A methods assessment and recommendations for improving calculations and reducing uncertainties in the determination of 210Po and 210Pb activities in seawater

In marine systems, 210Po and 210Pb disequilibria are being increasingly used to examine oceanic particle formation and export. Here, an updated assessment of current methods for determining 210Po and 210Pb activity in marine samples is provided and includes a complete description of the vast number of calculations and uncertainties associated with Po and Pb loss, decay, and ingrowth during sample processing. First, we summarize the current methods for the determination of 210Po and 210Pb activities in dissolved and particulate seawater samples and recommend areas for improvement. Next, we detail the calculations and associated uncertainties using principles of error propagation, while also accounting for radionuclide ingrowth, decay, and recovery. A spreadsheet reporting these calculations is included as a downloadable Web Appendix. Our analysis provides insight into the contributions of the relative uncertainty for each parameter considered in the calculation of final 210Po and 210Pb activities and gives recommendations on how to obtain the most precise final values. For typical experimental conditions in open seawater, we show that our method allows calculating 210Pb activity with a relative uncertainty of about 7%. However for 210Po activities, the final relative uncertainty is more variable and depends on the 210Po/210Pb activity ratio in the initial sample and the time elapsed between sampling and sample processing. The lowest relative uncertainties on 210Po that can be obtained by this method is 6% and can only be obtained for samples with high 210Po/210Pb activity ratios (>1) that were rapidly processed. *Corresponding author: E-mail: [email protected] Phone: (302) 831-2558, Fax: (302) 831-4575 Acknowledgments Full text appears at the end of the article. DOI 10.4319/lom.2013.11.561 Limnol. Oceanogr.: Methods 11, 2013, 561–571 © 2013, by the American Society of Limnology and Oceanography, Inc. LIMNOLOGY and OCEANOGRAPHY: METHODS ing the accuracy and precision of 210Po and 210Pb measurements in seawater (Church et al. 2012). An initial assessment of the precision and accuracy of current procedures for 210Po and 210Pb measurement was conducted as part of a recent intercalibration exercise using dissolved and particulate seawater samples (Church et al. 2012). One of the major conclusions was that while the results reported by laboratories agree relatively well (relative standard deviation, RSD < 50%) for samples with high 210Po and 210Pb activities (> 0.1 dpm), this agreement became rather poor (RSD up to 200%) for lower activity samples. Although the authors were not able to precisely identify the sources of the disagreements, they suggested that one possibility includes the manner in which the 210Po and 210Pb ingrowth, decay, and recovery calculations were conducted. Their study further revealed that there were various methodologies in how uncertainties and error propagation were considered, which resulted in a large range in the specific activity uncertainties reported. The intercalibration effort by Church et al. (2012), therefore, suggests that there is a need for the scientific community to concur on “best practices” for 210Po and 210Pb measurement as well as final data calculations. In this context, the aims of this paper are (1) to review the protocols used for 210Po and 210Pb measurements in seawater and provide recommendations for improving the method’s accuracy, (2) to detail the calculations necessary for including isotopic recoveries and decay/ingrowth corrections during sample processing, (3) to develop a protocol for error propagation and to identify the main sources of uncertainty in the final data, and (4) to recommend methods for lowering the relative uncertainty. A practical spreadsheet, which follows step-wise the complex formulations reported in the paper, has been made available as a downloadable Web Appendix. Materials and procedures General procedure for sample collection and processing A typical protocol used for seawater sample processing of 210Pb and 210Po is presented in Fig. 1 and assumes analysis of 210Po and 210Pb by α spectrometry as described by Fleer and Bacon (1984). The seawater sample is collected as either total (unfiltered) or dissolved (filtered) with the particulate fraction measured separately. After collection, the dissolved or total sample is acidified to pH 1-2 with HCl, spiked with a well-calibrated 209Po tracer solution (T1/2 = 102 y) and a wellstandardized stable lead carrier added to monitor the losses of Po and Pb during sample processing. Some laboratories also use 208Po (T1/2 = 2.9 y) in a double spike technique, the former added to monitor the initial yield and the latter to act as a second yield tracer (Friedrich and Rutgers van der Loeff 2002). In the following, we limit our discussions to the single-209Po spike method, as tailing/peak overlap corrections for 208Po (α energy of 5.11 MeV) and 210Po (5.31 MeV) add another level of complexity not necessary for this discussion (Fleer and Bacon 1984). For both total and dissolved samples, Po and Pb can be preconcentrated from large volumes of seawater via co-precipitation with Fe(OH)3 (Thomson and Turekian 1976; Nozaki 1986), Co-APDC (Fleer and Bacon 1984) or MnO2 (Bojanowski et al. 1983). The precipitate is then dissolved in an acid solution (generally HCl for Fe(OH)3 and MnO2, HNO3 for Co-APDC) and, after evaporation to near-dryness, recovered in a 0.5-2M HCl solution. For particulate samples, the solid phase is completely dissolved using a mixture of strong acids (including HF) and, after evaporation to near-dryness, also recovered in 0.5-2 M HCl solution. The Po nuclides are then plated by spontaneous deposition onto a silver disc (Flynn 1968). Silver discs, typically 1-2 cm in diameter, can be obtained with greater than 99.99% purity. They are first shined with a commercial silver Rigaud et al. 210Po and 210Pb activity calculations in seawater 562 Fig. 1. Sample processing scheme for the determination of total, dissolved, and particulate 210Po and 210Pb activities. The times term (t) required for each step used in the calculation are provided. polish and then washed using water and ethanol. One side of the disc is covered by an inert substance, such as rubber cement, electronic spray (e.g glyptol) or plastic tape, so that Po nuclides are plated on only one side. For samples processed using Fe(OH)3 co-precipitation, ascorbic acid should be added to the plating solution before plating in order to avoid Fe(OH)3 formation on the plate. The Po activities are measured after deposition by α spectrometry. Any remaining 210Po and 209Po in solution is removed by either a second deposition onto another silver disc or scrap silver and/or using anion exchange resin such as AG-1 × 8 (Sarin et al. 1992) or Sr Spec resin (Vajda et al. 1994). Note that the Po and Pb separation using Sr Spec resin can also be conducted prior to the first plating (Bojanowski et al. 1983). After separation, the final eluate containing the 210Pb is re-spiked with 209Po and stored for greater than 6 months to allow in-growth of 210Po from 210Pb. At that time, the 210Pb activity of the sample is determined by replating the eluate solution on a new silver disc and measuring the in-growth of 210Po (Fig. 1). The determination of the initial activities of 210Po and 210Pb in the sample at the time of collection requires several corrections that account for decay and ingrowth between the time of collection and processing, together with corrections for Po and Pb chemical recoveries (detailed in Calculations and associated uncertainties section). Improved accuracy of the method Use of ion exchange resin for Po and Pb separation Complete removal of Po isotopes after the initial plating procedure is a key component for increasing the accuracy of the method. Indeed, incomplete removal of Po isotopes prior to storage will affect the final calculated 210Pb activity, and thus that of 210Po. There are two methods to remove the residual Po isotopes: replating the solution or separation onto ion exchange resin. Replating of samples may not be sufficient to ensure complete removal of residual Po as a fraction of the Po nuclides may remain in solution. Based on the Po recovery efficiency obtained on about 80 processed samples, we found that 17 ± 19% of the Po introduced into the plating solution can remain in solution after the first plating. Note that such results are in agreement with previous findings (Flynn 1968). Assuming the same efficiency for the cleaning plate, residual Po nuclides of 3 ± 3% will remain in solution. In contrast, ion exchange experiments with spiked solutions of known amounts of 209Po and 210Po showed a quantitative removal of Po (98.9 ± 1.4%, n = 6, for AG-1 × 8 in HCl 9M; Rigaud unpubl. data). Thus, although both methods are valid, we recommend the use of ion exchange resin to obtain the most accurate results. Precise determination of 210Pb recovery efficiency during sample processing The Pb recovery is quantified by measuring stable lead concentrations in known aliquots of the plating solution. Usually only one aliquot is collected after the second plating, providing information on the total Pb loss that occurs during complete sample processing. This loss is generally assumed to occur only during sample extraction (filtered or unfiltered) or dissolution (particulate). For example, experiments using the most common extraction protocols (Fe(OH)3 and Co-APDC methods) on about 80 dissolved seawater samples showed that the Pb extraction efficiency during initial seawater extraction were 70 ± 10% and 88 ± 18% for Fe(OH)3 and Co-APDC, respectively. However, Pb losses have also been shown to occur during ion exchange resin procedure. For comparison, the Pb recovery during resin separation was 90 ± 17% and 96 ± 16% for the Fe(OH)3 and Co-APDC method, respectively. Thus, although the Pb recovery during the separation onto resin is significantly higher than during the extraction procedure, it is not always complete and can constitute a non-negligible loss of Pb during sample processing. In the case of particulate samples, for which no extraction is required (Fig. 1), ion exchange separation is the main step for Pb loss during sample processing. Thus precise assessment of Pb isotope loss during both extraction and ion exchange procedure needs to be considered to accurately correct for the actual 210Pb fraction that leads to 210Po ingrowth between the extraction and the first plating dates as well as during the storage period. Therefore we strongly recommend collecting another aliquot of plating solution between the first plating and the resin procedure (Fig. 1). Note also that the fraction of the solution removed with the aliquot, although small, must also be accounted for during the activity correction process. Indeed, the fraction of 210Pb removed with the aliquot will not contribute to 210Po ingrowth during the storage period, resulting in an underestimation of 210Pb and an overestimation of 210Po (detailed in Calculations and associated uncertainties section).