Home

Search

Collections

Journals

About

Contact us

My IOPscience

Isotopic composition of water used for triple point of water cells

This article has been downloaded from IOPscience. Please scroll down to see the full text article. 1996 Metrologia 33 265 (http://iopscience.iop.org/0026-1394/33/3/9) View the table of contents for this issue, or go to the journal homepage for more

Download details: IP Address: 200.27.124.194 The article was downloaded on 15/09/2010 at 16:52

Please note that terms and conditions apply.

metrologia

Short Communication Isotopic composition of water used for triple point of water cells
J. V. Nicholas, T. D. Dranseld and D. R. White

Abstract. Measurements of the isotopic composition of water used in the production of triple point of water cells show that state-of-the-art triple point realizations must consider corrections and uncertainties due to variations in the isotopic composition of water. Triple point cell production processes yield cells with triple point temperatures 60 K below that implied by the denition, with a similar sized uncertainty in the value. The use of isotopically impure water for the triple point leads to a fundamental uncertainty in the denition of the kelvin of around 14 K because of the equilibrium isotope fractionation between the ice and liquid water in the triple point cell.

1. Introduction The 18th Meeting of the Comit e Consultatif de Thermom etrie (CCT) recommended further investigation into the triple point of water because improvements in the accuracy of realization will allow full advantage to be taken of the high reproducibility of the International Temperature Scale of 1990 [1]. During the preparation of triple point cells for comparisons with overseas laboratories we realized that isotopic variations in New Zealand water may give rise to differences from triple point cells made with other sources of water. Quite high precision is possible in comparisons of water triple point cells. A high-precision survey of triple point of water cells [2] shows individual cells are reproducible to 10 K and a group of quality cells match to within 50 K. At this level of accuracy variations in the isotopic composition of the water arising from the production of the cell by different manufacturers must also be considered. While isotopic variations are often considered to be small they do have a direct bearing on the denition of the kelvin because the isotopic composition of the water is not dened. Natural water is isotopically impure, consisting of a mixture of at least three species with different thermodynamic properties, and as such does not have a triple point. The guidelines

[3] for realizing the International Temperature Scale of 1990 (ITS-90) recommend that the water has substantially the isotopic composition of ocean water. While an analysis is given of the uncertainty arising from variations in the amount of heavy hydrogen, 2 H, there is no analysis for heavy oxygen, 18 O. No analysis is given for the departure from a pure triple point. Here we report measurements made on the isotopic composition of water samples taken from various stages in the production of triple point cells at our laboratory. This information gives an improved basis on which to make uncertainty estimates for the effects of the isotopic variation on high quality triple point of water realizations. 2. Isotope effects We assume here that Standard Mean Ocean Water (SMOW) meets the requirements of the reference to ocean water given in the ITS-90 guidelines. International standards have been set for reference samples of SMOW, under the auspices of the International Atomic Energy Agency (IAEA). The isotopic composition of SMOW is dened and has 0,158 mmol of 2 H per mole of 1 H and 1,989 mmol of 18 O per mole of 16 O [4]. Other isotopes, 17 O and 3 H, are not important for the present purposes. Note that the isotope abundances imply that 1 in 500 water molecules can be considered heavy. Barber et al. [5] estimate changes in the triple point temperature due to variations in the concentration of heavy hydrogen, 2 H. On the basis of the heavy

J. V. Nicholas, T. D. Dranseld and D. R. White: Measurement Standards Laboratory, New Zealand Institute for Industrial Research, P.O. Box 31-310, Lower Hutt, New Zealand. Metrologia, 1996, 33, 265-267

265

J. V. Nicholas et al.

hydrogen concentration they estimate that the triple point of natural water is about 600 K greater than that of pure light water. Natural variations of 10 % in hydrogen isotopic concentrations are common, thus isotopic determinations are important if uncertainty estimates less than 100 K are claimed. Extreme variations in the hydrogen isotopic composition of natural water amounting to an equivalent of 250 K in the triple point temperatures also occur [3, 6]. Because Barber et al. do not include the effect of heavy oxygen [7], their temperature estimates for isotopic variations are low. Pure light water, 1 H2 16 O, for example, has a triple point about 1,2 mK less than SMOW. Water treatment processes such as distillation and freezing change the isotopic composition. Under equilibrium conditions the 2 H fractionation factor for vapour-liquid is 1,03 at 100 C and 1,09 at 15 C, that is, water near its boiling point is 3 % richer in 2 H than the vapour [8]. It is of interest to note that at the time of the denition of the kelvin in 1960, opinion was divided on the fractionation factor for freezing water, with most research indicating no signicant fractionation [9]. Subsequent research shows that the 2 H equilibrium fractionation factor for water-ice is about 1,02 [9]. In spite of the variation in the source of water and the fractionation effects occurring during the manufacture and preparation of a triple point cell, it is not usual to measure the isotopic composition of the water used in producing the cells [5, 6]. For example, there appear to be no published claims of any triple point cell having water of the isotopic composition of standard ocean water as required by the ITS-90 guidelines. The guidelines do indicate that if there is no modication of the water isotopic composition in production, then cells produced with continental surface water are likely to be 40 K low. These considerations indicate that the variations in triple point temperature due to the isotopic composition arise from three possible causes. (a) Natural isotopic variations occurring in the source of the water used by the laboratory. (b) Variations arising from the treatment of the water during production of the cell. (c) Variations inside the completed cell due to fractionation between phases, mainly between water and ice. 3. Isotopic analysis Five samples of water were collected from various stages in the manufacture and use of triple point cells in our laboratory. They were submitted to the Institute of Geological and Nuclear Sciences, Lower Hutt, for analysis of the 2 H and 18 O concentrations [10]. The results in Table 1 are expressed as differences in the amount of the isotope per mole of the atomic species, with respect to SMOW. Also given is the estimated

change in temperature of the triple point of water from that for SMOW assuming a linear change in triple point value with respect to the isotope concentration. The values used are for 2 H, 3,8 K per mol, and for 18 O, 0,30 K per mol, based on the triple point values of 276,95 K for 2 H2 16 O [4] and 273,46 K for 1 H2 18 O [7], respectively. Note that the temperature dependence of the triple point of water on the isotope concentration has not been accurately measured at the low concentrations typical of triple point cells.
Table 1. Isotopic analysis of ve water samples showing the deviation with respect to SMOW. The equivalent temperature change in the triple point is also given. The uncertainties in the concentration differences are estimated to be 1,0 mol/mol in d18 O and 0,1 mol/mol in d2 H.
Sample d2 H/ d18 O/ no. ( mol/mol) ( mol/mol) 1 2 3 4 5 12 8,5 12 + 5,2 26 5,3 3,7 5,3 + 6,2 13,5 dT/ K 24 17 24 + 25 59 Description Tap water Melted ice from ice machine Water from laboratory still Residue water in degassing ask Water from a triple point cell

The result for sample 1 for tap water indicates that local groundwater differs from SMOW by the equivalent of 24 K. Samples taken at other times indicate that the deviation can vary by as much as 8 %. The result for sample 2 from melted ice shows an increase in the concentration of heavy isotopes that corresponds to a 7 K shift in a triple point cell. The achieved fractionation factor of 1,01, compared with the expected 1,02, indicates that full equilibrium was not established in the ice-making machine. In a triple point cell an equilibrium ice-water interface could give a variation of 14 K in the triple point value between the ice and the water depending on the fraction of water frozen. This variation is a fundamental limit to the realization of the denition of the kelvin because of the use of isotopically impure water. The result for sample 3 from the output of the laboratory still is almost identical to the tap water indicating that the still is driven hard in order to obtain a sufcient throughput of clean water. On the other hand, the result for sample 4 shows that the distillation from the degassing ask to the cell has a marked effect. Sample 4 is the water residue after the production of two triple point cells, and shows a large increase in the heavy isotope concentration. Both the degassing and distillation processes under vacuum result in the build-up of the heavier isotopes. The build-up during the distillation process results in different isotope concentrations in the two triple point cells produced.
Metrologia, 1996, 33, 265-267

266

Isotopic composition of water used for triple point of water cells

Sample 5 is from a water triple point cell, made previously on our equipment, for investigating the pressure at the water triple point. A temperature equivalent shift of 35 K from the starting tap water is obtained and this is about 60 K lower than SMOW. The total overall variation in the waters investigated is 85 K. 4. Conclusions The present results, while only directly applicable to our cells, are indicative of the uncertainties arising from isotopic effects on the triple point of water. They are larger than the estimates of Barber et al. [5] mainly because of the addition of estimates for the effects of heavy oxygen, 18 O. Isotopic effects on the triple point are of the same order of magnitude as other effects such as strain, residual air and pressure. Elimination of the effects by using isotopically pure water is therefore of considerable interest in order to establish if such cells have a more consistent behaviour under a variety of preparation and handling techniques. The denition of the kelvin has a fundamental limit of accuracy because the associated guidelines allow the use of isotopically impure water. Practical realizations are therefore not a true triple point. At present the effect has no practical consequences because comparisons between laboratories are not yet at this level of accuracy. There is a clear need, however, to determine accurate values for the isotope effects in triple point of water cells. Acknowledgements. We thank Mike Stewart and John Hulston of the Institute of Geological and Nuclear Sciences for useful discussions on the isotope effects. We also thank the referee for pointing out [7] which shows that heavy oxygen is an important contributor to the isotope effect.

References
1. BIPM Com. Cons. Thermom etrie, 1993, 18, Recommendation T 2. 2. Furukawa G. T., Bigge W. R., Reproducibility of some triple point of water cells, In Temperature: Its Measurement and Control in Science and Industry, Vol. 5 (Edited by J. F. Schooley), New York, American Institute of Physics, 1982, 291-297. 3. Supplementary Information for the International Temperature Scale of 1990, S` evres, Bureau International des Poids et Mesures, 1990, 29. 4. Kestin J., Sengers J. V., Kamgar-Parsi B., Levelt Sengers J. M. H., Thermophysical properties of uid D2 O, J. Phys. Chem. Ref. Data, 1984, 13, 601-609. 5. Barber C. R., Handley R., Herington E. F. G., The preparation and use of cells for the realization of the triple point of water, Br. J. Appl. Phys., 1953, 5, 41-44. 6. Stewart M. K., Taylor C. B., Environmental isotopes in New Zealand hydrology 1: Introduction: The role of oxygen-18, deuterium, and tritium in hydrology, New Zealand J. Science, 1981, 24, 295-311. 7. Nagano Y., Miyazaki Y., Matsuo T., Suga H., Heat capacities and enthalpy of fusion of heavy oxygen water, J. Phys. Chem., 1993, 97, 6897-6901. 8. Stewart M. K., Friedman I., Deuterium fractionation between aqueous salt solutions and water vapour, J. Geophys. Res., 1975, 80, 3812-3818. 9. ONeil J. R., Hydrogen and oxygen isotope fractionation between ice and water, J. Phys. Chem., 1968, 72, 3683-3684. 10. Hulston J. R., Taylor C. B., Lyon G. L., Stewart M. K., Cox M. A., Environmental isotopes in New Zealand hydrology 2: Standards, measurement techniques, and reporting of measurements for oxygen-18, deuterium, and tritium in water, New Zealand J. Science, 1981, 24, 313-322.

Received on 26 March 1996.

Metrologia, 1996, 33, 265-267

267