Production scheme for diagnostic-therapeutic radioisotopes by accelerator neutrons

Abstract

Interest has been growing in the development of medical radioisotopes used for noninvasive nuclear medicine imaging of disease and cancer therapy. Especially the development of an alternative production scheme of ⁹⁹Mo, the mother radioisotope of ^99mTc used for imaging, is required, because the current supply chain of the reactor product ⁹⁹Mo is fragile worldwide. We have proposed a new production scheme of ⁹⁹Mo as well as therapeutic radioisotopes, such as ⁶⁴Cu and ⁶⁷Cu, using accelerator neutrons provided by the ^natC(d,n) reaction. Based on this scheme we have obtained high-quality ^99mTc, ⁶⁴Cu, and ⁶⁷Cu suitable for clinical use by developing both production and separation methods of the radioisotopes. We proposed a new facility to constantly and reliably produce a wide variety of high-quality, carrier-free radioisotopes, including ⁹⁹Mo, with accelerator neutrons. We report on the development of the proposed scheme and future prospects of the facility toward the domestic production of medical radioisotopes.

1. Introduction

More than 150 different radioisotopes (RIs) have been used in medical, industrial, research as well as commercial applications, such as food processing, agriculture, and structural safety.¹⁾ In these applications, the radioindicator tracer method, which was discovered by George de Hevesy, has been playing an important role.

1.1. Discovery of a radioindicator tracer concept.

George de Hevesy, known as the founder of radioanalytical chemistry and nuclear medicine, published a paper in 1913 claiming that radioindicator tracers play a unique role in chemical analysis.²⁾ The concept of radioindicator tracer chemistry was born through a talk with Ernest Rutherford.³⁾ He encouraged Hevesy to separate the natural radioisotope of lead-210 (²¹⁰Pb) from its admixture with large amounts of nonradioactive lead stored at that time in Rutherford’s laboratory. Hevesy used chemical methods for such separation. After failing at this, Hevesy had an idea to use ²¹⁰Pb as an indicator of lead. Hevesy studied the transfer of lead from soil in different parts of bean plants using ²¹⁰Pb, which was the first application of the radioactive tracer technique to biology.³⁾ The extreme sensitivity of physical radioassay methods allowed him to carry out these experiments with such minuscule concentrations of lead so as to avoid its toxic properties. In 1935 Hevesy investigated the distribution and kinetics of the exchange of phosphorus in different parts of animals using the ³²P radioisotope, which was the first radioindicator study in life sciences.³⁾

The essence of the radioindicator tracer method is that radioactivity is of such minute quantities that it will not cause any toxic affect on the system, and can be detected with high sensitivity by discriminating any conceivable background events as low as possible. A radiolabeled tracer allows noninvasive measurements of the distribution and function in a biological system. Thus, a door to so far unexplored nuclear medicine had been opened.

1.2. Medical radioisotopes for diagnostics and therapy in nuclear medicine.

Medical RIs in the form of radiopharmaceuticals have been used for noninvasive diagnostic imaging studies of diseases and therapeutic applications in cancer treatments, resulting to account for the majority of the applications of RIs.¹⁾ Recently, an approach involving the merging of therapeutic and diagnosis (imaging) treatments, called the “theranostics” approach, has become very important to make a personalized medicine treatment for a specific patient.⁴⁾ Personalized medicine is a form of medicine that uses information about a patient’s own genes or proteins to prevent, diagnose, or treat some disease. A personalized medicine treatment in nuclear medicine is being made using theranostic radioisotopes in the same patients as follows. Firstly low-dose molecular imaging is performed to obtain necessary pre-treatment information concerning the patient’s biodistribution and dosimetry in patients, and secondly higher-dose target molecules therapy is carried out.⁵⁾ In order to promote this theranostics approach while aiming at personalized medicine treatment, the development of a variety of medical radioisotopes for the imaging and therapy of various diseases (especially cancers) is essential. The term “theranostics” was coined by Funkhouser.⁶⁾

Radioisotope imaging is divided into two modalities based on the type of decay and resultant particle emission and detection.⁴⁾ Single-photon emission-computed tomography (SPECT) is performed using γ-ray-emitting radioisotopes with energy below about 400 keV, which can penetrate the patient’s body and be detected by a gamma camera. Technetium-99m (^99mTc) with a half-life (T_1/2 = 6 h) is the most widely used radioisotope in diagnostic imaging studies with SPECT.⁴⁾ Positron emission tomography (PET) imaging is the second-most common imaging modality with the use of a positron (β⁺)-emitting radioisotope. Fluorine-18 (¹⁸F) with T_1/2 = 1.8 h has excellent nuclear properties for a PET radioisotope to provide optimal resolution, and an appropriate half-life for making radiopharmaceuticals.⁴⁾

Cancer treatments have been achieved by performing surgery, chemotherapy, radiation therapy, immunotherapy, targeted therapy, and hormone therapy etc. Radiation therapy is carried out using a certain amount of radiation doses to kill cancer cells and shrink tumors. It is divided into external radiation therapy, which is performed using such particles as X-rays, electrons, protons, and carbons, and internal radiation therapy which is carried out using therapeutic radiopharmaceuticals, including radioisotopes and pharmaceuticals, by employing techniques involving either brachytherapy or radioimmunotherapy.⁵⁾ In brachytherapy, radioisotopes are placed inside or near tumors of patients to treat cancer cells. Note that growing research activities include combining radiopharmaceuticals with conventional treatments, such as chemotherapy and external beam radiotherapy, that would need a variety of medical RIs to form effective radiopharmaceuticals.

Medical RIs used for internal radiation therapy should have the following physical and chemical properties⁴⁾: firstly, physical properties, such as the physical half-life (T_1/2), particle decay mode (β^±-ray, Auger electron, γ-ray, α-particle emission), and the energy of the emitted particle are important. The half-life of RIs must match with the pharmacokinetics of the radioactive drug for uptake and clearance from normal versus targeted (disease site) tissues. It is a crucial point for any radiopharmaceuticals to maximize the dose to the target tissue and minimize the dose to normal tissue. An excellent medical radioisotope used for diagnostics and therapeutics has a short half-life, typically within several hours, like ¹⁸F, and within ten days, respectively. Secondly, the chemical properties should be suitable for allowing the radioisotope to be incorporated into all sorts of molecules, including such as protein compounds for labeling the RIs with pharmaceuticals having high radiochemical yields and a high radionuclide purity. Thirdly, we have to prepare carrier-free RIs (without any other isotope of a sample nuclide) with high specific activity, because a typical activity of a ^99mTc radiopharmaceuticals solution administered to a patient is as high as about 740 MBq/(a few ml).⁴⁾ Note that the specific activity, which is defined as the activity per quantity of the atoms of a particular radionuclide, is usually given in units of Bq/g.

1.3. Production of medical radioisotopes in reactors and accelerators.

Most of medical radioisotopes are produced in reactors or accelerators. In reactors neutron-rich RIs are produced via the fission reaction of ²³⁵U or the thermal neutron capture reaction by a sample nucleus. The specific activity of RIs produced by the former reaction is expected to be high because carrier-free RIs can be obtained by a chemical separation process owing to the fact that an atomic number of the RIs differs from that of ²³⁵U. The specific activity of the RIs produced by the latter reaction is mostly very low because an atomic number of the RIs is the same as that of a sample, and therefore the RIs are not possible to be chemically separated from the sample. Neutron-rich nuclei decay by emitting β⁻-rays and γ-rays, and therefore radiopharmaceuticals containing β⁻-ray emitting RIs are used for radioimmunotherapy (RIT), and those containing γ-ray emitting RIs are used to diagnose the dynamics in a living body via SPECT. In accelerators, a variety of carrier-free RIs including many proton-rich light RIs, such as ¹⁸F and ¹¹C, are produced using mostly proton beams.⁵⁾ These RIs decay mostly by emitting β⁺-rays or alpha particle, which are used in imaging and RIT, respectively. Typical medical RIs produced in reactors and accelerators are listed in Table 1.⁷⁾

Table 1. Radioisotopes typically used for nuclear medicine and their production routes. Here, r, g, and a in brackets indicate reactor, generator, and accelerator

Imaging: conventional single photon emission tomography SPECT, SPECT/CT	Imaging: positron emission tomography (PET,PET/CT,PET/MRI)	Therapy, pain palliation, radioimmunotherapeutics
67Ga(a)	18F(a)	67Cu(a)	131I(r)
99mTc/99Mo(r)(g)	61Cu(a)	89Sr(r)	153Sm(r)
111In(a)	64Cu(a)	89Zr(a)	169Er(r)
123I(a)	68Ga/68Ge(a)(g)	90Y/90Sr(r)(g)	177Lu(r)
131I(r)	82mRb/82Sr(a)(g)	90Y(r)	186Re(r)
133Xe(r)	89Zr(a)	117mSn(r)	188Re/188W(r)(g)
201Tl(a)	124I(a)	123I(a)	alpha emitters, e.g. 213Bi/225Ac(g)(a)

1.4. Challenges in developing medical radioisotopes.

Currently there are important challenges in the productions of medical RIs used for diagnosis and therapy, respectively. Firstly, the supply chain of ⁹⁹Mo (T_1/2 = 66 h) is vulnerable and unreliable, which has frequently caused a shortage of ⁹⁹Mo worldwide since 2008.⁸⁾ More than 80% of all diagnostic procedures in the world are carried out using ^99mTc obtained from a ⁹⁹Mo/^99mTc generator. In Japan approximately 0.9 million procedures/year (about 2,750 procedures/day) are performed using ^99mTc.⁹⁾ Japan imports all of its required ⁹⁹Mo several times per week. Therefore, a reliable and constant supply of ⁹⁹Mo is a key issue for ensuring the routine application of ^99mTc worldwide. About 95% of ⁹⁹Mo is produced by the fission reaction of enriched ²³⁵U in several aging research reactors around the world. The vulnerable situation of the supply chain of ⁹⁹Mo is the impetus to study alternative methods for producing ⁹⁹Mo and/or ^99mTc in reactors or accelerators worldwide.⁸⁾ In fact, strong efforts have been undertaken to develop an alternative production route of ⁹⁹Mo. Until now, as far as we know, no alternative method has yet succeeded to secure a constant supply of ⁹⁹Mo. Secondly, in order to promote a theranostics approach aimed at a personalized medicine treatment for a specific patient, the development of theranostic radioisotopes is highly required. Recent great success in the development of therapeutic RI of ¹⁷⁷Lu (in reactors)¹⁰⁾ and ²²⁵Ac (decay product of ²²⁹Th of nuclear waste in reactors)¹¹⁾ has been prompting further development of a wide variety of medical RIs by innovative production methods.

This review focuses on our studies, started in 2010, that aim to develop a new scheme of medical radioisotopes production and separation to secure a reliable supply chain of ⁹⁹Mo for domestic use and to proceed the theranostics approach while aiming at personalized nuclear medicine. We discuss ^99mTc imaging and the supply chain of ⁹⁹Mo in chapter 2. The new production scheme of medical radioisotopes, such as ⁹⁹Mo, ⁶⁴Cu and ⁶⁷Cu, using accelerator neutrons is introduced in chapter 3. The development of separation methods of ^99mTc from ⁹⁹Mo, and ⁶⁴Cu and ⁶⁷Cu from Zn samples are given in chapter 4. The deuteron accelerator and the accelerator neutron source are described in chapter 5. Conclusions and future prospects are given in chapter 6.

2. ^99mTc diagnostic imaging and ⁹⁹Mo supply chain

The excited state of ^99mTc at 143 keV is populated by the decay of ⁹⁹Mo, as shown in Fig. 1.^{12) 99m}Tc has physical and chemical properties suitable to perform a diagnostic procedure.⁴⁾ First, a short half-life of ^99mTc (T_1/2 = 6 h) allows one to use a large quantity of ^99mTc activity with a low radiation dose to a patient for obtaining clear imaging. Second, 141 keV γ-rays emitted from ^99mTc to the ground state of ⁹⁹Tc are detected by a gamma camera with a high detection efficiency with a low-energy collimator. Third, ^99mTc has a versatile chemistry that allows it to be incorporated into all sorts of molecules. Fourth, ^99mTc is routinely produced in ⁹⁹Mo/^99mTc generators over a period of about one week.

Fig. 1.

Partial decay scheme of ⁹⁹Mo (T_1/2 = 66 h) to ^99mTc (T_1/2 = 6 h). Parenthesis: transition intensity (%). Parenthesis in square brackets: transition intensity (%) from the 181 keV state to the 141 keV one. (2004) Table of Radioactive Isotopes. Version 2.1. available on http://ie.lbl.gov/toi/index.asp (Ref. 12).

2.1. Discoveries of Tc and a Mo/Tc generator.

Tc is an unstable element. In 1937, an artificial element, Tc (^99mTc), was discovered by E. Segrè and C. Perrier in a sample obtained by bombarding a natural Mo sample with deuterons in the 37-inch cyclotron of the Berkeley Radiation Laboratory (now Lawrence Berkeley National Laboratory).¹³⁾ A radioactive Mo species (T_1/2 = 66 h) was obtained, identified as ^99mTc.¹⁴⁾ Note that a cyclotron was invented by L. O. Lawrence in 1929–1930.¹⁵⁾ In order to separate element-43 from Mo metal a thermo-separation (sublimation) method, discussed later, was first employed by Perrier and Segrè.

A ⁹⁹Mo/^99mTc generator, currently used to elute ^99mTc from ⁹⁹Mo, was developed by W. Tucker and M. Greene at Brookhaven National Laboratory in 1958.¹⁶⁾ When they developed a system to separate ¹³²I (T_1/2 = 77 h) from its parent ¹³²Te (T_1/2 = 2.3 h), aiming at medical studies, they happened to detect a trace contaminant, which was later proved to be ^99mTc. It was then realized that ^99mTc was coming from its parent, ⁹⁹Mo. Similarities between the chemistry of the tellurium-iodine parent-daughter pair and the molybdenum-technetium pair led to the development of a ⁹⁹Mo/^99mTc generator system.

Note that Tc, presumably ⁹⁹Tc with a relatively short half-life of 2.1 × 10⁵ y, compared to the age of the universe of 13.7 billion years, was discovered by P. W. Merrill at the surface of red giant stars in 1952.¹⁷⁾ The observation of Tc provided the first powerful evidence that Tc has been synthesized recently within stars. This was a great contribution for understanding the origins of elements heavier than iron observed in stars.

2.2. Production of ⁹⁹Mo and supply chain of ⁹⁹Mo.

About 95% of ⁹⁹Mo is produced by the fission reaction of enriched ²³⁵U [hereafter: fission-⁹⁹Mo] in a limited number of aging research reactors. In around 2008, just before the shortage of ⁹⁹Mo began, five aging nuclear research reactors in The Netherlands, Australia, South Africa, Belgium, and France contributed to meet nearly all (about 95%) of the world’s supply of ⁹⁹Mo/^99mTc.⁸⁾ Between 2008 and 2010 a supply crisis of ⁹⁹Mo, caused by repeated shutdowns, resulted in many diagnostic tests being cancelled or delayed.⁸⁾ The incidents of the reactors highlighted shortcomings and unreliability in the supply of ^99mTc. Currently, ⁹⁹Mo has been mostly produced in seven research reactors, listed in Table 2.^18),19)

Table 2. Current irradiation facility for producing ⁹⁹Mo around the world. A ‘6-day curie’ is defined as the amount of ⁹⁹Mo activity left six days after the generator has left the producer’s facilities.¹⁹⁾ HEU and LEU indicate highly enriched uranium and low enriched uranium used in the fuel of reactors

Reactor	Operating days/year	99Mo production weeks/year	Available capacity/year (6-day Ci 99Mo) by 2024	Fraction (%)	Expected end of operating
BR-2 (HEU) Belgium	147	21	136,500	17	2026
HFR(LEU) the Netherlands	275	39	241,800	30	2026
LVR-15(LEU) Czech Republic	210	30	90,000	11	2028
MARIA(LEU) Poland	200	36	79,200	10	2040
OPAL(LEU) Australia	300	43	92,450	12	2057
RA-3(LEU) Argentina	230	46	23,000	3	<2027
SAFARI(LEU) South Africa	305	44	130,700	16	2030
99Mo demand	46,800		793,650

In addition to concerns related to aging reactors, it is worth noting that the global supply chain of ⁹⁹Mo for the ⁹⁹Mo/^99mTc generator production is complex, as shown in Fig. 2.²⁰⁾ Namely, after being produced in reactors, ⁹⁹Mo is transferred to a processing facility (as listed in Table 3) to be chemically separated and purified. The finished ⁹⁹Mo product material is then isolated and shipped to one of eight generator-manufacturing facilities, located in different countries, that supply ⁹⁹Mo in the form of a ⁹⁹Mo/^99mTc generator to end users, such as nuclear pharmacies and hospitals. In addition, the time frame to deliver purified ⁹⁹Mo to generator manufacturers is limited because of the short half-life of ⁹⁹Mo (T_1/2 = 66 h). These complex supply chains would cause transportation obstacles, such as customs, government regulations, flight schedules, weather delays, and pilot refusal. Natural disasters also have the potential to result in significant product shipment delays. A volcano that erupted in Iceland also reminded us of the risk of relying on a small number of ⁹⁹Mo production facilities in the world. There are currently only five ⁹⁹Mo-processing facilities in the world, as listed in Table 3, in which three of them will end their operations within a decade.^18),19)

Fig. 2.

Schematic view of the supply chain of ⁹⁹Mo from producers to users. At a processing facility ⁹⁹Mo is chemically separated and purified from a fission product of ²³⁵U. IAEA, Belgium, INIS-BE-10K0001, 2010 (Ref. 20).

Table 3. Current ⁹⁹Mo processing facilities in the world

Reactor	99Mo production weeks/year	Available capacity/year (6-day Ci 99Mo) by 2024	Fraction (%)	Expected end of operating
ANSTO Health Australia	43	92,450	14	2057
CNEA Argentina	46	23,000	3	<2027
IRE Belgium	49	171,500	25	2028
Curium the Netherlands	52	260,000	38	Not known
NTP South Africa	44	130,700	19	2030
Total		677,650

2.3. Action by Organisation for Economic Cooperation and Development (OECD).

Under such a circumstance of fragility of the ⁹⁹Mo supply chain, the OECD Nuclear Energy Agency (NEA) has started global efforts to ensure reliable supplies of ⁹⁹Mo and ^99mTc.²¹⁾ The NEA Steering Committee for Nuclear Energy established the High-level Group on the Security of Supply of Medical Radioisotopes (HLG-MR) in 2009 to strengthen the reliability of ⁹⁹Mo and ^99mTc supply in both the short (between 2010 and 2017), medium (2017–2025) and long terms (after 2025). Among the actions requested by the HLG-MR was a review of other potential methods for producing ⁹⁹Mo in reactors and accelerators.

The OECD efforts have succeeded to significantly improve the global demand for ⁹⁹Mo so as to meet at a near-to-full-service level. However, the fragility of the current production chain of ⁹⁹Mo has been remaining, as mentioned above.¹⁸⁾ This vulnerable situation is an impetus to study alternative methods for producing ⁹⁹Mo and/or ^99mTc in reactors or in accelerators worldwide. In fact, many efforts are being made for the production of ⁹⁹Mo or ^99mTc worldwide. Typical proposed reactions are the (n,γ) reaction on ⁹⁸Mo or natural Mo samples in reactors and the ¹⁰⁰Mo(p,pn)⁹⁹Mo, ¹⁰⁰Mo(d,p2n)⁹⁹Mo, ¹⁰⁰Mo(p,2n)^99mTc, ²³⁵U(n,fission)⁹⁹Mo, ⁹⁸Mo(n,γ)⁹⁹Mo, ²³⁸U(γ,fission)⁹⁹Mo and ¹⁰⁰Mo(γ,n)⁹⁹Mo reactions in accelerators.²²⁾ Reaction routes reported in the OECD report are shown in Fig. 3.²³⁾

Fig. 3.

Typical examples of alternative production methods of ⁹⁹Mo and ^99mTc. OECD. Nuclear Development (2010) The Supply of Medical Radioisotopes: Review of Potential Molybdenum-99/Technetium-99m Production Technologies. OECD. p. 17 (Ref. 23).

3. New production routes of medical radioisotopes using accelerator neutrons

3.1. ⁹⁹Mo production via the ¹⁰⁰Mo(n,2n)⁹⁹Mo reaction.

The current route of fission-⁹⁹Mo provides a large amount of ⁹⁹Mo with a high specific activity of ∼370 TBq/(g ⁹⁹Mo) in a single research reactor. Hence, a high specific activity of ^99mTc used for formulating ^99mTc radiopharmaceuticals is obtained from ⁹⁹Mo using a commercially available ⁹⁹Mo/^99mTc generator. Such highly specific activity of fission-⁹⁹Mo is provided by high neutron flux of reactors, a large target volume of enriched ²³⁵U, and the high probability of the nuclear fission reaction of ²³⁵U. However, a specific activity of ⁹⁹Mo based on any alternative routes, other than that of fission-⁹⁹Mo, is as low as about 1/5,000 of the fission-⁹⁹Mo. In fact, all alternative ⁹⁹Mo production routes would face the challenges of lower reaction rates and lower specific activity, which must be overcome by fundamental technical breakthroughs. So far, a variety of nuclear reactions have been proposed to produce ⁹⁹Mo in accelerators, as discussed in chapter 2. However, a production scheme of medical radioisotopes with fast neutrons from an accelerator (hereafter accelerator neutrons) without using a fissionable element U sample had not yet been considered.

In 2009, we proposed a new route to produce ⁹⁹Mo by the ¹⁰⁰Mo(n,2n)⁹⁹Mo reaction using accelerator neutrons provided from an accelerator.²⁴⁾ In proposing a scheme of ⁹⁹Mo production using this route, we considered the following points. Firstly, any scheme of ⁹⁹Mo production is required to meet all or a significant part of the domestic demand of ⁹⁹Mo from an economic view point. We have kept in mind that because the requirement might be hardly met using existing accelerators, one has to propose the production scheme by taking into account the cost for an infrastructure, including an accelerator. Secondly, the safety and efficacy of the ^99mTc radiopharmaceutical preparation based on the proposed scheme should be ensured. Thirdly, we also took into account a criterion concerning the potential for other isotopes co-production at the same time, which was introduced in the OECD report.²³⁾ It is considered to provide an indication of the economic sustainability, demand risk mitigation and the ability to avoid creating some other isotope shortage.

3.1.1. Experiments and evaluations of the ¹⁰⁰Mo(n,2n)⁹⁹Mo reaction product.

In order to meet these requirements, the activity of ⁹⁹Mo produced by a single accelerator should be as high as possible by considering the key parameters, such as the nuclear reaction cross section, beam flux, energy, irradiation time, number of sample nuclei, and half-life of the radioisotope, as discussed next. Firstly, the cross section of the ¹⁰⁰Mo(n,2n)⁹⁹Mo reaction has been measured²⁵⁾ and evaluated²⁶⁾ by many groups at a neutron energy of around 8 MeV up to 20.5 MeV. The ¹⁰⁰Mo(n,2n)⁹⁹Mo reaction cross section is the largest one (except for an elastic scattering cross section) in the neutron-induced reaction of ¹⁰⁰Mo, about 1.5 barn in the neutron energy (E_n) between 12 and 20 MeV, as shown in Fig. 4.²⁶⁾ The cross sections of the ¹⁰⁰Mo(n,α), ¹⁰⁰Mo(n,3n), and ¹⁰⁰Mo(n,p) reactions for producing impurity radioisotopes are less than a few mb at E_n ∼ 14 MeV. Hence, ⁹⁹Mo could be produced with a minimum level of radioactive waste. Secondly, high-flux accelerator neutrons are expected to be provided by a neutron source based on the deuteron breakup reaction of light elements, such as carbon and beryllium etc. In fact, owing to the progress in accelerator technology, neutrons with a high flux of 10¹⁵ n/s having a most probable energy of 14 MeV are produced by the ^natC(d,n) reaction using 40 MeV, 5 mA deuterons at SPIRAL2 in GANIL in France.²⁷⁾ The fluxes are compared with the thermal neutron flux of the reactor at Oak Ridge National Laboratory, having a factor × 10¹⁵ n/cm²/s.²⁸⁾ The accelerator neutrons are characterized to be emitted in the forward-direction with respect to the deuteron beam direction. Hence, most of the emitted neutrons will be used effectively to produce ⁹⁹Mo by placing a sample after a neutron target (discussed later) in the direction of the deuteron beam. Thirdly, a quantity of ¹⁰⁰Mo samples of over 100 g weight mass can be used, because the neutron has no charge, and therefore the traveling range in a sample is much longer than that of a charged particle. When one uses a proton beam to produce ⁹⁹Mo or ^99mTc via the ¹⁰⁰Mo(p,pn)⁹⁹Mo reaction or the ¹⁰⁰Mo(p,2n)^99mTc reaction, the quantity of the ¹⁰⁰Mo sample mass would be less than 1 g, owing to the short proton range in the ¹⁰⁰Mo sample. In addition, proton beams have a heat problem of the sample because protons are stopped inside a ¹⁰⁰Mo sample material, and therefore high-intensity proton beams of above a few hundred µA are hardly used to produce medical radioisotopes.

Fig. 4.

(Color online) Neutron-induced reaction cross sections on ¹⁰⁰Mo.²⁶⁾ J. Phys. Soc. Jpn. 82, 064201 (Ref. 46).

3.1.2. ⁹⁹Mo yield.

In obtaining the specific activity of the produced ⁹⁹Mo as high as possible we calculated the angular and thickness distributions of the yield of ⁹⁹Mo produced by irradiating a ¹⁰⁰Mo sample having a large surface area with accelerator neutrons provided by the ^natC(d,n) reaction at a deuteron energy (E_d) of 40 MeV.²⁹⁾ In the calculation we assumed that the ¹⁰⁰Mo sample would be placed 2 cm downward from the carbon target, as shown in Fig. 5a, and the 40 MeV deuteron beam size to be a point. We used the latest data of neutrons from the ^natC(d,n) reaction at E_d = 40 MeV³⁰⁾ and the evaluated cross section of ¹⁰⁰Mo(n,2n)⁹⁹Mo given in the Japanese Evaluated Nuclear Data Library (JENDL-4.0).²⁶⁾ As we can see in Fig. 5b, the calculated ⁹⁹Mo yield is mostly distributed in a narrow region at an extremely forward angle with respect to the deuteron beam direction and within a sample thickness of as thick as about 4 cm. From this result, an appropriate shape of the ¹⁰⁰Mo sample to obtain a high specific activity of ⁹⁹Mo is determined to be cylindrical. Next, we measured the yield of ⁹⁹Mo produced by the ^natMo(n,2n)⁹⁹Mo reaction to make a rigorous test of the measured energy and angular distributions of the accelerator neutrons, including the evaluated cross section.^31),32) We used four pellet ^natMoO₃ samples of 25.869, 25.868, 25.483 and 25.220 g mass (in total 102.440 g mass) with dimensions of 30 mmϕ × 11.6 mm (total length 46.4 mm), as shown in Fig. 6.³²⁾ The isotopic composition of ^natMo is 14.8% for ⁹²Mo, 9.25% for ⁹⁴Mo, 15.92% for ⁹⁵Mo, 16.68% for ⁹⁶Mo, 9.55% for ⁹⁷Mo, 24.13% for ⁹⁸Mo and 9.63% for ¹⁰⁰Mo.³³⁾ The accelerator neutrons were provided by the ^natC(d,n) reaction using 40 MeV deuterons at the azimuthally variable field (AVF) cyclotron at Cyclotron and Radioisotope Center (CYRIC), Tohoku University.³⁴⁾ The distance, d, between the carbon target and the ^natMoO₃ sample was 9 mm.

Fig. 5.

(Color online) (a) Schematic view of the setup of the carbon target and ¹⁰⁰Mo sample. J. Phys. Soc. Jpn. 82, 064201 (Ref. 46). (b) Calculated ⁹⁹Mo yield distribution given in terms of the radius and thickness of a ¹⁰⁰Mo sample. The color label of the right vertical axis indicates the yield of ⁹⁹Mo (arbitrary unit). J. Phys. Soc. Jpn. 79, 093201 (Ref. 29).

Fig. 6.

Schematic view of the experimental setup at the ^natMoO₃ sample position. Sample numbers are indicated in parentheses. J. Phys. Soc. Jpn. 87, 043201 (Ref. 32).

The activity of ⁹⁹Mo at the end of irradiation (EOI) was determined, as given in Table 4, by considering the branching ratio of the observed γ-rays and the γ-ray detection efficiency of the HPGe detector. The self-absorption of the γ-rays in the irradiated ¹⁰⁰MoO₃ sample was corrected by using a photon cross-sectional database provided by the National Institute of Standards and Technology.³⁵⁾ Next, we estimated the yield of ⁹⁹Mo by numerical calculations for a comparison with the measured yield, as follows. We used the neutron-nucleus reaction cross sections given in the fourth version of the Japanese Evaluated Nuclear Data Library (JENDL-4.0) for molybdenum and oxygen in the ¹⁰⁰MoO₃ sample. In calculating the neutron flux we used the latest data of the cross section, which was obtained by irradiating a 15-mm-thick carbon target with 40 MeV deuteron beams.³⁰⁾ We corrected for the difference in the attenuation of the neutron flux inside the carbon target because the thickness of the carbon target used in the aforementioned study³⁰⁾ was 15 mm, compared with 10 mm in this measurement. The correction was made by using the simulation code Particle and Heavy Ion Transport code System (PHITS).³⁶⁾

Table 4. Measured ⁹⁹Mo yield (⁹⁹Mo_meas.) at the EOI compared with the calculated yield (⁹⁹Mo_cal.). The number in brackets in the first line is the number of the sample shown in Fig. 6

natMoO3 (g)	25.869	25.868	25.483	25.220
	(1)	(2)	(3)	(4)
99Momeas.(104 Bq)	3.9 ± 0.2	2.6 ± 0.1	1.7 ± 0.1	1.3 ± 0.1
99Mocal.(104 Bq)	3.6 ± 0.7	2.3 ± 0.4	1.5 ± 0.3	1.0 ± 0.2

The estimated yield of ⁹⁹Mo for each set of the ^natMo samples is in good agreement with the measured yield, as given in Table 4, which reflects the accuracies of the measured neutron energy and the angular distributions of the neutrons from the ^natC(d,n) reaction and the accuracy of the evaluated cross section of the ¹⁰⁰Mo(n,2n)⁹⁹Mo reaction.

Based on the good agreement between the measured ⁹⁹Mo yield and the calculated yield, we calculated the activities of ⁹⁹Mo at the EOI, produced by the ¹⁰⁰Mo(n,2n)⁹⁹Mo reaction for an enriched ¹⁰⁰MoO₃ sample, in terms of the weight and radius (r_s) of the ¹⁰⁰MoO₃ sample, the distance (d) between the carbon target and the ¹⁰⁰MoO₃ sample, and the irradiation time (i_t).³¹⁾ In the calculation, we used the latest data on the angular and energy distributions of neutrons from ^natC(d,n) at E_d = 40 MeV (assuming a beam intensity of 2 mA) and evaluated cross sections given in JENDL-4.0.²⁶⁾ Some of the calculated ⁹⁹Mo yields are listed in Table 5. Here, we took the radius of the deuteron beam, r_d, of 0.5 cm and i_t = 24 h as a typical setup to reduce the heat power density in the carbon target deposited by the deuteron beam and the decay loss of ⁹⁹Mo produced during the neutron irradiation period. We can see that both the ⁹⁹Mo and ^99mTc yields have a maximum at r_s = 2.0 cm independent of d in most cases.

Table 5. Calculated activities (in units of gigabecquerel) of ⁹⁹Mo at the EOI for an enriched ¹⁰⁰MoO₃ sample (100% enriched in ¹⁰⁰Mo) for 24 h in terms of the radius and thickness of ¹⁰⁰Mo, and the distance (d) between the natural carbon target and the ¹⁰⁰Mo sample position

Radius of d-beam (cm)	Distance between C and sample (cm)	Radius of sample (cm)	99mTc			99Mo
			100 g	150 g	200 g	100 g	150 g	200 g
0.5	0.5	0.5	1.4E+11	1.4E+11	1.4E+11	2.4E+11	2.4E+11	2.4E+11
		1.0	3.7E+11	3.7E+11	3.8E+11	6.1E+11	6.3E+11	6.4E+11
		2.0	3.3E+11	4.3E+11	4.9E+11	5.5E+11	7.2E+11	8.3E+11
		3.0	1.8E+11	2.6E+11	3.3E+11	3.0E+11	4.3E+11	5.6E+11
	1.0	0.5	1.0E+11	1.0E+11	1.0E+11	1.8E+11	1.8E+11	1.8E+11
		1.0	3.0E+11	3.1E+11	3.2E+11	5.1E+11	5.3E+11	5.3E+11
		2.0	3.0E+11	3.9E+11	4.5E+11	5.1E+11	6.6E+11	7.6E+11
		3.0	1.7E+11	2.5E+11	3.1E+11	2.8E+11	4.1E+11	5.3E+11
	1.5	0.5	8.0E+10	8.0E+10	8.0E+10	1.3E+11	1.3E+11	1.3E+11
		1.0	2.5E+11	2.6E+11	2.6E+11	4.2E+11	4.4E+11	4.5E+11
		2.0	2.7E+11	3.6E+11	4.1E+11	4.6E+11	6.0E+11	6.9E+11
		3.0	1.6E+11	2.3E+11	3.0E+11	2.7E+11	3.9E+11	5.0E+11
	2.0	0.5	6.3E+10	6.3E+10	6.3E+10	1.1E+11	1.1E+11	1.1E+11
		1.0	2.1E+11	2.2E+11	2.2E+11	3.6E+11	3.7E+11	3.8E+11
		2.0	2.5E+11	3.2E+11	3.7E+11	4.2E+11	5.4E+11	6.2E+11
		3.0	1.5E+11	2.2E+11	2.8E+11	2.5E+11	3.7E+11	4.7E+11

3.1.3. Capability to meet demand for ⁹⁹Mo/^99mTc in Japan.

In order to calculate the activity of ^99mTc, which is obtained daily using ⁹⁹Mo produced every day by the ¹⁰⁰Mo(n,2n)⁹⁹Mo reaction, we first calculated the ⁹⁹Mo activity reserved daily. Secondly, we compared the calculated activity with the current demand of ⁹⁹Mo, which is estimated by considering the number of diagnostic procedures by using ^99mTc-radiopharmaceuticals every year in Japan.³¹⁾

As given in Table 5 we calculated that a certain amount of ⁹⁹Mo is produced every day by the ¹⁰⁰Mo(n,2n)⁹⁹Mo reaction with 40 MeV, 2 mA deuteron beams (24 h irradiation). Through our studies discussed later we consider that an enriched ¹⁰⁰MoO₃ sample of mass 150 g (100 g of ¹⁰⁰Mo) is one of the feasible cases in terms of the present production efficiency of ⁹⁹Mo, and the elution performance of ^99mTc from ⁹⁹Mo, and the shortest distance between the carbon target and the ¹⁰⁰MoO₃ sample will be practically d = 1.0 cm by taking account of the mechanical structure of a rotating carbon target system (discussed later). Since we obtain the maximum (calculated) yield of ⁹⁹Mo for a sample radius of r = 2.0 cm for most cases (Table 5), the calculated yield of ⁹⁹Mo at the EOI per day is 657 GBq (150 g ¹⁰⁰MoO₃ sample mass, r_s = 2.0 cm, and d = 1.0 cm). The daily production of 657 GBq (18 Ci) of ⁹⁹Mo provides 2.30 TBq (63 Ci) of ⁹⁹Mo on average after 10 days in steady-state operation of the cyclotron, as shown in Fig. 7. It should be mentioned that we can also obtain the ⁹⁹Mo yield of 700 GBq using a ¹⁰⁰MoO₃ sample of 100 g by employing a following neutron irradiation scheme of the sample. The calculated yields of ⁹⁹Mo for one-pellet (100 g) and two-pellet (2 × 100 g) ¹⁰⁰MoO₃ samples are 505 and 756 GBq (Table 5), respectively. Hence, the yield of ⁹⁹Mo for one of the two samples, which was placed on the far side of the carbon target [hereafter the second sample] during the first irradiation time of 24 h, is 251 GBq. During the second neutron irradiation time of 24 h, the second sample was placed on the near side of the carbon target, resulting in a ⁹⁹Mo yield of 505 + 251 × 0.78 = 700 GBq. This yield is approximately the same (657 GBq) as that using a 150 g ¹⁰⁰MoO₃ sample mass. The factor 0.78 is due to the decay of ⁹⁹Mo, which was produced in the first irradiation. Hence, a 100 g ¹⁰⁰MoO₃ sample mass would be a better choice from an economical view point.³²⁾

Fig. 7.

Calculated ⁹⁹Mo activity after starting irradiation of the ¹⁰⁰MoO₃ sample with neutrons as a function of time (day). When 657 GBq (18 Ci) of ⁹⁹Mo is produced daily for 6 days per week, 2.30 TBq (63 Ci) of ⁹⁹Mo (in total) is obtained daily on average (dotted line). J. Phys. Soc. Jpn. 86, 114803 (Ref. 31).

Hence, by eluting ^99mTc from the reserved ⁹⁹Mo of 2.30 TBq once a day we obtain on average 1.63 TBq (44 Ci) of ^99mTc daily. Note that the activity of ^99mTc at time t, (A_Tc)_t, is calculated by using the following equation³⁷⁾:   

\begin{equation} (A_{\text{Tc}})_{t} = \frac{0.875\lambda_{2}}{(\lambda_{2} - \lambda_{1})}(A_{\text{Mo}})_{0}(\mathrm{e}^{-\lambda_{1}t} - \mathrm{e}^{-\lambda_{2}t}), \end{equation} [1]

where λ₁ = 0.0105/h and λ₂ = 0.1155/h are the decay constants for ⁹⁹Mo and ^99mTc, respectively, and (A_Mo)₀ is the activity of ⁹⁹Mo at t = 0.

Next, the ^99mTc activity that might be obtained daily at the end of thermochromatographic separation, discussed later, is evaluated. After the EOI, the irradiated ¹⁰⁰MoO₃ sample will be placed in a thermochromatography apparatus to separate ^99mTc from the neutron irradiated sample containing ⁹⁹Mo. Taking into account the decay loss of ^99mTc during the separation procedure for about 2 h and the separation efficiency of 80%, the ^99mTc activity immediately after separating ^99mTc from the “old” and “new” ⁹⁹Mo is 1.30 TBq (35 Ci). Here, “old” and “new” ⁹⁹Mo indicate the activity of ⁹⁹Mo produced on previous days and that day, respectively. Note that a ^99mTc solution is usually eluted twice per day, and therefore the obtained ^99mTc activity is 2.21 TBq (60 Ci), 1.7-times that in the case of elution once a day (see Fig. 8).

Fig. 8.

Decay-growth curve of the ⁹⁹Mo and ^99mTc. ^99mTc is eluted either every 12 h or every 24 h. J. Phys. Soc. Jpn. 86, 114803 (Ref. 31).

We next discuss the demand of ⁹⁹Mo in Japan by considering the number of diagnostic procedures currently carried out by using ^99mTc radiopharmaceuticals. About 0.9 million diagnostic procedures (2,750 procedures/day) have been performed every year using ^99mTc radiopharmaceuticals with an average dose of about 740 MBq (20 mCi) at the time of injection to a patient: 2.05 TBq (55 Ci) of ^99mTc is used every day. The activity of ^99mTc needed in Japan every day at a radiopharmaceutical company to prepare ^99mTc radiopharmaceuticals is calculated to be 4.10 TBq (110 Ci) by referring to the reports by Pillai et al.,³⁸⁾ Bennett et al.,³⁹⁾ and Ross et al.⁴⁰⁾ as follows. In the U.S.A., about 50,000 diagnostic procedures are carried out daily using ^99mTc radiopharmaceuticals. The U.S.A. requires 55.5 TBq (1,500 Ci) of ^99mTc daily assuming that 1.1 GBq (30 mCi) of ^99mTc is injected into a patient, and by considering the decay loss during the transportation of ^99mTc from the radiopharmaceutical company to hospitals, a total ^99mTc activity of 111 TBq (3,000 Ci) is required every day. Similarly, the total ^99mTc activity required in Japan is calculated to be 4.10 TBq (110 Ci) every day.

As discussed above, 2.21 TBq (60 Ci) of ^99mTc is obtained at the end of the separation of ^99mTc, i.e., immediately before forming the ^99mTc radiopharmaceuticals at a radiopharmaceutical company. When one can inject the ^99mTc radiopharmaceuticals prepared by using the 2.21 TBq ^99mTc within 6 h to a patient, about 50% of the daily procedures using ^99mTc radiopharmaceuticals can be performed in Japan. In fact, because about 35% of Japan’s population live in the capital-area (a region within about 150 km radius from the center of Tokyo, e.g., Kanagawa, Saitama, Chiba, Ibaraki, Tochigi, Gunma, and Yamanashi prefectures), and the number of people who live in the capital-area, Kansai-area, and Chukyo-area is about 60% of the total population of Japan, ^99mTc radiopharmaceuticals can be delivered from a radiopharmaceutical company to those who live there within 6 h. The proposed delivery of ^99mTc would be possible by using the current delivery system of ¹⁸F-FDG,⁴¹⁾ a radiopharmaceutical fluorodeoxyglucose containing ¹⁸F with a half-life of 1.8 h shorter than ^99mTc (T_1/2 = 6 h). In Japan, deliveries of ¹⁸F-FDG are being carried out three times per day by road transport from the ¹⁸F-FDG-producing radiopharmaceutical company to hospitals about 200 km away within 3 h. The activity of ¹⁸F prepared at the radiopharmaceutical company is three-times stronger than that needed for injection into a patient at a hospital by considering a decay loss of 68% during the transport of ¹⁸F in 3 h. Note that the ^99mTc activity needed at hospitals for the ^99mTc procedures to satisfy 50% of the demand in Japan is 10.2 TBq (27.5 Ci). The decay loss of ^99mTc activity in transportation for 3 h is 30%. Hence, the ^99mTc activity of 60 Ci obtained at a radiopharmaceutical company would be enough to perform the mentioned procedures. Note that we decided to harvest the produced ⁹⁹Mo every day by considering the ¹⁰⁰Mo inventory and the ⁹⁹Mo decay during the ¹⁰⁰Mo sample irradiation. A harvest frequency of ⁹⁹Mo once every six days would decrease the instantaneous production rate by about 40% relative to that of one day.

3.1.4. Radionuclide purity of ⁹⁹Mo.

The radionuclide purity of ⁹⁹Mo should be high in order to obtain high-quality ^99mTc to perform the separation process of ^99mTc from ⁹⁹Mo under radionuclides with little impurities, and not to create a problem for the storage of long-lived radioactivity. We measured the radionuclide purity of ⁹⁹Mo produced by using an enriched ¹⁰⁰MoO₃ sample of 0.578 g. The isotopic composition of Mo in the enriched ¹⁰⁰MoO₃ sample was 0.15 at% for ⁹⁴Mo, 0.1 at% for ⁹⁴Mo, 0.18 at% for ⁹⁵Mo, 0.21 at% for ⁹⁶Mo, 0.17 at% for ⁹⁷Mo, 3.29 at% for ⁹⁸Mo, and 95.90 at% for ¹⁰⁰Mo. The neutrons were provided by the ^natC(d,n) reaction using 40 MeV, 1.75 µA deuterons at the Takasaki Ion Accelerators for Advanced Radiation Application of the National Institutes for Quantum and Radiological Science and Technology (hereafter TIARA-QST).⁴²⁾ The ¹⁰⁰MoO₃ sample was irradiated for 5 h.

A typical γ-ray spectrum of the ¹⁰⁰MoO₃ sample irradiated by neutrons is shown in Fig. 9. We clearly observed only several γ-rays from the decay of ⁹⁹Mo (at 181.1 keV) and ^99mTc (at 140.5 keV), and impurity γ-rays from ⁹⁷Zr (T_1/2 = 16.9 h at 743.4 keV) and ^97gNb (T_1/2 = 1.2 h at 658.1 keV).

Fig. 9.

γ-ray spectrum of the ¹⁰⁰MoO₃ sample taken 4 h after the EOI. The γ-ray peaks are from the decay of ⁹⁹Mo (open circles), ^99mTc (filled circle), ⁹⁷Zr (open triangle), and ⁹⁷Nb (filled triangle). J. Phys. Soc. Jpn. 86, 114803 (Ref. 31).

The activities of ⁹⁹Mo and ⁹⁷Zr at the EOI were determined to be (3.16 ± 0.12) × 10⁶ Bq for ⁹⁹Mo and (31.5 ± 1.6) × 10³ Bq for ⁹⁷Zr, which is 1% of the ⁹⁹Mo activity, as shown in Table 6. Namely, ⁹⁹Mo was produced with a minimum level of radioactive waste and without radioisotopes of Tc other than ^99mTc and ⁹⁹Tc (T_1/2 = 2.1 × 10⁵ y). They are important because the irradiated enriched ¹⁰⁰MoO₃ sample can be recycled.

Table 6. Activities of radionuclides produced by the ¹⁰⁰Mo(n,2n)⁹⁹Mo reaction. b and Y_γ are the γ-ray emission probability and the γ-ray yield, respectively

Nuclides	Reaction	T1/2 (h)	Eγ (keV)	εγ (%)	b (%)	Yγ (cps)	EOI activity (Bq)
							Measurement
99mTc	99Mo decay	6.0	140.5	0.040	89.1	633.1 ± 1.4	—
99Mo	100Mo(n,2n)	65.9	181.1	0.036	6.0	53.8 ± 0.4	(3.08 ± 0.11) × 106
			739.5	0.012	12.1	41.5 ± 0.3	(3.21 ± 0.12) × 106
			777.9	0.012	4.3	14.2 ± 0.2	(3.19 ± 0.13) × 106
97Zr	100Mo(n,α)	16.9	743.4	0.012	93.1	2.14 ± 0.09	(31.5 ± 1.6) × 103
97Nb	97Zr decay	1.2	658.1	0.014	98.4	2.53 ± 0.10	—

3.2. Therapeutic radioisotopes production by the (n,xp) or (n,αx) reaction.

3.2.1. Medical radioisotopes produced in reactors and accelerators.

In the treatment for patients with cancers, medical RIs are used first to obtain pre-therapy imaging information concerning biodistribution and dosimetry in patients, and second to perform higher dose targeted molecular therapy in the same patients. Most medical RIs used for imaging and therapy are, respectively, being produced in accelerators (except ⁹⁹Mo) and in reactors (except ²²⁵Ac). In the production of therapeutic RIs in reactors by the fission reaction of ²³⁵U or the thermal neutron capture reaction of a sample, high thermal neutron fluxes on the order of 10¹⁴ n/(cm² s) and the use of a large quantity of a sample plays a key role. Note that carrier-free (without any isotope of a sample nuclide of ²³⁵U) RIs suitable for medical use are obtained by the fission reaction. Currently, carrier-free RIs of ⁹⁰Y (T_1/2 = 2.67 d), the daughter radioisotope of ⁹⁰Sr (T_1/2 = 28.8 y), and ¹³¹I (T_1/2 = 8.02 d) have been used for therapy. On the other hand, carrier-added (with a sample) RIs are usually generated by the thermal-neutron capture reaction. They cannot be separated from a neutron-irradiated sample because the atomic number of RIs is the same as that of a sample. However, there are several cases in which the produced RIs can decay by emitting β⁻-rays, resulting to obtain carrier-free RIs by chemical separation. ¹³¹I and ¹⁷⁷Lu are, respectively, produced by the ¹³⁰Te(n,γ)¹³¹Te → ¹³¹I and ¹⁷⁶Yb(n,γ)¹⁷⁷Yb → ¹⁷⁷Lu reactions. Here, it must be noted that currently constantly available RIs produced by the fission reaction and thermal neutron capture reaction are, ⁸⁹Sr, ⁹⁰Y, ¹³¹I, ¹⁷⁷Lu, ¹⁹²Ir, and ¹⁹⁸Au etc. We might expect that a wide variety of medical radioisotopes can be produced in reactors by the two reactions mentioned above. A limitation of available numbers comes mainly from the fission yield curve of ²³⁵U having maxima at masses of around 90–100 and 133–143, as shown in Fig. 10, and a sample mass dependence of the thermal neutron capture reaction having a large cross section.⁴³⁾

Fig. 10.

²³⁵U Fission product mass yield curve. No. JAEA-DATA/CODE--2011-025. Tokai, Japan: Japan Atomic Energy Agency, 2012 (Ref. 43).

In accelerators a wide variety of carrier-free RI with a high specific activity, mostly used for diagnostics, have been produced by using proton beams. In proton irradiation on a sample, it must be noted that the whole proton energy is transformed into heat in the sample, and the traveling range of protons in a sample is much shorter than that of neutrons, which limit both the proton beam intensity and the quantity of the sample for producing RIs. Hence, therapeutic RIs are hardly produced using proton beams. We first proposed a new route to produce ⁹⁹Mo by using accelerator neutrons, and then proposed new methods to produce therapeutic RIs.^44)–46)

3.2.2. Production for theranostic radioisotopes.

A charge-exchange reaction, such as (n,p), (n,x), and (n,α), of a sample nucleus with a medium-weight mass, has a sizable cross section of from ∼50 to ∼500 mb at a neutron energy of between ∼10 and ∼30 MeV, which is almost independent of the mass number of the sample. Here, (n,x) denotes the (n,n′p) and (n,d) reactions. The cross section of the (n,2n) reaction of a neutron-rich nucleus at ∼10 < E_n < 20 MeV is also quite large, and does not depend so much on the nuclear mass; it is in the range between 500 and 2,000 mb. Note that the neutron has no charge, and therefore the traveling range in a sample is much longer than that of a charged particle. Therefore, using high flux accelerator neutrons and a large amount of a sample, a large quantity of a wide variety of carrier-added and carrier-free radioisotopes (without any other isotope of a sample nuclide) can be produced, which would lead to a new era in theranostic RIs production. A schematic view of the production of a variety of therapeutic radioisotopes using accelerators and many stable isotopes is shown in Fig. 11.

Fig. 11.

Schematic view of the production of a wide variety of therapeutic radioisotopes. A hot nucleus is produced by bombarding a sample with accelerator neutrons, followed proton, neutron or α-particle emission. When a charged particle, such as a proton or α, is emitted, one can obtain carrier-free RIs by employing a chemical separation technique.

We have proposed new routes to produce carrier-free medical radioisotopes of ⁹⁰Y,^{43) 64}Cu, and ⁶⁷Cu^44),45) using accelerator neutrons provided by the ^natC(d,n) reaction. Successful PET with the use of ¹⁸F for assessments of tumor characterization has triggered a search for a longer half-life PET RI to diagnose the dynamics of a medicine in a living body that has a slow reaction time. ⁶⁴Cu with a half-life of T_1/2 = 12.7 h, longer than that of ¹⁸F (T_1/2 = 1.8 h), is considered to be a promising RI suitable for labeling many radiopharmaceuticals for PET imaging,²⁾ since ⁶⁴Cu decays by positron (β⁺) emission. The Cu radioisotope is known to have unique potentials useful for diagnostic imaging and in targeted radionuclide therapy.^4),5) In radioimmunotherapy (RIT) for tumor treatments, ⁹⁰Y (T_1/2 = 64 h), a pure β⁻-ray emitter with an average β⁻-ray energy of 935 keV, is most widely used to kill large tumor masses, since the range of β⁻-rays in H₂O is as long as 12 mm. ⁶⁷Cu (T_1/2 = 62 h), a pair radioisotope of ⁶⁴Cu, is considered to be a promising radionuclide for treating small distant metastases of up to 4 mm in size in radioimmunotherapy.^{6) 67}Cu has unique nuclear properties and chemical behavior for use in RIT.^4),5) Namely, ⁶⁷Cu emits β⁻-rays with an average energy of 141 keV, which allow radiopharmaceuticals of ⁶⁷Cu to provide a lethal dose of radiation to target cancer cells. ⁶⁷Cu also emits 185 keV γ-rays, which permit SPECT imaging during therapy. In addition, ⁶⁷Cu has a sufficiently long half-life (T_1/2) of 62 h, allowing it to be delivered to tumors, which may take 24 to 48 h to reach their peak concentration in tumors.^{4) 64}Cu should be noted to be used for pre-therapeutic PET studies for accurate evaluations of the dose delivered to a normal organ before the injection of ⁶⁷Cu for RIT to patients, since the irradiation of vital organs should be minimized. The coordination chemistry of copper applied to the production of radiopharmaceuticals has been well established. On the basis of a successful clinical study on a radiopharmaceutical containing ⁶⁷Cu for B-cell non-Hodgkin lymphoma, about 450 TBq (12,000 Ci) of ⁶⁷Cu is considered to be required per year in the U.S.A.⁴⁷⁾ Currently, there exists no technology to meet such a demand.

Thus far, many studies have been carried out to produce ⁶⁴Cu and ⁶⁷Cu in reactors or accelerators. ⁶⁴Cu was produced by the ⁶⁴Zn(n,p)⁶⁴Cu reaction in reactors using accelerators by the ⁶⁴Ni(p,n)⁶⁴Cu, ⁶⁴Ni(d,2n)⁶⁴Cu, ⁶⁴Zn(d,2p)⁶⁴Cu, ⁶⁶Zn(d,α)⁶⁴Cu, ⁶⁸Zn(p,αn)⁶⁴Cu, and ⁶⁴Zn(n,p)⁶⁴Cu reactions in accelerators.⁴⁷⁾ Among the studies, the generally adopted production route is the ⁶⁴Ni(p,n)⁶⁴Cu reaction, which provides a high specific activity ⁶⁴Cu using a highly enriched ⁶⁴Ni target. The maximum production yield of ⁶⁴Cu is expected to be about 37 GBq (1 Ci) by bombarding a ⁶⁴Ni sample with 50 µA proton beams for 12 h.⁴⁸⁾ It is very greatly encouraged to increase the availability of ⁶⁴Cu. ⁶⁷Cu has been produced by the ⁶⁷Zn(n,p)⁶⁷Cu reaction in both reactors and accelerators,¹³⁾ and in accelerators via the ⁶⁸Zn(p,2p)⁶⁷Cu, ⁷⁰Zn(p,α)⁶⁷Cu, ⁶⁴Ni(α,p)⁶⁷Cu, ⁶⁸Zn(γ,p)⁶⁷Cu and ⁶⁷Zn(n,p)⁶⁷Cu reactions. Note that the isotopic components of natural Zn are 48.6% ⁶⁴Zn, 27.9% ⁶⁶Zn, 4.1% ⁶⁷Zn, 18.8% ⁶⁸Zn, and 0.62% ⁷⁰Zn.³³⁾ Among the reactions, the ⁶⁸Zn(p,2p)⁶⁷Cu and ⁶⁸Zn(γ,p)⁶⁷Cu reactions are currently used for the production of ⁶⁷Cu. Since the proton energy used in the ⁶⁸Zn(p,2p)⁶⁷Cu reaction is high, a large amount of impurity RI of ⁶⁴Cu is produced by the ⁶⁸Zn(p,αn)⁶⁴Cu reaction at EOI. Regarding the ⁶⁷Cu production induced by accelerator neutrons, the spectrum-averaged cross section of the ⁶⁷Zn(n,p)⁶⁷Cu reaction was measured at E_n = 4.95 MeV,⁴⁹⁾ but the route has not yet been adopted because of the small cross section and the lack of an intense neutron source. Note that recently high-quality ⁶⁷Cu has been produced by the ⁶⁸Zn(γ,p)⁶⁷Cu reaction at Argonne National Laboratory.⁵⁰⁾

Concerning the ⁶⁴Cu and ⁶⁷Cu productions using accelerator neutrons, the main drawback comes from the low neutron flux, but not from the nuclear reaction processes, such as the cross section of a required RI or the high production yield of an impurity RI. As mentioned above a high neutron flux of ∼10¹⁵ n/s at an average neutron energy of E_n ≈ 14 MeV can be obtained owing to recent progress in accelerator technology. These findings led us to propose new routes to produce carrier-free radioisotopes of ⁶⁴Cu by the ⁶⁴Zn(n,p)⁶⁴Cu reaction and ⁶⁷Cu via the ⁶⁷Zn(n,p)⁶⁷Cu and ⁶⁸Zn(n,x)⁶⁷Cu reactions using accelerator neutrons. In order to calculate the production yields of ⁶⁴Cu and ⁶⁷Cu we first re-measured the cross sections of the ⁶⁴Zn(n,p)⁶⁴Cu, ⁶⁷Zn(n,p)⁶⁷Cu, and ⁶⁸Zn(n,x)⁶⁷Cu reactions using neutrons at E_n ≈ 14 MeV. Although many studies were carried out to measure those cross sections at E_n ≈ 14 MeV, there remained significant differences between different data sets. The measurement was carried out using ∼14 MeV neutrons produced via the ³H(d,n)⁴He reaction at the Fusion Neutronics Source (FNS) facility of Japan Atomic Energy Agency (JAEA).⁵¹⁾ The obtained results led us to estimate the yields of ⁶⁴Cu produced by the ⁶⁴Zn(n,p)⁶⁴Cu reaction and ⁶⁷Cu by the ⁶⁷Zn(n,p)⁶⁷Cu and ⁶⁸Zn(n,x)⁶⁷Cu reactions using high neutron fluxes. The neutrons could be provided by the ^natC(d,n) reaction with 40 MeV, 5 mA deuteron beams.²⁷⁾ The estimation was performed using the evaluated cross section of the neutron induced reaction on Zn isotopes given in the Japanese Evaluated Nuclear Data Library. The yield of ⁶⁴Cu was calculated to be 1.8 TBq/(175 g of 100%-enriched sample of ⁶⁴Zn) for an irradiation time of 12 h, which is much larger than the expected ⁶⁴Cu yield of 37 GBq via ⁶⁴Ni(p,n)⁶⁴Cu, and that of ⁶⁷Cu via the ⁶⁸Zn(n,x)⁶⁷Cu reaction was calculated to be 287 GBq/(186 g of 100%-enriched sample of ⁶⁸Zn) at EOI for an irradiation time of 2 days,⁴⁵⁾ which is much larger than a reported yield of 10 GBq by ^natZn(p,2p)⁶⁷Cu at E_p = 192 MeV at an average beam current of 43 µA for a period of 5–6 days.⁵²⁾

3.2.3. Radionuclide purity of ⁶⁷Cu.

We have studied the radionuclide purity of ⁶⁷Cu,⁵³⁾ the vital importance for its medical use and the recycling of an enriched sample of ⁶⁸Zn, produced by bombarding an enriched ⁶⁸ZnO sample of 0.329 g for 5 h with neutrons from the ^natC(d,n) reaction at E_d = 40 MeV at an average beam current of 1.84 µA at TIARA-QST. The isotopic compositions of Zn in the enriched ⁶⁸ZnO sample were 0.047 at% for ⁶⁴Zn, 0.104 at% for ⁶⁶Zn, 0.466 at% for ⁶⁷Zn, 99.291 at% for ⁶⁸Zn, and 0.092 at% for ⁷⁰Zn. A typical background-subtracted γ-ray spectrum of the irradiated ⁶⁸ZnO sample is shown in Fig. 12. We observed γ-rays from the decay of ⁶⁷Cu (at 91, 93, 185, 209, 300, and 394 keV), ⁶⁵Ni (T_1/2 = 2.52 h at 366, 508, 610, and 1,116 keV), ⁶⁶Ni (T_1/2 = 54.6 h at 1,039 keV), ⁶⁵Zn (T_1/2 = 244 d at 1,116 keV), and ^69mZn (T_1/2 = 13.8 h at 439 keV). The isotope assignments of the γ-rays were made on the basis of their energies and decay curves. ⁶⁵Zn was identified by the 1,116 keV γ-ray.

Fig. 12.

γ-ray spectrum of a ⁶⁸ZnO sample irradiated with neutrons. The γ-ray peaks come from the decays of ⁶⁷Cu (open circles), ⁶⁵Ni (open triangles), ⁶⁶Ni (filled triangle), and ^69mZn (filled diamond), and ⁶⁵Zn (filled square). The spectrum was taken 40 h after EOI. J. Phys. Soc. Jpn. 83, 073201 (Ref. 53).

The radionuclide purity of ⁶⁷Cu was determined to be extremely low compared with those produced by the ⁶⁸Zn(p,2p)⁶⁷Cu and ⁷⁰Zn(d,αn)⁶⁷Cu reactions, as given in Table 7, and estimated ones to investigate a possible reaction for producing impurity radionuclides. The estimation was made using the isotope composition of the enriched ⁶⁸ZnO sample mentioned above, the neutron energy spectra from the ^natC(d,n) reaction, and neutron nuclear reaction cross-sectional data on Zn isotopes.²⁶⁾ The estimated activity ratios of the impurity radioisotopes agree with the experimental ratios within the experimental uncertainties, as given in Table 7.⁵³⁾ The obtained information is also important when purchasing an expensive sample with a variety of isotopic compositions.

Table 7. Activity ratios of impurity radionuclides to ⁶⁷Cu produced by ⁶⁸Zn(p,2p)⁶⁷Cu, ⁷⁰Zn(d,αn)⁶⁷Cu, and ⁶⁸Zn(n,x)⁶⁷Cu reactions at EOI together with estimated ones

	Beam energy (MeV)	Ratio to 67Cu
		64Cu	66Ga	67Ga	69mZn	65Zn	65Ni	66Ni
68Zn(p,2p)67Cu	100 → 20	10	∼12	∼2.5		∼0.1
70Zn(d,αn)67Cu	19.5 → 18.4	0.1	0.03	0.07	2.3
Present Exp.		<0.016	0	0	0.14	6.7 × 10−4	2.6	(2.4–5.0) × 10−3
68Zn(n,x)
Estimation		7.5 × 10−4			0.12	5.1 × 10−4	2.9	1.0 × 10−3
68Zn(n,x)
66Zn(n,2n)						1.7 × 10−4
70Zn(n,2n)					0.03
64Zn(n,p)		8.7 × 10−3

3.3. Production of radioisotopes in polyethylene blocks.

We have proposed another production route by using accelerator neutrons backscattered by materials, such as polyethylene or lead blocks.

3.3.1. Experiment for producing RIs using polyethylene blocks.

In studying the production routes of medical RIs with accelerator neutrons by using a sample that was covered with polyethylene blocks to reduce the neutron background in an experimental room, we happened to find much larger yields of some of the RIs than those without polyethylene blocks.⁵⁴⁾

This study was performed by irradiating five stacked samples of ⁹³Nb, enriched ⁶⁸ZnO, enriched ⁶⁴ZnO, natural ^natZnO, and enriched ⁹⁰ZrO₂ as well as two stacked samples of ⁹³Nb and enriched metallic ⁶⁸Zn with accelerator neutrons. The ⁹³Nb sample was used as a high-energy neutron-fluence monitor.⁵⁴⁾ The masses for the ⁶⁸ZnO, ⁶⁸Zn, and ⁶⁴ZnO samples were about 360 mg, and the enrichment of the ⁶⁸ZnO, ⁶⁸Zn, and ⁶⁴ZnO samples was over 99%. These samples were covered with polyethylene (or lead) blocks, as shown in Fig. 13a. The size of the individual polyethylene or lead block was 200 × 100 × 50 mm³. The distance (d) between the Al holder and the polyethylene (or lead) shown in Fig. 13b was set to be either 3 or 6 cm for investigating a possible effect of the blocks on the yields of produced RIs. The neutrons were provided from the deuteron breakup reaction on a ⁹Be target using a 50 MeV, 0.5 µA deuteron beam at TIARA-QST. The samples were irradiated for about 15 min.

Fig. 13.

Schematic view of the experimental setup at the position of samples in polyethylene blocks (a), and the five stacked samples (b). d (cm) is the distance between the Al holder and the polyethylene block. J. Phys. Soc. Jpn. 89, 034201 (Ref. 45).

Hereafter we focus on the results of the two samples, ⁶⁸ZnO and ⁶⁸Zn; ⁶⁸Zn(PE) and ⁶⁸Zn(Pb) indicate ⁶⁸Zn samples covered with polyethylene and with lead, and ⁶⁸Zn(no) stands for a ⁶⁸Zn sample with neither polyethylene nor lead, respectively. Typical γ-ray spectra of the irradiated ⁶⁸ZnO(no), ⁶⁸ZnO(PE), and ⁶⁸Zn(PE) samples placed at d = 3 cm are shown in Figs. 14a, 14b, and 14c, respectively. Identifications of the produced radioisotopes were made based on the γ-ray energies and/or the absolute γ-ray branching ratio (I_γ), as given in Table 8.

Fig. 14.

γ-ray spectra from the decay of produced radioisotopes that were taken about 15 min after the EOI for the ⁶⁸ZnO sample (a) without polyethylene blocks and (b) with polyethylene blocks, and (c) for a metallic ⁶⁸Zn sample with polyethylene blocks placed at d = 3 cm. J. Phys. Soc. Jpn. 89, 034201 (Ref. 45).

Table 8. γ-ray energy of ⁶⁷Cu, ⁶⁷Ga, ⁶⁶Ga, ^69mZn, ⁶⁴Cu, ⁶⁵Zn, and ⁶⁵Ni, and the absolute γ-ray branching ratio, I_γ (in parenthesis)

	67Cu (61.8 h)	67Ga (78.6 h)	66Ga (9.5 h)	69mZn (13.8 h)	64Cu (12.7 h)	65Zn (244 d)	65Ni (2.52 h)
Eg (keV) and Ig (%)	93 (14.8)	93 (39)	834 (5.89)			1,116 (50.6)	1,116 (15.4)
	185 (44.2)	185 (20.9)	1,039 (37)	439 (94.8)	1,346 (0.47)		1,482 (24)
	300 (0.743)	300 (16.8)

An anomalous nuclear reaction phenomenon was found in this study. Namely, significant amounts of proton-induced reaction products of ⁶⁶Ga and ⁶⁷Ga were observed when irradiating an enriched ⁶⁸ZnO(PE) sample with accelerator neutrons. In fact, the γ-ray peak intensities of ⁶⁶Ga (at 834 and 1,039 keV), ⁶⁷Ga (at 300 keV), ^69mZn (at 439 keV), and ⁶⁴Cu (at 1,346 keV) of the ⁶⁸ZnO(PE) sample in Fig. 14b (and those of the ⁶⁸ZnO(Pb) sample) are much larger than those of the ⁶⁸ZnO(no) sample in Fig. 14a. We also found that the γ-ray intensities for the metallic ⁶⁸Zn(PE) sample in Fig. 14c were much smaller than those of the oxide ⁶⁸ZnO(PE) sample in Fig. 14b. However, the γ-ray intensities of ⁶⁷Cu, ⁶⁵Ni, and ⁶⁵Zn of the ⁶⁸ZnO(PE) and ⁶⁸Zn(PE) samples were approximately the same as those of the ⁶⁸ZnO(no) sample.

The activities for various isotopes of the ⁶⁸ZnO and ⁶⁸Zn samples at the end of irradiation (EOI) were obtained, as shown in Table 9, where the corresponding nuclear reaction path and the reaction threshold energies are also indicated. Here, as an example, in order to know the dependence of the activities for the ⁶⁸ZnO(PE) sample on d we took the difference between the activities for the ⁶⁸ZnO(PE) and for the ⁶⁸ZnO(no) samples and divided this difference by the activity for ⁶⁸ZnO(no) (columns F and G for d = 3 and 6 cm, respectively, in Table 9). The enhancement factors for d = 3 and 6 cm are compared by taking their ratio (column H in Table 9).

Table 9. Activity (in units of kBq) of radioisotopes produced from the ⁶⁸ZnO(no), ⁶⁸ZnO(PE), ⁶⁸ZnO(Pb), and ⁶⁸Zn(PE) samples at the end of irradiation in columns A–E. Columns F and G give the difference of B, C and A divided by A, and column H is a ratio between G and F. In most cases the uncertainties of measured activities are less than 10%

Radioisotope		67Cu	65Ni	65Zn	69mZn	67Ga	66Ga	64Cu
Reaction		(n,x)	(n,α)	(n,4n)	(n,γ)	(p,2n)	(p,3n)	(p,αn)
Ethr.(MeV)		7.9	0	28.7	0	12.2	23.6	7.9
Sample
A	68ZnO(no)	5.7	21	0.016	3.3	1.5	0.5	0.81
B	68ZnO(PE) (d = 3 cm)	5.4	33	<0.021	67	65	11	62
C	68ZnO(PE) (6 cm)	5.9	23	0.018	14	14	6.9	16
D	68ZnO(Pb) (3 cm)	6.1	35	0.022	75	72	15	65
E	68Zn(PE) (3 cm)	6.5	19	0.016	1.4	0.6	0.1	0.99
F	(B − A)/A	−0.5	0.6		19	42	20	76
G	(C − A)/A	0.04	0.1		3.2	8.3	12	19
H	G/F = (C − A)/(B − A)	−0.7	0.2		0.2	0.2	0.6	0.25

The results given in Table 9 are summarized in terms of the activity A(X) of a particular radioisotope X at the end of irradiation (EOI), as follows. Firstly, A(⁶⁷Ga), A(⁶⁶Ga), A(^69mZn), and A(⁶⁴Cu) obtained for the ⁶⁸ZnO(PE) sample placed at d = 3 cm are more than twenty-times larger than those for the ⁶⁸ZnO(no) sample (column F). Their yields of the ⁶⁸ZnO(PE) sample placed at d = 6 cm are about one-fourth of those for the sample at d = 3 cm, indicating a strong dependence of the yields on d. Secondly, their yields of the metallic ⁶⁸Zn(PE) sample at d = 3 cm (column E) are approximately 1/100, 1/100, 1/50, and 1/50 of those of the ⁶⁸ZnO(PE) sample at d = 3 cm, respectively. Thirdly, A(⁶⁷Cu), A(⁶⁵Ni), and A(⁶⁵Zn) of the ⁶⁸ZnO(PE) and ⁶⁸Zn(PE) samples are approximately the same as those of the ⁶⁸ZnO(no) sample within an uncertainty factor of 2.

Based on these findings we consider that they indicate a main reaction process to generate protons and neutrons, which play a key role in the large productions of ⁶⁷Ga, ⁶⁶Ga, ⁶⁴Cu, and ^69mZn. The second part of the summary, their smaller yields at d = 6 cm, suggests that they might be generated by any interaction between primary neutrons (hereafter n_prim) produced by the Be(d,n) reaction and nuclei in the polyethylene blocks. On the other hand, the third part of the summary indicates that these protons and neutrons should be dominantly generated in the ⁶⁸ZnO(PE) sample not in the polyethylene blocks via some type of nuclear interaction between neutrons scattered backwards by the polyethylene blocks (hereafter n_sc) and oxygen nuclei in the ⁶⁸ZnO sample. The anomalous large yields are considered to be dominated by n_sc. This is apparently unexpected, since the flux of n_sc is much reduced compared with that of n_prim, and the energy of n_sc is lower than that of n_prim. When the energy of n_sc is low, that of protons produced by any reaction process between n_sc and nuclei in the samples is also low. The cross sections of the ⁶⁸Zn(p,2n)⁶⁷Ga reaction with a threshold energy of 12.2 MeV and the ⁶⁸Zn(p,3n)⁶⁶Ga reaction with a threshold energy of 23.6 MeV decrease with a proton energy of less than about 26 MeV.⁵⁵⁾

Next, we studied the neutron energy dependence of anomalous large yields observed at E_d = 50 MeV by using 40 MeV deuterons. Enriched ⁶⁸ZnO samples with and without polyethylene blocks were irradiated with 40 MeV deuterons. Note that the neutron energy from 40 MeV deuterons is smaller than that from 50 MeV deuterons. The measured activities of ⁶⁷Cu, ^69mZn, and ⁶⁴Cu for the ⁶⁸ZnO sample are given in Table 10. Contrary to the results for 50 MeV deuterons, the yields of the observed radioisotopes for 40 MeV deuterons were independent of the existence of polyethylene blocks. It is evident that the primary-neutron energy played a key role in the large yields of the various radioisotopes mentioned above.

Table 10. Activity (in units of kBq) of radioisotopes produced from ⁶⁸ZnO(no) and ⁶⁸ZnO(PE) samples at the end of irradiations using 40 MeV deuterons. ND = not detected

Radioisotope		67Cu	65Ni	65Zn	69mZn	67Ga	66Ga	64Cu
Reaction		(n,x)	(n,α)	(n,4n)	(n,γ)	(p,2n)	(p,3n)	(p,αn)
Ethr.(MeV)		7.9	0	28.7	0	12.2	23.6	7.9
Sample
A	68ZnO(no)	6.9	ND	ND	1.6	ND	ND	1.2
B	68ZnO(PE) (d = 3 cm)	8.1	ND	ND	2	ND	ND	1.9

3.3.2. Estimation with Particle and Heavy Ion Transport code System (PHITS).

We calculated the produced activities, A(X), for ⁶⁸ZnO(no) and ⁶⁸ZnO(PE) samples for 50 and 40 MeV deuterons to compare with the measured activities. The calculation was performed using the PHITS code with the geometry of the experimental setup shown in Fig. 13, and the evaluated production cross section data of neutron- and proton-induced reactions given in the fourth version of the Japanese Evaluated Nuclear Data Library (JENDL-4.0/HE). The neutron production rates from the ⁹Be(d,n) reaction at E_d = 50 and 40 MeV were determined so that the PHITS simulation gives the same angular and energy distributions of neutrons as those of the latest data reported by Meulders et al.⁵⁶⁾ and Saltmarsh et al.⁵⁷⁾ Note that the neutron data at E_d = 50 MeV and at E_d = 40 MeV were obtained by using 10-mm-thick and 6.3-mm-thick Be targets, respectively. We used a 9-mm-thick Be target for 50 and 40 MeV deuterons. Hence, we corrected for any difference in the attenuation of the neutron fluence inside the Be target between the previous studies and the present study by using the PHITS code.

As given in Table 10, the calculated activities of ⁶⁷Cu, ⁶⁷Ga, ⁶⁶Ga, ^69mZn, and ⁶⁴Cu from the ⁶⁸ZnO(no) sample for 50 and 40 MeV deuterons are consistent with the measured ones within a factor of 3. Here, ⁶⁷Cu are produced by the ⁶⁸Zn(n,x) reaction. ⁶⁷Ga and ⁶⁶Ga are generated by the ⁶⁸Zn(p,2n) and ⁶⁸Zn(p,3n) reactions in which protons that passed the Be target after the deuteron breakup reaction on Be are considered to be bombarded on the ⁶⁸ZnO(no) sample. The ⁶⁸Zn(p,nα)⁶⁴Cu and ⁶⁴Zn(n,p)⁶⁴Cu reactions contribute to generate ⁶⁴Cu. Note that the isotopic composition of ⁶⁴Zn in the ⁶⁸ZnO sample is 0.047 at.%. Contrary to the cases for the ⁶⁸ZnO(no) sample, the calculated activities of ⁶⁷Ga, ⁶⁶Ga, and ^69mZn were smaller by a factor of about 15 to 75 than the measured ones, as given in Table 11, and almost the same as those obtained from the ⁶⁸ZnO(no) sample. The reason why the measured activities differ from the calculated ones is unclear at present. Some sort of nuclear reaction process between oxygen nuclei in the ⁶⁸ZnO sample and the scattered neutrons could be the origins of the discrepancy.

Table 11. Measured and calculated activities of radioisotopes produced by an irradiating ⁶⁸ZnO sample without and with polyethylene blocks with neutrons from 50 MeV deuterons (in brackets 40 MeV deuterons). ND = not detected

68ZnO(no) sample
	Activity (kBq) at EOI Ed = 50 MeV [40 MeV]
	67Cu	65Ni	65Zn	69mZn	67Ga	66Ga	64Cu
Exp	5.7 [7.6]	21 [53]	0.016 [ND]	3.3 [1.6]	1.5 [ND]	0.53 [ND]	0.81 [1.2]
Cal	14 [11]	26 [38]	0.045 [0.013]	0.81 [0.88]	3.9 [0.036]	0.74 [0.001]	1.6 [0.12]
Exp/Cal	0.41 [0.69]	0.87 [1.4]	0.35 [ND]	4.1 [1.8]	0.38 [—]	0.71 [—]	0.51 [10]

68ZnO(PE) sample
	Activity (kBq) at EOI Ed = 50 MeV
	67Cu	65Ni	65Zn	69mZn	67Ga	66Ga	64Cu
Exp	5	33	ND	67	73	11	62
Cal	14	26	0.045	0.9	3.9	0.74	1.7
Exp/Cal	0.36	1.27	—	74	19	15	36

In conclusion we discovered anomalously large yields of ⁶⁷Ga, ⁶⁶Ga, ^69mZn, and ⁶⁴Cu by irradiating a ⁶⁸ZnO sample that was covered with polyethylene blocks with neutrons provided from the ⁹Be(d,n) reaction by 50 MeV deuterons, but not by 40 MeV deuterons. This finding will be important for the production of radioisotopes in large quantity with accelerator neutrons by using not only incident accelerator neutrons on a sample, but also neutrons scattered backward from polyethylene blocks simultaneously. Especially, we could produce several radioisotopes which are normally generated by proton-induced reaction from a single sample irradiated with the backscattered neutrons.

4. Separation

4.1. Separation of ^99mTc from ⁹⁹Mo.

The specific activity of fission-⁹⁹Mo is high, ∼370 TBq/(g Mo),³⁷⁾ and therefore ^99mTc with high specific activity that is sufficient for performing medical diagnostics is recovered by using quite small volumes of saline with high separation efficiencies of 80 to 90% under repeated elution (milking) by using a ⁹⁹Mo/^99mTc generator. In fact, about 1 ml of saline is enough to recover ^99mTc with an activity of 740 MBq. On the other hand, a typical specific activity of ⁹⁹Mo produced by any alternative production method of ⁹⁹Mo without the fission reaction of ²³⁵U is very low, about 1/5,000 of fission-⁹⁹Mo.²²⁾ Hence, when one uses the ⁹⁹Mo/^99mTc generator, about 5 l of saline is required to collect ^99mTc of 740 MBq, which would result in an unacceptably low concentration of eluate for the direct formulation of radiopharmaceuticals. Note that a typical activity of a solution of ^99mTc radiopharmaceuticals administered to a patient is high, about 740 MBq/(a few ml). So far, in order to obtain a high specific activity, ^99mTc, from ⁹⁹Mo with such a low specific activity various method, such as chromatographic,⁵⁸⁾ solvent extraction,⁵⁹⁾ and thermo-separation methods,¹³⁾ have been developed.

Among these methods, upon considering the following potential we employed the thermo-separation procedure, which utilizes the different volatility of technetium oxide and MoO₃ to separate ^99mTc from ⁹⁹Mo. Namely, with this method it is expected that a large quantity of an enriched ¹⁰⁰MoO₃ sample can be used. ^99mTc with a high radioactive concentration can be obtained free from any chemical impurities, and an irradiated ¹⁰⁰MoO₃ sample can be recovered with high efficiency for recycling. So far, on the basis of thermo-separation, the procedures of sublimation and thermo-chromatography have been developed.⁶⁰⁾ In sublimation, ^99mTc produced in a MoO₃ powder sample is separated, since it volatilizes at a temperature much lower than the melting point of MoO₃ at 795 °C. In thermo-chromatography, Tc and Mo oxides volatilized from a molten MoO₃ sample are condensed in different temperature zones along a column where the temperature gradient is kept constant in a furnace (see Fig. 15). In reality, before we started to develop a thermoseparation system, the aforementioned potentials of the thermoseparation method had not yet been materialized, owing to the challenges discussed below. In fact, it was not even clear which is better for separating ^99mTc from an irradiated Mo sample of over 100 g, a sublimation method or a thermo-chromatography method. We decided to employ the latter by noting a previous study by Tachimori et al.⁶¹⁾ They first determined a diffusion coefficient, D, of ^99mTc in MoO₃ using Fick’s equation of diffusion when considering the release mechanism of ^99mTc from MoO₃ powder samples of 10 to 20 mg. A typical value of D was 7.71 × 10⁻¹² cm²/s at a furnace temperature of 780 °C.⁶¹⁾ Using the D at 780 °C and Brownian motion theory, the mean diffusion distance, x, which is given as x = (2Dt)^1/2, was calculated to be ∼10⁻⁴ cm for one hour of diffusion (t = 1 h). Since D is so low, we understood that it is impossible to obtain a high separation efficiency of ^99mTc from a thick MoO₃ powder sample which is used for ⁹⁹Mo production by the ¹⁰⁰Mo(n,2n)⁹⁹Mo reaction.⁶²⁾ Hence, we employed a thermochromatographic separation method.^62)–64) A schematic of the experimental setup of the thermochromatographic separation is shown in Fig. 15.

Fig. 15.

(Color online) Schematic of the experimental setup of the thermochromatographic separation. ① MoO₃ sample, ② Platinum boat, ③ MoO₃ needle crystal, ④ CZT detector, ⑤ Quartz wool, ⑥ Heater, ⑦ Heat insulator, ⑧ Water cooling jacket and ⑨ Crumpled gold wire. J. Phys. Soc. Jpn. 84, 043202 (Ref. 63).

So far, by using the thermochromatographic separation technique many studies were undertaken to measure the separation efficiency by using a molten MoO₃ sample.^37),65),66) However, there remain several challenging problems concerning separation. First is that the separation efficiency of ^99mTc, ε_sp, which is the proportion of ^99mTc separated during the thermochromatographic-separation process, is low. This diminishes markedly with repeated sublimation tests (repeated milking tests) at a constant furnace temperature, T_fur, and decreases with an increasing mass of MoO₃ loaded into a sublimation furnace at a time. For example, the efficiency at furnace loading of a 200 g Mo sample generated via the ⁹⁸Mo(n,γ)⁹⁹Mo reaction was 25% on average, and rarely exceeded 50%.³⁷⁾ Note that a milking process of ^99mTc from ⁹⁹Mo is usually performed twice per day for at least one week. Typical separation efficiencies of a 5 g MoO₃ sample were 85, 12, 19 and 24%,⁶⁵⁾ which were markedly diminished with repeated milking tests. An Idaho group⁶⁶⁾ used a molten MoO₃ sample with a thickness of 0.8 mm in order to obtain a high separation efficiency of ⁹⁹Tc, a pure β-emitter. Having such a thin sample is, however, not desirable for the large-scale production of ⁹⁹Mo.

4.1.1. Thermoseparation efficiency of ^99mTc from ⁹⁹Mo.

We have challenged these problems by measuring the diffusion coefficient and separation efficiency of ^99mTc from a molten MoO₃ sample of two different thicknesses using a home-made electric furnace, shown in Fig. 15.^62),63)

A set of three-stage quartz tubes was enclosed by a four-zone tube furnace. A platinum boat was used to hold the irradiated MoO₃ sample in a high-temperature region throughout the milking. The first two zones were heating sections used to melt the irradiated MoO₃ sample in a stream of oxygen carrier gas at around 830 °C so as to form gaseous materials containing vaporized ^99mTc and Mo oxides. The third zone was an intermediate section used to condense any vaporized MoO₃ as a needle crystal and to transfer gaseous materials containing vaporized ^99mTc oxide from the heating sections to a final section. The fourth zone was a final section to collect the separated ^99mTc. Since the temperature of the intermediate section was set so as to decrease gradually as the distance in the intermediate section from the heating section increases, the temperature of the gaseous materials at the final section became sufficiently low so as to produce condensation of the gaseous ^99mTc products. A quartz wool filter was placed within the tube at a temperature below the melting point of MoO₃ so as to stop the migration of any volatilized MoO₃ towards the final section. Crumpled gold wire was placed in the tube to increase the surface area as much as possible for ^99mTc collection. The distribution of the ^99mTc activity along the tube was investigated using the cadmium zinc telluride (CZT) detector; its peak activity was in the condense region, clearly separated from the ¹⁰⁰MoO₃ sample, as shown in Fig. 16.

Fig. 16.

Distribution of the 141 keV γ-ray intensity along a quartz tube measured by the CZT detector after thermochromatographic separation. The vertical axis is the 141 keV γ-ray intensity. The horizontal axis represents the distance from the MoO₃ sample position along the quartz tube. The deposited ^99mTc is clearly separated from the sample. J. Phys. Soc. Jpn. 84, 043202 (Ref. 63).

A sequential milking process used to separate ^99mTc from molten ⁹⁹MoO₃ samples of 4.0 and 8.5 mm thicknesses was carried out every for ∼24 hours at a furnace temperature (T_fur) of T_fur = 845–855 °C for a heating time (t_heat) of 15–30 min. We used two CZT γ-ray detectors to study the separation efficiency and the diffusion coefficient of ^99mTc in a molten MoO₃ sample by detecting the 141 keV and 181 keV γ-rays from the decays of ^99mTc and ⁹⁹Mo, respectively. The separation efficiency was determined by comparing the 141 keV γ-ray yield of the separated ^99mTc with that of the untreated ^99mTc within the irradiated MoO₃ sample before the separation. CZT detectors were placed at the quartz boat (CZT-1) and quartz wool (CZT-2) positions, respectively. The 141 γ-rays yields obtained by the CZT-1 detector for the 4.0 and 8.5 mm thick samples during the milking processes are shown in Figs. 17a and 17b, respectively. For the first 18 min, i.e., before melting of the samples, the γ-ray yield was normalized to 100%. A marked release of ^99mTc, which then condensed on the inner wall of the cooled quartz tube, was observed, after the furnace temperature reached the melting point of MoO₃ at 795 °C after t ≈ 20 min. The deposition of ^99mTc that thermally diffused through the quartz tube was confirmed by CZT-2. It should be noted that all of the 141 keV γ-ray yields of a 4.0 mm (8.5 mm) thick sample obtained by CZT-1 showed the same dependence on the temperature and time, which indicates that the high separation efficiencies of ^99mTc remained constant during the sequential milking processes, which differed from the previously reported results. The 141 keV γ-ray was observed after the release of ^99mTc from the molten MoO₃ sample, which included γ-ray emission from the 141 keV state of ⁹⁹Tc fed by γ-decay from the 181 keV state, which was populated by the decay of ⁹⁹Mo, not via the 143 keV state (^99mTc) (see Fig. 1), as well as the remaining ^99mTc in the sample.

Fig. 17.

(a) The 141 keV γ-ray yields of ^99mTc milking processes obtained by the CZT-1 detector for a 4.0 mm thick MoO₃ sample at T_fur = 845 °C for t_heat = 15 min. Both the 141 and 181 keV γ-rays from ^99mTc and ⁹⁹Mo are clearly observed (inserted). Four milking runs (filled diamonds, open diamonds, filled squares, and open squares) are shown. (b) The γ-ray yields of ^99mTc for the 8.5 mm thick MoO₃ sample at T_fur = 845 °C for t_heat = 15 min {Run 1 (filled triangles), Run 2 (open triangles)} and for t_heat = 30 min {Run 3 (filled circles), and at T_fur = 855 °C for t_heat = 15 min {Run 4 (open circles)}. The dotted and dashed lines indicate the furnace temperature at T_fur = 845 °C for t_heat = 15 min and for t_heat = 30 min, respectively. The γ-ray yield in the time range between 0 and 18 min was normalized to 100%. J. Phys. Soc. Jpn. 83, 083201 (Ref. 62).

The separation efficiency of ^99mTc, ε_sp, was derived by comparing the 141 keV γ-ray intensity obtained by the CZT-1 detector after separation, Y_sep, with that before the separation Y_unt. Y_sep and Y_unt are given as the sum of the intensities of the 141 keV γ-ray emitted from decays of the 143 keV state (^99mTc), Y(Tc), and the 181 keV state: Y(Mo). Using the Bateman equation for the parent-daughter decay, Y(Tc) and Y(Mo) are given as follows:   

\begin{equation} Y_{\text{unt}} = [Y\text{(Mo)} +Y \text{(Tc)}], \end{equation} [2]

  

\begin{equation} Y_{\text{sep}} = [Y\text{(Mo)} + (1-\varepsilon_{\text{sp}}) \times Y\text{(Tc)}], \end{equation} [3]

  

\begin{equation} Y\text{(Mo)} = A\text{(Mo)}_{0} \times \exp(-\lambda_{1}\mathrm{t}) \times 0.047, \end{equation} [4]

  

\begin{align} Y\text{(Tc)} &= [0.88 \times \lambda_{2} \times (\lambda_{2} - \lambda_{1})^{-1} \times A\text{(Mo)}_{0} \\ &\quad \times \{\exp(-\lambda_{1}\mathrm{t}) - \exp(-\lambda_{2}\mathrm{t})\}] \times 0.89, \end{align} [5]

where A(Mo)₀ is the ⁹⁹Mo activity at t = 0; λ₁ and λ₂ are the decay constants of ⁹⁹Mo and ^99mTc, respectively. In Eq. [4], the last number of 0.047 comes from the intensity of the 40 keV transition from the 181 keV state to the 141 keV one, 0.053, and the total internal conversion coefficient of the 141 keV transition, 0.119; hence, the 141 keV γ-ray transition intensity is 0.053/1.119 = 0.047.⁶²⁾ In Eq. [5], the first 0.88 in the brackets comes from the fact that 88% of the parent ⁹⁹Mo decays to ^99mTc, and the last 0.89 outside the brackets is obtained by taking account of the 141 keV transition intensity per disintegration of ^99mTc and the total internal conversion coefficient of the 141 keV transition. Hence, the ratio R of Y_sep to Y_unt is given as follows:   

\begin{align} \mathrm{R}& = \frac{Y_{\text{sep}}}{Y_{\text{unt}}} \\&= \frac{[\exp (- 0.0105t) \times 0.047 + (1 - \varepsilon_{\textit{sep}}) \times 0.85 \times \{\exp (- 0.0105t) - \exp(- 0.1155t)\}]}{[\exp (- 0.0105t) \times 0.047 + 0.85 \times \{\exp (- 0.0105t) - \exp(- 0.1155t)\}]}. \end{align} [6]

Using Eqs. [2] to [6], high separation efficiencies of close to 90 and 70% on average were obtained for molten MoO₃ samples of 4.0 and 8.5 mm thickness, respectively, in repeated milking processes at T_fur = 845 °C for t_heat = 15 min, as given in Table 12.⁶²⁾

Table 12. Obtained value of R, the separation efficiency and the diffusion coefficient of ^99mTc from molten MoO₃ samples with 4.0 and 8.5 mm thicknesses for each run. ⟨D⟩ is the average value of the diffusion coefficient. T_fur and t_heat are the furnace temperature and the heating time of the furnace, respectively

Run	Sample depth (mm) (Mass (g))	Tfur	theat	R	εsp (%)	D (cm2 sec−1)	⟨D⟩ (cm2 sec−1)
1	4.0 (9.87)	845	15	0.105	95	7.9 × 10−5	6.3 × 10−5
2		845	15	0.165	89	6.5 × 10−5
3		845	15	0.163	89	5.8 × 10−5
4	(7.69)	845	15	0.172	88	5.0 × 10−5
1	8.5 (14.3)	845	15	0.333	71	1.4 × 10−4	1.4 × 10−4
2		845	15	0.354	68	1.4 × 10−4
3		845	30	0.203	84	1.5 × 10−4
4	(11.9)	855	15	0.285	76	1.4 × 10−4

The diffusion coefficient, D, of ^99mTc within the molten MoO₃ samples was derived as follows. So far, the D value of a radioactive ion in a molten sample had not yet been obtained by a release measurement of the ion. In this study, by referring to the work on foil targets for the production of radioactive ion beams at the Isotope Separator On-Line (ISOLDE) at CERN,⁶⁷⁾ we divided the release process of ^99mTc into the following two steps: 1) the transport of ^99mTc to the surface of the molten MoO₃ sample, and 2) the evaporation of ^99mTc from the surface of the molten MoO₃ sample. In Ref. 18 the D of radioactive ions in metal targets, which were heated to high temperatures, was determined by assuming that step 2) can be neglected. Based on the same assumption, and by referring to work on the self-diffusion of radioactive ions in a sodium tungstate solution using ¹⁸⁷W (T_1/2 = 24 h) with a capillary method,⁶⁸⁾ in which the evaporation process of ¹⁸⁷W was not included, the diffusion coefficient of ^99mTc could be derived. A fraction of the original amount of ¹⁸⁷W, which is left in a capillary cell at the end of the diffusion, δ, is given using Fick’s law of diffusion:   

\begin{equation} \delta \approx (8/\pi^{2}) \times \mathrm{e}^{-\theta}, \end{equation} [7]

where the parameter θ is given by θ = π²Dt/(4h²), t is the time of diffusion, and h is the length of the capillary cell.

Using Eq. [7] and neglecting the evaporation process of ^99mTc, the averaged values of D, ⟨D⟩, of ^99mTc for the molten MoO₃ samples with 4.0 and 8.5 mm thicknesses could be derived, as presented in Table 12, where ⟨D⟩ is given by correcting the change of the sample thickness during the repeated milking process. The ⟨D⟩ value of about 1 × 10⁻⁴ cm²/s is much larger than a reported value of 7.71 × 10⁻¹² cm²/s at 780 °C for a powder sample.

In this study high separation efficiencies of about 90 and 70% were successfully obtained through a repeated milking process by the thermo-separation of ^99mTc from 10 and 14 g molten MoO₃ samples with thicknesses of 4.0 and 8.5 mm. By further developing the thermo-chromatography separation system we achieved a higher separation efficiency of over 90% for an irradiated MoO₃ sample of about 100 g. In these studies, the irradiated MoO₃ samples were melted in every milking process, and therefore the process could be performed under the same condition as that of the MoO₃ sample irrespective of the number of processes.

The diffusion coefficients of ^99mTc were found to be very large, and therefore ^99mTc could diffuse very rapidly in a thick molten sample within a reasonable heating time of 15 to 30 min, and subsequently evaporate from the sample. The present result solves the long-standing problems concerning the thermo-separation of ^99mTc from a MoO₃ sample with an increase in the sample mass or with repeated sublimation, and will bring a major breakthrough in the production of high-quality ^99mTc by using a massive ¹⁰⁰Mo sample.

4.1.2. Quality test of ^99mTc from ⁹⁹Mo separated by thermochromatography.

Here, we discuss the safety and efficacy of the “desired” ^99mTc radiopharmaceutical to assure parenteral administration to a patient. The United States Pharmacopeia (USP) contains regulatory requirements concerning the radionuclide purity and the radiochemical purity of ^99mTc and the concentration of aluminum (Al) in the ^99mTc product used to prepare radiopharmaceuticals.⁶⁹⁾ Namely, the amount of ⁹⁹Mo and the total concentration of all other β⁻ and γ-ray emitters in the ^99mTc product must be less than 0.015% and 0.01%, respectively. The radiochemical purity of ^99mTcO₄⁻ (pertechnetate) in a saline solution must be above 95% and the chemical purity of ^99mTc must be above 90%. Note that chemical impurities generate “undesired” ^99mTc compounds, such as free ^99mTcO₄⁻, which does not bind to a ligand and thus do not accumulate in a targeted organ of a patient, and thus lead to an extra radiation dose to non-targeted organs of the patient and to cause serious errors in diagnosis. The Al concentration must be less than 10 ppm. Endotoxin, known to be a pyrogen, is another important item, since even small amounts of endotoxin can cause illness in humans. The USP sets the maximum endotoxin concentration limit to be 175 EU/V, where EU is endotoxin units and V is the maximum recommended total dose in milliliters (mL).

Currently, ^99mTcO₄⁻, which meets the USP requirements, is obtained from a ⁹⁹Mo/^99mTc generator, in which the fission-⁹⁹Mo is loaded to an alumina column and ^99mTc in the form of ^99mTcO₄⁻ is repeatedly eluted from the column in a saline solution. Chemical impurities, which inhibit the labeling of the ^99mTc radiopharmaceutical complex, generate “undesired” ^99mTc compounds, such as free ^99mTcO₄⁻ and hydrolyzed-reduced ^99mTc.

It should be noted that the current USP sets those pharmacopeia standards for ^99mTcO₄⁻ obtained from a ⁹⁹Mo/^99mTc generator, but there are no pharmacopeia standards for ^99mTcO₄⁻, which is obtained by a ⁹⁹Mo (or ^99mTc) production method other than one with fission-⁹⁹Mo.³⁾ Therefore, it is important to investigate the pharmaceutical equivalence of ^99mTcO₄⁻ to that obtained from the alumina-based ⁹⁹Mo/^99mTc generator. In fact, such a study has been conducted concerning the production of ^99mTc by the ¹⁰⁰Mo(p,2n)^99mTc reaction⁷⁰⁾ and ⁹⁹Mo by the ¹⁰⁰Mo(γ,n)⁹⁹Mo reaction,⁷¹⁾ but it has not yet been performed in the production of ^99mTc by the ¹⁰⁰Mo(n,2n)⁹⁹Mo reaction. Hence, we have tested the pharmaceutical equivalence of ^99mTcO₄⁻ obtained by ^99mTc from the thermochromatographic separation procedure to that obtained from the alumina-based ⁹⁹Mo/^99mTc generator. In addition, we have studied quality-control specifications associated with new variables, such as the contamination arising from an enriched ¹⁰⁰Mo sample. The USP does not set any criteria concerning the nonradioactive (stable) Mo content in ^99mTc-radiopharmaceuticals because an enriched ²³⁵U sample, which does not contain a stable Mo, and the ⁹⁹Mo/^99mTc generator are used for the production of ^99mTc using the fission-⁹⁹Mo. We have adopted a dosage limit of 1,700 µg/day, which is given as the injection agent of stable Mo in a report of the International Conference on Harmonization Guideline for elemental impurities (ICH Q3D),⁷²⁾ which does not cover radiopharmaceuticals, but “is intended to provide guidance for registration applications on the content and qualification of impurities in new drug substances produced by chemical syntheses”. We have also studied quality-control tests of ^99mTc-radiopharmaceuticals commonly used for the imaging of brain perfusion (^99mTc-ECD), myocardial perfusion (^99mTc-MIBI), and kidney (^99mTc-MAG3), to ensure the safe clinical use of ^99mTc obtained by the ¹⁰⁰Mo(n,2n)⁹⁹Mo reaction. The separation of ^99mTc from an irradiated ¹⁰⁰MoO₃ sample was carried out by the thermochromatographic method,^62),63) discussed in chapter 4.1.

The quality-control tests of a ^99mTcO₄⁻ saline solution on the radiochemical purity and the radiochemical yields of the ^99mTc-radiopharmaceuticals, such as ^99mTc-ECD, ^99mTc-MIBI, ^99mTc-MAG3, and ^99mTc-MDP, were performed by paper chromatography and by thin-layer chromatography, respectively. The radionuclide purity was studied by taking a γ-ray spectrum of the purified ^99mTcO₄⁻ solution using a high-purity Ge (HPGe) detector. The Al concentration of separated ^99mTc was checked by using an Al test paper. Details of preparing ^99mTc radiopharmaceuticals using commercially available labelling kits (FUJIFILM RI Pharma Co., Ltd., Japan) are given in Ref. 73. The stable Mo content in a 3 mL ^99mTcO₄⁻ saline solution was measured by inductively coupled plasma-atomic emission spectroscopy (ICP-AES). Tests of the endotoxin concentrations were carried out by following the statement that the pyrogenicity of a ^99mTcO₄⁻ solution from a particular production procedure should be verified by having a portion of it tested by an independent qualified professional using accepted procedures. The results of the quality assessments of the ^99mTcO₄⁻ saline solution and ^99mTc-radiopharmaceuticals were shown to satisfy the USP requirements listed in Table 13.⁷³⁾ The endotoxin concentrations were below the limit of detection (0.03 EU/mL), much less than the established limit in pharmacopoeias. The measured maximum value of the stable Mo content was 0.138 ppm, i.e., 0.138 µg/mL, which is much less than the permitted daily exposure of 1,700 µg/day given in the ICH Q3D guideline. These results provide important evidence that ^99mTc prepared by thermochromatographic separation using ⁹⁹Mo produced by the ¹⁰⁰Mo(n,2n)⁹⁹Mo reaction can be a promising substitute for the fission product ⁹⁹Mo.

Table 13. Results of quality control tests of a ^99mTcO₄⁻ saline solution and ^99mTc-radiopharmaceuticals along with USP specifications

Parameter	USP	Exp. 1	Exp. 2	Exp. 3	Exp. 4
pH	4.5 to 7.5	7.23	7.16	6.66	6.58
Endotoxins	<175 EU/V	<0.03 EU/mL	>0.03 EU/mL	<0.03 EU/mL	<0.03 EU/mL
Radionuclidic purity - non fission	99Mo/99mTc < 0.015%	<0.015%	<0.015%	<0.015%	<0.015%
	Other gammas/99mTc < 0.05%	<0.05%	<0.05%	<0.05%	<0.05%
Aluminium	<10 ppm	<10 ppm	<10 ppm	<10 ppm	<10 ppm
Molybdenum	Not specified	—	0.138 ppm	0.020 ppm	0.034 ppm
Radiochemical purity of 99mTcO4−	>95%	99.80%	99.90%	99.59%	99.52%
Radiochemical yield of 99mTc-radiopharmaceuticals	>90%	99mTc-MIBI	99mTc-ECD	99mTc-MAG3	99mTc-MDP
		99.98%	98.27%	97.23%	99.95%

4.1.3. Recovery of an irradiated ¹⁰⁰MoO₃ sample.

It is important to recycle an enriched (expensive) ¹⁰⁰Mo sample irradiated by neutrons, because a loss fraction of the quantity of the ¹⁰⁰Mo sample of 100 g during neutron irradiation for 24 h is estimated to be small, about 0.0001%. Here, accelerator neutrons are assumed to be produced by the ^natC(d,n) reaction using 40 MeV, 2 mA deuteron beams, similarly to the estimation given in Table 5. We have therefore developed a recovery method of ¹⁰⁰MoO₃ having a recovery efficiency, ε_rec, higher than 99% to mitigate the cost of a ¹⁰⁰MoO₃ sample loss.⁷⁴⁾ So far, the recovery efficiency, ε_rec, of the ¹⁰⁰MoO₃ samples in the range of 84–97% has been reported in a study of the thermochromatography of ^94mTc (T_1/2 = 52 min) from ⁹⁴MoO₃⁷⁵⁾; a recovery efficiency of 87%⁶⁶⁾ or 90%⁷⁶⁾ was obtained in a ⁹⁹Mo or ^99mTc production study based on the ¹⁰⁰Mo(γ,n)⁹⁹Mo or ¹⁰⁰Mo(p,2n)^99mTc reactions by employing a chemical process.

The present recovery test was performed using a home-made electric furnace, shown in Fig. 18.

Fig. 18.

(Color online) (a) Schematic view of thermochromatography system. 1, Electric furnace (three individual heating zones); 2, Crucibles for MoO₃ samples; 3, MoO₃ needle crystal holder; 4, Quartz wool; 5, ^99mTc condensation region; 6, CZT detector; 7, Moist oxygen inlet. (b) Photograph of sample shelf holding three crucibles with ¹⁰⁰MoO₃. J. Phys. Soc. Jpn. 86, 053201 (Ref. 74).

An irradiated enriched ¹⁰⁰MoO₃ sample of 26.450 g was divided into three platinum crucibles in a vertical three-zone tubular electric furnace along with 103.253 g of non-neutron-irradiated ¹⁰⁰MoO₃ to bring the sample mass up to 129.703 g so as to develop a recovery method of over 100 g of ¹⁰⁰MoO₃. After the milking process was carried out eight times in total, ¹⁰⁰MoO₃ was recovered in two batches from the first five and the following three milking processes. The recovery was focused on the two main deposition sources (the sample remaining in crucibles and needle crystals) in the first batch (run 1), and the detailed distribution of ¹⁰⁰MoO₃ was investigated in the second batch (run 2), including in the quartz tube, the platinum shelf, and the quartz wool. The recovery was studied gravimetrically by measuring the weight of ¹⁰⁰MoO₃ remaining in the crucibles and being deposited onto any quartz pieces that ¹⁰⁰MoO₃ could travel through. All of the crucibles and quartz pieces were weighed before and after the thermochromatography to determine the amount of recovered ¹⁰⁰MoO₃ free of any possible contamination caused by the process. Firstly, the ¹⁰⁰MoO₃ sample remaining in the crucibles was melted at 830 °C in an electric furnace and collected into a quartz test tube using a funnel. Secondly, the ¹⁰⁰MoO₃ needle crystals were collected by washing them off the holder with pure water, and then evaporated to dryness to measure the weight. Thirdly, the quartz wool that trapped ¹⁰⁰MoO₃ crystals was heated to above 830 °C, and thereby the ¹⁰⁰MoO₃ alone was separated from the quartz wool by thermochromatography. Any quartz pieces, including the funnel used to channel molten ¹⁰⁰MoO₃ into the quartz test tube, were washed with pure water in an ultrasonic bath to recover the most ¹⁰⁰MoO₃ possible in run 2. The collected ¹⁰⁰MoO₃ crystals were evaporated to dryness to measure the weight.

The distribution and recovery yield of the ¹⁰⁰MoO₃ mass after thermochromatography is summarized in Table 14. Concerning run 1, after thermochromatography, 118.870 g (92%) out of the initial ¹⁰⁰MoO₃ mass of 129.703 g was found to remain in the crucibles, while 9.477 g (87%) out of the 10.833 g of ¹⁰⁰MoO₃ that was vaporized from the crucibles was trapped as needle crystals. During the recovery process of 118.870 g of ¹⁰⁰MoO₃ in the crucibles, a certain amount of ¹⁰⁰MoO₃ was vaporized and then trapped as needle crystals. Hence, the recovery from the needle crystals was 10.296 g, which was more than the needle crystal mass of 9.477 g measured after thermochromatography. Finally, 117.502 g (99%) out of 118.870 g of ¹⁰⁰MoO₃ was recovered from the crucibles and 10.296 g from the needle crystal holder, which gave a recovery yield of 98.5% (127.798/129.703 = 0.985). For run 2, 110.551 g (94%) out of the initial ¹⁰⁰MoO₃ mass of 117.490 g was found to remain in the crucibles after thermochromatography. The vaporized ¹⁰⁰MoO₃ from the crucibles (6.939 g) was deposited onto the quartz tubes (0.073 g), quartz wool (0.376 g), and a needle crystal holder (6.380 g), and was 6.829 g in total. The recovered ¹⁰⁰MoO₃ from the crucibles was 108.640 g (92%), and that from other than the crucibles was 8.426 g, giving a total recovery of 117.066 g (99.6%). Note that 1.911 g of ¹⁰⁰MoO₃ was vaporized from the crucibles and then trapped as needle crystals during the recovery process, which resulted in the total amount of recovered ¹⁰⁰MoO₃ deposited, other than in the crucibles (8.740 g) being larger than the amount of needle crystals measured after the separation.

Table 14. Recovery yield and distribution of ¹⁰⁰MoO₃ after ^99mTc thermochromatography and the recovery procedure

	Run 1 (mass/g)	Run 2 (mass/g)
Distribution
Before separation
1) Crucibles	129.703	117.490
After separation
1) Crucibles	118.870	110.551
2) Other than crucibles	10.833	6.939
a) Platinum shelf		0
b) Quartz tubes		0.073
c) Quartz wool		0.376
d) Needle crystal holder	9.477	6.380
Recovery yield
1) Crucibles	117.502	108.640
2) Other than crucibles (Needle crystal holder)	10.296	8.426
Total	127.798 (98.5%)	117.066 (99.6%)

A high recovery yield of 99% was obtained, which would significantly reduce any financial damage due to the loss of the enriched ¹⁰⁰MoO₃ sample. We consider that the newly developed home-made thermochromatography system should have a capability of nearly 100% recovery, because all of the ¹⁰⁰MoO₃, including the small unrecovered amount, is kept within the thermochromatographic apparatus inside the electric furnace.

4.2. Separation of ⁶⁴Cu and ⁶⁷Cu and biodistribution of ⁶⁷CuCl₂ in tumor bearing mice.

4.2.1. Separation of ⁶⁴Cu and ⁶⁷Cu from Zn.

We have developed a radiochemical separation of ⁶⁴Cu and ⁶⁷Cu produced by the ⁶⁴Zn(n,p)⁶⁴Cu, ⁶⁷Zn(n,p)⁶⁷Cu and ⁶⁸Zn(n,x)⁶⁷Cu reactions using ^natZnO or ⁶⁴ZnO samples.⁷⁷⁾ A flowchart of the separation steps of the irradiated samples is given in Fig. 19.

Fig. 19.

Flow chart of ^64,67Cu separation from irradiated ZnO samples and the recovery of ZnO. J. Radioanal. Nucl. Chem. 303, 1205–1209 (Ref. 77).

The irradiated sample of 5.225 g was dissolved in 20 ml of 36 wt% HCl, which was passed through an ion-exchange column for adsorbing the Cu ions, and thus separating Zn. The ^64,67Cu was then eluted with 20 ml of 2.0 M HCl, passed through an anion-exchange column to remove traces of Zn, followed by washing with 10 ml 2.0 M HCl to obtain purified ^64,67Cu radionuclides. The collected efficiency of ⁶⁴Cu separated from the irradiated ⁶⁴ZnO sample was 96%. The time required for the column separation process was 3–4 h. We also developed a method for recycling irradiated enriched ⁶⁴Zn and ⁶⁸Zn samples after radiochemical separation. The recovery efficiency of the ^natZnO sample was demonstrated to be over 95% in a cold (non-radioactive) run using an alkaline precipitation method. The purified ^64,67Cu solution was reacted with a bifunctional ligand used for antibody labelling. The labelling yield was determined by thin-layer chromatography (TLC) to be good at 92–97% which was satisfactory for clinical radiotherapy applications.

4.2.2. Biodistribution of ⁶⁷CuCl₂ in tumor-bearing mice.

We developed new production routes to improve the low-availability of the promising radionuclide of ⁶⁴Cu and ⁶⁷Cu, and to establish a radiochemical method for obtaining high-quality ⁶⁴Cu and ⁶⁷Cu from neutrons irradiated Zn samples. Cu-based radiopharmaceuticals that can accumulate in cancer cells, such as ⁶⁴Cu-labeled proteins, peptides, and antibodies, have been developed and widely used.⁴⁾ However, currently ⁶⁴Cu complexes are considered to have relatively low stability in vivo, which could cause the loss of ⁶⁴Cu from the complexes, leading to less accumulation of ⁶⁴Cu in targeted cancer cells by producing free radioactive ⁶⁴Cu.^{5) 64}Cu chloride (⁶⁴CuCl₂) has been identified as a potential agent for PET imaging and radionuclide therapy.⁷⁸⁾ In a study using ⁶⁴CuCl₂ relevant to radionuclide therapy, it was demonstrated that Cu metabolism is important for many cancers. Here, it is worth noting that compared with ⁶⁴Cu-labeled complexes, ⁶⁴CuCl₂ has simple radiochemistry without a radiolabeling process. The results prompted us to measure the biodistribution of ⁶⁷CuCl₂ in colorectal tumor-bearing mice. Colorectal cancer is a major cause of death in Japan.⁷⁹⁾

⁶⁷Cu was produced by irradiating an ⁶⁸ZnO (99.935% enriched in ⁶⁸Zn) sample with neutrons at TIARA-QST.⁴²⁾ The chemical separation of ⁶⁷Cu from a neutron-irradiated ⁶⁸ZnO sample was performed by slightly modifying the previously reported method to separate Ga ions.⁷⁷⁾ The radionuclide purity of ⁶⁷Cu was 99.8% at the time of injection. The specific activity [MBq/(µg Cu)] of ⁶⁷Cu was determined to be 4.5 MBq/(µg Cu) at EOI by the titration method. This value is much smaller than the typical specific activity of ⁶⁴Cu produced by the ⁶⁴Ni(p,n)⁶⁴Cu reaction in the range of 2.4–11 GBq/(µg Cu) quoted from a recent study on the biodistribution of ⁶⁴CuCl₂ in rats.⁷⁸⁾ It is considered that the specific activity plays an important role in radiolabeling and the in vivo biodistribution of radioactive tracers. Hence, it is very interesting to study the role of ⁶⁷CuCl₂ with low specific activity in the biodistribution of ⁶⁷Cu ions in colorectal tumor-bearing mice. Animal procedures were carried out according to a protocol approved by the QST Institutional Animal Care and Use Committee. The ⁶⁷Cu solution after radiochemical purification was diluted with a physiological saline solution for injection into mice. When the tumors were palpable, the mice were intravenously injected with 35 or 50 kBq of ⁶⁷CuCl₂ dissolved in 100 µl saline via a tail vein. After the mice were sacrificed at 0.5, 1, 4, 8, 24, and 48 h post-injection (n = 4, four mice at a time), their blood and organ samples of interest (liver, kidney, intestine without content, spleen, pancreas, stomach, heart, lung, muscle, bone, brain, and tumor) were removed and weighed.⁸⁰⁾

The radioactivities in the blood and organ samples were measured by using a well-type NaI(Tl) detector. The biodistribution of ⁶⁷CuCl₂ in tumor-bearing mice was determined, as shown in Fig. 20. Note that it is common in animal studies to express the biodistribution of radiotracers using the parameter %ID/g of tissue, defined as the radioactivity in a particular tissue at each time point as a percentage of the total radioactivity injected into the animal, which was further divided by the weight of each tissue. It is very interesting that a high uptake of ⁶⁷Cu in the tumor was found, which may indicate an important role of Cu metabolism in colorectal cancer. The accumulation of ⁶⁷Cu in the tumor was 7.0 ± 1.4%ID/g at 48 h, comparable to that of ⁶⁴Cu, ∼5%ID/g, in spite of the difference in the specific activities. A high uptake of ⁶⁷Cu was also observed in the organs, such as the liver and kidney. The ⁶⁷Cu uptake in the liver and kidney gradually decreased over time from 0.5 to 48 h. The biodistribution of ⁶⁷CuCl₂ determined by using very low-specific-activity ⁶⁷Cu is similar to the recent biodistribution of ⁶⁴CuCl₂ obtained by using high-specific-activity ⁶⁴Cu in malignant melanoma tumor-bearing mice.⁷⁸⁾ The observed uptake of ⁶⁷Cu in these organs is considered to be due to copper metabolism being independent of the specific activity of ⁶⁷Cu.

Fig. 20.

(Color online) Biodistribution of ⁶⁷CuCl₂ in tumor-bearing mice with standard deviation. J. Phys. Soc. Jpn. 86, 023201 (Ref. 80).

In summary, ⁶⁷CuCl₂ was used for the first time to determine the biodistribution in colorectal tumor-bearing mice. A high uptake of ⁶⁷Cu in the tumor was found, although the specific activity of ⁶⁷Cu was low owing to the neutron intensity currently available. This result suggests that ⁶⁷CuCl₂ can be a potential radionuclide agent for cancer radiotherapy.

5. Deuteron accelerator and neutron source

5.1. Deuteron accelerator.

Deuteron beams, provided mostly by linear accelerators, are used to produce high neutron fluxes by irradiating a light element, such as carbon, liquid Li, and Be, for example in the projects SPIRAL2, the Soreq Applied Research Accelerator Facility (SARAF) project in Israel,⁸¹⁾ and the International Fusion Materials Irradiation Facility (IFMIF)⁸²⁾ et al., where a fixed neutron energy of 14 MeV is mostly needed. However, in order to produce a particular medical radioisotope by neutron-induced reaction using accelerator neutrons, a wide variety of the neutron energy is necessary, because a neutron-induced reaction cross section on samples depends on the neutron energy. Therefore, the neutron energy needs to be easily tuned to an energy suitable for the production of medical radioisotopes by changing the energy of the deuteron beams. In addition, we will use the same accelerator to produce not only accelerator neutrons, but also proton and deuteron beams, for generating a wide variety of medical radioisotopes via proton- and deuteron-induced reactions on a sample. In considering that accelerators have an active lifespan of over 30 years, and that interest in new medical radioisotopes will be continuously growing, the expected newly installed accelerators must have a capability for producing a wide variety of medical isotopes. We choose AVF cyclotrons with 50 MeV, 2 mA deuteron beams to meet the mentioned requirements. A fixed radiofrequency AVF cyclotron is robust in operation, compact in size, and relatively cheap compared to a linear accelerator. Such cyclotrons can be constructed by many cyclotron companies around the world. In fact, Sumitomo Heavy Industries, Ltd. has been constructing AVF cyclotrons, which can provide a 30 MeV, 1 mA H⁻ beam for Boron Neutron Capture Therapy (BNCT).⁸³⁾ Here, it is worth mentioning that the beam intensities from cyclotrons are limited by an extraction device (deflector), and therefore negative deuteron D⁻ ions should be accelerated up to 50 MeV. The principal advantage of a D⁻ ions cyclotron is the ease and low loss in extraction by the stripping of D⁻ ions into positive deuteron (D⁺) ions on a thin carbon foil with a thickness of about 500 µg/cm². The D⁺ ions can be extracted to a beam line through a residual magnetic field in the AVF cyclotron.

5.2. Accelerator neutron source.

So far, various types of accelerator-based neutron sources with high neutron fluxes that have kinetic energy above a few MeV have been developed for fundamental studies in nuclear physics and nuclear astrophysics, radiation-resistant materials irradiation testing for fusion reactors, boron neutron capture therapy, and slow neutron scattering etc. In these cases, accelerator neutrons are generated by the ⁷Li(p,n)⁷Be, ³H(d,n)⁴He, and ⁹Be(p,n)⁹C reactions, and the spallation reaction by bombarding a liquid mercury target, a liquid bismuth-lead target with high-energy proton beams.

In medical radioisotope productions in reactors, thermal neutron fluxes of a factor × 10¹⁴ n/(cm² s) have been used. Accelerator neutrons with a quasi-monoenergy of 14 MeV with about 10¹² n/(cm² s) were used in the field of nuclear engineering at Fusion Neutronics Facility (FNS) of the Japan Atomic Energy Agency (JAEA),⁵¹⁾ and at a facility with a rotating target neutron source (RTNS-II) in U.S.A.,⁸⁴⁾ respectively. Because the ³H target is radioactive, and neutrons from the ³H(d,n)⁴He reaction at E_d = 300 keV are emitted isotopically with respect to the deuteron beam direction, this leads to a disadvantage concerning the effective use of neutrons for RIs productions. Another intense neutron source based on the deuteron breakup reaction was proposed by P. Grand and A. N. Goland.⁸⁵⁾ Note that Helmholz et al. first observed the breakup process,⁸⁶⁾ and also found that an intense forward-directed beam of neutrons is emitted when a target with a low-atomic number, such as Li or Be, is bombarded with deuteron beams. They also designed a high-flux neutron generator system composed of a 35 MeV high-current deuteron linear accelerator and a molten Li target configuration. Based on the proposed neutron source, at the IFMIF, intense neutron fluxes of greater than 10¹⁵ n/(cm² s) with the energy spectrum peaking at around 14 MeV are expected to be produced by bombarding a liquid lithium jet target with intense deuteron beams of about 35–40 MeV. At the SPIRAL2 facility, neutrons with a high flux of 10¹⁵ n/s are planned to be produced by ^natC(d,n) using 40 MeV 5 mA deuterons provided from a linear accelerator. At the SARAF facility, there is an ongoing project with the superconducting light/heavy-ion LINAC, with a potential of about 40 MV, capable of accelerating 5 mA deuterons up to 40 MeV. The neutrons are produced by irradiating deuterons on a liquid Li target.

We have also installed a mini-type rotating carbon target system used for producing accelerator neutrons provided by the ^natC(d,n) reaction in collaboration with Sumitomo Heavy Industries, Ltd. A schematic view of the neutron target is shown in Fig. 21. The carbon target was shown to work successfully under a thermal power of 40 kW using the JAERI Electron Beam Irradiation System (JEBIS) at Japan Atomic Energy Agency (now QST), which can provide 20–100 keV electron beams with an output beam power of 400 kW.⁸⁷⁾

Fig. 21.

(Color online) Schematic view of the neutron target system.

6. Conclusions and future prospects

We proposed an innovative method to produce a wide variety of medical radioisotopes. By overcoming challenges in the proposed method by fundamental studies, we have successfully carried out all of the important steps necessary to obtain high-quality ^99mTc suitable for formulating ^99mTc radiopharmaceuticals. In addition, we could validate the high capability of accelerator neutrons so far unexplored to produce a large amount of high-quality therapeutic radioisotopes conducting detailed studies of ⁶⁴Cu and ⁶⁷Cu. In order to secure a constant and reliable supply chain of ⁹⁹Mo for domestic use and to promote a theranostics approach in personalized nuclear medicine, we have presented a proposal for a prototype facility for the Generation of Radioisotopes with Accelerator Neutrons by Deuterons (GRAND).⁴⁶⁾ The proposed system consists of an AVF cyclotron with a 50 MeV, 2 mA deuteron beam intensity and a rotating carbon target system to produce intense accelerator neutrons. In the cyclotron negative deuteron ions (D⁻) are accelerated up to 50 MeV, and by passing them through a stripper foil, D⁻ beams are converted to D⁺ beams, which are extracted from the cyclotron to a beam-transport system for irradiating a carbon target to produce accelerator neutrons. The principal advantage of a negative deuteron-cyclotron is the ease and low loss in extraction by the stripping of negative deuteron ions into positive deuteron (D⁺) ions on a thin carbon foil with a thickness of about 500 µg/cm². The D⁺ ions can be extracted to two beam lines through a residual magnetic field in the AVF with a possibility to irradiate two different targets simultaneously. The layout of the accelerator is shown in Fig. 22.

Fig. 22.

(Color online) Layout of the accelerator facility for radioisotope production. J. Phys. Soc. Jpn. 82, 064201 (Ref. 46).

In April 2020, the two-year “Deuteron Accelerator for Theranostics mEdicine (DATE)” project at Tohoku University was started. In this project we plan to accelerate 25–40 MeV deuteron beams with an intensity of 100 µA by newly setting up a negative deuterium ion source and a stripper foil for the existing AVF cyclotron at CYRIC at Tohoku University.³⁴⁾ The deuteron beam intensity will be about twenty-times stronger than the presently available D⁺ beam intensity of about 5 µA, and about one-twentieth of the intensity that will be obtained at the GRAND project. At CYRIC, a negative hydrogen beam of 50 MeV with 22 µA was successfully obtained in 2003, which demonstrated the feasibility of high-current acceleration and extraction for proton beams with H⁻ acceleration. The DATE project will play an important role in the on-demand medical RIs production for promoting a theranostics approach in personalized nuclear medicine and also in detailed planning for a prototype facility, GRAND.

Acknowledgements

I would like to express my sincere thanks to all collaborators who have contributed in exciting developments discussed in this article. Special thanks to K. Hashimoto, K. Tsukada, S. Motoishi, H. Saeki, M. Kawabata, F. Minato, S. Hashimoto, C. Konno, Y. Kawauchi, N. Takeuchi, Y. Adachi, and T. Ogawa for their great contributions to these developments. I am grateful to operating crews of the FNS-JAEA accelerator, TIARA-QST cyclotron, and CYRIC accelerator for their skillful operation. The present work was supported by a Grant-in-Aid for Specially Promoted Research of the Japan Ministry of Education, Culture, Sports, Science and Technology (23000005) and in part by JSPS KAKENHI Grant Numbers JP 226540311, JP16K05383, JP19K03903, and the Adaptable and Seamless Technology Transfer Program through the Target driven R&D (A-STEP) of the Japan Agency of Medical Research and Development (AMED), and in part by Program on Open Innovation Platform with Enterprises, Research Institute and Academia, Japan Science and Technology Agency (JST, OPERA, JPMJOP1721).

Notes

Edited by Toshimitsu YAMAZAKI, M.J.A.

Correspondence should be addressed: Y. Nagai, Quantum Beam Science Research Directorate, National Institutes for Quantum and Radiological Science and Technology, Tokai, Naka-gun, Ibaraki 319-1106, Japan (e-mail: [e-mail]).

References

• 1)  IAEA. (2004) Nuclear Technology Review 2004. IAEA, Vienna. See pp. 62–77.
• 2)   Hevesy,  G. and  Paneth,  F. (1913) RaD als “Indikator” des Bleis. Z. Anorg. Chem. 82, 322–328.
• 3)   Niese,  S. (2017) George de Hevesy (1885–1966): Discoverer of hafnium, founder of radioanalytical chemistry and X-ray fluorescence analysis and father of nuclear medicine. J. Radioanal. Nucl. Chem. 311, 1035–1041.
• 4)   Cutler,  C.S.,  Hennkens,  H.M.,  Sisay,  S.,  Huclier-Markai,  S. and  Jurisson,  S.S. (2013) Radiometals for combined imaging and therapy. Chem. Rev. 113, 858–883.
• 5)  Srivastava, S.C. and Mausner, L.F. (2013) Therapeutic radionuclides: Production, physical characteristics, and applications. In Therapeutic Nuclear Medicine, Medical Radiology (eds. Baum, R.P.). Springer, Berlin, Heidelberg, pp. 12–50.
• 6)   Funkhouser,  J. (2002) Reinventing pharma: The theranostic revolution. Curr. Drug. Discov. 2, 17–19.
• 7)  European Commission Health and Consumers Directorate-General (2009) Preliminary Report on Supply of Radioisotopes for Medical Use and Current Developments in Nuclear Medicine. European Commission, Luxembourg.
• 8)   Ruth,  T. (2009) Accelerating production of medical isotopes. Nature 457, 536–537.
• 9)  Japan Radioisotope Association. (2019) Statistics on the distribution of radioisotopes in Japan. Japan Radioisotope Association, Tokyo.
• 10)   Dash,  A.,  Pillai,  M.R.A. and  Knapp,  F.F. Jr. (2015) Production of ¹⁷⁷Lu for targeted radionuclide therapy: Available options. Nucl. Med. Mol. Imaging 49, 85–107.
• 11)   Engle,  J.W. (2018) The production of Ac-225. Curr. Radiopharm. 11, 173–179.
• 12)  Firestone, R.B. and Ekström, L.P. (2004) Table of Radioactive Isotopes. Version 2.1. available on http://ie.lbl.gov/toi/index.asp.
• 13)   Perrier,  C. and  Segrè,  E. (1937) Some chemical properties of element 43. J. Chem. Phys. 5, 712–716.
• 14)   Segrè,  E. and  Seaborg,  G.T. (1938) Nuclear isomerism in element 43. Phys. Rev. 54, 772.
• 15)   Lawrence,  E.O. and  Edelfson,  N.E. (1930) On the production of high speed protons. Science 72, 376–377.
• 16)   Richards,  P.,  Tucker,  W.D. and  Srivastava,  S.C. (1982) Technetium-99m: An historical perspective. Int. J. Appl. Radiat. Isot. 33, 793–799.
• 17)   Merrill,  P.W. (1952) Spectroscopic observations of stars of class. ApJ 116, 21–26.
• 18)  OECD/Nuclear Energy Agency (NEA). Steering Committee for Nuclear Energy on Security of Supply of Medical Radioisotopes. (2019) The Supply of Medical Radioisotopes 2019 Medical Isotope Demand and Capacity Projection for the 2019–2024 Period. NEA/SEN/HLGMR(2019)1.
• 19)  National Research Council. (2009) Medical Isotope Production Without Highly Enriched Uranium. The National Academies Press, Washington, D.C.
• 20)  Ponsard, B. (2010) Mo-99 supply issues: Status report and lessons learned. In International Topical Meeting on Research Reactor Fuel Management (RRFM) (Marrakech, 2010). Report No. INIS-BE--10K0001.
• 21)  OECD/Nuclear Energy Agency (NEA). (2010) The Supply of Medical Radioisotopes An Economic Study of the Molybdenum-99 Supply Chain. OECD, NEA No. 6967.
• 22)  Bertsche, K. (2010) Accelerator production options for ⁹⁹Mo. In Proceedings of IPAC’10, Kyoto, Japan, pp. 121–123.
• 23)  OECD. Nuclear Development (2010) The Supply of Medical Radioisotopes: Review of Potential Molybdenum-99/Technetium-99m Production Technologies. OECD.
• 24)   Nagai,  Y. and  Hatsukawa,  Y. (2009) Production of ⁹⁹Mo for nuclear medicine by ¹⁰⁰Mo(n,2n)⁹⁹Mo. J. Phys. Soc. Jpn. 78, 033201.
• 25)   Plompen,  A.I.M.,  Smith,  D.L.,  Reimer,  P.,  Qaim,  S.M.,  Semkova,  V.,  Cserpák,  F. et al. (2002) Recent neutron activation cross section measurements. J. Nucl. Sci. Technol. 39, 192–197.
• 26)   Shibata,  K.,  Iwamoto,  O.,  Nakagawa,  T.,  Iwamoto,  N.,  Ichihara,  A.,  Kunieda,  S. et al. (2011) JENDL-4.0: A new library for nuclear science and engineering. J. Nucl. Sci. Technol. 48, 1–30.
• 27)   Fadil,  M. and  Rannou,  B., the SPIRAL2 project team. (2008) About the production rates and the activation of the uranium carbide target for SPIRAL 2. Nucl. Instrum. Methods Phys. Res. B 266, 4318–4321.
• 28)  Oak Ridge National Laboratory. (2015) High Flux Isotope Reactor (HFIR) USER GUIDE, Revision 2.0 (November 2015).
• 29)   Minato,  F. and  Nagai,  Y. (2010) Estimation of production yield of ⁹⁹Mo for medical use using neutrons from ^natC(d,n) at E_d = 40 MeV. J. Phys. Soc. Jpn. 79, 093201.
• 30)   Lhersonneau,  G.,  Malkiewicz,  T.,  Kolos,  K.,  Fadil,  M.,  Kettunen,  H.,  Saint-Laurent,  M.G. et al. (2009) Neutron yield from carbon, light- and heavy-water thick targets irradiated by 40 MeV deuterons. Nucl. Instrum. Methods Phys. Res. A 603, 228–235.
• 31)   Minato,  F.,  Tsukada,  K.,  Sato,  N.,  Watanabe,  S.,  Saeki,  H.,  Kawabata,  M. et al. (2017) Measurement and estimation of the ⁹⁹Mo production yield by ¹⁰⁰Mo(n,2n)⁹⁹Mo. J. Phys. Soc. Jpn. 86, 114803.
• 32)   Tsukada,  K.,  Nagai,  Y.,  Hashimoto,  K.,  Kawabata,  M.,  Minato,  F.,  Saeki,  H. et al. (2018) ⁹⁹Mo yield using large sample mass of MoO₃ for sustainable production of ⁹⁹Mo. J. Phys. Soc. Jpn. 87, 043201.
• 33)   Rosman,  K.J.R. and  Taylor,  P.D.P. (1998) Isotopic compositions of the elements 1997 (Technical Report). Pure Appl. Chem. 70, 217–235.
• 34)  Fujita, M., Terakawa, A., Suzuki, H., Endo, T., Yamazaki, A., Tanaka, E. et al. (2004) Present status of the K = 110 MeV AVF cyclotron at Tohoku University. In Proceedings of the 17th International Conference on Cyclotrons and Their Applications. Tokyo, pp. 143–145.
• 35)  Berger, M.J., Hubbell, J.H., Seltzer, S.M., Chang, J., Coursey, J.S., Sukumar, R. et al. (2010) XCOM: Photon Cross Sections Database; National Institute of Standards and Technology: Gaithersburg, MD.
• 36)   Sato,  T.,  Iwamoto,  Y.,  Hashimoto,  S.,  Ogawa,  T.,  Furuta,  T.,  Abe,  S. et al. (2013) Heavy Ion Transport code System, PHITS, version 2.52. J. Nucl. Sci. Technol. 50, 913–923.
• 37)   Boyd,  R.E. (1982) Technetium-99m generators—The available options. Int. J. Appl. Radiat. Isot. 33, 801–809.
• 38)   Pillai,  M.R.A.,  Dash,  A. and  (Russ) Knapp,  F.F. (2015) Diversification of ⁹⁹Mo/^99mTc separation: Non–fission reactor production of ⁹⁹Mo as a strategy for enhancing ^99mTc availability. J. Nucl. Med. 56, 159–161.
• 39)   Bennett,  R.G.,  Christian,  J.D.,  Petti,  D.A.,  Terry,  W.K. and  Grover,  S.B. (1999) A system of ^99mTc production based on distributed electron accelerators and thermal separation. Nucl. Technol. 126, 102–121.
• 40)   Ross,  C.,  Galea,  R.,  Saull,  P.,  Davidson,  W.,  Brown,  P.,  Brown,  D. et al. (2010) Using the ¹⁰⁰Mo photoneutron reaction to meet Canada’s requirement for ^99mTc. Phys. Canada 66, 19–24.
• 41)   Fujiwara,  T. (2013) Delivery PET fundamental and clinical studies. Month. Bull. N. Med. (Gekkan Shin-iryo) 40, 49–52 (in Japanese).
• 42)  Arakawa, K., Nakamura, Y., Yokota, W., Fukuda, M., Nara, T., Agematsu, T. et al. (1992) Construction and first year’s operation of the JAERI AVF cyclotron. In Proceedings of the 13th International Conference on Cyclotrons and Their Applications. Vancouver, Canada, pp. 119–122.
• 43)  Katakura, J.I. (2012) JENDL FP Decay Data File 2011 and Fission Yields Data File 2011. No. JAEA-DATA/CODE--2011-025. Japan Atomic Energy Agency, Tokai, Japan.
• 44)   Nagai,  Y.,  Iwamoto,  O.,  Iwamoto,  N.,  Kin,  T.,  Segawa,  M.,  Hatsukawa,  Y. et al. (2009) Production of an isomeric state of ⁹⁰Y by fast neutrons for nuclear diagnostics. J. Phys. Soc. Jpn. 78, 113201.
• 45)   Kin,  T.,  Nagai,  Y.,  Iwamoto,  N.,  Minato,  F.,  Iwamoto,  O.,  Hatsukawa,  Y. et al. (2013) New production routes for medical isotopes ⁶⁴Cu and ⁶⁷Cu using accelerator neutrons. J. Phys. Soc. Jpn. 82, 034201.
• 46)   Nagai,  Y.,  Hashimoto,  K.,  Hatsukawa,  Y.,  Saeki,  H.,  Motoishi,  S.,  Sato,  N. et al. (2013) Generation of radioisotopes with accelerator neutrons by deuterons. J. Phys. Soc. Jpn. 82, 064201.
• 47)   Smith,  N.A.,  Bowers,  D. and  Ehst,  D.A. (2012) The production, separation, and use of ⁶⁷Cu for radioimmunotherapy: A review. Appl. Radiat. Isot. 70, 2377–2383.
• 48)   Obata,  A.,  Kasamatsu,  S.,  McCarthy,  D.W.,  Welch,  M.J.,  Saji,  H.,  Yonekura,  Y. et al. (2003) Production of therapeutic quantities of ⁶⁴Cu using a 12 MeV cyclotron. Nucl. Med. Biol. 30, 535–539.
• 49)   Spahn,  I.,  Coenen,  H.H. and  Qaim,  S.M. (2004) Enhanced production possibility of the therapeutic radionuclides ⁶⁴Cu, ⁶⁷Cu and ⁸⁹Sr via (n, p) reactions induced by fast spectral neutrons. Radiochim. Acta 92, 183–186.
• 50)   Merrick,  M.J.,  Rotsch,  D.A.,  Tiwari,  A.,  Nolen,  J.,  Brossard,  T.,  Song,  J. et al. (2021) Imaging and dosimetric characteristics of ⁶⁷Cu. Phys. Med. Biol. 66, 035002.
• 51)  Nakamura, T., Maekawa, H., Ikeda, Y. and Oyama, A. (1983) A D-T neutron source for fusion neutronics experiments at the JAERI. In Proc. Int. Ion Eng. Congress–ISIAT ’83/IAPT ’83. Kyoto, Japan, Vol. 1, pp. 567–570.
• 52)   Mausner,  L.F.,  Kolsky,  K.L.,  Joshi,  V. and  Srivastava,  S.C. (1998) Radionuclide development at BNL for nuclear medicine therapy. Appl. Radiat. Isot. 49, 285–294.
• 53)   Sato,  N.,  Tsukada,  K.,  Watanabe,  S.,  Ishioka,  N.S.,  Kawabata,  M.,  Saeki,  H. et al. (2014) First measurement of the radionuclide purity of the therapeutic isotope ⁶⁷Cu produced by ⁶⁸Zn(n,x) reaction using ^natC(d,n) neutrons. J. Phys. Soc. Jpn. 83, 073201.
• 54)   Tsukada,  K.,  Nagai,  Y.,  Hashimoto,  S.,  Minato,  F.,  Kawabata,  M.,  Hatsukawa,  Y. et al. (2020) Anomalous radioisotope production for ⁶⁸ZnO using polyethylene by accelerator neutrons. J. Phys. Soc. Jpn. 89, 034201.
• 55)   Szelecsényi,  F.,  Boothe,  T.E.,  Takács,  S.,  Tárkányi,  F. and  Tavano,  E. (1998) Evaluated cross section and thick target yield data bases of Zn+p processes for practical applications. Appl. Radiat. Isot. 49, 1005–1032.
• 56)   Meulders,  J.P.,  Leleux,  P.,  Macq,  P.C. and  Pirart,  C. (1975) Fast neutron yields and spectra from targets of varying atomic number bombarded with deuterons from 16 to 50 MeV (for radiobiology and radiotherapy). Phys. Med. Biol. 20, 235–243.
• 57)   Saltmarsh,  M.J.,  Ludemann,  C.A.,  Fulmer,  C.B. and  Styles,  R.C. (1977) Characteristics of an intense neutron source based on the d+Be reaction. Nucl. Instrum. Methods 145, 81–90.
• 58)  Arino, H., Kramer, H.H., McGovern, J.J. and Thornton, A.K. (1974) Production of high purity fission product molybdenum-99. U.S. Patent No. 3799883.
• 59)   Chattopadhyay,  S.,  Barua,  L.,  De,  A.,  Das,  S.S.,  Kuniyil,  R.,  Bhaskar,  P. et al. (2012) A computerized compact module for separation of 99mTc-radionuclide from molybdenum. Appl. Radiat. Isot. 70, 2631–2637.
• 60)   Dash,  A.,  (Russ) Knapp,  F.F. Jr. and  Pillai,  M.R.A. (2013) ⁹⁹Mo/^99mTc separation: An assessment of technology options. Nucl. Med. Biol. 40, 167–176.
• 61)   Tachimori,  S.,  Nakamura,  H. and  Amano,  H. (1971) Diffusion of Tc-99m in neutron irradiated molybdenum trioxide and its application to separation. J. Nucl. Sci. Technol. 8, 295–301.
• 62)   Nagai,  Y.,  Kawabata,  M.,  Sato,  N.,  Hashimoto,  K.,  Saeki,  H. and  Motoishi,  S. (2014) High thermo-separation efficiency of ^99mTc from molten ¹⁰⁰MoO₃ samples by repeated milking tests. J. Phys. Soc. Jpn. 83, 083201.
• 63)   Hashimoto,  K.,  Nagai,  Y.,  Kawabata,  M.,  Sato,  N.,  Hatsukawa,  Y.,  Saeki,  H. et al. (2015) SPECT imaging of mice with ^99mTc-radiopharmaceuticals obtained from ⁹⁹Mo produced by ¹⁰⁰Mo(n,2n)⁹⁹Mo and fission of ²³⁵U. J. Phys. Soc. Jpn. 84, 043202.
• 64)  Novgorodov, A.F., Rösch, F. and Korolev, N.A. (2011) Radiochemical Separations by Thermochromatography. In Handbook of Nuclear Chemistry (eds. Vértes, A., Nagy, S., Klencsár, Z., Lovas, R.G. and Rösch, F.). Springer, Boston, MA, pp. 2429–2458.
• 65)  Tomicic, M. (1977) Separation of ^99mTc from ⁹⁹MoO₃. A High Performance Sublimation Generator: Roskielde. Risø National Laboratory, Roskilde, Denmark, Risø-M No. 1943.
• 66)   Christian,  J.D.,  Petti,  D.A.,  Kirkham,  R.J. and  Bennett,  R.G. (2000) Advances in sublimation separation of technetium from low-specific-activity molybdenum-99. Ind. Eng. Chem. Res. 39, 3157–3168.
• 67)   Beyer,  G.J.,  Hagebø,  E.,  Novgorodov,  A.F. and  Ravn,  H.L., ISOLDE Collaboration. (2003) The role of diffusion in ISOL targets for the production of radioactive ion beams. Nucl. Instrum. Methods. Phys. Res. Sect. B 204, 225–234.
• 68)   Anderson,  J.S. and  Saddington,  K. (1949) The use of radioactive isotopes in the study of the diffusion of ions in solution. J. Chem. Soc. S381–S386.
• 69)  United States Pharmacopeial Convention. (2005) Official monographs: USP 28, sodium pertechnetate Tc^99m injection. United States Pharmacopeia (USP) 28 — national formulary (NF) 23, p. 1861.
• 70)   Guérin,  B.,  Tremblay,  S.,  Rodrigue,  S.,  Rousseau,  J.A.,  Dumulon-Perreault,  V.,  Lecomte,  R. et al. (2010) Cyclotron production of ^99mTc: An approach to the medical isotope crisis. J. Nucl. Med. 51, 13N–16N.
• 71)   Galea,  R.,  Wells,  R.G.,  Ross,  C.K.,  Lockwood,  J.,  Moore,  K.,  Harvey,  J.T. et al. (2013) A comparison of rat SPECT images obtained using ^99mTc derived from ⁹⁹Mo produced by an electron accelerator with that from a reactor. Phys. Med. Biol. 58, 2737–2750.
• 72)  International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH). (2014) ICH Harmonized Guideline Q3D: Guideline for Elemental Impurities, Step 4.
• 73)   Nagai,  Y.,  Nakahara,  Y.,  Kawabata,  M.,  Hatsukawa,  Y.,  Hashimoto,  K.,  Saeki,  H. et al. (2017) Quality of ^99mTcO₄⁻ from ⁹⁹Mo produced by ¹⁰⁰Mo(n,2n)⁹⁹Mo. J. Phys. Soc. Jpn. 86, 053202.
• 74)   Kawabata,  M.,  Motoishi,  S.,  Saeki,  H.,  Hashimoto,  K. and  Nagai,  Y. (2017) Recovery efficiency of enriched ¹⁰⁰MoO₃ irradiated by accelerator neutrons. J. Phys. Soc. Jpn. 86, 053201.
• 75)   Rösch,  F.,  Novgorodov,  A.F. and  Qaim,  S.M. (1994) Thermochromatographic separation of ^94mTc from enriched molybdenum targets and its large scale production for nuclear medical application. Radiochim. Acta 64, 113–120.
• 76)   Gagnon,  K.,  Wilson,  J.S.,  Holt,  C.M.B.,  Abrams,  D.N.,  McEwan,  A.J.B.,  Mitlin,  D. et al. (2012) Cyclotron production of ^99mTc: Recycling of enriched ¹⁰⁰Mo metal targets. Appl. Radiat. Isot. 70, 1685–1690.
• 77)   Kawabata,  M.,  Hashimoto,  K.,  Saeki,  H.,  Sato,  N.,  Motoishi,  S.,  Takakura,  K. et al. (2015) Production and separation of ⁶⁴Cu and ⁶⁷Cu using 14 MeV neutrons. J. Radioanal. Nucl. Chem. 303, 1205–1209.
• 78)   Peng,  F.,  Lu,  X.,  Janisse,  J.,  Muzik,  O. and  Shields,  A.F. (2006) PET of human prostate cancer Xenografts in mice with increased uptake of ⁶⁴CuCl₂. J. Nucl. Med. 47, 1649–1652.
• 79)  The Editorial Board of the Cancer Statistics in Japan. (eds.) (2020) CANCER STATISTICS IN JAPAN 2019, (published by Foundation for Promotion of Cancer Research, Tokyo, March, 2020) p. 15.
• 80)   Sugo,  Y.,  Hashimoto,  K.,  Kawabata,  M.,  Saeki,  H.,  Sato,  S.,  Tsukada,  K. et al. (2017) Application of ⁶⁷Cu produced by ⁶⁸Zn(n,n′p+d)⁶⁷Cu to biodistribution study in tumor-bearing mice. J. Phys. Soc. Jpn. 86, 023201.
• 81)  Nagler, A., Mardor, I., Berkovits, D., Dunkel, K., Pekeler, M., Piel, C. et al. (2006) Status of the SARAF project. In Proc. LINAC 2006. Knoxville, Tennessee, U.S.A., pp. 168–170.
• 82)   Ehrlich,  K.,  Bloom,  E.E. and  Kondo,  T. (2000) International strategy for fusion materials development. J. Nucl. Mater. 283, 79–88.
• 83)  Mitsumoto, T., Fujita, K., Ogasawara, T., Tsutsui, H. and Yajima, S. (2010) BNCT system using 30 MeV H-cyclotron. In Proceedings of the 19th International Conference on Cyclotrons and Their Applications. Lanzhou, China, pp. 430–432.
• 84)   Davis,  J. (1979) Research applications of the Livermore RTNS-II neutron sources. IEEE Trans. Nucl. Sci. 26, 1240–1244.
• 85)   Grand,  P. and  Goland,  A.N. (1977) An intense neutron source based upon the deuteron-stripping reaction. Nucl. Instrum. Methods 145, 49–76.
• 86)   Helmholz,  A.C.,  McMillan,  E.M. and  Sewell,  D.C. (1947) Angular distribution of neutrons from targets bombarded by 190-MeV deuterons. Phys. Rev. 72, 1003–1007.
• 87)  Kasuya, K., Ozawa, S., Suzuki, S., Esato, K. and Akiba, M. (2007) Material surface ablation with ELM-like E-beam thermal load. In Proceedings of the 22nd IEEE/NPSS Symposium on Fusion Engineering. Albuquerque, NM, pp. 1–4.

Profile

Yasuki Nagai was born in 1943. He graduated from Tokyo Institute of Technology and received his Ph.D. in applied physics in 1971. Following an assistant professor position at Osaka University, he moved to Tokyo Institute of Technology in 1985 and then to Osaka University in 1999. He was director of the Research Center for Nuclear Physics, Osaka University (1999–2001). He studied nuclear properties at high-spin states and giant resonance states using hadron beams and laser Compton photons as well as double β decays of ⁷⁶Ge and ¹⁰⁰Mo to search for the neutrino mass. From 1990 he began to measure the accelerator neutron (also α) capture cross sections of nuclei that are necessary to construct models of stellar nucleosynthesis relevant to the origin of the elements in stars. In 2007 he moved as an invited researcher to Japan Atomic Energy Agency (JAEA) (2007–2015) and to National Institutes for Quantum and Radiological Science and Technology (QST) (2016–2018). Currently, he is a QST Associate, Research Professor of Cyclotron Radioisotope Center at Tohoku University, and Research Fellow at Chiyoda Technol Corporation. In 2009 he proposed a new production route for the medical radioisotope ⁹⁹Mo using accelerator neutrons. In 2010 he contributed to form a special group within JAEA to develop technology to produce medical radioisotopes using accelerator neutrons together with private firms. In 2020 he started up the DATE project at Tohoku University (with Hiroshi Watabe, QST, Sumitomo Heavy Industries, Ltd. and Chiyoda Technol Corporation). He is an emeritus professor of both Osaka University and Tokyo Institute of Technology. He received the 2002 Nishina Memorial Prize (with Masayuki Igashira).