2. EXPERIMENTAL SETUP AND METHOD
Our experiments were aimed at quantifying the difference inδ 18O between VSMOW and SLAP by producing a surrogate VSMOW by gravimetrical mixing of a SLAP-like water with highly18O enriched water, and compare this surrogate with real VSMOW. To this end, several instruments and waters and procedures were used, which are described in the next section.
2.1 Water portions.
For these experiments, a large batch (20 liter) of Antarctic water was made available to us by the isotope hydrology laboratory of the IAEA in Vienna. Its δ 18O value was even slightly more negative than that of SLAP. Portions of 1 liter of this batch were mixed with demineralized Groningen tap water to mimic SLAP. Such large amounts of water were needed to reach the accuracy goal of ≤ 0.05‰ in the final result for SLAP, because of gravimetric/weighing and sample handling precision limitations. Obviously, such quantities of the real SLAP were out of the question.
The reference waters SLAP and VSMOW (ampoules with 1 mL), for the isotopic measurements, were provided by the IAEA Terrestrial Environment Laboratory in Seibersdorf. In order to avoid additional uncertainty contributions, the IAEA provided us with the real VSMOW and SLAP and not their replacements VSMOW2 and SLAP2. The real VSMOW and SLAP were used for isotopic comparison measurements with the SLAP-like and VSMOW-like waters that were produced in the experiments.
For this study 6 highly 18O enriched water portions were obtained from two manufacturers: three from Cortec (CortecNet, Voisins le Bretonneux, France, specification 18O > 99%) and three from Rotem (Rotem industries Ltd., Arava, Israel, specification 18O > 98%). All six water portions were from different production batches.
Furthermore, one virtually pure 2H2O water (10 times 1 mL ampoules) was obtained from Sigma-Aldrich2H ≥ 99.96% (certificate of analysis specified 99.978%, determined via NMR analysis).
The SLAP-like product of mixing Antarctic water and Groningen demineralized tap water, with the same δ 18O as SLAP, will be referred to as SLAP-replicate-Oxygen henceforth in the manuscript, and in short SLAP-rep-O. δ18OSLAP-rep-O is ≈ -55.5‰ on the VSMOW-SLAP scale. Similarly, VSMOW-rep-O refers to a VSMOW-like water in18O, δ18OVSMOW-rep-O is ≈ 0‰. The other produced replicates are VSMOW-rep-D (δ2HVSMOW-rep-D ≈ 0‰) and VSMOW-rep-OD (δ 2HVSMOW-rep-OD ≈ 0‰,δ 18OVSMOW-rep-OD ≈ 0‰). So, the last replicate matches VSMOW in both water isotopes.
2.2 Instruments.
Accurate determination of the 18O concentration of the highly enriched water was key to our efforts: to achieve an accuracy of ≤ 0.05‰ in the δ 18O value for SLAP, the18O concentration of the highly enriched H218O water had to be determined at ≤ ± 0.1%. We were able to reach this precision and accuracy by performing detailed mass scans using a quadrupole mass spectrometer (QMS) equipped with an electron ionization (EI) ion source (Extorr XT100, Extorr Inc., USA), in combination with a bespoke spectral fitting program. The measurements were carried out at an electron energy of 70 eV. For the uncertainty in our signal we use the standard deviation of the instrument’s background signal-to-noise at m/z 5, as no peak is expected at m/z 5, which was around 2 x 10-9 Pa. The total integrated signal of m/z 1 to 41 was approximately 2 x 10-4 Pa. The base peak signal at m/z 20, [H218O]+, was almost 1.3 x 10-4 Pa.
All water samples were analyzed using a LGR Liquid Water Isotope Analyzer (LGR LWIA 912-0050), which is an off-axis integrated cavity output spectrometer, to determine the triple-stable isotope composition:δ 18O, δ 17O andδ 2H. Typically sample measurements are bracketed with local references as well as international references, details of which has been presented later in the manuscript.
The portion of H218O water (approximately 125 mg) was weighed on a Sartorius BP210 D (210 g, readability 0.01 mg) analytical balance. The SLAP-like water used for mixing (approximately 1000 g) was weighed on a precision balance from Kern 572 (4210 g, readability 0.01 g).
To verify the NMR specification of the supplier of2H2O (and check in general that sample handling of such highly enriched waters had a negligible influence on the abundances), the 1H abundance of2H2O was analyzed using a NMR (Bruker Avance NEO 600 MHz).
2.3 Procedure
2.3.1 Approach 1
All the steps taken to prepare the various water-replicate samples leading to the precise determination ofδ 18OSLAP with respect toδ 18OVSMOW is illustrated in Figure 1. The flow diagram illustrates the mixing steps from Antarctic water via a SLAP-replicate to the two VSMOW-replicates created by adding well-characterized H218O (left-hand side), and with an extra step in which also the2H-side is modified (right-hand side). The most critical part of the process entails the characterization of the highly enriched 18O-water that is added. Important other, but more standard determinations, are the initial creation of the SLAP-rep-O water, as well as several additional determinations (such as the determination of the 17O and 2H content of the 18O-water, and the optical measurements of the isotopic differences between the created SLAP-rep-O and SLAP, and between the VSMOW-rep-O (or VSMOW-rep-OD) and VSMOW. The steps indicated on the right-hand side of Figure 1 is further described in 2.3.2.
<Figure 1>