Investigation on effects of bottle material on water evaporation and isotopic changes in δ2H and δ18O during sample storage

The interpretation of results of water stable-isotope analyses is less reliable if the water storage conditions are not suitable. One very important aspect in this respect is the type and material of the sample bottles used for sampling and storage. The IAEA has a long experience with projects carried out worldwide within its Technical Cooperation (TC) activities and within the Global Network of Isotopes in Precipitation (GNIP). For use within its GNIP activities the IAEA recommends and supplies coloured glass bottles with proper seal and closure for mid- and long-term storage of water samples. However, because of the obvious restrictions of glass (breakable, heavy in transport), and because in most other cases the analysis commences relatively soon after sampling, different types of high-density polyethylene (HDPE) bottles with inner seals and proper caps were found to be adequate. The IAEA has used HDPE bottles for many years now and supplies them to TC counterparts. In the past, and occasionally also to date, water samples are also taken and shipped to IAEA in locally available, various types of containers and bottles. A survey recently revealed around 60 different types of containers (many of which being obviously inadequate), that were used to send samples to the IAEA Isotope Hydrology Laboratory (IHL) in the past decades. Requests were made by local counterparts to investigate some locally and globally available sample bottles on their suitability for sampling, shipping and storing samples for stable isotopes. These considerations lead to a series of tests on sample bottles in the IAEA-IHL. The following tests were performed: 1. The δ18O composition was re-measured in old samples, stored at the IAEA-IHL for 2 - 7 years 2. Three of each kind of old sample bottles were emptied, dried and filled with a well known water, fairly different in isotopic composition from Vienna rainwater and corresponding air moisture (distilled lake surface water, δ2H =-38.7 per mille vs. VSMOW, δ18O = -4.16 per mille vs VSMOW). Similarly, 6 pieces of the IAEA 50 ml glass- and HDPE bottles were prepared. 3. Internationally available HDPE sample bottles, and some sample bottles supplied by counterparts, were prepared as in under 2); results will be available and presented at the symposium. For all bottles the initial tare and net water weight was taken and the content was analysed for stable isotopic composition after 6, 12 and 18 months. Loss of sample weight and differences in the δ2H and δ18O values with respect to the initial isotopic values were recorded. The isotope ratio measurements were carried out on a Finnigan Delta plus mass spectrometer with a self constructed 48-port water equilibration system. In the first preliminary study (re-measurements of δ18O values after 2 - 7 years storage in a range of containers) showed a clear effect of the bottle material (generally minor change in δ18O in glass and HDPE, considerably more positive values in PET and PC) and of the amount of water left in the bottle. Surprisingly there was no clear indication for the importance of an inner cap for proper sample sealing. The second systematic study recorded the weight change during sample storage period (0.5 - 1.5 years). The highest weight loss was noted for bottles with wide necks, followed by PET and PC bottles. No, or minimal weight change was recorded for glass bottles. HDPE bottles had a slightly higher loss of weight than glass bottles. No positive influence of an inner cap for both glass and HDPE bottles on weight loss and isotopic composition was found during undisturbed storage. Generally, the changes in δ2H and δ18O agreed with each other and followed the patterns of the loss of weight and showed changes of up to 4 per mille for δ18O and 10 per mille for δ2H in PET-, LDPE- and PC bottles, indicating higher evaporation/diffusion rates in these materials. In most cases, a clear time trend was observed for weight loss and the enrichment of δ18O a nd δ2H, i.e. gradual changes in less suitable bottles, and constant values in appropriate containers after 6, 12 and 18 months storage. Conclusions: Glass bottles are highly recommendable for water sample storage in hydrology applications. Alternatively, bottles made of HDPE are found to be most suitable for sample transport and midterm storage up to one year. Commercial PET and PC bottles (mainly used for beverages), as well as wide-neck bottles were found to be unsuitable for storage and transport of samples for water stable isotopes analysis