Quantification of Technetium-99 in Complex Groundwater Matrixes

Jan 6, 2009 - A preconcentrating minicolumn sensor for technetium-99 detection in water consists of a packed bed containing a mixture of anion-exchang...
0 downloads 8 Views 1MB Size
Download PDF
Anal. Chem. 2009, 81, 1068–1078

Quantification of Technetium-99 in Complex Groundwater Matrixes Using a Radiometric Preconcentrating Minicolumn Sensor in an Equilibration-Based Sensing Approach Matthew J. O’Hara,† Scott R. Burge,‡ and Jay W. Grate*,† Pacific Northwest National Laboratory, P.O. Box 999, Richland, Washington 99352, and Burge Environmental, Incorporated, 6100 South Maple Avenue, Suite 114, Tempe, Arizona 85283 A preconcentrating minicolumn sensor for technetium99 detection in water consists of a packed bed containing a mixture of anion-exchange resin and scintillating plastic beads. The column materials are contained in a transparent plastic flow cell placed between two photomultiplier tubes for radiometric detection. Upon retention of pertechnetate anions, the radioactive decay of Tc-99 results in detectable scintillation pulses that are counted in coincidence. In equilibration-based sensing mode, the sample is pumped through the packed bed until complete chromatographic equilibrium is achieved between the activity concentration in the water sample and the concentration on the anion-exchange resin. The analytical signal is the observed steady-state count rate at equilibrium. The sensitivity is related to a measurement efficiency parameter that is the product of the retention volume and the absolute radiometric detection efficiency. This sensor can readily detect pertechnetate to levels 10 times below the drinking water standard of 0.033 Bq/mL. The potential for other anions in natural groundwater and contaminated groundwater plumes to interfere with pertechnetate detection and quantification has been examined in detail, with reference to the groundwater chemistry at the Hanford site in Washington state. Individual anions such as nitrate, carbonate, chloride, and iodide, at natural or elevated concentrations, do not interfere significantly with pertechnetate uptake on the anion-exchange resin. Elevated chromate or sulfate anion concentrations can interfere with pertechnetate uptake by the resin, but only at levels substantially higher than typical concentrations in groundwater or contamination plumes. Nevertheless, elevated anion concentrations may reduce pertechnetate uptake and sensitivity of the sensor when present in combination. Chromate is retained on the anion-exchange resin from water at parts-per-billion levels, leading to an orange stain that interferes with pertechnetate detection by the absorption of scintillation light pulses (color quench). Radioactivity from radioiodine, tritium, and uranium is not expected to create a significant positive bias in ground* To whom correspondence should be addressed. E-mail: [email protected]. † Pacific Northwest National Laboratory. ‡ Burge Environmental.

1068

Analytical Chemistry, Vol. 81, No. 3, February 1, 2009

water analyses. A method of automated fluidic standard addition is demonstrated that corrects for matrix interferences leading to accurate analyses over a wide range of groundwater compositions. This method is developed for automated groundwater monitoring applications. There is a continuing and challenging need for sensors to monitor the concentrations of soluble contaminants in groundwater. Such sensors are needed to support scientific investigations of contaminant transport and to meet monitoring needs during remediation actions or for long-term stewardship. The contaminant concentrations and required detection limits can be extremely low. This challenge is particularly significant in the sensing of radionuclides in groundwater. Drinking water standards or maximum contaminant levels, typically defined in radioactivity units, convert to chemical mass detection limits that are below parts-per-billion (ppb) levels.1-4 For example, the 0.033 Bq/mL (900 pCi/L) drinking water level standard for 99Tc2,4,5 translates to 0.05 µg/L (0.05 ppb). Required mass detection limits for other radionuclides such as 90Sr, 129I, and various transuranic actinides are from 1 to 6 orders of magnitude lower. The development of sensors for radionuclides that emit R or β particles is also challenging due to the short penetration ranges of these particles in water. Furthermore, the radiation characteristics alone do not provide sufficient energy information to discriminate among radionuclides. A chemically selective separation is necessary so that the observed radioactivity can be ascribed to a particular chemical species. (1) Egorov, O. B.; Fiskum, S. K.; O’Hara, M. J.; Grate, J. W. Anal. Chem. 1999, 71, 5420–5429. (2) Hartman, M. J., Morasch, L. F., Webber, W. D., Eds. Hanford Site Groundwater Monitoring for Fiscal Year 2005; PNNL-14670; Pacific Northwest National Laboratory: Richland, WA, 2006. (3) Hartman, M. J.; Morasch, L. F.; Webber, W. D. Summary of Hanford Site Groundwater Monitoring for Fiscal Year 2003; PNNL-14548-SUM; Pacific Northwest National Laboratory: Richland, WA, 2004. (4) Hartman, M. J., Dresel, P. E., Eds. Hanford Site Groundwater Monitoring for Fiscal Year 1997; PNNL-11793 UC-402, 403, 702; Pacific Northwest National Laboratory: Richland, WA, 1998. (5) The drinking water standard for a β-emitter in water as established by the EPA is 4 mrem/year. Using the dose conversion factor from National Bureau of Standards Handbook 69 (U.S. Department of Commerce, as amended August 1963) and the other parameters established by the EPA, one can calculate an equivalent concentration of 33 Bq/L (900 pCi/L) assuming 99Tc is the sole β-emitter. 10.1021/ac8021604 CCC: $40.75  2009 American Chemical Society Published on Web 01/06/2009

In this paper, we are concerned with the development of sensors and automated analyzers for the determination of 99Tc in groundwater. Technetium is a significant radioactive contaminant at the U.S. DOE sites and elsewhere.6,7 It is a pure β-emitter with a half-life of 2.11 × 105 years. It was generated from the thermal neutron fission of 235U with a high production yield of 6% and released to the environment by various pathways.8-11 It is highly mobile in the subsurface as the heptavalent pertechnetate anion, TcO4-, which travels essentially unimpeded by adsorption on mineral surfaces.12-14 At the Hanford site in Washington state, a groundwater area of more than 2.3 km2 exists with 99Tc concentrations above drinking water standards (900 pCi L-1).3 Conventional determination entails groundwater sampling and laboratory analysis by a multistep approach including (a) initial sample treatment using nitric acid and H2O2, (b) a sequence of hydroxide and carbonate precipitation steps, (c) pertechnetate purification using a column of a strongly basic anion-exchange resin, and finally, (d) radiochemical determination by liquid scintillation counting.1,4 Past reviews have indicated a historical absence of chemically selective radiochemical sensors for radionuclides in water. For example, a review article with a section on “Radiochemical Sensors”15 actually described general radioactivity detection techniques, inductively coupled plasma mass spectrometry (ICPMS), inductively coupled plasma atomic emission spectroscopy (ICPAES), neutron activation analysis (NAA), and X-ray fluorescence (XRF) spectrometry for analysis of radionuclides and heavy metals. However, in recent years groups at the Pacific Northwest National Laboratory1,16,17 and Clemson University18-20 have developed a new class of sensors based on small-volume flow-through columns that collect and concentrate radionuclides from larger volumes of sample water. These “preconcentrating minicolumn sensors”, shown in Figure 1, are configured with a whole-column radiometric detection method for radionuclide sensing. The column packing material provides selective uptake of the radionuclide analyte of interest, while simultaneously providing a scintillation signal for radiometric detection. The (6) Hu, Q.-H.; Rose, T. P.; Zavarin, M.; Smith, D. K.; Moran, J. E.; Zhao, P. H. J. Environ. Radioact. 2008, 99, 1617–1630. (7) Hu, Q.-H.; Weng, J.-Q.; Wang, J.-S. J. Environ. Radioact., in press. (8) Browne, E.; Firestone, R. B. Table of Radioactive Isotopes; John Wiley & Sons: New York, 1986. (9) Lieser, K. H. Radiochim. Acta 1993, 63, 5–8. (10) Ashley, K. R.; Ball, J. R.; Abney, K. D.; Truner, R.; Schroeder, N. C. J. Radioanal. Nucl. Chem. 1995, 194, 71–79. (11) Del Cul, G. D.; Bostick, W. D.; GTrotter, D. R.; Osborne, P. E. Sep. Sci. Technol. 1993, 28, 551–564. (12) Lieser, K. H.; Bauscher, C. Radiochim. Acta 1987, 42, 205–213. (13) Lieser, K. H.; Bauscher, C. Radiochim. Acta 1988, 44/45, 125–128. (14) Schroeder, N. C.; Morgan, D.; Rokop, D. J.; Fabryka-Martin, J. Radiochim. Acta 1993, 60, 203–209. (15) Koglin, E. N.; Poziomek, E. J.; Kram, M. L. Emerging Technologies for Detecting and Measuring Contaminants in the Vadose Zone. In Handbook of Vadose Zone Characterization & Monitoring; Wilson, L. G., Everett, L. G., Cullen, S. J., Eds.; Lewis Publishers: Ann Arbor, MI, 1994. (16) Egorov, O. B.; O’Hara, M. J.; Grate, J. W.; Knopf, M.; Anderson, G.; Hartman, J. J. Radioanal. Nucl. Chem. 2005, 264, 495–500. (17) Egorov, O. B.; O’Hara, M. J.; Grate, J. W. Anal. Chem. 2006, 78, 5480– 5490. (18) DeVol, T. A.; Roane, J. E.; Harvey, J. T. IEEE Nucl. Sci. Symp. Conf. Rec. 1997, 415-419. (19) DeVol, T. A.; Roane, J. E.; Williamson, J. M.; Duffey, J. M.; Harvey, J. T. Radioact. Radiochem. 2000, 11, 34–46. (20) DeVol, T. A.; Egorov, O. B.; Roane, J. E.; Paulenova, A.; Grate, J. W. J. Radioanal. Nucl. Chem. 2001, 249, 181–189.

Figure 1. Radiometric preconcentrating minicolumn sensor configuration for 99Tc sensing in groundwater, consisting of a transparent flow cell containing the packed bed placed between two photomultiplier tubes. The sensor assembly is shown connected to the automated fluidic system for sensor column performance characterization. The guard column contains hydroxylapatite to adsorb organic matter.

selective uptake chemistry and scintillating properties may be incorporated together in each individual particle of column packing material or by intimately mixing scintillating particles with chemically selective particle materials. These sensors were comprehensively reviewed in 2008.21 The fluidic presentation of the sample to these sensors can be carried out in a “quantitative capture” mode or an “equilibrationbased sensing” mode.17 In the former approach, a specific volume of sample is passed through the sensor minicolumn, quantitatively capturing the radionuclide of interest while washing away unretained interferences. The sensor response is proportional to the amount of radioactivity in the specific volume of sample, and the sensor is regenerated using a volume of a reagent that releases the radionuclide prior to the next measurement. A flow injection or sequential injection system22-31 is typically used to automate the delivery of samples and regenerating reagents to the sensor flow cell. Alternatively, sequential injection methods developed for the fluidic manipulation and observation of functionalized microspheres32-38 can be used to capture a fresh sample of beads in a flow cell or column for each measurement. This approach, dubbed renewable surface sensing or bead injection,32 has been discussed with reference to pertechnetate sensing.1 (21) Grate, J. W.; Egorov, O.; O’Hara, M. J.; DeVol, T. A. Chem. Rev. 2008, 108, 543–563. (22) Valcarcel, M.; Luque de Castro, M. D. Flow-Through (Bio)Chemical Sensors; Elsevier: Amsterdam, The Netherlands, 1994. (23) Ruzicka, J.; Hansen, E. H. Flow Injection Analysis, 2nd ed.; WileyInterscience: New York, 1988; Vol. 62, p 498. (24) Fang, Z. Flow Injection Separation and Preconcentration; VCH: Weinheim, Germany, 1993. (25) Ruzicka, J.; Marshall, B. D. Anal. Chim. Acta 1990, 237, 329. (26) Ruzicka, J. Anal. Chim. Acta 1992, 261, 3–10. (27) Ivaska, A.; Ruzicka, J. Analyst 1993, 118, 885–889. (28) Ruzicka, J. Collect. Czech. Chem. Commun. 2005, 70, 1737–1755. (29) Miro, M.; Hansen, E. H. Trends Anal. Chem. 2006, 25, 267–281. (30) Yoshimura, K.; Matsuoka, S. Lab. Rob. Autom. 1993, 5, 231–44. (31) Miro, M.; Frenzel, W. Trends Anal. Chem. 2004, 23, 11–20. (32) Ruzicka, J.; Scampavia, L. Anal. Chem. 1999, 71, 257A–263A. (33) Dockendorff, B.; Holman, D. A.; Christian, G. D.; Ruzicka, J. Anal. Commun. 1998, 35, 357–359. (34) Holman, D. A.; Christian, G. D.; Ruzicka, J. Anal. Chem. 1997, 69, 1763– 1765. (35) Willumsen, B.; Christian, G.; Ruzicka, J. Anal. Chem. 1997, 68, 3482–3489. (36) Ruzicka, J.; Ivaska, A. Anal. Chem. 1997, 69, 5024–5030. (37) Mayer, M.; Ruzicka, J. Anal. Chem. 1996, 68, 3808–3814. (38) Egorov, O.; Ruzicka, J. Analyst 1995, 120, 1959–1962.

Analytical Chemistry, Vol. 81, No. 3, February 1, 2009

1069

In the equilibration-based approach,16,17,39-42 a sufficient volume of sample is passed through the column to completely equilibrate the entire bed of column material with the radionuclide activity concentration in the water, resulting in complete breakthrough at the column output. The sensor response is the steadystate amount of activity on column packing material at equilibrium, and the sensor need not be regenerated. It is simply reequilibrated with a new volume of sample for the next measurement. The signal is proportional to the activity concentration in the water sample because the amount of adsorbed analyte at trace concentrations is also proportional to the concentration in water. In the trace concentration range, the amount of analyte on the sorbent in the column is much less than the capacity, the sorption isotherm is linear, and the partition ratio at equilibrium is constant within that sample matrix. In principle, such a sensor can be used for repeated analytical determinations without requiring consumable reagents for analysis or regeneration. However, an automated fluidic system may still be useful for periodic calibration as we discuss and demonstrate below. In this paper we are concerned with the performance of a sensor for the detection of 99Tc in groundwater. Several papers have described pertechnetate sensing using the preconcentrating minicolumn sensor approach.1,16,17,19-21,39-43 The β decay of 99Tc is readily detected with reasonable absolute detection efficiencies (e.g., ca. 30-40%) using scintillating material in the column packing. The selective material for pertechnetate uptake is typically either a liquid anion exchanger impregnated in polymer beads or solid-phase anion-exchange resins. In general, the anion-exchange chemistry is selective because most metal ions and radionuclide metals do not form anions and because, among anions, the large poorly hydrated pertechnetate oxyanion is retained better than many small hydrated inorganic anions.44,45 A prior paper described the concept, theory, and principles for equilibration-based preconcentrating minicolumn sensors for radionuclides and metals, and briefly demonstrated 99Tc quantification in Hanford groundwater.17 We now focus specifically on analytical quantification, selectivity, and interferences in the performance of a pertechnetate sensor in contaminated groundwater. Because it is a radiometric sensor, it is selective for radioactive species whose radiation decay characteristics and specific activity lead to a scintillation signal. Furthermore, it is most selective for such radioactive species that are retained on the anion-exchange resin. However, stable species that do not generate a signal may interfere with the uptake or signal (39) Egorov, O.; O’Hara, M. J.; Grate, J. W. Spectrum 2002: Explor. Sci.-Based Solutions Technol., Bienn. Int. Conf. Nucl. Hazard. Waste Manage. 2002, 9, 928-931. (40) Grate, J. W.; Egorov, O. B. Automated Radiochemical Separation, Analysis, and Sensing. In Handbook of Radioactivity Analysis, 2nd ed.; Elsevier: 2003; pp 1129-1164. (41) Egorov, O. B.; O’Hara, M. J.; Addleman, R. S.; Grate, J. W. ACS Symp. Ser. 2004, 868, 246–270. (42) Grate, J. W.; Egorov, O. B.; O’Hara, M. J. ACS Symp. Ser. 2004, 904, 322– 341. (43) Ayaz, B.; DeVol, T. A. Nucl. Instrum. Methods Phys. Res., Sect. A 2003, 505, 458–461. (44) Bonnesen, P. V.; Brown, G. M.; Alexandratos, S. D.; Bavoux, L. B.; Presley, D. J.; Patel, V.; Ober, R.; Moyer, B. A. Environ. Sci. Technol. 2000, 34, 3761–3766. (45) Gu, B.; Brown, G. M.; Bonnesen, P. V.; Liang, L.; Moyer, B. A.; Ober, R.; Alexandratos, S. D. Environ. Sci. Technol. 2000, 34, 1075–1080.

1070

Analytical Chemistry, Vol. 81, No. 3, February 1, 2009

detection of the target radioactive analyte. This paper examines these stable anionic interferences in detail with reference to both a natural groundwater matrix at the Hanford site in Washington state and the matrix characteristics of groundwater plumes containing multiple cocontaminants. We demonstrate that stable anions whose concentrations are elevated in contamination plumes may reduce the uptake of pertechnetate if present at concentrations that are sufficiently elevated over the natural groundwater anion concentrations or if present in combination. Hence, the sample matrix may alter the partition ratio of the analyte on the sorbent in the column. In addition, the colored anion, chromate, may interfere by the optical absorption of scintillation light. Nevertheless, accurate determination of 99Tc in the presence of varying matrix composition can be achieved using the method of standard addition in the context of an automated fluidic analyzer system. Radioactive cocontaminants are not expected to lead to a significant positive bias. THEORETICAL CONSIDERATIONS Sensitivity. Upon equilibration of the stationary sensing phase in the sensor column with the concentration of analyte in the mobile aqueous phase, the equilibrium amount of the analyte, Meq, captured by the sensor column with the stationary phase volume, Vs, can be calculated according to eq 1. Meq ) CsVs ) CaKVs

(1)

The concentration of analyte in the stationary phase, Cs, is related to the concentration in the aqueous phase, Ca, by the analyte partition ratio K.

K)

Cs Ca

(2)

In the trace concentration range where the sorption isotherm is linear, the amount of analyte sorbed is much less than the sorbent capacity, and the value of K is a constant; however, the value of the constant may vary with the matrix composition. The amount of the analyte captured by the sensor column after equilibration is approximately equivalent to the amount of analyte contained in a sample volume equal to the retention volume Vr. The retention volume, Vr, is related to Meq and the analyte retention factor, k, according to eq 3. Meq ) CaVmk ) Ca(Vr - Vm) = CaVr

(3)

The mobile phase volume in the sensor column is given by Vm. The retention factor is equivalent to the partition ratio multiplied by the phase volume ratio, k ) K(Vs/Vm). Since Vm is typically small relative to Vr, eq 3 simplifies to the expression on the far right. For a given radionuclide, the number of radioactive decay events per second is directly proportional to the number of atoms and their specific activity. (A Becquerel, Bq, is the SI unit of radioactivity defined as 1 decay/s.) Equations describing the total amount of analyte present on the sensor column can be converted

to radiometric sensor response according to eq 4 taking into account that the measurement system detects only a fraction of the total number of decay events. Req ) EdVrAa ) EmAa

(4)

The analytical signal from the column is the radiometric count rate, Req, in counts/s, from a sensor column that is fully equilibrated with a water sample containing an analyte activity, Aa, expressed in Bq/mL. The absolute detection efficiency is represented by Ed with units of (counts/s)/Bq. The retention volume times the absolute detection efficiency, VrEd, defines the measurement efficiency, Em, in units of (counts/s)/(Bq/ mL). The slope of the calibration curve, plotting the response versus the activity concentration of standards, gives this measurement efficiency, which is thus equivalent to the sensitivity. (Measurement efficiency can also be determined by the method of standard addition, as we describe below, or by fitting data to models to obtain VrEd as we have published previously.17) In the analysis of an unknown, the inverse of the measurement efficiency is used to convert Req to Aa, the activity concentration of the sample, as shown explicitly here in eq 5 for comparison with equations of similar form below. Aa ) ReqEm-1

(5)

Limit of Detection. The minimum detectable activity concentration in water (MDA) for the radiometric preconcentrating minicolumn sensor depends on the measurement efficiency according to eq 6.

MDA(Bq/mL) )

Ld -1 E t m

(6)

Here, Ld is the radioactivity measurement limit of detection in counts, which is determined at the 95% confidence level using the Curie equation given in eq 7.46 Ld ) 4.653√Cbt + 2.706

(7)

Hence, the MDA depends on (1) signal acquisition time, t, (2) background count rate, Cb, and (3) the measurement efficiency, Em, for the particular radionuclide analyte in the given groundwater matrix composition, using a particular sensor geometry, flow cell, and bed packing of sorbent and scintillator material. Note that the MDA of the sensor, eq 6, is almost identical in form to eq 5, except that the Ld per count time, t, is substituted for the sensor’s equilibrium count rate, Req. The MDA for a given 99Tc sensor can be improved by either increasing the data integration (count) time, t, or shielding the detector in order to decrease its background count rate. Data integration time, t, greatly affects the MDA over relatively short count times, but the advantages decrease asymptotically as count times are increased. The detector background signal can only be shielded to a point that the PMTs’ dark current becomes dominant. With regard to analyte uptake, the mea(46) Currie, L. A. Anal. Chem. 1968, 40, 586–591.

surement efficiency Em depends in part on the selection of sorbent material in the column and the effect of the groundwater matrix composition on analyte uptake. Potential Interferences. Accurate determinations of sample activity can be made in various groundwater matrixes provided that the measurement efficiency is determined for the analyte in that matrix by calibration. However, a groundwater plume migrating through the subsurface may contain a variety of cocontaminants, some at high concentrations. If the sample matrix changes in a way that alters the measurement efficiency, correct interpretation of a change in signal is no longer certain, since it could be due to a change in analyte concentration, a change in the matrix, or both. The presence of other anions that suppress pertechnetate uptake by the anion-exchange resin may lower the measurement efficiency, and hence sensitivity, by changing the retention volume, Vr. Colored anions, such as chromate, that accumulate on the anion-exchange resin in the column, may absorb scintillation light and reduce the measurement efficiency via their effect on absolute detection efficiency, Ed. If other radionuclides are in the sensor flow cell, they could lead to a positive bias if their radioactivity generates a significant scintillation signal. This signal would depend on the quantity of the radionuclide in the flow cell (either in the mobile aqueous phase in the void volume of the column or retained and concentrated on a solid phase) and the detection efficiency. In a radionuclide sensor with a mixture of chemically selective beads and scintillating plastic beads, the radioactive decay particle must penetrate condensed media before encountering a scintillating bead, losing energy along the way. R-Emitting radionuclides will have much lower detection efficiencies than β-emitting radionuclides, and β-emitting species with lower maximum energies of the β particle will have lower efficiencies than those with higher maximum β energies. Specific potential radioactive interferences will be discussed further in a later section. Standard Addition. Accurate calibration of the sensor in complex groundwater matrixes that may change with time can be obtained using the method of standard addition. As implemented in this paper, a sample measurement is followed immediately by a second sample that is spiked to a known additional 99 Tc activity. The signal of the spiked sample, Req,sp is elevated relative to that of the original sample, Req, and the measurement efficiency for the specific sample matrix (for the condition of the sensor at that time) can be determined according to eq 8.

Req,sp Em )

(

)

Vs - Vsp Req Vs

Amatrixsp

=

Req,sp - Req Amatrixsp

(8)

In our experimental setup described below, the volume of the spike, Vsp, is defined by a fixed loop in hardware, and the total volume of the sample-containing solution (whether spiked or unspiked) is a fixed value, Vs, defined by a sample chamber. The ratio of volumes in the numerator of eq 8 approaches a value of 1 as Vsp becomes insignificant relative to Vs. The activity concentration in the spiked sample, due to the spike material, is Amatrixsp, defined in eq 9 where Asp is the activity concentration of the solution used to spike the sample. Analytical Chemistry, Vol. 81, No. 3, February 1, 2009

1071

Amatrixsp )

AspVsp Vs

(9)

EXPERIMENTAL SECTION Reagents and Standards. All commercial reagents used in this study (nitric acid; sodium salts of nitrate, chloride, sulfate, iodide, and bicarbonate; potassium dichromate) were of analytical grade or higher. Ultima Gold scintillation cocktail was obtained from Perkin-Elmer Life and Analytical Sciences (Boston, MA). The radiotracer 99Tc was obtained in-house from a high-purity stock solution traceable to NIST. It was diluted to a working stock solution (1.48 Bq/mL) in a matrix of Hanford groundwater (the pH of the groundwater was not affected by the addition of the 99 Tc standard). Sensor Column. The packed column or the minicolumn sensor was composed of homogeneous mixture of a 0.68:1 dry weight ratio of 100-200 mesh AG 4-X4 (Bio-Rad Laboratories, Hercules, CA) and 100-250 µm BC-400 (St. Gobain Crystals and Detectors, Newbury, OH). Fast-flow hydroxylapatite (Calbiochem, La Jolla, CA) was packed into a 1 cm3 column and placed immediately upstream of the sensor cell. It was used as a means of removing trace organics in the groundwater. The sensor flow cell was machined from a block of BC-800 (St. Gobain Crystals and Detectors). The block consisted of a machined column with dimensions of 4 mm × 29 mm with 1/4-28 female fittings machined above and below the column. Overall dimensions of the block were 51 mm (H) × 22 mm (W) × 9 mm (D). Details have been described previously;17 however, the specific flow cell used here was newly machined. The BC-800 block was fitted into a customized aluminum support that positioned the minicolumn sensor between the photomultiplier tubes (PMTs) of an FSA 610TR flow-through scintillation counter (Perkin-Elmer, Boston, MA). Counting data was acquired using Flo-One software (version 3.65) running on a laptop computer. The FSA detector configuration was set up to have a region of interest between 0 and 800 channels, and the cell type was specified as “glass”. The time-resolved counting feature was disabled. Groundwater Sample Preparation and Characterization. Uncontaminated groundwater from the Hanford site (well 69919-88, sampled December 2005) was used for all batch studies, instrument development, and testing. Groundwater sample was not acidified upon collection; rather, the water was stored cool and in the dark. Prior to use, groundwater samples were filtered using a 47 mm diameter, 0.45 µm type HA filter (Millipore, Billerica, MA). Radiotracers were always added postfiltration. The activity concentration of the 99Tc samples was verified by direct analysis using a liquid scintillation analyzer Tri-Carb 3100TR (Perkin-Elmer, Boston, MA). Either 2.5 or 5.0 mL of groundwater standard was mixed with 15 mL of Ultima Gold prior to counting. Distribution Constant Measurements. The 99Tc distribution constants (Kd) were measured on AG 4-X4 (Bio-Rad) using uncontaminated, freshly filtered, and freshly spiked Hanford groundwater. The salt spikes augmented the anion concentration by 10, 100, 1000, 10 000, 40 000, and 100 000 ppb (to 200 000 ppb in the case of bicarbonate). Approximately 50 mg of resin (moisture content ) 67.6%) was weighed into a tared 20 mL glass scintillation vial. The vial was transferred into a 1072

Analytical Chemistry, Vol. 81, No. 3, February 1, 2009

radiological area, and a 10 mL volume of 99Tc-spiked groundwater (with and without salt concentrations) was delivered to the sample vial via electronic pipet. The sample was placed on an orbital shaker and allowed to mix (∼150 rpm) with the sorbent for a minimum of 4 h. Replicate experiments in natural Hanford groundwater using 25-190 mg of resin were carried out to determine that approximately 50 mg of resin leads to approximately 50% uptake. Replicate measurements were also made for salt-augmented groundwater samples, varying the resin weight around a midpoint of about 50 mg, and error bars are shown in plots to be described below. After the completion of the contact period, the samples were removed from the orbital shaker. An aliquot of each sample was filtered into a clean vial using a PS 0.45 µm 13 mm diameter syringe filter (Whatman, Inc., Florham Park, NJ) to remove all sorbent from the liquid. Finally, a precise volume (typically 2.5 mL) of the filtrate was delivered to 15 mL of Ultima Gold scintillation cocktail for determination of the radioactivity remaining in the filtrate. Analysis was performed using the Tri-Carb 3100TR liquid scintillation analyzer. The calculation of distribution constant, Kd, was performed using the following equation:

Kd )

(

)(

Ai - Af V Af M

)

(10)

where Ai and Af are the activities of the solution prior to and after contact with sorbent, respectively; M is the mass of the sorbent (dry resin weight); V is the volume of contact solution. (A previously published value of Kd for pertechnetate in Hanford groundwater16 was given per the wet weight of anionexchange resin.) Fluid Handling System for Sensor Characterization and Interference Studies. An automated fluid handling system was configured as shown in Figure 1 using a variety of off-the-shelf components. Fluid handling and reagent selection were made possible using a Kloehn 48 000 step digital syringe pump utilizing a zero-dead-volume syringe (10 mL volume) (Kloehn Company, Las Vegas, NV). The pump featured a six-position distribution valve at its head. Fluid lines were constructed of 1/16 in. o.d./ 0.03 in. i.d. Teflon FEP tubing and were held in place using 1/4-28 and 10-32 flangeless fittings made from PEEK (Upchurch Scientific, Oak Harbor, WA). To eliminate light piping into the detector flow cell, fluid lines leading to and from the Packard FSA detector were constructed of black Teflon FEP tubing. Fluid Handling System for Standard Addition Experiments. Burge Environmental, Inc. assembled a prototype 99Tc groundwater monitoring instrument based on past Burge Environmental experience in developing and deploying the Burge Universal Sensor Platform for Cr(VI) contaminants at the Hanford site,47 joint PNNL/Burge Environmental experience in adapting this platform to radiochemical determination of 90Sr groundwater,48 and specifications from PNNL specifically for a pertechnetate monitoring system based on the precon(47) Burge, S. R.; Hoffman, D. A.; Hartman, M. J.; Venedam, R. J. Sensors 2005, 5, 38–50. (48) O’Hara, M. J.; Burge, S. R.; Grate, J. W. Anal. Chem. 2008, in press.

Figure 2. Automated fluidic system for standard addition experiments. W ) waste line. The pneumatic system to drive solutions with air pressure (open-headed arrows) is not shown. Three-way solenoid valves are indicated with normally open (open circles) and normally closed (solid circles) ports. Volumes of chambers are not shown to scale.

centrating minicolumn sensor. The system is designed to acquire a precise volume of groundwater sample, perform periodic automated standard addition to the groundwater sample, and enable the periodic delivery of a stripping solution to the sensor flow cell. The essential fluidic components are shown in the schematic in Figure 2. Fluid flow was driven by two means: (1) pneumatic fluid transfer using regulated gas pressure from a nitrogen gas cylinder and (2) by a custom-built syringe pump. Fluids were routed along desired pathways using solenoid valves with Teflon diaphragms (N Research, Inc., Caldwell, NJ). Computer control of the instrument was accomplished using custom software prepared in Microsoft Visual Basic 6.0 running on a laptop computer. The system was delivered to PNNL and tested using 99Tc standards prepared in Hanford groundwater. CAUTION! Radioactive solutions used in this work present radiological hazards. RESULTS AND DISCUSSION Sensor Response and Groundwater Chemistry Issues. The preconcentrating minicolumn sensor for 99Tc detection in water consists of a packed bed containing a mixture of anionexchange resin and scintillating plastic beads in a transparent plastic flow cell placed between two PMTs, as shown schematically in Figure 1 and described in the Experimental Section. The mixed bed of anion-exchange resin and scintillating beads is expected to detect 99Tc decay with reasonable efficiency because the maximum and average ranges of the 99Tc β particles (∼750 and ∼130 µm in water, respectively) are roughly equal to or greater than the diameter of the sorbent bead, (e150 µm). The selectivity of the sensor derives from the fact that only radioactive species generating a signal are detected, and the large poorly hydrated pertechnetate anion, 99TcO4-, is collected more efficiently on anion-exchange resins than smaller hydrated anions.44,45 The resin, AG 4-X4, is a weakly basic anionexchange resin that was selected over a strongly basic anionexchange resin because dissolved soil organic matter, such as humic acids, can foul a sensor column that has processed large volumes of groundwater; these bind much more strongly to a strongly basic anion-exchange resin than a weakly basic one.16

Figure 3. Response from four 99Tc calibration standards in Hanford groundwater delivered to the sensor in a random order with regard to concentration, with blank groundwater at the beginning and end of the series. The smaller plot illustrates the strict linearity of the calibration curve.

Figure 4. Plot of the 99Tc minicolumn sensor MDAs (2) as a function of detector count time. The upper horizontal bar represents the regulatory 99Tc DWL of 0.033 Bq/mL; the lower horizontal bar indicates an MDA that is 10 times below the DWL.

We further protected the sensor column from fouling by employing a guard column containing hydroxylapatite upstream as shown in Figure 1 and described previously.16 This material adsorbs soil organic matter without impeding the transport of pertechnetate. In the initial sensor characterization, it was observed that ∼90 mL of sample was sufficient to equilibrate the column. All subsequent experiments delivered a minimum of 130 mL to ensure complete equilibrium. Figure 3 shows sensor equilibration to a series of standards in Hanford groundwater, delivered in random sequence relative to 99Tc activity concentration, with blanks at the beginning and end. The 130 mL samples were delivered at a flow rate of ∼0.915 mL/min for 142 min, followed by 60 min of stop-flow detection (total time ) 202 min for each sample). On the basis of the integrated count rate for this 60 min interval, the sensor responses yielded a calibration line with a slope, Em ) 16.6 (counts/s)/(Bq/mL), and an intercept of 0.013 counts/s (R2 ) 0.9998). Using the observed measurement efficiency, Em, and a background count rate, Cb, of 0.468 counts/s, the MDA as a function of count time was calculated according to eqs 6 and 7; the results are shown in Figure 4. (The average background count rate was obtained as the average of the signal observed from a blank run with natural Hanford groundwater.) For this sensor at equilibrium with 99Tc, less than 10 min of data integration time is required to achieve an MDA at the DWL of 0.033 Bq/mL, while detecting 99Tc at 10 times below the DWL Analytical Chemistry, Vol. 81, No. 3, February 1, 2009

1073

Table 1. Concentrations of Various Major Anionic Species in Pristine Natural Hanford Groundwater and Highly Contaminated Groundwater within a 99Tc Plumea,b anion present in groundwater pertechnetate nitrate chloride sulfate bicarbonate (as alkalinity) chromate a

concn in uncontaminated groundwater (699-19-88)

concn in contaminated groundwater (299-W22-83)

concn in contaminated groundwater (299-W22-46)

1700 3900 13 500 123 000

0.411 81 500 5500 15 100 84 000

0.363 72 200 6000 16 300 92 000

4

120

123

b

Values as of March 2005. Values are in units of µg/L (ppb) except for pertechnetate, which is in Bq/mL.

Figure 6. Pertechnetate uptake on the anion-exchange resin AG 4-X4 in natural Hanford groundwater (well 699-19-88, as of March 2005) as influenced by the elevated levels of anions spiked into the sample: (a) upper plotsnitrate, chloride, and iodide; (b) lower plotscarbonate, sulfate, and chromate. The data trace for each anion begins at the natural groundwater concentration; the horizontal line in each plot indicates the Kd of 99Tc in natural groundwater at 705 ( 83 mL/g. Figure 5. Concentrations of anions (µg/L) in contaminated Hanford groundwater from well 299-W22-83 as a function of time, based on quarterly sampling intervals.

(0.0033 Bq/mL) would require only ∼1 h. These are detection times only; it requires an additional ∼2.4 h to deliver the 130 mL sample volume to the sensor cell. This analysis rate is sufficiently rapid to allow for nine and seven analyses to be performed within a 24 h period for groundwater with 99Tc at the DWL and at 1/10 the DWL, respectively. We have previously demonstrated that the measurement efficiency in groundwater for a similar pertechnetate sensor was lower than that observed in dilute nitric acid solution,17 an effect we attribute to reduced pertechnetate uptake in the groundwater matrix. In some groundwater plumes containing pertechnetate, the concentrations of other anions are elevated over the natural groundwater concentrations, and they change with time. Table 1 presents the concentration of some of the major anions found in groundwater in the Hanford site. It also presents the concentrations found in a groundwater contamination plume that contains significant quantities of 99Tc. Figure 5 illustrates changes in anion concentrations with time in one of the 99Tc-containing plumes on the Hanford site as measured from samples at a fixed location (well 299-W22-83). As the pertechnetate concentration increased, the nitrate and chromate concentrations increased as well. Chloride showed much smaller increases, while the 1074

Analytical Chemistry, Vol. 81, No. 3, February 1, 2009

sulfate background was relatively unchanged over time. It was not known if matrix components elevated over the natural matrix concentrations would alter the sensor sensitivity. Distribution Constant Determination of 99TcO4- in the Presence of Competing Anions. The potential for background anions to interfere with pertechnetate uptake on the anionexchange resin was determined using batch uptake experiments to determine the Kd in mL/g for 99Tc in pristine natural groundwater and groundwater with artificially elevated anion concentrations. Each such series of determinations for a given background anion, and hence each trace in Figure 6, parts a and b, began at the concentration of the anion in natural groundwater. The Kd of 99Tc on the AG 4-X4 in the natural groundwater was measured to be 705 ± 83 mL/g. The data traces in Figure 6a show that nitrate, iodide, or chloride, even at elevated concentrations, do not significantly affect the uptake of 99Tc by the resin material. Figure 6b also shows that elevating the bicarbonate concentration from the natural level of 122 000-322 000 ppb did not alter the Kd. On the other hand, the divalent sulfate and chromate anions, shown in Figure 6b, could potentially suppress pertechnetate uptake if present at sufficiently elevated concentrations in a groundwater plume. Sulfate reduces the 99Tc Kd at concentrations above 20 000 ppb. This concentration is above the natural level of 13 500 ppb (well 699-19-88) and elevated levels of 16 300 ppb and 15 100 ppb in contaminated groundwater (wells 299-W22-46 and 299-W22-83,

Table 2. Test Compositions and Results for Calibration of the Pertechnetate Sensor in Natural and Altered Groundwater Matrixes n-HGW natural groundwater Anion nitrate chloride sulfate chromate Analyte Tc-99 Tc-99 Tc-99 Result Em, (counts/s)/(Bq/mL) intercept, counts/s R2

HGW 2 artificially elevated major anions

HGW 3 elevated major anions plus chromate

Natural and Elevated Anion Concentrations, ppb 1700 71 900 71 900 3900 8000 8000 13 500 20 500 20 500 4 4 300 99 Tc Spike Levels, Bq/mL 0 0 0 0.17 0.17 0.17 0.50 0.50 0.50 99 Tc Sensor Calibration Results 16.5 11.8 7.9 0.006 0.032 0.013 1.000 1.000 1.000

respectively). As noted from the data in Figure 5, the sulfate concentration is relatively unchanging over time. Chromate is observed to reduce 99Tc Kd at levels of ca. 1000 ppb and above. This chromate concentration is much higher than the natural and contaminated levels indicated in Table 1 and Figure 5. Uptake Effects on Sensor Response. Although elevated levels of anions, spiked individually, do not appear to represent significant interferences in terms of suppressing pertechnetate uptake (relative to uptake in natural groundwater matrix), this does not preclude the possibility that multiple anions in combination at elevated levels may suppress pertechnetate uptake and hence sensor response. Indeed, this has proven to be the case. Natural Hanford groundwater (n-HGW) was modified to contain artificially elevated concentrations of the major anions in combination, spiked with pertechnetate, and analyzed with the preconcentrating minicolumn sensor. The fluidic approach shown in Figure 1 was used to deliver blank and spiked samples through the sensor to obtain the baseline signal and the 99Tc sensor response at equilibrium, Req. Table 2 indicates the water compositions analyzed and the resulting calibration sensitivities as the measurement efficiency parameter, Em. Sensor response traces are shown in Figure 7. The top trace in Figure 7 shows the sensor response to pertechnetate spiked into n-HGW, delivering 150 mL samples at 0.915 mL/min over a 164 min period to equilibrate the sensor to each sample. The initial portion of the trace shows the baseline signal while equilibrating the sensor anion-exchange material with the sample matrix. Next, the sensor was equilibrated in sequence with samples spiked with 99Tc to 0.17 and 0.50 Bq/mL. After each spiked sample, a stopped-flow interval of 60 min was used to determine the count rate. The n-HGW series resulted in a response slope (Em) of 16.5 (counts/s)/(Bq/mL), which is nearly identical to the Em determined previously on the initial calibration. Sample HGW 2 contains major anions elevated in combination (nitrate, chloride, and sulfate) at concentrations similar to or greater than those in the contaminated groundwater given in Table 1. The HGW 2 sensor response seen in the middle trace is suppressed by 29% relative to that of HGW. The slope was Em ) 11.78 (counts/s)/(Bq/mL). This decrease in signal response

Table 3. Concentrations of Major Anions in Hanford Groundwater Samples Used for Standard Addition Experiments anion nitrate chloride sulfate

HGW a (ppb) HGW b (ppb) HGW c (ppb) HGW d (ppb) 1700 3900 13 500

72 600 8400 20 700

143 800 13 000 27 800

231 200 19 200 37 000

is taken as evidence that the combination of anions in the solution suppresses the uptake of pertechnetate ion (TcO4-) on the anion-exchange resin within the sensor cell; these anions are not colored and should therefore not interfere with the scintillation signal. The results for HGW 3, which contained the same elevated levels of major anions as HGW 2, but with the addition of chromate, showed a further suppression of the sensitivity to 99Tc by a color quench mechanism and will be described further below. Analysis in Diverse Matrixes Using Standard Addition. We performed a series of tests using a fluidically automated method of standard addition. The apparatus for these experiments

Figure 7. Responses of the minicolumn sensor to pertechnetate in three groundwater compositions. The blue upper trace shows the response of the sensor to natural groundwater (n-HGW), while the green middle trace shows the sensor response to water with elevated levels of major anions (HGW 2, Table 2). The lower orange trace shows the effects of elevated major anions plus elevated chromate (HGW 3, Table 2) that result in color quench.

Figure 8. Series of 99Tc standard traces illustrating standard addition of 99Tc into samples containing varying amounts of dissolved salts (Table 4). Traces 1, 4, 7, and 10 are groundwater samples “a” through “d” without 99Tc, where “a” is n-HGW. Traces 2, 5, 8, and 11 show the sensor response of 99Tc at the DWL, while traces 3, 6, 9, and 12 show the response of the corresponding matrix spike run for samples “a” through “d”, respectively. Analytical Chemistry, Vol. 81, No. 3, February 1, 2009

1075

Table 4. Results of Matrix Spike Addition Runs for Groundwater Compositions Listed in Table 3 and Sensor Traces Shown in Figure 8

sample HGW HGW HGW HGW

a b c d

sensor response, prespike (counts/s)

sensor response, postspike (counts/s)

measd Em [(counts/s)/(Bq/mL)]

calcd activity (Bq/mL)

actual activity (Bq/mL)

bias (%)

0.520 0.352 0.269 0.227

2.411 1.748 1.285 0.919

15.62 11.65 8.25 5.84

0.0333 0.0302 0.0326 0.0389

0.0335 0.0335 0.0335 0.0335

-0.49 -9.85 -2.64 16.24

is shown in Figure 2. The key features of this apparatus are a fixed loop for measuring out a precise and consistent quantity of spike solution and a sample chamber that loads a fixed volume of sample (or sample plus spike) based on the geometry of the chamber and the level sensor. This sample chamber thus contains the sample to be delivered to the sensor for equilibration, and in standard addition procedures, it is where the spike was mixed with the sample. Groundwater standards in four matrix compositions (given in Table 3) were prepared, each with (1) no 99Tc (blank), (2) 99Tc at the DWL of 0.033 Bq/mL (sample), and (3) a 99Tc spike at ∼1 Bq/mL. Sample HGW a is natural Hanford groundwater (equivalent to n-HGW above), whereas solutions “b” through “d” contain elevated levels of the major anionic constituents nitrate, chloride, and sulfate. HGW b is similar to HGW 2 above, whereas HGW c and HGW d have even higher levels of major anionic constituents. The levels in HGW c and d are elevated beyond those found in the contaminated groundwater indicated in Table 1. For each groundwater composition a through d, the system was programmed to first equilibrate the sensor to that groundwater composition by delivering ∼150 mL of blank solution, then counting the sensor for 2.5 h to establish the baseline count rate. Next, ∼150 mL of the 0.033 Bq/mL 99Tc “sample” was delivered to the sensor, followed by a 2.5 h detection interval. Finally, the same 0.033 Bq/mL 99Tc sample was loaded with an interruption midway through the process for the automatic addition of the ∼1 Bq/mL 99Tc spike to the sample via a 24.4 mL standard loop. (This large sample loop volume was necessitated by the low activity of the standards available at the time.) The total sample chamber volume was 193.7 mL; the resulting final matrix spike value was therefore ∼0.126 Bq/ mL above the 0.033 Bq/mL of 99Tc already present in the sample (calculated from eq 9). The combined sensor traces for all four matrix compositions are shown in Figure 8, and the results are given in Table 4. In Figure 8 the numbers below the detector trace aid in discussion. Samples 1, 4, 7, and 10 are blank groundwater samples containing no 99Tc but increasing levels of nitrate, chloride, and sulfate (“a” through “d”). Samples 2, 5, 8, and 11 are groundwater samples containing ∼0.033 Bq/mL 99Tc. Samples 3, 6, 9, and 12 are standard addition runs that were automatically spiked with an additional 0.126 Bq/mL 99Tc via the standard loop; these spiked samples clearly gave higher signals, as expected. The signals for the unspiked and spiked HGW b, c, and d are clearly lower than the HGW a signals due to the elevated anion concentrations that lower the measurement efficiency, Em, for 99 Tc. We subsequently confirmed that the combination of elevated anion concentrations in these samples suppresses 1076

Analytical Chemistry, Vol. 81, No. 3, February 1, 2009

pertechnetate uptake, measuring Kd values of 472 ± 50, 383 ± 52, and 299 ± 32 for HGW b, c, and d, respectively (compared to 705 ± 83 for HGW a which is the same as n-HGW). Nevertheless, the ability of the system to perform automated standard addition by spiking the sample enables the sensor response to be accurately calibrated so that 99Tc activity concentrations can be determined. Table 4 provides the sensor responses, determined Em values, and calculated activities for these samples. The Em values for the four groundwater samples decreased from 15.62, 11.65, 8.25, and 5.84 (counts/s)/(Bq/ mL) as the level of combined dissolved salts in the samples increased. However, the standard addition enabled the correct 99 Tc values to be calculated within an analytical bias of less than 20% (last column). Color Quench Effects of Cr(VI) on Sensor Response. Hexavalent chromium, expected to be present as chromate anions (CrO42-) in alkaline Hanford groundwater plumes, could lead to interference via both uptake effects and color quench effects. Cr(VI) is strongly retained on anion-exchange resins, but this does not appear to significantly influence pertechnetate retention below 1000 ppb based on the uptake data presented above. As a cocontaminant with pertechnetate in Hanford plumes, elevated chromate levels are more likely to be on the order of 100 ppb. However, columns equilibrated with chromate are visibly colored orange. Chromate has a weak UV absorption peak with maximum at ∼375 nm. The wavelength of maximum emission of the BC-400 scintillating plastic bead materials is 423 nm. Scintillation photons generated by 99Tc β-decay events can be absorbed prior to detection by the PMTs; this interference mechanism is called “color quench”. The HGW 3 sample (see Table 2) was spiked to contain Cr(VI) at a concentration of 300 ppb, in addition to having elevated

Figure 9. Response of the minicolumn sensor first to 0.17 Bq/mL 99 Tc in 130 mL of n-HGW, followed by a 60 min stopped-flow counting interval. Next, a solution of groundwater with 0.17 Bq/mL 99Tc and 300 ppb Cr(VI) in n-HGW was introduced. The inset plot in the upper right shows the pulse height spectra for the two steady-state responses.

Table 5. Decay Characteristics of Selected Radionuclides and Their Affinity for Weakly Basic Anion-Exchange Resin (AG 4-X4) in Groundwater analyte/interference -

99

Tc (TcO4 )

3

H (H2O) I (I-)

129

238

235

U, U (various anionic uranyl species)

radiationa

half-life, specific activity

retained on anion-exchange resin

β decay Eβ ) 294 keV β decay Eβ ) 18.6 keV β decay Eβ ) 154 keV R decay

2.11 × 10 year 1.7 × 10-2 Ci/g 12.32 year 9.7 × 102 Ci/g 1.57 × 107 year 1.8 × 104 Ci/g 238 U: 4.5 × 109 year 3.4 × 10-7 Ci/g 235 U: 7.0 × 108 year (235U) 2.2 × 10-6 Ci/g

yes Kd ) 705 ± 83 mL/g no

5

yes Kd ) 180 ± 26mL/gb yes Kd > 105

a The β energy is the maximum energy of the β decay. b The Kd of iodide on the AG 4-X4 resin in natural Hanford groundwater was determined using 125I tracer.

concentrations of major anions like HGW 2. HGW 3 samples with and without pertechnetate were delivered to the sensor as described above, and the traces are shown in Figure 7. The sensor responses were visibly lower than those seen for HGW 2, and the Em value determined from the slope was 7.93 (counts/s)/ (Bq/mL). The effect of chromate is further illustrated in Figure 9. The sensor was equilibrated with blank n-HGW followed by equilibration with n-HGW containing 0.17 Bq/mL 99Tc, resulting in a net equilibrium count rate of 2.79 counts/s. Subsequently, the sensor flow cell was perfused with n-HGW containing the same 99Tc activity while also having 300 ppb Cr(VI). Although the 99Tc activity on the sensor should not change at this level of chromate (i.e., the uptake of pertechnetate is not suppressed at this chromate concentration, see Figure 6b), the observed signal is seen to decrease as chromate is accumulated onto the column. Because the distribution constant of chromate is higher than pertechnetate, a larger volume of groundwater is necessary to fully equilibrate the sorbent in the column. At about 1200 min into the experiment, the observed sensor signal reaches a steady state, indicating that chromate has come to equilibrium within the sensor material and the color quench effect becomes constant. The new net sensor response was 1.77 counts/s, which represents a ∼37% decrease in signal due to color quench by retained chromate anions. Further evidence that the chromate effect is due to color quench was seen in the pulse height spectra from the detector’s multichannel analyzer (shown in the inset Figure 9). It was observed that, when chromate was present on the column, the spectral distribution of detected pulses was clearly altered toward pulses with lower numbers of photons reaching the PMT and the overall intensity (from the area) was lower. Although chromate can cause a decrease in the Em for 99Tc analysis, the sensor is still highly sensitive to 99Tc β-decay events. At the rather high level of 300 ppb chromate, the sensor was capable of reaching the drinking water limit of 99Tc in less than 1 h, and 10× below the drinking water limit in less than 4 h. Standard addition, as described above, can be used to ensure that the monitor system remains accurate for 99Tc even when chromate has equilibrated with the sensor material.

Potential Radiometric Interference with Sensor Response. The presence of radioactive cocontaminants could lead to a positive bias in the sensor signal, particularly if they are concentrated on the anion-exchange resin and the radiation from the interfering radionuclide is detected with good detection efficiency. Three radionuclides of potential concern in plumes on the Hanford site are tritium (3H), radioiodine (129I), and uranium. Relevant characteristics of these radionuclides are given in Table 5. Anions containing uranium are strongly retained. However, the detection efficiency of a mixed bed column for R particles is very poor due to their short ranges in water. Therefore, uranium isotopes are not expected to create a positive bias. In initial experiments, we have observed that no signal is detectable from the sensor column when it has retained 3 times more uranium activity than the activity of 99Tc retained on the column in equilibrium with 99Tc at the DWL. Radioiodine is retained on anion-exchange resin and emits a β particle that can be detected in a mixed bed column. However, the retention of iodide is lower than pertechnetate, and the maximum energy of the 129I β particle is significantly lower, which will lead to reduced detection efficiency. We equilibrated the sensor column with 0.037 Bq/mL 129I in Hanford groundwater and determined an Em for 129I of only ∼0.56 (counts/s)/(Bq/mL); this sensitivity is a factor of ∼30 times lower than that for 99Tc (Em is ∼16 (counts/s)/(Bq/mL)). 3 H is a highly mobile contaminant present in large plumes on the Hanford site. However, it is not retained by the anionexchange resin, and it decays by a β emission with a low maximum energy. Hence, the detection efficiency for 3H will be quite low. In an initial experiment, we observed no measurable scintillation signal upon delivery of a 3H standard of 14.7 Bq/mL to the minicolumn. This 3H activity was >400 times higher in activity concentration than 99Tc at the DWL. Although further research will be required to fully characterize potential radiometric interferences, the radioactive species described above are not expected to be of significant concern. DISCUSSION We have assessed the impact of complex groundwater matrixes on the quantitative response of the radiometric preconcentrating minicolumn sensor for 99Tc, operating in equilibration-based sensing mode. We have shown that this sensor can respond Analytical Chemistry, Vol. 81, No. 3, February 1, 2009

1077

to changing concentrations of 99Tc without the use of consumable reagents between measurements. However, anions present in natural groundwater influence the sensitivity of the sensor requiring calibration for the matrix of interest. If the overall matrix is sufficiently constant with time, changes in 99Tc concentration in groundwater could be determined using an in situ sensor calibrated for that matrix. On the other hand, if matrix composition changes significantly due to cocontaminants or other causes, then periodic sensor recalibration will be necessary. This may be achieved using an automated fluidic system that performs standard addition, as we show in Figure 2 and discussed above. Such a fluidic system is suitable for use as an “at-site” monitor which could be deployed at a sampling well or as part of a sampling network. Such monitors could also be used to determine 99Tc concentrations in process waters associated with pump and treat operations. Automated monitors of this type could significantly reduce labor and analytical turnaround time if used to replace conventional manual procedures in analytical laboratories and could make possible frequent sampling and analysis intervals to support field studies of contaminant transport. The monitor design in Figure 2 includes a provision for periodically stripping adsorbed species from the sensor column material. We have observed that carbonate solution can remove accumulated anions, including 99TcO4- and the colored CrO42-, but it is not effective at removing a brown stain at the inlet of

1078

Analytical Chemistry, Vol. 81, No. 3, February 1, 2009

the column that we attribute to traces of soil organic matter that have passed through the hydroxylapatite guard column. However, 3 M nitric acid can remove some (but not all) of this stain. We also anticipate that stripping solutions could remove any uranium that accumulates on the column. After “rejuvenation” by such stripping solutions, the sensor would need to be re-equilibrated and recalibrated with the relevant groundwater matrix. ACKNOWLEDGMENT The authors gratefully acknowledge funding from U.S. DOE Office of Science Environmental Management Science Program and the Environmental Remediation Science Program as well as funding from the U.S. DOE Office of Science STTR program. We thank Dr. Anne Farawila for a detailed reading of this manuscript. J.W.G. acknowledges that a portion of the research was performed using EMSL, a national scientific user facility sponsored by the Department of Energy’s Office of Biological and Environmental Research located at Pacific Northwest National Laboratory. The Pacific Northwest National Laboratory is a multiprogram national laboratory operated for the U.S. Department of Energy by Battelle Memorial Institute. Received for review November 26, 2008. AC8021604

October

12,

2008.

Accepted