Journal of Low Temperature Physics

, Volume 167, Issue 5, pp 991–1003

Neutrino Physics with Low-Temperature Detectors


    • Centre de Spectrométrie Nucléaire et de Spectrométrie de Masse

DOI: 10.1007/s10909-012-0576-9

Cite this article as:
Giuliani, A. J Low Temp Phys (2012) 167: 991. doi:10.1007/s10909-012-0576-9


In the last years, neutrino physics has provided exciting discoveries that for the first time have cracked the solid building of the Standard Model. However, many mysteries remain, and prospects are even more appealing. Low temperature detectors can give fundamental contributions to this field. They already play a major role in the study of neutrinoless double beta decay, a rare nuclear process that can ascertain if neutrino is a self-conjugate elementary fermion and fix its mass scale. Cuoricino, a project based on macro-bolometers, is the most sensitive double-beta-decay search in the world together with the much debated Heidelberg–Moscow experiment. CUORE, its natural continuation, is one of the most promising experiments under construction or commissioning, capable to start to attack the so-called inverted hierarchy region of the neutrino mass pattern. Several ideas based on low-temperature detectors (among which the simultaneous detection of phonon and scintillation light) are among the most promising approaches for next-generation experiments, capable to cover fully the inverted hierarchy region. In other sectors of neutrino physics, like the direct measurement of the neutrino mass in the MARE and ECHO projects or the detection of coherent neutrino-nucleus elastic scattering, low temperature detectors look less mature scientifically. However, they remain extremely promising devices to address these very challenging searches.


Neutrino massDouble beta decayMacrobolometersMicrocalorimeters

1 Neutrinos Today

The difficulty in detecting and studying neutrinos explains why much of what we know now on their properties [1] was totally obscure only twenty years ago. We have today overwhelming proofs that neutrino flavors oscillate. Oscillations can take place since neutrinos of definite flavor (νe, νμ, ντ) are not necessarily states of a definite mass (ν1, ν2, ν3). On the contrary, they are generally coherent superpositions of such states, to which they are connected by a unitary 3×3 mixing matrix U. As a consequence, the neutrino flavor is no longer a conserved quantity and the amplitude of the process νlνl is not vanishing. Due to the unitarity of U there is no flavor change if all the three masses vanish or are exactly degenerate. From oscillations, we can get information only on the mass square differences, but not on the masses themselves. We know that they are much smaller than charged lepton masses, but the mass pattern is unknown. Anyway, the discovery that neutrinos have a mass is a breakthrough by itself. This is the first clear signal of new physics beyond the Standard Model (SM), after 30 years of almost boring successes.

Two mass square differences have been determined reasonably well [1]: the 99% C.L. range are \(\varDelta m_{12}^{2} = (7.58 \pm0.21)\times10^{-5}~\mathrm{eV}^{2}\), |Δm23|2=(2.40±0.15)×10−3 eV2. The sign of \(\varDelta m_{13}^{2}\) is however unknown. If it is positive, one speaks of normal mass hierarchy (NI), with m1<m2<m3; if it is negative, of inverted mass hierarchy (IH), with m3<m1<m2. If m1m2m3, the pattern is quasi-degenerate. The absolute neutrino mass scale is not accessible to oscillation experiments. Three methods can address directly this important parameter: analysis of CMB temperature fluctuations [2], neutrinoless double beta [3] decay (0ν-ββ) and single β decay [4]. The quantities probed in these three approaches are however different, and given respectively by:
where Uei are the elements of the first row of U, and αi are the so-called Majorana phases, related to CP violation. The first method is observational, and performs a purely kinematical estimation of the neutrino masses. Even if very sensitive, it depends critically on cosmological and astrophysical assumptions and requires therefore independent checks. There is a large spread in the limits on mcosm (ranging in the interval 0.5–1.0 eV) according to different authors [2]. The second and third methods (on which this review is focused) are based on laboratory searches and will be discussed in the next sections.

It is important to stress that the parallel study of the three variants of the mass scale is a crucial task. The quantities in (1) depend on different combinations of the neutrino mass values and oscillation parameters. A complete neutrino physics program should comprise all these three observational/experimental approaches, which are not redundant but rather complementary. If, ideally, a positive measurement were reached in all of them one could test the results for consistency and with a bit of luck determine the CP violating Majorana phases.

We have described the baseline framework. Other possible scenarios, suggested by anomalies observed in oscillation experiments and unexplained controversial experimental data, envisage the existence of sterile neutrinos [5], hypothetical particles that interact via gravity, but however mix with the ordinary weakly interaction neutrinos. For example, a recent re-analysis of existing reactor data shows a 3σ deviation from theoretical predictions. A combined analysis using data from various sources leads to an additional mass splitting of \(\varDelta m_{\mathit{sterile}}^{2} > 1.5~\mathrm{eV}^{2}\) with a mixing parameter sin2(2θ)=0.17±0.08 at 95% C.L. This effect may well be due to subtle SM effects; however, the existence of at least one sterile neutrino remains a viable explanation. At much higher mass scales, 1 keV sterile neutrinos [7] have been suggested as attractive “warm” dark matter candidates.

Other peculiarities of neutrinos, beyond the standard scenario, could emerge from the way they interact with ordinary matter. A specially interesting case is the coherent neutrino-nucleus scattering. In this process, the nucleons in a given nucleus cooperate coherently to the neutrino diffusion, under the condition that the incident neutrino energy is below 50 MeV (in this region the de Broglie wave length of the neutrino is in the same order of magnitude as the diameter of a nucleus), with a momentum transfer Q2 so small that Q2R2<1, R being the nuclear size. The weak cross-section is then amplified, as it scales with the square of the mass number of the target nucleus. The price to pay is of course a very tiny energy deposition, due to the large mass of the recoiling body. In this extreme low-momentum transfer region data do not exist. The cross section \(\sigma^{\mathit{coh}}_{\nu N}\) is well predicted by the SM. Therefore, coherent elastic neutrino-nucleus scattering is an ideal tool to search for new physics in the neutrino sector.

In several frontier sectors of neutrino physics, low temperature calorimeters [6] play and have the potential to play a crucial role. These devices consist of an energy absorber, where the energy of the event under study is released, and of a temperature sensor in thermal contact with it. The temperature variation measured by the sensor provides a precise estimation of the released energy, limited only by the so-called thermodynamic limit [6]. Some peculiarities of these devices with respect to conventional nuclear detectors explain their relevance in the solution of open issues in neutrino physics:
  1. 1.

    The wide flexibility in the choice of the detector material (the only stringent requirement is a low specific heat) allows to incorporate special unsteady nuclei in the detecting medium (for example 130Te or 187Re), enabling the so-called “source=detector” approach in several favorable situations;

  2. 2.

    The high energy resolution makes easier the detection of peaks (as required in the search for 0ν-ββ) or of tiny distortions of the energy spectra (as required in the direct measurement of the neutrino mass);

  3. 3.

    The low energy threshold may allow the detection of extremely slow nuclear recoils, as those induced by the neutrino coherent scattering;

  4. 4.

    The detection of other elementary excitations (in addition to phonons) produced in the energy absorber can provide a method to identify the interacting particle and therefore to discriminate dangerous background sources.

In the next sections, practical applications of these general features will be illustrated, by reviewing the neutrino-physics topics in which low temperature calorimeters give relevant contributions.

2 Double Beta Decay

0ν-ββ [3] is a rare nuclear process, energetically possible for a few tens of candidate nuclides and experimentally interesting for less than ten nuclides. It consists in the transformation of an even-even nucleus into an isobar by the simultaneous emission of two electrons and nothing else. This process violates by two unities lepton number conservation. Of primary interest is the case in which the transition is driven by the exchange of light Majorana neutrinos interacting through the left-handed VA weak currents. The decay rate is then proportional to the square of mββ, the so-called effective neutrino mass defined by the second expression of (1). Experimentally, mββ can be determined or at least constrained within the uncertainties of the nuclear matrix elements. We ignore Nature’s choice about the neutrino mass ordering at the moment, but 0ν-ββ has the potential to provide this essential information. In fact, if mββ is measured to be greater than ≈50 meV, the QD pattern holds and an allowed range of mmin values can be extracted. On the other hand, if mββ lies in the range 20–50 meV, the pattern is likely IH. Eventually, if one determined that mββ<10 meV but non-vanishing (which is unlikely in a foreseeable future), one would conclude that the NH pattern holds.

From the experimental point of view, the shape of the two electron sum energy spectrum enables to distinguish among the neutrinoless process and that allowed by the SM, which foresees the simultaneous emission of two antineutrinos (2νββ). In the latter case, this spectrum is expected to be a continuum between 0 and Q (nuclear transition energy) with a maximum around 1/3⋅Q. For 0ν-ββ, the spectrum is just a peak at the energy Q, enlarged only by the finite energy resolution of the detector. Q ranges from ≈2 to ≈4 MeV for the most promising candidates.

The current experimental sensitivity to mββ is around 0.2–0.5 eV, corresponding to half lives in the range 1024–1025 y. A much debated claim of evidence in 76Ge by part of the Heidelberg–Moscow collaboration suggests that mββ≈0.3 eV [8]. Experiments under commissioning or construction (CUORE, GERDA-I, EXO-200, KamLAND-ZEN, SNO+, MAJORANA) can hardly start to explore the IH region [3], with sensitivities around 0.1–0.05 eV. In order to deeply analyze it, relevant expansions or improvements of these experiments are needed, or new technologies need to be developed.

The experimental strategy pursued to investigate 0ν-ββ consists of the development of a proper nuclear detector, with the purpose to reveal the two emitted electrons in real time and to collect their sum energy spectrum as a minimal information. The detector must be operated underground in order to shield the cosmic ray flux. Radioactive background must be kept under control through passive and/or active shielding and through a careful selection of the detector components in terms of radio-purity. Low-temperature calorimeters are particularly suitable to 0ν-ββ search, as they provide high energy resolution and large flexibility in the choice of the sensitive material. They represent the most advanced and promising application of the so-called calorimetric approach, in which the source is embedded in the detector itself, maximizing therefore the detection efficiency. Several interesting bolometric candidates were proposed and tested. By far the most advanced results were obtained with TeO2 low-temperature calorimeters. With this compound, it is possible to grow large crystals with reasonable mechanical and thermal properties together with a very large (27% in mass) content of the 0ν-ββ candidate 130Te. This property makes the request of enrichment not compulsory, as it is for the other interesting isotopes. Moreover, the reasonably high transition energy (Q≃2530 keV) and the favorable nuclear matrix elements make this nuclide one of the best candidate to 0ν-ββ search. A large international collaboration has been running for five years (2003–2008) an experiment based on this approach, named Cuoricino and installed underground in the Laboratori Nazionali del Gran Sasso (LNGS) [9].

Cuoricino consisted of a tower of thirteen modules, containing 62 TeO2 crystals for a total mass of ∼41 kg, corresponding to a source strength of 5.0×1025130Te nuclei. Eleven modules contained 4 TeO2 single crystals with a mass of ≈790 g each, while two modules contained 9 crystals with a mass of ≈340 g each. They were inserted in a copper frame and held by several small PTFE elements which provided the mechanical support. The Cuoricino operation temperature was in the range 10–12 mK. The temperature signal was registered by a 40 mg neutron transmutation doped (NTD) thermistor, read out by thin gold wires which also constituted the main thermal link to the heat sink. The good point of Cuoricino detectors is that they are simple, robust and exactly adapted to their scope. They are built with a restricted variety of materials (TeO2, Cu, PTFE, thermistors and some epoxy) in order to make easier the control of residual radioactive contamination. In most of the channels, the read out electronic consisted of room-temperature low noise voltage amplifiers, and the load resistors used to apply the DC bias to the thermistors were also placed at room temperature, with a very high value (≈50 GΩ) in order to make the related parallel noise negligible. This configuration allowed to limit to just two the number of read-out wires for each channel. Heaters were used to stabilize the detector response, but they were grouped and read out in parallel to save wires.

Cuoricino is still in a leading position in the 0ν-ββ search. It provides limits on mββ in the range 0.2–0.7 eV [9], depending on the choice of the nuclear matrix elements. A background of the order of 0.18 counts/(keV kg y) was obtained in the 0ν-ββ decay region, similar to the one achieved in the Heidelberg–Moscow set-up. The energy resolution is about 8 keV FWHM, quite reproducible in all the crystals. Unfortunately Cuoricino, despite a sensitivity comparable to that of the Heidelberg–Moscow experiment, cannot disprove the 76Ge claim due to the discrepancies in the nuclear matrix element calculations.

These results show how low-temperature detection of particles dominates the present experimental scenario of 0ν-ββ. They will play a major role in the future of this crucial search as well. In fact, one of the most promising projects in preparation, capable to approach and perhaps start to probe the IH region with a sensitivity down to ≈50 meV, is the CUORE experiment [10], a natural expansion of Cuoricino. It will be an array of natural TeO2 bolometers arranged in 19 towers and operated at 10–15 mK, in a custom helium-free low-radioactivity dilution refrigerator located in LNGS. The source will correspond to ≈200 kg of the isotope 130Te. The CUORE single module is similar to that of Cuoricino in its essential elements. However, a careful work was done to make all the thermal couplings reproducible, to decrease the amount of copper (source of α surface radioactivity) facing the detectors, and to prepare an assembly line which allows to speed up the tower construction under protected atmosphere. The technical results on the detector performance are excellent: in the CUORE R&D tests, which involve two tens of crystals, the average energy resolution is 4.6±1.2 keV at 2615 keV. The baseline resolution is often below 1 keV, and a threshold of 3 keV can be set for many channels realized in the CUORE R&D program. CUORE is in the construction phase and data taking is foreseen to start in 2014. A general test of the CUORE detector, comprising a single tower and named CUORE-0, will take data in 2012.

The background target for CUORE is about 20 times better than that in Cuoricino. The most limiting factor in the CUORE approach is the background induced by surface radioactive contamination [11], which populates the region of interest for 0ν-ββ for 130Te and in general for high Q-value candidates. This form of background, not completely under control, prevents CUORE from exploring deeply the IH region. For this reason, there has been a proliferation of interesting experimental ideas on how to eliminate α events (which dominate the surface background) or directly surface events, irrespectively of the particle type. This operation can be realized either keeping 130Te as isotope under study, or changing isotope. The former approach has the advantage to preserve all the exceptional advantages offered by TeO2 bolometers. Four possibilities have been considered to get rid of α/surface background in a TeO2 experiment:
  1. 1.

    Realization of the so-called surface-sensitive composite macrobolometers [12, 13]. The idea is to glue on each face of a rectangular TeO2 crystal a thin TeO2 slab, which can or cannot be equipped with a dedicated thermistor. When energy is totally or partially released in the slab (surface event), an abnormous signal is revealed by the slab thermistor (in terms of high pulse amplitude and fast rise time) and the event can be rejected. The case in which the added slab is totally passive is of course much more attractive. The slab would work as a pulse-shape modifier: the thermistor on the main crystal would see a pulse with different shape, and in particular with a longer decay time, for a surface event. Both approaches have been clearly validated with dedicated prototypes.

  2. 2.

    Deposition of superconductive films on the surfaces of a rectangular crystal [14]. If the energy is released close to the deposited film, typically within 1 mm distance (surface event), long living quasi-particles are produced in the film by the athermal phonons generated by the impinging particle. The quasiparticles will recombine with a characteristic time that can be of the order of a few milliseconds in ultrapure films, creating a delayed component in the phonon signal read out by the sensor on the main absorber. This modified temporal structure makes surface event identification possible. A very promising prototype, equipped with a 10 μm thick Al film on one side, has clearly shown that this method is viable, provided that the phonon sensor is fast enough.

  3. 3.

    Encapsulation of the TeO2 bolometer with a scintillating foil and addition of a light detector [15].

  4. 4.

    Detection of Cerenkov light in TeO2 crystals [16, 17].

The last two points will be discussed in more detail after introducing the concept of scintillating bolometers for 0ν-ββ.

When the energy absorber in a bolometer scintillates at low temperatures, the simultaneous detection of scintillation light and heat is possible, providing a very powerful tool to identify the nature of the interacting particle and therefore to suppress background. In particular, a massive charged particle can be distinguished from a β or γ due to the different light yield for the same amount of deposited heat, as pointed out a long time ago [18] and then demonstrated [19, 20]. Recently, this approach was proposed to study 0ν-ββ of 82Se (LUCIFER project) with the help of ZnSe crystals [21]. In the current technology, the scintillation photons are detected by a dedicated bolometer facing the main one, in the form of a thin slab of Ge or Al2O3.

The additional value brought by scintillating bolometers can be appreciated thanks to the following considerations. The end point of the natural γ radioactivity, which represents an important source of background, lies at 2615 keV (position of a γ line of 208Tl, belonging to the 232Th decay chain). Beyond this point, there are only extremely rare high energy γ rays from 214Bi. Therefore, a detector based on a double beta decay emitter with a Q-value above 2615 keV represents an excellent starting point for a future experiment. However, this energy region, which can also be populated by β processes, βγ and γγ coincidences, is dominated at first by energy-degraded α particles emitted by surface contamination, as already pointed out. The experiences brought by Cuoricino and CUORE-R&D show clearly that this generates a flat component in the spectral region between 2.5 and 4 MeV at the level of 0.1–0.05 counts/(keV kg y) [11]. Hence derives the power and the potential of scintillating bolometers. They offer a reasonable freedom in the choice of the candidate, that can be selected for its high Q-value. In addition, α particles are recognized and rejected, killing the dominant background component in the relevant spectral region.

Several compounds were successfully tested in LNGS [2224] as bolometers containing a high mass fraction of a high Q-value candidate: CdWO4, CdMoO4 (for 116Cd); PbMoO4, CaMoO4, SrMoO4, ZnMoO4 (for 100Mo); CaF2, CaMoO4 (for 48Ca) and ZnSe (for 82Se). Practically all these tests were realized with a detector technology similar to that employed in CUORE and based on NTD Ge temperature sensors. An interesting exception is represented by CaMoO4. Two prototypes of this material have been recently realized and successfully tested in Korea using a meander metallic magnetic calorimeter as a phonon sensor [25].

In terms of scintillating bolometers, the preliminary results show that best candidates are CdWO4, ZnMoO4 and ZnSe. The first of these three compounds has shown excellent bolometric performance and α/β rejection factor [22], but also considerable drawbacks: the relatively low mass fraction in Cd, the difficulty and the high cost in enriching Cd in 116Cd, a huge neutron capture cross-section for 113Cd and considerable internal contamination from natural radioactivity. The second compound looks immune to these defects and is now subject of an intense R&D program in LNGS [23] and Orsay (CSNSM) [26]. An α/β rejection factor better than 99.9% was demonstrated in the region of interest. The third compound is at the center of the LUCIFER project. Excellent α/β discrimination was achieved with ZnSe scintillating bolometers, in spite of the unusual behaviour of these devices, in which α light yield is larger than β light yield. The LUCIFER program is on a good track, and the collaboration is very close to procure ≈10 kg of selenium enriched in 82Se, that will enable the realization of a sensitive pilot experiment. In all the scintillating bolometers realized so far, the α/β discrimination can be obtained also by pulse shape analysis [27], both in the light and the heat channel, offering a second handle for background identification. The performance of the prototypes up to now realized and preliminary background evaluations show that an experiment with background close to zero in the ton×year scale exposure is viable with the scintillating bolometer technology. This is the level required for a full investigation of the IH region.

In a compound in which the scintillation light, if exists, is close or below the detectability threshold (as in the excellent bolometric material TeO2), a light detector could anyway help in reducing the background, as anticipated above. A surface α particle releasing part of its energy in the crystal and part in a scintillating foil encapsulating it can be tagged by analyzing the coincidence signal of heat (in the absorber) and light (in the light detector) and rejected as background. This technique has been proposed and a test is in preparation [15]. An alternative very elegant approach consists in detecting the Cerenkov light (a universal phenomenon), to which TeO2 is transparent down to 350 nm [16]. The Cerenkov light emitted in visible wavelengths is predicted to be about 350 eV at the 130Te Q-value (≈2.5 MeV). This value is extremely low compared to usual scintillating bolometers, which provide signals in the keV range at the same energy. Of course, α particles are well below the Cerenkov threshold. In spite of the extreme difficulty of this approach, a recent test performed with a TeO2 crystal coupled to a LUCIFER light detector showed luminescence light escaping the crystal in coincidence with its bolometric heat signal [17], confirming results achieved in Orsay (IAS) years ago [28]. The features of the emitted light is compatible with Cerenkov emission. This opens the way to a new fascinating method to improve CUORE sensitivity.

3 Single Beta Decay

The most sensitive direct neutrino mass measurement to date, involving electron type neutrinos, is based on studying the shape of the β spectrum. In such measurements the quantity mβ reported in the third expression in (1) is determined or constrained. Most of the information on neutrino mass is contained in proximity of the Q-value of the β transition, where the fraction of events occurring is very small and scaling as 1/Q3.

The traditional approach to the direct determination of the neutrino mass is based on the study of the tritium β decay (Q=18.6 keV) [4], extensively performed in the last thirty years with magnetic and electrostatic spectrometers. The present more stringent limit on mβ is 2.2 eV and was achieved by two experiments (Mainz and Troitsk) based on electrostatic spectrometers with adiabatic magnetic collimation, which select the interesting portion of the β spectrum. A very promising next-generation experiment, named KATRIN [29] and based on a large expansion of this approach, is under construction in the Karlsruhe tritium laboratory and the first tritium runs are planned in 2013–2014. The simultaneous improvement of resolution and luminosity with respect to the previous searches, together with a better control of background, allow to foresee for KATRIN a sensitivity one order of magnitude better than the present limit: mβ<0.2 eV 90% C.L., with a discovery potential of mβ=0.35 eV at 5σ. Due to the extreme difficulty of this kind of measurement, an alternative technique to measure the neutrino mass in that range, possibly with a totally different systematic, would be welcome, especially if the sensitivity could be pushed further at least in principle. Low-temperature calorimeters can provide this opportunity.

The isotope 187Re is an ideal candidate for a bolometric neutrino mass experiment [30]. It undergoes β decay with a unique first forbidden transition, for which the nuclear matrix element is computable, even if the calculation is not straight forward as in the case of tritium. The transition energy is 2.47 keV (the lowest known) and the half lifetime is 43.2 Gy. These values were determined with bolometric searches. The large isotopic abundance (62.8%) of 187Re in natural rhenium allows to get useful source without any isotopic separation process. The β decay rate in natural rhenium is of the order of ∼1 Bq/mg, almost ideally suited to bolometric detection. Thanks to the much lower transition energy, the useful fraction of events close to the end-point is ∼350 times higher in rhenium than in tritium. In an ideal calorimetric experiment, the source is embedded in the detector and therefore only the neutrino energy escapes to detection. The part of the energy spent for the excitation of atomic or molecular levels is measured through the de-excitation of these states, provided that their lifetime be negligible with respect to the detector time response. There is however an important inconvenience which represents a serious limitation for this approach. In a calorimeter, the whole β spectrum is acquired. This forces to keep a low counting rate in order not to generate distortions of the spectral shape due to pulse pile-up. The pulse rise time, directly related to the pulse-pair resolving time, becomes therefore one of the crucial parameters together with the energy resolution. In spite of these difficulties, elegant experiments were performed several years ago with Re-based bolometers, which served as a proof-of-principle and allowed to set rather stringent limits on neutrino mass (down to 15 eV) with a statistics of 106 events. The sensitivity to the neutrino mass scales roughly as the fourth root of the total number of events.

The preliminary bolometric searches can be considered as precursors of the current MARE project, which is foreseen to occur in two phases. Phase I will consist of an array of ∼300 detectors. The statistic gathered in three years (of the order of 1010 events) will allow to reach a sensitivity of ∼2.5 eV, scrutinizing the Mainz/Troitsk results with a completely independent approach. This phase is planned in Milano-Bicocca with square matrices of 6×6 elements, obtained with Si micromachining and provided by the NASA/GSFC and Wisconsin groups. The single Si pixel is doped by implantation close to the metal-to-insulator transition. Single AgReO4 crystals (a dielectric compound of rhenium) with a mass of about 500 μg will be epoxied to each pixel. Energy resolutions down to 30 eV FWHM and ≈250 μs rise time were demonstrated, with detectors operating in the 50–70 mK range. This performance complies with the MARE-I target and now the full array is under assembly. Phase II of MARE will take advantage of a R&D activity on faster and more sensitive temperature sensors and of the specific experience on β decay gathered in phase I. The aim of phase II is to collect a statistics of 1013–1014 events. The basic array planned for phase II will be a matrix of typically ∼10000 elements with 1–5 eV FWHM energy resolution, ≈1 μs time resolution, and a viable multiplexing scheme. This improved performance is mandatory for the success of the project and can be achieved only by changing the thermometer technology and by using metallic rhenium as energy absorber. Three options are under study for the phonon sensor:
  1. 1.

    Transition Edge Sensors (TES) based on Ir-Au multilayer film grown on a SiN membrane (Genova-Miami): the state of the art of this approach consists of an energy resolution of 11 eV FWHM at 5.9 keV and ≈160 μs rise time.

  2. 2.

    Metallic Magnetic Calorimeter (MMC) consisting of Au:Er sensors [31], which has provided so far an energy resolution of 44 eV FWHM (Heidelberg).

  3. 3.

    Microwave Kinetic Inductance Detectors (MKIDs), which are quasi-particle detectors promising high energy resolution and easy multiplexing for a large number of pixels. An R&D program on this subject will start soon in Milano-Bicocca.

Unfortunately, the MARE-II target is still far from the present performance. The difficulty in improving the energy resolution with TES and MMC may be related to the thermalization mechanism in superconducting rhenium, which is systematically studied by the Heidelberg group. The MKID approach is quite interesting in this respect, since the elementary excitations which mediate the detection in these devices are quasi-particles rather than phonons.

Once established the winning technology for the sensor, the MARE program foresees the realization of a first array as a pilot step, able anyway to attain a sub-eV sensitivity to the neutrino mass. The final set-up will consist of several arrays for a total of tens of thousands elements. The sensitivity to neutrino mass achievable by this experiment is below 0.2 eV [32]. MARE is a long way behind KATRIN at the moment. However, it could play a crucial role in the future of neutrino physics. In fact, while KATRIN is technologically the ultimate experiment mainly because of the physical size of the spectrometer, the intrinsic modularity of MARE is compatible with an arbitrary expansion of the experiment.

Unlike KATRIN, the MARE approach is sensitive to sterile neutrinos in the keV mass range, which would produce a deformation in the central region of the 187Re spectrum. The MARE precursors already allow to set a limit of the order of a few 10−3 for sin2(θ) at a neutrino mass of 1 keV [33]. The full MARE experiment would bring the sensitivity to the mixing parameter in the 10−5–10−6 range.

The difficulties in operating rhenium absorbers has lead the MARE collaboration to reconsider other approaches to the direct measurement of the neutrino mass, based on electron capture processes. One of the most interesting candidate is 163Ho. Since the 1980s, this isotope has been the subject of many experimental investigations as a powerful means for neutrino mass determination, thanks to its very low transition energy (QEC≈2.5 keV). 163Ho decays to 163Dy with a half life of about 4570 years and, because of the low transition energy, capture is only allowed from the M shell or higher. The EC decay may be detected only through the mostly non-radiative atomic de-excitation of the Dy atom and from the Inner Bremsstrahlung (IB) radiation. The most interesting approach to the measurement of the neutrino mass consists in studying the end-point of the total absorption spectrum (where all the energy released in the decay is measured except that carried away by the neutrino), as proposed by De Rujula and Lusignoli [34]. The spectrum, made up of peaks with Breit-Wigner shapes, is truncated at QECmν , in analogy with β decay spectra. In this case too, the shape at the end point depends on the neutrino mass. It is not easy to compare the sensitivity of this method (which by the way measures the neutrino mass and not the anti-neutrino mass as happens in β decay) with that based on 187Re, especially because it depends strongly on QEC, which is not precisely known [35]. In case of low QEC (of the order of ≈2.2 keV), this approach can be competitive and even superior to 187Re beta decay. In addition, there is a seductive experimental advantage: due to the relatively short half life, few nuclides (of the order of 1011) are necessary to get 1 Bq. This allows to optimize independently the energy absorber in terms of energy resolution and rise time, and then to produce, separate and introduce—this is the new experimental challenge—the desired amount of 163Ho in it. This approach is pursued by the MARE collaboration, which has realized preliminary tests using microcalorimeters with 163Ho embedded in a metallic absorber read out by a TES. A similar approach is pursued by the Heidelberg group, in the framework of the just formed ECHO collaboration. In this case, the housing detector is a microstructured MMC with 163Ho implanted in it. An excellent preliminary resolution of 11.8 eV FWHM was obtained with this device, with a rise time of only 90 ns.

4 Coherent Neutrino-Nucleus Scattering

The coherent neutrino-nucleus cross section favors very low recoil energies, up to tens of keV, undetectable with current low energy neutrino detectors. Low temperature calorimeters, which exploit elementary excitations in the target material of the order of μeV (thermal phonons) or meV (athermal phonons and quasiparticles) are ideal devices to measure so tiny energy depositions. Even if the detection of this process remains extremely challenging, proposals that define with precision the required detector properties and the related experimental configuration have now appeared.

A possible experimental configuration exploits dark matter detectors designed for WIMP searches [36]. The evolution of the present CDMS detectors, to be employed in the DUSEL project GEODM, should lead to devices consisting of 5 kg Ge crystals operated at 40 mK with a recoil threshold of the order of 5–10 keV. An array of 300 of these modules would provide a ton-scale dark matter cryogenic experiment. Similar scale and performance are foreseen for the European project EURECA. A high intensity pion- and muon- decay-at-rest (DAR) neutrino source (maximum neutrino energy ≈52 MeV) was recently proposed for oscillation physics at DUSEL. Assuming that this source is actually realized and is placed at a distance of 2.3 km from the dark matter apparatus (at the surface in correspondence of the underground laboratory), an experiment like GEODM, with baseline detector performance, could discover coherent scattering after 4.5 ton×y exposure with 3–4 signal events above a background expectation of 0.15 events. Improved, but reasonable, detector performance could lead to a signal rate of 2.0 events/(ton y). This observation would be a general test of the detector performance (WIMP and coherent neutrino signals are identical), but also an opportunity to test the Standard Model in the neutrino sector with an unprecedented low momentum transfer.

Another even more challenging approach envisions the design of a dedicated bolometric experiment for the detection of neutrino-induced nuclear recoils with a threshold of only 10 eV [37]. In this case, neutrinos would be provided by an intense electron capture source, for example a 5 MCi 37Ar source, which emits ≈800 keV nearly monochromatic neutrinos. A possible detector structure would consist of 10,000 elements. The single module would be a 50 g Si crystal coupled to a TES. A detailed analysis of the signal-to-noise ratio shows that a threshold of 10 eV is indeed possible operating the devices at 15 mK. The signal rate, from the threshold to the maximum deposited energy (50.3 eV), is expected to be 3.82 events/day in each module with the source at 10 cm distance. In this configuration, the experiment would provide not only the first detection of the neutrino-nucleus coherent scattering, but also a powerful tool to study oscillation to sterile neutrinos in the eV mass scale. In fact, for 1 MeV neutrinos the oscillation length would be of the order of few meters. The presence of an oscillatory behaviour in a pure neutral current process would indicate mixing solely to non-active neutrinos. This makes neutrino-nucleus coherent scattering an ideal method to ascertain the existence of sterile neutrinos.

5 Conclusions

The study of frontier subjects in neutrino physics require the use of frontier detectors. Typical requests, often to be fulfilled together, are energy resolutions of the order of ‰ in the MeV or in the keV energy range, energy thresholds as low as a few eV, strict requirements on the detector material, which often needs to contain a specific isotope, and finally the capability to identify the type of interating particle in order to control the background. Low-temperature detectors today are mature devices which often represent the only technical opportunity to satisfy these challenging demands. This explains their major and increasing role in neutrino physics, especially in the sectors where the tools to disantangle the properties of this elusive particle are low-energy nuclear processes.

Copyright information

© Springer Science+Business Media, LLC 2012