CAST’s first decade of solar-axion research

31 March 2010

The ingenious project based on X-ray astronomy, particle physics and an LHC magnet.

In 1983, when I was thinking about how axions may be produced and detected by their conversion to photons in a magnetic field, it struck me suddenly that there is no need to produce axions because the Sun does that for us. The solar axion flux is much larger than any that we could produce on Earth, and it is here free of charge. Our job is simply to detect these solar axions.

– Pierre Sikivie of University of Florida.


Axions are one of the favoured candidates for the mysterious dark matter created in the early universe. A variety of observatories located on Earth and in outer space form a quasi-network that can target specific places in the search for these particles, such as the galactic centre, the inner Earth and the Sun’s hot core. The CERN Axion Solar Telescope (CAST) points at the Sun – its aim being the direct detection of axions or other exotic particles with similar properties.

While relic axions from the early universe should propagate with a velocity of about one thousandth of the speed of light, solar axions – with a broad spectral shape of around 4–5 keV kinetic energy – are relativistic. The open window for the axion rest mass is currently in the micro-electron-volt to electron-volt range. The several orders of magnitude difference in kinetic energy associated with the two origins make for different experimental search techniques: microwave cavities for relic axions versus X-ray detectors for solar axions. However, both techniques use a magnetic field as the catalyst that allows axions to become photons

Accelerator laboratories, with their powerful magnets are natural locations for axion helioscopes – the instruments used to search for axions from the Sun. The first experiment to look at the Sun, which incorporated a 2.2-m iron-core magnet, was set up by a Rochester-Brookhaven-Fermilab (RBF) collaboration in the early 1990s. It was followed by the Sumico experiment based on a 2.3-m long superconducting magnet at the University of Tokyo, which is still in operation. The CAST helioscope at CERN uses a decommissioned LHC-dipole test magnet, with a field of 9 T and two tubes – originally designed to house the beam pipes – that are 9.2 m long and have an aperture of 43 mm. The dipole is one of four original prototypes and was rescued at the last minute before it was about to be scrapped along with the others. A comparison of CAST’s performance with its two predecessors in Brookhaven and Tokyo shows that the LHC magnet was good choice.

The possibility that a bending magnet could be used to make visible the “dark” Sun was – and still is – inspiring and motivating. To transform the multi-tonne superconducting, superfluid-helium-cooled magnet from a static LHC prototype dipole into a helioscope that can track the Sun with millimetre precision involved delicate engineering work and cryo-expertise. Thankfully, Louis Walckiers in the Accelerator Technology Division supported the idea, even though we had both just failed to prove with the same magnet that the biomechanics of cell-structure formation becomes confused in a 9 T environment.

Recycling space technology


Position-sensitive X-ray detectors of the MicroMegas type, invented by Georges Charpak and Ioannis Giomataris at CERN, now cover three of the ends of the tubes through the magnet, making CAST the only axion helioscope to have implemented such technology. For the fourth exit, together with Dieter Hoffmann and Joachim Jacoby of TU Darmstadt we were able to recover an excellent X-ray imaging telescope from the German space programme, which was delivered by Heinrich Bräuninger from the Max Planck Institute for Extraterrestrial Physics in Garching. With state-of-the-art X-ray optics and low-noise X-ray pixel detectors at the focal plane, this not only improves the signal-to-noise ratio substantially but also allows for the unambiguous identification of the axion signal. Its CCD imaging camera simultaneously measures the expected solar-axion signal spot and the surrounding background. This is an important feature that makes CAST unique as an axion helioscope. With most of the components located, CAST received formal approval at CERN in April 2000.

In the same way that much of the CAST equipment was recycled from particle physics so, too, was its working principle: the Primakoff effect, known since 1951, which regards the production of neutral pions by the interaction of high-energy photons with the high electric field of the nucleus as the reverse of the decay into two photons. The expectation is that the quasi-stable axion should “decay” in the presence of a magnetic field into a photon emitted exactly along the axion’s trajectory. In principle this allows for a perfect axion telescope thanks to the spatial resolution of the X-ray telescope.

The Primakoff effect deserves to be a textbook example of macroscopic quantum-mechanical coherence, which, in astrophysical magnetic fields, can extend over kiloparsecs – although only for very small axion rest masses. For CAST, coherence holds over the whole length of the magnet, around 9 m, provided that the particle rest mass is below 0.02 eV/c2 when the two pipes are vacuum-pumped. To extend the detection sensitivity to higher masses, adding a certain amount of helium as a refractive gas to the 1.8 K cold magnetic pipes restores coherence for a rest mass up to around 1 eV/c2 from a few millimetres up to 9 m but for a narrow range in solar axion rest mass. With this adaptation, suggested in 1988 by two collaboration members Karl van Bibber and Georg Raffelt, and implemented during 2005 and 2006, CAST has become a scanning experiment. The rest-mass range for solar axions that will be scanned by the end of 2010 fits the cosmologically derived upper limit of about 1 eV/c2, from the Wilkinson Microwave Anisotropy Probe (WMAP) data, and the lower limit around 1 μeV/c2, which arises because axions with lower rest mass would be produced earlier in the early universe, with a total mass exceeding that of the critical density (“overclosure”).


The precise pressure settings for the helium gas and controlled changes in the very cold magnet pipes are highly demanding and are not without risk. CAST has benefited greatly from CERN’s world-class cryogenic expertise in this respect, with its reliable user-friendly gas system designed by Tapio Niinikoski and his PhD student Nuno Elias. At present an extensive thermodynamic simulation is being performed with the aim of reconstructing the changing conditions of the helium gas as the magnet tracks the Sun. For example, to achieve the homogeneity in gas density necessary to keep coherence, the temperature variations along the 9-m long pipes should be in the milli-kelvin range; this is made possible by the surrounding bath of superfluid liquid helium at about 1.8 K.

CAST is also a “special” experiment when compared with others because its highly sensitive magnet and low-background detectors must operate while in motion, even though the speed of about 2 m an hour is almost imperceptible. In addition, CAST’s equipment must withstand quenches of the superconducting magnet. After each quench the gas control system must cope with extreme conditions within seconds. However, during 15,000 hours of operation with the magnet on, and more than 2000 hours of solar tracking, CAST has survived potentially catastrophic events because its safety features have – thanks to the careful work of CERN’s Martyn Davenport – never failed simultaneously.

Scientific return


While CAST has failed so far to find direct evidence for solar axions, it has been able to provide new robust limits on the interaction of solar axions with a magnetic field, i.e. the sea of virtual photons (figure 1). Its experimentally derived limit dominates the relevant phase space and competes with the best astrophysically derived lower value for the coupling constant, g. CAST is now moving into a theoretically motivated region, having almost fulfilled the original expectations set a decade ago with all of the input uncertainties at that time.

Moving beyond the initial proposal, CAST has in parallel explored – for the first time for a solar axion search – the region of high-energy solar axions, following the proposal of collaboration member Juan Collar. It has also made the first measurements below 1 keV, covering so far the range of around 1–3 eV. Moving to energies above this is possible; however, it will require larger energy steps and some new state-of-the-art detector technology to explore this interesting energy region that covers of most Sun’s puzzling X-ray activity.

Without detecting any solar-axion signature so far, the question arises: what is the scientific return from CAST? Certainly, the first benefit is educational, with students completing some 10 PhD theses and an equal number of diploma theses. There have also been several CAST summer students at CERN. On the research side, CAST has helped to revive axion activities around the world, fitting between pure axion searches in the laboratory and a variety of astrophysical/cosmological observatories that usually did not have axions in their original list of objectives. The state-of-the-art detectors in these observatories cover photon energies from micro-electron-volts upwards. With CAST, the implementation of X-ray optics in axion helioscopy has become widely accepted as a necessary ingredient for future scaled-up versions.

While CAST’s results have became a reference in the relevant field, they have also been used by other teams to search, for example, for “paraphotons” – sterile massive photons from the “hidden sector”. Furthermore, two members of the CAST collaboration, Milica Krĉmar and Biljana Lakić, have used the experiment’s results to explore theories of large extra dimensions, which predict “massive” axions of the Kaluza-Klein type. Interestingly, such massive exotica could be gravitationally trapped in the Sun and could build a bright halo, as a result of their spontaneous decay, as we have suggested with Luigi Di Lella of CERN.

The axion signal that the CAST collaboration aims to observe while tracking the Sun consists of excess X-rays emerging from the magnet tubes. Interestingly, there is abundant solar X-ray emission of otherwise unknown origin, which is further enhanced just above the magnetized photosphere. For more than 70 years, known physics has failed to explain this intriguing behaviour, which could, however, arise from the conversion or decay of axions or other similar exotica near the Sun’s restless surface. The outermost solar layers, i.e. the photosphere, might act occasionally as scaled-up and highly effective catalysts of axions or similar particles, emitting large numbers of X-rays (like a fine-tuned CAST might do one day). Then, extending Sikivie’s original idea, the otherwise mysterious solar surface makes these axions visible as X-rays. New X-ray observatories in space are already providing more and more exciting evidence that something new and interesting is going on in the Sun’s outer layers. The complete axion scheme may make the Sun even more special than it already is.

Such a solar scenario might eventually point to a “superCAST”, which in 5 to 10 years may well make the present CAST look like an old fashioned miniature device – provided that Sikivie’s pioneering idea behind CAST is not replaced by a novel conceptual design. For example, together with Andrzej Siemko of CERN we have proposed using a quadrupole magnet as a potentially better axion catalyst than the dipole magnets used at present in almost all axion experiments. This idea, which was also discussed theoretically by Eduardo Guendelman in 2008, is motivated observationally because otherwise puzzling solar X-ray activity correlates not only with magnetic fields but even more with places of varying field vector.

Alvaro De Rújula commented in 1998 that “axion searches are mandatory, fun, creative – and proceeding”. His words are just as true today, as the CAST project continues into its second decade.

• I am very grateful to all members of the CAST collaboration, to CERN for its hospitality and support, including the librarians, and to my colleagues at the University of Patras for their real help.

This article is dedicated to the memory of the following members of the CAST collaboration who have sadly passed away since the project’s inception: Engin Abat, Engin Arik, Fatma Senel Boydag, Ozgen Berkol Dogan, Angel Morales and Julio Morales.

Further reading

For the latest news on axions, see and

bright-rec iop pub iop-science physcis connect