Isomer and decay studies for the rp process at IGISOL

This article reviews the decay studies of neutron-deficient nuclei within the mass region \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$\ensuremath A=56\mbox{--}100$\end{document} performed at the Ion-Guide Isotope Separator On-Line (IGISOL) facility in the University of Jyväskylä over last 25 years. Development from He-jet measurements to on-line mass spectrometry, and eventually to atomic mass measurements and post-trap spectroscopy at IGISOL, has yielded studies of around 100 neutron-deficient nuclei over the years. The studies form a solid foundation to astrophysical rp -process path modelling. The focus is on isomers studied either via spectroscopy or via Penning-trap mass measurements. The review is complemented with recent results on the ground and isomeric states of 90Tc . The excitation energy of the low-spin isomer in 90Tc has been measured as \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$\ensuremath E_x=144.1(17)$\end{document} keV with JYFLTRAP double Penning trap and the ground state of 90Tc has been confirmed to be the (8+) state with a half-life of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$\ensuremath T_{1/2}=49.2(4)$\end{document} s. Finally, the mass-excess results for the spin-gap isomers 53Com and 95Pdm and implications from the JYFLTRAP mass measurements for the (21+) isomer in 94Ag are discussed.


Introduction
Decay studies of neutron-deficient nuclides have a long tradition in the IGISOL group in Jyväskylä. Even before the era of the ion-guide method, several studies had been performed with the helium-jet technique at the MC-20 cyclotron in Jyväskylä. Those studies included the T Z = −1 beta-delayed proton/alpha emitters 20 Na [1], 24 Al [2], 24 Al m [2], 28 53 Co [6], and 59 Zn [7]. After the invention of the new ion-guide method [8], these decay studies were continued at IGISOL and focused on the T Z = −1/2 mirror nuclei 43 Ti [6], 51 Fe [9], and 55 Ni [9].
After commissioning of the K-130 cyclotron, another milestone in the decay studies of neutron-deficient nuclides has been the application of the heavy-ion ion-guide HIGISOL in the production of these nuclides [10,11]. Previously, the ions of interest had been produced by light a e-mail: anu.k.kankainen@jyu.fi b Present address: Turku PET Centre, Accelerator Laboratory,Åbo Akademi University, FI-20500 Turku, Finland. ion (p or 3 He) fusion-evaporation reactions, which limited the studies close to the stability line. With HIGISOL, the heavy-ion fusion-evaporation reactions became available, and a broader range of radioactive isotopes could be studied in the heavier mass region.
The third major step forward has been the application of the JYFLTRAP Penning-trap mass spectrometer for the mass measurements of neutron-deficient nuclides at IGISOL. This has also made it possible to measure the excitation energies of the isomers very precisely. JYFLTRAP has also been used for isobarical, and in some cases, isomerical purification of the IGISOL beams. The mass measurements of neutron-deficient nuclides are reviewed in the following paper of this special issue of the European Physical Journal A.
We hope that this contribution provides a view to the history, presence and opportunities of the ion-guiderelated methodology in studying neutron-deficient nuclei beyond the doubly magic 56 Ni. In the following sections, decay studies of mirror nuclei and the isospin triplet at A = 58 will be discussed at first. Then, the mass region around A = 80-90 rich of isomers is discussed with the main focus on the isomers. In addition, the three spin-gap isomers studied at JYFLTRAP are reviewed in sect. 4. The investigated nuclides lie on the path of the astrophysical rapid proton capture (rp) process, a sequence of Sr (38) Zr (40) 34 35 36  Nuclides studied via beta-decay spectroscopy with He-jet, IGISOL and HIGISOL techniques in Jyväskylä. For the lower mass region, the main path of the rp process is drawn for one-zone X-ray burst model according to ref. [12]. For the higher mass region, the time-integrated reaction flow has been plotted for conditions where the thermonuclear burning proceeds in steady state [13]. There, the solid line represents the major path (more than 10% of the reaction flow through 3α reaction) and the dashed line the minor path (1-10% of the reaction flow through 3α reaction). Recent mass measurements at JYFLTRAP [14] have shown that the SnSbTe cycle is not so strong as plotted here. 53 Co has been studied both with the He-jet technique [6] and with JYFLTRAP at IGISOL [15]. The energy of the spin-gap isomer in 94 Ag has been estimated based on JYFLTRAP mass measurements. Decay studies of 23 Al, 31 Cl, and 41 Ti are discussed in a separate article [16].
proton captures and β + decays occurring at high temperatures and hydrogen densities, such as in X-ray bursts [17,18] (see fig. 1). The rp process has been one motivation for these studies as the detailed knowledge of beta decays, half-lives, energy levels and isomers of the involved nuclides is required for an accurate modeling of the process. Long-lived isomers could play a role by increasing the reaction flow and thus reducing the timescale for the rp-process nucleosynthesis during the cooling phase [19].  [21] have been determined for the pf and sd shells, respectively. At IGISOL, beta-decay study of 55 Ni was one of the first experiments performed with the ion-guide method [9]. There, a 27 MeV 3 He beam on enriched 54 Fe target was applied from the old MC-20 cyclotron to produce 55 Ni. With a beta telescope, an endpoint energy of 7.70 (18) MeV and a half-life of 208(5) ms were determined [9]. The γ spectrum was recorded with a 19% Ge(Li) detector but no feeding to other states than the ground state was observed. Also the beta decay of 51 Fe produced with the same beam on 50 Cr 2 O 3 target was studied with the same setup. The obtained half-life and the endpoint energy were 310(5) ms and 7.26(15) MeV [9], respectively. A beta feeding of 5.0(13)% to a 7/2 − state at 237 keV was determined from the intensity ratio to the annihilation peak [9]. The ion-guide method was also used for a half-life determination of 43 Ti with a beta telescope [6]. The ions of interest were produced via 40 Ca(α, n) 43 Ti reactions with an 18 MeV alpha beam from the MC-20 cyclotron. The measured half-life was 509(5) ms [6].

T = 1 triplet at A = 58
In addition to beta decays, Gamow-Teller strength can be studied via charge-exchange (CE) reactions. The CE reactions, such as 58 Ni( 3 He, t) 58 Cu [22], are not limited by the Q value as beta-decay strength studies, and thus, GT strengths over a broader energy range can be investigated. However, the charge-exchange reactions rely on the proportionality between the cross-section near 0 • scattering angle and the GT strength. This proportionality has to be normalized, for example, at A = 58 with the GT strength value from the beta decay of 58 Cu.
In order to provide a more precise value for the normalization, beta decay of 58 Cu was measured at IGISOL. About 4000 ions/s of 58 Cu were produced at IGISOL with an 18 MeV, 15 μA proton beam from the K-130 cyclotron impinging on an enriched 58 Ni target [23]. At the implantation point, the yield of 58 Cu and corresponding γ transitions were monitored by two HPGe detectors and a plastic ΔE β detector. The tape was moved to the measurement position, where the ratio of the intensities of the 511 keV annihilation radiation to the 1454 keV γ-transition in 58 Ni was precisely measured. A branching ratio of 80.8(7)% to the 58 Ni ground state was obtained [23].
The T = 1 triplet at A = 58 offers a possibility to study the isospin symmetry of transitions. Namely, the Gamow-Teller strength to the states in 58 Cu obtained from the ( 3 He, t) charge-exchange reactions on 58 Ni (T Z = +1) [22] can be compared with the strenghts of the analogous transitions from the beta decay of 58 Zn (T Z = −1). For this purpose, beta-decay of 58 Zn has been studied with a betadelayed gamma and proton setups at ISOLDE and at IGISOL [27]. At ISOLDE, the earlier results [28] concerning the half-life and two lowest γ-transitions were confirmed. No beta-delayed protons were observed with the ISOLDE Silicon Ball detector [29] similarly to the previous work [28]. Therefore, the experiment was also tried at IGISOL with a 50 MeV, 1 μA 3 He beam on a nat Ni target. The estimated yield of 58 Zn at IGISOL was about 0.6 atoms/s based on the 203 keV γ-transition [30]. Unfortunately, no beta-delayed protons were observed at IGISOL either.

Studies around A = 80-90 for the rp process
3.1 Yttrium isotopes: 80 Y m , 81 Y, and 82 Y 80 Y m 80 Y has a long-lived isomeric state at 228.5 keV relevant for the astrophysical rp process. Since the spin assignment and internal conversion coefficient were not certain, an experiment aiming for the spin identification of this isomer was performed at HIGISOL [10,11] by employing a 150 MeV 32 S 7+ beam on a 2.7 mg/cm 2 thick 54 Fe target [31]. The isomeric transition was studied with a magnetic conversion electron transporter spectrometer ELLI [32] and a low-energy Ge detector (LeGe). ELLI transported electrons from the implantation point to a cooled Si(Au) surface barrier detector in a remote detection area, which helped to reduce the background. The observed 228.5 keV γ-rays were not in coincidence with the beta particles. Moreover, the coincidence of yttrium K α X-rays with the 228.5 keV transition confirmed the isomeric character. The internal conversion coefficient for the isomeric transition was determined from the ratio of the conversion electrons to the corresponding γ-rays as well as from the ratio of yttrium K X-rays to the 228.5 keV γ-rays. The contribution from the beta decay of 80 Zr to the yttrium X-rays is negligible since the production rate of 80 Zr is very small. The obtained values α K = 0.51(9) [31] and α K = 0.48 (13) [31] agree well with each other and with the result of ref. [33], α K = 0.47 (15). The average value of α K = 0.50(7) [31] gives M 3 + (E4) multipolarity for the 228.5 keV transition and a spin assignment 1 − for the isomer.
The measured half-life of the isomeric transition, T 1/2 = 5.0(5) s [31], is in agreement with the previous result of T 1/2 = 4.7(3) s [34]. The weighted mean for the half-life of the 228.5 keV isomeric transition is T 1/2 = 4.8(3) s, which converts into a bare half-life of T 1/2 = 6.8(5) s [31] corresponding the fully ionized 80 Y m in the stellar plasma of an X-ray burst. In the rp process, 80 Y m is produced via beta decay of 80 Zr which dominantly populates the (1 + ) states at 623 keV and at higher energies [35]. De-excitations of these states lead finally to the 228.5 keV 1 − isomeric state. About 19(2)% of the isomeric decays proceed via beta decay and the rest 81(2)% via internal transition to the 80 Y ground state [36]. Since the decay of the isomeric state is many orders of magnitudes slower than the proton captures on it during most of the rp process, the proton captures will predominantly occur on the isomeric state [31] and not on the ground state of 80 Y as previously assumed. This is a clear example how isomers can play a role in the rp process and should be known for an accurate modeling. 81 Y 81 Y was produced at HIGISOL with a 150-170 MeV 32 S 7+ beam impinging on an enriched 54 Fe (1.8 mg/cm 2 ) target [37]. The beta-decay of 81 Y was studied at HIGISOL with a setup consisting of two HPGe detectors, a LeGe detector and the magnetic conversion-electron transporter spectrometer ELLI [32]. Before this HIGISOL experiment [37], the beta-decay of 81 Y had been studied already in refs. [38][39][40]. All known γ-transitions [41] following the beta-decay of 81 Y could be observed at HIGISOL except the 216.6 keV transition. Conversion electrons from the 43.2, 79.23, 101.05, 115.39, 124.16, and 155.20 keV transitions in 81 Sr were measured. The internal conversion coefficient for the 43.2 keV transition was determined as α K = 1.5(3), which supports an M 1+E2 multipolarity for the transition [37]. The obtained low conversion coefficient (α K = 0.03(1)) for the 124 keV transition supports an E1 multipolarity for the transition. The measured conversion coefficient for the 79.23 keV transition, α K = 2.3(1), agrees nicely with the literature value of 2.3(6) [39], and supports an E2 multipolarity. Figure 2 shows an example

of conversion electron spectrum measured with ELLI at
Y was produced at HIGISOL using the reaction 165 MeV 32 S + nat Ni [10]. A 37% Ge detector was used for detecting γ-rays, a LeGe detector for X-rays and a 1 mm thick plastic scintillator for detecting beta particles [10]. The beta-decay of 82 Y had been studied via 60 Ni( 24 Mg, pn) 82 Y and 60 Ni( 28 Si, αpn) 82 Y reactions with beam energies of 90 and 110 MeV [38] and via 54 Fe( 32 S, 3pn) 82 Y reaction at an energy of 123 MeV [39]. The measured half-lives of 9.5(4) s [38] and 9.5(5) s [39] agree with each other but the values for the relative intensity of the 602 keV γ-transition, 41% [38] and 13(1)% [39], differ a lot. At HIGISOL, the half-life was determined from the beta particles, annihilation γ-rays, as well as from the 573, 602, 737, 1176 and 1291 keV γ-transitions following the beta-decay of 82 Y [10]. The shorter-lived 82 Zr might distort the half-lives obtained from the beta particles and from the annihilation radiation, and since the measured half-life values of 82 Zr vary a lot, the value of 8.30 (20) s based on the most intensive, 573 keV γtransition was adopted for 82 Y [10]. The statistics of the other γ-transitions were too poor for precise half-life fitting. The half-life measured at HIGISOL disagrees with the former values of [38,39] but is much more precise.
The obtained relative intensity of 4.4(4)% [10] for the 602 keV transition deviates from the previous He-jet studies [38,39]. In the He-jet studies no mass separation is used and contamination for example from the 602 keV γtransitions following the beta-decay of 84 Y, (6 + ) state, cannot be excluded. Therefore, the obtained relative intensities for the 602 keV transition in [38,39] can be higher than at HIGISOL, where a clear spectrum at mass A = 82 was obtained.
The mass of 82 Y has been measured precisely with the JYFLTRAP Penning-trap mass spectrometer [42]. (AME03) [43] value is 127(100) keV. The updated Q EC value for 82 Y is 7946(8) keV, which is 130 keV higher and 13 times more precise than in the AME03. The updated log ft values of the 82 Y beta-decay are little higher than previously (see fig. 3). Neutron-deficient yttrium isotopes have typically isomers. No clear evidence for an existence of an isomer in 82 Y with a significantly different half-life than the 1 + ground state was observed in ref. [10]. On the other hand, for example the production of 84 Y was found to be almost totally concentrated on the 5 − , 39.5 min isomeric state: only about 1% of the 84 Y production was due to the 1 + ground state based on the intensities of the 793 keV and 974 keV γ-transitions [10]. 85 Nb m A 37% Ge detector for detecting γ-rays, a LeGe detector for X-rays and a 1 mm thick plastic scintillator for detecting beta particles were employed in a spectroscopic survey at A = 85 carried out at HIGISOL by using 165 MeV 32 S + nat Ni reactions [10]. Several unknown γrays were found with energies of 166, 272, 423, 434, 484, 532, 538, 590, 610, 660, 709, and 759 keV at A = 85. The most intense, 759 keV γ transition had a half-life of 12(5) s agreeing with the beta-decay half-life of 85 Zr m (T 1/2 = 10.9(3) s [44]), and disagreeing with the half-lives of 85 Mo (T 1/2 = 3.2(2) s [45]) and 85 Nb (T 1/2 = 20.9(7) s [46]). However, the transition could not be associated with the beta-decay of 85 Zr m since such an intense transition should have been observed in the detailed beta-decay study of ref. [47]. At HIGISOL, only the two most intensive peaks at 416 and 454 keV following the 85 Zr beta decay were observed. The half-life for the unknown 484 keV transition was determined as T 1/2 = 4 +8 −2 s [10] consistent with the reported half-life of 85 Mo. This transition could not be associated with the beta decays of 85 Nb and 85 Zr due to the different half-lives. Similarly to the 759 keV transition, the 484 keV transition should have been detected in ref. [47] if it was associated with 85 Zr m . As a summary, the 759 keV, 12(5) s transition follows either the beta-decay of 85 Nb m or 85 Mo m . In addition to the isomer in Nb or Mo, the 484 keV transition could originate from the beta decay of 85 Mo.

Niobium isotopes: 85 Nb m and 86 Nb
Unknown γ-rays at energies of 272, 484, 530, 536, 590, 660, 709, and 758 keV were also observed with two HPGe detectors at the implantation point in a run employing a 150-170 MeV 32 S 7+ beam on a nat Ni target at HIGISOL [37]. After 15 or 40 seconds, the activity was moved to the LeGe detector station. There, none of these γ-rays could be observed supporting short half-lives for these transitions. It should also be noted that about half of the observed 12 unknown γ-transitions in ref. [10] at A = 85 have an energy difference of 50 keV: 484−434 = 50 keV, 660 − 610 = 50 keV and 759 − 709 = 50 keV. In addition, the differences 532 − 484 = 48 keV and 590 − 538 = 52 keV are quite close to 50 keV which corresponds to the energy of a well-known 50 keV transition in 85 Zr. One possibility is that these γ-rays follow the beta decay of an isomer in 85 Nb to the possible states at 484 keV, 660 keV and 759 keV (and possibly at 590 keV), which de-excite to the (7/2 + ) ground state and to the (7/2 + , 9/2 + ) state at 50 keV in 85 Zr. However, no coincidences have been checked between these γ-rays or between the Zr, Nb or Mo X-rays and these transitions. Thus, we cannot conclude to which nucleus these transitions belong. Trap-assisted spectroscopy at A = 85 would enlighten the situation provided the yields are high enough.
In the latter run at A = 85 [37], conversion electrons at 32.1 and 50 keV were observed with ELLI. The 32.1 keV conversion electrons were in coincidence with Zr K Xrays, thus confiming that they belong to the 50 keV transition in 85 Zr. The determined internal conversion coefficient for the 50 keV transition, α K = 1.7(2), suggests a mixed M 1 + E2 multipolarity for the transition. The half-life determined for this 50 keV transition fed by the beta decay of 85 Nb was 17(2) s, which is little shorter than T 1/2 = 20.9(7) s given for 85 Nb in ref. [46].
The 50 keV conversion electrons were in coincidence with the Nb K X-rays and had a half-life of 3.3(9) s. Since the electrons were not in coincidence with beta particles and because the production rate of 85 Mo is much smaller than for 85 Nb, we identify this transition as a 69 keV isomeric transition in 85 Nb. Based on the internal conversion coefficients, the preferred multipolarity for the transition would be E2 or M 2, which would mean a much shorter half-life than 3.3 s. An indication of beta-decay of this isomer to the 292 keV isomeric state in 85 Zr is observed in the time behaviour of the 292 keV γ-line [37].
At JYFLTRAP, the Q EC value of 85 Nb has been measured as 6898(9) keV [42] which is 900(200) keV higher than the previous value [46]. There, the beta-endpoint energy was determined from the beta spectrum in coincidence with the 50 keV transition in 85 Zr. If the 50 keV transition is fed by de-excitations from possible states at 484 keV, 660 keV and 759 keV in 85 Zr, this would increase the deduced Q EC value of ref. [46] and move it closer to the JYFLTRAP value. Although the beta-decay experiments tend to underestimate Q EC values, such a big difference suggests that it should be verified for which state the mass of 85 Nb has been measured. For example, a half-life measurement using the isobarically purified beta emitters after JYFLTRAP would allow us to distinguish between the 20.9 s ground state and the suggested isomers with halflives of 3.3(9) s [37] and 12(5) s [10].  86 Mo was determined from the time behaviour of γ, Nb K X-ray, and electron peaks [37]. This is in agreement with the value of 19.6(11) s obtained in ref. [48]. No evidence for the 56 s isomer was found in ref. [37]. No converted low-energy transitions with an energy E 0 were observed. Thus, the lowest low-spin state E 0 fed by the 86 Mo beta-decay, should be either highly excited or decay mainly via beta-decay to 86 Zr.
At JYFLTRAP, the mass-excess value of 86 Nb [37] was found to be 700(90) keV higher than in the AME03 [43]. Here, again, the ground-state nature of the measured mass should be verified for example via half-life determination after the trap. In addition, the production ratio of the isomer to ground state at HIGISOL can shed light on which mass has been measured.

90 Tc m
The beta decay of 90 Tc has been studied in refs. [49] and [50]. A 1 + state with a half-life of 8.7(2) s and a high-spin state (6 + ) with a half-life of 49.2(4) s were found in ref. [50]. The corresponding excitation energy of the suggested (6 + ) isomer has been estimated to be 124(390) keV [51]. In ref. [52], the ground state was suggested to be an 8 + state based on the systematics of all neighbouring N = 47, 49 nuclei. In a study of the 90 Ru 0 + ground-state beta-decay [53], feeding to three low-spin states in 90 Tc was found. There, a high-spin 8 + ground state was obtained, when the low-spin and high-spin results of shell-model calculations were compared to the experimental results. 90 Tc has been studied previously at HIGISOL via a JYFLTRAP mass measurement [51]. There, the measured mass-excess value was assigned for the (8 + ) ground state. Since the identification of the ground or isomeric state is uncertain, a new experiment was performed with a 40 Ca 8+ beam on a nat Ni target at 210 MeV. In that experiment, a half-life for the mass A = 90 was determined by measuring the time behaviour of beta particles with a silicon detector after the IGISOL dipole magnet. A half-life of T 1/2 = 50.7(63) s was obtained with a one-component fit (see fig. 4). This is in agreement with the half-life of the high-spin state (8 + ), T 1/2 = 49.2(4) s [50] and shows that the contribution from the low-spin 1 + , 8.7(2) s state [50] is small. The half-lives of 90 Nb (14.60(5) h [54]) and 90 Mo (5.56(9) h [54]) can be considered to be constant within the time scale of 200 s. 90 Ru (T 1/2 = 11(3) s) is much less produced than 90 Tc and thus, does not contribute much to the beta-particle time behaviour. The Mo and Nb isomers decay via internal transition and do not contribute to the beta spectrum.
The isomer in 90 Tc was also observed with JYFLTRAP in the experiment employing 40 Ca 8+ + nat Ni at HIGISOL, as shown in fig. 5. Three measurements of the 90 Tc ground state and five measurements of the isomeric state against the 86 Kr reference were performed with the time-of-flight ion-cyclotron resonance (TOF-ICR) method [55,56]. In the precision trap, a quadrupolar RF excitation time of 800 ms was applied for the ions of interest. In total, there were 14331 ions of 90 Tc and 4229 ions of 90 Tc m in the analysis. The count-rate class analysis [57] was performed for the ground state with three classes. For the isomer, no count-rate class analysis could be performed. Therefore, the ion number was limited to 1-2 ions per bunch in the analysis of the isomeric state. The uncertainty of the cyclotron frequency of the isomer was multiplied by a factor    86 Kr + as a reference. The result for the ground-state mass agrees with the previous experiments at JYFLTRAP (ME = −70723.6(39) keV) [51] and SHIPTRAP (ME = −70724.1 (74) keV) [51], which employed 85 Rb + ions as a reference.
Nuclide r ME (keV) Ex (keV) 90 Tc 1.046717034(13) −70724.7(11) 0 90 Tc m 1.046718834(16) −70580. 6(13) 144.1 (17) of two. This was based on a comparison of the non-Zclassed results to the Z-classed results of the ground state and the reference. The weighted mean of the frequency ratios and the internal and external errors [58] were calculated. The internal errors were bigger than the external errors showing that the fluctuations of the frequency ratios were purely statistical. The mass-dependent and residual uncertainties [59] were added to the measured frequency ratios quadratically. The resulting frequency ratios and mass-excess values are given in table 1.
When the time-of-flight spectrum for 90 Tc is fitted with two peaks (see fig. 5(a)), the fractions of the lowermass state and the higher-mass state are 89(5)% and 11(5)%, respectively. A very similar result is obtained if a two-component fit is performed to the beta-particle time behaviour at A = 90: half-lives of 60(21) s and 8 (14) s are obtained with fractions of 84 (19)% and 16(16)%, respectively. This indicates that the low-spin state is much less produced than the high-spin state at HIGISOL. Thus, we confirm that the previously measured mass-excess value of 90 Tc belongs to the more favorably produced (8 + ) ground state with a half-life of T 1/2 = 49.2(4) s.
The excitation energy obtained for the isomer, E x = 144.1 (17) keV, fits well with the shell-model prediction of a 2 + state (143 keV [52], 160 keV [53]). However, log ft values of around 5-6 to the 0 + ground state and to the 2 + state at 948.1 keV in 90 Mo [49,50] support a 1 + assignment. The 2 + assignment is possible if the ground-state feeding has not been correctly determined.

93 Ru m
The beta-decay of 93 Ru m (T 1/2 = 10.8(3) s, J π = (1/2 − )) has been studied via a 24 MeV 3 He beam from the University of Jyväskylä MC-20 cyclotron on an isotopically enriched 92 Mo target employing the He-jet technique [61]. The total decay energy of the isomer was measured as 7070 (85) keV and the proton binding energy of 93 Tc as 4086.5(10) keV [61]. The excitation energy of the isomer is 734.4(1) keV [62]. The transition was suggested as an isomeric transition in 93 Ru since it was not in coincidence with any other γ-ray or annihilation radiation. Later, in ref. [53], the beta-decay of 93 Rh was studied via the reaction 58 Ni( 40 Ar, p4n) 93 Ru. There, no evidence for the (1/2) − isomeric state in 93 Ru was found. Thus, the lowspin isomeric state is not populated at all or only weakly populated in these kind of reactions.
The mass of 93 Ru has been studied at JYFLTRAP [51] and it agrees with the values of ref. [61]. The ground state should be well resolved from the isomeric state at 734.4.(1) keV. In future, it would be interesting to measure the mass of the isomeric state by using similar reaction as in ref. [62] at IGISOL. These measurements could have implications for the rp process, because such long-lived isomers can play a role in it.

53 Co m
The first observation of proton radioactivity was from the high-spin (19/2 − ) isomer in 53 Co [65,71,72]. This state with a half-life of 247(12) ms [72] is the isobaric analogue state of the (19/2 − ) isomeric state in its mirror nucleus 53 Fe [64]. The excitation energy of 53 Co m as determined from the energy of the observed protons, is 3197(29) keV [71,72]. The isomeric state decays mainly via beta-decay to its analogue state in 53 Fe -only about 1.5% decays to 52 Fe via proton decay [72].
In an experiment at the JYFLTRAP mass spectrometer, both the ground and the isomeric state of 53 Co could be produced with a 40 MeV proton beam impinging on an enriched 2 mg/cm 2 thick 54 Fe target [15]. At JYFLTRAP, a Ramsey excitation with a time pattern of 25-150 (wait)-25 ms was applied for 53 Co and its reference 53 Fe. The ground state and the isomeric state were measured against each other and against the 53 Fe ground state. The obtained mass-excess values were −42657.3 (15) and −39482.9 (16) keV for the ground and isomeric state, respectively [15]. The mass-excess value for the ground state agrees well with the earlier results [43,73]. The massexcess value of the isomer agrees with the latter protondecay experiment [72] but deviates from the first experiment [71]. The excitation energy for the (19/2 − ) isomeric state obtained as a weighted average from the frequency ratio 53 Co m -53 Co and from the frequency ratios of 53 Co m -53 Fe and 53 Fe-53 Co, was 3174.3(10) keV [15]. This new value agrees well with the previously adopted value 3197(29) keV [43] but is 29 times more precise. With the 52 Fe mass from ref. [43], this would correspond to a proton energy of E p,lab = 1530(7) keV [15].
Since 53 Co and 53 Co m were measured against their beta-decay daughter 53 Fe at JYFLTRAP, also Q EC values could be determined precisely. The obtained Q EC values from the ground and isomeric state of 53 Co to the ground state of 53 Fe were 8288.12 (45) keV and 11462.2(12) keV, respectively [15]. These precisely measured mirror decay Q EC values are important for an accurate determination of the corrected ft values. In addition, a Coulomb energy difference of 133.9(10) keV [15] was obtained for the (19/2 − ) states in 53 Fe and 53 Co, which improves the adopted value 157(29) keV [43] a lot.
In future, a direct measurement of the proton separation energy of 53 Co at JYFLTRAP would yield a sub-keV value for this separation energy. This, in turn, would give a precise calibration value for proton spectroscopy experiments. In addition, precise mass measurements of 53 Fe, 53 Fe m and 52 Fe m could be considered in future at JYFLTRAP.

94 Ag m
Two high-spin states, (7 + ) and (21 + ) have been observed in the N = Z nucleus 94 Ag. After beta-delayed γ-ray and proton studies of the isomeric states in 94 Ag performed at GSI [66,67,74,75], direct one-proton decay from the (21 + ) state to the high-spin states in 93 Pd was observed by detecting protons in coincidence with γ-γ correlations and applying γ gates based on the known 93 Pd levels [76]. A year later also direct two-proton radioactivity was observed from the (21 + ) isomer by detecting fourfold coincidences between proton-proton coincidences measured with the Si detectors and γ-γ coincidences measured with the Ge detectors at GSI [77]. The γ-γ gates were set to two γ-transitions in the daughter nucleus 92 Rh. A twoproton energy of 1900(100) keV [77], an excitation energy of 5780(30) keV [76], and a half-life of 0.39(4) s [75] has been obtained for this (21 + ) isomer.
At JYFLTRAP, we have measured the mass excesses of 92 Rh and 94 Pd [51], the two-proton decay daughter of the (21 + ) isomer, and the beta-decay daughter of 94 Ag, respectively. The mass-excess value of 94 Ag was not directly measured but it was extrapolated to −53330(360) keV based on the linear behaviour of the Coulomb displacement energies of odd-odd N = Z nuclei [78]. The extrapolated mass of 94 Ag yields a two-proton separation energy S 2p = −Δ( 94 Ag) + Δ( 92 Rh) + 2Δ( 1 H) = 4910(360) keV [78], where Δ refers to the mass excess. If this value is now combined with the two-proton decay data [77], an excitation energy of 8360(370) keV [78] is obtained for the (21 + ) state.
The one-proton decay daughter 93 Pd was not measured directly at JYFLTRAP. Instead, two-proton separation energies of several N = 47 isotones have been determined at JYFLTRAP. A parabolic fit to the smooth behaviour of these S 2p values yields S 2p = 5780(160) keV [78] for 93 Pd. Since the mass excess of 91 Ru has been measured at JYFLTRAP, we obtain from the S 2p ( 93 Pd) value a mass excess of −59440(160) keV for 93 Pd. Finally, we can calculate the proton separation energy of 93 Pd with the mass of 92 Rh measured at JYFLTRAP, S p ( 93 Pd) = 3730(160) keV [78]. From the mass-excess values of 93 Pd and 94 Ag, a proton separation energy of 1180(390) keV [78] is obtained for 94 Ag. If we now combine this value with the one-proton decay data [76], we obtain an excitation energy of 6960(400) keV [78] for the (21 + ) isomer. This disagrees with the value based on the two-proton decay data and suggests that either the oneproton or two-proton decay data of the (21 + ) state in 94 Ag needs revision.
The two-proton decay of the (21 + ) isomer in 94 Ag has raised a lot of discussion recently. Shell-model calculations using a gds model space have shown that the additional binding energy due to the large attractive pn interaction in the 0g 9/2 orbit lowers the energy of the 21 + state primarily but the level inversion is caused eventually by the mixing with the 1d 5/2 configurations [79]. The gds shellmodel calculations [79] do not support strong deformation for the 21 + state in contrast to ref. [77], where strong prolate deformation was suggested to explain the unexpectedly high two-proton decay probability [77]. In addition, recent calculations based on statistical theory of hot rotating nucleus combined with the macroscopic-microscopic approach do not support a strong prolate deformation [80].
The direct two-proton decay has been questioned in many papers. For example, the level scheme of 92 Rh has been improved from ref. [81]: the tentative 575 keV γtransition [81] was not observed, the 307 keV transition was assigned as the 20 (−) → 19 (−) transition, the 632 keV γ-rays as the 19 (−) → 18 (−) transition and the order of the 939 and 1034 keV transitions was reversed [82]. No γrays at energies of 565 and 833 keV in coincidence with the known γ-transitions in 92 Rh were observed in ref. [82] in contrast to ref. [77]. The data of ref. [82] suggest that the spin difference between the two-proton decay mother and daughter should be greater than 10 and that the isomer should be even more deformed. When the in-beam data of ref. [82] is combined with the (extrapolated) mass-excess data of AME03 [43], the two-proton decay scenario seems to be impossible. Later, these considerations have been discussed in refs. [83,84].
Protons could not be unambigously identified in ref. [77] since two 1 mm thick Si detectors, which are also sensitive to electrons and positrons, were used. Detailed discussion in ref. [85] shows that the 1.9 MeV two-proton peak could originate from Compton-scattered γ-rays following the beta-decay of the 7 + isomer in 94 Ag and annihilation radiation, and not from 92 Rh. However, these considerations are based on the assumption that Compton scattering events between adjacent Ge crystals had not been reduced in ref. [77]. In a recent paper [86], it is clarified that the Compton scattering effects in γ-γ coincidence events were reduced by excluding double hits in adjacent Ge crystals while they were accepted in the other crystals [67,86]. In addition, the coincidence events with the sum energy of the two γ-rays corresponding to 511 ± 1.5 keV were excluded for all crystals [86].
To identify the protons from the 21 + isomer in 94 Ag, a measurement employing an array of 24 ΔE1(gas)-ΔE2(gas)-E(Si) detectors was performed at Lawrence Berkeley National Laboratory [87]. There, the total twoproton coincidence yield for the used reaction of 197 MeV 40 Ca beam on 58 Ni target should have been comparable to the yield at GSI. However, only the lower energy one-proton decay group at 0.79(3) MeV with its branching ratio of 1.9(5)% was confirmed [87]. No evidence for the direct two-proton decay was found [87]. Thus, ref. [87] suggests that the one-proton decay data should be correct and we should adopt the excitation energy of 6960(400) keV [78] for the (21 + ) isomer.
At IGISOL, developments to produce 94 Ag continue [88,89]. The hot cavity laser ion source has been successfully commissioned at IGISOL and the laser ionization has been found to be fully saturated [89]. If the tests with stable 107 Ag 21+ beam from the K-130 cyclotron will be successfull, the on-line experiments with 94 Ag will be carried out at IGISOL. In future, the mass measurements of the ground and isomeric states of 94 Ag with JYFLTRAP Penning-trap mass spectrometer would solve the two-proton decay energy puzzle. The measurement of the hyperfine structure of the isomeric states would experimentally determine the spectroscopic quadrupole moment, and thus, also the shape of the isomer.

95 Pd m
The (21/2 + ) isomer in 95 Pd with a half-life of 13.3(3) s [90] was first observed at Munich MP tandem via beta-delayed proton and gamma spectroscopy [69]. The total beta branching to proton-emitting states was determined as 0.74 (19)%. The final state to be populated after the beta decay was found to be 8 + in 94 Ru. The most strongly populated level in the beta decay of 95 Pd m was the (21/2 + ) level at 2449 keV in 94 Rh. Soon after this experiment, the known properties of 95 Pd were confirmed at GSI [90]. Shell-model calculations [91] suggest that this isomer lies at 1.90 MeV and is a spin-gap isomer in nature. Other shell-model calculations have predicted excitation energies of 1803 keV [92,93] and 1973 keV [94] for this spingap isomer. The extrapolated value in the AME03 is 1860(500)# keV [43]. The energy of the isomeric state was experimentally determined as 1876 keV [68] based on the γ-transitions connecting the states built on the ground and isomeric states of 95 Pd.
At JYFLTRAP, a 170 MeV 40 Ca beam on a nat Ni target was used for the production of the ground and isomeric states of 95 Pd [51]. For these rather long-lived states with an energy difference of about 2 MeV, a quadrupole RF excitation time of 800 ms in the precision trap allowed separate mass measurements for the ground and isomeric state. The reference ion used in these measurements was 94 Mo. The measured mass-excess values for the ground and isomeric state of 95 Pd were −69961.6(4.8) keV and −68086.2(4.7) keV [51], respectively. This yields an excitation energy of 1875.4(6.7) keV [51] for the (21/2 + ) isomer in 95 Pd in agreement with the result of ref. [68]. It further confirms the spin-gap character of the transition as the 21/2 + state lies lower in energy than the 15/2 + (1973 keV) and 17/2 + (1879 keV) states [68]. A recent mass measurement of the reference 94 Mo against 85 Rb has shown that the mass excess of 94 Mo is off by −3.0 (21) keV. This has an effect on the mass-excess values of the ground and isomeric states but does not change the value for the excitation energy of the isomer.
The obtained Q EC value for the isomer, 10256.1(6.3) keV [51], together with the branching ratio of 35.8% [90] for the beta decay to the (21/2 + ) state at 2449 keV in 94 Rh, yields a log ft = 5.5 for this most dominant betatransition. This confirms the allowed character of the transition. The measurement of 95 Pd m was the first direct Penning-trap mass measurement of such a high-spin isomer at JYFLTRAP. It proved that JYFLTRAP can also be used for the studies of spin-gap isomers.

Conclusions and outlook
The studies of neutron-deficient nuclides at the rp-process path have been an important part of the physics performed at IGISOL. Decay studies and Q EC -value measurements of the mirror beta decays as well as the studies of the T = 1 triplet at A = 58 with the light-ion ion guide have yielded information necessary for precise and corrected ft values and for the studies of the isospin symmetry of transitions. The HIGISOL method has shown its strength in the production of Y, Nb and Zr isotopes. JYFLTRAP Penning-trap mass spectrometer has been employed for Q EC -value measurements as well as for measuring the masses or excitation energies of the long-lived isomers, such as 53 Co m and 95 Pd m .
By combining the latest state-of-the-art detectors for spectroscopy experiments and employing JYFLTRAP for isobaric purification, many of the older experiments performed with He-jet or IGISOL techniques could be superseded in future. Other future challenges are, for example, the identification of isomeric and ground states in the mass A = 80-90 region via post-trap spectroscopy. Since the states are populated differently in the heavy-ion and light-ion fusion-evaporation reactions, some isomeric states, such as 93 Ru m or 91 Tc m , could also be produced with the light-ion ion-guide method similarly to the earliest studies of these isomers. There are still many open questions in this mass region, such as the observed γ-rays at A = 85, and the excitation energies of 85 Nb m , 86 Nb m and 94 Ag m . For the production of 94 Ag m (21 + ) isomer, the development and testing of the hot-cavity laser ion source at IGISOL are essential.
Spectroscopic experiments provide information on the ground, excited and isomeric states (spins, energies, halflives), and possible decay modes of the studied nuclides. These data are relevant for the rp and recently proposed νp process [95] network calculations, which should also take into account the contribution from the isomers. Close collaboration with nuclear astrophysicists modeling the rp and νp processes will continue in future. It would be beneficial to establish a large campaign to measure the interesting spectroscopic details within the relevant mass region at the new IGISOL4 facility in collaboration with the rp and νp process modelers. In this way, the impact of new data on the modeling of the rp or νp processes would be seen immediately.