed in the carbonaceous clots in samples 96NP452 and T-58. Such heterogeneity was reported in Phanerozoic eukaryotic fossils (Qu et al. 2015).

Biological origin

The R_3/2 values for the carbonaceous clots in the present study (0.22–0.51) generally fall within the range for extant bacterial and archaeal cellular compounds (Fig. 7a, c). The observed low R_3/2 values below ~ 0.4 cannot be explained by the degraded cellular compounds other than degraded bacterial lipids, possibly representing the relatively bacteria-dominated fractions of the microbial communities. On the other hand, there are carbonaceous clots with higher R_3/2 values (> ~ 0.4). Such carbonaceous clots may be caused by thermal alteration of organic matter with lower R_3/2 values or be derived from the archaea-dominated assemblages as well as compounds with short-chain aliphatic components. Such a possible contribution of aliphatic components to geomacromolecules by living organisms without resistant aliphatic precursors (e.g., algaenan) has been suggested by Gupta et al. (2007a, 2007b, 2014).

A comparison of the R_3/2 values of the carbonaceous clots with those of other Proterozoic prokaryotic fossils and carbonaceous matter in the stromatolites (~ 830 Ma Bitter Springs Group, ~ 1500 Ma Wumishan Group, and ~ 1900 Ma Gunflint Formation) reveals that the carbonaceous clots have similar R_3/2 values as other Proterozoic samples (Fig. 7c), while the Raman spectra of the carbonaceous clots show a more graphitized nature than those of the Proterozoic microfossils (i.e., narrower bandwidth of especially the D band and higher D/G band intensity ratio). This suggests the possible interpretation of the prokaryotic origin of the carbonaceous clots, assuming that the R_3/2 value either increases or remains unchanged during the thermal degradation processes.

Given that the aliphatic moieties in the carbonaceous clots might indicate the presence of cellular organic aggregates of the prokaryotic population in the ancient microbial communities, the IR spectroscopic features of the carbonaceous clots likely suggest a contribution from the bacterial lipids in the ancient microbial communities associated with the hydrothermal activities. Our results thus afford additional important insights into the indigenous microbial communities in the ~ 3.5 Ga Dresser seafloor hydrothermal system that consists of not only archaeal components, but also significant bacterial populations. This supports the previous pyrolysis analysis of the isolated organic matter, suggesting microbial origin of the organic matter (Derenne et al. 2008; Duda et al. 2018).

Conclusions

FTIR microspectroscopic analysis was conducted on carbonaceous clots taken from two silica veins in the ~ 3.5 Ga Dresser Formation. We found the following:

• 1. The chemical structure of the carbonaceous clots comprises randomly arranged layered polyaromatic domains with very minor aliphatic straight chain moieties.

• 2. It appears unlikely at this stage that the observed R_3/2 values of the carbonaceous clots and the mineral assemblages in the analyzed samples do not support the only abiogenic origin of the carbonaceous clots.

• 3. Thermal alteration may increase or not change the R_3/2 value of organic matter. Therefore, it may explain the varying R_3/2 values, but is unlikely to be the cause of the substantially lower R_3/2 values when compared with those of primary organic matter.

• 4. The R_3/2 values of the carbonaceous clots (0.22–0.51) generally fall within the range of the values for extant prokaryotic cellular components. In particular, the low R_3/2 values (< ~ 0.4) could have originated from bacterial lipid, assuming that the R_3/2 value would either increase or remain unchanged during the thermal degradation processes. This suggests the preservation of remnants of the bacterial population, mainly as aliphatic C–H groups, in the macromolecular structure of the analyzed carbonaceous matter.

The new FTIR microspectroscopic data of the carbonaceous clots combined with the previous studies based on their isotopic features (e.g., Ueno et al. 2001, 2004, 2006) may possibly suggest that both bacteria and archaea were colonized in the ~ 3.5 Ga Dresser hydrothermal system.