Anal. Chem. 2008, 80, 8688–8693
Method for Purification of Krypton from Environmental Samples for Analysis of Radiokrypton Isotopes Reika Yokochi,* Linnea J. Heraty, and Neil C. Sturchio Department of Earth and Environmental Sciences, University of Illinois at Chicago, 845 West Taylor Street, Chicago, MC186, SES2440, Illinois Radiokrypton isotopes (81Kr and 85Kr) are ideal tracers and chronometers of various environmental processes. Atom trap trace analysis (ATTA) is capable of determining the ultralow isotopic abundances of radiokryptons (98% purity from 5-125 L STP (standard temperature and pressure) of bulk gas with >90% yield within several hours. This system is generally useful for separation of microliter amounts of unreactive trace volatile compounds from large-volume gas samples. Krypton has two extremely low-abundance radioactive isotopes. These are 81Kr (81Kr/Kr ) 10-12, t1/2 ) 229 000 years), which is produced by cosmic-ray-induced spallation and neutron activation of stable Kr isotopes in the upper atmosphere,1 and 85Kr (85Kr/ Kr ) 10-11, t1/2 ) 10.8 years), which is a nuclear fission product of 235U and 239Pu and is released from nuclear-fuel reprocessing activities. Due to its atmophile nature, most Kr (>98%) resides in the atmosphere and becomes isotopically well-mixed within a few years. Radiokrypton isotopes may serve as chronometers where atmospheric Kr is incorporated in a reservoir such as glacial ice, groundwater, or seawater and subsequently isolated from exchange with the atmosphere. Radiokrypton chronologies are potentially important in diverse studies of hydrology and paleoclimate; the inertness of Kr, being a noble gas, makes radiokrypton-based chronometers superior to other hydrological tracers for * Corresponding author. E-mail:
[email protected]. Fax: 312-413-2279. (1) Loosli, H. H.; Oeschger, H. Earth Planet. Sci. Lett. 1969, 7 (1), 67–71.
8688
Analytical Chemistry, Vol. 80, No. 22, November 15, 2008
many such applications (e.g., 3H-He, 14C, 36Cl, CFCs, SF6). Furthermore, monitoring atmospheric 85Kr has important applications: (i) constraining mesoscale atmospheric transport and dispersion and (ii) verifying compliance with the Nuclear NonProliferation Treaty.2 The analysis of radiokrypton isotopes has been challenging because their low isotopic abundances fall outside the dynamic range of conventional mass spectrometry. Krypton-85 can be analyzed by low-level decay counting, as are other short-lived radionuclides, provided a sufficient amount of purified Kr (i.e., ∼20 µL of Kr extracted from ∼400 L of water).3 However, because of its much longer half-life and lower specific activity, 81Kr cannot practically be measured by low-level decay counting, and therefore its determination requires other methods that remain experimental rather than routine, including accelerator mass spectrometry (AMS),3 resonance ionization spectroscopy (RIS),4 and atom trap trace analysis (ATTA).5-7 These analytical methods for 81Kr all require relatively large quantities of purified Kr gas (e.g., 500 µL for AMS, 50 µL for ATTA). Efficient sampling and purification methods for Kr in environmental matrixes are thus essential for 81 Kr to become a practical tracer. In most systems of interest in radiokrypton studies, the atmospheric component has a chemical composition between those of atmosphere and air-saturated water (ASW), from which O2 may be consumed by biological activity. Krypton typically constitutes 1-4 ppm of this component, and separation of Kr from N2 and Ar is the major challenge in Kr purification. Lithospheric sources of volatiles (e.g., magmatism, water-rock reaction, and biological activity) may add CO2 and CH4 to dilute the atmospheric component. There are several successful methods of separating Kr from bulk gas extracted from seawater and groundwater samples for the determination of 85Kr activity, typically involving about 5 L of (2) von Hippel, F.; Albright, D. H.; Levi, B. G. Sci. Am. 1985, 253, 40. (3) Collon, P.; Antaya, T.; Davids, B.; Fauerbach, M.; Harkewicz, R.; Hellstrom, M.; Kutschera, W.; Morrissey, D.; Pardo, R.; Paul, M.; Sherrill, B.; Steiner, M. Nucl. Instrum. Methods Phys. Res., Sect. B 1997, 123, 122–127. (4) Thonnard, N.; Willis, R. D.; Wright, M. C.; Davis, W. A.; Lehmann, B. E. Nucl. Instrum. Methods Phys. Res., Sect. B 1987, 29, 398–406. (5) Chen, C. Y.; Li, Y. M.; Bailey, K.; O’Connor, T. P.; Young, L.; Lu, Z.-T. Science 1999, 286, 1139–1141. (6) Du, X.; Purtschert, R.; Bailey, K.; Lehmann, B. E.; Lorenzo, R.; Lu, Z.-T.; Mueller, P.; O’Connor, T. P.; Sturchio, N. C.; Young, L. Geophys. Res. Lett. 2003, 30 (20), 2068 DOI:, 10.1029/2003GL018293. (7) Sturchio, N. C.; Du, X.; Purtschert, R.; Lehmann, B. E.; Sultan, M.; Patterson, L. J.; Lu, Z.-T.; Muller, P.; Bigler, T.; Bailey, K.; O’Connor, T. P.; Young, L.; Lorenzo, R.; Becker, R.; El Alfy, Z.; El Kaliouby, B.; Dawood, Y.; Abdallah, A. M. A. Geophys. Res. Lett. 2004, 31, L05503 DOI:, 10.1029/2003GL019234. 10.1021/ac801804x CCC: $40.75 2008 American Chemical Society Published on Web 10/24/2008
Figure 1. Schematic of the purification system. Crossed circles represents valves. Circles with filled triangles are three-way valves. Acronyms AC, MS, and P represent activated charcoal, molecular sieve, and port, respectively.
extracted gas.8-10 Studies of 81Kr in groundwater have involved extraction of more than 100 L of bulk gas,3,7 which was purified in both cases by using an existing Ar purification system involving several gas chromatographic separation steps and a CH4-burning process.3 This approach was designed specifically for the separation of Ar for 39Ar studies, but it can be simplified when only Kr needs to be extracted. In this contribution, we report a relatively simple new method developed to purify Kr from a wide range of bulk gas quantities (5-125 L STP (standard temperature and pressure)). Our Kr purification system uses a quadrupole mass spectrometer (QMS) to monitor gas effluent composition during separation, which enables (i) small-scale cryogenic distillation in a controlled manner and (ii) gas chromatographic separation of parts-per-million by volume (ppmv) level Kr from a significantly large quantity of gas (up to a few liters). A further advantage of analyzing the gas effluent is the applicability of this method to natural groundwater samples characterized by variable chemical compositions unlike atmospheric air. The purified Kr product is analyzed by QMS to determine its purity and yield, as calibrated by an internal Kr standard. All procedures are performed on a single vacuum manifold. EXPERIMENTAL SECTION: DESCRIPTION OF THE VACUUM LINE AND PURIFICATION PROCEDURE The Kr purification system can be subdivided into three sections: cryogenic distillation, gas chromatography, and Tigettering. The schematic is shown in Figure 1. The vacuum line is composed of 6 mm diameter stainless steel tubing, compression fittings, and air-actuated valves. Each section has one or two ports that are connected to the QMS (RGA200, Stanford Instruments) to allow real-time monitoring of the separation procedures as described below. Note that for simplicity some recirculation pathways and traps are not shown in Figure 1. (8) Smethie, W. M., Jr.; Mathieu, G. Mar. Chem. 1986, 18, 17–33. (9) Held, J.; Schuhbeck, S.; Rauert, W. Int. J. Radiat. Appl. Instrum., Part A 1992, 43, 939–942. (10) Sidle, W. C.; Fischer, R. A. Environ. Geol. 2003, 44, 781–789.
Cryogenic Distillation Section. This section simply consists of a light-duty vacuum compressor and two commercially available empty gas containers both cooled at a constant temperature (77 K) using liquid N2. In the distillation phase, sample gas (from cylinder S in Figure 1) is first condensed in the condensation container (300 cm3 volume, cylinder C) at liquid N2 temperature assisted by a vacuum compressor. More than 95% of the gas is condensed. Molecular sieve 4A beds (GAB, 700 g) are placed between the sample cylinder (cylinder S) and the compressor to remove water vapor and carbon dioxide. Subsequently, the gas from cylinder C in equilibrium with the condensed phase (mostly liquid N2) is slowly released and sent to the exhaust container (cylinder E) by the vacuum compressor. Because the vapor pressure of Kr is lower than that of major components such as N2, O2, and Ar, Kr is strongly partitioned into the condensed phase and thereby significantly enriched in the residual fraction remaining in cylinder C. This distillation process also fractionates other elemental ratios (e.g., N2/Ar, O2/Ar, CH4/Ar, and Kr/Ar); thus, the chemical composition of the effluent from cylinder C is continuously monitored through port 0 (P0 in Figure 1) to determine the progress of the distillation. Gas Chromatography Section. The distillation residue from cylinder C is collected in an activated charcoal trap (AC1) at 77 K. Sample gas proceeds from there through molecular sieve 3A and 4A columns (GAC; grain size, no. 60-80; diameter, 6 mm; 1.8 m length each) with He carrier gas to remove residual water and carbon dioxide. The first separation column that contains activated charcoal (AC2, no. 60-80; diameter, 6 mm; length, 1.8 m) separates N2-O2-Ar from Kr-CH4 at room temperature. The former group of gas components is sent to an activated charcoal trap (AC3) at 77 K while the latter flows downstream to the molecular sieve 5A column (MS5A, no. 60-80; diameter, 6 mm; length, 1.8 m) for further separation. The gas effluent from the port 2 (P2) is continuously analyzed to cue switching of the three-way valve. Kr breakthrough from MS5A occurs before CH4; once the passage of Kr is observed at port 3 (P3), indicating that Kr is being trapped at the final activated charcoal trap (AC4) at Analytical Chemistry, Vol. 80, No. 22, November 15, 2008
8689
77 K, the second three-way valve is switched and CH4 sent to the outlet. Ti-Gettering Section. The He carrier gas remaining in AC4 (77 K) is pumped down to ∼1 mTorr using the rotary pump and turbomolecular pump. Subsequently, AC4 is heated at 393 K to release the trapped Kr and trace impurities, which is then introduced into the final section where it is exposed to a Ti-sponge getter at 930 K. Ti is highly reactive metal that readily reacts at high temperature to form titanium oxides and titanium nitrides in contact with gaseous N2 and O2, respectively. As mentioned above, the distillation residue moving through the gas chromatographic separation section is dominated by N2 or O2 depending on the initial chemical composition of the sample gas. Ti-gettering is aimed to remove the trace residues of these reactive gases in the gas chromatographic separation line trapped together with Kr in AC4. The chemical composition of the purified gas is analyzed through port 4 (P4) to evaluate the purity and quantity of Kr; then the pure Kr is collected in the Kr sample container (∼0.7 cm3, containing activated charcoal held at 77 K). RESULTS AND DISCUSSION Cryogenic Distillation. Theory. The evolution of the chemical composition of gas during the cryogenic distillation process described above can be modeled assuming equilibrium partitioning between gaseous and condensed phase. As a gas mixture is cooled and compressed, substances are partitioned into gas and condensed phases. Provided the total quantity of each component and the volume of the system, following equations reveal the equilibrium state of this system.
ni ) Pi,0Xi
Vtot )
Vg + LXi RT
mi (∑ni - L)RT + ∑LXi ∑Pi,0Xi di
with the following definitions: T, temperature (77 K); R, gas constant; Vtot, total volume of the system (300 cm3); Vg, volume of the gas phase; L, total quantity of the condensed phase (mol); ni, total quantity (mol) of a substance i; Pi,0, vapor pressure of a pure substance i; Xi, molar fraction of i in the condensed phase; mi, molar mass of i; di, density of a condensed phase of i. The first equation represents the quantity of a substance i in gas (the first term) and condensed (the second terms) phases, respectively. The second one constrains that the sum of the volumes of the gas and condensed phases equals to that of the system. The iteration was programmed so that a fraction (5-20%) of the gas phase is removed from the system to attain new equilibrium. Ten substances (N2, O2, CO2, CH4, H2O, and five noble gases) are considered in the model to simulate natural samples as explained in the introduction. This model calculation confirms that a cryogenic distillation of airlike gas at liquid nitrogen temperature significantly enriches the residual phase in Kr as shown in Figure 2. The inset box of this figure is a close-up view of the upper left corner, showing that the distillation residue consisting of 0.5% of the initial inventory is expected to retain >95% of the Kr for CH4-poor (95% of the Kr for CH4-poor (55% of Kr would still remain in the residual phase for the gas initially containing up to 99% CH4 when the residual gas fraction reaches 1%. The major component of the distillation residue is expected to be the least volatile substance among major (>1%) substances in the initial gas composition, that is, the first that presents in the order of CH4, O2, and N2 in practice. As distillation proceeds and the concentration of Kr in the condensed phase increases, Kr will eventually be partitioned into the gas phase. In order to retain the majority of the Kr in the residual phase, it is therefore essential to predict the timing at which significant Kr evaporation occurs and to terminate the distillation process prior to this occurrence. Kr has a low initial concentration of a few parts-per-million and is concentrated in the condensed phase by more than 2 orders of magnitude. The Kr signal is thus below the detection limit of QMS while the favorable process is occurring. It is only after a significant fraction of Kr is lost from the condensed phase that the Kr signal becomes visible in the effluent monitored by the QMS. Thus, the progress of distillation must be traced using other parameters than the peak height of Kr. The quantity of total residual gas may be monitored either by the pressure, which reflects the vapor pressure of the system, or by elemental ratios of two substances characterized by significantly different vapor pressures. Among those, the N2/Ar ratio of the gas effluent is an appropriate proxy to trace the progress of the distillation because of the relatively high abundances (thus, they are easy to measure) and the relatively constant ratio of N2 and Ar within naturally occurring gases. As shown in Figure 3, the model indicates that the N2/Ar ratio of the gas effluent should decrease with the total quantity of residual condensate to provide a cue for the end of the distillation step. Note, however, that this method cannot be applied for CH4-rich samples because N2 and Ar are excluded from the residual phase, whereas the majority of
Figure 3. Modeled evolution of N2/Ar ratio in gas effluent of cryogenic distillation plotted as a function of residual total gas quantity. Dashed lines are model results of initially CH4-poor gases, representing chemical compositions of atmospheric air, air-saturated water (ASW), and O2-free ASW. Solid lines are those of CH4-containing gas, and the numbers in the figure represent the percentage of CH4 in the initial gas composition. Chemical composition of the balancing gas was assumed to be a 1:1 mixture of ASW and atmospheric air. Inset box magnifies the final purification stage.
Figure 4. Observed evolution of N2/Ar ratio in gas effluent during cryogenic distillation of atmospheric air plotted as a function of residual total gas quantity (open circles). The expected evolution curve based on the theoretical model in the Theory section is also shown.
CH4 remains condensed (Figure 3). In such case, the vapor pressure of the system becomes so low that distillation process is too slow to be practical (see below for countermeasures). Experimental Results. The distillation model above suggested that the N2/Ar ratio is a reasonable indicator of the distillation process for CH4-poor gas. To evaluate the degree of elemental fractionation in an actual system compared to the theoretical one, the evolution of the N2/Ar ratio and the total quantity of condensed residual gas was analyzed during the distillation of atmospheric air. To minimize chemical heterogeneity in the condensed liquid caused by continuous evaporation at the liquid-vapor interface, it was stirred magnetically during cryogenic distillation. The experimental results are shown in Figure 4 together with the expected evolution curve based on the model above. The N2/ Ar ratio of the gas effluent and the quantity of condensed residual gas covary mostly as predicted. Exceptions are found during the first 30% of the distillation process where the N2/Ar ratio is up to
15% lower than the theoretical value, indicating that the fractionation is less than predicted. Possible explanations for this apparent disequilibrium are either relatively slow kinetics or thermal inhomogeneity. Slow kinetics are indicated by the fact that the system pressure is always lower than the equilibrium pressure, i.e., the pressure increases as soon as the distillation process is interrupted. Varying the flow rate, however, did not resolve this discrepancy. The cause of this deviation may be a mixing of gases from a warmer region of the system into the effluent. It was not possible to determine the fraction of the Kr in the residual phase due to the slow mixing and heterogeneity within cylinder C. A similar experiment was performed using a gas mixture enriched in CH4. As expected from the theory, the distillation process enriches the residual phase in CH4 compared to N2 and Ar. Once the residual phase becomes dominated by CH4, unfortunately, the vapor pressure of the system becomes so low (98% purity of Kr after Ti-gettering that removes the trace reactive impurity such as N2, O2, H2O, and CO2. The determination of the purity is limited by the background noise. Purification Yield of Kr from Natural Gas Samples. Krypton from compressed air and gases extracted from groundwater have been successfully purified using the system described here. Overall yields were determined using a calibrated pure Kr standard. The chemical compositions, quantities, and purification yields of Kr are shown in Table 1. The purification yield varied between 63% and >90%. Up to 5% of sample gas remained in the sample cylinder. This is because the vacuum compressor currently used can only reach ∼60 Torr. The charcoal trap near the outlet of the gas chromatographic separation line (AC3) does not contain a detectable amount of Kr, indicating that the gas chromatographic process does not result in significant Kr loss. As most Kr loss appears to have occurred during the cryogenic distillation process, the gases from the exhaust cylinder were recovered and reprocessed (corresponding to “Kr extracted, second distillation” in Table 1). It turned out that the last quarter of the distillation exhaust contained about 20% of the sample Kr. Moreover, the ratio of gas quantities of the distillation residue to the initial sample (excluding CO2 as it is removed before distillation) roughly correlated with the total Kr yield (open circles in Figure 6). This correlation does not match the theoretical expectation (curves in Figure 6), but the Kr yield is significantly lower for a given residual gas fraction. This data demonstrates that Kr is not as well-retained at the end of the distillation as modeled even when the 5% of bulk gas left behind in the sample cylinder during the condensation is taken into account. Further, the deviation from the model seems to enlarge as the fraction of distillation residue decreases. A viable explanation for this deviation is that the actual pressure of the gas in contact with condensed phase is lower than the equilibrium vapor pressure of the condensed phase due to the pumping (by the vacuum compressor), whereas the model assumes equilibrium. The correlation of decreasing Kr yield with decreasing residual gas amount indicates that it is necessary to retain ∼4% of the sample
Table 1. Chemical Composition, Total Quantity of Processed Gas, and Kr Extraction Efficiency chemical composition
sample no.
gas type
84 110 111 114 115 118 119 125 129 211
O2-poor air O2- and CO2-rich O2- and CO2-rich O2- and CO2-rich O2- and CO2-rich O2- and CO2-rich O2-poor O2-poor air
8692
Kr Kr total total Kr extracted extracted second Kr distillation distillation µL yield residue L Ar % N2 % O2 % CO2 % CH4 % Kr ppm gas L content µL total µL 2.6 0.9 1.4 1.2 1.3 1.6 1.4 2.6 1.3 0.9
97.3 78.1 62.0 52.3 60.2 63.2 62.6 95.2 98.3 78.1
