Hostname: page-component-586b7cd67f-t8hqh Total loading time: 0 Render date: 2024-11-29T01:01:39.940Z Has data issue: false hasContentIssue false

Comparing MICADAS Gas Source, Direct Carbonate, and Standard Graphite 14C Determinations of Biogenic Carbonate

Published online by Cambridge University Press:  29 April 2024

Jordon Bright*
Affiliation:
School of Earth and Sustainability, Northern Arizona University, Flagstaff, AZ, 86011, USA
Chris Ebert
Affiliation:
Center for Ecosystem Sciences and Society, and Department of Biological Sciences, Northern Arizona University, Flagstaff, AZ, 86011, USA
Carola Flores
Affiliation:
Department of History and Social Science, Faculty of Liberal Arts, Universidad Adolfo Ibañez, Viña del mar, Chile School of Archaeology, University Austral de Chile. Liborio Guerrero 1765. Puerto Montt, Chile
Paul G Harnik
Affiliation:
Department of Earth and Environmental Geosciences, Colgate University, Hamilton, NY, 13346, USA
John Warren Huntley
Affiliation:
Geological Sciences, University of Missouri, Columbia, MO, 65211, USA
Michał Kowalewski
Affiliation:
Florida Museum of Natural History, University of Florida, Gainesville, FL, 32611, USA
Roger W Portell
Affiliation:
Florida Museum of Natural History, University of Florida, Gainesville, FL, 32611, USA
Michael Retelle
Affiliation:
Department of Earth and Climate Science, Bates College, Lewiston, ME, 04240, USA Arctic Geology Department, University Centre in Svalbard, Longyearbyen, 9171 Svalbard, Norway
Edward A G Schuur
Affiliation:
Center for Ecosystem Sciences and Society, and Department of Biological Sciences, Northern Arizona University, Flagstaff, AZ, 86011, USA
Darrell S Kaufman
Affiliation:
School of Earth and Sustainability, Northern Arizona University, Flagstaff, AZ, 86011, USA
*
*Corresponding author: Jordon Bright; Email: [email protected]
Rights & Permissions [Opens in a new window]

Abstract

Northern Arizona University, Flagstaff, Arizona, USA, recently installed a MIni CArbon DAting System (MICADAS) with a gas interface system (GIS) for determining the 14C content of CO2 gas released by the acid dissolution of biogenic carbonates. We compare 48 paired graphite, GIS, and direct carbonate 14C determinations of individual mollusk shells and echinoid tests. GIS sample sizes ranged between 0.5 and 1.5 mg and span 0.1 to 45.1 ka BP (n = 42). A reduced major axis regression shows a strong relationship between GIS and graphite percent Modern Carbon (pMC) values (m = 1.011; 95% CI [0.997–1.023], R2 = 0.999) that is superior to the relationship between the direct carbonate and graphite values (m = 0.978; 95% CI [0.959-0.999], R2 = 0.997). Sixty percent of GIS pMC values are within ±0.5 pMC of their graphite counterparts, compared to 26% of direct carbonate pMC values. The precision of GIS analyses is approximately ±70 14C yrs to 6.5 ka BP and decreases to approximately ±130 14C yrs at 12.5 ka BP. This precision is on par with direct carbonate and is approximately five times larger than for graphite. Six Plio-Pleistocene mollusk and echinoid samples yield finite ages when analyzed as direct carbonate but yield non-finite ages when analyzed as graphite or as GIS. Our results show that GIS 14C dating of biogenic carbonates is preferable to direct carbonate 14C dating and is an efficient alternative to standard graphite 14C dating when the precision of graphite 14C dating is not required.

Type
Research Article
Copyright
© The Author(s), 2024. Published by Cambridge University Press on behalf of University of Arizona

Introduction

The ability to rapidly measure the 14C content of sub-milligram quantities of biogenic carbonate without conversion to graphite has risen in popularity over the past decade (e.g., Lougheed et al. Reference Lougheed, Snowball, Moros, Kabel, Muscheler, Virtasalo and Wacker2012, Reference Lougheed, Metcalfe, Ninnemann and Wacker2018; Bush et al. Reference Bush, Santos, Xu, Southon, Thiagarajan, Hines and Adkins2013; Longworth et al. Reference Longworth, Robinson, Roberts, Beaupre, Burke and Jenkins2013; Wacker et al. Reference Wacker, Fahrni, Hajdas, Molnar, Synal, Szidat and Zhang2013a, Reference Wacker, Lippold, Molnár and Schulz2013b; Dominguez et al. Reference Dominguez, Kosnik, Allen, Hua, Jacob, Kaufman and Whitacre2016; Kosnik et al. Reference Kosnik, Hua, Kaufman, Kowalewski and Whitacre2017; Ritter et al. Reference Ritter, Erthal, Kosnik, Ciombra and Kaufman2017, Reference Ritter, Erthal, Kosnik, Kowalewski, Coimbra, Caron and Kaufman2023; Gottschalk et al. Reference Gottschalk, Szidat, Michel, Mazaud, Salazar, Battaglia, Lippold and Jaccard2018; Kowalewski et al. Reference Kowalewski, Casebolt, Hua, Whitacre, Kaufman and Kosnik2018; Tuna et al. Reference Tuna, Fagault, Bonvalot, Capano and Bard2018; Fagault et al. Reference Fagault, Tuna, Rostek and Bard2019; Lindauer et al. Reference Lindauer, Friedrich, van Gyseghem, Schöne and Hinderer2019; New et al. Reference New, Yanes, Cameron, Miller, Teixeira and Kaufman2019; Parker et al. Reference Parker, Yanes, Hernández, Hernández Marreno, Paris and Surge2019; Albano et al. Reference Albano, Hua, Kaufman, Tomašových, Zuschin and Agadi2020, Reference Albano, Sabbatini, Lattanzio, Päßler, Steger, Hua, Kaufman, Szidat, Zuschin and Negri2023a, Reference Albano, Hua, Kaufman and Zuschin2023b; Missiaen et al. Reference Missiaen, Wacker, Lougheed, Skinner, Hajdas, Nouet, Pichat and Waelbroeck2020; Dolman et al. Reference Dolman, Groeneveld, Mollenhauer, Ho and Laepple2021; Mollenhauer et al. Reference Mollenhauer, Grotheer, Gentz, Bonk and Hefter2021; Nawrot et al. Reference Nawrot, Berensmeier, Gallmetzer, Haselmair, Tomašových and Zuschin2022; Sanchez et al. Reference Sanchez, Yanes, Linstädter and Hutterer2022; Steger et al. Reference Steger, Bosnjak, Belmaker, Galil, Zuschin and Albano2022). Dating small carbonate samples has become increasingly popular in time averaging or population structure studies (e.g., Kowalewski et al. Reference Kowalewski, Casebolt, Hua, Whitacre, Kaufman and Kosnik2018; Nawrot et al. Reference Nawrot, Berensmeier, Gallmetzer, Haselmair, Tomašových and Zuschin2022; Ritter et al. Reference Ritter, Erthal, Kosnik, Kowalewski, Coimbra, Caron and Kaufman2023) and in the field of marine geochronology where a few dozen or even single large foraminifera can be dated (e.g., Lougheed et al. Reference Lougheed, Metcalfe, Ninnemann and Wacker2018; Mollenhauer et al. Reference Mollenhauer, Grotheer, Gentz, Bonk and Hefter2021). Two rapid 14C methods are currently employed: direct carbonate and gas ion source (GIS). The direct carbonate method produces carbon ions (C-) by directly sputtering powdered carbonate and metal binder (typically iron or niobium) mixtures with cesium atoms (Bush et al. Reference Bush, Santos, Xu, Southon, Thiagarajan, Hines and Adkins2013; Longworth et al. Reference Longworth, Robinson, Roberts, Beaupre, Burke and Jenkins2013; Hua et al. Reference Hua, Levchenko and Kosnik2019). Alternatively, C- ions are produced by the GIS method through the acid dissolution of carbonate and then sputtering the released CO2 gas with cesium atoms in the presence of titanium (Middleton Reference Middleton1984; Bronk and Hedges 1987). The GIS method has become increasingly popular with the development of the MIni CArbon DAting System (MICADAS) (Synal et al. Reference Synal, Stocker and Suter2007) coupled with a carbonate handling system (CHS) inlet to the GIS (Ruff et al. Reference Ruff, Wacker, Gäggeler, Suter, Synal and Szidat2007, Reference Ruff, Fahrni, Gäggeler, Hajdas, Suter, Synal, Szidat and Wacker2010; Wacker et al. Reference Wacker, Fahrni, Hajdas, Molnar, Synal, Szidat and Zhang2013a; Lindauer et al. Reference Lindauer, Friedrich, van Gyseghem, Schöne and Hinderer2019).

The Arizona Climate and Ecosystems (ACE) Isotope Laboratory at Northern Arizona University (NAU) brought a MICADAS system online in June 2021 (Ebert et al. Reference Ebert, Schuur, Kaufman, Brown, Propster, Kelley, Bright, Carbone, McKay and Koch2022). The ACE Lab includes elemental analyzers, a GIS, a CHS, and both automated graphitization equipment (AGE) and a manual cryogenic purification line.

This report compares the percent Modern Carbon (pMC) values and analytical precision from 48 samples of biogenic carbonate each analyzed as graphite, GIS, and as direct carbonate. The primary goal of our study is to compare the accuracy and reproducibility of carbonate pMC values from NAU’s MICADAS and GIS system against standard graphite and direct carbonate analyses.

Materials and Methods

This study features predominantly mollusk shells and fewer echinoid tests. Several of the standard graphite and direct carbonate analyses were included in an earlier study published prior to the arrival of the MICADAS at NAU (Bright et al. Reference Bright, Ebert, Kosnik, Southon, Whitacre, Albano, Flores, Frazer, Hua, Kowalewski, Martinelli, Oakley, Parker, Retelle, Ritter, Rivadeneira, Scarponi, Yanes, Zuschin and Kaufman2021). The compilation featured here comprises pMC values ranging from approximately 99 to 0.4 (radiocarbon ages approximately 0.1 to 45.1 ka BP), based on prior AMS analysis of samples archived at NAU (Bright et al. Reference Bright, Ebert, Kosnik, Southon, Whitacre, Albano, Flores, Frazer, Hua, Kowalewski, Martinelli, Oakley, Parker, Retelle, Ritter, Rivadeneira, Scarponi, Yanes, Zuschin and Kaufman2021). To fill in gaps in this dataset, new graphite and direct carbonate powders were processed and analyzed at NAU or at the W.M. Keck Carbon Cycle Accelerator Mass Spectrometer Facility at the University of California – Irvine (UCI), respectively. To obtain equivalent subsamples for the three analyses, each mollusk shell and echinoid test was sampled parallel to growth bands and/or sub-adjacent to each other. This approach minimizes the risk of sampling different aged shell material. For example, some mollusks are slow growing and can live for several hundred years (Moss et al. Reference Moss, Ivany, Judd, Cummings, Bearden, Kim, Artruc and Driscol2016).

All samples were sonicated briefly to remove loose debris before being exposed to ACS grade 2N hydrochloric acid which removed 30% of their mass. The leaching process was considered complete when effervescence stopped. All samples were rinsed three times with deionized reverse osmosis water (16.7 Mohm*cm) before being dried at 50°C in a convection oven.

Carbonate samples (0.5–1.5 mg; 60–180 μg C) destined for GIS 14C analysis were placed in baked (3 hr at 500° C) glass reaction vials sealed with screw-top caps containing a rubber septum (Exetainer #VW101). The vials were flushed with N2 gas and sealed until analyzed. The CHS automatically flushes each vial with helium using a double-walled needle, after which 85% phosphoric acid is injected to dissolve the shell material. The evolved CO2 is carried by helium through a water trap and then to the zeolite trap of the GIS. The CO2 is released from the zeolite at 450°C and transferred to a syringe where it is mixed with helium to a final CO2 concentration of 5% by volume, after which it is injected into the ion source. The zeolite trap is optimized for 80–100 μg C (Ruff et al. Reference Ruff, Fahrni, Gäggeler, Hajdas, Suter, Synal, Szidat and Wacker2010), which is equivalent to about 0.6–0.8 mg of calcium carbonate. Excess carbon is not analyzed. Detailed description of the MICADAS GIS and CHS analytical method is provided in Synal et al. (Reference Synal, Stocker and Suter2007) and Ruff et al. (Reference Ruff, Wacker, Gäggeler, Suter, Synal and Szidat2007, Reference Ruff, Fahrni, Gäggeler, Hajdas, Suter, Synal, Szidat and Wacker2010).

Ten new samples (0.3–0.5 mg; 36–60 μg C) for direct carbonate AMS analyses at UCI were manually ground to a fine powder with an agate mortar and pestle. The powdered shell was mixed with 6 to 7 mg of niobium powder (Alfa Aesar Puratronic, –325 mesh, 99.99%) in baked (3 hr at 500°C) Kimble borosilicate glass culture tubes (6 mm OD × 50 mm), flushed with nitrogen gas, and capped with Supelco plastic column caps (1/4” OD) until the carbonate-niobium mixture was manually pressed into pre-drilled (4.1 mm depth) aluminum targets before being sent to UCI.

Thirteen new samples for graphite analysis (8-10 mg; 960–1200 μg C) at NAU were placed in baked (3 hr at 500°C) acid-washed glass vials sealed with rubber septa from BD Vacutainer® plastic collection tubes (No. 366704). Ambient atmosphere was removed via vacuum before a small-bore needle added approximately 0.8 mL of ACS grade 85% phosphoric acid to each vial. Vials were warmed in a heating block set to 70°C until the shell material entirely dissolved. The evolved gas was removed via vacuum on a manual cryogenic purification line. Water vapor was removed by passing the gas through a mixture of ethanol and liquid nitrogen at approximately –80°C. Carbon dioxide was condensed to a solid using a liquid nitrogen bath and the remaining gases were drawn off. The purified CO2 was converted to graphite by reaction with iron powder (Alfa Aesar, -325 mesh, reduced 98%) in a hydrogen reducing environment at 550°C for 3 hr (Vogel et al. Reference Vogel, Southon, Nelson and Brown1984). The graphite-iron mixture was packed into targets using an automated press before being loaded onto the MICADAS.

Radiocarbon (14C) concentrations are reported as pMC following the conventions of Stuvier and Polach (Reference Stuvier and Polach1977). Sample preparation backgrounds have been subtracted based on measurements of 14C-free calcite processed in the same fashion as the unknowns. All 14C determinations have been corrected for isotopic fractionation according to the conventions of Stuvier and Polach (Reference Stuvier and Polach1977) with δ13C values being measured on the AMS. These δ13C values are superior when used to correct for fractionation but can differ from the actual value of the original material and are thus not reported. Machine performance was monitored by repeat analysis of the IAEA standards C1 (14C-dead marble; consensus pMC = 0.00 ± 0.02; Rozanski et al. Reference Rozanski, Stichler, Gofiantini, Scott, Buekens, Kromer and van der Plicht1992) and C2 (travertine; consensus pMC = 41.14 ± 0.03; Rozanski et al. Reference Rozanski, Stichler, Gofiantini, Scott, Buekens, Kromer and van der Plicht1992), the CAHI coral standard provided by UCI (pMC = 94.44 ± 0.19; Mollenhauer et al. Reference Mollenhauer, Grotheer, Gentz, Bonk and Hefter2021), and a Pliocene 14C-dead Tridacna shell provided by the Florida Museum of Natural History (specimen UF 143174) which accounts for possible matrix effects that the marble C1 blank might not. All GIS 14C determinations of unknowns have been calibrated by assigning a blank pMC value of 0 and forcing the IAEA C2 travertine standard and the CAHI coral standard runs in the same batch through their consensus values (e.g., Gottschalk et al. Reference Gottschalk, Szidat, Michel, Mazaud, Salazar, Battaglia, Lippold and Jaccard2018). The blank and standard compilations were analyzed using boxplot and whisker diagrams to identify outliers, which were removed, without iteration, from the mean and standard deviation calculations. The number of outliers is included in Table 1. Data from the MICADAS was processed using BATS software (version 4.0; Wacker et al. Reference Wacker, Bonani, Friedrich, Hajdas, Kromer, Nemec, Ruff, Suter, Synal and Vockenhuber2010). When reported, radiocarbon ages are uncalibrated years BP, without consideration of the marine reservoir effect, with BP meaning conventional radiocarbon years before AD 1950 (Stuvier and Polach Reference Stuvier and Polach1977).

Table 1 Mean pMC values for reference material (IAEA) and laboratory standards analyzed at Northern Arizona University ACE lab. IAEA reference values are from Rozanski et al. (Reference Rozanski, Stichler, Gofiantini, Scott, Buekens, Kromer and van der Plicht1992)

N.A. – not applicable.

* Coral standard provided by UC-Irvine. Graphite reference value as cited in Mollenhauer et al. (Reference Mollenhauer, Grotheer, Gentz, Bonk and Hefter2021) from 294 analyses performed at the Keck Carbon Cycle Radiocarbon Laboratory at UC-Irvine. GIS “reference” value is based on 47 analyses reported in Table 3 of Mollenhauer et al. (Reference Mollenhauer, Grotheer, Gentz, Bonk and Hefter2021).

** Calibrated values.

The relationship between GIS and graphite or direct carbonate and graphite pMC values was evaluated using a reduced major axis regression (RMA) analysis, which minimizes the residual variation across both the X- and Y-axes (Quinn and Keough Reference Quinn and Keough2002; Smith Reference Smith2009). An RMA regression avoids assumptions about the dependent and independent variables between GIS and graphite or direct carbonate and graphite pMC values (Smith Reference Smith2009). The PAST 4.13 statistical program (Hammer et al. Reference Hammer, Harper and Ryan2001) was used for the RMA with 95% bootstrapped confidence intervals [n = 1999 replications]. The Akaike Information Criterion (AIC) (Akaike Reference Akaike1973; Cavanaugh and Neath Reference Cavanaugh and Neath2019) using the R language version 4.3.1 (R Core Team 2023) and R package “rcompanion” version 2.4.30 (Mangiafico Reference Mangiafico2023) was used to compare the performance of the GIS versus graphite and direct carbonate versus graphite models. Scores were calculated by AIC = 2K – 2ln(L), where K is the number of model parameters and ln(L) is the log-likelihood of the model. The lowest AIC score is considered the better fit.

Results and Discussion

ACE Laboratory Blank (IAEA C1 and Pliocene Tridacna Shell) and Holocene Standard (IAEA C2 and CAHI) Performance

The quality of our carbonate 14C determinations is monitored by repeat analysis of certified and internal reference blanks and standards (Table 1). The pMC values overlap published values at one standard deviation (Rozanski et al. Reference Rozanski, Stichler, Gofiantini, Scott, Buekens, Kromer and van der Plicht1992; Mollenhauer et al. Reference Mollenhauer, Grotheer, Gentz, Bonk and Hefter2021). At the ACE Lab, the C1 procedural blank yields graphite and GIS ages of 50.6 ± 1.9 and 40.2 ± 3.9 ka BP, respectively. The Tridacna procedural blank yields graphite and GIS ages of 49.2 ± 2.4 and 42.2 ± 4.3 ka BP, respectively.

GIS versus Graphite and Direct Carbonate versus Graphite FMC Determinations

There is considerable time savings when using the GIS method. Processing samples for the direct carbonate method requires manual powdering of the samples, followed by manual subsampling, weighing, and manipulation of that powder plus a metal binder into small borosilicate tubes. The powder and metal mixture is then carefully poured into manually drilled aluminum cathodes before being manually pressed into targets. In contrast, the GIS preparation method is more streamlined and only requires weighing the cleaned carbonate samples into glass reaction vials. Processing a shell for GIS analysis takes roughly half as much time as processing a shell for direct carbonate analysis. For example, an undergraduate worker typically produces a batch of 51 unknowns for direct carbonate analysis in approximately 12 hours. That same worker can produce a batch of 48 unknowns for GIS analysis in approximately five to six hours. Converting carbonate to graphite is more time consuming. It typically takes approximately three hours to process eight unknowns to where the purified CO2 is converted to graphite.

Forty-one blank-corrected pMC values range from 98.68 to 20.51 (Supple. Info.), or spanning 14C ages of approximately 0.1 to 12.6 ka BP. One mid-Pleistocene Rangia shell that yields measurable carbon was analyzed in triplicate using each of the three AMS methods. The triplicates were averaged into a single value for each method to provide a lower limit to the compilation at a pMC value of approximately 0.36 (Supple. Info.), or approximately 45.1 ka BP. Six additional 14C-dead mollusk and echinoid samples were analyzed using each AMS method to determine and compare the limits of detection (Supple. Info.).

The RMA regression of the graphite versus GIS pMC values yields a slope with a 95% confidence interval that includes 1.000 (Figure 1; Table 2), which outperforms the RMA regression of the graphite versus direct carbonate values (Figure 1; Table 2). The RMA regression of graphite versus GIS and graphite versus direct carbonate yield slopes that are statistically indistinguishable from each other (p-value same slope = 0.003). Analyzing the graphite versus GIS and graphite versus direct carbonate data as linear regressions using AIC, where graphite pMC values are the dependent variable, yields a lower AIC score for graphite versus GIS (102.5) than for graphite versus direct carbonate (147.2). The lower AIC score indicates that graphite values are better explained by a model in which GIS values are the predictor than a model in which direct carbonate values are the predictor (Akaike Reference Akaike1973; Cavanaugh and Neath Reference Cavanaugh and Neath2019).

Figure 1 Reduced major axis (RMA) regression of paired rapid and graphite pMC determinations of biogenic carbonate. A – relationship between gas ion source (GIS) and graphite. B – relationship between direct carbonate and graphite. Analysis performed using PAST 4.13 statistical software (Hammer et al. Reference Hammer, Harper and Ryan2001). Fine dashed lines are 95% bootstrapped confidence intervals (n = 1999). Inset diagrams are frequency histograms of pMC differences, calculated as “GIS – graphite pMC” in A and “direct – graphite pMC” in B.

Table 2 Comparison of reduced major axis regression (RMA) of paired graphite and gas interface (GIS) or graphite and direct carbonate pMC determinations

* This study.

** Uncalibrated. Not normalized using a blank pMC value of 0.00 and IAEA C2 and UCI CAHI consensus values.

Missiaen et al. (Reference Missiaen, Wacker, Lougheed, Skinner, Hajdas, Nouet, Pichat and Waelbroeck2020). Radiocarbon ages converted to pMC using the equation pMC = (e14C yr/−8033) × 100.

An alternative method of data reduction that does not calibrate the unknown pMC values by forcing the IAEA C2 and CAHI standards in each batch through their consensus values does not appreciably change the outcome of the RMA comparison (Table 2). We conclude that GIS outperforms direct carbonate 14C determinations, acknowledging the relatively small number of paired analyses (n = 42) used in this comparison.

A larger study of 150 graphite versus direct carbonate pairs of mollusk, echinoid, and brachiopod carbonate yielded results that are more comparable to the graphite versus GIS comparison presented here (Table 2) (Bright et al. Reference Bright, Ebert, Kosnik, Southon, Whitacre, Albano, Flores, Frazer, Hua, Kowalewski, Martinelli, Oakley, Parker, Retelle, Ritter, Rivadeneira, Scarponi, Yanes, Zuschin and Kaufman2021). The RMA regression slope of graphite versus direct carbonate data in Bright et al. (Reference Bright, Ebert, Kosnik, Southon, Whitacre, Albano, Flores, Frazer, Hua, Kowalewski, Martinelli, Oakley, Parker, Retelle, Ritter, Rivadeneira, Scarponi, Yanes, Zuschin and Kaufman2021) and the graphite versus GIS regression slope in this study are not statistically distinguishable (p-value = 0.167). Additional comparisons using foraminifera have shown that graphite and GIS analyses also yield comparable results (Table 2) (Wacker et al. Reference Wacker, Lippold, Molnár and Schulz2013b; Gottschalk et al. Reference Gottschalk, Szidat, Michel, Mazaud, Salazar, Battaglia, Lippold and Jaccard2018; Missiaen et al. Reference Missiaen, Wacker, Lougheed, Skinner, Hajdas, Nouet, Pichat and Waelbroeck2020). Our mollusk and echinoid graphite versus GIS comparison yields slightly better results than the foraminifera comparisons, which we suspect reflects true age heterogeneity within the collections of analyzed foraminifera tests, which can be prone to mixing and reworking in some environments (Fagault et al. Reference Fagault, Tuna, Rostek and Bard2019; Dolman et al. Reference Dolman, Groeneveld, Mollenhauer, Ho and Laepple2021). Mollusk shells and echinoid tests have the advantage of being large enough to routinely date single specimens (e.g., Scarponi et al. Reference Scarponi, Kaufman, Amorosi and Kowalewski2013; Harnik et al. Reference Harnik, Torstenson and Williams2017; Kowalewski et al. Reference Kowalewski, Casebolt, Hua, Whitacre, Kaufman and Kosnik2018; Nawrot et al. Reference Nawrot, Berensmeier, Gallmetzer, Haselmair, Tomašových and Zuschin2022; Ritter et al. Reference Ritter, Erthal, Kosnik, Kowalewski, Coimbra, Caron and Kaufman2023) or individual growth lines (e.g., Lindauer et al. Reference Lindauer, Friedrich, van Gyseghem, Schöne and Hinderer2019; Towers Reference Towers2022).

The accuracy of the GIS 14C determinations outperforms the direct carbonate method when compared to graphite. Sixty percent (25/42) of the GIS determinations are within ±0.5 pMC of their graphite counterparts, compared to 26% (11/42) of the direct carbonate determinations (Figure 1). Furthermore, 86% (36/42) of the GIS determinations are within ±1.0 pMC of their graphite counterparts, compared to 67% (28/42) of the direct carbonate determinations (Figure 1). A larger study found that 39% (59/150) and 77% (116/150) of direct carbonate determinations were within ±0.5 pMC and ±1.0 pMC, respectively, of their graphite counterparts (Bright et al. Reference Bright, Ebert, Kosnik, Southon, Whitacre, Albano, Flores, Frazer, Hua, Kowalewski, Martinelli, Oakley, Parker, Retelle, Ritter, Rivadeneira, Scarponi, Yanes, Zuschin and Kaufman2021), indicating that GIS accuracy outperforms the direct carbonate method regardless of the number of samples analyzed. We suspect that part of the difference in performance is due to the smaller carbon content of the direct carbonate samples (36–60 μg C) compared to the GIS samples featured in this study (60–100 μg C), which may render the direct carbonate 14C determinations more vulnerable to any contamination introduced during the powdering process or with the addition of the metal binder. In addition, the direct carbonate 14C determinations were conducted over a roughly 5 yr span whereas most (38/42) of the GIS 14C determinations were run in a single day (Supple. Info.), thus, direct carbonate 14C measurements might have captured additional long-term variability in AMS performance. However, the average graphite and GIS pMC values over roughly two years of analysis for the IAEA C2 standard and the CAHI coral standard differ by 0.55 and 0.14 pMC, respectively (Table 1), suggesting the typically small differences between graphite and GIS pMC values measured in this study are robust and are not necessarily a function of the short analytical timeframe for the GIS samples.

Graphite 14C determinations have comparatively small, typically decadal-scale precision due in part to the long sputter times (approximately 75 minutes) and large number of carbon counts per analysis, which can be on the order of one million counts. In contrast, the GIS method analyzes a sample in roughly 15 minutes and generates a few thousand to a few tens of thousands of carbon counts, thus the lower precision is based on counting statistics. In our compilation, the GIS ages have precision on the order of ±60 to ±80 14C yrs back to about 6.5 ka BP. At around 9 ka BP, the precision decreases to about ±110 14C yrs, and decreases further to about ±130 14C yrs at 12.5 ka BP (Figure 2). The precision of the GIS is on par with that of the direct carbonate method and averages four to five times worse than graphite precision (Figure 2). Although the precision for the GIS and direct carbonate method both increase with age (Figure 2), the variation in the GIS precision (stdev = 17 14C years) is similar to that of graphite (stdev = 13 14C years), and both are better than for direct carbonate errors (stdev = 30 14C years). This is likely because the GIS and graphite methods spall C- ions from comparatively pure CO2 gas and graphite sources, respectively, which promotes more homogenous ion formation conditions (e.g., Middleton Reference Middleton1984; Bronk and Hedges 1987), whereas the direct carbonate method spalls C- ions from a mixture of powdered calcium carbonate and metal binder.

Figure 2 Cross-plot comparing analytical precision as a function of 14C age to 12.7 ka BP for gas ion source, direct carbonate, and graphite. White circles – graphite. Dark gray circles – gas ion source. White diamonds – direct carbonate.

Six Plio-Pleistocene carbonate samples, determined to be 14C-dead by graphite analysis, yielded finite pMC values and 29.4 to 36.7 ka BP ages when analyzed using the direct carbonate method but yielded non-finite pMC values when analyzed by GIS (Supple. Info). Bush et al. (Reference Bush, Santos, Xu, Southon, Thiagarajan, Hines and Adkins2013) also reported a direct carbonate result from an old coral sample that was approximately 6 14C kyr younger than the paired graphite result. They attributed the difference to a variety of issues, including extended storage (five years) of their carbonate materials between the graphite and direct carbonate analyses, modern contamination of the carbonate powder during processing, and heterogeneity in the carbonate sample. They concluded that lower beam currents and lower count rates, coupled with modern contamination of the carbonate powders during processing led to less favorable results on samples with ages over 30 ka BP. In our study, there is a six-year difference between the direct carbonate analyses of the six 14C-dead samples and their graphite counterparts, whereas the GIS counterparts were processed five months after the graphite analyses. Thus, the more similar GIS and graphite pMC results on the 14C-dead shells could also be a function of reduced sampling time between the two analyses, as proposed by Bush et al. (Reference Bush, Santos, Xu, Southon, Thiagarajan, Hines and Adkins2013). The IAEA C1 blank yielded a pMC value of 0.67 ± 0.33 (n = 39) when run by GIS at ACE, whereas 188 direct carbonate analyses of IAEA C1 processed at ACE and analyzed at UCI between 2017 and 2021 yielded a pMC value of 1.67 ± 0.69. We suspect that the poorer performance of the direct carbonate analyses results from carbon contamination during powdering and the addition of a metal binder that is known to contain some carbon (Bush et al. Reference Bush, Santos, Xu, Southon, Thiagarajan, Hines and Adkins2013; Hua et al. Reference Hua, Levchenko and Kosnik2019). Similarly, 142 direct carbonate analyses of the IAEA C2 standard processed at ACE and analyzed at UCI between 2017 and 2021 yielded a pMC value of 40.81 ± 0.63, whereas 89 GIS analyses of C2 analyzed at ACE over the past two years yielded a pMC value of 41.16 ± 0.64 (Table 2). The two mean values are statistically distinguishable at a 95% confidence interval (p-value = 6 x 10−5) and the GIS average falls closer to the consensus value of 41.14 ± 0.03 (Rozanski et al. Reference Rozanski, Stichler, Gofiantini, Scott, Buekens, Kromer and van der Plicht1992).

Thus, collectively, we conclude that the GIS method is superior to the direct carbonate method, especially when it comes to older samples (e.g., > 20 ka BP).

Conclusions

This study compared 42 samples of biogenic carbonate (mollusk shells and echinoid tests) that were dated using a MICADAS plus gas interface system (GIS) at Northern Arizona University’s Arizona Climate and Ecosystems Lab, by direct carbonate methods, and by standard graphite.

  • Preparing carbonate samples for GIS analysis is roughly 50% less time-consuming than required for the direct carbonate method.

  • The GIS method yields pMC values that are virtually indistinguishable from graphite (slope = 1.011 ± 0.006, 95% CI [0.997–1.023]; R2 = 0.999) and outperforms the direct carbonate method (slope = 0.978 ± 0.009, 95% CI [0.959 – 0.999]; R2 = 0.997).

  • Sixty percent (25/42) and 86% (36/42) of the GIS determinations are within ±0.5 and ±1.0 pMC of their graphite counterparts, respectively, compared to 26% (11/42) and 67% (28/42) of the direct carbonate determinations, respectively.

  • Errors on the GIS determinations are on par with those generated by the direct carbonate method and are roughly four to five times worse than errors derived from standard graphite determinations.

  • Six Plio-Pleistocene shells, 14C-dead as determined by graphite analysis, yielded finite direct carbonate ages but yielded expected non-finite GIS ages.

  • Our evaluation of graphite, GIS, and direct carbonate 14C determinations reveals that GIS outperforms direct carbonate, especially for older samples (e.g., > 20 ka BP) and at smaller sample sizes.

  • MICADAS + GIS 14C determinations of small biogenic carbonate samples are an effective alternative to graphite 14C determinations, especially when the precision of graphite is not required. This rapid and reliable dating method is useful for a broad range of applications, including the dating of minute specimens (e.g., small mollusks, foraminifera tests, ostracode valves, fish otoliths, etc.), determining degrees of time averaging across environments and taxa, and inferring temporal changes in population dynamics from post-mortem age distributions.

Supplementary material

To view supplementary material for this article, please visit https://doi.org/10.1017/RDC.2024.45

Acknowledgments

We appreciate the efforts of Aibhlin Ryan, Fallon Born, and Michelle Olichney as undergraduate laboratory technicians at Northern Arizona University. We thank John Southon for the direct carbonate analyses, Dan Muhs for providing several mollusk shells for the comparative analysis, and Sheila Griffen and Ellen Druffel for sharing the CAHI standard. DSK and EAGS acknowledge funding from the National Science Foundation for the GIS-CHS system (EAR 1855381 and 1919506). Additional support from the National Science Foundation was provided to PGH (EAR CAREER 2041667), JWH (EAR CAREER 1650745), MK (EAR 2127623), and MR (OPP-1744433). PGH acknowledges funding from the Gulf Research Program of the National Academies of Science, Engineering, and Medicine (GRP 2000008410). CF thanks FONDECYT 11200953 and Millennium Nucleus UPWELL NCN19 153. The authors declare no competing interests with this manuscript.

References

Akaike, H. 1973. Information theory and an extension of the maximum likelihood principle. In: Petrov, BH, Csaki F, editors. 2nd international symposium of information theory. Akademica Kaido, Budapest, Hungary. p. 267–281. doi: 10.1007/978-1-4612-1694-0_15.CrossRefGoogle Scholar
Albano, PG, Hua, Q, Kaufman, DS, Tomašových, A, Zuschin, M, Agadi, K. 2020. Radiocarbon dating supports bivalve-fish age coupling along a bathymetric gradient in high-resolution paleoenvironmental studies. Geology 48:589593. doi: 10.1130/G47210.1.CrossRefGoogle Scholar
Albano, PG, Sabbatini, A, Lattanzio, J, Päßler, JF, Steger, J, Hua, Q, Kaufman, DS, Szidat, S, Zuschin, M, Negri, A. 2023a. Alleged Lessepsian foraminifera prove native and suggest Pleistocene range expansions into the Mediterranean Sea. Marine Ecology Progress Series 700:6578. doi: 10.3354/meps14181.CrossRefGoogle Scholar
Albano, PG, Hua, Q, Kaufman, DS, Zuschin, M. 2023b. Young death assemblages with limited time-averaging in rocky and Posidonia ocanica habitats in the Mediterranean Sea. In: Nawrot R, Dominici S, Tomašových A, Zuchin M, editors. Conservation Palaeobiology of Marine Ecosystems. Geological Society, London, Special Publications 529:41–48. doi: 10.1144/SP529-2022-224CrossRefGoogle Scholar
Bright, J, Ebert, C, Kosnik, MA, Southon, JR, Whitacre, K, Albano, PG, Flores, C, Frazer, TK, Hua, Q, Kowalewski, M, Martinelli, JC, Oakley, D, Parker, WG, Retelle, M, Ritter, MN, Rivadeneira, MM, Scarponi, D, Yanes, Y, Zuschin, M, Kaufman, DS. 2021. Comparing direct carbonate and standard graphite 14C determinations of biogenic carbonates. Radiocarbon 63:387403. doi: 10.1017/RDC.2020.131.CrossRefGoogle Scholar
Bronk CR, Hedges REM. 1987. A gas ion source for radiocarbon dating. Nuclear Instrument and Methods in Physics Research 29:4549. doi: 10.1016/0168-583X(87)90201-1. CrossRefGoogle Scholar
Bush, SL, Santos, GM, Xu, X, Southon, JR, Thiagarajan, N, Hines, SK, Adkins, JF. 2013. Simple, rapid, and cost effective: a screening method for 14C analysis of small carbonate samples. Radiocarbon 55:631640. doi: 10.1017/S0033822200057787.CrossRefGoogle Scholar
Cavanaugh, JE, Neath, AA. 2019. The Akaike information criterion: background, derivation, properties, application, interpretation, and refinements. WIREs Computational Statistics 11:e1460. doi: 10.002/wics.1460.CrossRefGoogle Scholar
Dolman, AM, Groeneveld, J, Mollenhauer, G, Ho, SL, Laepple, T. 2021. Estimating bioturbation from replicated small-sample radiocarbon ages. Paleoceanography and Paleoclimatology 36, e2020PA004142. doi: 10.1029/2020PA004142.CrossRefGoogle Scholar
Dominguez, JG, Kosnik, MA, Allen, AP, Hua, Q, Jacob, DE, Kaufman, DS, Whitacre, K. 2016. Time-averaging and stratigraphic resolution in death assemblages and Holocene deposits: Sydney Harbour’s molluscan record. Palaios 31:564575. doi: 10.2110/palo.2015.087.CrossRefGoogle Scholar
Ebert, C, Schuur, EAG, Kaufman, D, Brown, J, Propster, J, Kelley, A, Bright, J, Carbone, MS, McKay, N, Koch, G. 2022. Capabilities, procedures, and summary statistics of the MICADAS and GIS at ACE Laboratory, Northern Arizona University. 24th Radiocarbon Conference; 10th 14C and Archaeology Conference, Zurich, 11–16 September 2022.Google Scholar
Fagault, Y, Tuna, T, Rostek, F, Bard, E. 2019. Radiocarbon dating small carbonate samples with the gas ion source of AixMICADAS. Nuclear Instrument and Methods in Physics Research B 455:276283. doi: 10.1016/j.nimb.2018.11.018.CrossRefGoogle Scholar
Gottschalk, J, Szidat, S, Michel, E, Mazaud, A, Salazar, G, Battaglia, M, Lippold, J, Jaccard, S. 2018. Radiocarbon measurements of small-size foraminiferal samples with the Mini Carbon Dating System (MICADAS) at the University of Bern: implications for paleoclimate reconstructions. Radiocarbon 60:469491. doi: 10.1017/RDC.2018.3.CrossRefGoogle Scholar
Hammer, Ø, Harper, DAT, Ryan, PD. 2001. PAST: paleontological statistics software package for education and data analysis. Palaeontologia Electronica 4(1). 9 p.Google Scholar
Harnik, PG, Torstenson, ML, Williams, MA. 2017. Assessing the effects of anthropogenic eutrophication on marine bivalve history in the northern Gulf of Mexico. Palaios 32:678688. doi: 10.2110/palo.2017.033.CrossRefGoogle Scholar
Hua, Q, Levchenko, VA, Kosnik, MA. 2019. Direct AMS 14C analysis of carbonate. Radiocarbon 61:14311440. doi: 10.1017/RDC.2019.24.CrossRefGoogle Scholar
Kosnik, MA, Hua, Q, Kaufman, DS, Kowalewski, M, Whitacre, K. 2017. Radiocarbon-calibrated amino acid racemization ages from Holocene sand dollars (Peronella peronii). Quaternary Geochronology 39:174188. doi: 10.1016/j.quageo.2016.12.001.CrossRefGoogle Scholar
Kowalewski, M, Casebolt, S, Hua, Q, Whitacre, KE, Kaufman, D, Kosnik, MA. 2018. One fossil record, multiple time-resolutions: comparative time averaging of mollusks and echinoids on a carbonate platform. Geology 46:5154. doi: 10.1130/G39789.1.CrossRefGoogle Scholar
Lindauer, S, Friedrich, R, van Gyseghem, R, Schöne, BR, Hinderer, M. 2019. Highly-resolved radiocarbon measurements on shells from Kalba, UAE, using carbonate handling system and gas ion source MICADAS. Nuclear Instruments and Methods in Physics Research B 455:146153. doi: 10.1016/j.nimb.2018.12.020.CrossRefGoogle Scholar
Longworth, BE, Robinson, LF, Roberts, ML, Beaupre, SR, Burke, A, Jenkins, WJ. 2013. Carbonate as a sputter target material for rapid 14C AMS. Nuclear Instruments and Methods in Physics Research B 294:328334. doi: 10.1016/j.nimb.2012.05.014.CrossRefGoogle Scholar
Lougheed, BC, Snowball, I, Moros, M, Kabel, K, Muscheler, R, Virtasalo, JJ, Wacker, L. 2012. Using an independent geochronology based on palaeomagnetic secular variation (PSV) and atmospheric Pb deposition to date Baltic Sea sediments and infer 14C reservoir age. Quaternary Science Reviews 42:4348. doi: 10.1016/j.quascirev.2012.03.013.CrossRefGoogle Scholar
Lougheed, BC, Metcalfe, B, Ninnemann, US, Wacker, L. 2018. Moving beyond the age-depth model paradigm in deep-sea palaeoclimate archives: dual radiocarbon and stable isotope analysis on single foraminifera. Climate of the Past 14:515526. doi: 10.5194/cp-14-626-2018.CrossRefGoogle Scholar
Mangiafico, SS. 2023. rcompanion: functions to support extension education program evaluation. Rutgers Cooperative Extension, New Brunswick, New Jersey, version 2.4.30. https://CRAN.R-project.org/package=rcompanion/ Google Scholar
Middleton, R. 1984. A review of ion sources for accelerator mass spectrometry. Nuclear Instruments and Methods in Physics Research B 5:193199. doi: 10.1016/0168-583X(84)90508-1.CrossRefGoogle Scholar
Missiaen, L, Wacker, L, Lougheed, BC, Skinner, L, Hajdas, I, Nouet, J, Pichat, S, Waelbroeck, C. 2020. Radiocarbon dating of small-sized foraminifer samples: insights into marine sediment mixing. Radiocarbon 62:313333. doi: 10.1017/RDC2020.13.CrossRefGoogle Scholar
Mollenhauer, G, Grotheer, H, Gentz, T, Bonk, E, Hefter, J. 2021. Standard operation procedures and performance of the MICADAS radiocarbon laboratory at Alfred Wegner Institute (AWI), Germany. Nuclear Instruments and Methods in Physics Research B 496:4551. doi: 10.1016/j.nimb.2021.03.016.CrossRefGoogle Scholar
Moss, DK, Ivany, LC, Judd, EJ, Cummings, P, Bearden, CE, Kim, WJ, Artruc, EG, Driscol, JR. 2016. Lifespan, growth rate, and body size across latitude in marine Bivalvia, with implications for Phanerozoic evolution. Proceedings of the Royal Society B 283:20161364. doi: 10.1098/rspb.2016.1364.CrossRefGoogle ScholarPubMed
Nawrot, R, Berensmeier, M, Gallmetzer, I, Haselmair, A, Tomašových, A, Zuschin, M. 2022. Multiple phyla, one time resolution? Similar time averaging in benthic foraminifera, mollusk, echinoid, crustacean, and otolith fossil assemblages. Geology 50:902906. doi: 10.1130/G49970.1.CrossRefGoogle Scholar
New, E, Yanes, Y, Cameron, RAD, Miller, JH, Teixeira, D, Kaufman, DS. 2019. Aminochronology and time averaging of Quaternary land snail assemblages from colluvial deposits in the Madeira Archipelago, Portugal. Quaternary Research 92:483496. doi: 10.1017/qua.2019.1.CrossRefGoogle Scholar
Parker, WG, Yanes, Y, Hernández, EM, Hernández Marreno, JC, Paris, J, Surge, D. 2019. Scale of time-averaging in archaeological shell middens from the Canary Islands. The Holocene 30:258271. doi: 10.1177/0959683619883020.CrossRefGoogle Scholar
Quinn, GP, Keough, MJ. 2002. Experimental design and data analysis for biologists. Cambridge University Press, Cambridge, UK. 537 p. doi: 10.1017/CBO9780511806384.CrossRefGoogle Scholar
Ritter, MN, Erthal, F, Kosnik, MA, Ciombra, JC, Kaufman, DS. 2017. Spatial variation in the temporal resolution of subtropical shallow-water molluscan death assemblages. Palaios 32:572583. doi: 10.2110/palo.2017.003.CrossRefGoogle Scholar
Ritter, MN, Erthal, F, Kosnik, MA, Kowalewski, M, Coimbra, JC, Caron, F, Kaufman, DS. 2023. Onshore-offshore trends in the temporal resolution of molluscan death assemblages: how age-frequency distributions reveal Quaternary sea-level history. Palaios 38:148157. doi: 10.2110/palo.2021.041.CrossRefGoogle Scholar
Rozanski, K, Stichler, W, Gofiantini, R, Scott, EM, Buekens, RP, Kromer, B, van der Plicht, J. 1992. The IAEA 14C Intercomparison Exercise 1990. Radiocarbon 34:506519. doi: 10.1017/S0033822200063761.CrossRefGoogle Scholar
Ruff, M, Wacker, L, Gäggeler, HW, Suter, M, Synal, H-A, Szidat, S. 2007. A gas ion source for radiocarbon measurements at 200 kV. Radiocarbon 49:307314. doi: 10.1017/S0033822200042235.CrossRefGoogle Scholar
Ruff, M, Fahrni, S, Gäggeler, HW, Hajdas, I, Suter, M, Synal, H-A, Szidat, S, Wacker, L. 2010. On-line radiocarbon measurements of small samples using elemental analyzer and MICADAS ion source. Radiocarbon 52:16451656. doi: 10.1017/S003382220005637X.CrossRefGoogle Scholar
Sanchez, W, Yanes, Y, Linstädter, J, Hutterer, R. 2022. Chronology, time averaging, and oxygen isotope composition of harvested marine mollusk assemblages from Ifri Oudadane, northeast Morocco. Quaternary Research 106:147161. doi: 10.1017/qua.2021.60.CrossRefGoogle Scholar
Scarponi, D, Kaufman, D, Amorosi, A, Kowalewski, M. 2013. Sequence stratigraphy and the resolution of the fossil record. Geology 41:239242. doi: 10.1130/G33849.1.CrossRefGoogle Scholar
Smith, RJ. 2009. Use and misuse of the reduced major axis for line fitting. American Journal of Physical Anthropology 140:476486. doi: 10.1002/ajpa.21090.CrossRefGoogle ScholarPubMed
Steger, J, Bosnjak, M, Belmaker, J, Galil, BS, Zuschin, M, Albano, PG. 2022. Non-indigenous molluscs in the Eastern Mediterranean have distinct traits and cannot replace historic ecosystem functioning. Global Ecology and Biogeography 31:89102. doi: 10.111/geb.13415.CrossRefGoogle Scholar
Stuvier, M, Polach, HA. 1977. Discussion: reporting of 14C data. Radiocarbon 19:355363. doi: 10.1017/S0033822200003672.CrossRefGoogle Scholar
Synal, H-A, Stocker, M, Suter, M. 2007. MICADAS: A new compact radiocarbon AMS system. Nuclear Instruments and Methods in Physics Research B 259:713. doi: 10.1016/j.nimb.2007.01.138.CrossRefGoogle Scholar
Towers, E. 2022. Arctica islandica – Annually banded mollusc offers high temporal resolution record into the end of North Sea Little Ice Age. Degree Project at the Department of Earth Sciences. Uppsala University, Uppsala, Sweden. 94 p.Google Scholar
Tuna, T, Fagault, Y, Bonvalot, L, Capano, M, Bard, E. 2018. Development of small CO2 gas measurements with AixMICADAS. Nuclear Instruments and Methods in Physics Research B 427:9397. doi: 10.1016/j.nimb.2018.09.012.CrossRefGoogle Scholar
Vogel, JS, Southon, JR, Nelson, DE, Brown, TA. 1984. Performance of catalytically condensed carbon for use in accelerator mass spectrometry. Nuclear Instruments and Methods in Physics Research B 5:289293. doi: 10.1016/0168-583X(84)90529-9.CrossRefGoogle Scholar
Wacker, L, Bonani, G, Friedrich, M, Hajdas, I, Kromer, B, Nemec, M, Ruff, M, Suter, M, Synal, H-A, Vockenhuber, C. 2010. MICADAS: Routine and high-precision radiocarbon dating. Radiocarbon 52:252262. doi: 10.1017/S0033822200045288.CrossRefGoogle Scholar
Wacker, L, Fahrni, SM, Hajdas, I, Molnar, M, Synal, H-A, Szidat, S, Zhang, YL. 2013a. A versatile gas interface for routine radiocarbon analysis with a gas ion source. Nuclear Instruments and Methods in Physics Research B 294:315319. doi: 10.1016/j.nimb.2012.02.009.CrossRefGoogle Scholar
Wacker, L, Lippold, J, Molnár, M, Schulz, H. 2013b. Towards radiocarbon dating of single foraminifera with a gas ion source. Nuclear Instruments and Methods in Physics Research B 264:307310. doi: 10.1016/j.nimb.2012.08.038.CrossRefGoogle Scholar
Figure 0

Table 1 Mean pMC values for reference material (IAEA) and laboratory standards analyzed at Northern Arizona University ACE lab. IAEA reference values are from Rozanski et al. (1992)

Figure 1

Figure 1 Reduced major axis (RMA) regression of paired rapid and graphite pMC determinations of biogenic carbonate. A – relationship between gas ion source (GIS) and graphite. B – relationship between direct carbonate and graphite. Analysis performed using PAST 4.13 statistical software (Hammer et al. 2001). Fine dashed lines are 95% bootstrapped confidence intervals (n = 1999). Inset diagrams are frequency histograms of pMC differences, calculated as “GIS – graphite pMC” in A and “direct – graphite pMC” in B.

Figure 2

Table 2 Comparison of reduced major axis regression (RMA) of paired graphite and gas interface (GIS) or graphite and direct carbonate pMC determinations

Figure 3

Figure 2 Cross-plot comparing analytical precision as a function of 14C age to 12.7 ka BP for gas ion source, direct carbonate, and graphite. White circles – graphite. Dark gray circles – gas ion source. White diamonds – direct carbonate.

Supplementary material: File

Bright et al. supplementary material

Bright et al. supplementary material
Download Bright et al. supplementary material(File)
File 29.9 KB