Dark Energy May Leave a Twist in the Oldest Light

On the Tibetan Plateau, at an altitude of about 5,250 meters, a telescope called AliCPT is learning to read the faint polarization of the cosmic microwave background (CMB), the oldest linearly polarized light in the universe. In a preprint posted to arXiv on November 6, 2025, Hua Zhai and colleagues at the Institute of High Energy Physics, Chinese Academy of Sciences, and collaborators ask whether dark energy might leave a mark on that light: a slight rotation of its polarization plane.

The paper is not a detection. It is a forecast—a careful calculation of what AliCPT could see if dark energy interacts with photons through a Chern-Simons coupling, a proposed link between a dark-energy scalar field and electromagnetism. The payoff, if it works, would be a new way to test whether dark energy changes over cosmic time.

An Old Light with a Slight Twist

The CMB is linearly polarized. Its polarization patterns are usually split into E-modes, which look like gradients, and B-modes, which look like swirls. In the standard cosmological model, certain correlations are expected to vanish: the TB spectrum (temperature with B-mode polarization) and the EB spectrum (E-mode with B-mode polarization).

Cosmic birefringence can make them nonzero. If dark energy couples to photons via a Chern-Simons term, the polarization plane of CMB light rotates as it travels. The rotation angle depends on how the dark-energy field changes along the line of sight. A uniform rotation mixes E-modes into B-modes and creates TB and EB correlations. A patchy rotation instead makes the rotation direction-dependent and couples different angular scales.

From DESI’s Surprise to a Photon’s Rotation

The motivation comes from recent Dark Energy Spectroscopic Instrument (DESI) results, which support a dynamical dark energy. DESI data favor an equation of state—the ratio of pressure to energy density—that changes with time and crosses w = -1, the value for a cosmological constant. That behavior is a hallmark of the so-called Quintom scenario. If dark energy is a field rather than a rigid constant, it is natural to ask whether it interacts with ordinary matter.

The paper’s answer is to look for that interaction in the CMB. Previous analyses of Planck, WMAP, and ACT data have reported a small isotropic rotation angle, but the signal is tangled with an instrumental effect: a miscalibrated detector polarization angle can mimic a cosmic rotation. The CMB is affected by the sum of the true rotation and the miscalibration, while Galactic foreground emission is affected only by the miscalibration. That difference lets the team separate the two.

What AliCPT Could Detect

AliCPT-1 observes at 95 and 150 GHz with a 72-centimeter aperture and saw first light in early 2025. Its wide survey covers about 50% of the sky; after masking ±5° around the Galactic plane, the forecast uses about 44%. The team combined AliCPT-1 with Planck HFI and LFI data and calculated how precisely the rotation angle β could be measured.

Taking β = 0.35°, the best-fit value from Planck 2018 polarization data, they project that 11 module-years of AliCPT-1 observations plus Planck HFI data could detect the isotropic rotation at about 5σ, provided the polarization-angle calibration uncertainty is 0.1°. A module-year means one detector module observing for one season. AliCPT-1 alone would need about 18 module-years for the same significance.

The result depends heavily on calibration. If calibration uncertainty is larger than about 0.02°, improving it tightens the measurement. If calibration becomes extremely precise, the statistical noise of the CMB data itself sets the floor. Adding more well-calibrated frequency channels also helps.

The Difference a Wide View Makes

The paper also forecasts a more ambitious measurement: spatial fluctuations in the rotation angle. A patchy dark-energy field would rotate polarization by different amounts in different directions, breaking statistical isotropy and coupling different angular scales. This anisotropic rotation can be reconstructed with a technique analogous to CMB lensing reconstruction.

No experiment has yet detected this anisotropic signal. The current best 95% upper bound on its scale-invariant amplitude, A_CB, is 0.044 from BICEP/Keck. The new forecast finds that a possible 6-meter large-aperture version of AliCPT, called AliCPT-LAT, would reach σ_A_CB ≈ 9 × 10^-3 after 50 module-years. That is about six times tighter than the small-aperture AliCPT-1 design and would allow competitive tests of spatial fluctuations in the dark-energy field.

Caveats in a Forecast

These numbers are projections, not measurements. They assume the true rotation is 0.35°, a value that itself depends on how Galactic foregrounds are modeled. Planck and WMAP reported a 3.6σ result of β = 0.342° +0.094°/−0.091°, while ACT reported β = 0.215° ± 0.074°. The paper neglects some possible biases, including intrinsic EB correlations from foregrounds, and focuses on statistical uncertainty.

The anisotropic forecast is even more exploratory. AliCPT-LAT is a proposed large-aperture configuration, not the baseline instrument. The reconstruction assumes realistic noise and sky coverage, but the signal has not been seen. The path to a discovery runs through better calibration, more frequency channels, years of data, and—for spatial fluctuations—a large-aperture configuration.

The Cosmic Stake

If the effect is real, the CMB’s faint twist would be more than a curiosity. It would turn dark energy from a gravitational mystery into a field that interacts with ordinary photons. It would also give cosmologists a new microphysical probe of the dark-energy field, complementing surveys that map how dark energy stretches the expansion of the universe.

For now, the oldest light still holds its secrets. But as AliCPT and other telescopes accumulate data, they may begin to notice a subtle rotation—a whisper from dark energy, written across the sky.