Chromatographic Separation of Selenium and Arsenic: A Potential ⁷²Se/⁷²As Generator

Summary

An anion exchange method was developed to separate selenium and arsenic for potential utility in a ⁷²Se/⁷²As generator. The separation of the daughter ⁷²As from the ⁷²Se parent is based on the relative acid-base behavior of the two oxo-anions in their highest oxidation states. At pH 1.5, selenate is retained on strongly basic anion exchange resin as HSeO₄⁻ and SeO₄²⁻, while neutral arsenic acid, H₃AsO₄, is eluted.

1. Introduction

Positron emission tomography (PET) has become a routine nuclear medicine imaging technique with the approval of 2-deoxy-(¹⁸F)-fluoroglucose (¹⁸F-FDG). PET imaging allows for better resolution nuclear medicine images than single photon emission computed tomography (SPECT) and allows for quantitation. Several positron emitting radionuclides are currently available but most are short-lived (e.g., t_1/2 = 110 min for ¹⁸F, t_1/2 = 20 min for ¹¹C, t_1/2 = 12.7 h for ⁶⁴Cu) [1]. Longer-lived positron emitting radionuclides would allow the imaging of biological processes that require longer localization times such as radiolabeled antibodies. Zirconium-89 is being developed as a longer-lived, metallic positron emitter (t_1/2 = 3.27 d), and enhanced chelate development is currently underway [2]. Iodine-124 is a long-lived, non-metallic, positron emitter (t_1/2 = 4.176 d) being developed for molecular imaging applications, but minimizing radioiodine impurities and maximizing ¹²⁴I production yield during either proton or deuteron irradiation remains a trade-off [3–6]. An alternative non-metallic, positron emitting radionuclide is ⁷²As (t_1/2 = 26.0 h, 88%, β⁺, 2.49 MeV endpoint energy, 12% EC), which is available from the decay of ⁷²Se (t_1/2 = 8.4 d, EC, 46 keV γ (58% abundant)) and makes this parent-daughter pair suitable for development of a ⁷²Se/⁷²As generator [7]. Compared to most PET radionuclides, the longer half-life of ⁷²As allows for attachment to antibodies and proteins for quantitative imaging of biochemical and physiological processes and receptor mapping [8–12]. The chemistry of arsenic is diverse (oxidation states of 0, +3 and +5) allowing for the development of potentially valuable PET radiopharmaceuticals. Direct labeling of the monoclonal antibody Bavituximab with ⁷²As for tumor-specific PET imaging has been reported [13].

Selenium-72, the parent of ⁷²As, can be produced from charged particle reactions on Ge, Br, or As targets [14]. The most common production routes involve α or ³He irradiation of ⁷⁰Ge, ⁷²Ge, or ^natGe, generally as the oxide (e.g., ⁷⁰Ge(α,2n)⁷²Se or ⁷⁰Ge(³He,n)⁷²Se), or a proton induced nuclear reaction on stable arsenic (⁷⁵As(p,5n)⁷²Se) [15–18]. Cross-section data suggest that these are the best production methods [15,19,20]. Alternatively natural NaBr targets irradiated with higher energy protons (72–92 MeV) generate Se radioisotopes (⁷²Se, ⁷³Se, ⁷⁵Se) with fewer by-products than higher energy reactions [7,21,22]. The higher beam currents available (average of 101.4 μA) at the Los Alamos National Laboratory Isotope Production Facility (LANL-IPF) make this a viable route although radioselenium impurities are generated [22]. Any of these routes will produce ⁷²Se for generator development, with the ⁷²Se generally separated from its target as Se(0) [15,17] or Se(IV) [16,18,20].

Development of a column generator for the separation of ⁷²As from ⁷²Se, similar to the Mo-99 alumina column generator for ^99mTc (e.g., Ultra-TechneKow^®) or the Sr-82 stannic oxide column for ⁸²Rb (e.g., Cardiogen-82^®) would be advantageous. Selenium-72/Arsenic-72 (⁷²Se/⁷²As) radionuclide generators have been proposed in the literature [14], however complex separation/isolation methods are required. Separation of the daughter from the parent can occur via repeated distillation of “grown-in” ⁷²AsCl₃ from carrier added ⁷²Se stock solutions [16,21]; electroplating of ⁷²Se as Cu₂Se on Cu backings [23], solvent extraction [22] or solid phase extraction of ⁷²Se [17,24]. However, with the exception of the solid phase extraction method, the reported methods require laborious and/or significant handling. The solid phase approach as reported in the literature is limited to very small dimensions (500 μL cartridge volume, few μCi of ⁷²Se activity) involving reagents such as HF, requires Se in its elemental state under an inert atmosphere to prevent oxidation, and scale up to clinically useful 10 mCi batches may not be feasible [17,22,24].

Extraction chromatographic methods, including batch distribution ratio methods and column chromatography, have provided simple, robust methods for the separation of parent/daughter radionuclide pairs [25–27] including the development of a higher activity ¹³⁷Cs/^137mBa generator system for industrial and environmental applications [27]. Using several radioisotopes of arsenic and selenium, an anion exchange method was developed for the separation of As from Se with each element in its highest oxidation state, namely as selenate (Se(VI)O₄²⁻) and arsenic acid (H₃As(V)O₄). A low activity generator was prepared and evaluated.

2. Experimental

Caution! Selenium-75 (⁷⁵Se), ⁷²Se, ^72,73,74,76As, and ²²Na are radioactive and all work involving these radionuclides were carried out in approved laboratories following the appropriate radiation safety procedures. Selenium-75 (t_1/2 = 119.78 d) used as a Se tracer in column and batch studies was produced at the University of Missouri Research Reactor (MURR) via the ⁷⁴Se (n, γ) ⁷⁵Se reaction of an encapsulated ⁷⁴Se (as SeO₂; 98.85%) target, which was dissolved in 1 M HNO₃. Arsenic-76 (t_1/2 = 1.0942 d) used as an As tracer in batch studies was produced at the MURR via the ⁷⁵As (n, γ)⁷⁶As reaction of an encapsulated ⁷⁵As (as ^natAs₂O₃) target, which was dissolved in 1 M NaHCO₃. A mixture of radioactive selenium (^72,75Se), arsenic (^72,73,74As), and sodium (²²Na) in 0.1 M HCl and ~3 M NaCl (from processing) [22], produced by the irradiation of natural NaBr targets with 72–92 MeV protons at the Los Alamos National Laboratory – Isotope Production Facility (LANL-IPF), was evaluated for producing a ⁷²Se/⁷²As generator. On receipt, the sample contained 19.34 μCi of ²²Na, 2.8 μCi of ⁷²As (with ⁷²Se in equilibrium), 62.51 μCi of ⁷³As, 56.26 μCi of ⁷⁴As, and 35.24 μCi of ⁷⁵Se as determined by gamma spectroscopy (HPGe). Table 1 lists the radionuclides produced at LANL during irradiation of the NaBr target, their half-lives in days, the initial activities measured at LANL, and the activities measured at the University of Missouri after processing and receipt.

Table 1.

Activities of sample from LANL.

Nuclide	Literature28 t1/2 (days)	Measured Initial Activity at LANL on 07/24/201122	Measured Activity at MURR on 08/23/2011
22Na	950.5	Not determined	19.34 μCi/mL
71As	2.72	0.19 ± 0.01 Ci	Decayed
72As	1.08	Not determined	2.8 μCi/mL
73As	80.3	a Not determined	62.51 μCi/mL
74As	17.77	0.046 ±0.001 Ci	56.26 μCi/mL
72Se	8.4	0.03 ± 0.01 Ci	b Not determined
75Se	119.78	0.050 ± 0.002 Ci	35.24 μCi/mL
76Br	0.68	8.1 ± 0.7 Ci	c Not present
77Br	2.38	4.2 ± 0.6 Ci	c Not present
82Br	1.47	0.27 ± 0.06 Ci	c Not present

^aAccurate measurement of the low energy (56 keV) was impossible with the high activities present.

^bAccurate measurement of the low activity, low energy (46 keV) gamma was not possible.

^cBr isotopes had been removed prior to our receipt.

Table 2 lists the various radionuclides, their nuclear properties, and their specific activities as used in these studies.

Table 2.

Evaluated sodium, arsenic, and selenium isotopes and their associated gammas [28].

Isotope	Half-Life	Gamma Emissions (Abundances)	Specific Activity
22Na	2.6027 a	511 (180%), 1274.5 keV (99.9%)	1.65 MBq/mg (4.5 × 10−5 Ci/mg)*
72As	26.0 h	511 (176%), 834.0 keV (81.0%)	nca (6.18 GBq/mg; 1670 Ci/mg)
73As	80.30 d	53.4 keV (10.3%)	nca (825 MBq/mg; 22.3 Ci/mg)
74As	17.77 d	511 (58%), 595.8 keV (59.0%)	nca (3.67 GBq/mg; 99.3 Ci/mg)
76As	26.3 h	559.1 keV (45.0%)	5.74 MBq/mg (0.155 mCi/mg) at EOB*
72Se	8.4 d	45.89 keV (57.2%)	nca (7.99 GBq/mg; 216 Ci/mg)
75Se	119.78 d	121 (17%), 136.0 keV (58.3%) 264 (59%), 279 (25%), 400 (11%)	nca in LANL sample 9.36 MBq/mg (0.253 Ci/mg)*

nca = no carrier added; specific activities calculated.

^*Specific activities determined from target mass irradiated.

2.1 Materials

Reagents and solvents were purchased from Alfa Aesar (Ward Hill, MA), Fisher Scientific (Pittsburg, PA), Mallinckrodt (St. Louis, MO), and Sigma-Aldrich (St. Louis, MO) and used as received. Dowex anion exchange resins were obtained from Bio-Rad Corporation (Hercules, CA) and Sigma-Aldrich (St. Louis, MO). Alumina basic form (Sigma Aldrich), alumina neutral form (Sigma-Aldrich), alumina acidic form (Fisher Scientific), silica gel 60 (Brinkman, Des Plaines, IL), magnesium oxide (Mallinkrodt), Florisil (Fisher Scientific), and zirconium oxide (Alfa Aesar) were evaluated. All water used was purified on-site (deionized water from a Millipore system to >18MΩcm).

2.2 Instrumental analysis

Ion pairing HPLC analysis was performed using a Perkin-Elmer Model 410 LC pump with controller and Turbochrome software equipped with a Rainin Dynamax UV-1 variable UV-visible detector. An analytical reversed phase C18 column (Alltech Hypersil, 120 Å, 0.46 cm × 25 cm × 5 μm) with an isocratic mobile phase of 20:80 acetonitrile: phosphate buffer (60 mM, pH 7.2, 20 mM tetrabutylammonium hydroxide as the ion pair agent) at a flow rate of 1 mL/min and UV detection at 210 nm was utilized for HPLC. Retention times for selenite, arsenate and selenate were 3.1, 3.4 and 3.9 minutes, respectively.

Radiochemical assays for samples containing solely ⁷⁵Se were determined by gamma counting using a Harshaw integral line NaI(Tl) solid scintillation well detector with a Canberra HV power supply and Ortec preamp.

Radiochemical assays for ²²Na, ^{72, 73,74}As, and ⁷⁵Se were performed by γ-ray spectroscopy using a Canberra Model GC2018S HPGe detector system. The detector diameter was 60.5 mm, detector length was 30.5 mm, and the distance from the window was 5 mm. The detector’s specified FWHM at 1.33 MeV was 1.8 keV. Spectral analyses were performed by a Canberra Model 9600 multichannel analyzer.4

2.3 Determination of optimum chromatographic media

For the development of a ⁷²Se/⁷²As generator, various support media were evaluated to elute arsenate while retaining selenate. Silica gel, acidic alumina, neutral alumina, basic alumina, Florisil, zirconium oxide, magnesium oxide, Dowex 1-X8 and Dowex 1-X2 were tested for their ability to retain selenate while allowing arsenate to elute under neutral to slightly acidic conditions (Table 3). Columns were separately loaded with non-radioactive sodium selenate and sodium arsenate dibasic heptahydrate dissolved in water, and were then evaluated for ion retention by elution with DI H₂O and either dilute HCl (<1 M) or dilute HNO₃ (<1 M). The collected elution fractions were evaluated for selenate and arsenate using ion pairing reversed phase HPLC analysis described above.

Table 3.

Column and eluent conditions evaluated.

Matrix (volume)	Eluent	Arsenate	Selenate
Silica (1.5 mL)	Water or 0.12 M HCl	Elutes rapidly	Elutes rapidly
Acidic Alumina (1.5 mL)	Water, neutral saline, 0.12 M or 0.24 M HCl	Did not elute	Did not elute
Neutral Alumina (1.5 mL)	Water or normal saline	Did not elute	Elutes
Basic Alumina (1.5 mL)	Water, normal saline, 0.12 M or 0.24 M HCl	Did not elute	Elutes
Florisil (1.5 mL)	Water, normal saline, 0.012 M or 0.12 M HCl	Elutes	Elutes
Zirconium Oxide* (0.5 mL)	0.001 M, 0.01 M or 0.1 M HNO3	Elutes	Elutes
Magnesium Oxide* (0.5 mL)	0.012 M HCl	Elutes	Elutes
Dowex 1-X2 or X8 (1.5 mL)	Water or normal saline	Did not elute	Did not elute
	0.012 M or 0.12 M HCl	Elutes	Did not elute
	1.2 M HCl	Elutes	Begins to elute very slowly

^*Fine powders; smaller volumes for reasonable flow rates.

2.4 Optimization of elution parameters

To optimize the elution parameters on Dowex anion exchange resin, sorption studies were performed via the batch method for selenite, selenate, and arsenate. Prior to use, samples were verified to contain purely selenate, selenite, or arsenate by comparison to standards using HPLC. All experiments were performed at room temperature (20°C). The removal of selenium and arsenic from varying pH (−1 to 7) solutions was measured by mixing 50 mg of AG 1-X8 resin with 1.5 mL of a ⁷⁵Se- or ⁷⁶As-spiked solution. The selenium tracer was 117 kBq (3.16 μCi)/sample and the arsenic tracer was 146 kBq (3.95 μCi)/sample. The liquid-solid system was mixed by vortexing for 2 minutes and immediately centrifuged for 2 minutes at 7500 rpm. Two 500-microliter aliquots of the supernatant (A_s) were transferred into clean counting vials. Additionally, a 500-microliter aliquot of the original solution (A₀) was transferred to a clean counting vial to determine the original activity. The distribution ratio, K_d, is calculated as follows:

$$K_d=\frac{(A_0-A_s)}{A_s}\ast \frac{V}{m}$$

A₀ is the original selenium or arsenic activity of the aqueous solution. A_s is the selenium or arsenic activity remaining in solution following contact with the resin. (A₀ − A_s) is the amount of activity adsorbed by the resin. The volume (V) of the aqueous solution is measured in milliliters and the mass (m) is measured in grams resulting in the distribution ratio having units of mL/g. A larger distribution ratio correlates to a more effective removal of selenate, selenite, or arsenate from the aqueous solution.

2.5 Oxidation of Se(IV) to Se(VI)

An oxidation method was developed to treat the samples to ensure that selenium was present as Se(VI). The optimum method was as follows: samples of ⁷⁵Se were treated using 1 mL of 30% H₂O₂ and refluxed for 1 hour. Each sample was quantitatively loaded onto a 1 mL bed volume of Dowex AG1-X8 (100–200 mesh) column and washed with DI H₂O (4 × 1 mL), 0.03 N HNO₃ (4 × 1 mL), and 0.6 N HNO₃ (4 × 1 mL) with all fractions collected and counted using the NaI(Tl) well detector. Selenous acid (H₂SeO₃) eluted with 0.3 N HNO₃ and selenic acid (H₂SeO₄) eluted with 0.6 N HNO₃, allowing determination of Se oxidation state present.

The LANL-IPF mixture (222 kBq; 6 μCi/sample) was evaluated by this technique to determine the Se oxidation state. Treated and untreated control samples were eluted from Dowex columns in the manner described above and the fractions were counted on the HPGe detector, and evaluated for ²²Na, ⁷³As, and ⁷⁵Se content (Table 1).

2.6 Application of method to Se/As generator column

A large, preparative column (2.2 cm i.d. × 12.5 cm) packed with a 48 mL bed volume of Dowex 1-X2 (50–100 mesh) resin was conditioned with 3 M HCl followed by DI H₂O (~10 bed volumes). Approximately 11.1 MBq (300 μCi) of the LANL-IPF sample was neutralized, oxidized and diluted (total volume = 300 mL), and then added to this pre-concentrating column followed by DI H₂O washes (2 × 50 mL). Arsenate was eluted from the column with 0.03 N HNO₃ (6 × 50 mL) followed by DI H₂O (2 × 50 mL). Selenate was then eluted from the column with 0.6 N HNO₃ (6 × 50 mL) followed by DI H₂O (2 × 50 mL). The 2^nd and 3^rd fractions of the 0.6 N HNO₃ elution (containing the majority of the selenate) were combined, the volume slowly reduced to 0.1 mL by evaporation in a hot water bath under air flow, and then neutralized (1 M NaHCO₃), oxidized (30% H₂O₂), and diluted with DI H₂O to 32 mL. This sample was loaded onto a generator Dowex 1-X2 column (50–100 mesh; 4 mL wet resin volume; 0.7 cm i.d. × 10 cm; conditioned as above), washed with DI H₂O (10 mL), and then eluted after 1, 3, and 8 days with 0.03 N HNO₃ (5 × 5 mL) and DI H₂O (2 × 10 mL). The fractions were counted on the HPGe to analyze for ^72,73,74As and ⁷⁵Se content.

2.7 Effect of radiation on Dowex resin

With the anticipated, large radiation dose to the resin expected from a ⁷²Se/⁷²As generator, the Dowex 1-X8 resin was exposed to 2880 Gy (288 MRad) over 2 weeks using a ⁶⁰Co source. The elution profile of two columns (0.6 cm i.d. × 4.2 cm; 1.2 mL wet volume), one containing the exposed resin and one the control unexposed resin, were prepared as described above to evaluate any difference in breakthrough of ⁷⁵Se. The same activity of ⁷⁵Se (1850 Bq (50 nCi); 0.24 ng; 0.206 Ci/mg) was applied to each column, with ⁷⁵Se used as radiotracer for ⁷²Se for easier counting.

3. Results and discussion

Several methods have been reported to separate ⁷²As from ⁷²Se, none are straight-forward and most require extractions or harsh chemical treatments (e.g., HF) to generate ⁷²As in the +3 oxidation state, which is readily oxidized at the tracer level to the +5 oxidation state. To this end, an anion exchange method has been developed that takes advantage of the difference in acid strength between selenic acid (H₂SeO₄) and arsenic acid (H₃AsO₄) with the ⁷²As in its +5 oxidation state.

3.1 Determination of optimum chromatographic media

Several chromatographic materials were evaluated for their ability to retain selenate and elute arsenate, with Dowex anion exchange resins (1X-8 and 1X-2) determined to be the best suited (Table 2). At pH = 1.5 the arsenate ion is nearly fully protonated and is eluted from the anion exchange resin while the selenate ion retains a negative charge and is not readily displaced. None of the hydrous metal oxides evaluated gave a suitable separation as they retained arsenate and eluted selenate or both species were similarly retained.

3.2 Optimization of elution parameters

To determine the best pH for selenate retention and arsenate elution, equilibrium distribution ratios were acquired as a function of pH. The retention of selenate, selenite, and arsenate was impeded by the high ionic strength of the pH −1 and 0 solutions. Since selenic acid is a strong acid, selenate was readily retained by the resin and quickly reached its maximum K_d. Maximum K_d is a function of the minimal detectable activity of the detector and the activity of the spike; the maximum K_d in this experiment is 3570 mL/g. Since selenous and arsenic acids are weaker acids, selenite (pK_a1 = 2.62) and arsenate (pK_a1 = 2.19) are weakly retained until reaching a solution pH where the anionic species is the primary form. The several orders of magnitude difference in the retention of selenite and selenate highlights the necessity of an oxidation step to prevent loss of selenium from the column. Additionally, the results indicate eluting at pH 2 allows the parent, selenate, to be retained on the column while the arsenate daughter is removed (Figure 1).

Figure 1.

Distribution coefficient for selenate, selenite, and arsenate as a function of pH (n = 6; the error bars are included).

3.3 Oxidation of Se(IV) to Se(VI)

The first pK_a value of selenous acid (H₂SeO₃) is close to that of arsenic acid (H₃AsO₄) (2.62 and 2.19, respectively), which would make their separation very difficult as shown in Figure 1. Thus, a method was developed to ensure that all of the selenium was present as selenic acid to afford a separation from the weak arsenic acid. An ion pairing reversed phase HPLC method was used to determine the speciation of selenate, selenite and arsenate present and to optimize the conditions for the oxidation of Se(IV) to Se(VI). Refluxing 30% H₂O₂ was found to completely convert selenite to selenate. Variations in pH had no effect on the conversion or elution.

The LANL-IPF sample, which contained radioisotopes of Se, As and Na, and had been oxidized with hydrogen peroxide/HCl to remove bromide isotopes as Br₂ prior to shipping to the University of Missouri Research Reactor (MURR), required additional treatment with refluxing 30% H₂O₂ to convert all of the selenium to selenate. Either the original oxidation was insufficient to convert the selenium to Se(VI) or over time the selenium originally converted was reduced to a range of lower oxidation states (Figure 2). The sample contained principally ²²Na, ⁷³As, ⁷⁴As and ⁷⁵Se as observed by HPGe gamma spectroscopy with ⁷⁵Se in the greatest abundance. At low levels of ⁷²Se activity the ⁷²Se gamma signal is not readily quantitated from the high Compton background at 45 keV. Chromatographic separation of the non-oxidized sample showed selenium in at least three different forms: an uncharged or cationic species eluting with water, selenite eluting with 0.03 N HNO₃ and selenate eluting with 0.6 N HNO₃. The arsenic present eluted with water as an uncharged or cationic species and the sodium also eluted with water.

Figure 2.

Un-oxidized aliquot of the LANL-IPF sample evaluated on Dowex 1X-8 (100–200 mesh). Eluent of fractions: 0 = sample addition; 1–4 = water; 5–8 = 0.03 M HNO₃ (pH 1.5); 9–12 = 0.6 M HNO₃ (pH 0.2). Selenium is present in three different chemical forms: Se(0) in fraction 0; selenite in fractions 5–6; selenate in fractions 9–11.

Treatment with refluxing 30% H₂O₂ oxidized the arsenic to arsenate, and the selenium to selenate. No lower oxidation state As or Se species were observed (Figure 3).

Figure 3.

Oxidation of LANL-IPF sample with 30% H₂O₂ evaluated on Dowex 1X-8 (100–200 mesh). Eluent of fractions: 0 = sample addition; 1–4 = water; 5–8 = 0.03 M HNO₃ (pH 1.5); 9–12 = 0.6 M HNO₃ (pH 0.2). Arsenate elutes in fractions 5 and 6, while selenate elutes in fractions 9–11.

3.4 Application of method to Se/As generator column

To examine the practical application of the separation method to a generator preparation, a portion of the neutralized, oxidized, diluted LANL-IPF sample was pre-concentrated to reduce the volume and remove Na and As radioisotopes produced during irradiation with 72–92 MeV protons. A portion of the neutralized, oxidized, diluted LANL-IPF sample was directly added to a preparative column; the eluent resulting from the addition yielded 481 kBq (13 μCi) of ²²Na and trace amounts of ⁷⁴As. No selenium was detected during the addition. Elution with 0.03 N HNO₃ removed approximately 24 μCi of ⁷⁴As and 41 μCi of ⁷³As and a much smaller amount of ⁷²As. Again no Se was detected. Elution with 0.6 N HNO₃ yielded over 1.61 MBq (43.5 μCi) of ⁷⁵Se and trace quantities of ^72,73,74As. From the concentrated fractions, more than 97 % of the ⁷⁵Se was transferred to the 4 mL (bed volume) generator column. No radioisotopes were detected in the resulting eluent from addition of the sample to the generator column. Results from elution with 0.03 N HNO₃ are illustrated in Figure 4. After the first elution, which removed the residual arsenic isotopes, ⁷²As was the only arsenic radioisotope eluted. The activity of the LANL sample was measured by HPGe to contain 35.2 μCi/mL of ⁷⁵Se and 2.8 μCi/mL of ⁷²As (and also ⁷²Se in equilibrium) on 08/23/2010. These activities were decay corrected to the expected activities during the generator column experiment, 26.4 μCi/mL of ⁷⁵Se and 0.045 μCi of ⁷²As (⁷²Se). Note the significantly higher activity for ⁷⁵Se over ⁷²Se at this time. The final elution of the generator column yielded 50 nCi of ⁷²As and 1.5 nCi of ⁷⁵Se. The expected ⁷²Se breakthrough was determined from the ⁷⁵Se:⁷²Se ratio present in the sample at this time. Only 0.0026 nCi of ⁷²Se would have co-eluted with the ⁷²As, which represented 0.005% of the ⁷²Se activity on the column.

Figure 4.

Elution profile of generator column. Each elution included 0.03 M HNO₃(4 × 5 mL) followed by a water rinse (2 × 5 mL).

3.5 Effect of radiation on Dowex resin

Additionally, a test was performed to assess possible changes in selenium breakthrough upon exposure of the Dowex anion exchange resin to gamma radiation. After irradiation with 2880 Gy (288 MRads) cumulative dose over two weeks using a ⁶⁰Co source, the resin was darker in color and did not wet as well as the untreated resin. The column prepared from the irradiated resin appeared to have a slightly higher flow rate than the column prepared from the original resin (visually observed; not measured). Breakthrough of Se(VI) occurred on day 16 (elution 12) for the ⁶⁰Co irradiated resin, and on day 23 (elution 16) for the untreated resin. The column capacity to retain Se(VI) was only marginally affected by the strong extended gamma radiation conditions. The complete breakthrough curves are shown in Figure 5.

Figure 5.

Se(VI) breakthrough comparison of irradiated and non-irradiated Dowex 1×8 resins. The columns were eluted with 0.03 M HNO₃ (4 × 1 mL) followed with a water rinse (4 × 1 mL).

4. Conclusions

The development of new drugs utilizing longer-lived radionuclides will allow for the continued growth and increased application of PET imaging for quantitation, especially radioimmunoimaging. Arsenic-72, a long-lived, non-metallic positron emitter, is best produced in high specific activity from the decay of its ⁷²Se parent. A ⁷²Se/⁷²As generator system was developed and will greatly increase the availability and application of ⁷²As. Dowex 1-X2 and 1-X8 resins retained selenate while allowing no-carrier-added ⁷²As arsenate to be eluted.

Distribution coefficient studies showed selenite has over an order of magnitude weaker retention on Dowex 1-X8 than selenate, which would lead to undesired ⁷²Se breakthrough and product loss. A simple oxidation method was developed to ensure selenium is completely in the selenate form.

A ⁷²Se/⁷²As generator, prepared from an irradiated NaBr target, demonstrated the production and elution of ⁷²As from the decay of ⁷²Se. It was further demonstrated that Dowex 1-X8 held up relatively well to radiation damage, which should allow for elutions to be performed over a longer time period without ⁷²Se breakthrough due to radiolysis.

Future work will include preparation and evaluation of a ⁷²Se/⁷²As generator containing higher activities similar to those used in a clinical setting. This would allow determination of radiation resistance of the polystyrene-divinylbenzene copolymer backbone of Dowex AG 1-X8 or 1-X2 resins or whether other resin backbones are needed. A ⁷²Se/⁷²As generator would make ⁷²As available as a tracer for medical, toxicological and environmental studies.

Highlights.

• Ion exchange separation for ⁷²Se and ⁷²As developed.

• Oxidation method to Se(VI) for column loading developed.

• H₃AsO₄ eluted at pH 1.5 while SeO₄²⁻/HSeO₄⁻ retained.

• K_d values reported for As(V), Se(IV) and Se(VI).

• Resin gamma irradiation has minimal effects on Se breakthrough.