UC San Diego's cross-check reproduces the standard calibration pattern across eight Planck maps

The trouble is a degeneracy. From the CMB signal alone, a uniform rotation of the polarization and a common error in how the detectors are oriented produce exactly the same observed effect ^[3]. The hinted rotation is small, a fraction of a degree built up over the nearly 14 billion years the light has traveled ^[2], and detectors miscalibrated by a similarly small angle would produce the same rotation ^[3]. If the rotation is confirmed, cosmic birefringence could be evidence for physics beyond the Standard Model and a clue to dark matter and dark energy ^[1].

The UC San Diego design uses the one difference between the two explanations that shows up across instruments. A rotation imposed by the universe would appear in every map, whichever detectors made it. Compare maps from different detector groups and that shared rotation cancels, leaving only the differences in how each group was calibrated ^[5]. I like this piece of experimental logic. It swaps a question the data cannot answer for one it can: how far are Planck's detector sets from one another?

In my view, the most useful result is the match with the Minami-Komatsu analysis. The two methods rely on different assumptions ^[6]. For both to return the same calibration pattern on eight maps, a flaw in one method's assumptions would have to be copied by a flaw in the other's ^[6].

The thing this doesn't tell you is whether the twist is real. The estimator is deliberately insensitive to any rotation common to all maps ^[8]. A cosmic rotation is common to all maps, and so is a miscalibration that every detector shares. The method can measure how well the detector sets are calibrated relative to one another. It cannot by itself establish the overall polarization angle or the absolute birefringence signal, and that still takes an independent absolute calibration reference ^[8].

The 0.37-degree figure needs the same care. As a conditional demonstration, the team anchored its differential reconstruction to the common calibration mode inferred by the existing analysis and got 0.37 plus or minus 0.12 degrees, consistent with Minami-Komatsu ^[9]. Read as a detection, that sits about 3.1 standard deviations from zero ^[1]. Its absolute angle came from the existing analysis, though. The number shows the two methods fit together; it cannot count as an independent measurement of the twist ^[9].

Lonappan, a postdoctoral fellow who led the study with Brian Keating and Kam Arnold ^[4], described the work as a check. "The signal we are looking for is incredibly small, so we have to be certain that we are seeing the universe and not our instrument," Lonappan said ^[10]. "Before interpreting a tiny rotation as new physics, we want to know that the calibration itself can be trusted." ^[10]

The same check has a second use. Gravitational lensing has produced B modes, the curl-like polarization patterns, that have already been observed. The extra B-mode signal that primordial gravitational waves could produce has not been detected ^[11]. According to the phys.org account, the method could support those primordial searches by identifying relative angle-calibration errors ^[12].