The Grenoble Axion Haloscope platform (GrAHal): development plan and first results

Thierry Grenet,1 Rafik Ballou,1 Quentin Basto,1 Killian Martineau,2 Pierre

Perrier,1 Pierre Pugnat,3 Jeremie Quevillon,2 Nicolas Roch,1 and Christopher Smith2

1Institut Neel, Universite Grenoble-Alpes, CNRS, Grenoble INP,38000 Grenoble, France

2Laboratoire de Physique Subatomique et de Cosmologie (LPSC), Universite Grenoble-Alpes, CNRS/IN2P353, avenue des Martyrs, 38000 Grenoble, France

3Laboratoire National des Champs Magnetiques Intenses (LNCMI),European Magnetic Field Laboratory (EMFL), Universite Grenoble-Alpes, CNRS, 38000 Grenoble, France

(Dated: October 28, 2021)

In this note we report on the development plans and first results of the Grenoble Axion Haloscope(GrAHal) project. It is aimed at developing a haloscope platform dedicated to the search for axiondark matter particles. We discuss its general framework and the plans to reach the sensitivityrequired to probe well known invisible axion models, over particularly relevant axion masses andcoupling regions. We also present our first haloscope prototype and the result of its test run atliquid He temperature, setting a new exclusion limit gaγγ ≤ 2.2×10−13 GeV−1 (gaγγ ≤ 22×gKSVZ)around 6.375 GHz (ma ' 26.37 µeV).

I. INTRODUCTION

New particles in unexplored territories of mass and in-teractions are predicted in many extensions of the Stan-dard Model (SM) of particle physics, in particular in thelow energy frontier of weakly interacting sub-eV particles.These include the QCD axion and the axion-like particles(ALPs). An utmost interest of these particles is to pro-vide serious candidates for the observed dark matter inthe universe for specific ranges of mass and density. TheQCD axion is a pseudo-scalar particle associated withthe spontaneous breaking, at an energy scale fa, of a newchiral symmetry (the so-called Peccei-Quinn symmetry)initially postulated to explain the lack of experimentalyknown charge-parity (CP) violation in quantum chromo-dynamics (QCD) [1–4]. The aforementioned symmetrybeing anomalous, the axion couples to gluons and ac-quires a mass ma that scale with f−1

a at the QCD phasetransition. Depending on the exact model, axions caninteract with other particles, as the SM fermions and thephotons with the typical scaling factor f−1

a . It was laterrealized that the QCD axion could be a serious Dark Mat-ter (DM) candidate through coherent oscillations [5–7] ofthe associated field. ALPs can also be produced from amisalignment mechanism and thereby provide contribu-tions to DM.

A number of experiments have been proposed and con-ducted to search for the QCD axion and ALPs based ontheir (possible) interactions with the SM particles [8].Among those exploiting their coupling to two photons,the largest experimental effort is presently made towardsthe use of radio-frequency (RF) cavities for resonant con-version of axions into photons in the 10−6−10−4 eV massrange, where QCD axions are supposed to contributedominantly to DM. In the world several projects of this

kind are underway aiming at different masses regions,namely: ADMX [9, 10], HAYSTACK [11, 12], CAPP [13],ORGAN [14], CAST-RADES [15] and QUAXaγ [16], aswell as MADMAX, a project of dielectric haloscope [17].

Axions detection via a RF cavity haloscope relies onthe assumption that they constitute a significant part ofthe DM galactic halo. In this case it is possible to detecttheir conversion into photons in a resonant cavity im-mersed in an intense magnetic field. The signal is poweramplified as

P = g2aγγρa1

ma

β

(1 + β)2Q0B

2V C1

1 +(

2(ν−νc)Q0

νc(1+β)

)2(I.1)

where gaγγ is the axion-diphoton coupling constant, ρais the halo DM density, ma the axion mass, β the cavitycoupling to the receiver chain, Q0 the unloaded cavityquality factor, B the magnetic field, V the cavity volume,C the coupling factor of the electromagnetic mode, νc thecavity mode frequency and ν the measured frequency.

The instrumental challenge lies in the weakness of theexpected signal (of the order of 10−24−10−23 W) relativeto the thermal radiation of the cavity. The signal-to-noiseratio (SNR) can be approximated by the Dick radiometerformula

SNR =P

kBTsyst

√∆t

∆νa(I.2)

where kB is the Boltzmann constant, Tsyst = T +TN , i.e.the sum of the detector temperature and the amplifiernoise temperature TN , ∆t the integration time and ∆νathe analysis bandwidth (set to the axion bandwidth).The SNR can be optimised by lowering Tsyst, which in-volves cooling the cavity at ultra-low temperature andusing amplifiers operating at the quantum limit of noiseor even below.

arX

iv:2

110.

1440

6v1

[he

p-ex

] 2

7 O

ct 2

021

2

The most sensitive results have been obtained so far bythe ADMX experiment [10]. Using quantum amplifiers,this experiment has shown that the technology is nowmature and sensitive enough to test the so called axionDFSZ model [18, 19] through photon couplings in theµeV range.

The Grenoble Axion Haloscope project (GrAHal) aimsat developing a haloscope platform in Grenoble (France),able to run detectors of different sizes and designs for thesearch of galactic axions and ALPs at the best sensitivityin the 0.3 − 30 GHz frequency range (1.25 - 125 µeV).It draws on Grenoble’s internationally recognised know-how on key experimental aspects of the project: highmagnetic fields and fluxes, ultra-low temperatures andultra-low noise quantum detectors.

II. THE GRAHAL PROJECT

A. Magnet systems

One of the key tools of GrAHal is the Grenoble hybridmagnet (cf. Fig.1) close to completion at LNCMI [20]and planned to be fully operational in 2023 after firstcooling down to 1.8 K and commissioning tests in 2022.It is a modular experimental platform offering variouspossibilities of maximum field values and useful warmbore diameters. From the combination of resistivepolyhelix and Bitter coils inserted within a large boresuperconducting one, a maximum field of at least 43 Twill be produced in a 34 mm diameter aperture with 24MW of electrical power. By combining the Bitter insertalone with the superconducting coil, another hybridmagnet configuration will allow to produce 17.5 T ina 375 mm diameter aperture with 12 MW. With thesuperconducting coil alone, a maximum field of 9.5 Tin 812 mm diameter bore can be reached. Compared toother running haloscopes the superconducting outsertmagnet alone offers an unprecedented value for thefigure of merit B2V ≈ 60 T2 m3. Other configurationsof magnetic field and warm bore diameter offered arelisted in Table.I. To highlight the unique opportunitiesoffered by this magnet to probe a large domain ofaxion mass and diphoton coupling, the inner diametersof cylindrical RF-cavities that can be inserted insidethe various magnet apertures are also given with theirTM010 resonance frequency and corresponding axionmass.

In addition to the Grenoble hybrid magnet platform,several superconducting coils at CNRS-Grenoble produc-ing magnetic fields up to 20 T in 50 mm diameter canbe used as test beds for RF cavities and allow the opera-tion of several haloscopes in parallel. Finally, within fourto five years, GrAHal will also benefit from the recently

FIG. 1. Status of the construction of the Grenoblehybrid magnet. In the front: assembled magnetcryostat with its warm bore diameter of 812 mm.It contains the superconducting coil immersedinside its 1.8 K He vessel and surrounded by threethermal shields at 4 K, 30 K and 100 K. On thetop of the magnet cryostat, one of the watercooling pipe for resistive magnets is beingconnected. In the back: the cryogenic line to beconnected to the magnet cryostat and to thecryogenic satellite producing the superfluid He canalso be seen. Final assembly: beginning of 2022.First operation: beginning of 2023.

funded FASUM project offering a full superconductingcoil producing 40 T in 30 mm aperture with the possi-bility of also using the 20 T outsert of 150 mm diameterbore alone.

B. Detector developments

Large magnet bores allow large cavity volumes to en-hance the axion signal to be detected. These cavities canbe developped in collaboration with IBS-CAPP in Korea[21]. The TM010 frequency of a cylindrical cavity, and

3

Magnetic Warm RF-Cavity Resonant AxionField Diameter Diameter Frequency Mass

(Tesla) (mm) (mm)* (GHz) (µeV )43 34 8 29 11840 50 23 10 4127 170 110 2 8.6

17.5 375 315 0.7 39.5 812 675 0.34 1.4

* From 1st cut design and after subtraction of the cryostatplus cavity walls thickness.

TABLE I. Main characteristics of the modularGrenoble hybrid magnet user platform. Forreference, corresponding inner diameters ofcylindrical RF cavities are also given. Dedicatedtuning mechanisms will allow frequency scans. Thepossibility to insert several matched cavities ofsmaller diameter for each hybrid magnetconfiguration is also considered.

thus the axion mass searched for, scales inversely withthe radius (cf. Table I) and a cavity mode frequency canonly be tuned over typically one octave. Thus, highermode frequencies imply smaller volumes and degradedsensitivities. Probing the whole dark matter axion massrange will imply the development of more involved de-signs to circumvent this difficulty, like multiple-cell cav-ities [22], networks of single cavities, or detectors basedon the conversion of axions to specific excitations such ase.g. plasmons [23].

To increase sensitivity, RF cavities with quality factorsof order Q ∼ 105 − 106 (higher than the values obtainedwith copper) will have to be built, which will requirededicated developments such as superconducting cavitywalls compatible with the magnetic field environment.

The total noise temperature of the haloscope Tsystfrom Eq.(I.2), i.e. of the cavity and first amplifier stage,is an important parameter as the time required to ob-tain a given signal to noise ratio is proportional to T2

syst.The follow up detectors of our path-finding haloscopedescribed below will work at He3/He4 dilution tempera-tures. As for amplifiers, one of their main figures of meritis their input noise temperature, which contributes to thesystems total noise. Josephson junction based amplifiers,such as Josephson parametric amplifiers (JPAs), have be-come the technology of choice for amplification near thequantum noise limit. GrAHal will benefit from the lat-est quantum amplifiers developments at Institut Neel:broad-band josephson travelling wave parametric ampli-fiers can be routinely fabricated using standard lithogra-phy techniques [24].

C. Reachable axion-diphoton couplings

As schematized in Fig.2 a rather large axion-diphotoncoupling exclusion domain will be accessible to the GrA-Hal experiments. However its full exploration withpresent technologies would involve a considerable inte-gration time. GrAHal will first focus on essentially un-explored mass regions, namely 1.4-2 µeV and 3.5-5 µeVdown to the DFSZ limit for the axion-diphoton coupling,as well as above 26 µeV near the KSVZ limit. Fur-ther progress in detector sensitivity, combined with anincrease of the number of haloscopes in operation and aworldwide coordination of experimental efforts will be ofprime importance to reduce integration times.

FIG. 2. Exclusion limits for the axion diphotoncoupling constant versus axion mass, presentlyestablished by haloscopes [25]. The blue bandrepresents the area of theoretical interest based onthe KSVZ and DFSZ models. The purple dashedline schematizes the exclusion domain which can beaccessed by the GrAHal experiments, using thevarious magnetic field and RF cavity configurationsdescribed in the text. The vertical purple linerepresents the first exclusion limit obtained by theGrAHal path finding run at liquid He around 6.375GHz (26.37 µeV ). It will be improved by loweringthe temperature and using a JPA.

III. THE PATHFINDING RUN

GrAHal’s path-finding run aimed at designing andoperating a first haloscope prototype working at liquidHe temperature. It consists of an oxygen free coppercylindrical cavity (inner diameter 36 mm, length 150mm), presently without mechanical tuning system, witha TM010 mode frequency close to 6.375 GHz at low tem-perature (axion mass close to 26.37µeV). It is maintained

4

FIG. 3. Main: power spectra of the different tuningsteps after subtraction of the background. Curvesare shifted vertically for clarity. Inset: rawspectrum with its background shape created by asmall thermal imbalance between the cavity andthe slightly warmer low T stage RF circuit.

at liquid He temperature inside a 52 mm bore 14 Teslamagnet. Its two antennas (main and secondary ports)allow transmission and reflection characterization mea-surements. Its unloaded quality factor Q0 is 37000±10%,the main antenna coupling constant is β = 1.2 and wetake C = 0.69 for the TM010 mode coupling factor. Thesignal emerging from the cavity main port is first am-plified by a Low Noise Factory cryogenic high-electron-mobility transistor amplifier, before further amplificationand injection in a Rhode & Schwarz ZVL13 analyzer forsignal processing. The total gain of the receiver chain iscalibrated using the power emitted by a heated load con-nected in place of the cavity. A few tuning steps couldbe taken by increasing the He exchange gas pressure. Byincreasing the inner dielectric constant of the cavity thisallows an easily controlled downward tuning of the modefrequency, in a limited range. The proper operation ofthe haloscope is checked by measuring fake axion sig-nals injected in the cavity through the weakly coupledantenna port.

The path finding run took place during 6 days in July2021. Figure 3 shows the data obtained after subtractionof the backgrounds. We obtained a single combined spec-trum by computing a weighted sum of the six flat spec-tra, taking into account their different integration timesand frequency dependent detection sensitivity given bythe lorentzian power resonance of the cavity in Eq.I.1[26], and applied a three points smoothing. For the ax-ion peak detection, a 3σ threshold was applied to thecombined spectrum. Any single point found above thisvalue would constitute a rescan candidate. We performed

FIG. 4. Individual (thin lines) and combined (thickblue line) exclusion limits on the axion-diphotoncoupling relative to the KSVZ benchmark.

numerical simulations adding synthetic virialized axionpeaks of various heights to the data, and found that peaksof height 3.65σ are detected at 95% efficiency (i.e. 5%of them produce no points above the 3σ threshold andare thus undetected). As no points were observed abovethe threshold, an exclusion limit was established for theaxion-diphoton coupling constant, assuming axions con-stitute all of the local dark matter density. This limit isshown in Figure 4.

IV. CONCLUSION

The GrAHal collaboration has undertaken its firstpathfinding run at liquid He temperature, which con-strains axion-diphoton coupling to gaγγ ≤ 2.2 ×10−13GeV−1 (gaγγ ≤ 22 × gKSV Z) around 6.375 GHz(axion mass of 26.37 µeV) with 95% confidence. Thesubsequent developments of the project were briefly de-scribed and are underway. The GrAHal main goal is todevelop a platform open to axion searches probing darkmatter axions in the highly promising region of 0.3 - 30GHz (1.25 - 125 µeV).

Acknowledgments

The authors would like to thank F. Caspers, A. Siemko,W. Venturini-Delsolaro and W. Wuensch from CERN forfruitful discussions, C. Simon, Director of LNCMI, forhis support and constructive approach toward sharing

5

the future use of the Grenoble hybrid magnet betweenGrAHal and other scientific projects, David J. E. Marshand the Grenoble “Amis des Axions” working group forhelpful discussions. The pathfinder run was partiallyfunded by the IDEX Universite Grenoble Alpes and bythe CNRS Institut National de Physique. The work ofK.M., J.Q. and C.S. is supported by the IN2P3 Mas-ter project “Axions from Particle Physics to Cosmology”and the Enigmass Labex. J.Q.’s work is also supportedby the IN2P3 Master project UCMN. The Grenoble hy-brid magnet project is supported by Universite Grenoble-Alpes (UGA), CNRS, French Ministry of Higher Educa-tion and Research in the framework of “Investissementspour l’avenir” Equipex LaSUP (Large SuperconductingUser Platform), European Funds for Regional Develop-ment (FEDER) and Rhone-Alpes region.

[1] R. D. Peccei and H. R. Quinn, CP Conservation in thePresence of Instantons, Phys. Rev. Lett. 38 (1977) 1440.

[2] R. D. Peccei and H. R. Quinn, Constraints Imposed byCP Conservation in the Presence of Instantons, Phys.Rev. D 16 (1977) 1791.

[3] S. Weinberg, A New Light Boson?, Phys. Rev. Lett. 40(1978) 223.

[4] F. Wilczek, Problem of Strong P and T Invariance inthe Presence of Instantons, Phys. Rev. Lett. 40 (1978)279.

[5] J. Preskill, M. B. Wise and F. Wilczek, Cosmology ofthe Invisible Axion, Phys. Lett. B 120 (1983) 127.

[6] L. F. Abbott and P. Sikivie, A Cosmological Bound onthe Invisible Axion, Phys. Lett. B 120 (1983) 133.

[7] M. Dine and W. Fischler, The Not So Harmless Axion,Phys. Lett. B 120 (1983) 137.

[8] P. Sikivie, Invisible axion search methods, Reviews ofModern Physics 93 (2021) .

[9] ADMX collaboration, A Search for Invisible AxionDark Matter with the Axion Dark Matter Experiment,Phys. Rev. Lett. 120 (2018) 151301 [1804.05750].

[10] ADMX collaboration, Axion dark matter experiment:Run 1b analysis details, Phys. Rev. D 103 (2021)032002.

[11] B. M. Brubaker et al., First results from a microwavecavity axion search at 24 µeV, Phys. Rev. Lett. 118(2017) 061302 [1610.02580].

[12] K. M. Backes, D. A. Palken, S. A. Kenany, B. M.Brubaker, S. B. Cahn, A. Droster et al., A quantumenhanced search for dark matter axions, Nature 590(2021) 238–242.

[13] J. Choi, S. Ahn, B. R. Ko, S. Lee and Y. K.Semertzidis, CAPP-8TB: Axion dark matter searchexperiment around 6.7 µev, Nucl. Instrum. Meth. A1013 (2021) 165667 [2007.07468].

[14] B. T. McAllister, G. Flower, E. N. Ivanov,M. Goryachev, J. Bourhill and M. E. Tobar, TheORGAN Experiment: An axion haloscope above 15GHz, Phys. Dark Univ. 18 (2017) 67 [1706.00209].

[15] A. Alvarez Melcon, S. A. Cuendis, J. Baier, K. Barth,H. Brauniger, S. Calatroni et al., First results of the

cast-rades haloscope search for axions at 34.67 µev, J.High Energ. Phys. 075 (2021) .

[16] D. Alesini et al., Galactic axions search with asuperconducting resonant cavity, Phys. Rev. D 99(2019) 101101 [1903.06547].

[17] MADMAX Working Group collaboration, Dielectrichaloscopes: A new way to detect axion dark matter,Phys. Rev. Lett. 118 (2017) 091801.

[18] M. Dine, W. Fischler and M. Srednicki, A SimpleSolution to the Strong CP Problem with a HarmlessAxion, Phys. Lett. B 104 (1981) 199.

[19] A. R. Zhitnitsky, On Possible Suppression of the AxionHadron Interactions. (In Russian), Sov. J. Nucl. Phys.31 (1980) 260.

[20] LNCMI, “2020 annual report”, pp.90-93,http://lncmi.cnrs.fr/wp-content/uploads/2021/02/

LNCMI_AR2020vwBD.pdf, 2021.[21] P. Pugnat, R. Ballou, P. Camus, F. Caspers, B. R. Ko,

N. Roch et al., Preliminary Study for a New AxionDark-Matter Haloscope, in 12th ”Patras” Workshop onAxions, WIMPs and WISPs, PATRAS 2016 / Kim,Younglm, Lindner, Axel, Semertzidis, Yannis K. (eds.),Verlag Deutsches Elektronen-Synchrotron : 2017,DESY-PROC, (Hamburg), pp. 129–132.

[22] J. Jeong, S. Youn, S. Bae, J. Kim, T. Seong, J. E. Kimet al., Search for invisible axion dark matter with amultiple-cell haloscope, Physical Review Letters 125(2020) .

[23] M. Lawson, A. J. Millar, M. Pancaldi, E. Vitaglianoand F. Wilczek, Tunable axion plasma haloscopes, Phys.Rev. Lett. 123 (2019) 141802.

[24] L. Planat, A. Ranadive, R. Dassonneville,J. Puertas Martınez, S. Leger, C. Naud et al.,Photonic-crystal josephson traveling-wave parametricamplifier, Phys. Rev. X 10 (2020) 021021.

[25] P. D. Group, Review of Particle Physics, Progress ofTheoretical and Experimental Physics 2020 (2020) .

[26] B. M. Brubaker, L. Zhong, S. K. Lamoreaux, K. W.Lehnert and K. A. van Bibber, Haystac axion searchanalysis procedure, Phys. Rev. D 96 (2017) 123008.