Identification and Quantification of Radionuclides in Contaminated

---

Article pubs.acs.org/ac

Identification and Quantification of Radionuclides in Contaminated Drinking Waters and Pipeline Deposits Phil E. Warwick* and Ian W. Croudace GAU-Radioanalytical, Ocean and Earth Science, University of Southampton, National Oceanography Centre, Southampton, SO14 3ZH, United Kingdom ABSTRACT: There is broad international interest, particularly with Homeland Security and firstresponder organizations, in developing a range of effective, robust, and rapid analytical tools to identify terrorist assaults. Accidental or intentional radionuclide contamination of drinking water supplies would have significant public health, social, political, and financial implications even where the real risk might be small given public perception. Rapid identification and assessment of the magnitude of any contamination is critical in managing any threat and ultimately in allaying public and regulator concerns and in steering subsequent remediation operations. Conventional screening techniques do not provide information of the radionuclide present, and subsequent identification techniques are too time-consuming and require some prior knowledge of the nuclide identity to permit accurate quantification. The development described here presents a novel, rapid, and effective radiometric approach using industry-standard liquid scintillation counting equipment that can both identify and quantify alpha and beta radionuclide contamination within 1 h of sample receipt. The liquid scintillation counting (LSC) or liquid scintillation analysis (LSA) method, though widely used by the life science and the 14C scientific communities since the 1960s, has greater potential than is often used. The technique developed here, which uses multiple quench parameters for nuclide identification, has been tested on both contaminated drinking waters and pipeline scales with compositions typical of those that might be encountered. It is shown to be highly effective both in terms of rapidly identifying the radionuclide and providing a measure of the quantity of radionuclide present. The whole procedure is about to be developed into an integrated analytical system for use by untrained personnel. It is notable that the development could also be readily applied as a QC procedure in routine radioanalytical measurements.

C

capable of detecting and identifying radioactive contaminants in the water at least down to the maximum permitted levels defined in the Euratom Regulation 2218/89 (Table 1).5 Screening of pipeline deposits is also of importance as deposits can interact with the contaminant radionuclide, retarding the dispersion of the radionuclide and potentially acting as a longterm source of contaminant. In the UK at least, the routine screening of radionuclides in drinking water by Water Companies with laboratories is typically achieved using gas flow proportional counting.6 A volume of water (typically 1 L) is evaporated to dryness, and the residue is sulphated and ignited. The sulphated residue is then mounted on a planchette to produce a thin, evenly distributed source suitable for measurement. The technique is relatively time-consuming and unsuitable for rapid screening in emergency scenarios. A rapid technique has recently been developed6 which is based on the routine gas flow proportional technique but uses smaller sample volumes (to reduce evaporation times) and shorter measurement times more suitable for screening in emergency situations. The technique is capable of analyzing 150 samples per day (assuming the use of a 10 position counter) but still requires 3 h of source preparation/counting. Although the rapid technique is an

ontamination of drinking water supplies with chemical, biological, or radioactive substances (CBRN), either accidentally or through intentional sabotage, is a major concern in drinking water supply management.1 Such incidents can results in detrimental health effects and in significant remediation costs. Contamination of water supplies is not unprecedented although in general such incidents have involved chemical or biological substances. Introduction of radioactive materials into the drinking water supply is a much rarer occurrence; although the health, social, and economic impact could be significant if such an attack did occur.2 In 1985, a threat to contaminate New York drinking water supplies with Pu was received, and Pu was initially detected at elevated activity concentrations in a drinking water sample, although subsequent monitoring failed to confirm the contamination event.3 In such instances, rapid detection and quantification of the contaminant is essential in order to limit the health and financial impact of the incident or to provide public reassurance. Radionuclides that could potentially be used in a planned attack would typically be single nuclide sources associated with industrial/medical applications including irradiators/sterilizers, teletherapy machines, radiothermal generators,1,4 thickness gauges, or medical radiotherapy sources and include a range of alpha, beta, and beta/gamma emitting radionuclides (Table 1). In the event of radionuclide contamination in drinking water supplies, a rapid technique is required that must be © 2013 American Chemical Society

Received: April 17, 2013 Accepted: July 8, 2013 Published: July 8, 2013 8166

dx.doi.org/10.1021/ac401131t | Anal. Chem. 2013, 85, 8166−8172

Analytical Chemistry

Article

Table 1. Medical/Industrial Radionuclides Considered in This Study nuclide

a

half life

dominant decay mode

dominant emission energies (keV)

maximum permitted activity concentration [Bq/L]a

β 318 (99.89%) γ 1173 (99.86%) 1333 (99.98%) β 546 (100%) β 2279 (100%) β 606 (89.4%) 334 (7.36%) γ 364 (81.2%) β 512 (94.6%) γ 661 (90.1%) β 536 (41.6%) β 672 (48.1%) γ 317 (83%) α 5304 (100%) α 4602 (5.55%) 4785 (94.45%) γ 186 (3.28%) (also multiple daughter progeny) α 5443 (12.8%) 5486 (85.2%) γ 59.5 (35.9%) α 6076 (15.2%) 6118 (81.6%)

1000

cobalt-60

5.27 years

beta/gamma

strontium-90 +yttrium-90 iodine-131

28.64 years 2.67 days 8.04 days

beta beta beta/gamma

cesium-137 +barium-137m iridium-192

30.17 years 3.53 min 73.83 days

beta gamma beta/gamma

polonium-210 radium-226

138.38 days 1600 years

alpha alpha/gamma

americium-241

432.7 years

alpha/gamma

californium-252

2.645 years

alpha

125 500

1000 1000

1000 1000

20

20

Maximum permitted levels of radioactivity in drinking water supplies in the event of a radiological emergency.5,10

analysis is limited by the nature of beta decay whereby the energy of the decay is distributed between the β particle and antineutrino. This results in a broad beta spectrum with an asymmetric continuum of beta energies up to a maximum end point energy (Emax) corresponding to the beta decay energy and a peak maximum at approximately 1/3 Emax. Detection efficiencies are related to beta Emax and vary depending on the degree of quench in the sample. Conversely, internal conversion (as observed in 137Cs) produces monoenergetic electrons giving rise to a narrow-width peak in the spectrum and is detected with efficiencies approaching 100%. Alpha particles are also readily detected by liquid scintillation counting with efficiencies ∼100%. As with internal conversion electrons, alpha peaks are narrow. However, the photon yield is significantly lower than for beta of gamma interactions due to ionization quenching and alpha peaks therefore occur at apparent energies equivalent to ∼10% of the true alpha energy. The variation in peak shape and peak position can be used to identify a radionuclide contaminant by comparison with a reference spectrum. The end point energy (Emax) can also be determined to identify the radionuclide. However, the apparent beta Emax measured by liquid scintillation counting will be affected by the degree of quench present in the sample. Quenching reduces the number of photons produced for a given β particle energy resulting in a shift of the spectrum to lower energies and a reduction in the overall detection efficiency. The effect of quench must therefore be taken into account when determining the end-point energy of the spectrum. The current study has developed a novel approach to correct for the effect of sample quench and derive corrected end-point energies to permit nuclide identification. This approach was independently developed before it was discovered that De Filippis11 had presented a similar approach using a combination of internal and external quench parameters to determine beta

improvement on the conventional approach, the method does not provide any spectrometric information which would permit radionuclide identification. In general, nuclide-specific analysis requires knowledge of the nuclide identity and is too time-consuming to permit sufficiently rapid quantification of a contaminant in an emergency scenario. Liquid scintillation counting (LSC) offers numerous benefits for the rapid screening of water samples including high detection efficiencies for alpha and beta emitting radionuclides, sensitivity to low energy beta emitting nuclides, discrimination between alpha and beta events, ability to provide spectral information, and the widespread availability of instrumentation. The use of liquid scintillation analysis for the identification and quantification of anthropogenic radionuclide contamination in drinking waters following a deliberate contamination event has not been previously evaluated. Previous studies have applied LSC to the routine radionuclide screening in waters against WHO drinking water limits.7 Similarly, Jobbágy et al.8 have developed a LSC gross alpha/ beta method for waters that has subsequently formed an ISO standard procedure (ISO11704). Radionuclide identification using liquid scintillation counting has not been extensively reported. Mathematical deconvolution of multicomponent beta spectra has been developed9 but tends to be overly complicated to implement for routine screening and single nuclide identification. For emergency screening, a straightforward procedure which can be implemented using existing technologies is required for radionuclide identification and quantification. In this paper, intrinsic spectral parameters are used to derive correction factors that combine a quenchcorrected peak position with a peak shape factor in order to identify the radionuclide present. The majority of commercially available liquid scintillation counters utilize multichannel analyzer technologies which permit spectral information to be obtained. Beta spectrometric 8167

dx.doi.org/10.1021/ac401131t | Anal. Chem. 2013, 85, 8166−8172

Analytical Chemistry

Article

Figure 1. Relationship between beta Emax and SQPE/SQPI(95) over two quench ranges and between three Quantulus LS counters (Q1, Q2, and Q3).

Emax. In our study, we have extended the approach by introducing an additional peak shape factor measurement to provide an automatable system for radionuclide identification and quantification. The technique has been validated for the rapid identification and quantification of radionuclides in both waters and pipeline scales.

■

cocktail in a 22 mL polythene vial. All test samples were counted for 60 min. Three pipeline deposits identified as P45, APP1, and BayCa were supplied by IWW Rhenish-Westfalian Institute for Water, Mülheim/Ruhr, Germany. P45 and APP1 were Fe rich (containing 64% and 47% Fe, respectively). BayCa was predominantly CaCO3. A Mn-rich deposit (42% Fe and 5.7% Mn), identified here as Mn1, was supplied by National Institute of Health and Welfare, Kuopio, Finland. Samples, with a range of sample masses up to 0.1 g deposit, were digested with 3:1 HCl/HNO3 and ,filtered and the digest was spiked with 137Cs. The solution was evaporated to dryness, and the residue was dissolved in 1 mL of 2 M H3PO4. For Mn-rich samples, 2 drops of 40% H2O2 were added to aid dissolution of the MnO2 present in the residue following evaporation. The solution was transferred to a 22 mL polythene scintillation vial along with 2 × 1 mL water washings. 17 mL Ultima Gold AB cocktail was added, and the samples were counted by liquid scintillation counting to compare measured quench factors with those obtained for water samples. Blind trials using spiked deposit digests were performed to demonstrate nuclide identification and quantification using the multiple quench parameter approach. Approximately 0.1 g of each deposit was spiked with one of the target nuclides. The deposit was digested using 4 mL of 3:1 HCl/HNO3, and the mixture was evaporated to dryness. In the first trial (samples AX/Y, BX/Y, CX/Y), the residue was dissolved in 2 mL of 2 M H3PO4 to produce a colorless solution which was transferred to a 22 mL polythene scintillation vial with 1 mL water washings. Sufficient Ultima Gold AB or Gold Star scintillation cocktail was then added to obtain a total volume of 20 mL. In a second trial, samples of scale (AW/Z, BW/Z, CW/Z) were digested in HCl/HNO3 and the resulting solution was filtered prior to evaporation and dissolution in H3PO4. For Mn-rich samples, 2 drops of H2O2 were added to aid dissolution of MnO2. However, it is important that the resulting solution is heated prior to addition of the scintillation cocktail in order to

EXPERIMENTAL SECTION

Energy calibration was performed using 3H (18.6 keV), 14C (156.5 keV), 35S (167.5 keV), 63Ni (65.87 keV), 147Pm (224.1 keV), 99Tc (293.6 keV), and 36Cl (709.6 keV). All radionuclides are pure beta emitting radionuclides with a single beta emission. In addition, 129I (150 keV) and 131I (606 keV) were also included as the gamma emissions did not significantly affect the beta spectrum. A known activity of the nuclide was transferred in 0.1 mL of solution into a 22 mL polythene scintillation vial. A varying volume of water, ranging from 0 to 8 mL, was then added to the vials to create increasing levels of quench. Sufficient Gold Star scintillation cocktail (Meridian, UK) was then added to produce a total volume of 20 mL. To assess variation in quench parameters with different scintillation cocktails, the test was repeated using 99Tc, 8 mL of Milli-Q water, and 12 mL of Ultima Gold AB, Ultima Gold XR, or Hisafe 3 (Packard, UK). The effect of water hardness on quench factors was tested by mixing 99Tc with varying quantities of mineral water containing 230 mg/L [Ca2+ + Mg2+]. Sufficient Milli-Q water was added to give a total aqueous volume of 8 mL. The sample was then mixed with 12 mL of Gold Star. Peak shape factors were determined for each of the target radionuclides. A set of five standards with varying degrees of quench was prepared for each of the radionuclides identified in Table 1 using the approach described previously. Test contaminated water samples were prepared as part of a blind trial by diluting a traceable radionuclide standard with water to provide activity concentrations either at the action level or at 10% of the action level. Eight milliliters of the solution was then mixed with 12 mL of Gold Star scintillation 8168

dx.doi.org/10.1021/ac401131t | Anal. Chem. 2013, 85, 8166−8172

Analytical Chemistry

Article

Figure 2. Application of peak shape factors for radionuclide identification.

decompose excess H2O2 and avoid chemiluminescence. Ultima Gold AB was used for all second trial samples. All measurements were performed using a 1220 Quantulus liquid scintillation counter (Perkin-Elmer), although the approach would be easily adaptable for other liquid scintillation counters. Samples were counted for 60 min using a full energy counting window. The counter was configured in low bias mode with no pulse amplitude or pulse shape corrections activated. The external quench parameter, SQPE, was determined by positioning an in-built 152Eu standard adjacent to the sample and measuring the induced Compton spectrum. The sample was then counted in the absence of the 152Eu source, and the Compton spectrum was corrected for the contribution from the sample itself. The SQPE value was then calculated on the basis of the position of the Compton spectrum. The SQPE values for the quench calibration standards ranged from 740 to 840 and were independent of the radionuclide present. Internal quench parameters (SQPI) were determined using purpose-designed software (LSC+, Raddec Ltd., UK). The spectrum was divided such that a specified proportion of the sample spectrum lay below the division point. The position of the division point in terms of channel number was then calculated. For example, a measured SQPI(95) value of 500 would indicate that 95% of the counts associated with the spectrum lay below channel 500 and 5% lay above channel 500. The actual end point of the spectrum (equivalent to SQPI(100)) was not used as count rates as this point were low, and the measured position was highly dependent on the

signal-to-background ratio resulting in significant uncertainties in the measured end point value.

■

RESULTS AND DISCUSSION Identification of beta emitting radionuclides was achieved by determining the maximum beta energy of the spectrum, correcting for quench-induced shifts in the spectrum. The position of the sample spectrum is determined by both the beta decay energy and the degree of quench, whereas the external standard quench spectrum (SQPE) is determined only by the degree of quench and is independent of the sample nuclide. Analysis of a series of beta emitting radionuclide standards of varying energies and quench levels showed that the relationship between the sample peak position (SQPI(95)) and the external standard quench parameter varied depending on the energy of radionuclide present. The ratio of SQPE/SQPI(95) was indicative of beta Emax irrespective of sample quench over the quench range studied (Figure 1) and independent of water hardness (with the factor varying by