Introduction
Stromatolites are biologically induced organomineral layered sedimentary formations and their metabolic activities induce conditions for precipitation (Dupraz et al. Reference Dupraz, Reid, Braissant, Decho, Norman and Visscher2009). Its structure can host diverse microbial communities (Baumgartner et al. Reference Baumgartner, Spear, Buckley, Pace, Reid, Dupraz and Visscher2009), including bacteria, algae, and archaea (Lepot et al. Reference Lepot, Benzerara and Brown2008). Some bacteria within stromatolites may perform processes that can affect radiocarbon analyses (Barker and Fritz Reference Barker and Fritz1981). For example, methane-producing bacteria can introduce carbon from different sources, while sulfate-reducing bacteria can influence carbonate precipitation (Andres et al. Reference Andres, Sumner, Reid and Swart2006; Visscher et al. Reference Visscher, Reid and Bebout2000). These microbial processes are intimately linked with mineral formation (Reid et al. Reference Reid, Dupraz, Visscher, Sumner, Krumbein, Paterson and Zavarzin2003) but are also influenced by interconnected local factors such as light, pH, and temperature (Bowlin et al. Reference Bowlin, Klaus, Foster, Andres, Custals and Reid2012). Therefore, understanding the microbial community and their metabolic activities provides insights into mineralogical properties and potential variations in 14C content.
The reservoir effect on stromatolites is particularly important in calcareous regions rich in ancient carbonates. This effect is well-documented such as in Carreira et al. (Reference Carreira, Marques, Graça and Aires-Barros2008), where regions high in limestone are largely devoid of radiocarbon due to prolonged isolation from the atmospheric carbon cycle, hence, adding carbon with minimal 14C into surrounding waters. Stromatolites may assimilate this carbon during their growth, causing an overestimated age during radiocarbon dating due to the inclusion of older carbon sources that lack radiocarbon (Brook et al. Reference Brook, Cherkinsky, Railsback, Marais and Hipondoka2013; Jull et al. Reference Jull, Burr and Hodgins2013). As seen, organisms residing in areas with low carbonate influence are not subject to this effect (Macario et al. Reference Macario, Alves and Carvalho2016).
The study of the formation of these structures is very important because it is believed that stromatolites have dominated 80% of all geological formation on planet Earth and are the evidence of the earliest life found, dating to around 3.5 billion years (Grotzinger and Knoll Reference Grotzinger and Knoll1999; Hofmann Reference Hofmann1969, Reference Hofmann1973; Riding and Awramik Reference Riding and Awramik2000; Vologdin Reference Vologdin1962; Walter Reference Walter1976). Several paleoclimate reconstruction studies have been performed using the geochemistry proxies stored within the carbonate matrix over the last years (Bahniuk Reference Bahniuk2013; Birgel et al. Reference Birgel, Meister, Lundberg, Horath, Bontognali, Bahniuk, De Rezende, Vasconcelos and Mckenzie2015; Carvalho et al. Reference Carvalho, Oliveira, Macario, Guimarães, Keim, Sabadini-Santos and Crapez2017; Iespa et al. Reference Iespa, Iespa and Borghi2012; Silva e Silva and Senra 2000; Vasconcelos and McKenzie Reference Vasconcelos and McKenzie1997; Vasconcelos et al. Reference Vasconcelos, Warthmann, McKenzie, Visscher, Bittermann and van Lith2006). Radiocarbon dating is an important tool for a better understanding of these records. However, stromatolite has no well-defined pretreatment for being dated by the 14C-AMS technique.
In general, the 14C-AMS dating protocol for carbonate samples is based on a simple removal of the outer layer with a sandblaster or a chemical etching pretreatment using hydrochloric acid (HCl), to remove possible external contamination. Then, hydrolysis is performed by using orthophosphoric acid (H3PO4), to convert the carbonate into carbon dioxide, by the following equation (Burman et al. Reference Burman, Gustafsson, Segl and Schmitz2005):
$${\rm {3CaCO_3} + \;{2H_3}{PO_4} \rightarrow {Ca_3}{(PO_4)_2} \;+ 3H_2O\,+\,{3CO_2}}$$
This acid does not convert organic matter to carbon dioxide, but some authors affirm that through this reaction it is possible to produce this gas and/or other molecular gaseous species (Bowen Reference Bowen1966, Reference Bowen1991; Epstein et al. Reference Epstein, Buchsbaum, Lowenstam and Urey1951, Reference Epstein, Buchsbaum, Lowenstam and Urey1953; Falster et al. Reference Falster, Delean and Tyler2018; Oehlerich et al. Reference Oehlerich, Baumer, Lücke and Mayr2013; Weber et al. Reference Weber, Deines, Weber and Baker1976). Under certain conditions, CO2 can be produced from organic matter in the acid environment if the organo-mineral interaction bonds are broken and the organic matter is more accessible to hydrolysis reactions which is not the case because H3PO4 is a weak acid that is widely used to obtain carbon dioxide from carbonate samples for radiocarbon purposes (Wacker et al. Reference Wacker, Fülöp, Hajdas, Molnár and Rethemeyer2013a, Reference Wacker, Lippold, Molnár and Schulz2013b). Concerning stromatolites structure and composition it is necessary to consider the question: what is the impact of performing or not performing a chemical pretreatment to remove organic matter from the samples? Chaduteau et al. (Reference Chaduteau, Ader, Lebeau, Landais and Busigny2021) realized a similar test in δ13C analysis of carbonate samples and proved that it makes a difference in the results. Key et al. (Reference Key, Smith, Phillips and Forrester2020) also assess the impact of the most used pretreatment methods of organic matter removal on δ13C and δ18O ratios for carbonate samples from taxonomic groups with complex mineralogies. The main purpose of this work is to attempt to answer this question by presenting F14C results for 14C-AMS analysis, testing the extraction of organic matter and acid etching on subfacies of a stromatolite specimen which can have different mineral compositions in its growth layers.
Materials and methods
The stromatolite specimen used in this work was collected in the Salgada Lagoon (21°54'10''S e 41°00'30''W) in January 2016. This lagoon is located in the coastal area of Rio de Janeiro State, Brazil, being part of the Paraíba do Sul deltaic river complex (Lamego Reference Lamego1955). This is one of the few hypersaline lagoons that still have stromatolite formation nowadays. Initially, X-ray microtomography (micro-CT) was performed on the raw material to identify five stromatolite subfacies the analysis was performed at the Nuclear Instrumentation Laboratory at Rio de Janeiro State University (LIN-COPPE/UFRJ). This is a non-destructive technique that allows the inspection of the internal structure of the sample and the Phoenix Vtomex|m GE micro-CT system was used. Reconstruction of the sample’s volume was possible through software datos x (Version 2.5.0) and CTAn software (Version 1.18.4.0) was used to perform the image segmentation and quantitative and qualitative analyses. The segmentation allowed the division of the image, which was in different gray levels, into just black and white. The micro-CT examination of the raw material allowed the detailed identification of five subfacies as shown in Figure 1. SB01 layer is the oldest one, which corresponds to the lagoon floor, and SB05 is the most recent, on the top of the stromatolite. The subfacies structure showed different degrees of porosity, as follows: SB01=18.98%, SB02=40.73%, SB03=35.74%, SB04=11.85%, and SB05= 8.99%.
Stromatolite Subfacies limits from the bottom (SB01) to the top (SB05) obtained through micro-CT analysis.
The elemental and mineral composition were analyzed by X-ray fluorescence (XRF) and X-ray diffraction (XRD), respectively, before starting treatments to verify whether any differences could impact the radiocarbon results. For 14C pretreatment, the three intermediary samples were submitted to different organic matter treatments: hydrogen peroxide (H2O2), sodium hypochlorite (NaOCl), and control, without organic matter removal. The latter is a common carbonate protocol for radiocarbon dating. Then, the samples were subdivided into two groups: with and without etching using hydrochloric acid (HCl) and proceeded to the acid hydrolysis and graphitization protocol for 14C-AMS dating at the Radiocarbon Laboratory of the Fluminense Federal University (LAC-UFF). From the results obtained in the present work, we expect to better understand the stromatolite sample handling to improve the protocols for radiocarbon dating of biogenic carbonate samples by eliminating possible organic contamination. The stromatolite analyzed in this work was 10 cm high and 30 cm in diameter (Figure 2A).
(A) stromatolite sample of Salgada Lagoon; (B) sample extraction.
Five subsamples were extracted using a 7 mm wall drill (Figure 2B) and crushed in an agate mortar to be homogenized at LAC-UFF to perform XRF and XRD composition analysis.
XRF analysis was used to determine the elemental composition of subsamples. The sample preparation and analysis were made at Laboratório de Oceanografia Operacional e Paleoceonografia at Fluminense Federal University (LOOP-UFF). A benchtop Epsilon 3X energy-dispersive X-ray fluorescence spectrophotometer (EDXRF) from PANalytical was used. It has an X-ray tube with a silver anode and a 50 µm beryllium window, a power of 9 W, a current of 1 mA, and a voltage of 50 kV, using air and helium as carrier gas. The detector is a high-resolution Silicon Drift detector, typical of 135 eV, and a thin window of 8 µm (Be).
XRD analysis was used to investigate the crystalline structure of the powdered samples. The sample preparation and analysis were undertaken at the Geochemistry Department at Fluminense Federal University. A BrukerD2 Phaser (Cu Kα radiation) model was operated in a Bragg–Brentano θ/θ configuration, with the diffraction patterns being collected in a flat geometry with a range between 3 and 100 degrees, steps of 0.02 degrees, and accumulation time of 3.0 s per step. Phase identification was done with EVA® software. The XRD data were refined following the Rietveld method using the DIFFRAC.SUITE TOPAS® software (McCusker et al. Reference Mccusker, Von Dreele, Cox, Louer and Scardi1999; Toby Reference Toby2006).
Afterwards, the three intermediary subfacies were selected for the organic matter removal pretreatment test for 14C dating. For the test, all samples were prepared and analyzed at LAC-UFF. Each subfacie was subdivided to perform the organic removal pretreatment test as follows: [T-A]: Samples without organic matter removal; [T-B]: organic matter removal using a solution of 0.7 M of sodium hypochlorite that has pH ∼11 (NaOCl, Isofar, Brazil, 4–6% purum p.a., CAS 7722-84-1, Lot. n. 245/2015), and [T-C]: a solution of 8.8 M of hydrogen peroxide that has pH ∼5 (H2O2, Isofar, Brazil, 30% purum p.a., CAS 7681-52-9). NaOCl and H2O2 were selected because they are powerful oxidation agents and are the most used in organic matter treatments for stromatolite samples in geochemical analysis. A treatment diagram is presented in Figure 3. The treatment T-A-1 can be referred to as a “control” treatment not in the classical sense in terms of expected values. It can be referred to as a control treatment compared to the others since the sample was untreated. The T-A-2 treatment consists of etching the sample without organic matter removal. For both organic matter removal treatments, 3 ml of NaOCl or H2O2 were added to approximately 100 mg of dry sample and left to react for 1h at 60 °C in the dry bath. Each process was repeated 5 times and at the end, samples were washed 3 times with ultrapure water and dried at 90ºC to determine the mass loss percentage.
Steps of the treatment tests performed.
After the organic matter removal samples were split in two aliquots to proceed with the carbonate etching, as shown in Figure 3. An aliquot of the samples went straight to CO2 conversion without etching addition (T-B-1 and T-C-1), and another aliquot, with approximately 40 mg, was chemically treated with 1.7 mL of 0.1M HCl and stayed overnight at 90ºC in the dry bath (T-A-2, T-B-2, and T-C-2) for a 25% etching, following the LAC-UFF etching standard protocol. At the end of the process, samples were dried at 90ºC and proceeded to CO2 conversion.
The CO2 was obtained by acid hydrolysis using the orthophosphoric acid (H3PO4, 85%) protocol. After CO2 purification using temperature traps in the vacuum line, graphitization took place at 550°C (7h) with TiH2, Zn and iron used as catalysts in an inner tube (Macario et al. Reference Macario, Alves, Moreira, Oliveira and Chanca2017a, Reference Macario, Oliveira, Moreira, Alves and Carvalho2017b). Graphitized samples were measured in a NEC 250 kV single stage (SSAMS) system (Macario et al. Reference Macario, Gomes, Anjos, Carvalho, Linares, Alves, Oliveira, Castro, Chanca, Silveira, Pessenda, Moraes, Campos and Cherinsky2013, Reference Macario, Oliveira, Carvalho, Santos, Xu, Chanca, Alves, Jou, Oliveira, Brandão, Moreira, Muniz, Linares, Gomes, Anjos, Castro, Anjos, Marques and Rodrigues2015). Typical currents were 50 μA12C–1 (measured at the low energy Faraday cup). Graphite standard and calcite blanks yielded average 14C/13C ratios of 6×10–13 and 7×10–13, respectively. The average machine background was around 50 kyr for the unprocessed graphite, while the average precision ranged from 0.3 to 0.5%. Data analyses were carried out on LACAMS software (Castro et al. Reference Castro, Macario and Gomes2015).
Results and discussion
XRF analysis
From XRF analysis the elemental composition of subsamples was determined showing a low concentration of Fe and Mn in the stromatolite chemical composition when compared to Mg concentration was observed in XRF analysis results (Table 1). For more detailed information and to obtain the mineralogical composition we conducted XRD analysis.
XRF analysis results from stromatolite samples
XRD analysis
As shown by Moreira et al. (Reference Moreira, Macario, Guimarães, Dias, Araujo, Jesus and Douka2020), fossils of marine organisms are likely to undergo dissolution and recrystallization. During this process, carbonate exchanges may occur, which leads to the entry of exogenous carbon, thus altering the radiocarbon concentration. The same could occur to stromatolite material during the formation of layers. Therefore, we carried out an XRD analysis to verify the crystalline composition of our samples. The results have shown that the subfacies presented are mainly calcite composition but with three types according to the d104 typical peak. There is a mixture of calcite and Mg-calcite with different proportions. Table 2 presents d104 (A) results and MgCO3 (%) concentrations, where SB05 and SB04 presented low MgCO3 concentrations, 12% and 16%, respectively. SB02 presented 24% of MgCO3 and SB03 and SB01 presented high and close concentrations, 42% and 41%, respectively. It is important to mention that the balance between precipitation and dissolution of MgCO3 is a process that is influenced by the metabolic activity of the microbial community of the stromatolite during its growth, as well as by the physicochemical properties of the surrounding water, such as pH, temperature, saturation state, and the concentration of Mg2+ and carbonate ions in the environment. The concentration of MgCO3 in the growth layers can give information on the sedimentation environment in the past, but that is not the main goal of this work and will not be discussed in the present paper. Figure 4 shows the detailed d104 peaks from XRD analysis. Samples SB04 and SB05 have more calcite (the two blue peaks on the left of Figure 4). Otherwise, samples SB01, SB02, and SB03 are more Mg-calcite enriched. Samples SB01 and SB03 stand out as very Mg-enriched calcite samples (the black and green peaks on the right of Figure 4). The d104 values below 2.99 are typical Mg-enriched calcite samples. This appears to be related to a change in the crystallographic unit cell from the Mg-calcite to be like the one from the mineral kutnohorite (Ca (Mn, Fe, Mn)(CO3)2) a rare mineral from the dolomite group and not present here (Graff Reference Graff1961). Some authors also had a large shift in diffractogram peak (from 2.99A to 2.94A related to Mg-calcite) as a result of a very high-magnesian calcite (VHMC) or a disordered dolomite, as pointed out by Zhang et al. (Reference Zhang, Xu, Konishi and Roden2010).
Stromatolite subfacies d104 (A) results and MgCO3 (%) concentrations, where d104 (A) higher than 2.99 are related to low MgCO3 (%) concentrations (SB04 and SB05) and d104 (A) lower than 2.99 have high MgCO3 (%) concentrations (SB03, SB02 and SB01)
Comparison of diffractograms presenting the d104 calcite characteristic peak in Angstrom (A).
As can be observed in Table 3, subfacie SB04 had the lowest concentration of Mg-calcite (36%) and the highest concentration of calcite (57%). The highest concentration of Mg-calcite and lowest concentration of calcite were found in SB01 85% and 5% respectively. SB02 and SB03 have similar mineralogical values, although SB03 has a higher Mg-calcite concentration. Carbonate precipitation and recrystallization in stromatolites can manifest in various mineral forms, such as calcite, aragonite, or dolomite, and are influenced by physicochemical, biological, and environmental conditions. The subfacie SB02, which presented 12% of aragonite in its composition, was probably subjected to different conditions during the layer growth. If occurs, the recrystallization in the same layer while studying stromatolite well-defined sequential growth layers seems to not impact the radiocarbon dating discussion.
Mineralogical pattern of Salgada lagoon stromatolite layers. ND: not detected; trace: value below 1%
Chemical treatments
After the previous characterization, three inner layers were selected to move forward to the organic matter removal test: SB02 and SB03 which presented similar composition, and SB04 which presented lower Mg-calcite and higher calcite concentrations. The percentage of mass loss after the process is indicated in Table 4. Concerning weight loss during the organic matter removal, NaOCl bleaching presented similar results for all subfacies and a smaller weight loss than H2O2. This may be caused by the oxidizing strength of H2O2 compared to NaOCl, which changes the inorganic matrix of the samples beyond the consumption of the organic matter. The same results were found by Chaduteau et al. (Reference Chaduteau, Ader, Lebeau, Landais and Busigny2021), in which H2O2 altered the preservation of the carbonate, showing inorganic matrix dissolution, which did not occur while using NaOCl. The weight loss during the H2O2 use suggests dissolution in all samples. As pointed out by Key et al. (Reference Key, Smith, Phillips and Forrester2020), many researchers reported that H2O2 is strongly corrosive to CaCO3 and that H2O2 at ∼ 30% has a pH of ∼5 which can promote the dissolution of carbonates. The highest weight loss (57%) was observed after the H2O2 treatment on SB04 which presented the highest concentration of calcite (57%). Some other factors can increase the susceptibility of dissolution as the Mg content and the duration of the treatment but our study did not evidentiate this.
Mass loss during organic matter chemical pretreatment, where [T-B] is the organic matter removal treatment using sodium hypochlorite (NaOCl) and [T-C] using hydrogen peroxide (H2O2)
Following Reimer et al. (Reference Reimer, Brown and Reimer2004), radiocarbon results are reported by F14C values determined after the treatments and are shown in Table 5. It was not possible to determine the results of SB02 after T-C-1 due to probable loss during CO2 purification in the vacuum line. As a result, we would expect a reduction in F14C values for all samples after the organic matter removal step since it should remove any recent carbon that could have been incorporated into the original sample matrix. Comparing the results observed for group 1 (organic matter removal without etching) it is possible to observe this behavior when comparing T-A-1 (Control) and T-B-1 (NaOCl) results for all layers. When comparing T-C-1 (H2O2) with the control T-A-1 group the layer SB03, which has more Mg-calcite, presented a reduction in F14C values. On the other hand, layer SB04 presented a higher F14C value.
Radiocarbon results (F14C) and its uncertainty after pretreatment tests. [T-A]: Control samples without organic matter removal; [T-B]: organic matter removal using sodium hypochlorite (NaOCl) and [T-C]: hydrogen peroxide (H2O2). The numbers 1 and 2 represent results for samples without and with HCl etching, respectively
The F14C results after HCl etching (group 2) for Mg-calcite concentrated samples (SB02 and SB03) were similar to those without etching. The CaCO3 concentrated (SB04) presented F14C results higher than the control (T-A-2) in both combined tests: NaOCl + etching (T-B-2) and H2O2 + etching (T-C-2).
In the graph (Figure 5) the results represent the F14C levels for each treatment [T-A, T-B, and T-C], grouped by subfacie. The black triangles represent results after treatments with no etching and the gray circles represent results after etching. The F14C results for SB02 and SB03 are quite similar with low discrepancy. SB04 had a higher discrepancy, which seems to be an effect of its mineral composition containing higher calcite and low Mg-calcite. The results with etching (gray) for SB02 and SB04 are higher after all treatments, the same behavior is not observed for SB03 which has the highest Mg-calcite concentration.
F14C results for each treatment [T-A, T-B, and T-C], grouped by subfacie. The black triangles represent results with no etching treatment and the gray circles represent results after etching.
Many works discussed the effects of previous treatments on stable isotope ratios for biogenic and inorganic carbonate samples to remove organic matter showing its dependence on the mineralogical matrices, duration of the treatment, and the importance of avoiding isotopic fractionation or isotopic exchange during the treatments (Key et al. Reference Key, Smith, Phillips and Forrester2020; Zhang et al. Reference Zhang, Lin and Yamada2020). Since the selection of chemical pretreatment to be used depends on the mineral composition, wrong choices can also affect radiocarbon results, especially for stromatolite samples that could have a complex mineralogy. Carvalho et al. (Reference Carvalho, Oliveira, Macario, Guimarães, Keim, Sabadini-Santos and Crapez2017) dated Mg-calcite-rich stromatolite samples by applying H2O2, which was the most used treatment for this kind of sample, and afterward followed the radiocarbon carbonate etching protocol. From the present study, it is possible to say that it was a good choice to apply the H2O2 protocol to Mg-calcite-rich samples as it doesn’t seem to impact results in the case of high Mg-calcite composition. Additionally, comparing the present work with Chaduteau et al. (Reference Chaduteau, Ader, Lebeau, Landais and Busigny2021) who also observed shifted δ13C values for calcite-rich samples after being treated with H2O2, showing it is less efficient for removing organic matter than NaOCl, it’s possible to say that we have observed the same for SB04 subfacie in the present work. Furthermore, H2O2 pretreatment can make the remaining organic matter more reactive to H3PO4, which is used to convert carbonate into CO2 for radiocarbon analysis and could influence F14C results. On the other hand, NaOCl pretreatment presented good results for any stromatolite mineral matrix. The F14C results after etching seem to be dependent on the mineral matrix but were not significantly different for the Mg-calcite-rich layer.
Conclusions
Before radiocarbon dating stromatolite samples it is important to identify the mineralogical composition before organic matter removal. Even though H2O2 could not influence the results of Mg-calcite concentrate samples, it can promote the dissolution of the carbonate matrix showing an expressive weight loss for the calcite concentrated layer. The use of NaOCl appears to have been effective in preserving more material than H2O2 independent of the mineralogical composition of the stromatolite layers. The F14C results after HCl etching for Mg-calcite concentrated samples were similar to those without etching suggesting that the HCl etching does not impact the results in this case. Since stromatolites are biogenic carbonate samples with some porosity, the organic matter removal has shown to be more important than the etching procedure. The CaCO3 concentrated sample presented F14C results higher than the control in H2O2 pretreatment and both combined tests: NaOCl + etching and H2O2 + etching. In this case, NaOCl is more indicated to be used as chemical pretreatment for radiocarbon analysis purposes independent of the mineral matrix of samples.
Acknowledgments
The authors would like to thank Brazilian financial agencies CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico, Grant 315514/2020-5 and Grant 426338/2018-9 to Carvalho, C.; Grant 317397/2021-4 to Macario, K., Grant 309412/2019-6 to Oliveira, F.) and INCT-FNA 464898/2014-5), FAPERJ (Grant E-26/202.714/2018 to Carvalho, C.; Grant E-26/202615/2019 and E-26/200.540/2023 to Macario, K.; Grant E-26/201.320/2022 to Oliveira, F.), Project CLIMATE-PRINT-UFF (Grant 88887.310301/2018-00) for their support. This study was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - Brasil (CAPES) - Finance Code 001. We also thank CNPq for the PhD’s scholarship of Oliveira, M.I. We have a special acknowledgement to Professor Cleverson Guizan who collected and donated the stromatolite specimen making this research possible and the anonymous reviewers who identified some aspects that could be improved to clarify the work done.
