Rapid processing of 85Kr/Kr ratios using Atom Trap Trace Analysis

TL;DR

This paper introduces a rapid and efficient method using Atom Trap Trace Analysis to measure 85Kr/Kr ratios, significantly increasing throughput and reducing measurement time from 48 hours to about 4 hours per sample.

Contribution

The authors developed a new ATTA-based technique that boosts sample throughput for 85Kr/Kr measurements by over tenfold, enabling faster groundwater dating.

Findings

Measurement throughput increased to 6 samples per 24 hours.

Achieved 3-5% relative uncertainty in 4 hours per sample.

Reduced measurement time from 48 hours to approximately 4 hours.

Abstract

We report a methodology for measuring 85Kr/Kr isotopic abundances using Atom Trap Trace Analysis (ATTA) that increases sample measurement throughput by over an order of magnitude to 6 samples per 24 hours. The noble gas isotope 85Kr (half-life = 10.7 yr) is a useful tracer for young groundwater in the age range of 5-50 years. ATTA, an efficient and selective laser-based atom counting method, has recently been applied to 85Kr/Kr isotopic abundance measurements, requiring 5-10 microliters of krypton gas at STP extracted from 50-100 L of water. Previously a single such measurement required 48 hours. Our new method demonstrates that we can measure 85Kr/Kr ratios with 3-5% relative uncertainty every 4 hours, on average, with the same sample requirements.

Tables

Table 1. Table 1: Krypton calibration samples measured using LLC, ATTA, and the new rapid-processing procedure. a

	LLC Activityb	ATTA 85Kr${}_{\text{SR}}$	Rapid ATTA	Rapid ATTA
			85Kr${}_{\text{SR}}$ (raw)	85Kr${}_{\text{SR}}$ (corrected)c
J5	269 $\pm$ 13	8.0 $\pm$ 0.5d	7.7 $\pm$ 0.3	7.8 $\pm$ 0.3
J4	36.2 $\pm$ 3.1	1.04 $\pm$ 0.03	1.09 $\pm$ 0.05	1.09 $\pm$ 0.05
J3	32.1 $\pm$ 1.2	0.94 $\pm$ 0.01	0.95 $\pm$ 0.05	0.95 $\pm$ 0.05
J2	18.2 $\pm$ 0.6	0.53 $\pm$ 0.02	0.54 $\pm$ 0.03	0.54 $\pm$ 0.03
J1	8.9 $\pm$ 0.4	0.25 $\pm$ 0.01	0.26 $\pm$ 0.02	0.25 $\pm$ 0.02
J0	¡1.0	¡0.013	0.032 $\pm$ 0.006	¡0.021
	(90% C.L.)	(90% C.L.)		(90% C.L.)
a All ATTA results are expressed using the superratio (SR) defined in equation (1).
b Reported in decays per minute per cc of Kr gas at STP, adjusted to 3 March 2016.
c Corrected values include adjustment from the contamination model. Raw values do not.
d Measured using 1$\mu$L of Kr gas to prevent extensive contamination of 85Kr.

Equations

$${}^{85}{\text{Kr}}_{\text{SR}}=\frac{{}^{85}{\text{Kr}}_{\text{Sample}}{/}^{83}{\text{Kr}}_{\text{Sample}}}{{}^{85}{\text{Kr}}_{\text{Reference}}{/}^{83}{\text{Kr}}_{\text{Reference}}}.$$

$$C^{\prime }=C(1-x)+S_{1}x$$

$$M_S=(1-f)S+fC\text{where}f=\frac{\frac{1}{2}t{R}_{\text{Kr}}}{P_{\text{avg}}}.$$

$$M_{S_2}=(1-f)S_2+fC=(1-f)S_2+f{S}_1$$

$$M_{S_3}$$

Full text

\journalname

Water Resource Research

Physics Division, Argonne National Laboratory, Argonne, Illinois 60439, USA Department of Physics and Enrico Fermi Institute, University of Chicago, Chicago, Illinois 60637, USA Climate and Environmental Physics Division, Physics Institute, University of Bern, CH-3012 Bern, Switzerland

\correspondingauthor

Jake C. Zappalajczappala@uchicago.edu

{keypoints}

Krypton-85 is a useful tracer for young groundwater in the age range of 5-50 years

Throughput of krypton-85 abundance measurements increased twelvefold from previous state-of-the-art

1 Introduction

The noble gas isotope 85Kr is a radioactive nuclide with a half-life of 10.739 $\pm$ 0.014 years (Singh and Chen, 2014). It occurs naturally in the atmosphere, produced by cosmic radiation, but at a rate four orders of magnitude lower than current global emission from nuclear fuel reprocessing (Ahlswede et al., 2014). Due to this anthropogenically increased abundance in the atmosphere and a precise understanding of its input function, 85Kr can be applied as a tracer to date young groundwater on the order of 5-50 years old.

Tracers in this age regime are crucial to water resource management given the global increased dependency on groundwater, including instances of complete dependency on young, shallow groundwater for drinking water (Moran et al., 2005). 85Kr provides an excellent compl\replacediement for determining ages when taken with other existing tracers in this age regime, such as chlorofluorocarbons (CFCs) (Cook and Solomon, 1995) and 3H/3He, which have both independent input functions \replaced(Loosli et al., 1999) and unrelated corrections (Visser et al., 2007) from 85Krand corrections from 85Kr (Loosli et al., 1999; Visser et al., 2007). Moreover, CFCs are subject to local contamination (Plummer et al., 2006) and 3H/3He-dating is highly sensitive to natural degassing (Visser et al., 2007). In contrast, 85Kr is steadily released into the atmosphere in a manner that is both monitored and well understood (Ahlswede et al., 2014), making it a robust tool for dating. \replaced85Kr measurements also have an application towards nuclear non-proliferation efforts given that the isotope is released during nuclear fuel reprocessing efforts (Klingberg et al., 2010).85Kr also has a number of applications beyond groundwater dating, such as monitoring air for nuclear fuel processing activities (Kalinowski et al., 2004; Klingberg et al., 2010), monitoring gas transport in the unsaturated zones (which can differ significantly from water transport) (Cook and Solomon, 1995), and as a tracer of ocean water ventilation and shallow mixing (Schröder, 1975).

As a tracer for dating groundwater, 85Kr has been successfully applied on many occasions using low-level gas proportional counting (LLC), both on its own (Smethie et al., 1992) and in conjunction with other isotopic tracers \addedfor deconvolving the age distributions of mixed groundwater \replaced(Corcho Alvarado et al., 2007; Althaus et al., 2009; Mayer et al., 2014; Delbart et al., 2014; Alikhani et al., 2016)(Corcho Alvarado et al., 2007; Althaus et al., 2009; Visser et al., 2013; Mayer et al., 2014; Delbart et al., 2014; Alikhani et al., 2016). 85Kr samples are collected by degassing groundwater samples in the field, and then separating krypton from the bulk gas in the laboratory (Purtschert et al., 2013). Recent developments have both decreased seperation times and increased krypton yields of groundwater samples\added. 10 $\mu$L of krypton gas (STP) can be purified in the laboratory from 10 L of air in approximately 75 minutes. In the field, this amount of air can typically be degassed from 100 L of groundwater in 30-60 minutes (Yokochi, 2016).

However, despite these improvements, 85Kr-dating has not been applied routinely at a large scale due to the slow processing time and comparatively large sample volume requirements of LLC (Plummer and Friedman, 1999; Loosli and Purtschert, 2005). The development of Atom Trap Trace Analysis (ATTA), a laser-based atom counting method (Chen et al., 1999), has sought to provide the necessary tool to make large scale analysis of 85Kr viable. The ATTA-3 instrument at Argonne National Laboratory (ANL), has been used to routinely measure isotopic abundances of 81Kr and 85Kr in groundwater samples (Jiang et al., 2012) using 5-10 $\mu$L of krypton gas at STP extracted from 50-100 L of water \addedor 5-10 L of air at STP. In the past, a single sample measurement required 48 hours.

We report here on a new methodology for 85Kr analysis through ATTA. We demonstrate that, by using this method on a newly improved ATTA-3 system described in Zappala et al. (2017), we now have the ability to continuously measure 85Kr/Kr ratios with 3-5% error every 4 hours, on average, increasing the sample throughput by a factor of twelve from Jiang et al. (2012). We do so with no increase in sample size requirements. We show this method to be linear and repeatable, and present an understanding and control over systematic effects due to cross-sample contamination on the 0.8% level.

2 Atom Trap Trace Analysis Method

The ATTA technique is described fully in Chen et al. (1999); Jiang et al. (2012); Zappala et al. (2017), but summarized here briefly: krypton gas is injected into a vacuum system and passes through a liquid-nitrogen-cooled, radio-frequency plasma discharge. The cooling slows the atoms and the plasma transfers a fraction of the atoms into a metastable electronic state. From this state, the atoms are resonantly excited with 811 nm lasers used throughout the system to further slow and trap the atoms in a magneto optical trap (MOT). By measuring the loading rates of both the radioactive (81Kr and 85Kr) and stable (83Kr) isotopes into the trap, we can obtain an isotopic ratio. Furthermore, in order to remove any systematic effects from changes in efficiency that may occur in the system, we also measure a krypton reference gas immediately after measuring the sample that same day. With the isotopic ratios in both the sample and the reference, we ultimately report a “superratio” (SR) defined for 85Kr as

[TABLE]

Such a routine analysis of both the sample and the reference requires 6 hours of total atom trapping. At current, however, the improved ATTA-3 system described in Zappala et al. (2017) requires 24 hours to complete a single measurement. 16 hours of this time is devoted to “washing” the system to decrease cross-sample contamination caused by our plasma discharge, which implants krypton ions from the sample into our vacuum chamber walls. To remove this implanted krypton the plasma discharge is run using argon gas. This process requires that the system return to room temperature to completely remove frozen krypton, meaning 2 hours of the time are spent warming and later re-cooling the liquid-nitrogen source.

However, for measuring only 85Kr, we can employ a new method. Since 85Kr/Kr isotopic abundances are 10 times higher than 81Kr/Kr, a routine measurement would have sufficient statistics to reach a level of 2-3% error in 0.25 hours, subsequently reducing the amount of krypton being embedded in the system during such a short run. In addition, we can also remove the liquid-nitrogen cooling. This will increase the mean velocity of the atoms, reducing the efficiency of our trap by a factor of 4, lengthening the measurement time to 1 hour; however, it saves 2 hours by removing the heating/cooling cycle.

Here, we present a new measurement procedure without liquid-nitrogen cooling for rapid-processing of 85Kr/Kr ratios using the ATTA system. First we describe a contamination model for this new method that allows us to control the systematic effects caused from the residual cross-sample contamination. Then we apply that model to six calibration samples measured in a 24-hour period. Samples for this experiment were prepared at the University of Bern and ANL. Their activities were measured using LLC at the University of Bern, and their 85Kr superratios, defined in equation (1), were measured using the routine ATTA technique that includes liquid-nitrogen cooling and is described at the beginning of this section. These results are reported in first two columns of Table 1. \addedThe reference gas is represented by the sample J3.

For the proceeding sections, we perform all experiments under our liquid-nitrogen free “rapid-processing” conditions. Measurements are conducted in the manner illustrated in Figure 1: a sample is measured for 85Kr/Kr for 1 hour, followed by a 2.25 hour argon wash, another 1 hour sample measurement, a 1 hour reference measurement, and finally another 2.25 hour argon wash before the cycle is repeated. This timing permits us to measure the 85Kr/Kr ratio of one sample every 4 hours, on average. To define a shorthand for the following sections, a measurement of “$S_{1}$-$S_{2}$-$R$”, would mean a measurement of S1 as the first sample and S2 as the second, followed by a reference measurement $R$. During the sample measurements, gas is recirculated in the system (as it is for a typical ATTA measurement) due to the small size of the samples. During the washes and the reference measurements, the gas is flowed continuously and discarded. \addedDue to this systematic difference between measurements made with and without gas recirculation, the 85Kr superratio of our reference gas measured as a sample (J3) on ATTA is 0.94. However, this systematic difference is found to be consistent and is taken into account as a constant calibration factor. Thus, it does not effect the linearity or reliability of our measurements.

3 Contamination Model

\added

Due to the significantly reduced wash times in this rapid-processing procedure, it is crucial that we develop a model for the cross-sample contamination effects on our system. The goal of this section is to develop a relatively simple but reliable empirical model that quantitatively describes the data without going into the complexities of the cross-contamination mechanism such as specific implantation sites and chamber volumes.

Following the diagram in Figure 1, we\deleted can consider the collective surfaces of ATTA-3 affected by implantation and the volume of sample gas to be two distinct reservoirs. The former is filled with contaminant from previous samples and the latter is filled with our sample to be measured. The contaminant has its own 85Kr/Kr ratio, which we define as $C$. Due to the plasma discharge there is an exchange: sample gas enters the surface reservoir and contaminant leaks into the volume of the sample gas. The contamination that leaks into the sample becomes part of our measured value. The sample which enters into the surfaces replaces some\deleted fixed fraction $x$ of the current contaminant in the reservoir, reducing the influence of each previous sample’s contribution to the contaminant by some fraction 1-$x$. \replacedThe wash afterward then simply reduces all contributions in proportionThe wash procedure afterward reduces the overall number of contaminating krypton atoms by replacing them with argon atoms. However, the wash does not alter the 85Kr/Kr ratio of the contaminant since it affects all implanted krypton isotopes equally.

Thus, if there is some contaminant $C$ before $S_{1}$ is measured, then after the wash we now have a contaminant

[TABLE]

We test such a model by attempting to find a repeatable value for $x$. \addedThis $x$ is particular to our current vacuum system and will require reevaluation if changes are made to the chamber.

To find $x$ we first need to know how much contamination we have in our system. We define $f$, the average portion of the sample ($S$) volume that is replaced by the contaminant ($C$) during a measurement of the sample ($M_{S}$) as

[TABLE]

Here $t$ is the length of the measurement, $R_{\text{Kr}}$ is the linear outgassing rate of the contaminant, and $P_{\text{avg}}$ is the\added average partial pressure of\added krypton during the run. \addedThe factor $\frac{1}{2}tR_{\text{Kr}}$ gives us the integrated contamination injected into the gas volume, which is normalized by the partial pressure of krypton, $P_{\text{avg}}$. The only unknown here is the outgassing rate. To determine the rate we regularly conduct an outgassing test prior to each measurement: the system is filled with argon gas and the gas is recirculated with the plasma active. We then measure how much krypton leaches out of the wall over a few minutes using a SRS Residual Gas Analyzer and extrapolate a krypton outgassing rate due to the argon discharge, $R_{\text{Ar}}$. However, we need to determine the krypton outgassing rate in the presence of a krypton discharge, which should be proportional, but not equal to the rate we have measured, i.e. $R_{\text{Kr}}=bR_{\text{Ar}}$.

To determine $b$, we first clean the system for longer than the normal wash period such that the outgassing rate is more than a factor of 4 lower than the typical rates we expect in these measurements. In this “clean start,” if we measure $S_{1}$-$S_{2}$-$R$, then the contaminant $C\approx S_{1}$ when we measure $S_{2}$. Accordingly we obtain

[TABLE]

Using our calibration samples, we start with a clean system and then measure J5-J0-R. This simplifies the above equation even further, since J0 is devoid of 85Kr (”85Kr-dead”) and thus $S_{2}=0$. With only the second term, we can solve for $b$. We used three such measurements to determine that $b$ = 2.4 $\pm$ 0.3.

Now that we know the value of $f$, we can work to find $x$ by applying the model. If we consider measuring $S_{1}$-$S_{2}$-$R$-$S_{3}$ with our clean start, the model gives us the following for the third sample measurement

[TABLE]

From here, we can solve for $x$. We ran two separate sequences, J2-J5-R-J0 and J5-J0-R-J0-J0, to solve for $x$ and found that $x$ = 0.60 $\pm$ 0.02. Note that we have considered the reference to both be a sampling and a wash procedure. Yet, despite it only being 1 hour instead of 2.25, we still found consistent results. The reason is that, as shown by solving for $b$, a krypton wash is $\sim$2.4 times more effective at extracting krypton than an argon wash. \addedAs such, we could increase the efficiency of the wash by using 85Kr-dead krypton gas as our wash gas. However, sufficient amounts of 85Kr-dead krypton gas are not readily available.

With this repeatable value for $x$ we have determined a simple and consistent model for describing our contamination in this rapid-processing mode.

4 Rapid Processing Results

We measured six calibration samples in a 24-hour period (measured in the order J2-J5-R-J0-J4-R-J1-J3-R). The 85Kr superratios determined from these measurements are listed in Table 1 in the third column, and listed with corrections from the contamination model in the fourth column. The LLC activities of the samples are plotted against these corrected values in Figure 2\deletedPrevious caption: A demonstration of linearity for rapid-processing superratio measurements on ATTA via comparison with LLC results. Both axes are drawn on a log2 scale. J0 was measured to have a superratio of ¡0.021 (90% C.L.), but is not shown or applied to the fit. Instead the fit is forced through zero. The fit has a reduced chi-squared of 0.27. and fit to a line. The measurement of J0 does not appear in the figure \replacedsince it is reported as a limit, however, the fit shown in the figure is forced through zero in recognition that both our rapid-processing method and LLC yield below detection limit results after correctiondue to the log2 scaling, but is included in the fit. The reduced chi-squared of the fit is \replaced0.270.2 We also see that in the six samples we measured the contamination fraction per sample saturated below the 2.5% level, as seen in Figure 3. Based on this saturation level and the errors of our contamination model, the correction will add a maximum error of below 0.8% to the 85Kr/Kr isotopic abundance measurements, which typically have 3-5% statistical error.

The rapid-processing measurements’ agreement with typical ATTA measurements and their linear relationship with LLC activities, demonstrates the validity of this approach. This method increases the throughput of 85Kr/Kr isotopic abundance measurements on a single ATTA system by a factor of twelve. The agreement of this calibration over such a large range of activities (J5 being \replacedup to 3-5 times the typical value in the atmospherenearly 4 times higher than the typical 75 dpm/cc activity in the atmosphere of the northern hemisphere (Ahlswede et al., 2014)) also shows that our contamination model can even handle enrichment levels we would normally wish to avoid in the standard ATTA system. With this rapid-processing procedure validated, ATTA is ready to increase the capacity for 85Kr-dating in the geoscience community.

Acknowledgements.

We thank Zheng-Tian Lu for his comments and suggestions. This work is supported by Department of Energy, Office of Nuclear Physics, under Contract No. DEAC02-06CH11357. We also acknowledge funding from an Argonne/University of Chicago Collaborative Seed Grant. Interested readers can access the data reported in this paper by directly contacting the corresponding author.