INTRODUCTION
For the last three decades, 14C dating of sample material at molecular level (e.g., n-alkanes, alkenones, lignin oxidation products, fatty acids methyl esters, phospholipid fatty acids), so-called compound-specific radiocarbon analysis (CSRA), has been increasingly used. In particular, n-alkanes (leaf wax molecules) are proven biomarkers to detect vegetation changes in a landscape and the environment, as shown by studies in lake (Gierga et al. Reference Gierga, Hajdas, van Raden, Gilli, Wacker, Sturm, Bernasconi and Smittenberg2016) as well as in marine sediments (Pearson and Eglinton Reference Pearson and Eglinton2000), and loess (Haas et al. Reference Haas, Bliedtner, Borodynkin, Salazar, Szidat, Eglinton and Zech2017).
However, CSRA has two main challenges:
-
(1) The most significant is the input of unknown amounts of extraneous carbon to a sample with unknown mass and unknown radiocarbon content (here expressed as F14C-values as described in Hemingway (Reference Hemingway2021) during the analytical procedure (Shah and Pearson Reference Shah and Pearson2007). This so-called process or procedural blank—if large enough compared to the sample size—alters the sample’s radiocarbon age; hence a correction is mandatory. The preparation of a single substance to be dated requires many individual analytical steps, including compound extraction by Soxhlet, purification with solvents, isolation of single compounds by preparative gas chromatography and their concentration, and finally the radiocarbon determination by accelerator mass-spectrometry (AMS). Each step could potentially add extraneous carbon to the compound causing an unknown blank contribution. To make matters worse, the individual blanks of each step may not be constant throughout time. They might change with even slight modifications in the sample-processing procedure and instrument configurations. Hence, finding the total blank is challenging (Hanke et al. Reference Hanke, Wacker, Haghipour, Schmidt, Eglinton and McIntyre2017).
-
(2) CSRA samples typically have low amounts of carbon (< 100 µg C). That makes 14C dating at molecular level extremely sensitive to a blank contribution (Mollenhauer and Rethemeyer Reference Mollenhauer and Rethemeyer2009). The smaller the mass of the sample, the more the blank affects the 14C date. So, it is important to find a minimum mass carbon sample limit above which CSRA samples can be considered and below which the uncertainty becomes too great. Studies by Pearson et al. (Reference Pearson, McNichol, Schneider, von Reden and Zheng1997), Sarkar et al. (Reference Sarkar, Wilkes, Prasad, Brauer, Riedel, Stebich, Basavaiah and Sachse2014), and Shah and Pearson (Reference Shah and Pearson2007) give a concentration of 25 µg C as the limit. It is technically possible to measure carbon by accelerator mass spectrometry (AMS) in the range from 2 µg to 10 µg C (Santos et al. Reference Santos, Southon, Griffin, Beaupre and Druffel2007) when using a gas ion source. However, according to Ziolkowski and Druffel (Reference Ziolkowski and Druffel2009) 14C-dates of samples with less than 5 µg C are not interpretable.
There are diverse ways to correct the described blank contribution. One of these approaches is the direct determination of the process blank for each analytical step. Here, blank material runs through chemical analyses and preparative gas chromatography in the same way a sample would (Ziolkowski and Druffel Reference Ziolkowski and Druffel2009). An indirect approach is to find the total blank contribution of the whole analytical procedure independent of the individual procedural steps.
Samples and standards (substances with known F14C or known 14C age) have the same blank contribution provided they have undergone the same analytical procedure. Therefore, standards with different sample sizes and known F14C can be used to capture the mass and isotopic signature of the blank (Pearson et al. Reference Pearson, McNichol, Schneider, von Reden and Zheng1997; McNichol et al. Reference McNichol, Ertel and Eglinton2000; Hwang and Druffel Reference Hwang and Druffel2005; Shah and Pearson Reference Shah and Pearson2007).
A further standard dilution method to determine the blank mass, combined with a mass balance equation to find the isotopic signature of the blank reduced the uncertainty of the blank assessment (Hwang and Druffel Reference Hwang and Druffel2005). For samples with low mass Hwang and Druffel (Reference Hwang and Druffel2005) recommended a standard dilution method with a minimum of two different standards, so that the mass balance calculation is no longer necessary.
This kind of blank assessment was later taken up by Sun and co-worker (Reference Sun, Meyer, Dolman, Winterfeld, Hefter, Dummann, McIntyre, Montluçon, Haghipour, Wacker, Gentz, van der Voort, Eglinton and Mollenhauer2020) and developed into a simpler assessment of the total process blank. They devised a standardized blank assessment to determine unknown mass and F14C value of the blank contribution. The authors used modern 14C material (n-alkanoic acid and vanillin) extracted from apple peel and woodchips, respectively. As fossil standards Sun et al. (Reference Sun, Meyer, Dolman, Winterfeld, Hefter, Dummann, McIntyre, Montluçon, Haghipour, Wacker, Gentz, van der Voort, Eglinton and Mollenhauer2020) used n-alkanoic acids extracted from Messel shale and the commercially available standard material ferulic acid (Sigma-Aldrich, Prod. No.12,870-8, Lot STBB6360).
Correcting the blank contribution uses the mass balance equation that assumes a constant contamination independent of the amount of the sample (Ziolkowski and Druffel Reference Ziolkowski and Druffel2009). The measured F14C of a sample (F14Csample) consists of the true F14C of the unaltered sample (F14Ctrue) and the F14C of the blank (F14Cblank). It can be expressed as:
$${F^{14}}{C_{sample}} = {F^{14}}{C_{true}}*{\frac{{{m_{true}}}}{{{m_{sample}}}}} + {F^{14}}{C_{blank}}*{\frac{{{m_{blank}}}}{{{m_{sample}}}}}$$
where the mass of the sample is the sum of the true sample mass and the mass of the blank
${m_{sample}} = {m_{true}} + {m_{blank}}$
.
According to Sun et al. (Reference Sun, Meyer, Dolman, Winterfeld, Hefter, Dummann, McIntyre, Montluçon, Haghipour, Wacker, Gentz, van der Voort, Eglinton and Mollenhauer2020) the mass balance equation can be rearranged showing that F14Csample only depends on 1/msample if all other terms are constant and F14Csample can be calculated when F14Cblank and mblank are known:
$${F^{14}}{C_{sample}} = ({F^{14}}{C_{blank}}{\rm{*}}{m_{blank}} - {F^{14}}{C_{true}}{\rm{*}}{m_{blank}}){\rm{*}}{\frac{1}{{{m_{sample}}}}} + {F^{14}}{C_{true}}$$
Hence, aliquots of different sized standard (6–151 µg C) were used and graphically evaluated by plotting the results of 1/msample vs. F14Csample to determine F14Cblank and mblank.
The usage of standard material of different F14C value (modern and fossil) results in regression lines for each material that need to intersect each other in one point that defines the mass of the blank and its F14C.
Using a Bayesian model that considers the uncertainties of the F14C and mass determinations and is based on an inverse linear relationship between the measured F14C values, and the mass of the aliquots, Sun and co-workers (Reference Sun, Meyer, Dolman, Winterfeld, Hefter, Dummann, McIntyre, Montluçon, Haghipour, Wacker, Gentz, van der Voort, Eglinton and Mollenhauer2020) were able to determine the total contribution of the process blank and its uncertainty.
The objectives of our study are (1) whether the blank correction recommended by Sun et al. (Reference Sun, Meyer, Dolman, Winterfeld, Hefter, Dummann, McIntyre, Montluçon, Haghipour, Wacker, Gentz, van der Voort, Eglinton and Mollenhauer2020) can also be applied to n-alkanes, (2) to find the lower carbon mass limit for dependable CSRA, (3) application of blank correction for the dating of a sediment sample.
MATERIAL AND METHODS
Samples
Leaves of Festuca rubra agg (red fescue) and Fagus sylvatica (common beech) served as modern standard material. The plant material was collected in 2019 in Elfringhausen (Germany) (51°19’29.2’’N 7°10’11.9’’E). The material was homogenized by crushing in a mortar grinder before biomarker extraction.
Commercial petroleum-based standards (nC26, nC27 and nC28) (Sigma Aldrich, Steinheim, Germany) were used as fossil references. It turned out that the nC27 standard is not entirely of fossil origin but carries a 14C content of F14C = 0.6 (as determined by standard 14C analysis using graphite), so it was not included in the data set of fossil standards.
Three mg of sample material (equivalent to 1000 µg C) of those modern materials and fossil standards were dated by the standard radiocarbon procedure using graphite to determine their true F14C value.
In our investigations, we included a sample of a sediment layer of core HM1 from the Holzmaar lake archived in the Eifel Laminate Archive (ELSA) (Sirocko et al. Reference Sirocko, Dietrich, Veres, Grootes, Schaber-Mohr, Seelos, Nadeau, Kromer, Rothacker, Röhner, Krbetschek, Appleby, Hambach, Rolfi, Sudo and Grim2013). The Holzmaar is one of the best-studied maar lakes in the Eifel (Germany) (Kienel et al. Reference Kienel, Vos, Dulski, Lücke, Moschen, Nowaczyk and Schwab2013; Mollidor et al. Reference Mollidor, Tezkan, Bergers and Löhken2013; Musa Bandowe et al. Reference Musa Bandowe, Srinivasan, Seelge, Sirocko and Wilcke2014; Prasad and Baier Reference Prasad and Baier2014; Lehndorff et al. Reference Lehndorff, Wolf, Litt, Brauer and Amelung2015; Brunck et al. Reference Brunck, Sirocko and Albert2016; Sirocko et al. Reference Sirocko, Knapp, Dreher, Förster, Albert, Brunck, Veres, Dietrich, Zech, Hambach, Röhner, Rudert, Schwibus, Adams and Sigl2016; Birk et al. Reference Birk, Reetz, Sirocko, Wright and Fiedler2021). Since, CSRA samples require considerable amounts of carbon, enough core material is necessary (typically 15–25 g), so we decided to sample a 10 cm long section of the core (between 3.80 to 3.90 m). To compare our CSRA results with radiocarbon dates of bulk material from the core, we selected the depth in the core where radiocarbon dating above (3.30 m, 3342 BP) and below (4.20 m, 2580 BP) was carried out (Sirocko et al. Reference Sirocko, Martínez-García, Mudelsee, Albert, Britzius, Christl, Diehl, Diensberg, Friedrich, Fuhrmann, Muscheler, Hamann, Schneider, Schwibus and Haug2021). The sediment sample was dried at 40ºC, sieved < 2 mm and finely ground in an agate mill.
Laboratory Equipment
All laboratory materials (glass, metal, fiberglass tubes) for analytical procedure as well as quartz sand and boiling pearls, were heated in a muffle furnace at 300°C and at 450°C, respectively. Glass and metal ware were cleaned up with purifier Decon from Decon Laboratories (Sussex, United Kingdom) and Teepol from Bio-Connect (Huissen, Netherlands). Gas chromatography (GC)-ultra-grade solvents were from Carl Roth GmbH + Co. KG (Karlsruhe, Germany). The silica gel (60 Å, 0.063–0.200 mm, 70–230 mesh size; (Sigma-Aldrich Chemie GmbH, Taufkirchen, Germany) was used for solid phase extraction (SPE). Nitrogen (N2) (99.999 %, Messer AG, Münster, Germany) was used for solvent removal from samples.
Sample Extraction
All samples were treated with the same analytical procedure. Samples were directly weighed into glass-fiber tubes. The sample masses of modern material ranges between 2–14 g. Fossil standards were weighed-in at 2 mL each (1 mg mL–1). Soxhlet extraction was performed with n-hexane (200 mL) for 36 hours. Each time the extract was transferred, it was rinsed three times with n-hexane. Extracts were purified with solid phase extraction (SPE). For this purpose, the extract was reduced via a rotary evaporator (Rotavapor R-3, Büchi Labortechnik GmbH, Essen, Germany), transferred with n-hexane into a vial, reduced again under N2 and redissolved with 10 mL n-hexane. SPE columns (30 mL PP, Macherey–Nagel, Düren, Germany) were packed with 5 % deactivated silica gel and preconditioned with 30 mL n-hexane. The extract was transferred to the column and n-alkanes were eluted with 60 mL n-hexane. The solvent was removed via a rotary evaporator, n-alkanes were redissolved with n-hexane and transferred to a vial, the solvent was reduced again under N2 and a defined volume of n-hexane, depending on the desired mass carbon, was added. Here it has been shown that a volume between 2–5 mL is useful.
Instrumental Analysis
Identification and quantification of n-alkanes was performed by gas chromatography (GC) equipped with a flame ionization detector (GC-FID) (7890B G3440B), Agilent Technologies, Santa Clara, CA, USA). An HP5, 30 m fused silica capillary column, 0.250 mm i.d. and with a film thickness of 0.25 µm (Agilent Technologies, Santa Clara, CA, USA) was used. Helium (99.9999 %) was the carrier gas at a constant flow 30 mL min–1. The injection port was at 300°C (pressure front inlet 18 psi), and a sample volume of 1 µL was injected in splitless mode. The oven temperature program was 80ºC (1.5 min), ramp ran from 5ºC min–1 to 300ºC, hold time was 20 min. The detector temperature was 300ºC, and H2 flow was 30 mL min–1.
Individual n-alkane homologs (nC26, nC27, nC28, nC29, nC31, nC33) were isolated by preparative capillary gas chromatography (prepGC) (7890B G3440B), Agilent Technologies, Santa Clara, CA, USA). The prepGC was fitted with a cold injection system (CIS 4, Gerstel, Mühlheim a. d. Ruhr, Germany) and a preparative fraction collector (PFC, Gerstel, Mühlheim a. d. Ruhr Germany). A megabore column (HP5 30 m × 0.530 mm × 2.65 µm) (Agilent Technologies, Santa Clara, CA, USA) was used with a constant flow (5.5 mL min–1) and helium as carrier gas (99.9999%). The front inlet temperature program was 100ºC (start CIS), CIS ramp 12ºC s–1 to 350ºC, and hold time 2.11 min. The injection speed was 40,000 µL min–1, injection volume 14 µL, and transfer time (hold time oven) 2 min. The oven temperature was 100ºC, with a hold time of 2 min and a ramp of 10ºC min–1 to 300ºC, hold time of 33 min.
To reach the highest concentration of n-alkanes close to stop flow parameters were used with a vent flow of 90 mL min–1, vent time of 0.11 min and vent pressure of 1.10 psi. The transfer line was heated up to 310ºC and the switch to 320ºC. The switch was equipped with a converted trap heater. In this way, it was possible to use 100 µL traps. The trap heater ran constant at 250ºC during harvesting. Subsequently, analytes were washed out with 10 mL n-hexane, collected in glass tubes, and completely dried under a gentle flow of N2. Tubes were closed by a cap of aluminum foil and additionally sealed with parafilm to avoid contamination with dust or other larger particles.
Purity and amount of analyte were determined by GC-FID. The mean average recovery of individual n-alkanes after prepGC was 62 % (Table S1).
14C Analysis
The sealed tubes were transported to the radiocarbon laboratory in Mannheim (Kromer et al. Reference Kromer, Lindauer, Synal and Wacker2013). Within days after arrival, the tubes were prepared for the classic method of closed/sealed-tube combustion to convert the samples to carbon dioxide (CO2). The sample treatment and CSRA are described in detail in (Hoffmann et al. Reference Hoffmann, Friedrich, Kromer and Fahrni2017). The tubes containing the samples were attached to a custom-built vacuum system, the air was removed by vacuum pumping, and pure oxygen gas was added (instead of CuO to avoid the introduction of contamination) to the tubes before flame sealing the glass tubes. The tubes were placed into a muffle furnace and the sample material combusted at 900ºC overnight. Since the combustion of the n-alkanes produces not only CO2 but also water, the gaseous samples were cleaned by separating the carbon from other gases and quantifying the gas quantity within the aforementioned vacuum extraction system. A sequence of cold traps operated with acetone dry ice (for trapping the water vapor) and liquid nitrogen (for trapping the CO2) is used to separate the CO2 from all other gases and transfer the clean gas into small glass ampoules. The flame-sealed ampoules were inserted in and measured by the commercially available gas interface system (GIS, IonPlus, Dietikon, Switzerland) of the AMS of the type MICADAS (IonPlus, Dietikon, Switzerland). Determination of the process blank of the combustion and gas-cleaning procedure using empty glass tubes processed identical to samples on the vacuum extraction line showed no measurable carbon contribution. The GIS and MICADAS system was standardized by gas aliquots of pre-combusted oxalic acid II standard material (NIST SRM 4990C) and fossil CO2. Data evaluation was performed with the software BATS (Wacker et al. Reference Wacker, Bonani, Friedrich, Hajdas, Kromer, Nemee, Ruff, Suter, Synal and Vockenhuber2016).
Overall, 24 modern standards (modern plant material) with mass ranges 5–72 µg C and 24 fossil standards (nC26, nC28) with mass ranges 8–51 µg C were analyzed together with sample material. The sample masses were those determined during gas analysis by the GIS.
Blank Calculation
The calculation assumes that the blank is constant for a substance during the same analytical procedure. A constant blank is also hypothesized in prior studies before (Hwang and Druffel Reference Hwang and Druffel2005; Santos et al. Reference Santos, Southon, Griffin, Beaupre and Druffel2007). The detailed calculation of the F14Cblank and mblank using a Baysian model is described by Sun et al. (Reference Sun, Meyer, Dolman, Winterfeld, Hefter, Dummann, McIntyre, Montluçon, Haghipour, Wacker, Gentz, van der Voort, Eglinton and Mollenhauer2020).
Different-sized aliquots of modern (F14C ≅ 1) and fossil (F14C ≅ 0) standards are plotted against their inverse masses and regression lines are calculated. F14Cblank and mblank are derived from the intersection points of the modern and fossil regression lines.
The Bayesian model estimates the numerical bivariate distribution of the intersection of modern and fossil regression lines. Uncertainties of the linear fit and measurement uncertainties were considered.
The Bayesian model run with our data set of 24 modern and 24 fossil standards, in 3 Markov chains; the fitting process run with 3,500 iterations taking the uncertainties into account. Those calculated values of F14Cblank and mblank were used in the blank correction of the sample in order to determine the uncontaminated F14C_sample_blank_corrected of the sample as follows:
$${F^{14}}{C_{sample\,blank\ corrected}} = {{(({F^{14}}{C_{sample\ measured\,}} * {m_{sample}}) - ({F^{14}}{C_{blank}} * {m_{sample\,}}))} \over {{m_{sample\,}} - {m_{blank}}}}$$
RESULTS AND DISCUSSION
Blank Assessment
The 14C measured on the large-sized graphite targets resulted in F14C = 0.0027 ± 0.0005 (average ± standard deviation of 2 measurements) for the fossil standards (nC26, nC28), 0.71 ± 0.00003 (average and standard deviation of 2 measurements) for the nC27-standard and 1.00 ± 0.01 (average and standard deviation) of the F14C of the red fescue sample with F14C = 0.995 ± 0.003 and the common beech sample with F14C = 1.009 ± 0.002 for the modern plant material. Due to the enormous size of those samples, the material is virtually unaffected by a blank and we assume those numbers to be the “true” F14Ctrue values.
The nC27 standard is neither fossil nor modern, so we continuously process this standard with every CSRA batch (Table 1).
Table 1 Uncalibrated bulk and compound-specific radiocarbon dates of the standard nC27—the measured F14C, corrected F14C, and corrected 14C ages.
1 Measured on graphite targets.
2 Measured as gas sample.
Figure 1 shows the result of the blank assessment for n-alkanes resulting in a calculated blank mass of mblank = 4.1 ± 0.7 µg and 14C value of a F14Cblank = 0.25 ± 0.05. The mass of the blank, which is determined by the analytical equipment and sample handling procedure, is similar to the blank of n-alkanoic acid (mblank n-alkanoic acid 4.898 µg C) in the previous blank assessment by Sun and co-workers (Sun et al. Reference Sun, Meyer, Dolman, Winterfeld, Hefter, Dummann, McIntyre, Montluçon, Haghipour, Wacker, Gentz, van der Voort, Eglinton and Mollenhauer2020). In addition, our blank is close to the blank of n-alkanes from marine sediments (Druffel et al. Reference Druffel, Zhang, Xu, Ziolkowski, Southon, dos Santos and Trumbore2010). The analytical procedures for CSRA of n-alkanoic acids, lignin phenols, and n-alkanes are fundamentally different, so the blank values may differ. However the blank must be calculated for every substance and will change with another analytical protocol (Sun et al. Reference Sun, Meyer, Dolman, Winterfeld, Hefter, Dummann, McIntyre, Montluçon, Haghipour, Wacker, Gentz, van der Voort, Eglinton and Mollenhauer2020).
Figure 1 Blank assessment for n-alkanes: (a) Bayesian model ran with 3500 iterations, the visual check depicts 500 regression lines of modern and fossil standard, (b) the posterior distribution of F14C values and masses of the blank.
Lower Mass Limit for Dependable CSRA
The standard nC27 has been repeatedly measured both as a graphite sample (where the blank contribution is negligible due to the large sample mass) and as a small-weight gas sample. Contrary to the information received from the manufacturer of the nC27 standard, our dating results show a non-fossil age of F14C = 0.71 or a 14C age of approximately 2,706 ± 86 BP (Table 1). This standard with known F14C can now be used as a quality control standard that, when applying the blank correction, should lead to the correct value of F14C = 0.71. This standard offers the possibility to make a self-assessment of the blank assessment and to define a minimum threshold value for the mass of carbon at which a corresponding blank value correction is no longer useful because the proportion of the blank in the sample is too large. If we perform the blank assessment on the nC27 standard measured as a gas sample, considering the limit of 15 µg C, we get the radiocarbon age of about approximately F14C = 0.63 (2141 ± 844 BP) (Table 1). With increasing mass, the blank corrected ages deviate slightly from each other (Table 1). Sun and co-workers found that if the proportion of the blank is 30 % of the sample, the blank correction is not reliable (Sun et al. Reference Sun, Meyer, Dolman, Winterfeld, Hefter, Dummann, McIntyre, Montluçon, Haghipour, Wacker, Gentz, van der Voort, Eglinton and Mollenhauer2020). Accordingly, with our blank (mblank of 4.107 µg), at least 18 µg C would be necessary for reliable results. The standard nC27 (18 µg C) had a blank corrected F14C of 0.63 (14C age of 2383 ± 782 BP). This differs slightly from the known bulk radiocarbon value F14C = 0.71 (2706 ± 86 BP). Previous studies have already determined that 25 µg C is necessary for a dependable blank assessment (Pearson et al. Reference Pearson, McNichol, Schneider, von Reden and Zheng1997; von Reden et al. Reference von Reden, McNichol, Pearson and Schneider1998; Shah and Pearson Reference Shah and Pearson2007). The blank correction and the additional uncertainty resulting from the correction is strongly dependent on the mass of carbon. A lower relative blank contribution leads to more precise blank-corrected ages (Table 1). Even if it is technically possible to measure extremely low masses (< 10 µg C) on the AMS, the appropriate lower limit of at least 15 µg C should be considered.
Application of Blank Assessment to Sediment Samples
Individual n-alkane homologs nC27, nC29, nC31 and nC33 were isolated from sediment sample of the Holzmaar using prepGC. The peaks were baseline-separated. The separation of the different carbon chains from the lipid mix was checked with GC-FID. Traps 2 to 5 show the cleanly isolated analytes (Figure 2). Their F14C and mass was 0.465 ± 0.007 and 20.9 µg (nC27), 0.469 ± 0.006 and 31.5 µg (nC29), 0.479 ± 0.007 and 37.9 µg (nC31), 0.421 ± 0.011 and 10 µg (nC33). After the blank correction using the blank values above, ages of the four n-alkanes examined differed. Following uncalibrated ages resulted for the n-alkanes, homologs were 5286 ± 594 BP (nC27, F14C = 0.465 ± 0.007); 5548 ± 444 BP (nC29, F14C = 0.469 ± 0.006); 5451 ± 854 BP (nC31, F14C = 0.479 ± 0.007); 4960 ± 537 BP (nC33, F14C = 0.421 ± 0.011) (Table 2). One of the four dated n-alkanes had a mass < 15 µg C (nC33), which was determined to be the limit of excessive uncertainty in CSRA. Therefore, the radiocarbon age of the n-alkane nC33 is not dependable. In contrast, n-alkanes nC27, nC29 and nC31 with a mass ≥ 20.9 µg C are very close in age with a mean value of, 429 ± 600 BP (F14C = 0.47 ± 0.007). However, these CSRA are all older than the corresponding bulk and twigs samples determined by Sirocko and co-workers (Sirocko et al. Reference Sirocko, Knapp, Dreher, Förster, Albert, Brunck, Veres, Dietrich, Zech, Hambach, Röhner, Rudert, Schwibus, Adams and Sigl2016) (Figure 3).
Figure 2 The top trace shows the initial sample HM1_21 before prepGC. Trapped n-alkanes are isolated by prepGC (no. of runs = 40).
Table 2 Uncalibrated compound-specific radiocarbon dates of the sediment sample HM1_21—the measured F14C, corrected F14C, and corrected 14C ages.
Figure 3 14C-data from core HM1 Lake Holzmaar. Bulk sediment and twig uncalibrated radiocarbon data are already published in Sirocko et al. Reference Sirocko, Martínez-García, Mudelsee, Albert, Britzius, Christl, Diehl, Diensberg, Friedrich, Fuhrmann, Muscheler, Hamann, Schneider, Schwibus and Haug2021. All radiocarbon dating, including CSRA measurements, were performed at the Curt–Engelhorn Center for Archaeometry in Mannheim, Germany. N-Alkanes of the sample HM1_21 have the following uncalibrated CSRA dates: nC27 (5286 ± 594 BP), nC29 (5548 ± 444 BP), nC31 (5451 ± 854 BP) and nC33 (4960 ± 537 BP). Due to the time axis, the error bar for all the data shown here are so small that they are not visible in the figure.
An age offset is common and has been frequently described in other studies (Kusch et al. Reference Kusch, Rethemeyer, Schefuß and Mollenhauer2010; Feng et al. Reference Feng, Benitez-Nelson, Montluçon, Prahl, McNichol, Xu, Repeta and Eglinton2013; Douglas et al. Reference Douglas, Pagani, Eglinton, Brenner, Hodell, Curtis, Ma and Breckenridge2014; Gierga et al. Reference Gierga, Hajdas, van Raden, Gilli, Wacker, Sturm, Bernasconi and Smittenberg2016; Nelson et al. Reference Nelson, Nemiah Ladd, Schubert and Kahmen2018; Makou et al. Reference Makou, Eglinton, McIntyre, Montluçon, Antheaume and Grossi2018). Reasons for this may include: (1) Different transport processes of n-alkanes require contrasting times to be deposited in the sediment trap. Aeolian transport of n-alkanes occurs on a regional scale and can lead to an age offset of one year (Freimuth et al. Reference Freimuth, Diefendorf, Lowell, Schartman, Landis, Stewart and Bates2021). During fluviatile transport (Kusch et al. Reference Kusch, Rethemeyer, Schefuß and Mollenhauer2010; Galy et al. Reference Galy, Eglinton, France-Lanord and Sylva2011; Ponton et al. Reference Ponton, West, Feakins and Galy2014; Häggi et al. Reference Häggi, Sawakuchi, Chiessi, Mulitza, Mollenhauer, Sawakuchi, Baker, Zabel and Schefuß2016; Hemingway et al. Reference Hemingway, Schefuß, Dinga, Pryer and Galy2016; Simoneit et al. Reference Simoneit, Pisani, Ekpo, Fubara, Nna and Ekpa2017), n-alkanes from sediments of different riparian areas can be mixed and enter the sediment trap (Aichner et al. Reference Aichner, Gierga, Stolz, Mętrak, Wilk, Suska-Malawska, Mischke, Sachse, Rajabov, Rajabov and Rethemeyer2021). The river Sammetbach flows into the Holzmaar. Therefore, a mixture of terrestrial n-alkanes, eroded from several riparians, may enter the sediment of the Holzmaar. The n-alkanes can be incorporated into the soil in the lake’s catchment area via litter (Freimuth et al. Reference Freimuth, Diefendorf, Lowell, Schartman, Landis, Stewart and Bates2021). The age mixture of sources of the carbon pool in soils can range between hundreds to thousands of years before soil erosion transports the carbon including n-alkanes into the lake sediment (Diefendorf et al. Reference Diefendorf and Freimuth2017). Eroded soil with age-mixed n-alkanes may also result in an age difference in CSRA samples. This could be also a reason for the reverse chronology of bulk samples between 3.30 m (3342 BP) and 4.20 m (2580 BP) in HM1.
Additionally, (2) the hard-water effect can influence the radiocarbon measurements. Dissolved inorganic carbon (DIC) reaches the maar lake via groundwater (Sirocko et al. Reference Sirocko, Knapp, Dreher, Förster, Albert, Brunck, Veres, Dietrich, Zech, Hambach, Röhner, Rudert, Schwibus, Adams and Sigl2016). Additionally, rising mantle CO2 cannot be excluded (Dahm et al. Reference Dahm, Stiller, Mechie, Heimann, Hensch, Woith, Schmidt, Gabriel and Weber2020). The impact of hard water is not limited to bulk measurements, but also affects CSRA. Mischke and co-workers (Reference Mischke, Lai, Aichner, Heinecke, Mahmoudov, Kuessner and Herzschuh2017) determined the hard water effect by dating algae 14C. Modern algae had an F14C of 0.8914, thus a correction of the hard-water effect was necessary, which can cause an age shift of approximately 2000 years in submerged aquatic plants (Gross Reference Gross1957). The impact of hard water on the radiocarbon dating of bulk sediment is possible, as the heterogeneous bulk sample may contain a proportion of algae. However, it is unclear if this effect influences the terrestrial CSRA of n-alkanes.
(3) Inhomogeneous distribution of n-alkanes. The n-alkanes are never homogeneously distributed in a lake sediment, as evidenced by various samples within a lake (Sarkar et al. Reference Sarkar, Wilkes, Prasad, Brauer, Riedel, Stebich, Basavaiah and Sachse2014; Kou et al. Reference Kou, Lin, Wang, Yu, Kai, Laug and Zhu2021).
Intrinsic Blank Assessment
An alternative approach of blank assessment without using fossil and modern standards at all is by dating a single sample multiple times (or using multiple n-alkanes as subsamples) with different masses. The F14C values of the individual subsamples will obey the aforementioned mass-balance equation. Similar to the data in Figure 1, the individual F14C values will fall onto one line and define a linear equation with a given slope and intercept. While the mass and F14C value of the blank cannot be independently determined from the slope of the linear equation, its intercept equals the true F14C of the whole sample.
Figure 4 shows the results of Holzmaar samples. The F14C values of the individual n-alkanes lie on a straight line determining the intercept of the linear equation. Using this method, the true F14C of the Holzmaar sample from which the n-alkanes were extracted is 0.5 ± 0.01 resulting in a 14C age of (4464 ± 738 BP) which agrees well with the individual resulting blank corrected dates.
Figure 4 Intrinsic blank assessment for n-alkanes: The different masses of the n-alkanes are distributed in such a way that it is possible to use linear regression to determine an intersection with the y-axis. The intersection is the F14C value of the intrinsic blank. The intersection with modern and fossil samples is not included in the mblank assessment.
However, this approach also presents challenges, such as the amount of sample material available for subsampling and how many samples need to be analyzed. For example, when only a limited amount of sample material is available or a large number of samples need to be analyzed, it can be more efficient to use only a few fossil and modern standards to analyze many samples. In this way, it would require less effort and time to carry out the analysis and it would make sense to use the blank assessment according to (Sun et al. Reference Sun, Meyer, Dolman, Winterfeld, Hefter, Dummann, McIntyre, Montluçon, Haghipour, Wacker, Gentz, van der Voort, Eglinton and Mollenhauer2020). However, when very few samples are available with sufficient mass of carbon to extract more than one n-alkane, as in the Holzmaar sample, an intrinsic blank assessment can be performed.
Despite these challenges, this method offers advantages. For example, it allows more flexibility in the analytic process and could potentially be adapted to diverse types of samples. In addition, the sample itself serves as an intrinsic blank, which can increase the robustness of the method. In addition, this method could allow for different sample preparations without the need to use fossil and modern standards for each.
In summary, while blank assessment without fossil and modern standards by measuring a sample of different masses may have limitations, it has potential advantages that could make it a useful analytical tool.
CONCLUSION
Using fossil and modern standard material the contribution of the procedural blank to extracted n-alkanes could be determined in terms of blank mass and blank F14C. Thus, biomarker extraction and CRSA is successfully established at the participating laboratories. We conclude that we have a reliable blank assessment for samples with more than 15 µg C, supplemented by the internal control by the commercial standard nC27, but have an age offset in the Holzmaar sediment sample.
Three out of four isolated n-alkanes from a sediment sample from the Holzmaar have sufficient carbon for a reliable blank assessment. The CSRA ages of these three n-alkanes are close, but round about 1000–2000 years older than radiocarbon-dated twigs and bulk samples. We did not consider the age offset being caused by the hard-water effect. Instead, we assume that the input of organic matter with a long residence time in the soil determines the CSRA age of the terrestrial n-alkanes in the sediment as the decisive factor for the age offset. In addition, the Sammetbach transports a mix of terrestrial n-alkanes into the Holzmaar.
The intrinsic blank assessment using a sample of different masses offers advantages, such as increased flexibility in the analytical procedure and potentially being adaptable to different sample types, although it has limitations when limited sample material is available. Overall, the methods described in this text are useful for understanding the age and sources of organic matter in sediment samples.
ACKNOWLEDGMENTS
We thank Stefanie Klassen for laboratory assistance. Prof. Dr. Frank Sirocko, Johannes Gutenberg -University, is thanked for providing samples of the ELSA project. This work was supported by the Klaus Tschira Foundation Heidelberg (grand # 00.332.2027).
SUPPLEMENTARY MATERIAL
To view supplementary material for this article, please visit https://doi.org/10.1017/RDC.2024.26
Open access
