Heavy-ion production of 77Br and 76Br

Many radioisotopes with potential medical applications are difficult to produce routinely, especially those on the proton-rich side of the valley of stability. Current production methods typically use light-ion (protons or deuteron) reactions on materials of similar mass to the target radioisotope, which limits the elemental target material available and may require the use of targets with poor thermal properties (as is the case for the production of radiobromine). These reactions may also create significant amounts of proton-rich decay products which require chemical separation from the desired product in a highly radioactive environment. A promising alternative method using heavy-ion fusion-evaporation reactions for the production of the medically relevant bromine radioisotopes 76Br (t1/2 = 16.2 h) and 77Br (t1/2 = 57.0 h) is presented. Heavy-ion beams of 28Si and 16O were used to bombard natural chromium and copper targets just above the Coulomb barrier at the University of Notre Dame's Nuclear Science Laboratory to produce these bromine and precursor radioisotopes by fusion-evaporation reactions. Production yields for these reactions were measured and compared to PACE4 calculations. In addition to using more robust targets for irradiation, a simple physical–chemical separation method is proposed that will lead to very high radiopurity yields. A summary of accelerator facility requirements needed for routine production of these radioisotopes is also presented.

Si beam, the primary exit channel of interest is a (2p + n) fusion-evaporation reaction on 52 Cr: 52 Cr makes up 84% of natural chromium, so isotopically enriched targets are not required. Reactions on the other chromium isotopes have similar products, although with reduced production of the mass-77 isobars. The only strongly produced contaminant solely from the non-52 Cr isotopes is short-lived 74 Br (t 1/2 = 25 min), which would not interfere with routine production of 77 Br. The PACE4 13,14 fusion-evaporation code was used to predict production cross-sections. Calculations were performed at 10,000 cascades per energy to minimize simulation error, which is on the order of one millibarn. The cross-sections for various relevant radionuclides (bromine and radionuclides which decay to bromine isotopes) are presented in Fig. 1 as a function of laboratory bombarding energy. Production of the mass-77 isobar chain dominates at all energies. 16 O on nat Cu. For the oxygen on copper system, the reactions of interest are multi-nucleon-evaporation reactions on both 63 Cu and 65 Cu: These reactions are all available on natural copper, and again isotopically enriched targets are not required. If a mixed source of 76/77 Br is not desirable, then careful energy selection is required to favor the production of one isotope over the other. In that case, isotopically pure targets of either 63 Cu or 65 Cu could be advantageous. For the purposes of this proof-of-principle study, a mixed source was studied. The Stopping Range of Ions in Matter (SRIM 15 ) program was used to calculate the energy-loss-profile for the beam in the target. The energy loss was then used to calculate the beam energy as a function of target penetration depth, which was then combined with the PACE4 cross sections and a delivered-current profile to produce predicted yields. Figures 2 and 3 show the predicted cross sections and yields for the production of various Rb, Kr and Br isotopes from the 16 O bombardment of nat Cu as a function of laboratory bombarding energy (Figs. 10 and 11 in the supplemental section present similar plots for isotopically pure 63/65 Cu targets). The mass-77 and mass-76 isobars can both be strongly produced.

Separations.
Theoretically, krypton and bromine isotopes can be easily and rapidly separated from one another in vacuo via dry distillation 5 and differential cold-trapping. By carefully selecting when this separation occurs, it is possible to isolate specific radionuclides with high radiopurity. Figure 4 shows the calculated evolution of the number of atoms present of various relevant radionuclides as a function of time after start of irradiation, for a representative set of experimental parameters (1eμA, + 7 charge state, 72 MeV 16  . After the bromine is allowed to decay for some amount of time, it will contain only 76 Br and 77 Br with relatively high radiopurity and a small amount (on the order of 1%) of stable 79 Br. The exact decay time can be varied as needed to select acceptable radiopurities for a given application.
The krypton can simultaneously be allowed to decay for approximately 4 h, during which time 89% (of EOB amount) of the 77 Kr will decay to 77 Br, while 16% of the 76 Kr will decay to 76 Br and 7% of the 79 Kr will decay to 79 Br. Bromine and krypton can then be separated again, producing 76 Br and 77 Br with 51% and 48% respective radiopurity as measured by activity at 4 h EOB. The same procedure can be repeated after further decay to isolate 76 Br from the decay of 76 Kr. The exact decay lengths and radiopurities can again be tuned as needed for a given application.
To independently verify the statistical model predictions for these two proposed reactions mechanisms for radiobromine production, a set of complementary experiments were performed to measure the productions yield experimentally at relevant bombarding energies.   Table 1.
The targets-commercially available, high-purity sputtering targets (Kurt Lesker, 99.999% Cu and 99.95% Cr, 1 inch diameter and 0.125 inch thickness)-were mounted in electrical contact with an approximately 1 foot long, closed, conducting beampipe that was itself electrically isolated from the rest of the experimental set up (see Fig. 5). A constant magnetic field was placed across this section of beampipe to direct any ejected electrons into the wall of the isolated section of beampipe, which allows for accurate charge integration and continuous current monitoring without the need for a suppression voltage. Targets were not cooled.
Activity measurements were conducted at Notre Dame via off-line gamma spectroscopy using a high-purity germanium detector. Detector efficiency was determined using 133 Ba, 60 Co, and 137 Cs sources of known activity. The small (< 5%) dead-time losses were corrected by using the detector live-time in activity calculations. The target from the 28 Si run (run 1) was counted beginning approximately 5-h after end-of-beam (EOB), with 30-min measurements continuously repeated for 10 days. Due to the high activity immediately post irradiation, the 72 MeV 16 O irradiated copper target (run 2) was counted beginning approximately 18-h after EOB, with 30-min measurements continuously repeated for 10 days. Since much of the shorter-lived precursors would  www.nature.com/scientificreports/ have decayed during the cooling period, reported activities (corrected to EOB) for these runs will over-predict the direct production of longer-lived products and late-produced daughters (and consequently under-predict the production of shorter-lived products and precursors) and represent net yields-that is, how much of a given radionuclide can be produced if harvesting is delayed until each decay chain of interest has settled into some long-lived species. Offline counting for the 55 MeV 16 O run (run 3) began approximately 7 min after EOB with 1-min runs, so that over/under-attribution error does not apply to that run. Figure 6a shows a gamma spectrum measured for the 28 Si on nat Cr system 5-h after EOB. The dominant features are peaks from 77 Kr, 75 Br and positron annihilation. Figure 6b shows the same system after 10 further hours of decay. The contaminants and precursors have decayed away, leaving only the 77 Br, which retains 88% of the activity it had in 6a, 83% of EOB. Figure 7 shows a similar scenario for the 72 MeV 16 O on nat Cu system. The first measurements were further from EOB, so the initial precursors and contaminants are not visible, but the system decays such that eventually the desired products represent the majority of the radioactivity. A complete list of radioisotopes identified by gamma spectroscopy for each run is given in the supplemental information (Table 4). Figure 8 shows a spectrum obtained for the 55 MeV 16 O on nat Cu system, which was intentionally run at lower intensity to facilitate earlier gamma spectroscopy. The activity in the spectrum (which was only measured for 1-min, as opposed to the 30-min runs in Figs. 6 and 7) is almost entirely from 77 Kr and positron annihilation. Table 2 lists the runtimes, average currents, and predicted (from PACE and SRIM calculations) and measured yields for each experiment.

Results
It is apparent from Table 2 that the measured production from the 72 MeV 16 O run is far below what was predicted. Given the relatively good agreement between prediction and measurement for the 55 MeV 16 O run and previous experience 11,12 suggesting that PACE is more likely to overpredict than underpredict, it seems unlikely that the predictions were simply off to such a degree from a simple increase in bombarding energy. More detailed examinations suggest the most likely explanation is that krypton gas escaped from the target before and during gamma spectroscopy. Figure 9 shows the decay of two gamma signatures for 76 Kr for the first seven spectra taken for this system. The time constants for the decays correspond to an experimental half-life of 1.5 h. This is off by a factor of ten from the well-measured half-life of 76 Kr (14.8 h), so some additional physical effect on these isotopes must be in play. Since both gamma energies decay in tandem, the cause of the decreased lifetime must be affecting both peaks equally, so it cannot be independent contaminants for each energy. Since no other species have been identified which match this gamma signature, the cause cannot be spectral interference. Therefore, the most likely cause is krypton gas physically leaving the target prior to its radioactive decay. At later times (analyzed independently between t = 50 h and t = 100 h), the half-life of decay for these gamma peaks is measured to be t 1/2 = 14.7 ± 0.2 h, which is consistent with the established lifetime for 76 Kr. These observations are consistent with krypton which was produced near the surface of the target evaporating away, but krypton deeper in the target remaining trapped and decaying in place. In future experiments, this effect might be mitigated by cooling targets during irradiation to minimize evaporative losses.
These effects are not seen in the 55 MeV 16 O run. This is likely a combination of more immediate gamma spectroscopy and lesser beam energy, beam current and irradiation time resulting in substantially less target heating and, by extension, less evaporative loss. SRIM calculations suggest that a 72 MeV 16 O beam in natural copper will reduce to 55 MeV in approximately 6.4 μm, which implies that the increased krypton production would occur closer to the surface for the increased bombarding energy, and evaporation could justify the suppressed 76 Br activity as well.
An attempt was also made to perform an alpha bombardment of an arsenic target as an additional point of comparison, but even with only a few hundred nanoamperes of alpha-beam, target degradation was sufficient to severely impact the accuracy of post-irradiation gamma spectroscopy. Because of the difficultly of reproducing measurements with such a friable target, these data were excluded from this work.

Conclusions
These results demonstrate the utility of making proton-rich isotopes with heavy-ion reaction channels. Simple target preparation and simple radioisotope separation requirements offer advantages over current production methodologies, and do not require radiochemical purification. Although production yields are lower by 100-fold than current proton reaction methods (comparisons shown in Table 3), target robustness means that potentially higher (100-fold) beam currents can be used to achieve comparable yields to current techniques, but with more robust target material and enhanced radiopurity of the desired isotopes. On a per-incident-particle basis, yields from the heavy-ion reactions are suppressed by approximately two orders of magnitude, but much of this shortfall may be offset with increased beam current, with significantly easier targetry and purification processes. The short range (< 25 μm) of the heavy-ion projectiles means that all Targets made from natural copper are also reusable, meaning that targets can be cycled between irradiation and extraction (in vacuo) for routine production. With the observation of evaporative losses from the room temperature target, the possibility of continuous irradiation and on-line separation emerges. Krypton evaporating from the target could be pumped away and cold-trapped, and harvesting could potentially take place without ever having to stop irradiation, access the target or perform separation chemistry with hot targets. With modern ion sources capable of delivering abundant silicon or oxygen beams (see 17 , or a commercially available source, e.g. 18 ) it is feasible to construct a dedicated production facility that could routinely produce amounts of radiobromine suitable for clinical use. Such a facility may also be able to make large-scale production of several other proton-rich radioisotopes economically feasible.    77 Kr produced would have decayed to a level too low to be detected prior to gamma spectroscopy beginning. The measured 77 Br therefore represents both directly produced 77 Br and the product of 77 Kr decay. For both isotopes, the predicted activities are for physical yield at EOB (using the same delivered current profile as the actual irradiation). The effects of simulation error on the predicted yields are negligible compared to variations in delivered beam and are ignored. b The 37.7 kBq of 77 Br represents the predicted EOB activity that would be found if all 77 Kr produced was allowed to decay to 77 Br, and then an EOB activity for 77 Br was determined by backfitting the decay from that point forward without decay-correcting for 77 Kr feeding.  Figure 9. Activity vs time for two gamma signatures of 76 Kr for the first seven spectra taken for this system. The decays are fit with simple exponentials, and the half-life of the decays is found to be 1.5 ± 0.01 h. Table 3. Yield Comparison between proton and heavy-ion reactions (proton data from 5 ). a To account for indirect production, physical yields for the heavy-ion reaction are given for 4 h EOB, to allow ~ 90% of the 77 Kr to decay to 77 Br. b Since 3-h irradiations were not performed, the absolute yields are scaled from physical yield (accounting for saturation and decay effects) and reported per-particle-microamp of beam delivered (the proton data assumed a 40 μA max current for the entirety of the run). The 55 MeV data are based on data reported in this work. The 72 MeV data are a prediction from PACE calculations, since the measured data were likely depleted by krypton evaporation.