Axions in Astrophysics • Axion Detection

Axions and their detection

Igor G. Irastorza Center for Astroparticle and High Energy Physics (CAPA) University of Zaragoza

Les Houches Dark Matter School, August 2021 Outline of the lectures

• Axion theory (introduction) Axions in cosmology • Lectures 1 & 2 – Axion as DM candidate • Axions in astrophysics • Axion detection:

– Laboratory axions Lectures 3 & 4 – Axions from the Sun – Dark matter axions

Les Houches DM School Igor G. Irastorza / Universidad de Zaragoza 2 Some bibliography & references

• Notes for the lectures – Abundant references for further deepening (focus on • Some recent reviews: detection) – I. G. Irastorza and J. Redondo arXiv: 1801.08127 Di Luzio et al. arXiv: 2003.01100 (focus on theory & – phenomenology) Axions in wider contexts: – J. Billard et al., arXiv: 2104.07634 (axions among other DM searches) P. Agrawal et al., arXiv: 2102.12143 (feebly interacting particles (FIPS) – in extensions of SM)

– Textbook “Ultralight Bosonic Dark Matter”, edited by D. Kimball and K. van Bibber, May 2021. in press, to be published by Springer

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 3 Axions in a nutshell

• Most compelling solution to the Strong CP problem of the SM

Theory Cosmology • Axion-like particles (ALPs) predicted by many Strong CP Cold DM extensions of the SM (e.g. string theory) problem candidate • Axions, like WIMPs, may solve the DM problem for axions free. (i.e. not ad hoc solution to DM) Dark String ALPs radiation • Astrophysical hints for axion/ALPs? theory – Transparency of the Universe to UHE gammas Inflation Astrophysics Stellar anomalous cooling  g ~ few 10-11 GeV-1 / m Dark – aγ a Anomalous Energy? ~few meV ? stellar cooling + UHE γ • Relevant axion/ALP parameter space at reach of transparency current and near-future experiments • Still too little experimental efforts devoted to axions

Les Houches DM - Axions Igor G. Irastorza / Universidad de Zaragoza 4 Strong CP problem

• QCD Lagrangian contains the famous “ -term”:

𝜃𝜃 This term is CP violating.

input parameter of the SM

• Total derivative term: no effect in perturbative calculations • BUT it has observational consequences: e.g. it predicts electric dipole moments (EDM) to hadrons,most importantly, to neutrons:

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 5 Strong CP problem

• QCD Lagrangian contains the famous “theta term”:

This term is CP violating.

• From non-observation of neutron EDM in experiments:

• Why so small?

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 6 Strong CP problem

• QCD Lagrangian contains the famous “theta term”:

This term is CP violating.

input parameter of the SM Two contributions of very different origin

From the vacuum Quark mass structure of QCD matrix

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 7 Strong CP problem

Two contributions of very different origin

The essence of the strong CP-problem is why is so small if composed of two phases of completely unrelated origin 𝜃𝜃

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 8 Peccei-Quinn mechanism

• New U(1) symmetry introduced in the SM: Peccei Quinn symmetry of scale

• The symmetry is spontaneously𝐴𝐴 broken at the scale  a Nambu-Goldstone 𝑓𝑓boson appears

PQ symmetry 𝑓𝑓𝐴𝐴 Potential of a Higgs-like field PQ symmetry spontaneously broken below scale

𝐴𝐴 Radial𝑓𝑓 mode acquires a VEV:

Angular mode is the NG boson

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 9 Peccei-Quinn mechanism

• A new term appears in the lagrangian, formally identical to the theta term, but instead of parameter we have a field.

θ absorbed in the definition of “Axion lagrangian” the field A

= / in other words, the “constant” is promoted𝜃𝜃 to a 𝐴𝐴field𝑓𝑓𝑓𝑓 ( , ) … and therefore the original is rendered𝜃𝜃 unphysical.𝜃𝜃 𝑥𝑥 𝑡𝑡 𝜃𝜃

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 10 Peccei-Quinn mechanism

• A new term appears in the lagrangian, formally identical to the theta term, but instead of parameter we have a field.

θ absorbed in the definition of A “Axion lagrangian”

QCD-induces a potential to the = / relaxes to zero… field (*) 𝐴𝐴 CP conservation is preserved “dynamically” Minimum of 𝜃𝜃 𝐴𝐴 𝑓𝑓𝑓𝑓 potential is CP conserving (*) note that this potential is periodic because / is an angle

𝐴𝐴 𝑓𝑓𝑓𝑓 Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 11 The axion

• Weinberg and Wilczek independently noted that the quanta of the new field could be observable particles, baptized as axions.

“Axion lagrangian”

axion field

“…because it cleaned up a problem…”

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 12 The axion

• Basic properties: – Pseudoscalar particle – Neutral – Gets very small mass through mixing with pions axion field – Stable (for practical purposes). – Phenomenology driven by the PQ scale fa. (couplings inversely proportional to fa)

Note on notation: uppercase A when we refer to Axion mass is a QCD effect, and thus fully QCD axion; lowercase a when we refer to more generic axion-like particles determined (in principle) apart from the dependency on

𝐴𝐴 Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 𝑓𝑓 13 Axion models • When originally proposed, identified with electroweak scale

• This implies too large coupling𝐴𝐴 with signatures in accelerators that were not seen. 𝑓𝑓 • These Peccei-Quinn-Weinberg-Wilczek (PQWW) axions were soon ruled out.

• “invisible axion” models: is a new scale at much higher energies. (now all models are of this type) 𝑓𝑓𝐴𝐴

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 14 Axion models • Many models exist (specially recently, large model-building effort). But most can be linked to these “benchmark” models:

KSVZ axions DFSZ axions

• SM extended with: heavy q + singlet • No exotic fermions, SM fermions assigned complex scalar H PQ charges • SM extended with: 2 Higgs doublets + • H  implements spontaneours breaking singlet complex scalar H of PQ symmetry • 2 variants DFSZ-I or –II (depending on • More complex generalizations possible which Higgs is involved in lepton masses)

• They do not feature coupling to electrons • They do feature coupling to electrons

• Sometime called “hadronic axions” • Sometime called “GUT axions” as they are compatible with GUT scenarios

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 15 Axion phenomenology • Some phenomenology is directly expected from the PQ mechanism, while other depends on the particular way it is implemented in the SM, or the “axion model”.

Gluon coupling Mass Photon coupling Fermion couplings

Electron coupling Nucleon coupling … ×

generic generic generic but value Model dependent model dependent

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 16 Axion-photon coupling

• It contains both a model- independent contribution and a model–dependent one.

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 17 The “models band”

• For a given model, gAγ is proportional to ma. • Under rather general assumptions:

• This fixes a yellow band where QCD axions are typically expected

Note: however, more contrived modelling allows to bypass this constrain (axions outside the band are possible)

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 18 Axions beyond the “band”

• Conventional QCD axion models lie on the “yellow band” – KSVZ, DSFZ benchmarks • Outside the band typically ALPs

• BUT a lot of “model building” activity in recent years, leading to QCD axion models outside the conventional band…

– Normally populating higher gaγ. – Very interesting for experiments!

P. Agrawal, ESPP 2019 Granada

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 19 Axion-nucleon coupling

• It contains both model-independent contributions and model–dependent ones.

protons neutrons

• Note that cancellations between terms can occur (although for typical models not simultaneously for protons and neutrons)

is the ratio of VEVs of the 2 Higgs Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 20 Axion-electron coupling

≈ • The model-independent component is negligible (second order effect), so this coupling is model– dependent. E.g. DSFZ models feature it.

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 21 Other couplings

• The above cases are the most relevant axion coupling, but other exist:

• Coupling with pions and other mesons  relevant in cosmology

• Higher dimensional terms are also possible, in particular, terms of the type that leads to the existence of the neutron electric dipole moment

𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹 • More exotic, CP-violating scalar Yukawa couplings may be expected by e.g. new CP-violating physics beyond the SM that efectively shifts the minimum of the axion potential away from zero (strongly constrained by astrophysics)

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 22 Axion-photon mixing / conversion • vertex means that axions interact with a electromagnetic field (virtual photon) and 𝑎𝑎convert𝛾𝛾𝛾𝛾 into a photon (and viceversa).

 Axion-photon conversion in the presence of an electromagnetic field (Primakoff effect)

This is probably the most relevant of axion properties. Most axion detection strategies are based on the axion-photon coupling

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 23 Axion-photon mixing

• If the interaction happens coherently over a macroscopic EM field, the process is better seen as that of mixing,𝑎𝑎𝑎𝑎𝑎𝑎 similar to flavor mixing and neutrino oscillations. The probability of conversion oscillates between axion and photon along the line of propagation Length propagated Transverse B field N (*) 𝑎𝑎 S 𝛾𝛾 momentum difference (*) Approximations: 1D propagation, small mixing, relativistic velocities.

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 24 Axion-like particles (ALPs)

• Phenomenological definition: particles with similar phenomenology that the axion (light Weakly Interacting particles coupled to photons,…) Sub-eV Particles • Theoretical frameworks: Many extensions of (WISPs) SM predict axion-like particles Hidden Photons – Higher scale symmetry breaking Axion Like particles (HPs) (ALPs) String theory predicts a Stringy ALPs plenitude of ALPs AXIONS Arion Majoron Millicharged Familon Particles Chameleons • Not necessarily linked with PQ mechanism • WISP: similar (more generic) nomenclature

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 25 Beyond axions

• ALPs couplings and mass do not necessarily follow the ~1/ relation of axions. 𝑓𝑓𝐴𝐴 • ALPs can live anywhere in this plot, also outside the yellow band

Generic ALPs parameter space 

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 26 Axion Cosmology

¿How can it be that such a light particle can be a good DM candidate? (remember the WIMP miracle, it works for particle masses ~100 GeV)

Thermal production of axions lead to a relativistic population, that is, Hot DM (for ~1 eV) or Dark Radiation (for lower ). So, it does not solve the DM problem. 𝑚𝑚𝑎𝑎 If fact, one𝑚𝑚 can𝑎𝑎 use the maximum HDM density allowed by cosmological observations to constrain the axion mass:

Note: this limit applies only to axions not to generic ALPs (as coupling with mesons are invoked in the argument)

But the PQ mechanism itself offers a non-thermal production of non-relativistic axions in quantities large enough to fit the Cold DM required density

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 27 Vacuum realignment (VR)

• Let’s follow the evolution of the axion potential with the Universe temperature . • When ~ (PQ transition) the (massless!) axion first appears as a physical degree of freedom. 𝑇𝑇 𝐴𝐴 𝑇𝑇 𝑓𝑓 Before PQ transition After PQ transition > < Potential of a 𝑇𝑇 𝑓𝑓𝐴𝐴 𝑇𝑇 𝑓𝑓𝐴𝐴 Higgs-like field PQ symmetry spontaneously

broken below scale fa

The field is forced to take a particular (random) value in the angular component.

Radial mode acquires a VEV ~ PQ symmetry Angular mode is the axion. The potential is flat in the circular “valley” (the axion𝑓𝑓𝐴𝐴 is massless)

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 29 Vacuum realignment (VR)

• When ~ The QCD-induced axion potential “turns on”. • The axion acquires its mass. 𝑇𝑇 𝑇𝑇𝑄𝑄𝑄𝑄𝑄𝑄 Before QCD transition After QCD transition > <

The axion is 𝑄𝑄𝑄𝑄𝑄𝑄 The QCD axion potential 𝑇𝑇 𝑇𝑇𝑄𝑄𝑄𝑄𝑄𝑄 𝑇𝑇 𝑇𝑇 massless can be seen as a “tilting” of the Mexican hat potential It takes a random of . value , There𝜙𝜙 is now a minimum 𝑖𝑖 (remember𝜃𝜃 ∈ −=𝜋𝜋 𝜋𝜋/ ) that turns out to be CP

𝐴𝐴 conserving (that is the Note doesn𝜃𝜃 𝐴𝐴’t 𝑓𝑓 essence of the PQ need to the be same 𝑖𝑖 mechanism) is different𝜃𝜃 parts of the Universe.

The axion field starts rolling from whatever value had (randomly) taken before to the minimum of the potential. And keeps oscillating around it. 𝜃𝜃𝑖𝑖 Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 30 Vacuum realignment (VR)

• The oscillations of the axion field fills the Universe with non-relativistic axions ( Cold DM). Its density depends on the initial angle . • The VR mechanism is fully determined given and . 𝜃𝜃𝑖𝑖 𝜃𝜃𝑖𝑖 𝑓𝑓𝐴𝐴

For ~1 a ~6 eV axion would fill the needed DM Ratio of VR axion 𝑖𝑖 density over total axion𝜃𝜃 via VR mechanismμ DM density

Note the approx. inversely proportional relation O(1) factor accounting for between ~1/ . Contrary anharmonicities in the to thermal relics. axion potential (calculable Ω 𝑚𝑚𝐴𝐴 in principle – QCD effects)

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 31 Pre- and post-inflation scenarios

Pre-inflation scenario Post-inflation scenario Hubble scale > 2 < 2 at inflation 𝑓𝑓𝐴𝐴 𝜋𝜋𝐻𝐻𝐼𝐼 𝑓𝑓𝐴𝐴 𝜋𝜋𝐻𝐻𝐼𝐼 • The PQ symmetry is broken during • The PQ symmetry is broken after inflation and not restored afterwards inflation. • Inflation “selects” one , that is now • The angle takes different values in constant across the observable different “patches” of the Universe. Universe. 𝜃𝜃𝑖𝑖 𝜃𝜃𝑖𝑖

is “sampled” from , with equal The same in 𝜃𝜃probability𝑖𝑖 across the all the Universe.−𝜋𝜋 𝜋𝜋 observable𝜃𝜃𝑖𝑖 Universe

typical size of single patch nowadays:

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 32 Topological defects

• The value of ( ) smoothly connects between different patches. 𝜽𝜽 𝑥𝑥 • It may happen that in some = min boundaries it is topologically trapped to go around the domain , , 𝜽𝜽 𝒙𝒙𝟏𝟏 𝜽𝜽 where it stores a large energy. −𝜋𝜋 𝜋𝜋

= max = min 𝜽𝜽 𝒙𝒙𝟐𝟐 𝜽𝜽 𝜽𝜽 𝒙𝒙𝟑𝟑 𝜽𝜽

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 33 Topological defects (TD)

• The walls are the boundaries • TD take the shape of strings and walls. between two domains in different , minima. 𝟎𝟎 𝟐𝟐𝟐𝟐 • Their formation, dynamics and eventual • The field across the domain wall decay must be computed with complex takes all values , between the minima. cosmological simulations. • In the strings the𝟎𝟎 field𝟐𝟐𝟐𝟐 takes all There is a lot of research activity in this values , along any loop • enclosing the string. topic nowadays. 𝟎𝟎 𝟐𝟐𝟐𝟐 • There is no consensus on the importance of TD axions wrt VR axions.

• If > 2 inflation wipes out TD. Therefore they are only important in post- inflation𝑓𝑓𝐴𝐴 𝜋𝜋 scenarios𝐻𝐻𝐼𝐼

Saikawa, arXiv:1709.07091 Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 34 Pre-inflation scenario

Pre-inflation scenario • Inflation “selects” one , that is now > 2 constant across the observable 𝜃𝜃𝑖𝑖 𝐴𝐴 𝐼𝐼 Universe, but its value is unknown, is a 𝑓𝑓 𝜋𝜋𝐻𝐻 free parameter of the model.

The same in all the observable𝜃𝜃𝑖𝑖 Universe

• can be finetuned to get the right , for practically any up to ~meV 𝜃𝜃𝑖𝑖 Ω𝐴𝐴 𝑉𝑉𝑉𝑉 𝑚𝑚𝐴𝐴 • TD are wiped out by inflation, so VR is • Very low and very low could be the only mechanism of DM axion justified by anthropic reasons 𝜃𝜃𝑖𝑖 𝑚𝑚𝐴𝐴 production.

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 35 Post-inflation scenario

Post-inflation scenario • At cosmological scales, it is justified to < 2 assume an effective average:

𝑓𝑓𝐴𝐴 𝜋𝜋𝐻𝐻𝐼𝐼 (*) • And then:

• But in this case one has also to take into account , , Its computation 𝐴𝐴 𝑇𝑇𝑇𝑇 Ω is uncertain = , + ,

Ω𝐴𝐴 Ω𝐴𝐴 𝑉𝑉𝑉𝑉 Ω𝐴𝐴 𝑇𝑇𝑇𝑇 • Different estimates for , : from similar to , up to one order or even two 𝐴𝐴 𝑇𝑇𝑇𝑇 • The angle takes different values in orders of magnitude largerΩ …. different “patches” of the Universe. Ω𝐴𝐴 𝑉𝑉𝑉𝑉 𝜃𝜃𝑖𝑖 (*) = 2.15 when anharmonicities in the axion potential are taken into account Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 36 Domain wall problem

• For some models, the axion potential • The of different patches may roll down to features several ( ) physically-distinct physically different minima. In such cases the degenerate minima. TC formed𝜃𝜃𝑖𝑖 are stable and do not decay, leading 𝑁𝑁𝐷𝐷𝐷𝐷 •  “domain wall number” to a Universe very different, incompatible with observations… 𝑁𝑁𝐷𝐷𝐷𝐷 = This is a problem only in the post-inflation scenario (in pre-inf TD do not form) 𝑁𝑁𝐷𝐷𝐷𝐷 𝟒𝟒

• Possible solution: add new physics to make the minima non-degenerate. • The TD are longer-lived and the resulting higher. Pushes The original PQWW axion has = interesting to higher values (up to and the DFSZ models described 𝐴𝐴 above have = or . 𝑁𝑁𝐷𝐷𝐷𝐷 𝟑𝟑 few x100 meVΩ ). 𝑚𝑚𝐴𝐴 𝑁𝑁𝐷𝐷𝐷𝐷 𝟑𝟑 𝟔𝟔 Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 37 Best for axion DM?

• Traditional window to look for axionDM 1,100 meV (VR production with “natural” values of ) • But finetuned values of may be justified𝐴𝐴 by anthropic reasons. 𝑖𝑖 𝑚𝑚𝐴𝐴 ∈ 𝜃𝜃 • + TD is post-inflation models… 𝑚𝑚 𝜃𝜃𝑖𝑖 • Main message: right DM can be obtained in a large range of values. • Subdominant DM component <  corresponding larger values 𝑚𝑚𝐴𝐴 • Remember: thermal production of axions (as neutrinos) gives hot DM (upper limit ma~1 eV) Ω𝐴𝐴 Ω𝐷𝐷𝐷𝐷 𝑚𝑚𝐴𝐴

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 38 Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 39 Best for axion DM?

𝑚𝑚𝐴𝐴

• Distribution of our “confidence” on where the DM axion is: – Assumptions: DM is all axions, QCD axions, standard models,…

Bayesian analysis from Hoof et al. 1810.07192

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 40 ALP dark matter

• The axion DM production mechanism are common to other ALPs from HE symmetry spontaneous breaking. • Much of the ( , ) parameter space can potentially contain viable ALP DM models. Not very predictive 𝐴𝐴 𝐴𝐴𝛾𝛾 – Exception: ALP𝑚𝑚 miracle𝑔𝑔 models (ALP is also the inflaton)

Arias et al arXiv:1201.5902 Daido et al arXiv:1710.11107

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 41 Other cosmological implications of ALPs

• Axions can be hot DM: < 0.53 eV

• Lighter axions can be Dark𝐴𝐴 Radiation: ~0.027  detectable effect in future cosmological probes. 𝑚𝑚 ∆𝑁𝑁eff

• In pre-inflation models, the axion imprints isocurvature fluctuations in the CMB. • They are not seen so far, leading to dependent constraints 

𝐻𝐻𝐼𝐼

Visinelli & Gondolo PRL 2014 Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 42 Other cosmological implications of ALPs

• Axion decay is longer than the lifetime of the Universe. But can lead to visible𝑎𝑎 → 𝛾𝛾𝛾𝛾 -lines from DM rich regions. 𝛾𝛾 • Shorter decay times would have other cosmological consequences: – distortion of the CMB spectrum, – affect the result of the primordial nucleosynthesis, – Produce monochromatic X-ray and gamma-ray lines in the extragalactic background light,

– alter the H2 ionization fraction. Cadamuro et al. 1011.3694

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 43 Axions in astrophysics

• Stars as axion laboratories: – Axions modify the stellar evolution in ways that may produce observable consequences • Stars as axion factories (*): – Axions can be produced in stars and offer a direct detection opportunity (axion helioscopes)

• Effects in the propagation of photons over large distances: – The mix with axions in the large scale B-fields may produce observable effect

Still a fundamental reference !

Similar to the neutrino case: solar neutrinos have been crucial to dilucidate fundamental properties (neutrino oscillation) and at the same time are an excellent probe to confirm our understading of the solar interior (solar models)

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 44 Stellar evolution

• The life of a star is determined by a balance between the gravitational collapse and the pressure from nuclear fuel. The mass of the star determines its fate.

A massive star’s life cycle. From Eldridge 2008 Martínez-Pinedo, 2016

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 45 Stellar evolution

• As long as the gravitational pressure increases and the lighter nuclear fuel is exhausted, new stages start to burn H  He  C & O , etc…

A massive star’s life cycle. From Eldridge 2008

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 46 Les Houches DM - Axions Igor G. Irastorza / Universidad de Zaragoza 47 Color-magnitude Diagram

Properties of star’s surface

𝐿𝐿 observational CMD of the M5 globular cluster, which shows the luminosity and surface eff temperature of stars at a fixed 𝑇𝑇 time (isochrones).

Les Houches DM - Axions Igor G. Irastorza / Universidad de Zaragoza 48 Color-magnitude Diagram

H H He He C O

Asymptotic giant star Horizontal branch star

H He

Main sequence star Red Giant

Les Houches DM - Axions Igor G. Irastorza / Universidad de Zaragoza 49 The role of axions in stellar evolution

• Axions (or any other feebly interacting particle – like neutrinos) can be produced in the star’s core. If they escape without interacting, they may constitute a relevant energy drain mechanism for the star and alter its evolution. emission rate

Volume emission Surface Emission

coupling Feeble photons, neutrions, electrons, axions, … … • Which observables can best manifest the presence of such energy drain?

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 50 RGB and HB stars

• Additional cooling delays the He-ignition, moving the tip of the RGB to higher luminosities

• Particle emission reduces He burning lifetime, i.e. number of HB stars

• R-parameter: = 𝑁𝑁HB • The ratio depends on the𝑁𝑁RGB efficiency of the𝑅𝑅 energy loss in the two evolutionary stages

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 51 White dwarfs

• Last stage of evolution of low mass stars. Nuclear fuel exhausted. Evolution is a simple gravothermal process. Very well known evolution.

• Particle emission affects the distribution of number of WD of a given luminosity, i.e. the WDLF (WD luminosity function)

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 52 The “blue loop”

• Heavy stars in the He-burning stage have a particular evolution towards the blue (hotter) region of the HR diagram and back, a feature known as the blue loop.

• The time spent in this transient would be particularly sensitive to extra cooling mechanism, and indeed such a case could explain the observed defficitof blue versus red stars

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 53 Axion production mechanisms in stars

• Depending on the stellar environment ( , ) the relevant mechanism of axion production is different. 𝝆𝝆 𝑇𝑇 • Important. Different stars constraint different axion couplings

Production mechanisms:

Primakoff Compton Bremsstrahlung Low , high Low , high High

𝝆𝝆 𝑻𝑻 𝝆𝝆 𝑻𝑻 𝝆𝝆

𝑔𝑔𝑎𝑎𝑎𝑎 𝑔𝑔𝑎𝑎𝑎𝑎 𝑍𝑍𝑍𝑍 𝑔𝑔𝑎𝑎𝑎𝑎 𝑍𝑍𝑍𝑍 Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 54 R-parameter

• Cooling mechanism in HB stars can be due to or . Assuming negligible, one can extract a bound from the observation 𝑎𝑎𝛾𝛾 𝑎𝑎𝑒𝑒 𝑎𝑎𝑒𝑒 of 𝑔𝑔R-parameter𝑔𝑔 : 𝑔𝑔 Preference for non- zero coupling

HB bound

helium mass fraction HB hint

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 55 coupling – astro summary

< 0.65 ( ), best: 0.3 Straniero, Ayala, Dominguez, M.G., Mirizzi (2015) 𝑎𝑎𝛾𝛾 HB stars 2𝜎𝜎 𝑔𝑔 Ayala, Dominguez, M.G., Mirizzi, Straniero (2014) (R-parameter) < 0.66 (2 ), best: 0.4 𝜎𝜎 < 1 Raffelt, Dearborn (1987)

Blue loop Massive < 0.8 A. Friedland, M.G., M. Wise (2013) stars

N. Vinyoles, A. Serenelli, F. L. Villante, S. Basu, J. Redondo, J. Isern, (2015). — Global analysis < 4.1 (3 ) Sun P. Gondolo, G. G. Raffelt, Phys. Rev. D79 (2009) — neutrinos from𝜎𝜎 𝟖𝟖 < 7 (3 ) 𝑩𝑩 𝜎𝜎 × 10 Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 56 10 𝑔𝑔𝑎𝑎𝑎𝑎 𝐺𝐺𝐺𝐺𝐺𝐺 coupling

𝑔𝑔𝑎𝑎𝛾𝛾

HB bound

HB hint

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 57 White dwarfs: WDLF

• The WDLF strongly constrains the axion-electron coupling • Slight preference for a non- zero coupling can be translated to a “hint”:

For recent review of situation: Isern, arXiv:2002.08069

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 58 WD variables: period drift

Drift is visible in observations many • Measures of the period change rate in years apart WD variables offer a way to directly test the cooling rate in a single WD: / /

𝑃𝑃̇ 𝑃𝑃 ∝ 𝑇𝑇̇ 𝑇𝑇 • Observations over the past ~30 yr showed consistently > , which seems to imply an overly efficient cooling. 𝑃𝑃𝑜𝑜𝑜𝑜𝑜𝑜̇ 𝑃𝑃𝑡𝑡̇ ℎ

Córsico et al. 2019

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 59 RGB tip

RGB tip luminosity increases for larger • The luminosity of the RGB tip, the point of maximum values luminosity, when RBG stars reach the condition to (Straniero et al 20202) 𝑔𝑔𝑎𝑎𝑎𝑎 ignite Helium (known as the He-flash), is sensitive to extra cooling mediated by . • Recent analysis using better quality data: 𝑔𝑔𝑎𝑎𝑎𝑎

• Statistical combination with WD and RGB point to a almost ~3 hint for a non-zero coupling:

𝜎𝜎 Capozzi et al, 2020 ; Straniero et al. 2020

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 60 coupling – astro summary

𝑎𝑎𝑎𝑎 22 GC 𝑔𝑔 < 1.4 (2 ), best 0.6 O. Straniero et al., Astron.Astrophys. 644 (2020) RGB tip NGC 4258 < 1.6 (2 𝜎𝜎) F. Capozzi, G. Raffelt, Phys.Rev.D 102 (2020) 𝜎𝜎

WDV < 4.1 (2 ), best: 2.9 Córsico, Althaus, Miller Bertolami, Kepler, Astron. Astrophys. Rev. 27 (1) (2019) 𝜎𝜎 0.5-2.1 (2 ), best: 1.4 M. Bertolami, E. Melendez, L. G. Althaus, J. Isern (2014) WDLF

< 3.5 𝜎𝜎Raffelt (1986)

× 10 Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 13 61 𝑔𝑔𝑎𝑎𝑎𝑎 coupling

𝑔𝑔𝑎𝑎𝑎𝑎

Stellar bound WD hint

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 62 Global fits on ( , )

• A global analysis of RGB, WDLF, WDV and R-parameter gives a preference to some 𝑎𝑎𝑎𝑎 𝑎𝑎𝑎𝑎 energy loss unaccounted in the SM and 𝑔𝑔 𝑔𝑔 explainable by axions coupled to photons and electrons. • Couplings accessible to next gen. experiments (ALPS II, BabyIAXO), if sufficiently small (see experiments) 𝑚𝑚𝑎𝑎

Giannotti et al. 2017 (+ work in prep)

Les Houches DM - Axions Igor G. Irastorza / Universidad de Zaragoza 63 SN1987A neutrino signal

• Few neutrinos detected in coincidence of SN1987A explosion. Time spread of few seconds. • Emission of very weakly interacting particles would “steal” energy from the neutrino burst and shorten it. (Early neutrino burst powered by accretion, not sensitive to volume energy loss.) • Late-time signal most sensitive observable

Raffelt 2012

Les Houches DM - Axions Igor G. Irastorza / Universidad de Zaragoza 64 SN1987A neutrino signal

• Axion-nucleon interactions are the relevant ones in the core of a SN. • Combined on and : Carenza, et al, 2019 𝑔𝑔𝑎𝑎𝑝𝑝 𝑔𝑔𝑎𝑎𝑛𝑛 • Can be translated to a bound on :

𝑚𝑚𝑎𝑎 • Considerable uncertainty remains in the derivation of this bound, that stems from the supernova modelling itself, the sparse neutrino data on which it is based, and from the difficulty of describing the axion production in processes in a high density nuclear medium. A lot of literature on the topic…

• Cooling of neutron stars: similar constraints derived.

Les Houches DM - Axions Igor G. Irastorza / Universidad de Zaragoza 65 Summary of astro bounds & hints

To be updated…

Di Luzio et al 2020 Les Houches DM - Axions Igor G. Irastorza / Universidad de Zaragoza 66 Stars as axion factories

• Axions can be produced in stars and offer a direct detection opportunity: axion helioscopes (no counterpart in the WIMP sector)

• Axions from the Sun • Also axions from a future SN explosion… • Axions from massive stars back-converted into photons…

• More about this in the “experiments” part…

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 68 HE propagation over large distances

• Gama ray telescopes like MAGIC or HESS observe HE photons from very distant sources.𝛾𝛾 mixing in intervening -fields: – Reduced opacity of medium to HE  apparent hardening of spectra – Oscillatory𝛾𝛾 − features𝑎𝑎 in the photon spectrum𝐵𝐵 due to turbulent structure of -fields. 𝛾𝛾 𝐵𝐵

Sánchez-Conde, 2009

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 69 HE propagation over large distances

• These effects have been invoked both in claims of hints and𝛾𝛾 exclusions in the ballpark of:

• Lots of papers. Whether there is now a hint or not is controversial. In any case, region at reach of near- future experiments.

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 70 Axion/ALP searches motivation

“Focuses of interest” in the ALP parameter space

Theory Astrophysics Cosmology

Generic ALP DM models

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 71 Sources of axions Natural sources Laboratory sources Dark matter Dark radiation Stellar

• Photon-ALP conversion in strong magnetic fields (axion- photon coupling)

• ALP fields from macroscopic bodies (fermionic couplings)

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 72 Axion detection strategies

Most detection strategies rely on the axion-photon conversion

Table from Irastorza & Redondo 2018 Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 73 How to detect an axion?

a γ Axion source Photon detector

Igor G. Irastorza / Les Houches DM school 74 Universidad de Zaragoza Detection of axions

Model & Source Experiments Cosmology Technology dependency

ADMX, Relic HAYSTAC, CAPP, MADMAX, New ideas emerging, BRASS, ORGAN, RADES, High Active R&D going axions QUAX, CASPEr, on,… SHAFT, ABRA, DM-Radio, …

Lab ALPS, JURA, Ready for large OSQAR, PVLAS, Very low scale experiment axions ARIADNE,…

Solar SUMICO, CAST, Ready for large Low axions (Baby)IAXO scale experiment

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 75 Laboratory axions

γ a γ

Photon Photon source detector WALL

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 76 Light-shining-through-wall (LSW)

Standard configuration 

Enhanced “resonant” configuration (future) 

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 77 Axion-photon conversion in a B-field Length propagated Transverse B field N (*) 𝑎𝑎 S 𝛾𝛾 momentum difference

• Momentum difference: (*) Approximations: 1D propagation, small mixing, relativistic velocities.

• If 1, i.e. the length of magnet is much smaller than the oscillation length, the conversion𝑞𝑞𝑞𝑞 ≪ is “coherent”, and : 2 𝑃𝑃 ∝ 𝐿𝐿

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 78 LSW figure of merit (FOM)

• In a LSW we have a double conversion : 𝛾𝛾 → 𝑎𝑎 → 𝛾𝛾

• The signal will be proportional to the double probability:

Power build-up factors in production and reconversion cavities (*) • FOM proportional to signal-to-noise ratio:

Power of laser Detector noise

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 79 LSW experiments – current state-of-the-art

OSQAR @ CERN

Also: • GammeV & REAPR @ Fermilab, US • BMV @ Toulouse • …

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 80 From ALPS-I to ALPS-II

• Resonantly enhanced • L ~ 2 x 100 m (10+10 magnets) • Very close to commissioning! First data could come this year…

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 81 LSW at other wavelenghts

CROWS experiment @ CERN

• Microwave photons • Resonant implementation easier • Lose L enhancement…

• Large scale MW LSW studied and proposed in the literature (STAX exp) • LSW at X-rays also explored in the past (not large power)

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 82 Future LSW: JURA concept

• Concept being discussed at the Physics Beyond Colliders group at CERN • Piggybacking on development of future FCC dipole magnets. • L ~ 1km, B ~13T, P~2.5 MW,… Very challenging parameters… • Physics case to be settled (it may depend if positive signal in other exps)

A. Lindner, talk at PBC2019

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 83 Polarization experiments

PVLAS experiment: study QED vacuum birrefringence (standard effect), but also sensitivty to ALPs: Future project under discussion at PBC: VMB@CERN

Dichroism: Production of real particles

Ellipticity: Production of massive virtual particles

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 84 Laboratory experiments sensitivity

Also: • GammeV & REAPR @ Fermilab, US • BMV @ Toulouse • PVLAS @ Ferrara • CROWS @ CERN • …

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 85 Axion-mediated macroscopic forces

Axions could be detected as short-range deviation of gravity… (but traditionally though without enough sensitivity to QCD axions)

Recently proposed: ARIADNE experiment Short-range force by NMR technique Arvanitaki, Geraci Phys. Rev. Lett. 113, 161801 (2014) Good prospects for sub-meV axion

NMR sample Rotating source mass

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 86 Dark matter axions

a γ

Photon detector Axions from DM halo

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 88 Dark matter “axion field”

• If DM is (mostly) axions, they are very light, non-relativistic, very high number density  they can be better seen as a classical field For an observed DM density of oscillating at a frequency ~

𝑓𝑓𝐴𝐴 𝑚𝑚𝐴𝐴 The axion number density is • Energy of axions 1 + 10 . Small kinetic energy acquired by falling in the galactic potential− well.6 𝑚𝑚𝐴𝐴 𝑂𝑂 • Field coherent over lengths ~ de Broglie wavelength – The field can then be considered spatially constant over distances < 𝜆𝜆𝑐𝑐 𝜆𝜆𝑐𝑐 Axion DM field, non-relativistic

Oscillates at frequency ~

𝑓𝑓 𝑚𝑚𝐴𝐴 De Broglie wavelength

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 89 Dark matter “axion field”

• The velocity distribuition ( ) is given by galactic dynamics (falling into the potential well) • Standard Halo Model: ( ) is𝑓𝑓 a𝑣𝑣 𝐴𝐴Maxwellian distribution. • Other models predict other distributions. ( ) can also be 𝐴𝐴 extracted from DM simulations.𝑓𝑓 𝑣𝑣 More precise shape 𝐴𝐴 𝑓𝑓 𝑣𝑣 from DM sims

( ) determines spread Lentz et al 2017 in frequency ~10 𝑓𝑓 𝑣𝑣𝐴𝐴 Axion DM field, non-relativistic −6 𝑄𝑄𝐴𝐴

Oscillates at frequency ~

𝑓𝑓 𝑚𝑚𝐴𝐴

~10 −6 𝛿𝛿𝑓𝑓 𝑓𝑓

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 90 Detecting DM axions

If cavity tuned to the axion • Conversion of DM axions into photons in a B-field: too weak signal frequency, conversion is Energy of converted photon = (1 + 10 ) “boosted” by resonant factor • ( quality factor) Use of resonant cavity with resonant frequency−6 matching the axion • 𝛾𝛾 𝐴𝐴 mass: Axion haloscope concept,𝐸𝐸 Sikivie𝑚𝑚 , 1983𝑂𝑂 𝑄𝑄

Geometric factor: overlap of Axion DM field, non-relativistic cavity mode and external B-field Oscillates at frequency ~

𝑓𝑓 𝑚𝑚𝐴𝐴

Cavity dimensions smaller than 𝒆𝒆 𝜆𝜆𝑐𝑐 𝑩𝑩

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 91 Axion haloscopes

• Cavity geometry constrained by: – Maximizing geometric factor. E-field of the cavity mode (as much as possible) parallel to external B-field

– Data taking proceeds by scanning small ~ / mass steps and taking limited data a each step  need to mechanically modifyδ𝑓𝑓 𝑓𝑓 𝑄𝑄 the cavity geometry to change resonant frequency

• Figure of merit:

(proportional to “time needed to scan a given mass range”)

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 92 ADMX

• ADMX is the older axion haloscopes in - High cavity ~10 operation. Sited at U. of Washington - Large volume: 1m5 × 60cm𝑄𝑄 • Pioneer collaboration. Many years of R&D - Suited for resonance at few ×∅ eV - 8T superconducting • Low noise receivers based on SQUIDs + magnet 𝜇𝜇 dilution refrigeration at 100 mK. • Tuning achieved by set of movable rods

• Sensitivity to few µeV proven • Good support through Gen 2 DM US program

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 93 ADMX

Current program will surely • First results >10 years ago. • cover 1-10 µeV with high New data being released in the last few years • sensitivity (i.e. reaching even • First data down to DFSZ coupling… pessimistic coupling).

Bartram, talk at Patras2021 What about higher masses?

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 94 Higher-ma haloscopes

• Problematic: seeking a resonance at higher normally implies lower cavites  lower sensitivity 𝑚𝑚𝐴𝐴 𝑉𝑉 • One needs to overcome this by pushing other parameters of the FOM • Many development lines: Use of more powerful, Larger instrumented volumes: matching higher-B, magnets more cavities or new multicavity designs

Reducing noise: cryogenics, new detection techniques… but beware of the Larger can be obtained by • Many R&D lines: standard quantum limit superconducting cavities, but improvement𝑄𝑄 only up to < ~10 𝑇𝑇𝑆𝑆𝑆𝑆𝑆𝑆 6 𝑄𝑄 𝑄𝑄𝐴𝐴

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 95 Going for lower

SQUID amplifier from ADMX collaboration 𝑇𝑇𝑠𝑠𝑠𝑠𝑠𝑠

• Leading haloscopes use cryogenics to cool down well below 1 K. 𝑇𝑇𝑝𝑝ℎ𝑦𝑦𝑦𝑦 • But = + (contribution from the sensor technology). ≪ New technologies with lower are also needed. Quantum 𝑠𝑠𝑠𝑠𝑠𝑠 𝑝𝑝ℎ𝑦𝑦𝑦𝑦 𝑠𝑠𝑠𝑠𝑠𝑠 technology𝑇𝑇 𝑇𝑇 needed𝑇𝑇 to profit from at reach by dilution 𝑠𝑠𝑠𝑠𝑠𝑠 refrigerators (SQUID, JPA,…)𝑇𝑇 𝑇𝑇𝑝𝑝ℎ𝑦𝑦𝑦𝑦 • But Standard Quantum Limit (SQL): when approaches ~ , irreducible noise from vacuum fluctuations 𝑇𝑇𝑝𝑝ℎ𝑦𝑦𝑦𝑦 ℎ𝜈𝜈

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 97 Beyond the SQL: HAYSTAC Quantum information technologies for axion searches Squeezed photon state: one quadrature has lower variance, and the complementary one has higher-than-normal variance (to respect Heisenberg principle)

• If the cavity vacuum is prepared in a squeezed state, the axion signal can be read with a noise that is effectively lower than the SQL in an amount dependent of the degree of squeezing. Concept studied in Malnou et. al 2018 • First experimental application HAYSTAC at Yale (Backes et al 2021). Factor x2 effective faster scan rate obtained

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 98 Going for higher

HTS coated cavity NbTi coated cavity (RADES) (QUAX𝑄𝑄 experiment)

• Coating the walls of the cavity with a superconductor material will increase (decreases the losses on the walls) • Technological challenges: the material must remain SC in the 𝑄𝑄 magnetic field. HTS coated cavity (CAPP) • Novel high-T superconductor (HTS) are difficult to handle (brittle materials, etc.) • Physical limit to improving : above ~10 it does not translate in improved FOM (the axion peak 6must fit in the cavity bandwidth!) 𝑄𝑄 𝑄𝑄𝑎𝑎

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 99 CAPP

• Recently created “Center for Axion and • In addition, many R&D lines ongoing: Precision Physics” at South Korea – Ultrahigh field superconducting magnets • Main goal to “build a large axion DM – Superconducting films to get high G cavities experiment in Korea” – Low noise sensors (SQUIDs) – New cavity designs & multi-cavity phase locking • Recently release results from CAPP-PACE and schemes CAPP-8TB setups CAPP projected 5-y plan CAPP-8TB

CAPP-PACE

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 100 New cavity designs to increase V

• Combining the power of many smaller cavities is possible but ADMX R&D challenging -> “phase matching” – Probably only possible for “a few” cavities

• New cavity designs to “decouple” V from ma, and go for larger effective V and larger ma.

Long thin cavities: ma is fixed by small Properly designed arrays of cavities with couplings dimensions, but V can be ~larger

*Also ORGAN CAST-CAPP project RADES project Filter-like cavities project in in the CAST magnet “Pizza cavity” Autralia in the CAST magnet at CERN at CAPP

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 101 Magnetized dish antenna • DM field + B field + boundary condition in the dish  photon emission normal to surface Resonance versus area

• No resonance (loss a factor Q) BUT may be compensated with very large areas ?

Directionality

Implementation In the BRASS Project at Hamburg

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 103 Dielectric haloscopes • Effect of dish antenna “boosted” by the addition of many dielectric disks • Some “mild” resonance – Concept between a haloscope and a dish antenna – It needs tuning! (challenging) • Relevant sensitivity in the 10-4 eV ballpark for a ~m3 10T experiment (80 disks)

Similar high-ma structures: ORPHEUS and Electric-Tiger in US

MADMAX dielectric haloscope

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 104 Directional effects • DM directionality would be a powerful signature to confirm a putative signal • Long aspect-ratio cavities should show a directional dependence if L >

• Dish antennas: small parallel component𝑐𝑐 proportional to axion momentum𝜆𝜆 – pixelised detector at the focal point could image velocity distribution

An “axion astronomy” era would follow a discovery

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 105 Lower ma haloscopes • Large V haloscopes are technologically simpler, but expensive  huge magnet needed. • Use of existing magnets could be an effective strategy

3 KLASH proposal: use of 50 m , 0.6T, Future IAXO helioscope (see later) KLOE magnet at LNF will offer B2V > ~100 T2m3

Use of large toroidal magnets: ACTION proposal at CAPP

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 106 Pick-up coil & resonant circuit • DM-induced oscillating B in the center of a toroidal magnet • Resonance is achieved externally with a circuit (no cavity) • Both wideband search and resonance search possible

• Competitive at very low ma • ABRACADABRA at MIT • 10 cm prototype under preparation • Also DM Radio at Stanford • More recently BASE at CERN

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 107 Spin precession experiments • DM-induced spin precession  it can be detected with very sensitive NMR techniques • Directly sensitive to the gluon term (also to fermionic couplings)

• Maybe important at very low ma CASPEr experiment (Mainz-Berkeley)

• Also QUAX experiment: Phys. Rev. X 4, 021030 (2014) – Electron spin precession – Sensitive to “axion DM wind” through axion-electron coupling

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 108 DM-induced atomic transitions

• DM can induce atomic excitations equal to ma. • Sensitive to axion-electron and axion-nucleon coupling

• Zeeman effect  create atomic transitions tunable to ma • Detection of excitation via pump laser • AXIOMA  recent project aiming at an implementaiton

-4 Relevant sensitivity for ma ~ 10 eV seems possible for kg-sized samples

P. Sikivie PRL 113(14)

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 109 Pre-inflation models Post-inflation models > 1 = 1 = 1 𝑁𝑁𝐷𝐷𝐷𝐷 Finetunned 𝑁𝑁𝐷𝐷𝐷𝐷 𝑁𝑁𝐷𝐷𝐷𝐷 𝜃𝜃𝑖𝑖

• Summary of current status and future prospects… • A diverse experimental landscape has emerged with potential to cover a substantialll fraction of parameter space

• Caution: many of these prospects still rely on a prior succesful R&D phase • Caution: Green areas rely on axion as DM hypothesis…

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 110 Pre-inflation models Post-inflation models > 1 = 1 𝑁𝑁𝐷𝐷𝐷𝐷 Finetunned 𝑁𝑁𝐷𝐷𝐷𝐷 𝜃𝜃𝑖𝑖

• Summary of current status and future prospects… • A diverse experimental landscape has emerged with potential to cover a substantial fraction of parameter space

• Caution: many of these prospects still rely on a prior succesful R&D phase • Caution: Green areas rely on axion as DM hypothesis…

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 111 Solar axions

a γ

Photon detector

Axions from the Sun

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 112 Solar axions

• Solar axions produced by conversion of the solar plasma photons in the solar core. “Primakoff solar axions” 𝛾𝛾 → 𝑎𝑎

Solar axion flux at Earth

Spectral shape fully defined by solar physics

van Bibber PRD 39 (89) CAST JCAP 04(2007)010

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 113 Solar axions

• In addition to Primakoff, “ABC axions” may be x100 more intense… but model-dependent.

Primakoff

Non-hadronic “ABC” Solar axion flux at Earth JCAP 1312 008

“ABC” axions (*) * if the axion couples with the electron ( ) (non hadronic axion) 𝑔𝑔𝑎𝑎𝑎𝑎 Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 114 …and more solar axions

• Via axion-nucleon couplings: monochromatic lines from nuclear transitions: – E.g. 14.4 keV axions emitted in the M1 transition of Fe-57 nuclei, MeV axions from 7Li and D(p;γ)3He nuclear transitions or Tm-169.

• More recently proposed: photons interacting with macroscopic solar B-fields.

• longitudinal and transversal plasmons conversion. • Depend on ! TP depend on too • 𝒂𝒂𝜸𝜸 • Depend on solar𝒈𝒈 B-field. Poorly known but𝒎𝒎𝒂𝒂 constrained:

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 115 Axion helioscopes

Axion helioscope concept P. Sikivie, 1983 + K. van Bibber, G. Raffelt, et al. (1989) (use of buffer gas)

Coherence factor Photons keep the same momentum energy as incoming axions difference

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 116 Axion helioscopes

Figure of merit of axion helioscopes. Defined as proportional to SNR of a given signal:

Data taking efficiency and observation time Efficiencies of focalization and detection

Field strength, length and cross-sectional área of magnet Normalized background and signal spot size Photons keep the same energy as incoming axions

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 117 Buffer gas for higher

Coherence condition ( 1) is recovered for a narrow mass range around , the refractive 𝑎𝑎 mass acquired by the photon𝑞𝑞𝑞𝑞 ≪ in the gas 𝑚𝑚 𝑚𝑚𝛾𝛾

3 Ne: number of electrons/cm r: gas density (g/cm3)

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 118 Other types of helioscope • Instead of magnetic field, one can use as « target » the EM fields or electrons in large WIMP detectors: recent excess in XENON1T? Not sensitive to unconstrained models…

• ‘Primakoff-Bragg’ effect: +cristaline target • WIMP-like experiments provide limit to axions: SOLAX, COSME, DAMA, EDELWEISS, CDMS, etc… • Characteristical temporal pattern:

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 119 Excess in XENON1T E. Aprile et al., 2020

• The solar axion model has a 3.5 σ significance, but requires ~3 × 10 , … or a large coupling with photons plus a coupling with electrons (Gao et al. 2020) −12 𝑔𝑔𝑎𝑎𝑎𝑎 • Strong conflict with astrophysics: it cannot be solar axions

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 120 Other types of helioscope • « TPC in a magnetic field »: conversion and absorption happening in the gas • Competitive only for high axion mass ~0.1-10 eV • No coherence, but large volume can compensate. Also no preferred direction, so no tracking needed

Galán et al, arXiv:1508.03006

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 121 Axion Helioscopes • Previous helioscopes: – First implementation at Brookhaven (just few hours of data) [Lazarus et at. PRL 69 (92)] – TOKYO Helioscope (SUMICO): 2.3 m long 4 T magnet

Current state-of-the-art: CERN Axion Solar Telescope (CAST)

First helioscope using low background techniques & x-ray focusing

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 122 CAST experiment @ CERN

• Decommissioned LHC test magnet (L=10m, B=9 T) • 3 X rays detector prototypes used. • Moving platform ±8°V ±40°H (to allow up to 50 days / • X ray Focusing System to increase year of alignment) signal/noise ratio. • 4 magnet bores to look for X rays

Moving platform ±8°V ±40°H LHC test magnet

2 low background 1 Micromegas + Micromega XRT s 1 Ingrid + XRT

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 123 CAST hunting axions (movie credit: M. Rosu / CERN) Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 124 Latest CAST limit Enabled by the 2003 – CAST phase I IAXO pathfinder system 2004 • vacuum in the magnet bores

CAST phase II - 4He Run 2006 • axion masses explored up to 0.39 eV (160 P-steps) X-ray optics specifically built 2007 3He Gas system implementation for axions

CAST phase II - 3He Run 2008 - 2011 • axion masses explored up to 1.17 eV • bridging the dark matter limit Low background Micromegas •Revisit 4He Run with improved 2012 detectors

2013- •New vaccum phase with improved detectors 2015  Result released in 2017

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 125 IAXO pathfinder system in CAST Test MM detector + slumped-glass x- ray optics together

X-ray optics specifically built Detector: JCAP12 (2015) for axions Physics: Nature Physics Background spectrum (10.1038/nphys4109)

5 mm • Best SNR of any previous detector • 290 tracking hour acquired Low background (6.5 months operation) Micromegas at the focal point • 3 counts observed in RoI (1 Calibration data “axion” data expected)

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 126 IAXO – Concept

IAXO = International Axion Observatory

Enhanced axion helioscope: JCAP 1106:013,2011

4+ orders of magnitude better SNR that CAST

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 127 IAXO – Conceptual Design

• Large toroidal 8-coil magnet L = ~20 m

• 8 bores: 600 mm diameter each

• 8 x-ray telescopes + 8 detection systems

• Rotating platform with services

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 128 IAXO technologies – magnet IAXO magnet • Superconducting “detector” magnet. • Toriodal geometry (8 coils) • Based on ATLAS toroid technical solutions. • CERN+CEA expertise • 8 bores / 20 m long / 60 cm Ø per bore

Baseline developed at: IAXO Letter of Intent: CERN-SPSC-2013-022 IAXO Conceptual Design: JINST 9 (2014) T05002 (arXiv:1401.3233)

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 129 IAXO x-ray optics

• X-rays are focused by means of grazing angle reflection (usually 2) • Many techniques developed in the x-ray astronomy field. But usually costly due to exquisite imaging requirements

ABRIXAS spare telescope, in use in one of the 4 bores of CAST (pioneer use of x- ray optics in axion research)

Focal length

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 130 IAXO x-ray optics

• Each bore equipped with an x-ray optics • Exquisite imaging not required • BUT need cost-effective way to build 8 optics of 600 mm diameter each.

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 131 IAXO x-ray optics

• Technique of choice for IAXO: optics made of slumped glass substrates coated to enhance reflectivity in the energy regions for axions. • Same technique successfully used in NuSTAR mission, recently launched

• The specialized tooling to shape the substrates and NuSTAR optics assembly machine assemble the optics is available NuSTAR telescope ~400 mm Ø • Hardware can be easily configured to make optics with a variety of designs and sizes

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 132 IAXO x-ray optics

IAXO optics conceptual design AC Jakobsen et al, Proc. SPIE 8861 (2013)

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 133 IAXO low background detectors • 8 detector systems • Baseline: small Micromegas-TPC chambers: • Shielding • Radiopure components • Offline discrimination

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 134 IAXO low- Micromegas detectors History of background improvement of Micromegas detectors at CAST • Goal background level for IAXO: • 10-7 – 10-8 c keV-1 cm-2 s-𝑏𝑏1 • Already demonstrated: • ~8×10-7 c keV-1 cm-2 s-1 (in CAST 2014 result) • 10-7 c keV-1 cm-2 s-1 IAXO goals (underground at LSC)

Nominal values at CAST

• Active program of development. • IAXO-D0 test-platform to explore background sources and improve levels

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 135 Additional detector technologies for IAXO

Ingrid detectors (U. Bonn): MMC detectors (U. Heidelberg): • Micromegas on top of a CMOS chip • Extremely low threshold and (Timepix) energy resolution (~eV scale) • Very low threshold (tens of eV) • Low background capabilities under • Tested in CAST study

Typical X-ray event Single e- visible

Also: • Transition Edge Sensors (TES) • SDD- detetors

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 136 BabyIAXO

• Prototype: Intermediate experimental stage before IAXO • Two bores of dimensions similar to final IAXO bores  detection lines representative of final ones. • Magnet will test design options of final IAXO magnet • Test & improve all systems. Risk mitigation for full IAXO • Physics: will also produce relevant physics outcome (~100 times larger FOM than CAST)

Construction started. Commissioning ~100x CAST SNR expected for 2023-24

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 137 Helioscopes & astrophysics hints

… as well as a large fraction of the axion & IAXO will fully ALP models invoked in explore ALP models the “stellar cooling invoked to solve the anomaly” “transparency hint” But for this the gae is particularly interesting

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 138 IAXO & stellar cooling

• Multiple stellar anomalies (HB, RG, WD, NS,..). Overall 3σ effect.

gae channel

• IAXO will explore most of the relevant models (especially with IAXO+) Region of QCD axion M. Giannotti et al. models that solve the • Only experiment with JCAP 1710 (2017) 010 arXiv:1708.02111 stellar anomaly such capability

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 139 Overall conclusions

• Most compelling solution to the Strong CP problem of the SM

Les Houches DM - Axions Igor G. Irastorza / Universidad de Zaragoza 140 Overall picture (for )

• Helioscopes (IAXO) will 𝑎𝑎𝛾𝛾 • Helioscopes (IAXO) probe astrophisically 𝑔𝑔 will probe meV – eV motivated ALP models QCD axion models

• Haloscopes • … and most of the will soon probe region hinted by 1-10 µeV QCD stellar cooling axions

• Promising new haloscopes R&D to • In overall, a large substantially fraction of the ALP expand explorable mass range parameter space may be explored in the future

Les Houches DM school Igor G. Irastorza / Universidad de Zaragoza 141 Overall conclusions (II)

• Axion physics/search is an exciting field nowadays Theory Cosmology Cold DM Experimental and phenomenological efforts are Strong CP • problem candidate increasing axions – Consolidation of classical approaches into larger experiments + ALPs Dark new ingenious ideas to tackle new regions String radiation theory – Large impact still possible with relatively modest effort Inflation Astrophysics Dark Anomalous Energy? stellar cooling • Large fraction of parameter space at reach of near- + future experiments: chances of discovery! UHE γ transparency

I hope this course has been of your interest and will help you in your future career (be it in axion research or in neighboring topics)

Thank you very much for your attention

Les Houches DM - Axions Igor G. Irastorza / Universidad de Zaragoza 142