Cyclotron production of 225Ac from an electroplated 226Ra target

Purpose We demonstrate cyclotron production of high-quality 225Ac using an electroplated 226Ra target. Methods 226Ra was extracted from legacy Ra sources using a chelating resin. Subsequent ion-exchange purification gave pure 226Ra with a certain amount of carrier Ba. The radium target was prepared by electroplating. We successfully deposited about 37 MBq of 226Ra on a target box. Maximum activation was achieved using 15.6 MeV protons on the target at 20 µA for 5 h. Two functional resins with various concentrations of nitric acid purified 225Ac and recovered 226Ra. Cooling the intermediate 225Ac for 2–3 weeks decayed the major byproduct of 226Ac and increased the radionuclidic purity of 225Ac. Repeating the same separation protocol provided high-quality 225Ac. Results We obtained 225Ac at a yield of about 2.4 MBq at the end of bombardment (EOB), and the subsequent initial purification gave 1.7 MBq of 225Ac with 226Ac/225Ac ratio of < 3% at 4 days from EOB. Additional cooling time coupled with the separation procedure (secondary purification) effectively increased the 225Ac (4n + 1 series) radionuclidic purity up to 99 + %. The recovered 225Ac had a similar identification to commercially available 225Ac originating from a 229Th/225Ac generator. Conclusion This procedure, which involves the 226Ra(p,2n)225Ac reaction and the appropriate purification, has the potential to be a major alternative pathway for 225Ac production because it can be performed in any facility with a compact cyclotron to address the increasing demand for 225Ac. Supplementary Information The online version contains supplementary material available at 10.1007/s00259-021-05460-7.


Introduction
Targeted alpha therapy (TAT), which is a therapeutic regimen by radiopharmaceuticals labeled with alpha emitters, has received great interest due to its clinical impact such as 223 RaCl 2 and 225 Ac-PSMA-617 [1,2]. Compared to conventional therapy by beta emitters, alpha particles exhibit a high linear energy transfer (LET) in a short range (40-100 µm/5-9 MeV, [3]), which may provide a remarkable cytotoxic effect in a limited area of the target. Hence, unwanted radiation doses to other healthy tissues and organs may be limited. Actinium-225 ( 225 Ac, α = 100%, T 1/2 = 9.92 days) is a promising radionuclide applicable to TAT since it breeds multiple descendants via the net decays of 4α + 2β in a relatively short period (Fig. 1). This property enhances the therapeutic effects of 225 Ac-labeled compounds [4].
The accessibility of various radioisotopes for diagnostic nuclear medicine is well established. Both accelerators and nuclear reactors are used to produce medical-grade isotopes. The former is typically used to produce short-lived isotopes, while the latter is used for large-scale, centralized production. A shortage of actinium-225 is anticipated [5] because interest has drastically increased but largescale commercial production is still in the development phase. Currently, the most realistic path is the natural decay product by a 229 Th/ 225 Ac generator system [6,7]. However, only a few institutes such as the Joint Research Centre (JRC, Karlsruhe, Germany), Oak Ridge National Laboratory (ORNL, TN, USA), and the Institute of Physics and Power Engineering (IPPE, Kaluga Oblast, Russia) have such capabilities and their estimated total annual capacity of actinium-225 is approximately 63 GBq (1.7 Ci) [8]. Consequently, alternative production pathways are highly desired.
Many studies have investigated increasing actinium-225 production to meet the anticipated demand. Practical options include (1) high-energy protons on 232 Th (spallation channel, [9]), (2) moderate energy protons on 226 Ra (nuclear transmutation channel, [10]), and (3) high-intensity gammas on 226 Ra (photonuclear/Bremsstrahlung channel, [11][12][13]). Among these, option (1) holds promise because 232 Th is not a fissile material according to nuclear regulations. However, the reaction requires a projectile with an extraordinarily high energy and intensity (e.g., 100 MeV or higher proton). Consequently, few facilities can practically produce 225 Ac via the spallation route. On the other hand, options (2) and (3) use 226 Ra as their target material. Option (2) may be advantageous for actinium-225 production in general facilities. However, this material is difficult to handle due to safety concerns (i.e., radon ( 222 Rn) emanation and high-energy gamma emission from the descendants). For instance, as reported in [10], the reaction of 226 Ra(p,2n) 225 Ac can be performed efficiently with a relatively small amount of 226 Ra in a low-energy window provided by a compact medical cyclotron, Ep ≤ 20 MeV. In this study, we evaluate the production feasibility of 225 Ac from a 226 Ra target that includes (1) 226 Ra recovery from legacy needles, (2) radium target preparation, (3) activation, (4) separation, and (5) recycling of 226 Ra. This study should be useful not only from a production capability viewpoint but also from a quality control perspective for large-scale actinium-225 production.

General
All procedures were performed in a ventilated glove box with a pressure of -50 Pa. No system for 222 Rn handling was installed. A bag-in/out protocol with a polyethylene bag (thickness 100 µm) was employed when transferring samples across the glove box to avoid releasing 222 Rn and other possible radioactive materials into the laboratory [Online . The maximum daily permission for handling 226 Ra in our laboratory is 148 MBq (4 mCi).

Ra recovery from legacy needles
Radium needles (only information available is its size (Ø1.6 × 25 mm) and activity (1-2 mCi-226 Ra/needle)) were sectioned into 5-6 pieces by an ordinary tube cutter (nipper type, for 1/16″ stainless tubes). The pieces, which were collected in a 50-mL glass bottle with a polypropylene screw cap (Duran Wheaton Kimble, Germany), were mixed with 3 mL of a Chelex-100 resin slurry and 7 mL of pure water. The tightly capped bottle was sonicated daily for a period of 1 week to 1 month. Afterward, the Chelex-100 resin was filtered from the mixed materials by an empty cartridge (Bond Elut, 5 mL, Agilent Technologies, CA, USA), where the extracted 226 Ra adsorbed on the resin. Then, 1 M HCl (5 mL) and subsequent pure water (10 mL) for rinsing were loaded into the cartridge to elute 226 Ra from Chelex-100. The eluate was loaded into an anion exchange resin (16 mL, Monosphere) to remove chlorides. Then, the resin was washed with 10 mL pure water. The recovered solution was evaporated at 130 °C under a vacuum, yielding dried 226 Ra in the hydroxide form.
Extending the treatment time improved the recovery efficiency of 226 Ra. In our preliminary experiment, a maximum of 50% of the initial activity was recovered by the 1-week treatment. However, returning the mixture of needle pieces and Chelex-100 residue to the bottle again and repeating the same procedure for a month showed quantitative recovery of the initial activity. In the 1-month experiment, the slurry was adjusted to a pH ≥ 10 by adding a portion of conc. ammonium solution to ensure the Chelex-100 efficiently for absorbing 226 Ra.

Target box design and target preparation
The 226 Ra target was prepared by electrodeposition. Figure 2 shows the target box assembly. A Ti cylindrical cavity (#3 in Fig. 2) with a volume of about 3.5 mL was used for the electroplating reservoir and the dissolving vessel for the Ra target after activation. A Pt rod (Ø3 mm, anode for  ) was held by a polyimide screw with O-rings (#7). The bottom of the Ag cavity (#5) was assembled with both #3 and a polyimide electric insulator (#4). For chemical resistance, it should be noted that the Ra-depositing surface with a conical shape (#5) was fabricated with Au by hot isostatic pressing on the Ag body.
Purified dried 226 Ra, which ranged from 14.8 to 38.9 MBq (400-1050 µCi), was dissolved in 1 mL of 0.1 M HCl and 2 mL of 0.5 M ammonium acetate to prepare the electrolyte. The electrolyte was placed in the target box, and a constant current of 100 mA DC was applied in the pulse mode (5 Hz with a 0.1 s on-off cycle) for 3 h with a 15-mm gap between the cathode and the anode. After the electrodeposition process, the electrolyte was removed from the cavity by pipette work, and introduced in and removed from 2 mL of pure water twice to wash out the residual electrolyte in the target box. These rinsing solutions were collected as they may contain undeposited free 226 Ra. Then, the deposition efficiency was evaluated by measuring the 226 Ra activity. The cavity stood undisturbed overnight (> 15 h) to dry the Ra surface naturally in a ventilated glove box. Eventually, the Pt anode was withdrawn from the cavity and sealed with a thin Nb foil (50 µm). The cavity at the beam entrance was sealed with a 50-µm-thick Nb foil (#2 in Fig. 2) with #1.

Activation
Activations were carried out by 34 MeV H 2 + (ionized molecular hydrogen) provided by NIRS-AVF-930 cyclotron at a nominal intensity of 10 µA for 3-5 h. This condition increased the intensity of the lower energy particles accelerated by a relatively larger cyclotron to give 17 MeV protons at nearly 20 µA by splitting the kinetic H 2 + ion at the vacuum isolation window. The estimated proton energy on the target material by SRIM code [14] was 15.6 MeV after passing through the vacuum foil (Al, 100 µm), the He cooling layer (30 mm), and the target foil (Nb, 50 µm). To enrich the expected 225 Ac yield, the on-target energy of 15.6 MeV was set between two energies showing the highest crosssections for the 226 Ra(p,2n) 225 Ac provided by the ALICE code (ca. 700 mb at 15 MeV) and the previous study (ca. 710 mb at 16.8 MeV) [10]. Figure 3 shows our newly developed separation procedure, which was implemented 3-4 days after the end of bombardment (EOB). The activated target was dissolved in 3 mL of 0.7 M HNO 3 , and the solution was loaded slowly into a DGA cartridge (N,N,N′,N′-tetra-noctyldiglycolamide, 1 mL, Eichrom Technologies, IL, USA). To increase the leftover recovery of Ac/Ra, another 3 mL of 0.7 M HNO 3 was introduced into the target cavity and the rinsing fraction was loaded into the same DGA cartridge. This step was repeated twice.

Separation of 225 Ac from the target matrix
The DGA cartridge was washed with 20 mL of 0.7 M HNO 3 to remove residual 226 Ra. Then, 5 mM HNO 3 (10 mL) was loaded into the DGA to elute 225 Ac, which was the fraction collected in an intermediate reservoir.
Subsequently, the crude 225 Ac fraction was loaded into a LN cartridge (di(2-ethylhexyl)orthophosphoric acid, 2 mL, Eichrom Technologies); the cartridge was washed with 10 mL of 50 mM HNO 3 to eliminate trace amounts of 226 Ra and subsequently well purged. All the above waste fractions were collected as the Ra recovery fraction, which was recycled for the next use. Eventually, 225 Ac was stripped by loading 0.7 M HNO 3 (10 mL) and collected into another intermediate reservoir.
The actinium-225 solution in this separation step contained 226 Ac (β 83%, EC 17%, α 6 × 10 -3 %; T 1/2 = 29.4 h) as a byproduct because it was unavoidably generated via the 226 Ra(p,n)-channel in our activation condition. To increase the radionuclidic purity of 225 Ac, the intermediate product was allowed to cool for 2-3 weeks, which is equivalent to 10 half-lives or more for 226 Ac. After the cooling, the above separation protocol was repeated as the secondary purification.
Although the twice purified 225 Ac was free from or had negligible 226 Ac contamination in 0.7 M HNO 3 (10 mL), it was too acidic for further use. Thus, an anion exchange resin (AG1-X8, 100-200 mesh, Cl form) was employed to exchange the counter anion of 225 Ac with chloride and remove HCl from the product by evaporation of the above sample (130 °C under vacuum). The final product, which had a chemical form of 225 AcCl 3 , was reconstituted in 100 µL of 0.1 M HCl for further use. The recovery yield of 225 Ac was quantitative.

Recycling mode of Ra
All fractions possibly containing 226 Ra were collected into a single vessel. The solution was adjusted to a pH ≥ 10 by adding a conc. ammonium solution and then loading into a column filled with the Chelex-100 resin (0.5 mL, NH 4 form) to concentrate 226 Ra. After washing the column with 10 mL of pure water, 226 Ra was stripped by passing 1 M HCl (5 mL), and the eluant was led to an anion exchange resin column (16 mL, Monosphere, OH form). The Ra fraction was desalted by an anion exchanger, and an additional 10 mL of pure water was loaded to remove the residual 226 Ra. The collected 226 Ra with a volume of about 15 mL was subsequently evaporated at 130 °C under a vacuum to yield purified 226 Ra in the hydroxide form, which was ready for the next use as the electrolyte.

Gamma spectroscopy
Samples were subjected to a HPGe detector coupled with a well-calibrated 4096-ch multi-channel analyzer (39 cm 2 , EGC15-185-R, Eurisys (Mirion Technologies), Lingolsheim, France); RZMCA, Laboratory Equipment, Ibaraki, Japan). A measurement uncertainty of 9% was obtained as a positive square root of the sum of the following contributing factors in quadrature: counting statistics (5%), geometrical error (5%), and detector efficiency (6%). Table 1 lists the decay data for the radionuclides of interest [15]. The detection limit for these nuclides was 3.7 Bq (0.1 nCi) with an acquisition period of 60,000 s or longer, which was equivalent to 1.2 × 10 -3 % of 225 Ac activity in the most sensitive case.

Alpha spectroscopy
An alpha spectrometer (Alpha Duo, Ametek Ortec, Oak Ridge, TN) equipped with an ion-implanted-silicon charged-particle detector (Ultra-AS 450 mm 2 , Ametek Ortec) and 4096-ch pulse-height analyzer (Maestro-32, Ametek Ortec) was used to acquire the alpha spectra. The spectrometer was calibrated with a mixed source ( 148 Ga, 241 Am, and 244 Cm; Eckert & Zeigler, Valencia, CA). An aliquot of the sample (about 370-1 k Bq (10-50 nCi) of 225 Ac) was dropped on an Al sheet, and dried. The prepared sample without a cover was loaded inside the chamber. These samples were analyzed with an acquisition time of 1200 s.

Ra liberation from legacy needles
In most cases, 226 Ra prepared in the radium needle had a form of RaSO 4 . Although ionic compounds are typically water soluble, group II sulfates, including RaSO 4 , are practically insoluble in water. Our samples showed a nearly zero recovery of free 226 Ra 2+ when the Ra matrix was suspended in water or 1 M HCl, providing additional evidence that Ra was in the sulfate form. However, when Chelex-100 was allowed to sit long term, a remarkable recovery occurred. Trace amounts of Ra 2+ were gradually liberated from the sulfate as the chelation sites tightly held Ra 2+ .
In addition, the Ra matrix remained in the small sheath cavity, even though the legacy needle was cut into small pieces. Sonication seemed to crumble the solid Ra matrix, and free Ra was effectively released from the matrix. The recovery rate of 226 Ra was 30-50% after a week but quantitative after a month.

Electrodeposition and activation results of 226 Ra
The deposition yield of 226 Ra was satisfactory. The difference of the 226 Ra activity in the electrolyte between the initial and the post-deposition indicated that the yield ranged from 94 to 97% (Table 1). The deposited 226 Ra layer, which contained some amount of carrier Ba, was practically insoluble upon washing after the deposition process. The washing fractions of pure water showed a very small activity of 226 Ra (1-2% of the initial value).  Figure 4 shows the electrodeposition profiles of 1-mCi of 226 Ra (ca. 4.5 µmol) with an unknown amount of carrier and 5-mg of Ba (ca. 36 µmol) as an increasing challenge/evaluation for Ra deposition using a Ba surrogate. Due to safety regulations for 226 Ra handling and our limited Ra inventory, this study involved a small amount of 226 Ra. Consequently, we performed a cold experiment with Ba instead of 226 Ra to evaluate the electrodeposition performance based on the widely accepted chemical similarity between Ra and Ba. In the case of a small amount of Ra, 13.5-37.5 MBq (366-1010 µCi) in this study, a spot-like deposition profile appeared across the cathode surface (Fig. 4a), whereas a condensed layer covered from the top of the cone to the middle of the slope in a Ba rich condition (Fig. 4b).
These results can be explained by the principle of electrodeposition. The gap between the two electro-rods around the top of the cone was the shortest and formed the best condition for electroconductivity in this system. Consequently, effective deposition was expected around the top, and most of the solute (Ra or Ba) tended to be centralized on the surface. These results are encouraging evidence to obtain rich and high-density Ra depositions when there is sufficient Ra in the electrolyte.
We obtained 225 Ac at a yield of 522-2430 kBq (14-66 µCi, decay corrected) from a 226 Ra target of 13.5-37.5 MBq (366-1010 µCi) irradiated by 20 µA protons for 3-5 h (Table 1). A previous study employing a 226 Ra(p,2n)-channel reported the production yield of 225 Ac with various amounts of 226 Ra [10]. Although it is difficult to directly compare the previous study and our data due to the different conditions (amount of 226 Ra prepared, chemical form, target area, target thickness, beam energy, beam intensity, and irradiation time), the previously reported yield for 225 Ac/ 226 Ra (% activity) was 13-44% without considering the activation condition. By normalizing, this reported yield of 225 Ac was converted ca. 0.10-0.36 nCi-225 Ac/µg-226 Ra/ (µA h) (except for the result of 12.5 µg-Ra). Our results showed 0.63-0.65 nCi-225 Ac/µg-226 Ra/(µA h) by applying the same correction. Considering the above-mentioned differences, both production yields are within an acceptable range, suggesting that our production capacity is similar to that of the previous study. Hence, our data support the feasibility of 225 Ac production from an electroplated elemental 226 Ra target. In addition, the above production index demonstrates the good stability and repeatability of our target activation capability.

Separation
The initial separation of the 225 Ac sample contained 226 Ac and other radionuclidic impurities (Fig. 5a). Similar to 226 Ra, 226 Ac is a 4n + 2 series radionuclide, which generates many descendants during the cooling period (Fig. 1). Hence, repeated separation as a secondary purification removed the 4n + 2 impurities to yield high-quality 225 Ac. Although 224 Ac (EC 91%, α 9%, T 1/2 = 2.8 h) should be co-produced via the 226 Ra(p,3n)-channel (E TH = 13.6 MeV), the half-life of 224 Ac was too short to be detected at the end of separation at 4 days from EOB. However, the major distributions in the washing fraction and leftovers of the separation materials were a couple of 224 Ac descendants with favorable gamma emissions in the 4n series, 212 Bi (T 1/2 = 61 min, 727 keV, 6.7%) and 208 Tl (T 1/2 = 3.1 min, 2615 keV, 99%). Moreover, trace amounts were detected in the purified actinium-225 sample, providing evidence for 224 Ac generation. The presence of 212 Bi and 208 Tl in the Ac fraction was acceptable because Bi was partially similar to Ac in our separation conditions. As a result, other Bi isotopes, 214 Bi (originating from 226 Ra) and 213 Bi (a descendant of 225 Ac), should also be found in the initial actinium-225 fraction. Orphan 214 Bi should decay upon additional cooling (Fig. 5b). On the other hand, 212 Pb (T 1/2 = 10.6 h, 239 keV, 44%), the parent nuclide for 212 Bi, was not detected in the purified 225 Ac samples. All the 4n series-nuclides with the potential to be the parent for 212 Pb ( 224 Ac-216 Po) had shorter half-lives than 212 Pb. The exception was 224 Ra, which was removed along with 226 Ra. Hence, only the 4n + 2 series was considered the byproduct in the separation process.
Other notable byproducts were 135 La (EC, T 1/2 = 19.5 h) and 140 La (β, T 1/2 = 1.68 days). The former presumably originated from a carrier of natural Ba in the legacy Ra needle via the 135 Ba(p,n)-channel. However, the half-life of 135 La is much shorter than that of 225 Ac. Thus, an appropriate cooling time should gradually decrease the impact of 135 La on the 225 Ac even though the legacy Ra was not chemically purified. On the other hand, since the heaviest stable isotope of Ba is 138 Ba, the atomic mass of 140 La was too rich to be generated by proton activation, suggesting that fission on 226 Ra may occur in our activation condition. In addition, 140 Ba (β, T 1/2 = 12.6 days), which is a parent nuclide for 140 La, could also be generated as another fission product. Unfortunately, we were unable to directly confirm the presence of 140 Ba because most of the characteristic gamma lines for 140 Ba were close to those for 214 Bi (RaC) and the chemical similarity between Ba and Ra. However, the initial separation of the actinium-225 fraction would practically eliminate 140 Ba along with 226 Ra due to chemical similarity. Indeed, orphan 140 La in the actinium-225 fraction showed an acceptable half-life of 1.67 ± 0.10 days and decayed to a non-detectable level on the gamma spectrum by cooling for 2-3 weeks. This finding suggests that the carrier Ba in Ra needles does not affect the quality of 225 Ac, and Ra purification from carrier Ba does not provide a practical advantage. A discussion on the counter fragments is available in [Online Resource 4].
For example, we cooled the samples for 19-20 days after EOB or 2 weeks from the end of separation. The spectra of the cooled samples were similar to that of authentic 225 Ac originating from a 229 Th/ 225 Ac generator (Fig. 5b and c). The alpha spectrum of our 225 Ac product also showed the same profile as the reference (Fig. 6). Notably, neither 226 Ra (Eα = 4.78 MeV, 94%) nor 210 Po (Eα = 5.30 MeV, 100%) was detected. Hence, the double separation with an appropriate cooling period gave pure 225 Ac with a quality comparable to generator-made 225 Ac.

Recovery of 226 Ra for recycling
We developed a closed circuit for Ra recycling to minimize the loss of the 226 Ra inventory (Figs. 3 and 7). This process effectively reduced long-lived radioactive wastes. After single-runs of this circuit with a 37-MBq (1-mCi) 226 Ra, we evaluated the 226 Ra leftover in each measurable material (i.e., the Ra recovery fraction), separation material (cartridge), and reservoir. The Ra recovery fraction contained 90-98% 226 Ra, and other materials were negligible. For example, < 37-74 kBq (1-2 µCi) of 226 Ra activity was found in the respective materials, depending on the volume of the residual liquid presented in the small voids. In addition, any deposition/leftover on the target box could not be estimated due to its high radioactivity. This can explain the ~ 10% discrepancy in the activity distribution. Since the target box was used repeatedly, the practical loss of 226 Ra should be negligible.

Radionuclidic purity-impact of 226 Ac in 225 Ac
Actinium-226 disintegrates by pathways of β decay (83%), EC (17%), and trace α decay (6 × 10 -3 ), which breed 226 Th (α, T 1/2 = 30.6 min), 226 Ra, and 222 Fr (β, T 1/2 = 14.2 min), respectively. Among these radioisotopes, 226 Ra, which has the longest half-life, may be concern when preparing 225 Aclabeled injections. Although radiation safety related to both 226 Ac and 226 Ra is beyond the scope of this study, the amount of 226 Ra generation as a decay product of 226 Ac should be considered in discussions on clinical applications or the appropriate cooling time to address the acceptable quality on the actinium-225 product.
Actinium-226 with activities of 3.2 MBq (86 µCi) and 0.24 MBq (6.4 µCi) may generate 1.13 Bq and 0.084 Bq of 226 Ra, respectively. These values are equivalent to 226 Ra in the human body estimated (31 pg, where 27 pg is accumulated in the skeletal system) as well as those from dietary intake (2.3 pg/day) [16]. According to the reference [2], a maximum realistic single dose of actinium-225 injection is around 10 MBq (100 kBq/kg). Therefore, a 226 Ac/ 225 Ac sample in a 10-MBq actinium-225 injection may generate 226 Ra equivalent to the above-mentioned reference.
While the targeted alpha therapy field has yet to establish an acceptable limit for 227 Ac (β 98.6%, α 1.38%; T 1/2 = 21.8 y) [17], which is a major byproduct in the spallation pathway from 232 Th target, 227 Ac should be a good reference for rational considerations for long-life alpha-emitting byproducts. A previous study reported that 227 Ac was equivalent to 0.7% of 225 Ac at the time of injection (radioactivity-based estimation) [18]. One study reported that the 227 Ac contribution to the radiation dose delivered by an actinium-225 injection was negligible [19]. However, the presence of 227 Ac may pose a waste disposal issue. A recent study demonstrated the production feasibility of 225 Ac without 227 Ac ( 227 Ac/ 225 Ac = < 7.5 × 10 -5 %) using 225 Ra (β 100%; T 1/2 = 14.9 d), which is another spallation product from 232 Th, as the parent nuclide [17]. This report referenced the exemption activity recommended by the International Atomic Energy Agency (IAEA) [20] for a rational discussion about the long-life radionuclide.
Although our production yields were smaller than the assumed single dose of 10-MBq 225 Ac, the activity ratio of 226 Ac/ 225 Ac in our samples ranged 1.4-2.3% at the end of the initial separation (4 days from EOB). The ratio was comparable to the potential amount for 226 Ra delivery via dietary intake ( 226 Ac/ 225 Ac = 2.4%). The 226 Ac content of around 1-2% (100-200-kBq 226 Ac/10-MBq 225 Ac) was higher than the exemption activity for 226 Ac (100 kBq, [20]) but additional cooling for several days would gradually decrease the 226 Ac activity below the exemption. In addition, the potential 226 Ra amount in our purified samples at any time was much lower than the exemption activity (10 kBq, [20]), suggesting that radiation risks caused by 226 Ac would be negligible or exceedingly small, if the 226 Ac/ 225 Ac ratio is close to the range of our results.
As shown in Fig. 5b, secondary separation effectively increased the radionuclidic purity of 225 Ac, which reached > 99% within 2-3 weeks. In the future, we plan to evaluate the biological impact of 226 Ac since the physical decay loss of 225 Ac during the cooling period is a critical issue in the actinium-225 industry.

Conclusion
Actinium-225 purified from an electro-deposited 226 Ra target with two separation columns showed an acceptable quality without byproducts. The characteristics of the purified 225 Ac were similar to those of commercially available 225 Ac originating from a generator system. The production results showed a linear increase in the 225 Ac yield by increasing 226 Ra prepared.
Consequently, increasing the 226 Ra, beam intensity, or irradiation period can achieve the clinical requirement of 225 Ac yield, demonstrating that the proposed production method may be a viable alternative pathway to address the increasing demand for actinium-225.