The list of reactions evaluated in the present studies consists altogether of 69 charged-particle reactions for the production of 23 radionuclides of interest for PET imaging, including 11 generator systems for short-lived medically interesting radioisotopes (Table 2). There are 39 proton reactions, 16 deuteron reactions, one reaction for 3He, and 13 reactions for α particles. Energies of incident particles cover the range from a few MeV up to 100 MeV. Every subsequent subsection contains a summary of the most frequent use of each of the 23 medically relevant radionuclides, and the literature references found for each production route (given in both the text and figures), selected data (text and figures) and the characteristics of the Padé fit (text and figures). As mentioned previously, the physical yields are included in one or two additional figures at the end of each subsection if indirect and/or generator production is being considered.
Table 2 Evaluated nuclear reactions and decay data adopted for production of diagnostic PET radionuclides (2012–2016) [2]. Reaction thresholds for natural targets estimated approximately from the plots, and are given in bold Half-lives and limited decay-scheme data for the different radionuclides discussed in the following subsections can be found in Table 2. The γ-ray energies in keV and the corresponding absolute emission probabilities (absolute intensities, Pγ(%)) used to identify and quantify the activity of a given radionuclide in the experimental studies (and β+ decay fraction instead of the intensity of the 511 keV annihilation radiation) are listed, and have also been included within each of the primary subsections of this Section.
Reactions for radionuclides present in Table 1 but not considered during the course of this CRP will be evaluated in a similar manner as part of a future series of IAEA-sponsored studies and will be published elsewhere.
Production of 44gSc (T
1/2 = 3.97 h) and long-lived 44Ti parent (T
1/2 = 59.1 y)
Applications: 44Sc (av. Eβ+ = 632.0 keV, 94.27% intensity) has emerged as an attractive radiometal candidate for PET imaging by means of e.g., DOTA-functionalised biomolecules. 44Sc-labelled PET radiopharmaceuticals appear of interest for molecular imaging of medium-lasting physiological processes. Also forms a theranostic pair with therapeutic 47Sc.
44Sc (3.97 h): β+ (94.27%), and Eγ (keV) (Pγ(%)): 1157.020 (99.9).
44Ti (59.1 y): detected by means of radiation emitted by daughter 44Sc.
44Ca(p,n)44Sc, 44Ca(d,2n)44Sc and 43Ca(d,n)44Sc direct reactions and 45Sc(p,2n)44Ti-44Sc, 45Sc(d,3n)44Ti-44Sc generator production routes were evaluated.
44Ca(p,n)44gSc
The six experimental datasets available in the literature are shown in Fig. 1 [33,34,35,36,37,38], together with the TENDL calculations. Three sets were rejected Cheng et al. [34] and Mitchell [35] (too high values near the threshold), and Krajewski et al. [37] (strange overall shape)), while the remaining four datasets were used in the statistical fitting procedure. Both the selected data and their experimental uncertainties are shown in Fig. 2 together with the Padé fit (L = 14, N = 49, Χ2 = 1.63) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
44Ca(d,2n) 44gSc
Only one experimental dataset is available in the literature, and is shown in Fig. 3 [39] together with the TENDL calculations. These data and their experimental uncertainties are shown in Fig. 4 together with the Padé fit (L = 9, N = 9, χ2 = 0.53) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
43Ca(d,n)44gSc
Only one experimental dataset is available in the literature, and is shown in Fig. 5 [40] together with the TENDL calculations. These data and their experimental uncertainties are shown in Fig. 6 together with the Padé fit (L = 5, N = 16, χ2 = 1.49) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
45Sc(p,2n)44Ti
The five experimental datasets available in the literature are shown in Fig. 7 [36, 41,42,43] together with the TENDL calculations—Ref. [42] contains two sets labelled (a) and (b). Three datasets were rejected (both datasets of Ejnisman et al. [42], and McGee et al. [41] exhibit significant disagreement), while the remaining two datasets were used in the statistical fitting procedure. The selected data and their experimental uncertainties are shown in Fig. 8 together with the Padé fit (L = 7, N = 26, χ2 = 1.58) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
45Sc(d,3n)44Ti
Only one experimental dataset is available in the literature, and is shown in Fig. 9 [44] together with the TENDL calculations. These data and their experimental uncertainties are shown in Fig. 10 together with the Padé fit (L = 6, N = 18, χ2 = 0.406) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
Thick target yields for production of 44gSc, and long-lived 44Ti parent
See Figs. 11 and 12.
Production of 52mMn (T
1/2 = 21.1 min) and longer-lived 52Fe parent (T
1/2 = 8.275 h)
Applications: 52mMn has been suggested for myocardial and cerebral perfusion imaging, more recently for studies similar to Mn-enhanced neuronal MRI, and for diagnosis in other organ systems—bones, spinal cord and the digestive tract.
52mMn (21.1 min): β+ (96.6%), and Eγ (keV) (Pγ(%)): 1434.092 (98.2).
52Fe (8.275 h): detected by means of radiation emitted from daughter 52mMn.
Evaluations have been made of the 52Cr(p,n)52mMn and 52Cr(d,2n)52mMn direct production routes and natNi(p,x)52Fe, 55Mn(p,4n)52Fe and 50Cr(α,2n)52Fe reactions for indirect production through decay of the longer-lived parent.
natNi(p,x)52Fe
The four experimental datasets available in the literature are shown in Fig. 13 [45,46,47,48] together with he TENDL calculations. One set was rejected (Titarenko et al. [47], values too high), and the remaining three datasets were used in the statistical fitting procedure. These selected data and their experimental uncertainties are shown in Fig. 14 together with the Padé fit (L = 11, N = 41, χ2 = 0.57) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
55Mn(p,4n)52Fe
The four experimental datasets available in the literature are shown in Fig. 15 [46, 49,50,51] together with the TENDL calculations. All sets were used for the statistical fitting procedure. These data and their experimental uncertainties are shown in Fig. 16 together with the Padé fit (L = 17, N = 157, χ2 = 1.01) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
50Cr(α,2n)52Fe
The four experimental datasets available in the literature are shown in Fig. 17 [36, 52,53,54] together with the TENDL calculations. Two datasets were rejected (Akiha et al. [52], energy shift; Chowdhury et al. [53], unusual shape,with one outlying data point at 27.3 MeV not represented in Fig. 17), while the remaining two datasets were used in the statistical fitting procedure. Both the selected data and their experimental uncertainties are shown in Fig. 18 together with the Padé fit (L = 9, N = 52, χ2 = 0.616, solid line) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
52Cr(p,n)52mMn
The nine experimental datasets available in the literature are shown in Fig. 19 [36, 55,56,57,58,59,60,61,62] together with the TENDL calculations. Three datasets were rejected (Blosser and Handley [56], Wing and Huizenga [58], and West et al. [62], all values too high), while the remaining six datasets were used in the statistical fitting procedure. Both the selected data and their experimental uncertainties are shown in Fig. 20 together with the Padé fit (L = 14, N = 68, χ2 = 1.15) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
52Cr(d,2n)52mMn
The two experimental datasets available in the literature are shown in Fig. 21 [62, 63] together with the TENDL calculations. Both datasets were used for the statistical fitting procedure. These data and their experimental uncertainties are shown in Fig. 22 together with the Padé fit (L = 8, N = 16, χ2 = 0.71) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
Thick target yields for production of 52mMn, and long-lived 52Fe parent
See Figs. 23 and 24.
Production of 52gMn (T
½ = 5.591 d)
Applications: The longer-lived 52Mn ground state has potential as a PET tracer for preclinical in vivo neuroimaging and other applications such as cell tracking, immuno-PET and functional β-cell mass quantification. Unfortunately, a half-life of 5.591 d coupled with an extremely high radiation burden that arises from the resulting gamma-ray emissions has limited 52gMn clinical applications.
52gMn (5.591 d): β+ (29.4%), and Eγ (keV) (Pγ(%)): 744.233 (90.0), 935.544 (94.5), 1434.092 (100).
Evaluations have been made of the direct 52Cr(p,n)52gMn(m+) and 52Cr(d,2n)52gMn(m+) production routes, including the partial decay of the simultaneously produced short-lived 52mMn metastable state (IT = 1.78%, noted as (m+)) which has already been assessed and discussed in section “Production of 52mMn (T1/2 = 21.1 min) and longer-lived 52Fe parent (T1/2 = 8.275 h)”.
52Cr(p,n)52gMn (m+)
The thirteen experimental datasets available in the literature are shown in Fig. 25 [36, 55, 56, 58, 59, 61, 62, 64,65,66,67,68,69] together with the TENDL calculations. Six sets of data were rejected (Blosser and Handley [56], Tanaka and Furukawa [64], Lindner and James [65], Antropov et al. [66], Buchholz et al. [67], and Zherebchevsky et al. [69], all disagree significantly with the other datasets), while the remaining seven datasets were used in the statistical fitting procedure. These selected data and their experimental uncertainties are shown in Fig. 26 together with the Padé fit (L = 9, N = 103, χ2 = 1.84) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
52Cr(d,2n)52gMn(m+)
The six experimental datasets available in the literature are shown in Fig. 27 [62, 63, 70,71,72,73] together with the TENDL calculations. Two sets were rejected (Cheng Xiaowu et al. [71], values too low and no contribution from decay of metastable state marked as “g” in Fig. 27, and Nassiff and Münzel [72], values too high), and the remaining four datasets were used in the statistical fitting procedure. These selected data and their experimental uncertainties are shown in Fig. 28 together with the Padé fit (L = 8, N = 36, χ2 = 0.54) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
Thick target yields for production of 52gMn(m+)
See Fig. 29.
Production of 55Co (T
1/2 = 17.53 h)
Applications: 55Co is a typical example of a positron emitter of sufficient half-life to follow kinetic processes that function over a longer timescale. This radionuclide has been used to target the epidermal growth factor (EGFR) by means of labelled DOTA-conjugated Affibody. Exhibits lower liver and heart uptake for metal-chelate peptide complexes, with improved performance when compared with 68Ga. Also used as a Ca2+ analogue in imaging studies of Alzheimer disease, and shows promise in achieving improved imaging of cancer diseases.
55Co (17.53 h): β+ (76%), and Eγ (keV) (Pγ(%)): 931.1 (75), 1316.6 (7.1).
58Ni(p,α)55Co, 54Fe(d,n)55Co and 56Fe(p,2n)55Co production routes have been evaluated.
58Ni(p,α)55Co
The seventeen experimental datasets available in the literature are shown in Fig. 30 [36, 45, 59, 74,75,76,77,78,79,80,81,82,83,84,85,86,87] together with the TENDL calculations. One dataset was rejected (Haasbroek et al. [76], values too high), while the remaining sixteen datasets were used in the statistical fitting procedure. The selected data and their experimental uncertainties are shown in Fig. 31 together with the Padé fit (L = 10, N = 352, χ2 = 1.97) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
54Fe(d,n)55Co
The ten experimental datasets available in the literature are shown in Fig. 32 [88,89,90,91,92,93,94,95,96,97] together with the TENDL calculations. One dataset was rejected (Clark et al. [89], values too high), while the remaining nine datasets were used in the statistical fitting procedure (although some very discrepant points around 10 MeV from Hermanne [94] and the highest three points from Zhenlan [91] were also discarded). The selected data and their experimental uncertainties are shown in Fig. 33 together with the Padé fit (L = 13, N = 170, χ2 = 2.14) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
56Fe(p,2n)55Co
The fifteen experimental datasets available in the literature are shown in Fig. 34 [36, 59, 82, 98,99,100,101,102,103,104,105,106,107,108,109] together with the TENDL calculations. Four datasets were rejected (Michel et al. [82], Cohen and Newman [98], Williams and Fulmer [99], and Ditrói et al. [107], all show discrepant values), and the remaining eleven sets were used in the statistical fitting procedure. The selected data and their experimental uncertainties are shown in Fig. 35 together with the Padé fit (L = 8, N = 101, χ2 = 2.74) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
Thick target yields for production of 55Co
See Fig. 36.
Production of 61Cu (T
½ = 3.339 h)
Applications: Copper radionuclides form stable complexes with several chelators that can be conjugated to a wide variety of organic molecules for both imaging (61Cu, 62Cu, 64Cu) and radiotherapy (64Cu, 67Ci). Relatively longer-lived 61Cu (T½ = 3.339 h, 61% β+, 39% EC) possesses very good imaging properties that can be used for blood flow studies in a similar manner to 51Cr. Also has been applied to blood pool imaging (DOTA-human serum albumin) and the study of hypoxia in tumours (coupled to ATSM)—useful for following kinetics processes of the order of a few hours.
61Cu (3.339 h): β+ (61%), and Eγ (keV) (Pγ(%)): 282.956 (12.2), 656.008 (10.8), 1185.234 (3.7).
Evaluations have been made of the 61Ni(p,n)61Cu, 60Ni(d,n)61Cu and 64Zn(p,α)61Cu direct production routes.
61Ni(p,n)61Cu
The seventeen experimental datasets available in the literature are shown in Fig. 37 [45, 56, 59, 64, 78, 79, 84, 87, 110,111,112,113,114,115,116,117] together with the TENDL calculations. Ref. [112] contains two datasets, labelled (a) and (b). Five datasets were rejected (Blosser and Handley [56], Tanaka and Furukawa [64], Barrandon et al. [59], Michel et al. [78], and Al-Saleh et al. [84], all of these datasets exhibit maximum values that are too high), and the remaining twelve sets were used in the statistical fitting procedure. The selected data and their experimental uncertainties are shown in Fig. 38 together with the Padé fit (L = 12, N = 192, χ2 = 2.81) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
60Ni(d,n)61Cu
The five experimental datasets available in the literature are shown in Fig. 39 [90, 118,119,120,121] together with the TENDL calculations. All datasets were used in the statistical fitting procedure (Cogneau et al. [118] data were normalised, and data above 6-MeV particle beam energy discarded as inconsistent with model calculations). All of the data and their experimental uncertainties are shown in Fig. 40 together with the Padé fit (L = 16, N = 29, χ2 = 1.16) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale). This reaction is the main contributor to the formation of 61Cu on natural Ni by deuterons, adopted as a suitable beam monitor (see Ref. [30], Sect. 3.I).
64Zn(p,α)61Cu
The seven experimental datasets available in the literature are shown in Fig. 41 [36, 59, 122,123,124,125,126] together with the TENDL calculations. One dataset was rejected (Barrandon et al. [59], discrepant behaviour near maximum), and the remaining six sets were used in the statistical fitting procedure. The selected data and their experimental uncertainties are shown in Fig. 42 together with the Padé (L = 12, N = 72, χ2 = 0.88) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
Thick target yields for production of 61Cu
See Fig. 43.
Production of 62Cu (T
1/2 = 9.67 min) and longer-lived 62Zn parent (T
1/2 = 9.193 h)
Applications: As stated earlier, copper isotopes form stable complexes with several chelators that can be conjugated to a wide variety of organic molecules for both imaging (61Cu, 62Cu, 64Cu) and therapy (64Cu, 67Ci). Short-lived 62Cu has been proposed for the labelling of PTSM (pyruvaldehyde bis) to undertake myocardial and brain blood flow studies.
62Cu (9.67 min): β+ (97.83%), and Eγ (keV) (Pγ(%)): 875.66 (0.147), 1172.97 (0.342).
62Zn (9.193 h): β+ (8.2%), and Eγ (keV) (Pγ(%)): 548.35 (15.3), 596.56 (26).
Evaluations have been made of the 63Cu(p,2n)62Zn, 63Cu(d,3n)62Zn and natNi(α,xn)62Zn indirect, and 62Ni(p,n)62Cu and 62Ni(d,2n)62Cu direct production routes.
63Cu(p,2n)62Zn
Twenty-four experimental datasets available in the literature are shown in Fig. 44 [36, 82, 98, 99, 127,128,129,130,131,132,133,134,135,136,137,138,139,140,141,142,143,144,145,146] together with the TENDL calculations. Seven datasets were rejected (Ghoshal [127], Williams and Fulmer [99], Greene and Lebowitz [128], Greenwood and Smither [130], Aleksandrov et al. [132], Levkovskij [36], and Tárkányi et al. [142], all disagree significantly with the other datasets), and the remaining seventeen sets were used in the statistical fitting procedure. The selected data and their experimental uncertainties are shown in Fig. 45 together with the Padé fit (L = 16, N = 213, χ2 = 1.89) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale). As the only reaction known to contribute to the formation of 62Zn on natCu for protons below 30 MeV, the fitted data have been adopted as a beam monitor in this energy region (see Ref. [30], Sect. 2.6).
63Cu(d,3n)62Zn
While a known dataset by Bartell et al. [147] is not represented in Fig. 46 because the values are totally discrepant even after arbitrary normalisation, eight other experimental datasets available in the literature are shown [73, 95, 148,149,150,151,152,153] together with the TENDL calculations. The data by Fulmer and Williams [148] were subsequently rejected because they disagree significantly with the other datasets (attributed to the normalisation of inadequately defined low-intensity decay data). All of the remaining seven datasets were used in the statistical fitting procedure. The selected data and their experimental uncertainties are shown in Fig. 47 together with the Padé fit (L = 12, N = 82, χ2 = 1.89) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale). As the only reaction known to contribute to the formation of 62Zn on natCu for deuterons below 35 MeV, the fitted data have been adopted as a beam monitor in this energy region (see Ref. [30], Section III E).
natNi(α,xn)62Zn
The nine experimental datasets available in the literature are shown in Fig. 48 [36, 127, 154,155,156,157,158,159,160] together with the TENDL calculations. Three datasets were rejected (Neirinckx [155] (energy shift near threshold), Singh et al. [159] (discrepant values at energies below 35 MeV), and Yadav et al. [160] (value too low)), and the remaining six sets were used in the statistical fitting procedure. The selected data and their experimental uncertainties are shown in Fig. 49 together with the Padé fit (L = 21, N = 45, χ2 = 1.72) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
62Ni(p,n)62Cu
The seven experimental datasets available in the literature are shown in Fig. 50 [36, 45, 66, 110, 161,162,163] together with the TENDL calculations. All datasets were used in the statistical fitting procedure. The data and their experimental uncertainties are shown in Fig. 51 together with the Padé fit (L = 12, N = 77, χ2 = 1.33) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
62Ni(d,2n)62Cu
The single dataset available in the literature is shown in Fig. 52 [118] together with the TENDL calculations. This dataset was used in the statistical fitting procedure. The data and their experimental uncertainties are shown in Fig. 53 together with the Padé fit (L = 5, N = 16, χ2 = 0.83) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
Thick target yields for production of 62Cu, and 62Zn parent
See Figs. 54 and 55.
Production of 66Ga (T
½ = 9.49 h)
Applications: Both 66Ga and 68Ga are positron-emitting radionuclides that can be used in PET imaging. Longer-lived 66Ga has been coupled to monoclonal antibodies (e.g., for tumour angiogenesis studies) and to nanoparticles. This radionuclide has also been proposed in hadron therapy as an in situ marker for the incorporation of Zn in tumours. Obvious disadvantages are the rather high radiation burden and inferior imaging properties caused by the many gamma rays that accompany decay.
66Ga (9.49 h): β+ (57%), and Eγ (keV) (Pγ(%)): 833.5324 (5.9), 1039.220 (37.0).
Evaluations have been made of the 66Zn(p,n)66Ga and 63Cu(α,n)66Ga direct production routes.
66Zn(p,n)66Ga
The twenty experimental datasets available in the literature are shown in Fig. 56 [36, 56, 59, 124, 125, 164,165,166,167,168,169,170,171,172,173,174,175,176,177] together with the TENDL-2015 and TENDL-2017 calculations. Hermanne [173] contains two datasets labelled (a) and (b).Twelve datasets were rejected (Little and Lagunas-Solar [167] (values too low), Nortier et al. [171] (energy shift), Blosser and Handley [56] (only one data point that can not be checked), Howe [165] (energy shift), Kopecký [168] (values too low), Asad et al. [125] (values too low), Szelecsényi et al. [172] (preliminary results), Hermanne [173] set b (discrepant data points), Barrandon et al. [59] (values too low), Al-Saleh et al. [177] (values too low), Uddin et al. [124] (values too low), and Blaser et al. [164] (discrepant data points)), while the remaining eight sets were used in the statistical fitting procedure. The selected data and their experimental uncertainties are shown in Fig. 57 together with the Padé (L = 13, N = 188, χ2 = 1.87) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
63Cu(α,n)66Ga
The twenty-three experimental datasets available in the literature are shown in Fig. 58 [36, 54, 81, 166, 178,179,180,181,182,183,184,185,186,187,188,189,190,191,192,193,194,195,196] together with the TENDL calculations. Seven datasets were rejected (Porges [178] (values too low), Bonesso et al. [188] (values too low), Zhukova et al. [182] (values too low), Singh et al. [190] (values too low), Rizvi et al. [184] (values too low), Porile and Morrison [179] (values too low), and Nassiff and Nassiff [183] (discrepant data points)), and the remaining sixteen sets were used in the statistical fitting procedure. The selected data and their experimental uncertainties are shown in Fig. 59 together with the Padé (L = 13, N = 252, χ2 = 1.34) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale). This reaction is also used to monitor α-particle beams (see Ref. [30], Section V D).
Thick target yields for production of 66Ga
See Fig. 60.
Production of 68Ga (T
1/2 = 67.71 min) and long-lived 68Ge parent (T
1/2 = 270.95 d)
Applications: Rather short-lived 68Ga became the first widespread generator-produced positron emitter, thereby competing somewhat with 18F for preferred adoption in PET imaging. First introduced for the imaging of neuroendocrine tumours (68Ga-labelled DOTA-TOC), more recent significant success has been achieved in the form of very efficient imaging agents for prostate cancer diagnosis and staging (68Ga-DOTA-PSMA and derivatives).
68Ga (67.71 min): β+ (88.91%), and Eγ (keV) (Pγ(%)): 1077.34 (3.22).
68Ge (270.95 d): detected by means of radiation from daughter 68Ga.
Evaluations have been undertaken of the 68Zn(p,n)68Ga and 65Cu(α,n)68Ga direct routes and natGa(p,x)68Ge and 69Ga(p,2n)68Ge generator production.
68Zn(p,n)68Ga
The eighteen experimental datasets available in the literature are shown in Fig. 61 [36, 41, 56, 59, 111, 162, 164,165,166, 169, 170, 173, 177, 195, 197,198,199,200] together with the TENDL calculations. Five datasets were rejected (Hermanne et al. [170] (energy shift), Blosser and Handley [56] (value too high), McGee et al. [41] (value too low), Hermanne [173] (energy shift), and Barrandon et al. [59] (values too low)), and the remaining thirteen sets were used in the statistical fitting procedure. The selected data and their experimental uncertainties are shown in Fig. 62 together with the Padé (L = 20, N = 282, χ2 = 1.97) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
65Cu(α,n)68Ga
The fourteen experimental datasets available in the literature are shown in Fig. 63 [36, 54, 166, 178,179,180, 184, 186, 188, 190, 195, 196, 201, 202] together with the TENDL calculations. Four datasets were rejected (Porile and Morrison [179] (energy shift), Rizvi et al. [184] (energy shift), Bonesso et al. [188] (values too high), and Porges [178] (values too low)), and the remaining ten sets were used in the statistical fitting procedure. The selected data and their experimental uncertainties are shown in Fig. 64 together with the Padé (L = 10, N = 92, χ2 = 1.21) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
natGa(p,xn)68Ge
The six experimental datasets available in the literature are shown in Fig. 65 [36, 48, 98, 203, 204] together with the TENDL calculations. Hermanne et al. [48] contains two datasets labelled (a) and (b). One dataset was rejected (Cohen and Newman [98], single data point too low), and the remaining five sets were used in the statistical fitting procedure. The selected data and their experimental uncertainties are shown in Fig. 66 together with the Padé (L = 11, N = 101, χ2 = 1.42) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
69Ga(p,2n)68Ge
The four experimental datasets available in the literature are shown in Fig. 67 [36, 48, 203, 204] together with the TENDL calculations. All sets were used in the statistical fitting procedure. The data and their experimental uncertainties are shown in Fig. 68 together with the Padé (L = 8, N = 53, χ2 = 1.56) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
Thick target yields for 68Ga, and long-lived 68Ge parent for generator
See Figs. 69 and 70.
Production of 72As (T
1/2 = 26.0 h) and longer-lived 72Se parent (T
1/2 = 8.40 d)
Applications: 72As is a long-lived positron-emitting radionuclide suitable for imaging the bio-distribution of monoclonal antibodies with long biological half-lives that are promising in PET oncological research. Chemical properties offer the possibility of covalent bonding to thiol groups.
72As (26.0 h): β+ (87.8%), and Eγ (keV) (Pγ(%)): 629.92 (8.07), 833.99 (81).
72Se (8.40 d): detected by means of radiation emitted by daughter 72As.
Evaluations have been undertaken of the 75As(p,4n)72Se and natBr(p,x)72Se routes for parent production, and the natGe(p,xn)72As and natGe(d,xn)72As direct production routes.
75As(p,4n)72Se
The two experimental datasets available in the literature for the energy domain considered are shown in Fig. 71 [205, 206] together with the TENDL calculations. Both datasets were used in the statistical fitting procedure. The data and their experimental uncertainties are shown in Fig. 72 together with the Padé fit (L = 8, N = 33, χ2 = 1.30) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
natBr(p,x)72Se
The two experimental datasets available in the literature are shown in Fig. 73 [207, 208] together with the TENDL calculations. Both sets of measurements by Fassbender et al. [207] and de Villiers et al. [208] originate from the same experimental study, and should be identical. Therefore, the data of de Villers et al. [208] were set aside, while only the other dataset was used in the statistical fitting procedure. These selected data and their experimental uncertainties are shown in Fig. 74 together with the Padé fit (L = 10, N = 14, χ2 = 0.35) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale). The contributions of the similar (p,2pxn) reactions on the two stable isotopes of Br can be clearly distinguished (79Br: 50.69%; 81Br: 49.31%).
natGe(p,xn)72As
The four experimental datasets available in the literature are shown in Fig. 75 [36, 209,210,211] together with the TENDL calculations. All datasets were used in the statistical fitting procedure. The data and their experimental uncertainties are shown in Fig. 76 together with the Padé (L = 18, N = 123, χ2 = 1.97) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale). Contributions of similar (p,xn) reactions with increasing thresholds can be clearly distinguished for the higher abundance 72Ge,74Ge and 76Ge.
natGe(d,xn)72As
The single experimental dataset available in the literature is shown in Fig. 77 [212] together with the TENDL calculations. All data points and their experimental uncertainties are shown in Fig. 78 together with the Padé fit (L = 10, N = 25, χ2 = 1.13) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale). The contribitions of the 72Ge(d,2n) and 74Ge(d,4n) reactions on high natural abundance Ge isotopes can be seen in both figures.
Thick target yields for production of 72As, and 72Se parent
See Figs. 79 and 80.
Production of 73Se (T
1/2 = 7.15 h)
Applications: 73Se (T½ = 7.15 h; EC = 34.6%, β+ = 65.4%; Eβ+(max) = 1.65 MeV) is an interesting β+-emitting analogue of sulphur suitable for the imaging of enzymatic systems or sulphur-containing amino acids.
73Se(7.15 h): β+ (65.4%), and Eγ (keV) (Pγ(%)): 67.07 (70), 361.2 (97.0).
Evaluations have been made of the 75As(p,3n)73Se and 72Ge(α,3n)73Se direct production routes.
75As(p,3n)73Se
The four experimental datasets available in the literature are shown in Fig. 81 [36, 206, 213, 214] together with the TENDL calculations. One dataset was rejected (Mushtaq et al. [206] (values too low near maximum)), and the remaining three sets were used in the statistical fitting procedure. The selected data and their experimental uncertainties are shown in Fig. 82 together with the Padé (L = 9, N = 64, χ2 = 1.52) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
72Ge(α,3n)73Se
The two experimental datasets available in the literature are shown in Fig. 83 [36, 215] together with the TENDL calculations. Both datasets were used in the statistical fitting procedure. The data and their experimental uncertainties are shown in Fig. 84 together with the Padé fit (L = 8, N = 27, χ2 = 3.81) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
Thick target yields for production of 73Se
See Fig. 85.
Production of 76Br (T
1/2 = 16.2 h)
Applications: Longer-lived positron-emitting radiohalogens were some of the first radionuclides studied to follow processes with kinetics inappropriate for 18F application. 76Br was used in several studies to label monoclonal antibodies, although the large number of accompanying gamma rays that result in a relatively high radiation burden and poor imaging properties has seen a subsequent decline of interest in this radionuclide.
76Br(16.2 h): β+ (55%) and Eγ (keV) (Pγ(%)): 559.09 (74), 657.02 (15.9), 1853.67 (14.7).
Evaluations have been made of the 76Se(p,n)76Br, 77Se(p,2n)76Br and 75As(α,3n)76Br production routes.
76Se(p,n)76Br
The five experimental datasets available in the literature are shown in Fig. 86 [36, 216,217,218,219] together with the TENDL calculations. Two datasets were rejected (Kovács et al. [218] (values too low near maximum), and Hassan et al. [219] (discrepant data near maximum)), and the remaining three sets were used in the statistical fitting procedure. The selected data and their experimental uncertainties are shown in Fig. 87 together with the Padé fit (L = 8, N = 39, χ2 = 1.05) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
77Se(p,2n)76Br
The four experimental datasets available in the literature are shown in Fig. 88 [36, 219,220,221] together with the TENDL calculations. Two datasets were rejected (Janssen et al. [220] (values too low), and Hassan et al. [219] (values too high)), and the remaining two sets were used in the statistical fitting procedure. The selected data and their experimental uncertainties are shown in Fig. 89 together with the Padé fit (L = 9, N = 52, χ2 = 1.34) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
75As(α,3n)76Br
The five experimental datasets available in the literature are shown in Fig. 90 [216, 217, 222,223,224] together with the TENDL calculations. All datasets were used for the statistical fitting procedure. The data and their experimental uncertainties are shown in Fig. 91 together with the Padé fit (L = 10, N = 70, χ2 = 2.43) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale). An additional dataset published after the evaluation cut-off date was also included in the fit (Breunig et al. [224]) and is shown in Fig. 90.
Thick target yields for 76Br
See Fig. 92.
Production of 82Sr parent (T
1/2 = 25.35 d) of short-lived 82Rb (T
1/2 = 1.2575 min)
Applications: Generator-produced 82Rb is widely used in myocardial perfusion imaging, particularly in the USA. This isotope undergoes rapid uptake by myocardiocytes, and therefore is a valuable tool for identifying myocardial ischemia by means of PET. Such a short half-life allows one to perform both stress and rest perfusion studies within 30 min.
82Rb (1.2575 min): β+ (95.43%), and Eγ (keV) (Pγ(%)): 776.52 (15.08).
82Sr (25.35 d): detected by means of radiation emitted by daughter 82Rb.
Evaluations have been undertaken of the natRb(p,xn)82Sr and 85Rb(p,4n)82Sr parent production routes.
natRb(p,xn)82Sr
The seven experimental datasets available in the literature are shown in Fig. 93 [138, 225,226,227,228,229,230] together with the TENDL calculations. Two datasets were rejected (Horiguchi et al. [225] (values too high), and Deptula et al. [226] (values too high)), and the remaining five sets were used in the statistical fitting procedure. The selected data and their experimental uncertainties are shown in Fig. 94 together with the Padé fit (L = 13, N = 49, χ2 = 1.15) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
85Rb(p,4n)82Sr
The five experimental datasets available in the literature are shown in Fig. 95 [138, 225, 227, 229, 230] together with the TENDL calculations. All datasets were used in the statistical fitting procedure. The data and their experimental uncertainties are shown in Fig. 96 together with the Padé fit (L = 9, N = 49, χ2 = 1.60) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
Thick target yields for 82Sr parent of short-lived 82Rb
See Fig. 97.
Production of 82mRb (T½ = 6.472 h)
Applications: Longer-lived 82mRb isomeric state could possibly act as a substitute for generator-produced 82Rb in PET cardiology centres that operate a cyclotron. However, this isomer suffers from a relatively high radiation burden that arises from the longer half-life and gamma-ray emissions.
82mRb (6.472 h): β+ (21.2%), and Eγ (keV) (Pγ(%)): 554.35 (62.4), 619.11 (37.98), 698.37 (26.3), 776.52 (84.39), 827.83 (21.0), 1044.08 (32.07), 1317.43 (23.7), 1474.88 (15.5).
Evaluations have been undertaken of the 82Kr(p,n)82mRb and 82Kr(d,2n)82mRb reactions.
82Kr(p,n)82mRb
The four experimental datasets available in the literature are shown in Fig. 98 [231, 232] (each reference contains two datasets labelled (a) and (b)), together with the TENDL calculations. All datasets were used in the statistical fitting procedure. The data and their experimental uncertainties are shown in Fig. 99 together with the Padé fit (L = 9, N = 33, χ2 = 1.13) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
82Kr(d,2n)82mRb
A single experimental dataset available in the literature is shown in Fig. 100 [233] together with the TENDL calculations. This one dataset was used in the statistical fitting procedure. The data and their experimental uncertainties are shown in Fig. 101 together with the Padé fit (L = 5, N = 14, χ2 = 2.27) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
Thick target yields for production of 82mRb
See Fig. 102.
Production of 86Y (T
1/2 = 14.74 h)
Applications: Extensive studies of 86Y have been performed as a positron emitter (31.9%) with 14.74 h half-life that can adopted as a theranostic pair with clinically-established therapeutic beta-emitting 90Y. The role of 86Y is to monitor the localised therapeutic dose distribution in the body for dosimetry calculations. Has also been studied for prostate cancer imaging, and used to label monoclonal antibodies in EGFR targeting. However, interest in this radionuclide has declined because of the high radiation burden and resultant poor imaging properties.
86Y (14.74 h): β+ (31.9%), and Eγ (keV) (Pγ(%)): 627.72 (32.6), 1076.63 (82.5), 1153.05 (30.5).
Evaluations have been made of the 86Sr(p,n)86Y, 88Sr(p,3n)86Y and 85Rb(α,3n)86Y production routes.
86Sr(p,n)86Y
The four experimental datasets available in the literature are shown in Fig. 103 [36, 82, 234, 235] together with the TENDL calculations. One dataset was rejected (Rösch et al. [235] (scattered data, and values too high near maximum)), while the three remaining sets were used in the statistical fitting procedure. The selected data and their experimental uncertainties are shown in Fig. 104 together with the Padé fit (L = 9, N = 28, χ2 = 0.615) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
88Sr(p,3n)86Y
The two experimental datasets available in the literature are shown in Fig. 105 [36, 236] together with the TENDL calculations. Both datasets were used in the statistical fitting procedure. These data and their experimental uncertainties are shown in Fig. 106 together with the Padé fit (L = 8, N = 15, χ2 = 1.27) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
85Rb(α,3n)86Y
Five experimental datasets available in the literature are shown in Fig. 107 [36, 237,238,239,240] together with the TENDL calculations. Three datasets were rejected (Guin et al. [239], Iwata [237], and Agarwal et al. [240] (values refer to direct ground state production only, and are not cumulative). The data points of Demeyer et al. [238] below 45 MeV are discrepant, and were also deleted. Thus, the remaining data points for only two datasets were used in the statistical fitting procedure [36, 238]. These selected data and their experimental uncertainties are shown in Fig. 108 together with the Padé fit (L = 8, N = 32, χ2 = 0.91) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
Thick target yields for production of 86Y
See Fig. 109.
Production of 89Zr (T
1/2 = 78.41 h)
Applications: Long-lived positron-emitting 89Zr has been extensively studied with respect to following the in vivo behaviour of therapeutic monoclonal antibodies (mAbs) and other biomolecules with slow biokinetics. One significant disadvantage is the limited number of suitable 89Zr chelating agents and difficulties related to their development.
89Zr (78.41 h): β+ (22.74%), and Eγ (keV) (Pγ(%)): 909.15 (99.04).
Evaluations have been made of the 89Y(p,n)89Zr and 89Y(d,2n)89Zr production routes.
89Y(p,n)89Zr
The sixteen experimental datasets available in the literature are shown in Fig. 110 [36, 56, 82, 110, 234, 241,242,243,244,245,246,247,248,249,250,251] together with the TENDL calculations. Five datasets were rejected (Birattari et al. [244] (energy shift), Blosser and Handley [56] (value too high), Satheesh et al. [250] (energy shift), Delaunay-Olkowsky et al. [234] (value too low), and Saha et al. [242] (values too high)), and the remaining eleven datasets were used in the statistical fitting procedure. The selected data and their experimental uncertainties are shown in Fig. 111 together with the Padé fit (L = 11, N = 316, χ2 = 3.74) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
89Y(d,2n)89Zr
The seven experimental datasets available in the literature are shown in Fig. 112 [252,253,254,255,256,257,258] together with the TENDL calculations. Two datasets were rejected (La Gamma and Nassiff [253] (values too low), and Degering et al. [255] (energy shift)), and the remaining five sets were used in the statistical fitting procedure. The selected data and their experimental uncertainties are shown in Fig. 113 together with the Padé fit (L = 9, N = 64, χ2 = 2.95) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
Thick target yields for production of 89Zr
See Fig. 114.
Production of 90Nb (T
1/2 = 14.60 h)
Applications: As a non-conventional positron emitter, 90Nb with a half-life of 14.60 h can be used to visualise and quantify processes with medium and slow kinetics, such as tumour accumulation of antibodies and antibody fragments, or polymers and other nanoparticles. Exhibits promise in immuno-PET, although a search for appropriate chelators is desirable. Also emits several high-energy gamma rays that increase the radiation burden.
90Nb (14.60 h): β+ (51.2%), and Eγ (keV) (Pγ(%)): 132.716 (4.13), 141.178 (66.8), 1129.224 (92.7).
Evaluations have been undertaken of the 93Nb(p,x)90Nb and 89Y(α,3n)90Nb production.
93Nb(p,x)90Nb
The six experimental datasets available in the literature are shown in Fig. 115 [82, 249, 259,260,261,262] together with the TENDL calculations. All datasets were used in the statistical fitting procedure. The data and their experimental uncertainties are shown in Fig. 116 together with the Padé fit (L = 9, N = 94, χ2 = 3.18) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
89Y(α,3n)90Nb
The six experimental datasets available in the literature are shown in Fig. 117 [36, 263,264,265,266,267] together with the TENDL calculations. Four datasets were rejected (Singh et al. [266], Chaubey and Rizvi [265], Mukherjee et al. [264], and Smend et al. [263], all systematically lower values), while the remaining two sets were used in the statistical fitting procedure. The selected data and their experimental uncertainties are shown in Fig. 118 together with the Padé fit (L = 16, N = 33, χ2 = 1.29) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
Thick target yields for production of 90Nb
See Fig. 119.
Production of 94mTc (T
1/2 = 52.0 min)
Applications: Gamma-ray emitting 99mTc is the most widespread medical radionuclide for diagnosis, whereas 94mTc with a half-life of 52.0 min. is a positron emitter with a positron branch of 70.2% and Eβ+(max) of 2.44 MeV. Therefore, there has been interest in 94mTc as a PET analogue to 99mTc since they both undergo the same chemistry. Obvious disadvantages of 94mTc are the rather short half-life of 52.0 min., with many accompanying gamma rays and the inability to prepare the pure isomer without also generating significant amounts of ground state 94gTc.
94mTc (52.0 min): β+ (70.2%), and Eγ (keV) (Pγ(%)): 871.05 (94.2), 1522.1 (4.5), 1868.68 (5.7).
Evaluations have been made of the 92Mo(α,x)94mTc and 94Mo(p,n)94mTc production routes.
92Mo(α,x)94mTc
The four experimental datasets available in the literature are shown in Fig. 120 [36, 268,269,270] together with the TENDL calculations. Three datasets were rejected (Graf and Münzel [268], Denzler et al. [269], and Ditrói et al. [270], all contradictory sets of data), while the remaining single set of Levkovskij [36] was used in the statistical fitting procedure (and also accepted as a standard for the monitoring of α beams). The selected data and their experimental uncertainties are shown in Fig. 121 together with the Padé fit (L = 12, N = 28, χ2 = 1.33) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
94Mo(p,n)94mTc
The seven experimental datasets available in the literature are shown in Fig. 122 [36, 142, 271,272,273,274,275] together with the TENDL calculations. All datasets were used in the statistical fitting procedure. These data and their experimental uncertainties are shown in Fig. 123 together with the Padé fit (L = 9, N = 57, χ2 = 1.21) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).
Thick target yields for production of 94mTc
See Fig. 124.
Production of 110mIn (T
1/2 = 69.1 min) and longer-lived 110Sn parent (T
1/2 = 4.154 h)
Applications: 110mIn is a positron-emitting analogue for established SPECT 111In. Potential to provide more quantitative diagnostic information as well as in vivo quantification of the uptake kinetics of radiopharmaceuticals (e.g., applied along with 111In-labelled DTPA-D-Phe1-octeotride for neuroendocrine tumours).
110mIn can be produced directly and via parent 110Sn.
110mIn (69.1 min): β+ (61.3%), and Eγ (keV) (Pγ(%)): 2129.40 (2.15), 2211.33 (1.74), 2317.41 (1.285).
110Sn (4.154 h): Eγ (keV) (Pγ(%)): 280.459 (97.06).
Evaluations have been made of the natIn(p,xn)110Sn, 108Cd(α,2n)110Sn, 110Cd(p,n)110mIn, 110Cd(d,2n)110mIn and 107Ag(α,n)110mIn production routes.
natIn(p,xn)110Sn
The four experimental datasets available in the literature are shown in Fig. 125 [276,277,278,279] together with the TENDL calculations. All datasets were used in the statistical fitting procedure. The data and their experimental uncertainties are shown in Fig. 126 together with the Padé fit (L = 17, N = 112, χ2 = 1.50) and estimated uncertainty in percentages, including 4% systematic uncertainty (right-hand scale).