Integrated Extraction Chromatographic Separation of the Lithophile

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.6b04289. Information as mentioned ...
1 downloads 0 Views 734KB Size
Download PDF
Article pubs.acs.org/ac

Integrated Extraction Chromatographic Separation of the Lithophile Elements Involved in Long-Lived Radiogenic Isotope Systems (Rb− Sr, U−Th−Pb, Sm−Nd, La−Ce, and Lu−Hf) Useful in Geochemical and Environmental Sciences Christian Pin† and Abdelmouhcine Gannoun*,‡ †

Département de Géologie, C.N.R.S and Université Blaise Pascal, 5 rue Kessler, 63038 Clermont-Ferrand, Cedex, France Laboratoire Magmas et Volcans, Université Blaise Pascal, CNRS UMR 6524, 6 Avenue Blaise Pascal, 63178 Aubière, Cedex, France

‡

S Supporting Information *

ABSTRACT: A fast and efficient sample preparation method in view of isotope ratio measurements is described, allowing the separation of 11 elements involved, either as “parent” or as “daughter” isotopes, in six radiogenic isotope systems used as chronometers and tracers in earth, planetary, and environmental sciences. The protocol is based on small extraction chromatographic columns, used either alone or in tandem, through which a single nitric acid solution is passed, without any intervening evaporation step. The columns use commercially available extraction resins (Sr resin, TRU resin, Ln resin, RE resin, and again Ln resin for isolating Sr and Pb, LREE then La−Ce−Nd−Sm, Lu(Yb), and Hf, Th, and U, respectively) along with an additional, in-house prepared resin for separating Rb. A simplified scheme is proposed for samples requiring the separation of Sr, Pb, Nd, and Hf only. Adverse effects of troublesome major elements (Fe3+, Ti) are circumvented by masking with ascorbic acid and hydrofluoric acid, respectively. Typical recoveries in the 85−95% range are achieved, with procedural blanks of 10−100 pg, negligible with regard to the amounts of analytes processed. The fractions separated are suitable for high precision isotope ratio measurements by TIMS or MC-ICP-MS, as demonstrated by the repeat analyses of several international reference materials of basaltic composition for 87Sr/86Sr, 208,207,206Pb/204Pb, 143Nd/144Nd, 176Hf/177Hf, and 230Th/232Th. Concentration data could be obtained by spiking and equilibrating the sample with appropriate isotopic tracers before the onset of the separation process and, finally, measuring the isotope ratios modified by the isotope dilution process.

L

different behavior, producing variable degrees of mutual chemical fractionation, provides more complete insight into physicochemical processes which occurred in the source of the studied samples. It is therefore of interest to develop analytical protocols that allow as many as possible of those elements involved in radioactive systems to be isolated from a single sample aliquot. Beyond the obvious practical advantage of saving analyst time and effort and decreasing reagent volumes and running costs, measuring several isotopic systems on the very same sample aliquot is crucial when only a small mass of sample is available. Finally, this approach alleviates the risk of blurring fine correlations between the different systems studied by analyzing not strictly identical aliquots due to subtle heterogeneity at the sampling scale. Previous attempts based on combined cationexchange and extraction chromatography achieved the concomitant separation of Sr, Nd, and Hf,4 or Rb, Sr, Lu, Hf,

ong-lived, naturally radioactive isotopes of eight elements (uranium, thorium, rhenium, lutetium, samarium, lanthanum, rubidium, and potassium) decay into stable radiogenic isotopes belonging to eight other elements (lead, osmium, hafnium, neodymium, cerium (and barium), strontium, and argon, respectively). These radioactive/radiogenic (or “parent”/“daughter”) pairs of isotopes can place important constraints on a wealth of processes occurring on geological time scales. Indeed, besides invaluable pieces of chronological information, they provide a range of isotopic tracers that reflect time-integrated parent/daughter elemental ratios, thereby shedding light on ancient chemical fractionation events in the source reservoir of the studied samples. This property of radiogenic isotopes has proven to be most useful for dating and tracing a range of natural and anthropogenic processes.1−3 Each of the aforementioned radioactive/radiogenic systems may provide useful pieces of information on its own. However, multi-isotopic approaches combining several radiogenic systems offer much better potential because (i) using systems with variable half-lives allow processes with a wide range of time scales to be probed and (ii) using several pairs of elements with © 2017 American Chemical Society

Received: November 2, 2016 Accepted: January 25, 2017 Published: January 25, 2017 2411

DOI: 10.1021/acs.analchem.6b04289 Anal. Chem. 2017, 89, 2411−2417

Article

Analytical Chemistry

Figure 1. Flow chart of the full separation scheme, allowing the separation of all the lithophile elements involved in long-lived radioactive/radiogenic systems: Rb/Sr; U−Th/Pb; Sm/Nd; La/Ce−Ba; Lu/Hf. See text for details.

Sm, and Nd,5 but the important U−Th−Pb system was not comprised. Likewise, a miniaturized extraction chromatography method allowed Pb, Sr, and Sm−Nd to be isolated, but neither Hf nor the other elements possessing parent isotopes (Rb, U, Th, and Lu) were involved.6 In this article, we describe an integrated scheme entirely based on miniaturized extraction chromatography (EXC), allowing the online separation of the lithophile elements (i.e., with an affinity for silicates) having radiogenic isotopes, namely, Sr, [Ba], Ce, Nd, Hf, Pb, Th, and U from a single nitric acid solution of silicate samples. The elements separated in this way are suitable for high precision isotope ratio measurements by TIMS and/or MC-ICP-MS, providing data directly usable for studies of geologically young samples. Besides these elements with a radiogenic component, our protocol also offers the possibility to isolate the elements possessing their corresponding radioactive isotope (Rb, La, Sm, Lu, Th, U). This might allow, by using the isotope dilution method, the determination of parent/daughter ratios with a high degree of precision and accuracy, as required in investigations of geologically ancient samples.

Sample Digestion. Following the addition of enriched isotopic tracer(s) in case the total spiking version of the isotope dilution method is used for measuring concentration data, about 50 to 150 mg (depending on the concentrations of the target elements) of the sample powder is weight, wetted with water, then dissolved on a hot plate at ∼90 °C with 1 to 2 mL of 29 mol L−1 HF and 1 mL of 7 mol L−1 HNO3. Then, the sample is gently (ca. 50 °C) evaporated to dryness, before two times treatments with a mixture of 6 mol L−1 HCl and conc. HNO3. The solid residue left is treated again with conc. HNO3 to convert the sample to nitrates. These are taken up with 1.5 to 2 mL of 2 mol L−1 HNO3, and the resulting liquid is centrifuged 5 min at 5000 rpm, in order to separate a supernatant clear solution from a solid pellet consisting (occasionally) of refractory minerals and (almost invariably) of more or less gelatinous, whitish sparingly soluble fluorides. To bring them into solution, these compounds are resuspended in 100−300 μL (depending on the size of solid residue) of a 2 mol L−1 HNO3 solution saturated in boric acid by using a vortex mixer and treated in an ultrasonic bath and/or a hot plate. Alternatively, a saturated solution of Al(NO3)3 in 2 mol L−1 HNO3 can be used for the same purpose. The resulting solution is combined with the bulk (supernatant) sample solution and is ready for column separation. Column Separation. Sr−Pb Column. A few minutes before starting the column separation, 5 μL of 48% HF is added to the sample solution (thus, 2 mol L−1 HNO3 ∼0.07 mol L−1

■

ANALYTICAL METHODS This section presents a full description of the proposed protocol. Details on the reagents and materials, especially the extraction chromatography materials, are given in the Supporting Information. 2412

DOI: 10.1021/acs.analchem.6b04289 Anal. Chem. 2017, 89, 2411−2417

Article

Analytical Chemistry HF) which is loaded portion-wise on the first, Sr resin column, preconditioned with 0.5 mL of 2 mol L−1 HNO3 ∼0.07 mol L−1 HF. After complete draining of the sample solution, the sample beaker, pipet tip, and column reservoir are rinsed with 0.25 mL of 2 mol L−1 HNO3 ∼0.07 mol L−1 HF. Then, the column reservoir is rinsed again with 0.25 mL of the same medium. The loading plus rinsing effluent is collected in a clean PFA vial and left aside for subsequent extraction of the other analytes. Following this loading step, allowing the extraction of Ba, Sr, and Pb to be achieved by the Sr resin, 2 mL of 7 mol L−1 HNO3 are passed through the column to get rid of residual matrix elements and barium. Then, 2 × 0.25 mL of 2 mol L−1 HNO3 is passed through the column, before Sr is stripped with 2 × 1 mL of 0.05 mol L−1 HNO3, which is collected in a single use polystyrene pot and gently evaporated to dryness under a lamp. This Sr fraction is ready for mass spectrometric measurements by TIMS. Finally, 2 × 1 mL of 3 mol L−1 HCl is passed through the column, before Pb is stripped from the resin with 2 × 1 mL of 6 mol L−1 HCl. The Pb fraction is collected in a PFA vessel and evaporated to dryness. Before isotope analysis, it is further treated with 100−200 μL of 14 mol L−1 HNO3 in order to destroy any organic matter leached from the resin which might cause poor Pb recovery when the sample is finally dissolved with dilute nitric acid prior to MCICP-MS analyses. Rb Column (Optional). If an 87Rb-enriched tracer was added to the sample for measuring the Rb concentration by isotope dilution, then a small aliquot (∼100 μL) of the bulk sample solution left after Sr and Pb extraction by the Sr column is taken and loaded onto a microcolumn containing ∼50 μL of the Cs−Rb resin. Cs and Rb are extracted, while matrix elements (including any residual Sr not retained by the Sr−Pb column) pass through the column. The column is rinsed with 3 × 50 μL of 2 mol L−1 HNO3 ∼0.07 mol L−1 HF. About 250 μL loading + rinsing volumes, containing the other elements of interest (REE, Th, U, Hf), can be returned to the main solution. Then, the column is further rinsed with 3 × 200 μL of 1 mol L−1 HNO3, before Rb (and most of Cs) is stripped by using 3 × 0.5 mL of 6 mol L−1 HNO3 (alternatively, 6 mol L−1 HCl might be used) which is collected in a PFA vial and evaporated to dryness. HFSE and TRU Columns in Tandem. Before the next separation steps, 100−200 mg (depending on the quantity of iron in the sample) of ascorbic acid (C6H8O6) is added to the bulk sample solution left after extracting Sr and Pb, in order to reduce as much as possible of Fe(III). After allowing 10−15 min for the dissolution and reducing action of ascorbic acid to take place (conveniently, during the Sr−Pb column processing), the solution is loaded onto the HFSE column and the TRU column in tandem configuration (Figure 1) previously conditioned by passing through the superposed columns of 0.5 mL of 2 mol L−1 HNO3 ∼0.07 mol L−1 HF containing 50 mg mL−1 of ascorbic acid. Then, the sample beaker, pipet tip, and HFSE column reservoir are rinsed with 2 × 0.25 mL of 2 mol L−1 HNO3 ∼0.07 mol L−1 HF containing 50 mg mL−1 of C6H8O6. At that stage, the high field strength elements (Zr, Hf, Th, U, and part of Ti) and the HREE have been extracted by the upper HFSE column, while the LREE passed through and were extracted by the lower TRU column. Residual amounts of Fe and Ti are the only major elements accompanying Zr, Hf, Th, and U on the HFSE column.

The columns are then decoupled and processed separately, with priority given to the isolation of the LREE as a group on the TRU column because this step will be followed by a more time-consuming one aiming to split this group into individual lanthanide elements. TRU Column, after Decoupling from the HFSE olumn. Following decoupling, the TRU column is first rinsed with 0.25 mL of 2 mol L−1 HNO3 ∼0.07 mol L−1 HF; then, three fractions of 0.5 mL of 1 mol L−1 HNO3 are passed to remove matrix elements. In preparation for the back-extraction of the LREE, 2 × 100 μL of 0.05 mol L−1 HNO3 followed by 100 of 0.05 mol L−1 HCl are passed through the column and discarded, before the tip of the column is placed above the reservoir of the LREE column, which was previously conditioned with 2 × 100 μL 0.05 mol L−1 HCl. TRU and LREE Columns in Tandem. The LREE (i.e., from La to Eu) is stripped from the TRU column with 3 × 0.35 mL of 0.05 mol L−1 HCl, a medium which ensures the quantitative sorption of these elements by the downstream LREE column filled with the HDEHP-based Ln resin. LREE Column, after Decoupling from the TRU Column. The LREE column is then rinsed twice with 0.1 mL of 0.05 mol L−1 HCl, before starting the sequential elution of the LREE. This is made first by using 0.20 mol L−1 HCl from La to Nd and then increasing the acid strength to 0.60 mol L−1 to strip Sm. Specifically, 0.75 mL is passed through the column and discarded, before La is collected in a 0.9 mL fraction. Then, 0.30 mL is discarded before Ce is collected in the next 1.0 mL. Most of Pr is subsequently eluted with 0.3 mL (discarded), before Nd is stripped with a further 1.6 mL. The eluting agent is then switched to 0.60 mol L−1 HCl, the first 0.5 mL of which is discarded, before the Sm fraction is eluted with 0.45 mL of the same acid. The Ce and Nd fractions (and those containing La and Sm, if needed for isotope dilution measurements) are collected in disposable polystyrene pots and evaporated to dryness. HFSE Column, after Decoupling from the TRU Column. The column is first rinsed with 3 × 0.5 mL of 6 mol L−1 HNO3 0.145 mol L−1 HF, which is discarded. Then, a fraction containing Lu plus part of the Yb is stripped with 1.5 mL of the same medium. This fraction still contains some Fe and Ti. Optionally, it can be directly loaded onto the HREE column for a further cleanup of lutetium (see below). The elution is continued with 1.0 mL of 6 mol L−1 HNO3 0.145 mol L−1 HF, before Zr (accompanied by most of Mo) is stripped with a further 3.0 mL of the same medium. Then, 0.5 mL of 2 mol L−1 HNO3 0.0725 mol L−1 HF, followed by 0.5 mL of 2 mol L−1 HNO3 0.20 mol L−1 HF, is passed through the column and discarded because they do not contain appreciable quantities of Hf. This element is finally eluted (along with some residual Mo) with a further 1.0 mL of the same stripping agent and collected in a PFA vial. Th and U are still held by the resin. If these elements are needed, they can be recovered either sequentially by using 2 mL of 0.29 mol L−1 HF, followed by 2 mL of 2.9 mol L−1 HF, respectively, or together with 2 mL of 2.9 mol L−1 HF. HREE Column (Optional). The HREE fraction (1.5 mL of 6 mol L−1 HNO3 0.145 mol L−1 HF) obtained from the HFSE column contains essentially all the Lu, accompanied by ca. 50% of the Yb contained in the sample, along with a few tens of micrograms of residual Fe and Ti. If a purer fraction with a lower Yb/Lu ratio is desired, it is possible to elute this fraction onto the HREE column in order to (i) get rid of the unwanted 2413

DOI: 10.1021/acs.analchem.6b04289 Anal. Chem. 2017, 89, 2411−2417

Article

Analytical Chemistry residual major elements and (ii) further increase the Lu/Yb ratio, thereby diminishing the magnitude of the correction of the isobaric interference of 176Yb on 176Lu. After the loading step, the column reservoir is rinsed twice with 0.25 mL of 6 mol L−1 HNO3 0.145 mol L−1 HF, and 1.5 mL of the same solution is passed through the column to strip titanium. Then, 1 mL of 2 mol L−1 HNO3 is passed to reduce the nitric acid strength in preparation to the elution of residual iron (expressed by a bright yellow band on top of the column) with 1 mL of 2 mol L−1 HNO3 containing 50 mg/mL of ascorbic acid. This is followed by 1.7 mL of pure 2 mol L−1 HNO3 removing interstitial ascorbic acid from the column. Finally, a Lu fraction (with Lu ≫ Yb) is eluted with a further 1.5 mL of 2 mol L−1 HNO3. The whole process is depicted in Figure 1 and summarized in Table S1 of the Supporting Information. Simplified, “Daughter Only” Separation Scheme. An alternative version of the full separation protocol has been developed for those samples which do not require a precise measurement of parent/daughter elemental ratios. This simplified version allows the isolation of the elements possessing radiogenic isotopes: Sr, Pb, Nd, Ce, and Hf. Uranium and thorium, which have both radioactive and radiogenic isotopes, are also included in this scheme (Figure S1). See the text and Table S2 of the Supporting Information for a detailed description.

Hereafter, some characteristics of the proposed method are briefly discussed on an element by element basis. Strontium. Since its introduction8 as “Sr Spec” (for strontium specific) and its early use for geochemical applications,9 the Sr resin has rapidly become the standard method for Sr separation in the field of earth and environmental sciences by virtue of an outstanding advantage in terms of elemental selectivity (in particular, relative to the closely related major element Ca, which produced troublesome tailing in conventional chromatographic methods based on cationexchange) allowing a great degree of miniaturization to be achieved. Further, strontium can be stripped from the resin with very dilute nitric acid, making the method environmentally friendly. The resin is expensive, but this is largely compensated by gains in terms of simplicity and drastic reduction of reagent volumes. Provided memory effects are kept to satisfactory levels through adequate cleaning steps with alternating 6−8 mol L−1 hydrochloric and very dilute nitric acids, the method is rather cost-effective. Lead. In the scope of multi-isotope studies, the Sr resin is particularly attractive because, besides Sr, Pb is also strongly extracted from a very large range of nitric acid strength.7 Although back-extraction proves to be difficult, it has been found that 6−8 mol L−1 HCl offers a convenient eluting medium,10,11 and this has been used to set up methods allowing the concomitant separation of Sr and Pb.12−15 Albeit not offering a great advantage in terms of selectivity over conventional methods based on anion-exchange methods in hydrobromic acid medium,16 the EXC method proves to be more robust in terms of chemical yields because 2 mol L−1 nitric acid is a more efficient solvent than the dilute HBr solutions used to leach Pb from a still largely solid residue. Moreover, using ca. 2 mol L−1 HNO3 to bring the sample in solution permits an easy coupling with other extraction chromatography methods which are mostly based on nitrate media, whereas it would be necessary to evaporate to dryness the HBr solution of the non-Pb fraction and then convert it to the chloride form in preparation to the separation of the rest of the target elements (Sr, Hf, REE) by using conventional procedures, e.g. ref 39. Rubidium. This element cannot be separated in a straightforward way by using conventional cation exchange methods because its partition coefficient overlaps largely with those of several other elements, particularly the closely related major element potassium. Although K cannot cause isobaric interference on the much heavier Rb isotopes, the presence of large amounts of this easily ionizable major element decreases ionization efficiency (when TIMS is used), and is likely to affect the distribution of ions in the plasma (when MC-ICP-MS is used) and make the monitoring of mass bias not as straightforward as it could be with pure Rb solutions. For this reason, inorganic ion exchangers such as zirconium (or titanium) phosphate or ammonium molybdophosphate have been used.17,18 These materials allow Cs and Rb to be separated with a good selectivity over K, Na, and other major elements. However, the reagents able to strip Rb, namely, ammonium nitrate (or chloride) solutions, are far from ideal because they produce a solid residue when evaporated, which needs to be destroyed by tedious treatments with aqua regia before obtaining Rb in a suitable form for mass spectrometric measurements. Moreover, these inorganic exchangers are not mechanically very stable, and they release rather large amounts of matrix constituents (Zr, Ti, or Mo).19 This was not of special

■

RESULTS In order to demonstrate the validity of the separation procedure, five international standard reference materials have been analyzed in replicates for the elements possessing the four most widely used radiogenic isotopes, namely, Sr, Pb, Nd, and Hf. 230Th/232Th measurements were also performed in some cases. Emphasis was put on samples of basaltic composition because they contain large amounts of Fe and Ti, the two major elements which are known to be potentially troublesome during extraction chromatographic separations using the TRU and Ln resins. Further information and the measured 87Sr/86Sr (by TIMS), 208,207,206Pb/204Pb, 143Nd/144Nd, 176Hf/177Hf, and, in part, 230Th/232Th (by MC-ICP-MS) isotope ratios and procedural blanks are listed in Tables S3 and S4 of the Supporting Information, respectively.

■

DISCUSSION The commentaries pertaining to the sample dissolution steps can be found in the Supporting Information. Compared to a previous EXC protocol which did not afford the separation of Hf, nor that of the elements having “parent” isotopes,6 the new procedure involves a different bulk sample starting solution. Specifically, a 2 mol L−1 nitric acid ∼0.07 mol L−1 HF medium was used, instead of pure 1 mol L−1 nitric acid, in order to minimize hydrolytic problems which frequently jeopardize the behavior of HFSE in dilute aqueous solutions. This increased nitric acid strength did not produce a noticeable detrimental effect on the action of ascorbic acid used as a reducing agent of Fe(III) in the forthcoming steps of the separation protocol (i.e., extraction of Hf and the REE on the HFSE column and TRU column, respectively). It also has the additional benefit to significantly improve the uptake of Sr by the crown ether used in the Sr resin because the extraction efficiency steeply increases as a function of nitric acid molarity, specifically, by a factor of ∼4 from 1 mol L−1 to 2 mol L−1 HNO3.7 2414

DOI: 10.1021/acs.analchem.6b04289 Anal. Chem. 2017, 89, 2411−2417

Article

Analytical Chemistry

coefficients for these two elements. However, 141Pr does not produce isobaric interference during isotope ratio measurement using the Nd+ ion. Samarium. The Sm fraction separated in view of isotope dilution analyses occasionally contains some Eu, whose two isotopes at m/z 151 and 153 do not produce isobaric interference. Hafnium. Among many separation methods designed for geochemical applications, those based on the Ln resin are robust because the extraction of Hf is not impaired by the presence of fluorine in limited amount,25 as it happens when other EXC resins (e.g., TEVA, 26 U-TEVA, 27,28 RE,29 TODGA30) are used. When HF dissolution is used, variable amounts of F invariably remain in the forthcoming solutions,31 with drastic effects upon the subsequent behavior of Hf. Although enough F can be removed to allow the TEVA or RE Spec resin to be used for extracting Zr and Hf, this requires tedious treatments of repeated fumings with perchloric acid or using an excess of boric acid. Besides this, in the absence of fluorine, the HFSE are prone to hydrolysis problems. Indeed, it has long been recognized32 that it is far from straightforward to keep the right balance between retaining too much F in solution (with the risk of complexing Hf) and expelling too much F (with the risk of forming insoluble Ti compounds). For these reasons, the good tolerance of the extracting power of HDEHP to the presence of HF is particularly interesting. Indeed, media containing less than 0.2 mol L−1 HF can be used without triggering back-extraction of Hf,25 while Ti and Zr are released.33,34 Therefore, several methods used an HDEHP column to achieve the final separation of Hf from Zr and Ti in HCl-HF media after most major elements have been eliminated during earlier steps of cation-exchange chromatography. Later, a single column of Ln resin was used25 to separate Hf from most other elements in a hydrochloric acid based medium, involving hydrogen peroxide and citric acid as complexing agents. Alternatively, bearing in mind that HDEHP behaves as a cation-exchanger,35 nitric acid was used in this work, in order to make the Ln resin directly compatible with the already established EXC methods developed for isolating other trace elements of radiogenic isotope interest. Citric acid and H2O2 are not required in our scheme. The Hf separated in this way contains minor amounts of Mo and Zr, with the Zr/Hf ratio being typically reduced from ca. 40 in the starting sample to