Recovering Paleo-Records from Antarctic Ice-Cores by Coupling a

Oct 23, 2015 - Recently, the increasing interest in the understanding of global climatic changes and on natural processes related to climate yielded t...
0 downloads 5 Views 1MB Size
Article pubs.acs.org/ac

Recovering Paleo-Records from Antarctic Ice-Cores by Coupling a Continuous Melting Device and Fast Ion Chromatography Mirko Severi,* Silvia Becagli, Rita Traversi, and Roberto Udisti University of Florence, Chemistry Department “Ugo Schiff”, Via della Lastruccia, 3, 50019, Sesto Fiorentino, Florence, Italy ABSTRACT: Recently, the increasing interest in the understanding of global climatic changes and on natural processes related to climate yielded the development and improvement of new analytical methods for the analysis of environmental samples. The determination of trace chemical species is a useful tool in paleoclimatology, and the techniques for the analysis of ice cores have evolved during the past few years from laborious measurements on discrete samples to continuous techniques allowing higher temporal resolution, higher sensitivity and, above all, higher throughput. Two fast ion chromatographic (FIC) methods are presented. The first method was able to measure Cl−, NO3− and SO42− in a melterbased continuous flow system separating the three analytes in just 1 min. The second method (called Ultra-FIC) was able to perform a single chromatographic analysis in just 30 s and the resulting sampling resolution was 1.0 cm with a typical melting rate of 4.0 cm min−1. Both methods combine the accuracy, precision, and low detection limits of ion chromatography with the enhanced speed and high depth resolution of continuous melting systems. Both methods have been tested and validated with the analysis of several hundred meters of different ice cores. In particular, the Ultra-FIC method was used to reconstruct the high-resolution SO42− profile of the last 10 000 years for the EDML ice core, allowing the counting of the annual layers, which represents a key point in dating these kind of natural archives.

I

handling of ice cores. The decontamination techniques used for the preparation of discrete samples require the manual removal of the outer layers of ice samples by the use of a stainless steel or ceramic chisel11 or by successive immersions of the samples in baths of ultrapure water.12 Despite these decontamination methods that have proven to be reliable for the preparation of ice core samples, they are highly timeconsuming and this represents a huge drawback in their use when analyzing many hundreds of meters of ice. Starting from the pioneering work by Sigg et al.13 in the early 1990s, the coupling of a melting device with different analytical techniques has been widely used in ice core studies to retrieve the paleorecords of soluble and insoluble impurities,14−16 water isotopes,17 and gases.18 During the last decades, in order to reconstruct high-resolution chemical stratigraphies, several analytical techniques were coupled with an ice-core melting system: continuous flow analysis (CFA),15,19 inductively coupled plasma mass spectrometry (ICPMS),20−22 and ionchromatography (IC).16,23−25 Recently, a laser ablationinductively coupled plasma mass spectrometry system (LAICPMS)26 was set up which did not require sample melting and was able to achieve a submillimeter spatial resolution. All these methods proved to be extremely sensitive for several elements, but, among them, only ion chromatography is capable to

n the last years, the increasing interest in the understanding of global climatic changes and on natural processes related to climate yielded the development and improvement of new analytical systems dedicated to highly resolved measurements of chemical markers on samples from environmental archives. Particular attention was paid to the reconstruction of ice-core chemical stratigraphies from deep cores drilled in polar regions: indeed, snow layers deposited year after year on the Antarctic and Greenland plateau areas trap and preserve several markers able to provide information about past atmospheric composition and climatic variations. The isotopic composition of oxygen and hydrogen in ice is a reliable proxy of paleotemperature;1−3 air bubbles trapped in the ice/firn matrix represent so far the only means to reconstruct the temporal evolution of greenhouse gases for time periods spanning the whole Quaternary;4,5 soluble and insoluble impurities in the ice allow to retrieve unique pale-climatic and paleo-environmental information, including changes in the sea-ice extent, solar activity, marine biogenic activity, past volcanism, terrestrial aridity, and atmospheric transport processes.6−10 The determination of trace chemical species is a useful tool in paleoclimatology, and the techniques for the analysis of ice cores have evolved during the past few years from laborious measurements on discrete samples to continuous techniques allowing higher temporal resolution, higher sensitivity and, above all, higher throughput. The development of such methods is particularly important to minimize the contamination inevitably coming from the drilling, cutting, and © XXXX American Chemical Society

Received: August 3, 2015 Accepted: October 23, 2015

A

DOI: 10.1021/acs.analchem.5b02961 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry measure with a high sensitivity the concentration of Cl−, NO3−, and SO4 2− . Although ICPMS can measure the sulfur concentration, part of the total sulfur measured in ice cores with this technique comes from methanesulfonate (CH3SO3−) which is present at very variable concentrations depending on several factors (drilling site location, period of snow deposition, etc.). For these reasons, in the last decades, ion chromatography has been routinely used for the determination of Cl−, NO3−, and SO42− archived in ice cores. Particular attention was paid in recent years to the development of ion chromatographic methods coupled with flow injection systems able to produce an uncontaminated meltwater stream to be used for automated IC analysis using both isocratic23 and gradient25 elutions. In order to keep up with the high sample preparation speed of a melting system, Cole-Dai and colleagues23 used an IC method able to separate Cl−, NO3−, and SO42− in a 4 min run, using directly the meltwater stream, while Morganti et al.25 used 10 min chromatograms in order to separate the anionic content of a large number of discrete ice samples prepared using a melting device. Here we present the setup of two fast ion chromatographic (FIC) methods whose main feature is the very low analysis time with respect to previously shown methods: the first method was used for the high-resolution measurement of Cl−, NO3−, and SO42− in several ice cores, performing an analysis per minute with a typical sampling resolution of 2 cm; the second method (called Ultra-FIC) was set up for the very high-resolution measurement of SO42− in 30 s with a depth resolution of less than 1.0 cm. Sulfate (SO42−) represents one of the most abundant chemical species in the Earth’s atmosphere and plays a key role in many important atmospheric processes: its aerosols strongly affect the energy budget of the global atmosphere, mainly through the scattering of the incoming solar radiation and through other indirect effects involving clouds.27 The remoteness of the Antarctic continent makes it an ideal place to study the natural variability of the atmospheric SO42− without the contribution of the anthropogenic SO42− sources.28 Natural SO42− sources include sea spray, biogenic activity, biomass burning, stratospheric fallout, and volcanism. Sea salt (ss) SO42− can contribute over 25% to the total SO42− budget only at coastal or at low elevation sites,29−31 while its contribution is much lower in the inner Antarctic areas. Late spring-early summer marine biogenic activity in the southern oceans produces a large variety of sulfur species, which are progressively oxidized to methanesulfonic acid (MSA) and SO42−.28,32−34 Biomass burning as well contributes to the global sulfur budget by emission of carbonyl sulfide that can make its way to the stratosphere where it is converted to SO42− by photolysis.35,36 Moreover, stratospheric SO42−, which comprises volcanic, biogenic, and potentially anthropogenic contributions, provides a constant supply of background SO42− to the troposphere slowly depositing over the poles.37 In particular, volcanic activity represents an episodic but massive source of nonsea-salt(nss)-SO42−. Large volcanic eruptions inject large amounts of ash particles (e.g., tephra) and gases (mainly SO2) into the atmosphere and stratosphere, where oxidation and gas-to-particle conversion of SO2 to sulfuric acid (H2SO4) occurs, thereby modifying their radiative properties and leading to short-term (1−3 years) cooling at the global scale.38,39 Volcanic events are therefore recorded in ice cores mainly as acidic peaks,40 SO42− peaks,41,42 and tephra layers.43,44 While high-resolution measurements of tephra are not feasible, it is possible to carry out such measurements for

SO42− and for the acidic content of the ice. However, conductivity measurements on solid ice (electrical conductivity measurements (ECM) and dielectric profiling (DEP)) are not univocal markers of volcanic eruptions and can be weakened in ice samples with high dust levels due to their increased alkalinity and disturbance by the presence of insoluble particles. Therefore, large SO42− peaks have to be considered as the most reliable marker of volcanic activity recorded in ice cores. Cl− has been widely used for the reconstruction of the sea spray budget in the past atmosphere. In remote areas, the main sources of Cl− are sea spray (particulate) and HCl (gaseous) coming from volcanic emissions and from the exchange reaction of NaCl with H2SO4, HNO3, and many organic acids.45 In Central Antarctica, NO3− is supposed to be mainly deposited as gaseous HNO3 and dry deposition.46 Because of the great variety of sources of NO3− precursors (nitrogen oxides), including lighting, cosmic rays, biomass burning, stratospheric N2O oxidation, and sedimentation of polar stratospheric clouds (PSC), the evaluation of the relative contributions of the different NO3− sources in polar areas is still controversial. Cl−, NO3−, and SO42−, present at μg L−1 level in Antarctic snow and ice, have been measured in polar archives since the beginning of the 1980s by means of ion chromatography (IC).47 The methods here presented have been widely used for the measurement of these anions both in the field (at Dome C−East Antarctica) and in the cold laboratory at the Alfred Wegener Institute (AWI) in Bremerhaven (Germany). These methods have proved to be suitable for fast analysis of Cl−, NO3−, and SO42− at ppb levels in firn and ice core samples thanks to their high sensitivity, high reproducibility, and low detection limits. A few results coming from the high-resolution analysis of three deep ice core drilled in Antarctica are here shown and discussed.



EXPERIMENTAL SECTION The most challenging aspect of the set up of a FIC method is to obtain a sufficiently good separation between the analytes in a very short time; for this reason, particular attention was paid to the optimization of the analytical conditions (eluent strength and flow, column selection, etc.). On the other hand, the absence of relevant amounts of interfering analytes (such as fluoride, nitrite, bromide, phosphate, and some organic ions usually present at subppb levels in polar ice samples) is a basic feature for the success of FIC in the ice cores processing. Reagents. Standard solutions for calibrations were daily prepared in precleaned polyethylene vials by diluting stock standard solutions (1000 mg L−1) purchased from Merck (Darmstadt, Germany) with ultrahigh purity water (UHPH2O) of resistivity >18 MΩ cm (Milli-Q system by Millipore, Billerica, MA). UHP-H2O was also used for the preparation of the eluent and as the regenerant solution for the conductivity suppressor. The eluent was freshly prepared by dilution of two stock solutions of Na2CO3 (1 M) and NaHCO3 (0.1 M); these solutions were prepared by dissolving their respective pro analysis salts purchased from Merck (Darmstadt, Germany). Analysis System. Figure 1 shows the general setup of the ion chromatographic system which is the same for the FIC (A) and Ultra-FIC (B) methods here presented. It is built by coupling two Dionex DX-120 ion chromatographs (IC) with a flow analysis device. The system includes a peristaltic pump (Minipulse 3, Gilson, Middleton, WI) for sample/standard/ B

DOI: 10.1021/acs.analchem.5b02961 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

encoder (Baumer Electric),14 and the typical ice-core cross section was 32 × 32 mm2. Ice Core Samples. The FIC methods here presented have been used for the analysis of three deep ice cores: EPICA (European Project for Ice Coring in Antarctica) Dome C (EDC), EPICA Dronning Maud Land (EDML), and TALDICE. The drilling in Dome C reached the depth of 3260 m (a few meters above the bedrock) in January 2005, retrieving a core which covers a period of more than 800 kyr. All the EDC ice core was analyzed by FIC, including the firn section. EDML drilling operations started in 2000/2001 and were completed in January 2006, reaching liquid water at the bedrock interface at 2774 m depth. High-resolution SO42− measurements have been performed on this core from 113 m to the bottom. The deep drilling at Talos Dome (TALDICE) started during the 2004−2005 summer campaign and successfully reached a depth of 1620 m during the 2007− 2008 season; the FIC analysis on this core was carried out from 73 m to the bottom. The analysis of the EDC core was mainly performed in the field in the warm laboratory at Dome C East Antarctica, while the FIC analysis of the EDML and TALDICE cores were carried out in the laboratories at the Alfred Wegener Institute (AWI) in Bremerhaven (Germany).



RESULTS AND DISCUSSION Method A (Cl−, NO3−, and SO42− in 1 min). The two ion chromatographs shown in Figure 1 work in parallel: while IC 2 (see Figure 1a) is fed with the sample stream coming from the melter through the selection valve (SV), IC 1 is eluting and separating the analytes (and ultrapure water is washing its injection valve (Inj.V 1) and the selection valve (SV); after 30 s the two IC injection valves (Inj.V 1 and Inj.V 2) and the selection valve (SV) are switched, and IC 2 injects into the separation column (Dionex AG4A) the analytes previously trapped in the TAC-2 column, while the preconcentration column in the IC 1 is loaded with the following sample (see Figure 1b). These two steps are cyclically repeated for the time required to melt all the core section (typically 100 or 110 cm). Table 1 shows the position of the three valves at different times from the beginning of the analysis and a brief description of the different phases is also reported for each step. Therefore, the sample coming from the melter is loaded for 30 s in the first IC and for the following 30 s in the second IC. The overall resolution is 30 s, corresponding to a depth resolution of 2.0 cm, at the typical melting rate of 4 cm min−1. Each of the two ICs is calibrated separately and after the optimization of the chromatograms, the signals are processed in order to reconstruct the original continuous stratigraphy. Although the method is not substantially different from a classical IC method, the need for achieving a sufficient separation among Cl−, NO3−,

Figure 1. Scheme of the ion chromatographic system used in this work for the FIC and Ultra-FIC methods. IC1 and IC2 are two Dionex DX120 ion chromatographs; SV is a six-port valve and is called “selection valve” in the main text. The injection valves of the two ICs (Inj.V 1 and Inj.V 2) are also shown with their positions during the two steps of the methods (Inject or Load). The blue line shows the pathway of the ice-core sample during the two different steps, while the red line shows the UHQ pathway.

UHQ loading, run at a flow rate of 1.0 mL min−1 and a twoposition selection valve (six-port, electrically activated Rheodyne valve). Each of the two ion chromatographs was equipped with a Dionex TAC-2 (4 mm) preconcentration column, a Dionex AG4A (4 mm, 50 mm) guard column (used as a separation column), a Dionex ASRS suppressor (run at 100 mA), a Dionex CDM-3 conductivity detector, and an injection valve. Melting Unit. The melter head was made of copper plated with electroless nickel (5 μm) and gold (2 μm) to obtain a chemically inert surface with good thermal conductivity.14 To maintain a constant melting rate of typically 3.5−4.0 cm min−1, the melter head was kept at constant temperature (typically around 20 °C) and a weight was put upon every ice-core sample in order to have an additional pressure to maintain a uniform melting speed also toward the end of a measurement. The melt speed was continuously measured by an optical

Table 1. Position of the Inj.V 1, Inj.V 2, and SV Valves and Their Switching Time through the Two Steps Cyclically Carried out in the FIC (A) and Ultra-FIC (B) Methods time (min) method A method B

Inj.V1

Inj.V2

selection valve (SV)

0.0 0.5 1.0

0.0 0.25 0.5

inject load inject

load inject load

position 2 position 1 position 2

1.5

0.75

load

inject

position 1

phase sample is loaded in IC2 while the loading system of IC1 is rinsed with UHP-water separation of the analytes in IC2 starts while IC1 is loaded with the following sample separation of the analytes in IC1 starts; IC2 is ending the separation of the analytes while being loaded with a new sample New separation of the analytes in IC2 starts; IC1 is ending the separation of the analytes while being loaded with a new sample C

DOI: 10.1021/acs.analchem.5b02961 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry and SO42− peaks in just 1 min required an accurate optimization of the operative conditions (type and length of the separation column; type, concentration, and flow rate of the eluent; very small dead volumes of the chromatographic line, fast washing of the injection system, etc.). Column. In a typical IC method, the separation is performed connecting in series a guard column and a separation column. With this configuration, one can obtain the separation of well resolved peaks in 8−12 min,48,49 but this time is too long for the setup of a FIC method. Therefore, we decided to use a guard column as a very short separating column, in order to highly reduce the IC run time. After testing the commercially available guard columns, we found that Dionex AG4A (4 mm × 50 mm) was the best choice in terms of run time and resolution of the chromatographic separation. As concerning the column capacity, we found that the selected guard column has a capacity (4 μeq L−1) which is by far sufficient also in the case of extremely high sulfate peaks (up to 3000 μg/L−1 or higher). Eluent. The eluent used for the separation was a Na2CO3/ NaHCO3 buffer (2.5 mM and 0.5 mM, respectively) at a typical flow rate of about 2.3 mL min−1. We found that the optimal eluent concentration and flow rate were slightly different from one guard column to the other probably due to different packing, usage time, etc. Therefore, these experimental parameters were adjusted from time to time around the values reported above. The maximum variations observed during a processing campaign are generally within 5% for both the eluent concentration and flow rate. However, once the best conditions were found, separations were stable for several days. Figure 2 shows a series of consecutive chromatograms obtained from the analysis of an ice core section. The ion ratios

Table 2. Figures of Merit of the Two Methods Presented in This Papera method A total analysis time (s) sampling resolution (cm) sample throughput (samples/h) measured analytes detection limit (μg L−1) reproducibility (RSD %) sensitivity (nS μg−1 L) linear range (μg L−1)

method B

60 2 120 Cl− 5.0