Evidence of a Systematic Deviation of the Isotopic Composition of

This variation corresponds to a range of triple-point temperatures (Ttp) of more than 200 μK, much larger than the state-of-the-art uncertainty in th...
0 downloads 0 Views 144KB Size
Anal. Chem. 2005, 77, 5076-5080

Evidence of a Systematic Deviation of the Isotopic Composition of Neon from Commercial Sources Compared with Its Isotopic Composition in Air Franco Pavese,* Bernd Fellmuth,† David I. Head,‡ Yves Hermier,§ Kenneth D. Hill,| and Staf Valkiers⊥

CNR, Istituto di Metrologia “G.Colonnetti” (IMGC), strada delle Cacce 73, 10135 Torino, Italy

Results are reported of a study concerning the variation in isotopic composition of a limited number of neon samples of commercial origin and the resulting influence on the temperature of the triple point of this element. All seven neon samples investigated were found to contain more 22Ne than neon in air, and the amount fraction of 22Ne varied by as much as 0.2% from sample to sample. This variation corresponds to a range of triple-point temperatures (Ttp) of more than 200 µK, much larger than the state-of-the-art uncertainty in the realization of this phase transition for metrological purposes. Deviations in the amount fractions of 21Ne were irrelevant, as far as their effect on Ttp is concerned, though they may have relevance to other isotope studies. Ratios of amounts of neon isotopes at IRMM-Geel were obtained using the same measurement procedures, and instrumentation developed in the framework of the redetermination of the Avogadro constant and all significant sources of uncertainty were taken into account. The repeatability of the ion current ratio measurements on individual samples was 5 × 10-5 relative. All uncertainty statements are made following the ISO/BIPM Guide to the Expression of Uncertainty in Measurements. Whereas these results proved unexpected, a more comprehensive study will follow incorporating a much wider range of samples of commercial origin. The accepted international authority on variations in the fractions of amounts of isotopes of terrestrial materials is the IUPAC’s Commission on Atomic Weights and Isotopic Abundances (CAWIA). During their meeting in 1985, the Working Group on Natural Isotopic Variation (now the Sub-Committee on Natural Isotopic Fractionation) was formed to investigate the impact of the variation in the isotopic abundances of compounds of an element upon its atomic weight and isotopic composition. * To whom correspondence should be addressed. E-mail: [email protected]. † Present address: Physikalisch-Technische Bundesanstalt (PTB), Abbestrasse 2-12, 10587 Berlin, Germany. ‡ Present address: National Physical Laboratory (NPL), Hampton Rd., Teddington, Middlesex, TW11 0LW England. § Present address: Institut National de Me´trologie (INM/CNAM), 292 rue St Martin, 75003, Paris, France. | Present address: National Research Council (NRC), K1A 0R6 Ottawa, Canada. ⊥ Present address: Institute for Reference Materials and Measurements (IRMM), JRC-European Commission, B-2440 Geel, Belgium.

5076 Analytical Chemistry, Vol. 77, No. 15, August 1, 2005

Table 1. Values of IUPAC Neon Amount of Substance Fractions x(iNe) ) n(iNe)/n(∑jNe), with i ) 20-22 and Where the Summation Goes for j ) 20-22, for Samples of Terrestrial Origin 20Ne

genesis

primordial

recommended from best data range on earth

0.904 8 (3) 0.904 838 (90) 0.905 140.884 7

ratio to 20Ne

21Ne

22Ne

primordial + nucleogenic 18O(R,n) 0.002 7 (1) 0.002 696 (5) 0.017 10.002 66a 0.002 98

primordial + nucleogenic 25Mg(n,R) 0.092 5 (3) 0.092 465 (90) 0.099 60.092 0 0.102 19

a Means that a spread (from ranges indicated) on the isotope amount ratio n(21Ne)/n(20Ne) ) 0.0114 ( 0.0086 would be possible.

At regular intervals, the group publishes recommended values of the “natural isotopic composition” of the elements. In the most recent full edition of their report, published in 2003,1 the fractions of amount of neon isotopes x(iNe) ) n(iNe)/n(∑jNe) (with i ) 20-22 and where the summation goes for j ) 20-22), reported in Table 1, were published (see the reference for the meaning of “best”). The CAWIA recommended values for the isotopic composition of neon (dating back to 1984) are based on isotopic determinations of neon in air together with those of commercially purified neon gas derived from air. On the other hand, neon distilled from natural gas wells is reported to have different fractional amounts of neon isotopes (Best Measurement from a terrestrial single source)2(see above for the meaning of the notation)

x(20Ne) ) n(20Ne)/n(

∑ Ne)0.904 838 (90) j

x(21Ne) ) n(21Ne)/n(ΣjNe)0.002 696 (5) with n(21Ne)/n(20Ne) being 0.002 98, and

x(22Ne) ) n(22Ne)/n(ΣjNe)0.092 465 (90) with n(22Ne)/n(20Ne) being 0.102 19. (1) de Laeter, J. R.; Bo ¨hlke, J. K.; De Bie`vre, P.; Hidaka, H.; Peiser, H. S.; Rosman, K. J. R.; Taylor, P. D. P.: Atomic weights of the elements. Review 2000 (IUPAC Technical Report) Pure Appl. Chem. 2003, 75, 683. 10.1021/ac048083f CCC: $30.25

© 2005 American Chemical Society Published on Web 06/30/2005

Figure 1. Change in the isotopic ratios of neon isotopes from terrestrial sources.2 The inset shows a narrow region close to the isotopic composition of neon in air with the range of the values measured in this work.

Nucleogenic Ne and 4He are correlated in groundwater while crustal natural mantle gases are depleted in the isotope 22Ne. Crustal gases, however, are enriched in 21Ne and 22Ne compared to neon in air. Neon derived from natural gas (helium separation plants) will be isotopically divergent and less uniform in composition compared to neon derived from air.1 In general, the amount of substance ratio n(22Ne)/n(20Ne) can increase to ∼0.115, as shown in Figure 1. The neon isotopic composition affects some of the properties of neonsnotably its vapor pressuresresulting in compositiondependent P-T relations. Regarding the temperatures of the triple points of pure neon isotopes, the following values were obtained:3 Ttp (20Ne) - Ttp (natNe) ) (-13.0 ( 1.5) mK and Ttp (22Ne) - Ttp (natNe) ) (+134.0 ( 1.5) mK, where natNe composition refers to neon in air. Consequently, there is a variation of ∆Ttp ≈ 1.47 mK for each mole percent of 22Ne present. This would lead to a discrepancy as large as ∼0.9 mK between the Ttp measured for neon extracted from natural gas and neon obtained from air. State-of-the-art thermometry seeks to limit the uncertainty arising from variations in the isotopic composition of neon samples to ∼20 µK. Consequently, the 22Ne content is required with a total combined uncertainty equal to ∼1.4 × 10-4. Additionally, the 21Ne variability should also be ascertained. Therefore, a research program was set up to investigate the isotopic variability of commercial neon samples used to establish (2) Bottomley, D. J.; Ross, J. D.; Clarke, W. B. Geochim. Cosmochim. Acta 1984, 48, 1973. Emerson, D. E.; Stroud, L.; Meyer, T. O. Geochim. Cosmochim. Acta 1966, 30, 847. (3) Furukawa, G. T. Metrologia 1972, 8, 11.

national temperature standards, while taking advantage of a European Project addressing also isotopic issues.4 The results obtained to date were unexpected. Therefore, a Euromet Project was set up in 2004, to carry out a comprehensive study on a much larger set of samples of worldwide commercial origin.5 Literature data of acceptable accuracy describes the influence of variations in neon isotopic composition on Ttp. The main challenge was in finding a laboratory able to measure the isotopic composition of the (commercial) neon samples used in these thermometric measurements. This is a common difficulty, at least for gases, as was found, though to a lesser extent, for the determination of the ratio D/H of amount of protium isotopes.6 For neon, fewer institutes are able to provide such analysis, and most of those who do so offer a relative uncertainty in the order of a few percent, inadequate by more than 1 order of magnitude for this work. The neon ion current ratios were measured at IRMM-Geel (B) on the “Avogadro II amount of substance comparator”, a gas mass spectrometer developed for the redetermination of the Avogadro constant.7 Through a calibration process, the observed quantities (ratios of electrical currents) are converted into ratios of amount of isotopes (i.e., the functional relationship is established experimentally, including the uncertainty). In this specific field of isotope measurements, calibration is often synonymous with the use of synthetic isotopic mixtures and is defined as such in the Technical Document of the Commission of Atomic Weights and Isotope Abundances. The authors here have used such mixtures (for elements other than neon) to obtain calibrated measurements for Si,8 Xe,9 Kr,10 and S.11 These experiments have shown that the methods and instrumentation used (and applied to the redetermination of the Avogadro constant) are, to a high degree, dependent solely on the physical processes occurring in the mass spectrometer (mainly the effusion of the gas from the molecular inlet system). Moreover, procedures have been developed to assess whether other gases so introduced behave (during the experiment!) in the same manner. Proper calibration can be achieved in this way if appropriate uncertainties are assessed for each step of the measurement process. EXPERIMENTAL DETERMINATIONS Six neon bottles from different commercial suppliers (all from various batches) were used for the realization of the temperature standards at IMGC, INM-CNAM, NPL, PTB, and, outside the EU, NRC. Prior to the isotope abundance ratio measurements, all high(4) Pavese, F.; Fellmuth, B.; Head, D.; Hermier, Y.; Peruzzi, A.; Szmyrka Grzebyk, A.; Zanin, L. In Temperature, Its Measurement and Control in Science and Industry; Ripple, D., Ed.; AIP: NewYork, 2002; Vol. 8, p 161. (5) Euromet Project 770. Determination of an accurate Ttp vs x(isotopic) relationship for neon. Coordinator: F. Pavese, IMGC-CNR, Torino, Italy. www.euromet.org. (6) Fellmuth, B.; Wolber, L.; Hermier, Y.; Pavese, F.; Steur, P. P. M.; Peroni, I.; Szmyrka-Grzebyk, A.; Lipinski, L.; Tew, W. L.; Nakano, T.; Sakurai, H.; Tamura, O.; Head, D.; Hill, K. D.; Steele, A. G. Metrologia 2005, 42, 171. (7) De Bie`vre, P.; Lenaers, G.; Murphy, T. J.; Peiser, H. S.; Valkiers, S.; Metrologia 1995, 32, 103. (8) De Bie`vre, P.; Valkiers, S. Metrologia 1994, 31, 245. (9) Valkiers, S.; Aregbe, A.; Mayer, K.; De Bie`vre, P. A. Int. J. Mass Spectrosc. Ion Processes 1998, 173, 55. (10) Aregbe, Y.; Valkiers, S.; Poths, J.; Kipphardt, H.; Taylor, P. A. Int. J. Mass Spectrosc. 2001, 206, 129. (11) Valkiers, S.; Kipphardt, H.; Ding, T.; Damen, R.; De Bievre, P.; Taylor, P. D. P. Int. J. Mass Spectrosc. 1999, 193, 1.

Analytical Chemistry, Vol. 77, No. 15, August 1, 2005

5077

Table 2. Chemical Purity of the Seven Neon Samples Used in This Study, Measured by Gas Mass Spectrometry

Table 3. Observed Ion Current Ratios JI/j ) I(iNe+)/ I(jNe+) with j )20 and I ) 21 or 22 (1s, n ) 5), Measured on the 7 Neon Samples Used in This Worka

sample ID

neon source

supplying lab

chemical purity, %

sample ID

measurement sequence

J21/20 ) I(21Ne+)/ I(20Ne+)

J22/20 ) I(22Ne+)/ I20Ne+)

1 2 3 4 5 6 7

Messer Air Liquide Air Products Messer Messer Air Products Messer, used as working standard

IMGC INM NPL PTB PTB NRC IRMM

>99.999 >99.999 >99.999 >99.999 >99.999 >99.999 >99.995

7 5 1 7 3 4 2 6 7

1 2 3 4 5 6 7 8 9

0.002 911 41 (12) 0.002 936 64 (11) 0.002 930 81 (08) 0.002 914 32 (10) 0.002 930 81 (07) 0.002 937 14 (10) 0.002 942 14 (06) 0.002 941 20 (09) 0.002 913 35 (10)

0.102 886 4 (23) 0.103 842 3 (15) 0.103 275 2 (18) 0.102 900 3 (22) 0.104 554 0 (24) 0.103 862 1 (18) 0.104 922 0 (19) 0.104 794 1 (25) 0.102 896 3 (13)

purity gas samples were checked at IRMM for impurities, particularly for interferences in the mass region m/e 20-22, where the iNe+ ions are measured (see Table 2). The influence of interfering ions in this mass region was estimated to be less than 2 × 10-5 relative: in all samples, no traces of air components were detected, so details on chemical impurities are not reported in this paper. The realization of the triple point using gas from each of these samples was performed in the above laboratories for the purpose of realizing top-level temperature standards (ITS-90). These independent realizations were part of a recent international comparison of standards organized as Key Comparison CCT-K2 12 by the Consultative Committee for Thermometry, the international body charged with overseeing thermometry under the auspices of the Metre Convention. Therefore, the Ttp values for all of these samples were compared with 1997 state-of-the-art uncertainty. (The isotopic composition of these samples was unknown at the time of the comparison, but expected to comply with the temperature scale (ITS-90) requirement of “natural composition”. See refs 4 and 6 for a more comprehensive description of state-of-the-art techniques in temperature metrology.) An additional gas sample (an IRMM working standard) for which no temperature data are available was added to the isotope ratio measurement. The neon ion current ratio measurements were measured with the Avogadro II amount of substance comparator 7 using the procedure developed for the Avogadro project. The ion currents were measured for the most abundant ions iNe+ at mass positions 20, 21, and 22. All neon bottles were measured at least 5 times in individual runs, always with fresh gas. The observed ion current ratios Ji/j ) I(iNe+)/I(jNe+) with j ) 20 and i ) 21 or 22; a first approximation to the ratios of amount of neon isotopes n(iNe)/ n(jNe), are summarized in Table 3. The uncertainties indicated in parentheses (typically 4 × 10-5 relative uncertainty for J21/20 and 2 × 10-5 relative uncertainty for J22/20) are based on repeated measurements of an individual sample. To convert the ion current ratios Ji/j to ratios of amounts of isotopes n(iNe)/n(jNe) (given in Table 4), a conversion factor is applied (eq 1). The conversion factor must include all significant influence quantities and parameters with their associated uncertainties; otherwise the results reported would not be the correct ratios.

I(iNe + ) ) K ) KresJi/j res n(jNe) I(jNe + )

n(iNe)

5078

Analytical Chemistry, Vol. 77, No. 15, August 1, 2005

(1)

a The numbers in parentheses are standard uncertainties u , They c apply to the last digits of the value.

Table 4. Calibrated Ratios of the Amount of Isotopes n(iNe)/n(20Ne)a sample ID

measurement sequence

n(21Ne)/n(20Ne)

n(22Ne)/n(20Ne)

7 5 1 7 3 4 2 6 7

1 2 3 4 5 6 7 8 9

0.002 911 4 (15) 0.002 936 6 (15) 0.002 930 8 (15) 0.002 914 3 (15) 0.002 930 8 (15) 0.002 937 1 (15) 0.002 942 1 (15) 0.002 941 2 (15) 0.002 913 3 (15)

0.102 886 (51) 0.103 842 (51) 0.103 275 (51) 0.102 900 (51) 0.104 554 (51) 0.103 862 (51) 0.104 922 (51) 0.104 794 (51) 0.102 896 (51)

a The uncertainties are standard uncertainty estimated following ISO/BIPM Guide to the Expression of Uncertainty in Measurement (GUM), for a coverage factor k ) 1 (they are given in parentheses and apply to the last two digits). for sample identification, see Table 2.

Usually, mixtures of pure and highly enriched isotopes (neon in this case) with known isotopic compositions are required to determine an “overall” conversion factor with small uncertainties. However, the highly enriched isotopes needed to prepare synthetic isotope mixtures in sufficient amounts with adequately low uncertainties are very expensive, and their preparation is extremely time-consuming. A different approach is to investigate the measurement process until all steps are well understood (as was the case for IRMM’s xenon and krypton work), including the instrumentation, measurement procedures, and data algorithms. This approach cannot be used for every mass spectrometer technique requiring low uncertainties, but only for those where the processes are readily understood. Gas isotope mass spectrometry, where gases effuse from a batch volume via a molecular leak into an electron impact ion source, is such a measurement process and has been studied in extreme detail at the IRMM until all steps are well understood within the stated uncertainty. Measurements on synthetically prepared isotope mixtures at IRMM confirm that the conversion factor Kres for residual systematic effects of an unknown nature approaches unity to within 2 × 10-4 (often a factor of 5 or more less) with a typical uncertainty of 5 × 10-4 (or less). (12) Steele, A. G.; Fellmuth, B.; Head, D.; Hermier, Y.; Kang, K. H.; Steur P. P. M.; Tew, W. Metrologia 2002, 39, 551.

Table 5. Deviations of the Values in Table 2 from “Natural Isotopic Composition” (IUPAC, 2003)1 and the Corresponding Effect on ∆Ttp sample no.

n(21Ne)/ n(20Ne)

∆Ttp/µK a

u/%

n(22Ne)/ n(20Ne)

∆Ttp/ µK b

u/%

7 5 1 7 3 4 2 6 7

-0.000 069 -0.000 043 -0.000 049 -0.000 066 -0.000 049 -0.000 043 -0.000 038 -0.000 039 -0.000 067

-5 -3 -3 -5 -3 -3 -3 -3 -5

4,4 7,1 6,2 4,6 6,2 7,1 8,0 7,8 4,5

0.000 700 0.001 610 0.001 090 0.000 710 0.002 360 0.001 670 0.002 730 0.002 600 0.000 710

103 237 160 104 347 245 401 382 104

7.1 3.1 4.6 7.0 2.1 3.0 1.8 1.9 7.0

a

0.07 µK/ppm21Ne.15

b

0.147 µK/ppm22Ne.3

Since this conversion factor has even less dependence for the neon under investigation here, the conversion factor for small residual systematic effects in this work was estimated to be Kres ) 1.0000 with an uncertainty (type B evaluation) of 5 × 10-4 relative. This approach was chosen as the basis for the statement that the values for the isotopic compositions reported in this work have been calibrated. The authors do not claim to have performed a calibration by measuring synthetic isotope mixtures of the same element, and we are aware that using synthetic isotope mixtures could result in smaller combined uncertainties for the ratios of amounts of isotopes. An isotopic composition can be reported in various ways. In this work, the neon isotope amount ratios are reported as ratios with respect to 20Ne, following the ISO/BIPM Guide to the Expression of Uncertainty in Measurement (GUM). The values for the different neon bottles are compared in Table 5 to the IUPAC (2003) recognized “natural neon isotopic composition”. Often, clear uncertainty statements are missing or incomplete for such assessments, making comparison to published values difficult. RESULTS When the results from Table 4 are compared to the IUPAC’s recommended ratios of the amounts of isotopes for terrestrial neon with n(21Ne)/n(20Ne) ) 0.002 98 ( 0.000 11 and n(22Ne)/n(20Ne) ) 0.102 23 ( 0.000 33, it can be seen that the ratios of amounts of isotopes n(21Ne)/n(20Ne) in Table 4 are well within the recommended range, while all n(22Ne)/n(20Ne) ratios are clearly higher than the IUPAC valuesby up to 2%. The three IRMM measurements (on high-purity neon from Messer-D) that were randomly measured between the other neon gases are in good agreement with each other, a clear indication that the mass spectrometer was stable throughout the entire series of measurements. The ratios n(22Ne)/n(20Ne) and n(21Ne)/ n(20Ne) obtained for this sample are within the IUPAC recommended ranges at a coverage factor of k ) 2. The measurements performed at IRMM in 199414 using Linde gas were also consistent with the IUPAC recommended ranges, but not with the present data for the other samples. (13) In BIPM Key Comparison Data Base. URL: http://kcdb.bipm.org. (14) Valkiers, S.; Schaefer, F.; De Bie`vre, P. Sep. Technol. 1994, 965.

Figure 2. Correlation between the temperature differences calculated using Furukawa’s results3 (abscissas) arising from the deviations of the laboratory sample n(22Ne)/n(20Ne) composition from IUPAC composition1 (see Table 2 for the correspondence of the samples to the laboratories and Table 5 for the values), and the temperature differences measured between these samples in the intercomparison CCT-K212 (ordinates: positive values are for hotter realizations and the zero value of the ordinates corresponds to the comparison reference value). Circles: run 1. Squares: run 2, horizontally displaced for graphical convenience. Uncertainty bars are for k ) 1.

In fact, the origins of the neon samples are unknown. The scatter of the measured ratios of amounts of isotopes n(22Ne)/ n(20Ne) shows a relative variation of ∼0.2% for the different bottles. This corresponds to a maximum spread in the thermometric data (excluding the IRMM gas, for which no thermometric measurements were made) of ∼240 µK. On the contrary, the deviation in the ratio n(21Ne)/n(20Ne) and its dispersion is inconsequential for temperature metrology, though it may be relevant to other isotopic studies. Figure 2 shows the correlation for the known influence of n(22Ne)/n(20Ne) on Ttp 3 versus the temperature differences from CCT-K2.12 The uncertainties of the intercomparison differences (experimental data) are comparable to the effect of the isotopic composition; however, the figure shows a reasonable correlation. CONCLUSIONS Measurements made in this work initially undertaken for thermometric purposes indicate that the isotopic composition of commercially available neon samples deviates significantly from IUPAC’s “recommended best natural neon isotopic composition” (2003). (See also refs 16 and 17.) The ratios of the amount of isotopes of the samples show an isotopic composition more enriched in 22Ne than expected from fractionation occurring during industrial purification of gas derived from air. The scatter in the isotopic composition of 22Ne is even larger than indicated by IUPAC. Some relevant thermal properties are affected to such an extent that, for critical applications, a calibration against primary isotope gas standards is required to reduce the total combined uncertainty associated with the potential variability in isotopic (15) McConville G. T. J. Low Temp. Phys. 1974, 15, 647; US AEC Mound Laboratory Report MLM-2088 TID-4500 UC-22 1974. (16) Eberhardt, P.; Eugster, O.; Marti, K. Z. Naturforsch., A: Phys. Sci. 1965, 20, 623. (17) Walton, J. R.; Cameron, A. E. Z. Naturforsch., A: Phys. Sci. 1966, 21, 115.

Analytical Chemistry, Vol. 77, No. 15, August 1, 2005

5079

composition. Unfortunately, few institutes are able to provide analyses with the demanding uncertainty budgets required of those measurements because the preparation of primary standards (synthetic isotopic mixtures of highly pure and highly enriched isotopes) is very expensive and enormously time-consuming. Further studies are required to fully characterize the dispersion in the isotopic composition of 22Ne in most of the neon gas sources currently available. In addition, more effort is needed to reduce the uncertainties in the ratio measurements of the amount of neon isotopes, a prerequisite for accurate determination of the thermal properties of the gas. Some properties relevant to thermal studies, such as thermometric standards, are affected to such an extent that, for critical

5080

Analytical Chemistry, Vol. 77, No. 15, August 1, 2005

applications, certification of the gas is necessary to reduce the uncertainty arising from variations in isotopic composition. Further thermometric measurements, at lower uncertainties, are also necessary to validate, with sufficient accuracy, the corrections for the isotopic influence on thermal properties.13 ACKNOWLEDGMENT This work was partially funded under the EU Project MULTICELLS, Contract G6RD-CT-1999-00114. Received for review December 28, 2004. Accepted June 15, 2005. AC048083F