An ultra-clean technique for accurately analysing Pb isotopes and heavy metals at high spatial resolution in ice cores with sub-pg g⁻¹ Pb concentrations

Abstract

Measurements of Pb isotope ratios in ice containing sub-pg g⁻¹ concentrations are easily compromised by contamination, particularly where limited sample is available. Improved techniques are essential if Antarctic ice cores are to be analysed with sufficient spatial resolution to reveal seasonal variations due to climate. This was achieved here by using stainless steel chisels and saws and strict protocols in an ultra-clean cold room to decontaminate and section ice cores. Artificial ice cores, prepared from high purity water were used to develop and refine the procedures and quantify blanks. Ba and In, two other important elements present at pg g⁻¹ and fg g⁻¹ concentrations in Polar ice, were also measured. The final blank amounted to 0.2 ± 0.2 pg of Pb with ²⁰⁶Pb/²⁰⁷Pb and ²⁰⁸Pb/²⁰⁷Pb ratios of 1.16 ± 0.12 and 2.35 ± 0.16, respectively, 1.5 ± 0.4 pg of Ba and 0.6 ± 2.0 fg of In, most of which probably originates from abrasion of the steel saws by the ice. The procedure was demonstrated on a Holocene Antarctic ice core section and was shown to contribute blanks of only ∼5%, ∼14% and ∼0.8% to monthly resolved samples with respective Pb, Ba and In concentrations of 0.12 pg g⁻¹, 0.3 pg g⁻¹ and 2.3 fg g⁻¹. Uncertainties in the Pb isotopic ratio measurements were degraded by only ∼0.2%.

Introduction

Trace elements in ice and firn cores recovered from glaciers provide a record of atmospheric composition, reflecting changing climatic conditions and anthropogenic pollution levels over time (e.g. [1], [2], [3], [4] and references therein). For heavy metals, serious measurement inaccuracies can arise unless adequate precautions are taken to avoid contamination during sampling, transport, preparation and measurement. The problem is exacerbated for elements such as Pb where pollution levels can be high and natural concentrations are at or below the pg g⁻¹ level.

Patterson and co-workers [1] overcame this problem by collecting metre-sized blocks of Greenland ice from which they mechanically scraped the polluted surface layers before analysis. Ng and Patterson [5] applied this approach to ice cores drilled in Greenland and Antarctica. Using acid-cleaned stainless steel chisels they removed consecutive layers of ice from these cylindrical samples, keeping them frozen by manipulating them inside plastic trays lined with dry ice. By analysing the Pb concentration in each layer and the inner core, a profile was constructed allowing the success of the decontamination procedure to be assessed. A concentration plateau from at least the final veneer to the inner core demonstrated that the interior ice was not significantly contaminated. During the 1980s this approach was used in a series of pioneering studies to analyse metals in Antarctic ice from Adelie Land, Dome C, and Vostok [6], [7], [8].

In the 1990s Boutron and co-workers developed the technique further by constructing a polyethylene (PE) “lathe” to support the core above the plastic working surface and performing the decontamination inside a laminar flow hood sited in a regular cold room [9]. This reduced hand contact with the core and made chiselling easier. Generally, 20 cm or longer ice core sections of ∼5 cm diameter were used, thereby ensuring that the relative amounts of contaminant metals added during the process were negligible.

When much smaller samples of ice are available for analysis, such as in high spatial resolution studies, contamination becomes a much more significant issue. Measurements of Pb isotopic ratios are even more sensitive than concentrations to this contamination. To achieve accuracies of better than 0.5% in the ²⁰⁶Pb/²⁰⁷Pb isotopic ratio, the blank contribution needs to be less than ∼10%. Since some Antarctic ice contains concentrations as low as 0.1 pg g⁻¹[10], an extremely low decontamination blank is essential for such work. To date, the only attempt to analyse the isotopic composition of Pb in Antarctic firn at seasonal resolution has been made by Planchon et al. [10] who chiselled layers from ∼30 cm square snow blocks and investigated the seasonal variations. However, Pb concentrations in their samples were larger than 1.4 pg g⁻¹.

Another approach to sampling ice cores at high resolution has been to use melter systems which collect a meltwater sample from the inner part of a core sample while deflecting the contaminated outer part to waste. This method offers high throughput, but acceptably low blank contributions are yet to be demonstrated at the sub-pg g⁻¹ level necessary for reliable Pb isotopic analyses. The “melting” approach was first described by Sigg et al. and used to measure ionic species in ice cores recovered from Greenland and Antarctica [11], [12]. More recently this method has been used to analyse the concentration of Pb in Greenland ice [13], [14].

Here we describe procedures that allow ice cores to be sampled with low contamination and high spatial resolution permitting Pb isotopic ratios and concentrations of Pb, Ba and In to be measured with good accuracy and precision. Ba and In are important indicators of dust levels [15] and volcanic emission [16]. The traditional mechanical decontamination concept has been extended by the addition of a stainless steel circular saw for sectioning the ice and the apparatus housed inside a clean room operating below −10 °C. Very pure artificially prepared ice cores were used to refine the procedures and establish the blank, and an Antarctic ice core with Pb concentrations that reached less than 0.2 pg g⁻¹ was analysed to demonstrate the technique.