WorksPhysics and Astronomy
Dark Matter Constraints from Isomeric Hf
Abstract
We describe a first measurement of the radiation from a Hf sample to search for dark matter. The flux from this sample, possessed by Los Alamos National Laboratory nuclear chemistry, was measured with a Ge detector at a distance of 4 ft due to its high activity. We search for s that cannot arise from the radioactive decay of Hf, but might arise from the production of a nuclear state due to the inelastic scattering with dark matter. The limits obtained on this flux are then translated into constraints on the parameter space of inelastic dark matter. Finally, we describe the potential reach of future studies with Hf.
There is irrefutable evidence for the existence of dark matter arising purely from gravitational interactions. Understanding its particle nature is one of the burning questions of 21{sup}st century particle physics. Dark matter candidates at the weak scale arise naturally in theories beyond the Standard Model, such as supersymmetry. Furthermore, weak scale massive particles with weak scale cross-sections—the so-called “weakly interacting massive particles,” or WIMPs—are produced with the correct relic abundance when freezing out from thermal equilibrium in the early universe (WIMP miracle). Direct, indirect, and collider searches for WIMPs have reported repeated null results, setting stringent limits on their model space. A review of the present status of the search for dark matter can be found in {citep}[1]`. The derived limits are very restrictive and the lack of an observation has motivated dark matter considerations beyond the classic WIMP description.
Among other alternative dark matter models are inelastic dark matter (iDM) [2, 3, 4, 5] and strongly interacting dark matter (SIDM) [6, 7, 8]. While these models retain the salient features of thermal freeze-out, constraints on their parameter spaces are much less stringent because the threshold energy required for a detectable dark matter-nucleus scattering event is often unavailable—in the case of iDM, due to the large inelastic splitting; and in the case of SIDM, due to the loss of dark matter kinetic energy from its interactions with the overburden rock above deep underground experiments. More specifically, in iDM models, dark matter-nucleus elastic scattering is suppressed, and the dominant scattering process requires an internal transition to an excited dark matter state. If the transition energy is greater than the available kinetic energy of the dark matter-nucleus system, this process is completely shut-off, severely reducing the sensitivity to the iDM parameter space of present experiments focused on WIMPs. In the case of SIDM models, the large dark matter nuclear cross section causes dark matter particles to thermalize through interactions with the Earth, resulting in a velocity too low to produce a measurable interaction by the time they reach the detector’s location deep underground. In particular, large scattering
cross sections (much higher than electroweak-strength) are still viable for significant swaths of parameter space of both SIDM and iDM models. While this is difficult to obtain with perturbative models of new physics at the TeV scale, composite dark matter models from a strongly interacting dark sector can naturally accommodate these properties [9, 10, 11, 12, 13].
In this letter, we consider a recent proposal to use nuclear metastable states as exothermic-reaction targets for dark matter searches [14]. The isomer was suggested as a potential target and its large reservoir of available energy for transference to a dark matter particle makes it an intriguing candidate. Unfortunately, due to the relatively short half-life (31 y) of this manmade isotope, a large number of target atoms also means significant radioactivity. In this article, we describe a first study of the spectrum of to derive first limits on dark matter interactions with this isotope.
The sample used in this study was fabricated by the Los Alamos National Laboratory (LANL) nuclear chemistry division to produce for a / medical generator [15]. Later, this material was used to study the possibility of energy storage in nuclear isomers, specifically . Reports of triggered isomer decay in led to a further study [16] finding no evidence for the effect. The Hafnium sample used for that experiment, and for this measurement, was extracted from a Ta target at the LAMPF accelerator at LANL [17]. The sample studied here is the second set as described in [16] and is shown in Figure 1. It is now relatively old and Hf isotopes other than have decayed away, leaving a rather pure sample.

Figure 1. A photo of the source.
Due to the sample’s high activity (20-25 mR/hr on contact), the detector was placed at a distance of 1.2 m (4 ft) in order to minimize dead time from pile-up rejection. The detector was an ORTEC® Detective-X Ge detector [18]. The spectrum, shown in Figure 2, indicates some natural room background lines from the U/Th/K decay chains along with the known lines. A list of the identified lines are given in Table I. Note that above 600 keV, the spectrum is dominated by the natural background. Below 600 keV, the spectrum is dominated by emission.
Figure 2. The observed spectrum from the sample for a live time of 974.2 s. The inset shows the spectrum surrounding 1330 keV. A dark matter-induced with this energy proved to be the most sensitive test.
| [keV] | Origin | [keV] | Origin |
| 55.3 | x-rays | 725.9 | |
| 62.7 | x-ray | 768.4 | |
| 68.6 | Au x-ray det. mat. | 794.9 | |
| 77.9 | Au x-ray det. mat. | 860.6 | |
| 88.6 | 910.5 | ||
| 93.2 | 968.2 | ||
| 213.4 | 1093.5 | sum | |
| 216.7 | 1120.6 | ||
| 237.4 | 1238.4 | ||
| 257.6 | 1242.0 | 1 | |
| 277.3 | 1377.7 | ||
| 296.8 | 1460.2 | ||
| 325.6 | 1729.6 | ||
| 426.4 | 1764.8 | ||
| 454.0 | 1847.4 | ||
| 495.0 | 2103.6 | esc. peak | |
| 511.0 | annihilation | 2117.8 | |
| 535.0 | 2204.1 | ||
| 574.2 | 2448.4 | ||
| 583.5 | 2614.5 | ||
| 608.9 |
Table 1. The stronger lines observed in the spectrum.
The metastable state at 2446 keV, with a half-life of 31 yr, has a large spin . Table II lists a number of Hf states which are prevented from being populated by transitions originating from the metastable state due to the large spin change, . An interaction with a heavy and slow dark matter particle, on the other hand, could overcome the hindrance and catalyze transitions from the metastable state to those (otherwise unpopulated) lower spin states, which in turn could decay via emission. Figure 3 depicts the level diagram showing the radioactive decay pathways in contrast to the dark matter-induced pathways for a couple of the key transitions.

Figure 3. The level diagram of . To the left (and in black) are the transitions that occur from the SM radioactive decay of the isotope. To the right (and in blue), we illustrate two dark matter-induced transitions of significance: (i) the half-life of the transition (associated with the 1679 keV ) was the most strongly constrained by this work; (ii) the transition (associated with the 1330 keV ) provided the strongest constraint on the dark matter-nucleon cross section, , among the 11 s considered. This figure was adapted from Fig. 1 of [17].
When a dark matter particle scatters off the isomeric nuclear state, the metastable nuclear energy can be tapped to excite the dark matter to an elevated internal energy state. The primary advantage of in our study is its large metastable state energy, which allows us to probe, for the first time in a direct detection experiment, dark matter mass splittings as high as . In particular, dark matter-induced transitions of the isomeric state to the lower energy states would be sensitive to the largest dark matter mass splittings. Unfortunately, most of the low-lying states are also populated through the SM radioactive decay of , and therefore suffer from large backgrounds and result in a significant reduction in sensitivity to an iDM signal. Therefore, we focus on dark matter-induced transitions to higher energy states which are not populated by the SM decay pathways and therefore emit s in regions with reduced backgrounds; these s have energies above keV. In Table II, we only list the excited states and associated decay s determined to be the most sensitive given these constraints.2 These states balance the need for high dark matter-induced transition energy against the backgrounds at the energy.
For the candidate s listed in Table II, we obtain the number of observed and expected background counts within an optimized window centered on the energy. The expected background count (B) is obtained by performing a side-band fit of the data, assuming a linear background distribution. The width of the optimal window is deduced by maximizing , where is the signal acceptance within the chosen window, assuming it follows a gaussian distribution. For a continuum spectrum, the resulting signal acceptance within the optimal window is given by ; however, since we collected a binned spectrum, the signal acceptance for each received a small correction to account for the bin edges. There were no observable peaks at any of the smoking-gun energies in Table II.
To establish our notation, consider the process in which a dark matter particle upscatters off the isomeric nuclear state , causing the nucleus to transition to a lower energy level ,
We will denote the inelastic cross section for this process by . The produced state can then de-excite via emission of a of energy with a branching ratio
| label | state energy (keV) | state , | energy (keV) | branching ratio | acceptance | rel. eff. | background counts | observed counts | ( yrs) |
| 1 | 1635.6 | , 0 | 1542.2 | 0.83 | 0.41 | 17 | |||
| 2 | 1636.7 | , 5 | 1330.0 | 0.86 | 0.44 | 32 | |||
| 3 | 1640.5 | , 4 | 1333.8 | 0.86 | 0.44 | 40 | |||
| 4 | 1648.8 | , 2 | 1016.6 | 0.84 | 0.47 | 82 | |||
| 5 | 1651.5 | , 1 | 1344.9 | 0.86 | 0.44 | 41 | |||
| 6 | 1654.3 | , 0 | 1348.0 | 0.86 | 0.44 | 34 | |||
| 7 | 1691.1 | , 2 | 1059.0 | 0.83 | 0.47 | 50 | |||
| 8 | 1697.5 | , 8 | 333.4 | 0.83 | 1.24 | 6506 | |||
| 9 | 1747.1 | , 2 | 1440.6 | 0.84 | 0.44 | 31 | |||
| 10 | 1772.1 | , 0 | 1678.8 | 0.81 | 0.25 | 11 | |||
| 11 | 1788.6 | , 4 | 1156.3 | 0.81 | 0.47 | 67 |
Table 2. The input parameters and resulting 90% C.L. limits on the half-life of the dark matter-induced transition shown in eq. (1). The smoking gun signal for this process is the emitted in the decay of dark matter induced state (4th column). The relative detector efficiencies normalized to that for the 495-keV , , were obtained from [18] and are assumed to carry uncorrelated uncertainties of . The isomeric state of at 2446.1 keV has and .
given by . The expected signal count for the process described above, , is given by
where is the number of target nuclei; is the live time; is the dark matter flux; and and are, respectively, the signal acceptance within the region of interest and detection efficiency for the emitted in the decay of .
The number of target atoms can be deduced from the SM activity of the Hf sample. In particular, the SM decay chain of produces a 495-keV line with a probability [20]. The number of 495-keV counts observed during our live time, , relates to via:
where is the lifetime of and is the detection efficiency for the 495 keV line.
Combining (2) and (3), we can express the dark matter event rate (i.e., number of scattering events in (1) per unit time per target nucleus) as
The corresponding half-life for this dark matter-induced transition is simply related to (4) via
By performing a profiled log-likelihood fit of the signal strength for each of the 11 lines considered, we obtained 90% confidence level (C.L.) limits on the half-lives , given in Table II.
The inelastic cross section for the process in (1), , can be related to the model-dependent dark matter-nucleon cross section using the formalism of [14]. We used this relation to translate our limits into constraints on the parameter space of inelastic dark matter (iDM), namely, versus the iDM mass splitting for a benchmark iDM mass of TeV. Note that the transition with the most strongly constrained half-life does not necessarily provide the strongest constraint on . That is because the nuclear form factor, which suppresses the transition rate in (1), depends not only on the momentum transfer and change in angular momentum , but also on -selection rules. Specifically, each nuclear state has a -quantum number given by the projection of its angular momentum on its symmetry axis (see Table II), and transitions with greater than the multipolarity of the emitted radiation suffer from
an additional suppression, the so-called “-hindrance”. Among the 11 transitions considered, provides the strongest constraint on , since it has the second smallest , and the observed counts for its associated line of 1330 keV showed a deficit relative to the background expectation. We can contrast the constraining power of with that of other transitions. For example, while has the smallest , its associated 333.4 keV line lies in a region with substantial backgrounds ( orders of magnitude larger than the backgrounds for the other lines), which significantly weakens its dark matter constraining power. As another example, transitions and have the most strongly constrained half-lives, but also the largest , which suppresses their rate for keV. For further technical details, we refer the reader to [14].
Finally, we performed a joint log-likelihood fit of the signal strength for all the 11 s combined. Our results are shown in Figure 4.

Figure 4. The 90% C.L. exclusion limits on the parameter space of inelastic dark matter, assuming a standard halo model with local dark matter density and galactic escape velocity km/s. The black curve shows the combined limit from all 11 s, and the shaded gray regions show previous existing limits [21, 22, 23, 24, 25, 26, 27]. The 11 colored curves show the limits from each individual line in Table II. Following the -label convention of Table II, and fixing keV, the order of the 11 colored curves, from stronger to weaker exclusion, is: .
For keV, our experimental limit on is the best to date, albeit dark matter models with such large cross sections would necessarily come from composite dynamics and might face model building challenges. The strongest competing constraint in the large mass splitting region keV comes from searches using the tantalum metastable state , which has a very long half-life and is naturally abundant albeit with a low isotopic fraction. This has enabled experiments with samples containing a large number of nuclei, resulting in interesting limits on SIDM and iDM models [22, 28, 29], as well as ongoing experiments with significantly improved reach [30]. Still, the comparably lower metastable energy of (76.3 keV, in contrast to 2446 keV for ) limits its sensitivity to keV. Other existing experimental results are described in [21], with specific limits derived from data for PbWO [23], CaWO [24], CRESST-II [25], PICO-60 [26], and XENONnT [27].
A number of improvements and options for future measurements are possible. For a repeat of the measurements presented here, a longer run time and the use of a shielded detector could improve sensitivity by a factor of . A high-efficiency Ge-detector array similar to AGATA [31]—with a large solid angle acceptance and detectors distant enough from the source so as to have a manageable rate—could further improve sensitivity by an additional factor of .
The Hf measurements reported here were performed at a surface site, and therefore were not sensitive to viable parameter space in SIDM. By deploying the Hf sample and a detector underground and repeating these measurements, perhaps at multiple depths, one could probe the effects of a dark matter traffic jam [14] in models of SIDM (both elastic and inelastic).
Ideally, one would like a very large sample of the Hf isomer. For this Hf sample, 1 kg of Ta was irradiated for 60 days with an 800 MeV, 350 A proton beam. The Ta target was used as a dedicated beam stop from which the Hf was extracted by radiochemistry [15]. If feasible, processing additional targets could produce a large quantity of ; however, the cost of the required radiochemistry would have to be weighed against the science reach. Furthermore, practical difficulties associated with the high radioactivity of the sample would need to be overcome, such as a fast, highly efficient detection system in order to fully exploit the science scope of such measurements.
An alternative experimental setup could be arranged to search for the decay of the upscattered dark matter particle. Specifically, dark matter particles could inelastically scatter off to an excited state, and decay promptly or with displacements via emission of a monochromatic , which might be observed in a nearby detector (see, e.g., [32]). For this geometry, a large sample of (such as beam stops that have been irradiated for extended periods) could be located some distance from a detector. A Ge detector could be sited nearby to search for such an anomalous . While number of old tungsten and tantalum-clad tungsten beam stops are stored on location behind shielding but our early assessment is that none have enough for a useful measurement. A dedicated production would be required.
Acknowledgements
We thank Evelyn Bond and Athena Marie Marenco for assisting with access to the sample. We gratefully acknowledge support from the U.S. Department of Energy Office of Science, from the Los Alamos National Laboratory’s Directed Research and Development (LDRD) Program for this work.
References
References
- [1]A. Boveia, T. Y. Chen, C. Doglioni, A. Drlica-Wagner, S. Gori, W. H. Lippincott, M. E. Monzani, C. Prescod-Weinstein, B. Shakya, T. R. Slatyer, N. Toro, M. Williams, L. Winslow, P. Tanedo, Y.-T. Tsai, J. Yu, and T.-T. Yu, “Snowmass 2021 cross frontier report: Dark matter complementarity (extended version),” (2022).
- [2]D. Smith and N. Weiner, Phys. Rev. D 64, 043502 (2001).
- [3]B. Feldstein, P. W. Graham, and S. Rajendran, Phys. Rev. D 82, 075019 (2010).
- [4]M. Pospelov, N. Weiner, and I. Yavin, Phys. Rev. D 89, 055008 (2014).
- [5]J. Bramante, P. J. Fox, G. D. Kribs, and A. Martin, Phys. Rev. D 94, 115026 (2016).
- [6]D. Hooper and S. D. McDermott, Phys. Rev. D 97, 115006 (2018).
- [7]V. De Luca, A. Mitridate, M. Redi, J. Smirnov, and A. Strumia, Phys. Rev. D 97, 115024 (2018).
- [8]D. A. Neufeld, G. R. Farrar, and C. F. McKee, The Astrophysical Journal 866, 111 (2018).
- [9]S. Nussinov, Phys. Lett. B 165, 55 (1985).
- [10]J. Bagnasco, M. Dine, and S. D. Thomas, Phys. Lett. B 320, 99 (1994).
- [11]D. S. M. Alves, S. R. Behbahani, P. Schuster, and J. G. Wacker, Phys. Lett. B 692, 323 (2010).
- [12]M. Lisanti and J. G. Wacker, Phys. Rev. D 82, 055023 (2010).
- [13]D. Spier Moreira Alves, S. R. Behbahani, P. Schuster, and J. G. Wacker, JHEP 06, 113 (2010).
- [14]M. Pospelov, S. Rajendran, and H. Ramani, Phys. Rev. D 101, 055001 (2020).
- [15]W. Taylor et al., J. Radioanal. Nucl. Chem. 236, 155 (1998).
- [16]I. Ahmad, J. C. Banar, J. A. Becker, T. A. Bredeweg, J. R. Cooper, D. S. Gemmell, A. Kraemer, A. Mashayekhi, D. P. McNabb, G. G. Miller, E. F. Moore, P. Palmer, L. N. Pangault, R. S. Rundberg, J. P. Schiffer, S. D. Shastri, T. F. Wang, and J. B. Wilhelmy, Phys. Rev. C 71, 024311 (2005).
- [17]I. Ahmad, J. C. Banar, J. A. Becker, D. S. Gemmell, A. Kraemer, A. Mashayekhi, D. P. McNabb, G. G. Miller, E. F. Moore, L. N. Pangault, R. S. Rundberg, J. P. Schiffer, S. D. Shastri, T. F. Wang, and J. B. Wilhelmy, Phys. Rev. Lett. 87, 072503 (2001).
- [18]ORTEC, 801 South Illinois Ave., Oak Ridge, TN 37830 USA.
- [19]E. Achterberg, O. Capurro, and G. Marti, Nuclear Data Sheets 110, 1473 (2009).
- [20]M. B. Smith, P. M. Walker, G. C. Ball, J. J. Carroll, P. E. Garrett, G. Hackman, R. Propri, F. Sarazin, and H. C. Scraggs, Phys. Rev. C 68, 031302 (2003)
- [21]N. Song, S. Nagorny, and A. C. Vincent, Phys. Rev. D 104, 103032 (2021).
- [22]B. Lehnert, H. Ramani, M. Hult, G. Lutter, M. Pospelov, S. Rajendran, and K. Zuber, Phys. Rev. Lett. 124, 181802 (2020).
- [23]J. W. Beeman et al., Eur. Phys. J. A 49, 50 (2013).
- [24]A. Münster et al., Journal of Cosmology and Astroparticle Physics 2014, 018 (2014).
- [25]G. Angloher et al., Eur. Phys. J. C 76, 25 (2016).
- [26]C. Amole et al. (PICO Collaboration), Phys. Rev. D 93, 052014 (2016).
- [27]E. Aprile et al. (XENON Collaboration 7), Phys. Rev. Lett. 121, 111302 (2018).
- [28]B. Lehnert, M. Hult, G. Lutter, and K. Zuber, Phys. Rev. C 95, 044306 (2017).
- [29]R. Cerroni, S. Dell’Oro, A. Formicola, S. Ghislandi, L. Ioannuci, M. Laubenstein, B. Lehnert, S. S. Nagorny, S. Nisi, and L. Pagnanini, “Deep-underground search for the decay of 180m-ta with an ultra-low-background hpge detector,” (2023), arXiv:2305.17238 [nucl-ex].DOI
- [30]I. J. Arnquist, F. T. A. I. au2, A. S. Barabash, C. J. Barton, K. H. Bhimani, E. Blalock, B. Bos, M. Busch, M. Buuck, T. S. Caldwell, C. D. Christopherson, P. H. Chu, M. L. Clark, C. Cuesta, J. A. Detwiler, Y. Efre- menko, H. Ejiri, S. R. Elliott, G. Kiovanetti, J. Goett, M. P. Green, J. Gruszko, I. S. Guinn, V. E. Guiseppe, C. R. Haufe, R. Henning, D. H. Aguilar, E. W. Hoppe, A. Hostiuc, I. Kim, R. T. Kouzes, T. E. L. V. au2, A. Li, J. M. Lopez-Castano, R. Massarczyk, S. J. Meijer, W. Meijer, T. K. Oli, L. S. Paudel, W. Pettus, A. W. P. Poon, D. C. Radford, A. L. Reine, K. Rielage, A. Rouyer, N. W. Ruof, D. C. Schaper, S. J. Schleich, T. A. Smith-Gandy, W. Tedeschi, R. L. Varner, S. Vasilyev, S. L. Watkins, J. F. Wilkerson, C. Wiseman, W. Yu, C. H. Yu, D. S. M. Alves, and H. Ramani, “Constraints on the decay of 180mTa,” (2023), arXiv:2306.01965 [nucl-ex].DOI
- [31]A. Bracco, G. Duchêne, P. Del Zop, and P. Reiter, Prog. Part. Nucl. Phys. 121, 103887 (2021).
- [32]M. Pospelov, N. Weiner, and I. Yavin, Phys. Rev. D 89, 055008 (2014).
