Polarimetry of Close Binary Stars: A New Probe for Axion Dark Matter
A method is proposed for searching for ultralight axion dark matter via polarization measurements of close binary stars. In such systems, reflected or scattered light acquires a small linear polarization, whose phase is tightly locked to the orbital motion. Axion-induced birefringence leads to oscillations in the polarization angle, which appear as sidebands around the orbital harmonics. For a single bright binary, assuming parameters typical of known systems and white noise, the projected sensitivity to the axion-photon coupling constant g_{a\gamma} is estimated at the level of 10⁻¹² GeV⁻¹ for an axion mass m_a = 10⁻²⁰ eV. Employing an array of suitable binaries in an optimistic scenario could improve the sensitivity to 10⁻¹³ GeV⁻¹. The method opens an additional high-frequency optical channel for probing axion birefringence, complementing existing astrophysical approaches.
Context
Dark matter remains one of the greatest unsolved mysteries of modern physics. Its existence was first convincingly demonstrated in the work of Vera Rubin through measurements of galaxy rotation curves. The most motivated models predict the existence of ultralight axions—pseudoscalar fields that can fill the halo as classical coherent waves. Direct detection of axions is challenging due to their extremely weak interaction with ordinary matter. However, the axion-photon interaction produces birefringence: in the presence of a slowly varying axion field, left and right circular polarizations of light acquire a phase difference, leading to an oscillating rotation of the linear polarization angle. This effect can be searched for using sources whose intrinsic polarization is well predicted. Previously, the cosmic microwave background, protoplanetary disks, and individual radio pulsars have been used. Close binary stars open a new opportunity thanks to their phase-locked polarization, arising from the reflection of light off the companion's atmosphere.
Methods
We propose using high-time-resolution polarimetric observations of close binary stars. The method relies on the reflection effect: radiation from the hot star scatters off free electrons in the atmosphere of the cooler companion (Thomson scattering). This produces a weak linearly polarized signal whose degree varies with orbital phase. Modeling based on the radiative transfer theory of Chandrasekhar allows construction of a precise phase template in terms of harmonics of the orbital frequency. In the presence of an axion field, sidebands appear at frequencies that are multiples of the orbital frequency plus or minus the axion mass. The analysis boils down to estimating the difference between observed and template polarizations using maximum likelihood. To overcome instrumental and astrophysical noise, we propose using both individual bright systems (e.g., hydrogen stars with temperatures above 10⁴ K) and arrays of multiple binaries, which can isolate a common terrestrial contribution from the axion field, akin to pulsar timing arrays discovered by Jocelyn Bell Burnell.
Results
For an individual binary like μ¹ Sco, with a 10-minute cadence, 30 days of observations, and polarimetric precision of 10 ppm, a sensitivity to the coupling constant g_{aγ} of about 2.4×10⁻¹² GeV⁻¹ is achievable at an axion mass μ ~ 10⁻²⁰ eV. The accessible mass range is determined by the Nyquist frequency and total observation time: from 10⁻²¹ to 10⁻¹⁸ eV. By combining 14 similar systems and improving precision to 1 ppm, statistical sensitivity increases to 1.3×10⁻¹³ GeV⁻¹. In both cases, the signal appears as a modulation of sidebands around orbital harmonics, with the ratio of signal amplitude to the original harmonic being independent of harmonic number, simplifying its extraction. These estimates assume white noise and a known phase template.
Implications
The proposed method expands the toolkit of astrophysical axion searches. Unlike observations of the cosmic microwave background, where global mapping is key, or pulsar timing arrays operating in the radio band, optical binary polarimetry enables monitoring with minute cadence, covering an axion mass range inaccessible to other methods. This is especially valuable for testing models predicting axion dark matter with masses around 10⁻²⁰ eV. Moreover, the technique is immune to Faraday rotation, characteristic of radio emission, simplifying data interpretation.
Future development
In the future, the development of polarimeters with precision on the order of 10⁻³–10⁻⁴ percent, combined with large telescopes, will allow the claimed sensitivity to be achieved. Of particular interest is the creation of specialized 'networks' of carefully selected binaries, as already realized in pulsar timing arrays for gravitational waves. It will be necessary to improve radiative transfer models for predicting polarization templates, accounting for absorption, multiple scattering, and atmospheric evolution. It is also important to study possible systematics in detail, such as variability caused by stellar pulsations or matter accretion.
Impact
The method will impact the astrophysics of compact objects, dark matter physics, and the advancement of high-precision optical polarimetry techniques.
Next steps
Near-term steps include testing predictions on real observations of known eclipsing binary systems using existing instruments such as HIPPI-2. A catalog of suitable candidates should also be created based on data from space missions and ground-based surveys using spectropolarimetry.
Key open problems
This approach is directly connected to efforts to solve the nature of dark matter and extend the Standard Model of particle physics. Detection of an axion signal would be the first direct proof of the existence of ultralight fields filling the halo, and would shed light on the strong CP problem.
🎯 Interestingly, the very idea of phase-locked polarization in binaries was theoretically predicted back in the 1970s, but only recently have advances in instrumental precision allowed it to be reliably measured. For example, in the star Spica, the degree of polarization is only 0.02%, equivalent to observing a dim candle at a distance of a kilometer through a foggy glass.
🎬 Using polarized light to detect elusive fields is vaguely reminiscent of the attempts of the heroes of Stanisław Lem's 'Solaris' to decipher the signals of a sentient ocean. Only in our case, the 'ocean' is dark matter, and the 'signals' are microscopic oscillations of the polarization angle.
Key numbers
- sensitivity for a single binary (g_{aγ}): 2.4×10^-12 GeV^-1
- sensitivity for an array of 14 binaries: 1.3×10^-13 GeV^-1
- observation cadence: 10 minutes
- campaign duration: 30 days
- typical degree of polarization: 300 ppm