X-Ray Polarization from the Gamma-Ray Binary LS I +61 303

Abstract: The gamma-ray emitting binary stellar system LS I +61 303 was observed with the Imaging X-ray Polarimetry Explorer (IXPE) on two successive orbits over orbital phases of 0.74 to 1.05. Polarization is detected at a significance of 4.2$σ$ with an average polarization degree of $13.1\% \pm 3.0\%$ in the 2-8~keV band after background subtraction. This is the second detection of polarization of the X-ray synchrotron emission from a gamma-ray binary and, again, suggests that the magnetic field in the particle acceleration region has a significant ordered component. The orbital motion on the sky of LS I +61 303 is not well determined, which leads to ambiguity in interpretation of the X-ray electric vector polarization angle (EVPA) measurement. Use of orbital elements determined via radial velocity measurements combined with radio imaging of variable nebular emission, suggests an offset between the X-ray EVPA and the compact object-massive star axis on the order of ~30$<sup>{\circ}$. Such an offset could be produced by Coriolis forces due to binary motion. Use of two different sets orbital elements determined via optical polarimetry suggest either no offset or a perpendicular orientation, but require an unexpectedly high inclination. Use of orbital elements derived from modeling of the keV/TeV light curves suggest good alignment between the X-ray EVPA and the compact object-massive star axis. Such alignment was found for the gamma-ray binary PSR B1259-63. If the same physical situation holds for LS I +61 303, that would favor the orbital elements derived from the keV/TeV light curves.

• The paper reports the second X-ray polarization detection from a gamma-ray binary, measuring a background-corrected polarization degree of 13.1% ± 3.0% at 4.2σ significance with an EVPA of −23.8° ± 6.6° using IXPE observations.
• The measured polarization implies that roughly 20% of the intrabinary-shock magnetic-field energy is ordered, although geometric depolarization and shock curvature mean this estimate is a lower bound.
• The EVPA aligns with the pulsar–primary axis under some orbital solutions, particularly those of Chen et al., while radial-velocity-based orbits imply an approximately 30° offset potentially caused by Coriolis-deflected shocked flows.

The Imaging X-ray Polarimetry Explorer (IXPE) observations of the gamma-ray binary LS I +61°303 (hereafter LS61) reported by Kaaret et al. (2605.15972) provide the second detection of X-ray polarization from a gamma-ray binary, following the earlier measurement of PSR B1259–63. The measurement constrains both the degree of magnetic field ordering in the particle acceleration region and, more ambiguously, the orientation of that region relative to the binary axis.

Observations and polarization results

IXPE observed LS61 in three segments between 14 February and 19 March 2026, covering orbital phases 0.74–1.05 over two successive orbits, with roughly 714 ks exposure per detector unit. Source counts were extracted from a 60″ circular region with background from a concentric annulus; the source spectrum lies above the scaled background across the full 2–8 keV band.

The model-independent analysis yields an observed polarization degree (PD) of $11.0\% \pm 2.6\%$ at $4.2\sigma$ significance and an electric vector position angle (EVPA) of $-24\fdg1 \pm 6\fdg8$. After background subtraction using the additive Stokes parameters, the source PD is $13.1\% \pm 3.0\%$ with EVPA $-23\fdg8 \pm 6\fdg6$, from 33,284 source counts of which only ~4,258 are background. A spectropolarimetric fit with an absorbed power law and constant polarization gives consistent values ($\rm PD = 11.4\% \pm 2.8\%$, EVPA $-24\fdg1 \pm 7\fdg1$) with $\chi^2/\mathrm{DoF} = 238.6/240$ and photon index $\Gamma = 1.46 \pm 0.03$, within the previously measured range. No significant variation in PD or EVPA was found between the two orbits or between flux-selected intervals, despite the notably higher flux during the second orbit.

Magnetic field ordering

For synchrotron radiation, the observed PD relative to the maximum (~70% for this spectral index) measures the fraction of magnetic field energy in a uniform component. The measured PD of ~13% implies that roughly 20% of the field is ordered, which the authors emphasize is a lower bound: depolarization can also arise if the shock-cone axis is not perpendicular to the line of sight or from shock curvature. This confirms, as for PSR B1259–63, that the acceleration region contains a significant ordered magnetic component rather than a purely turbulent field.

Orbital geometry ambiguity

Interpreting the EVPA requires knowledge of the pulsar's position angle on the sky during the observation, but unlike PSR B1259–63—whose orbit is precisely determined by radio timing—LS61 has no direct astrometric measurement of the compact object's motion. The paper therefore compares the EVPA against pulsar position angles computed from three classes of orbital solutions:

Solution	Method	Offset from EVPA	Caveat
Casares / Aragona ($i=75^\circ$)	Radial velocity + VLBI radio ellipse	$4.2\sigma$0–$4.2\sigma$1	Inclination assumed; eccentricity varies among solutions
Kravtsov K20B ($4.2\sigma$2)	Optical polarimetry, $4.2\sigma$3 free	$4.2\sigma$4	Inclination likely biased high by noise
Kravtsov K20A ($4.2\sigma$5 constrained)	Optical polarimetry	Perpendicular within $4.2\sigma$6	Same inclination concern
Chen et al. ($4.2\sigma$7)	keV/TeV light-curve modeling + VLBI	Aligned (median PA $4.2\sigma$8)	Light curves carry no geometric information

The offsets of ~30° under the radial-velocity solutions exceed the EVPA uncertainty and would imply that the shock acceleration region is misaligned with the pulsar–primary axis, plausibly through Coriolis deflection of the shocked flows as seen in the hydrodynamical simulations of Bosch-Ramon et al. Conversely, the Chen orbital elements give good alignment, matching the geometry established for PSR B1259–63 where the X-ray EVPA aligned with the pulsar position angle and the dominant magnetic field lies perpendicular to the shock cone axis. If that physical situation holds for LS61, the IXPE result favors the Chen elements—an unusual instance in which X-ray polarimetry could adjudicate among competing orbital solutions.

Limitations and open questions

The interpretation rests on several assumptions stated plainly in the paper. The inclination of $4.2\sigma$9 used for radial-velocity and light-curve solutions follows from an assumed neutron star mass of $-24\fdg1 \pm 6\fdg8$0 and B0Ve mass of $-24\fdg1 \pm 6\fdg8$1; it ranges from $-24\fdg1 \pm 6\fdg8$2 to $-24\fdg1 \pm 6\fdg8$3 over allowed stellar masses, and the offset grows at lower inclinations (reaching $-24\fdg1 \pm 6\fdg8$4 for the Chen solution at $-24\fdg1 \pm 6\fdg8$5). The identification of the radio peak ellipse with the true orbit, needed to derive $-24\fdg1 \pm 6\fdg8$6, is itself an assumption given that the VLBI ellipse (semimajor axis ≈1.6 mas) far exceeds the optical orbit (~0.2 mas). The polarimetrically derived inclinations are biased toward high values in the presence of stochastic intrinsic polarization variations, so the apparent alignments from the Kravtsov solutions may be spurious. Finally, the FAST pulsar detection remains not fully conclusive because no Doppler shift was observed and the large field of view cannot exclude an interloper, though the authors adopt the neutron star interpretation on the weight of evidence. Whether the EVPA–pulsar-axis alignment found for PSR B1259–63 genuinely holds for LS61—and hence which orbital solution is correct—remains open.

Conclusion

This work establishes LS61 as the second gamma-ray binary with detected X-ray synchrotron polarization, at PD $-24\fdg1 \pm 6\fdg8$7 ($-24\fdg1 \pm 6\fdg8$8), demonstrating a substantially ordered magnetic field in the intrabinary shock. The EVPA measurement is consistent with alignment to the pulsar–primary axis only under specific orbital solutions, most naturally those of Chen et al., while radial-velocity-based orbits imply a ~30° offset attributable to Coriolis-driven shock asymmetry. Resolving this degeneracy—whether through improved astrometry of the pulsar or phase-resolved polarimetry—would allow the X-ray EVPA to serve as an independent probe of both the true orbit and the wind-interaction dynamics.