- Samples can be submitted in 30 mL wide mouth HDPE bottles (Fisher Scientific Cat # 02-893-5A) or any plastic bottle of your choice – Avoid tin foil lined caps.
- Samples must be filtered using 0.45 μm luer lock syringe filters (Fisher Scientific Cat # 50-109-8711).
- Bottles must be submitted with as little headspace as possible, ideally with no headspace.
- Samples do not need to be refrigerated, although, for long term storage, this is recommended.
- Samples should be sent as soon as possible to our facility where they will be refrigerated upon arrival.
5 mL of the saline water sample is transferred to a distillation vial and hooked up to a vacuum line. A second distillation vial that acts as the collector vessel is also hooked up. Sample is gradually frozen with liquid nitrogen, and both vessels are evacuated to baseline. Liquid nitrogen dewar is removed from the frozen sample and placed into the collector vessel, whilst completely submerged and a valve is opened between the two vessels to allow the transfer to take place. Distillation of the sample takes several hours, while leaving behind salt crystals from the sample side vessel. The process is complete when all of the water from the sample side has sublimated to the collector vessel. The collected sample is then thawed and the vessel is removed from the vacuum line, where the sample is then poured into a clean 2 mL screw cap vial and refrigerated.
Sample analysis is performed via automated injections using a CTC Analytics PAL autosampler into a Thermo Finnigan Thermal Conversion Elemental Analyzer (TCEA) set at 1400 °C and linked via continuous flow to a Thermo Fisher Delta V Isotope Ratio Mass Spectrometer. The autosampler syringe is rinsed twice with HPLC grade methanol, then subjected to a vacuum for five seconds to remove residual methanol. The syringe is then rinsed three times with the sample before injecting 0.5 μL into the TCEA reactor. This is repeated three times for each sample and standard. Sample is combusted to produce elemental hydrogen. Glassy carbon in the reactor produces carbon monoxide as the oxygen phase. These two analytes are carried by a continuous helium stream set at 100 mL/min and separated by a 1.5 m long stainless steel GC column before being admitted to the IRMS for isotope analysis of δ2H and δ18O. Of the three injections performed for each sample and standard, only the third is used for calibration purposes. Memory effect is monitored by calculating the standard deviation of all three peaks.
A set of four secondary water standards are analyzed every ten samples, at the beginning of a run, and at the end. Raw δ2H and δ18O values are plotted versus their known isotopic values to produce a four-point calibration curve. Sample data is calibrated to this curve and reported to the VSMOW scale. Typical linear regression is 0.9996 or better for both the hydrogen and oxygen calibration curves.
Two quality control standards are analyzed after every ten samples and calibrated to the VSMOW scale. Compiled results of EA Consumables Medium Natural Isotopic Water #B2193 and a Greenland Ice Core water show that sample data can be reported as better than ±0.4‰ for δ2H, and better than ±0.16‰ for δ18O.
Email Greg Cane (email@example.com) for discussion.