Mass measurements of neutron-deﬁcient nuclides close to A = 80 with a Penning trap

. The masses of 80 , 81 , 82 , 83 Y, 83 , 84 , 85 , 86 , 88 Zr and 85 , 86 , 87 , 88 Nb have been measured with a typical precision of 7 keV by using the Penning trap setup at IGISOL. The mass of 84 Zr has been measured for the ﬁrst time. These precise mass measurements have improved S p and Q EC values for astrophysically important nuclides.


Mass measurements of neutron-deficient nuclides close to A = 80 with a Penning trap 1 Introduction
The masses of neutron-deficient nuclides close to A = 80 are important for modeling the astrophysical rapid proton capture process (rp process) [1].The rp process occurs as a sequence of proton captures and β + decays in various astrophysical sites, such as on the surface of an accreting neutron star (see, e.g., ref. [2]).The flow of the rp process follows a path close to the N = Z nuclei up to 56 Ni.At higher masses, the path broadens and shifts by about one or two units towards stable nuclei.The path is wider for steady state burning (e.g., in X-ray pulsars) than for X-ray bursts.It is shown for relevant nuclei in steady-state burning conditions in fig. 1. Regardless of the burning type, the rp process can proceed up to A ≈ 100 ending in a closed SnSbTe cycle [3].Earlier it was proposed that the rp process would end in a Zr-Nb cycle in which 83 Nb is processed back to 80 Zr via (p,α) reactions [2].However, the existence of the Zr-Nb cycle depends on the α separation energy of 84 Mo which has not been measured.
The rp process is mainly determined by nuclear masses and β + decay half-lives of the nuclides on its path [2].The theoretical mass predictions for the nuclides in the A = 80 region give too high uncertainties for the rp process calculations and therefore, reliable experimental data on masses are needed.The rp process models depend critically on the Q values for proton captures which drive the path towards the proton drip line until a waiting-point nucleus is reached.In order to estimate the proton capture rates, a e-mail: anu.kankainen@phys.jyu.fi the masses and excitation energies should be known better than 10 keV [4].At a waiting point, an equilibrium between proton captures and photodisintegration is established and the rp process can continue only via a 2p capture or a β + decay.As the destruction rate of a waiting-point nucleus depends exponentially on its one-proton capture or two-proton capture Q values [2], nuclear masses play the main role close to the waiting-point nuclides, such as 80 Zr in the A ≈ 80 region.In this mass region, the rapid changes in nuclear structure make the extrapolation of masses and prediction of half-lives very difficult.As the half-life predictions depend severely on the Q EC values due to the energy dependence of the Fermi function, the Q EC values for β + decays with unknown half-lives are of special interest.In general, actual beta-decay schemes and therefore Q EC values, have to be known in order to determine the energy production of the rp process [4].In this paper, we present new mass measurement data on 13 neutron-deficient isotopes of Y, Zr and Nb approaching the Z = N line.
Isomeric states are common for nuclides close to A = 80.Previously, these isomers have been studied spectroscopically (see refs.[5,6]) using the HIGISOL technique (Heavyion Ion Guide Isotope Separator On-Line [7]).Nevertheless, excitation energies of many isomers in this region are still unknown, such as 85 Nb m and 86 Nb m .The identification of these isomeric states would be an important input to nuclear reaction network calculations modeling the rp process.

Experimental method
The isotopes of interest were produced by a 32 S beam from the Jyväskylä K-130 cyclotron impinging on an enriched 54 Fe or nat Ni target.The reaction products were stopped in the HIGISOL gas cell at the IGISOL facility [8].After the extraction from the gas cell by a helium flow and a small extraction potential, the ions were accelerated to 30 keV and mass-separated by a 55 • dipole magnet (M/∆M ≈ 500).The yields were about the same as in a previous experiment [6] which focused on the spectroscopy of the isomeric states in this mass region.
Mass-separated ions of interest were sent to the radiofrequency quadrupole ion beam cooler and buncher [9].After cooling and accumulating for a short time the ions were extracted as a short bunch and transported to the Penning trap system.The JYFLTRAP setup is shown in fig. 2. It consists of two cylindrical Penning traps in a superconducting magnet with a field strength of 7 Tesla [10].The traps are placed in the homogeneous regions of the magnet separated by 20 cm.The first trap, the purification trap filled with helium gas at a pressure of about 10 −4 mbar, is dedicated for isobaric cleaning using a mass selective buffer gas cooling technique.The second trap called the precision trap operates in vacuum and is used for high precision mass measurements employing the time-of-flight ion cyclotron resonance technique [11].In a Penning trap ions possess three different eigenmotions: one axial and two radial motions, the latter corresponding to magnetron (ν − ) and reduced cyclotron motion (ν + ) with frequencies summing up to the true cyclotron frequency Cooled and bunched ions from the buncher captured in the purification Penning trap were further cooled axially by collisions with helium atoms and stored at the center of the trap.Possible isobaric contaminations were removed by the application of RF dipole excitations in combination with helium buffer gas.Immediately after that a RF quadrupole excitation at a mass dependent cyclotron frequency (ν c ) was applied which leads to centering of the ions of interest.A total cycle time of about 400 ms was used and a mass resolving power of about 10 5 could be reached in this configuration.The cleaning procedure has been described in detail in ref. [12].
Isobarically purified and cooled ions were transported to the precision trap.In this trap, RF dipole excitations employing the phase locking technique [13] were applied in order to enhance the magnetron radius.Then the RF quadrupole excitations were applied to convert periodi- cally a slow magnetron motion to a fast reduced cyclotron motion with a higher radial energy.An excitation time of 900 ms was chosen with a corresponding excitation amplitude which finally allows a full conversion to the reduced cyclotron motion at the resonance frequency.Finally, the ions were ejected from the trap.They drift towards the micro channel plate (MCP) detector through a set of drift tubes where a strong magnetic-field gradient exists.Here the ions experience an axial force which converts the radial kinetic energy of the ions into longitudinal energy.As a result, an ion in resonance with the applied quadrupole excitation field moves faster towards the detector than an ion which was off resonance.By measuring the time of flight as a function of the applied excitation frequency, the cyclotron frequency of the ions under investigation was determined.The mass of an ion can be determined: where ν c is the cyclotron frequency of the ion, m is the mass and q is the charge state of the ion.The magnetic field B is calibrated with a well-known reference mass.Stable Mo isotopes 96 Mo and 98 Mo originating from the Havar window of the HIGISOL facility were used as references in this experiment.Examples of the measured timeof-flight resonances are shown for the odd-odd nucleus 80 Y (fig.3), for the even-even nucleus 84 Zr (fig.4) and for a reference nucleus 98 Mo (fig.5).
In summary, overall precision of 7×10 −8 was achieved in these measurements.Typical transmission from the focal plane of IGISOL through the cooler and trap system was about one percent.

Results
The masses of 80−83 Y, 83−86,88 Zr and 85−88 Nb were measured during this experiment.All measurements were done with 900 ms excitation time in the precision trap.The stable isotopes 96−98 Mo were used as reference masses.In order to test the reliability of the system, 96 Mo and 98 Mo  were measured relative to 97 Mo.As shown in table 1, the results for the Mo isotopes agree nicely with the values of the 2003 Atomic Mass Evaluation (AME) [14].Except for the stable Mo references, all nuclides were measured as oxides formed with 16 O in the HIGISOL.
The results of the analysis are given in table 1. Apart from the uncertainty of the reference masses, 1.9 keV for 96,97,98 Mo, and in the case of oxides that of 16 O (0.0001 keV), three systematic uncertainties were taken into account in the analysis.
Firstly, in an offline study comparing 129 Xe to 16 O 2 the resonance frequency was found not to be perfectly linear in mass, but to have a slight offset of 7 × 10 −10 per mass unit difference between the compared masses.This shift was accounted for by quadratically adding this uncertainty multiplied by the difference in mass number between the reference and the measured nucleus (molecule for the oxides) to the uncertainty of the final average for each studied nucleus.
Secondly, large numbers of ions in the trap can cause the frequency to shift.In this experiment, typically about 500-1000 ions were used for the analysis after setting time and countrate windows.To estimate the resulting uncer-tainty, an analysis of the frequency depending on the countrate was done for all measurements of 98 Mo.A shift of about 0.01 Hz per ion, mostly towards higher frequencies, was observed.Since 98 Mo is stable, only a few impurities are expected to be in the trap, and the effect is expected to be smaller than for unstable nuclei, where impurities are produced by decay.However, since all nuclides studied in this work have half-lives considerably longer than the employed excitation time, this difference is small.In the final analysis the maximum number of ions in the trap was therefore limited to three, and an uncertainty of 0.04 Hz was quadratically added to the uncertainty of the final average for each nucleus, corresponding to about 3 keV in this mass range.
Thirdly, to compensate for fluctuations of the magnetic field and of the electronics, the fluctuation of the measured frequency of 98 Mo from one measurement to the next was examined.The average instability was found to be of the order of 0.04 Hz.Since the fluctuation is assumed to be completely random, this value was quadratically added to the statistical uncertainty of each measured frequency; its influence decreases with an increasing number of measurements for the nucleus.

Discussion
In this section, the role of isomeric states in the measured nuclides is discussed.The results for Y, Zr and Nb isotopes are reported nuclide by nuclide and compared to earlier experiments and the values of the 2003 Atomic Mass Evaluation.Proton separation energies and Q EC values determined for the measured nuclides are collected in tables 3 and 4.

Isomers
Several of the measured nuclides have isomeric states (see table 2).The observation of these isomers is experimentally limited by the half-life, the excitation energy and by their relative production rate (isomer ratio).With typical production rates in this experiment, isomers with a halflife of less than 500 ms could not be measured due to the excitation time of 900 ms in the precision trap and a total cycle time of ∼1.5 s.The FWHM of the time-of-flight resonance was about 1.2 Hz, or about 100 keV in this mass region.Thus, isomers with low excitation energies could not be resolved, again depending on the relative production rates.
Due to half-life and excitation energy limitations, only the isomeric states of 80 Y, 85 Zr, 85 Nb, and possibly 86 Nb and 88 Nb, are interesting for further studies.Larger frequency ranges were scanned to find a second resonance.For 80 Y m , 85 Zr m and 85 Nb m the excitation energies are known, but no second resonance could be found at the corresponding distance from the measured resonance frequency.For 86 Nb, no second resonance was observed within about 450 keV and for 88 Nb the limit was about 310 keV.
Table 2. Isomeric states as listed in the NUBASE compilation [16].Given are the excitation energies, the half-lives and the spins and parities (J π ) for the isomeric and ground states.The energy and half-life of 85 Nb m are from ref. [6].The appearance of a 759.0(10) keV isomer with T 1/2 =12 s in ref. [16] is based on a misinterpretation of the result of ref. [32] where a β-coincident γ ray with an energy of 759 keV was reported.
Since in all of these cases only one state was observed, it is not possible to determine "a priori" whether the measured nucleus was the ground state or the excited state.In the case of 85 Nb, the measured mass would agree better with the tabulated value of the excited state than with the ground state (see ref. [16]).This is also the case for the other Nb isotopes.For Y and Zr isotopes, however, the measured mass is closer to the ground state.As only primary ions are mass-separated at IGISOL, the observed yields represent the isomeric and ground states as well.In general, heavy-ion reactions tend to favor the population of higher spin states.Therefore, they favorably populate ground states in all studied cases except for 87 Nb where the energy is very small anyway (see table 2).For example, the isomers 80 Y m , 85 Zr m and 85 Nb m are less produced than the ground states at HIGISOL.The isomeric to ground state ratios for these isotopes at HIGISOL are about 1:4 for 80 Y [5,17], 1:40 for 85 Zr [6] and 1:2 for 85 Nb [6].

Y isotopes
The results of the Y isotopes are presented in fig.6 in comparison with the earlier results and the AME values.Most of the masses are based on measurements of end-point energies of β + spectra.The corresponding mass excesses have been calculated from the Q EC values with the tabulated mass excesses of the daughter nuclides ( 80,81,82,83 Sr) [14].Our measured mass excesses agree with the AME value only for the isotope 80 Y for which a time-of-flight measurement by use of a cyclotron has been performed.

80 Y
Partly due to astrophysical motivation, the mass of 80 Y has been measured several times.The early beta-decay experiments yielded Q EC values of 6952(152) keV [18], Table 1.Overview of the measured mass excesses (ME) compared to the tabulated values [14].The reference nuclides and number of measurements (N) are given for the measured isotopes.All masses are for the ground states (see sect.4.1).The mass excesses are based on the frequency ratio of the reference nucleus to the measured nucleus (ν ref /ν) given in the table.The error for the difference AME-JYFL is dominated by the uncertainties in the AME mass excesses.[25] and to the TOF experiments at SARA by Issmer et al. [21] and at CSS2 by Chartier et al. [24].The QEC value of 80 Y in [22] is only a lower limit which is indicated by an arrow in the figure .6934(242) keV [19] and 6200(600) keV [20].Although these values were consistent, the calculated mass excesses were more than 2 MeV too low compared to the AME value based on the systematics of neighboring isotopic and isotonic chains.A direct mass measurement of 80 Y using the second cyclotron of SARA of the ISN as a time-offlight spectrometer gave a Q EC value of 9120(170) keV [21] which proved that the earlier Q EC values were too low.Due to this discrepancy, the results from [18], [19] and [20] have been left out from fig. 6.The Q EC value has also been recently measured via beta-gamma coincidence spectroscopy and a lower limit for the Q EC value was measured to be 8929(83) keV [22] which corresponds a mass excess larger than −61379(84) keV.The disagreement with the earlier β + end-point measurements is suggested to arise from the assumptions about β + feeding in the 80 Y decay scheme.The β + spectrum was earlier measured in coincidence with the two lowest-lying γ transitions from the levels at 386 keV and 981 keV [18,19] but in [22], the gating γ transition was from the highest fed level observed at 3284 keV.Thus, the β + decay of 80 Y should primarily feed a level or a set of levels above ≈3 MeV and not the low-lying levels.
There have also been two time-of-flight measurements using the second cyclotron of GANIL, CSS2, as a mass spectrometer of high resolution [23,24].Due to a number of experimental difficulties in [23], the measured mass excess of −62097(80) keV differs from the other existing experiments and is not included in fig.6.The mass was re-measured at CSS2 resulting a new mass excess of −60971(180) keV [24] which agrees well with the values of −61188(171) keV [21], the adopted AME value of −61220(180) keV and −61144 (7) keV from this work.Our measured mass corresponds to the ground state as the well-known isomer at 228.5 keV in 80 Y is much less produced than the ground state and no second resonance was found in a wider frequency range (see sect.4.1).The other isomeric state at 312.5 keV in 80 Y (4.7 µs) was too shortlived for JYFLTRAP.

81 Y
Our measured mass excess of −65709 (6) keV does not agree with the AME value which is based on beta-decay experiments of refs.[18] and [19].In ref. [18], a Q EC value of 5408(86) keV was determined from coincidences with 124 keV and 408 keV γ-rays in 81 Sr and the deduced mass excess is −66120(87) keV.In ref. [19], the mass and the 124 keV γ transition measured with a plastic scintillator were used to gate the β + spectrum, which resulted in a Q EC value of 5620(89) keV and a mass excess of −65908(90) keV.Although the adopted Q EC value of [19] disagrees with our value, the Q EC value with the 124 keV γ-ray gate in a Ge detector (Q EC =5704(128) keV [19]) yields a mass excess of −65824(129) keV in agreement with our value.As a summary, the differences in the mass excesses may be explained by the uncertainties in the determination of β + end-point energies.

82 Y
The isomer at 402.63 keV (268 ns) in 82 Y was too shortlived for JYFLTRAP and was not observed.The measured mass excess for the ground state, −68060 (6) keV, disagrees slightly with the AME value, −68190(100) keV.In ref. [18], a 574 keV γ transition in 82 Sr was used to gate the β + spectrum and the measured Q EC value was 7868(185) keV.The calculated mass excess of −68140(186) keV agrees with our value.However, the Q EC value of 7793(123) keV [19] with a mass excess of −68215(124) keV disagrees slightly with our value.Similar to 81 Y, the γ-transition gate in the Ge detector gives a higher Q EC value of 7820(162) keV [19] and a mass excess of −68188(163) keV in agreement with our value.

83 Y
The measured mass excess of 83 Y, −72170( 6) keV, deviates from the AME value, −72330(40) keV. 83Y has an isomeric state at 61.98 keV (2.85 min) which could not be resolved in the JYFLTRAP.The Q EC value of 4571(85) keV for the isomeric beta decay was derived from the β + spectrum gated by a 422 keV transition in 83 Sr in [18].It yielded a mass excess of −72286(86) keV for the ground state.Q EC values of 4260(80) keV [25] and 4411(50) keV [19] based on β + spectra in coincidence with a 36 keV γ transition in 83 Sr correspond to the ground-state mass excesses of −72535(81) keV and −72384(51) keV, respectively.In ref. [19], a Q EC value of 4644(84) keV for the isomeric beta decay was also measured.The mass excesses for the isomeric state, −72224(86) keV [18] and −72151(85) keV [19], agree with the measured JYFLTRAP value but the ground state values from all beta-decay experiments deviate from it.On the other hand, the errors in the endpoint energies are so large that it cannot be concluded that we had measured the isomeric state.In fact, the ground state (9/2 + ) should be more populated in heavyion fusion-evaporation reactions than the isomeric state (3/2 − ) which suggests that we have measured the ground state mass.

Zr isotopes
The measured Zr masses are shown in fig.7 and discussed below.

83 Zr
The isomeric state at 52.72 keV (530 ns) in 83 Zr is too short-lived to be observed at JYFLTRAP.The measured ground-state mass excess of −65908 (7) keV was 552 keV higher than the AME value of −66460(100) keV.The AME value is based on the end point of a β + spectrum in coincidence with 56 keV, 255 keV, 304 keV and 360 keV γ transitions in 83 Y [19].The resulting Q EC value for the beta decay to the 83 Y m (2.85 min) state was 5806(85) keV which converts to a mass excess of −66464(94) keV when the AME value for 83 Y m is applied.The Q EC value of [19] was already commented to be too low in [26].Namely, in ref. [26] the end-point energy of the beta-delayed proton spectrum from 83 Zr was determined.The measured Q EC − B p was 2750(100) keV corresponding to a mass excess of −65970(115) keV where the proton binding energy and the mass excess of 83 Y have been taken from the AME tables.With the mass of 83 Y measured in this work, this results in a mass excess of −65810(108) keV, still in agreement with our value.
The Q EC − B p value calculated from the Q EC value of [19] is only 2196(90) keV which is clearly too low for the beta-delayed proton spectrum presented in [26].Probable isomerism and complex decays of the nuclides were suggested to explain this discrepancy in [26].Further support for the work by Hagberg et al. [26] comes from this experiment.The mass excess measured at JYFLTRAP agrees with [26] and not with [19].In addition, the Q EC value determined from the mass excesses of 83 Y and 83 Zr in this experiment is 6263(9) keV which is much higher than Q EC =5868(85) keV from ref. [19].

84 Zr
The mass of 84 Zr was measured for the first time.The obtained mass excess agrees well with the AME value based on the systematics.As 84 Zr is situated in the rp process region, its mass is useful, e.g., for the estimation of proton capture reaction rates on 83 Y and 84 Zr.The beta decay Q EC value for 84 Zr can now be calculated from the AME mass excess of 84 Y, Q EC =2740(90) keV.The proton separation energy for 84 Zr can be determined as well from the experimental masses of 83 Y and 84 Zr for the first time, S p = 6536(9) keV.

85 Zr
The AME value for the mass excess of 85 Zr is −73150(100) keV.It is based on the end point of the β + spectrum in coincidence with 416 keV and 453 keV γ transitions in 85 Y [19].The mass excess obtained at JYFLTRAP, −73170 (6) keV, agrees nicely with the AME value.The measured mass corresponds to the ground state as it is about 40 times more produced at HIGISOL than the isomeric state at 292.2 keV [6] and no second resonance was observed when scanning a wider frequency range.

86 Zr
The measured mass excess of 86 Zr, −77958 (7) keV, differs significantly from the AME value, −77800 (30) keV.The AME value is based on a Q value of −40136 (30) keV for the reaction 90 Zr(α, 8 He) 86 Zr [27].The difference can be explained with low statistics in the momentum (position) spectrum in [27].Namely, there were only 7 counts in 80 channels (with a maximum of 2 counts/10 channels) in the momentum spectrum.A peak centroid with an accuracy of ±8.1 channels (±26 keV) was fitted on this spectrum.The estimated total error of the reaction Q value was given as 30 keV which took into account the errors caused by the uncertainties in the incident energy (±14 keV), in the target thickness (± 5 keV) and in the masses of 90 Zr (± 2 keV) and α-particle (less than 1 keV) [27].A more realistic error estimation with an error for the peak centroid of about ±40 channels summed with the systematic errors would give roughly an error of ±150 keV for the reaction Q value which would be in agreement with the JYFLTRAP value.

88 Zr
The measured mass excess of −83624 (7) keV agrees very well with the AME value: the difference is only 1 keV.The AME value is based on a Q value of −12805 (10) keV for the reaction 90 Zr(p,t) 88 Zr [28].

Nb isotopes
Figure 8 shows the measured Nb isotopes together with the AME values and previous mass measurements.The large deviations from the AME mass excesses in 85 Nb and 86 Nb might be explained if the masses of isomeric states instead of ground states had been measured, as discussed in sect.4.1.Only the mass of 88 Nb agrees with the AME value.The mass measurements of 85,86,88 Zr make it possible to determine Q EC values for 85,86,88 Nb.With the mass of 84 Zr, a proton separation energy for 85 Nb can be experimentally determined for the first time.The measured Zr masses have been taken into account in the derivation of the Nb mass excesses from beta-decay experiments for these nuclides.

85 Nb
The measured mass excess, −66273 (7) keV, is 877 keV higher than the AME value, −67150(220) keV.The AME value is based on the end point of a β + spectrum in coincidence with a 50 keV γ transition in 85 Zr [29].The Q EC value of 6100(200) keV [29] yields a mass excess of −67050(224) keV with the AME value for 85 Zr and a mass excess of −67070(224) keV with the mass excess of 85 Zr measured in this work, still about 800 keV off from our value.
The AME value for the mass excess of 85 Nb is based on the Q EC value given incorrectly as Q EC =6000(200) keV in the abstract of [29].The Q EC value determined from the mass excesses of 85 Nb and 85 Zr measured in this experiment is 6898(9) keV, about 900 keV from the adopted AME value.Also the proton separation energy determined from the measured masses of 84 Zr and 85 Nb, S p = 2144 (9) keV differed about 800 keV from the adopted AME value.
The differences in mass excesses and Q EC values can have several sources.Firstly, the energy calibration of the  [19] and Hagberg et al. [26] and to reaction studies done by Kato et al. [27] and Ball et al. [28].The 83 Zr value marked with an asterisk has been calculated with the mass excess of the daughter nucleus 83 Y measured in this experiment at the JYFLTRAP.The other mass excesses have been calculated with the respective AME values.Warburton et al. [34] and Oxorn and Mark [37].The values of 85,85,88 Nb marked with an asterisk have been calculated with the mass excesses of the corresponding daughter nuclei 85,86,88 Zr measured in this experiment at the JYFLTRAP.The other mass excesses have been calculated with AME values for daughter nuclides.beta detector in [29] was done with the Q EC values of 83 Y (4.411 MeV), 83 Zr (5.806 MeV) and 82 Y (7.793 MeV) which were found to be about 214 keV, 395 keV and 155 keV too low in this experiment.
Another explanation for the discrepancy in the masses could be that the beta decay of 85 Nb feeds mainly higherlying levels which then decay to the first excited state at 50 keV in 85 Zr.The high-spin studies of 85 Zr have revealed a 11/2 + state at 854 keV and a 13/2 + state at 872 keV [30,31].These states decay to the state at 50 keV by 804 keV and 822 keV γ transitions, respectively.
The third possibility is that the mass excess measured in this work is for an isomeric state in 85 Nb instead of the ground state.In a spectroscopic experiment [32], a 759 keV γ transition with a half-life of 12(5) s was observed at mass A = 85 and it was suggested to originate either from a β-decaying isomer in 85 Nb or in 85 Mo.In addition, a 69-keV transition with a half-life of 3.3 (9) s was observed at mass A=85 in [6].This transition belongs most likely to an isomeric state in 85 Nb.The origin of the 69 keV transition from β + decay of 85 Mo was unlikely as there was no coincidence with the annihilation radiation.As a summary, the energy of the isomeric state in 85 Nb is uncertain.The energy range covered in the JYFLTRAP was about ±900 keV from the resonance and no indication of an isomer was found.Therefore, the isomeric state suggested by the systematics of odd-A Nb isotopes (see ref. [6]) should lie above 900 keV, its half-life is below 500 ms or it is much less produced than the ground state.In ref. [6], the production ratio of isomeric state to the ground state was roughly 1:2 which supports the last option.Most likely the isomeric state was not produced enough for the observation in the trap.

86 Nb
The measured mass excess of 86 Nb, −69129 (6) keV, differs −701 keV from the AME value of −69830(90) keV.The AME value is based on [33] which measured the end point of a β + spectrum in coincidence with 752 keV, 914 keV and 1003 keV γ transitions in 86 Zr.The adopted Q EC value was 7978(80) keV which converts to a mass excess of −69822(86) keV with the AME mass excess for 86 Zr.With the JYFLTRAP mass excess of 86 Zr, the resulting mass excess is −69980(81) keV which is still far away from the measured mass excess of 86 Nb.The Q EC value determined from the mass excesses of 86 Zr and 86 Nb measured in this work is 8829(9) keV.
The Q EC value of [33] was claimed to be wrong already in ref. [34].Namely, an erroneous 100 % beta feeding was assumed to the 2670 keV level in 86 Zr in [33].In ref. [34] almost equally strong beta feedings to states at 3254 keV and 3418 keV were observed and the Q EC value was calculated from the weighted average for the final 86 Zr states.By using 752 keV, 916 keV and 1003 keV γ-transition gates for the β + spectrum, a Q EC value of 8150(200) keV was concluded in [34].The deduced mass excesses of 86 Nb, −69650(203) keV (from the AME mass excess of 86 Zr) and −69808(201) keV (from the mass excess of 86 Zr measured in this work), are still too low compared to our value.Thus, the discrepancy cannot solely be explained by the neglected beta-decay feedings to higher states than 2670 keV unless possible beta-decay feedings to even higher states of 86 Zr have been missed in [34].
An isomer with a half-life of 56 s in 86 Nb was suggested in ref. [35] based on the time behavior of Zr KX rays in coincidence with the sum of the 752 keV, 915 keV and 1003 keV γ transitions in 86 Zr.This isomeric state with an unknown energy was not confirmed in [6].In ref. [36], the existence of this isomer was considered as uncertain as the result has not been confirmed in later experiments.It is also surprising that the ground state assignment of 86 Nb is not experimentally verified, and thus, the lowest observed state with a half-life of 88 s (6 + ) is considered as an isomer at an energy of 0 + X keV [36].In this experiment, we could not find an isomeric state within 450 keV from the resonance in the JYFLTRAP.This means that the suggested isomer is located above 450 keV or is much less produced than the ground state.

87 Nb
The measured mass excess of 87 Nb, −73868 (7) keV, differs significantly from the AME value, −74180(60) keV.This discrepancy cannot be explained by the known isomeric state as it is located at 4 keV which falls within the error bars of the determined resonance.The AME mass excess is based on the end point of a β + spectrum in coincidence with 201 keV and 617 keV γ transitions of 87 Zr [33].The Q EC value of 5169(60) keV [33] is 311 keV lower than the Q EC determined from the masses of 87 Zr [14] and 87 Nb from this work, Q EC =5480 (11) keV.Either the energy calibration of the β + detector in [33] is off or the beta decay of 87 Nb to higher-lying states has been underestimated.Note that the mass excess of 86 Nb derived from [33] was also too low compared to this work.

88 Nb
The measured mass excess of 88 Nb agrees with the AME value.The AME mass excess is based on a measurement of the end point of a β + spectrum in coincidence with a 503 keV transition in 88 Zr [37].The Q EC value of 7550(100) keV [37] is in agreement with the value based on 88 Zr and 88 Nb masses measured in this work, Q EC =7476(9) keV.The isomer suggested to lie at 40(140) keV (7.8 min) was not observed within 310 keV from the resonance.Either the production rate of the isomer was too low or its energy is so close to the ground state that it could not be resolved.

Conclusions
With the measured absolute masses, the accuracies of proton separation energies and Q EC values for 12 neutrondeficient nuclides have been improved substantially (see Fig. 9. Proton separation energies for the measured nuclides relative to AME values.The two lines show the corridor of uncertainties taken from ref. [14].The mass of the (Z-1,N) nucleus is from this experiment for 83 Zr, 84 Zr, 85 Nb, 86 Nb and 87 Nb.For the others, the AME mass [14] of the (Z-1,N) nucleus has been used.Table 3. Proton separation energies for the measured nuclides.The mass of the (Z-1,N) nucleus is from this experiment for 83 Zr, 84 Zr, 85 Nb, 86 Nb and 87 Nb (marked with †).For the others, the AME mass of the (Z-1,N) nucleus is used [14].Proton separation energies for 84 Zr and 85 Nb have been experimentally determined for the first time.The last column lists the difference between the determined S p value and the corresponding AME value [14].
isotopes will have to be revised significantly which could even impact the estimated position of the proton dripline.These data will provide important contribution to more reliable predictions of binding energies at main rp-process nuclei.

Fig. 1 .Fig. 2 .
Fig. 1.The rp process path in the region close to A = 80 for steady state burning [3].The reaction flows of more than 10 % are shown by a solid line and of 1-10 % by a dashed line.The studied nuclides are highlighted by squares.Waiting-point nuclei 80 Zr and 84 Mo are labeled as WP.

Fig. 3 .
Fig. 3. Time-of-flight resonance of 80 Y with an excitation time of 900 ms.The fit of the experimental data points is shown by a solid line.

Fig. 4 .
Fig. 4. Time-of-flight resonance of 84 Zr with an excitation time of 900 ms.The fit of the experimental data points is shown by a solid line.

Fig. 5 .
Fig. 5. Time-of-flight resonance of a reference nucleus 98 Mo with an excitation time of 900 ms.The fit of the experimental data points is shown by a solid line.

Fig. 6 .
Fig. 6.Mass excesses of the measured Y isotopes relative to AME values [14].The error bars represent only the uncertainties of the corresponding experiments.The labels refer to beta-decay experiments of Barton et al. [22], Lister et al. [18], Della Negra et al. [19], Deprun et al.[25] and to the TOF experiments at SARA by Issmer et al.[21] and at CSS2 by Chartier et al.[24].The QEC value of 80 Y in[22] is only a lower limit which is indicated by an arrow in the figure.

Fig. 7 .
Fig. 7. Mass excesses of the measured Zr isotopes relative to AME values [14].The error bars represent only the uncertainties of the corresponding experiments.The labels refer to beta-decay experiments of Della Negra et al. [19] and Hagberg et al.[26] and to reaction studies done by Kato et al.[27] and Ball et al.[28].The 83 Zr value marked with an asterisk has been calculated with the mass excess of the daughter nucleus 83 Y measured in this experiment at the JYFLTRAP.The other mass excesses have been calculated with the respective AME values.

Fig. 8 .
Fig. 8. Mass excesses of the measured Nb isotopes relative to AME values [14].The error bars represent only the uncertainties of the corresponding experiments.The labels refer to beta-decay experiments of Kuroyanagi et al. [29], Della Negra et al. [33],Warburton et al.[34] and Oxorn and Mark[37].The values of 85,85,88 Nb marked with an asterisk have been calculated with the mass excesses of the corresponding daughter nuclei85,86,88 Zr measured in this experiment at the JYFLTRAP.The other mass excesses have been calculated with AME values for daughter nuclides.
This work has been supported by the EU 6th Framework program "Integrating Infrastructure Initiative -Transnational Access", Contract Number: 506065 (EURONS) and by the Academy of Finland under the Finnish Center of Excellence Program 2000-2005 (Project No. 44875, Nuclear and Condensed Matter Physics Program at JYFL).The support via the Finnish-Russian interacademy agreement (Project No. 8) is gratefully acknowledged.The authors thank H. Schatz for his valuable comments on the manuscript of this article.