Harvesting krypton isotopes from the off-gas of an irradiated water target to generate 76Br and 77Br

A flowing-water target was irradiated with a 150 MeV/nucleon beam of 78Kr at the National Superconducting Cyclotron Laboratory to produce 77Kr and 76Kr. Real-time gamma-imaging measurements revealed the mass transport of the krypton radioisotopes through the target-water processing, or “isotope harvesting”, system. The production rates were determined to be 2.7(1) × 10–4 nuclei of 76Kr and 1.18(6) × 10–2 nuclei of 77Kr formed per incident 78Kr ion. Utilizing an off-gas processing line as part of the isotope harvesting system, a total of 7.2(1) MBq of 76Kr and 19.1(6) MBq of 77Kr were collected in cold traps. Through the decay, the daughter radionuclides 76Br and 77Br were generated and removed from the traps with an average efficiency of 77 ± 12%. Due to the differences in half-lives of 76Kr and 77Kr, it was possible to isolate a pure sample of 76Br with 99.9% radionuclidic purity. The successful collection of krypton radioisotopes to generate 76Br and 77Br demonstrates the feasibility of gas-phase isotope harvesting from irradiated accelerator cooling-water. Larger-scale collections are planned for collecting by-product radionuclides from the Facility for Rare Isotope Beams.

The upcoming Facility for Rare Isotope Beams (FRIB) at Michigan State University provides the scientific community with a vast array of radionuclides that can be utilized for a variety of different basic and applied science experiments 1 . As FRIB fulfills its mission of providing rare isotope beams for fundamental nuclear science research, many useful by-product radionuclides will be formed through beam interactions with accelerator components. Of particular interest are interactions between FRIB's primary heavy-ion beam and a water-filled beam dump which will produce a wide variety of radionuclides. When the beam dump water is recirculated, it is possible to collect the radionuclides through a process termed "isotope harvesting". Isotope harvesting has been identified as a source of otherwise-difficult-to-obtain radionuclides for use in basic and applied research 2 .
To test the proposed harvesting system to be used at FRIB, a similar system was designed and implemented in several beam experiments at the National Superconducting Cyclotron Laboratory (NSCL) 3 . This system was designed with the same materials on a smaller scale and was tested at lower beam power than planned for FRIB as well as with higher areal power deposition. These test experiments have included irradiation with several different stable isotope beams at different beam powers including 1 H, 40 Ca, 48 Ca, and 78 Kr [3][4][5][6][7] . The 78 Kr irradiations are particularly interesting, because they provide an opportunity to simultaneously harvest radionuclides from the cooling water and its off-gas. The results from aqueous phase harvesting during 78 Kr irradiation were discussed by Domnanich et al., while the current manuscript describes the gas-phase harvesting results 7 . For this work, a gas processing system was designed to move dissolved gaseous radionuclides that are produced in the water to gas collection traps. The motivation for the development was to capture 76 Kr and 77 Kr and generate medically relevant bromine radioisotopes, 76 Br and 77 Br, respectively.
Radioisotopes of bromine, such as 76 Br and 77 Br, are uniquely suitable radiolabels for small molecule theranostic radiopharmaceuticals. As a radiohalogen theranostic pair, 76 Br and 77 Br could become a valuable tool for systemic targeted radiotherapy [8][9][10] . Bromine-76 possesses many advantageous properties as a positron emission tomography (PET) imaging tracer, including a high percentage of positron emission (57%) and a long half-life (16.2 h), which enables the delivery of the radioisotope from the production site to point-of-use 8,9,11 . Bromine-77 www.nature.com/scientificreports/ (t 1/2 = 57.0 h) is an Auger electron emitter which has promise for therapy in small disseminated tumors 10,12 . The advantages and disadvantages for the radiobromines in medical applications are thoroughly discussed in a recent review by Wilbur and Adam 13 . Currently, the proton-rich radioisotopes of bromine are produced by proton irradiation of metal-selenides (e.g. CuSe or NiSe) which are enriched in 76 Se or 77 Se [14][15][16] . Very elegant techniques have been developed to optimize production but they have not been widely adopted for clinical radiopharmaceutical use. An alternative approach was pursued by De Jong and coworkers, which was based upon high-energy irradiation of bromide salts to create radiokryptons to generate the bromine isotopes 17 . They found that the formation of generators from the shorter lived 76 Kr and 77 Kr parents (14.8 h, and 1.24 h respectively), made it possible to isolate and use the radioactive daughters for different applications, including radiolabeling peptides and antibodies. Interestingly, the radiokryptons could also be used to radiobrominate small molecules through a process called excitation labeling 18 . More recently, McGuinness, Wilkinson and Peaslee investigated production of the radiobromides and radiokrypton parents by heavy-ion fusion-evaporation 19 .
The current work demonstrates the successful use of an isotope harvesting gas processing line to collect radioactive noble gases and generate pure radiohalogens. With extension to other noble gases, the Kr/Br work presented here has implications for future gas-phase harvesting at FRIB to create generators like 122 Xe/ 122 I or 211 Rn/ 211 At 1 .

Experimental methods
Irradiation and experimental set-up. A water-filled target, described by Domnanich et al 3 , was irradiated at the NSCL with 0-4 particle-nanoAmperes (pnA) of 150 MeV/nucleon 78 Kr beam over the course of 11 h with a total integrated beam current of 16 pnA · h. Measurements of beam current from an unsuppressed target were recorded approximately every second; however, the values were calibrated against a Faraday cup which was inserted to intercept the beam at both low and high power periodically throughout the irradiation. The regularly recorded values from the unsuppressed target were corrected using a linear relationship between the accurate values measured with the Faraday cup and the relative measurements on the target. In addition to the periods of time where the Faraday cup intercepted the beam current, the beam was intercepted early in the beam line several times in order to collect water samples.
The irradiated water-filled target was installed at the end of an experimental beam line at the NSCL. The target was part of a flowing water system which included up to 36 L of water, circulated by a pump from a storage tank through multiple loops for different purposes. These loops included the main target loop which pumped water from the tank to the water-filled target and back, a peroxide decomposition loop to decrease the level of the potentially harmful radiolysis product, an aqueous chemistry loop where isotopes could be harvested from the water, and a gaseous chemistry line where isotopes could be harvested from the headspace described by Domnanich et al. 7 .
Using a helium purge-gas stream of 135 mL/min (Omega FMA5512A 0-500 mL/min Mass Flow Controller), gaseous radionuclides and radiolytically produced gases were sparged from the water into the headspace of the storage tank and carried downstream to the gaseous chemistry line. The gaseous chemistry line consisted of a series of gas traps to purify the gas stream and remove water vapor and from the mixture to optimize the collection of gaseous krypton isotopes. A schematic of the gas line is shown in Fig. 1. The first of these traps was a column filled with boric acid (Fisher Chemical, crystalline, certified ACS), which is known to capture ammonia 20 -a gaseous species which could be produced by reactions between radiogenically produced 13 N and radiolytically produced H 2 in the water during irradiation. Next in the series was another column filled with soda lime (Sigma Aldrich granular, ACS reagent, + 100 mesh), which is a highly efficient absorber for any carbon dioxide present as [ 11 C]CO 2 21 . Both of the first two columns were made from 14 cm of polypropylene tubing (McMaster Carr, OD ¾", 19 mm) with glass wool packing on both ends to contain the chemicals. Following the  The next segment of the gas line included a series of cold traps, used to trap krypton online. First in the series were two stainless steel traps each made with a fritted stainless-steel filter, featuring a large, porous surface (Swagelock, SS-6F-60, 316 stainless steel, 60 micron pore size) and approximately 60 cm of stainless steel tubing (316 stainless steel, ¼" OD, 6 mm, McMaster-Carr). Both traps were placed in a dewar filled with liquid nitrogen for cooling. Following the stainless-steel cold traps was a gas washing bottle trap, filled with molecular sieves (Sigma Aldrich, pellets, 4 Å) also placed in a dewar filled with liquid nitrogen. *Safety note: it should be noted that upon warming the traps pressurize due to desorption of gases from the molecular sieves. Therefore, care must be taken to not overpressurize the fittings and housings in the event of loss of cryogenics, or when allowing the traps to warm.
Following the traps was a hydrogen analyzer (Hy-Optima 700B Series In-Line Hydrogen Process Analyzer). During the irradiation of water, molecular hydrogen is radiolytically produced. By measuring the percentage of hydrogen in the gas stream, predictions of the mass transport of the other gases through the system could be made. Another important sensor in the gas line was a floating-ball flowmeter (Brooks 0 to 400 ccm Variable Area Mechanical Flowmeter), used to confirm a consistent flow rate of the gas at the end of the line after flowing across all traps. This was important to ensure that no obstructions were forming in the traps and all gas that was flowing into the gas line was also flowing out. *Safety note: H 2 can build up and pose an explosion hazard. Therefore, gas flow rates within the line were selected to ensure that the H 2 content in the components would not reach the lower explosive limit in air (4%) if released.
The final segment of the gas line contained a delay line of approximately 150 m of polypropylene tubing (OD ¼", 6 mm, McMaster Carr), leading to two air-tight collection bags (Restek Tedler 100 L Sampling Bags) to collect all gases that were not trapped online during the irradiation.

Radionuclide detection and quantification.
During the irradiation, a PHDS Nuclear Physics Imager (NPI) (PHDS Inc, Knoxville TN, USA) was used to measure the real-time distribution of activity through the harvesting system. This portable detector has the capability to detect, identify, and spatially quantify radioactivity by high-resolution, gamma-ray spectroscopy using a pinhole collimator and a position sensitive, segmented, high purity germanium crystal. During this experiment, NPI pinhole images were used to identify, locate, and quantify 77 Kr using the gamma energy of 129 keV. The detector was placed approximately 1.7 m away from the aqueous and gas chemistry boxes. In this configuration, the spatial resolution was approximately 3 cm. In order to reduce neutron fluence on the detector, three 5-gal water-filled carboys were used as shielding.
Pinhole collimated gamma spectra were collected with the NPI at different time points during beam-off periods (The interaction of neutrons with the detector impeded the collection of data while the beam was directed on-target). With the Imager 32 Software (PHDS Inc) dead-time corrected net photopeak count rates at 129 keV were integrated over regions of interest (ROIs): the water tank (including its headspace), the cold traps, and a background region between the collection boxes. The locations of the ROIs are shown in Fig. 2A, overlaid on the photograph of the experimental set-up. A representative heatmap showing the pinhole collimated 129 keV signal intensity in the field of view is shown in Fig. 2B. These measurements in-between irradiation periods www.nature.com/scientificreports/ were used to determine the movement of 77 Kr through the system and were coupled to a mass transport model to determine the 76/77 Kr isotope production rates. For comparison to the NPI data, the mass transport of 76/77 Kr through the harvesting system was modeled with a set of first order differential equations. Those equations are included in the Supplemental Materials.
After the irradiation, the boric acid and soda lime traps, the three cold gas traps, and the two air-tight gas collection bags were removed from the system and counted with a Canberra Broad Energy Ge Gamma-ray Detector (BE2022). The detector had been calibrated using a sealed 152 Eu source, and additional calibrations were made by placing the 152 Eu source inside one of the cold-trap dewars to accurately account for attenuation through the material. The Genie 2000 Software (Mirion Technologies) was used to analyze the data.
Trapping and collection of krypton isotopes. Using the gas line described above, gaseous isotopes of krypton ( 76 Kr and 77 Kr) were trapped on the cold molecular sieve trap and cold stainless-steel traps during irradiation and for another four hours after irradiation was stopped.
To allow safe removal and separation of the traps, all traps were plumbed with valved quick-release fittings (Colder NS4 series couplings, ¼" NPT) As an additional precaution, the stainless-steel cold traps contained Swagelock turn valves (316 stainless steel sealed valve, ¼" Swagelock tube fittings) as well, which were closed before unhooking the quick-release valves. In order to quantify the amount of 76/77 Kr gas which was trapped in the collection vessels, gamma-ray spectroscopy was performed on each of the traps approximately two hours after removing them from the gas line.
Generation of bromine isotopes. Following the post-irradiation gamma-ray measurements of the molecular sieve trap, both stainless steel traps, and each gas collection bag, an offline experimental set-up was developed to test the generation of 76 Br and 77 Br. All operations were carried out in a ventilated fume hood which was approved for radiological work. This was performed by connecting two traps using valves and polypropylene tubing and cryogenically transferring the collected gases from one trap to another. A schematic of this process is shown in Fig. 3.
An initial transfer was used to combine all the collected 76/77 Kr gas from the molecular sieve trap and the first stainless-steel trap (trap 1) onto the second stainless steel trap (trap 2), using the process shown in Fig. 4 Elution of generated radiobromines. In order to elute the trapped bromine which remained in the stainless-steel trap post-transfer, an experimental set-up was designed to flow water through each trap and to collect the eluent into individual sample vials. This was done by removing the quick release valves and Swagelok turn valves from the inlet and outlet of each trap. At the entrance of each trap, the valves were replaced by a syringe in order to dispense specific volumes of water through the trap to be collected on the other end. A schematic of this set-up can be seen in Fig. 4.
The activity of 76/77 Br in each trap, 1, 2, and 3, was assessed by gamma-ray spectroscopy before and after elution. The volumes of water used for each eluting each trap can be found in Table 1. There were two separate elutions of stainless-steel traps 1 and 2 and three elutions of trap 3. The percentage of the total 76/77 Br activity removed from each trap was calculated.

Results and discussion
Radionuclide production. Using the NPI, 13 position-sensitive gamma-ray measurements were made throughout the irradiation. The results and respective uncertainties are represented in Fig. 6a and b and are provided in tabular form in the Supplemental Material. By fitting the mass-transport predictions to the NPI data, the rate constant for 77 Kr transfer from the liquid phase to the gas phase, (defined as k 1 in the Supplemental Materials), was found to be 4.1(4) × 10 -5 Hz. The inefficiency of trapping, δ, was consistent with zero within an uncertainty of 4 × 10 -8 Hz. Once the krypton-specific constants were determined, it was possible to benchmark the NPI measurements against the calibrated, quantitative, BEGe gamma-ray spectra to find the production rates for 76 Kr and 77 Kr: 2.7(1) × 10 -4 nuclei of 76 Kr, and 1.18(6) × 10 -2 nuclei of 77 Kr were formed per incident 78 Kr ion. The complex time structures in Fig. 5A and B result from the intermittency of beam-on and beam-off periods that were necessary for beam tuning and system checks. The NPI data matched the time structure, which   www.nature.com/scientificreports/ was essential for determining the production rates of each 76/77 Kr isotope. The large number of gamma emitters present in the system made it difficult to window on any other gamma ray than the 129 keV gamma-ray from 77 Kr decay 7 . However, by determining the mass-transport behavior of krypton via 77 Kr mapping, it was possible to calculate the time structure for 76 Kr transport, shown in Fig. 6 in the absence of direct measurements. Around the 16-h time point in the graphs, the traps were removed from the harvesting system and the gas flow was diverted to the gas capture bags (causing the abrupt change in slope for the 76 Kr in traps curve). At that time there was residual 76 Kr in the tank and headspace which could have been recovered with prolonged trapping. In the future, the informed mass-transport model can be used to optimize the time at which the traps are removed from the system for processing. The traps were removed from the system in order to measure the amount of krypton gas which was collected. A total of 7.2(1) MBq of 76 Kr and 19.1(6) MBq of 77 Kr were trapped in the online collection system as shown in Table 2. The table demonstrates how much activity was contained in each trap, with the molecular sieves trapping the most gas. Although Kr is generally retained on liquid nitrogen cooled stainless-steel, it is likely that the perpetual heat load of the incoming carrier gas promoted Kr mobility through traps 1 and 2 resulting in low Kr retention there. Neither the boric acid nor the soda lime traps were radioactive after the experiment, indicating that volatile 7x Br species, such as [ 7x Br]HBr, did not readily evolve from the water. Largely, Br isotopes which were produced directly from the beam or by decay of dissolved Kr isotopes accumulated on the anion exchange resins and catalytic converter 7 .
Generation of bromine isotopes. 76 Br was isolated from 77 Kr, 76 Kr, and 77 Br through a series of transfers.
The first transfer, which involved combining the gaseous krypton isotopes that were collected online during irradiation, is qualitatively demonstrated with the BEGe gamma-ray spectra, shown in Figure S1. A small portion of the 76/77 Kr gas remained in the molecular sieve trap after the transfer, but overall, 0.99(1) MBq of 76 Kr and 3.8(2) MBq of 77 Kr were transferred onto trap 2.
Following the accumulation of all gaseous activity onto trap 2, the 76/77 Kr isotopes were allowed to decay for approximately 7 h. Then the activity of trap 2 two was re-measured before the transfer, and the gamma measurement results can be seen in Fig. 7A. The activity of 77 Kr had decreased significantly from 3.8 to 0.03 MBq, making it possible to prepare a sample of 76 Br without 77 Br as a contaminant. The transfer from stainless steel trap 2 to stainless steel trap 3 resulted in some 76 Kr remaining in trap 2 along with 76 Br and 77 Br, as shown in Fig. 7B and detailed in Table 3. The decay of 76 Kr immobilized in stainless steel trap 3 generated pure 76 Br ( Figure S2).
The final transfer from stainless steel trap 3 to stainless steel trap 4 left pure 76 Br in trap 3 and removed the 76 Kr gas that still remained, which is demonstrated by the absence of any contaminating gamma rays in the  www.nature.com/scientificreports/ gamma spectrum in Fig. 8 and Figure S3. Through the transfer process, it was demonstrated that 76 Br could be produced and isolated from 77 Br due to the differences in half-lives of their parent isotopes. The radionuclidic purity of the 76 Br was 99.9%.
Elution of generated radiobromines. Using the previously described elution procedure, the bromine remaining in each trap after the transfer of krypton, was eluted using different volumes of water. Performing quantitative gamma spectrometry measurements before the rinsing steps and of each eluent vial, the efficiencies   www.nature.com/scientificreports/ of the procedure were determined. An example of the differences in gamma spectrometry measurements before, during, and after the rinsing procedure can be seen in Figure S4. The rinsing procedure of trap 1, which was directly removed from the online set up, consisted of two separate elutions with 17 mL and 10 mL water, respectively. This was the first trial of the rinsing procedure, and as such, a larger volume of water was used to test the ability to remove the 76/77 Br from the trap with water. In total, 89% of the 76/77 Br contained in the trap was eluted with the 27 mL of water. The rinsing of trap 2 was also performed in two consecutive steps, while the water volume was reduced to 10 mL, facilitating a total elution efficiency of 79%. And finally, the rinsing procedure of trap 3 consisted of three separate elutions with 7 mL, 4 mL, and 10 mL of water, respectively. In this rinsing trial, only 76 Br was present in the trap to be removed, and the elution efficiency was 64%.
The elution efficiency of 76/77 Br decreased with decreasing total volume (Table S2) and had the lowest efficiency in the trial using a greater number of small volume washes. While it is important to use small volumes of water for eventually using the eluted bromine in radiolabeling procedures, it is also important to obtain as much bromine from the traps as possible to increase overall production yields using this method. Overall, given the high solubility of bromine in water, the elution efficiency was lower than expected, and this could be due to incomplete wetting of the generator surfaces. Additionally, speciation could play a role in the inefficiency. De Jong, Brinkman and Van Halteren observed formation of organic and other unidentifiable bromine species when eluting in vacuo produced radiobromines from generators. 22 They found that adding hydrogen containing gases to the krypton environment improved extraction of bromine isotopes and led to near-uniform speciation as bromide. In future experiments, similar approaches will be considered.

Conclusions
The irradiation of a novel beam blocker by a stable beam of 150 MeV/nucleon 78 Kr at an average beam current of 1.9 pnA for approximately 11 h facilitated the successful collection of 7.2(1) MBq of 76 Kr and 19.1(6) MBq of 77 Kr. This irradiation experiment utilized a harvesting system which has previously demonstrated the ability to collect radionuclides of interest from the aqueous phase; however, in this case, radionuclides were simultaneously harvested from the water and the gas line to be used as offline generators.
During this irradiation experiment, real-time gamma spectroscopy measurements were possible utilizing the NPI detector from PHDS. With this data, maps of the 76/77 Kr distribution, as it flowed through the system and accumulated on the cold traps, were created. In addition to the real-time radiation maps, the gamma spectra made it possible to predict and model the mass transport of both 76 Kr and 77 Kr through the system based upon the production in the tank water, Henry's law for the sparging of krypton into the gas phase, the flow of gas from the headspace to the traps for collection, and the decay of the 76/77 Kr isotopes over the irradiation time. From this, the production rates of 76 Kr and 77 Kr were determined to be 2.7(1) × 10 -4 nuclei of 76 Kr and 1.18(6) × 10 -2 nuclei of 77 Kr formed per incident 78 Kr ion. Looking forward to higher intensity irradiations at FRIB, it will be very beneficial to have the mass transport model of 76/77 Kr through this system to apply the model to higher production rates.
After the online collection of radioactive 76/77 Kr gas, utilizing differences in temperature, the gas was transferred between different vessels to generate samples of the daughter bromine isotopes. This 76 Kr/ 76 Br and 77 Kr/ 77 Br generator concept proved successful with the isolation of 76 Br from 77 Br. The 76 Br and 77 Br products were eluted from the vessels using water with an average efficiency of 77 ± 12%. Radionuclidically pure elutions of 76 Br could be obtained by a separation based on different half-lives of mother/daughter radionuclides.
While this collection and generation method proved successful, the future of this work will involve more novel methods of noble gas capture including metal-organic frameworks (MOFs). MOFs can be designed to selectively capture certain gases over others and through this method, the efficiency of gas collection of the harvesting system could be improved.
Using the mass transport constants and production rates from this work, it is possible to extrapolate to the conditions expected at FRIB. When FRIB reaches full power, the anticipated steady-state radioactivities are around 50 GBq for 76 Kr and 200 GBq for 77 Kr while the 78 Kr primary beam is in use. If harvested as described, around 5 GBq 76 Kr will be available for extraction each day, which will be accompanied by 3 GBq of 77 Kr. With the minor modification of adding membrane contactors to the harvesting system (in order to increase the degassing rate kinetics), the harvestable portion should exceed 25 GBq/day for 76 Kr which will be accompanied by 50 GBq 77 Kr . Although the 78 Kr beam will only be run for about 10 days per year, the same techniques (with some modification) can be employed to capture Kr, Xe, and Rn isotopes during other irradiations. Since the majority of the scientific program at FRIB will involve a 238 U primary beam, the major impact of gas phase harvesting will likely come from applying the techniques described here to collecting 211 Rn for 211 At generation. 211 At is one of the few viable radionuclides for targeted alpha therapy, and increased access to it through isotope harvesting at FRIB will facilitate translational research which is currently supply-limited. www.nature.com/scientificreports/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.