XRISM reveals sloshing-driven gas motions in the core of Abell 2029

Abstract: We investigate the velocity structure of the intracluster medium (ICM) in the core of the relaxed cool-core cluster Abell 2029 using XRISM Resolve spectroscopy. We analyze combined XRISM Resolve observations and divide the central region into several subregions. To account for photon mixing caused by the XRISM point spread function, we perform a spatial-spectral mixing analysis. We detect an ordered line-of-sight bulk-velocity gradient across the cluster core: the northern regions are blueshifted relative to the brightest cluster galaxy (BCG), while the southern regions are close to zero velocity or slightly redshifted. The maximum velocity difference is about $280~{\rm km\,s<sup>{-1}}$. In contrast, the turbulent velocity dispersion is smaller, with measured values and upper limits of $\lesssim150~{\rm km\,s<sup>{-1}}$, implying a non-thermal pressure fraction below $\sim2.5\%$. The velocity pattern is consistent with gas sloshing associated with the spiral structure seen in Chandra X-ray images. Averaged over all regions, the inferred turbulent heating rate is below the radiative cooling rate, indicating that turbulent dissipation alone is insufficient to offset cooling in the entire core. These results reveal that A2029 is not kinematically featureless: sloshing-induced bulk motions are present, while the observed line-of-sight velocity dispersion indicates only a limited contribution to pressure support and core heating.

• The paper detects a statistically significant north–south bulk-velocity gradient of 193 ± 17 km s⁻¹, reaching about 280 km s⁻¹ across Abell 2029’s core, using spatially resolved XRISM Resolve spectroscopy with full PSF mixing corrections.
• The measured subsonic gas motions, with velocity dispersions below roughly 150 km s⁻¹, indicate that sloshing coexists with an otherwise near-hydrostatic atmosphere and contributes less than about 2.5% non-thermal pressure support.
• The analysis finds turbulent dissipation supplies at most about 20% of the radiative cooling rate at the assumed driving scales, leaving AGN-scale motions, mixing, and the three-dimensional sloshing geometry as key questions for future observations.

Overview

This paper presents spatially resolved X-ray microcalorimeter spectroscopy of the intracluster medium (ICM) in the core of the relaxed cool-core cluster Abell 2029 ($z=0.0787$), using combined XRISM Resolve observations from the performance verification (PV) phase and a subsequent General Observer program. The authors divide the central region into six subregions and apply a spatial-spectral mixing (SSM) analysis to correct for photon redistribution by the broad XRISM point spread function, whose half-power diameter of 1.3 arcmin is comparable to the Resolve field of view. The principal result is a statistically robust north–south line-of-sight bulk-velocity gradient across the core—northern regions blueshifted relative to the brightest cluster galaxy (BCG), southern regions near zero or slightly redshifted—with a maximum velocity contrast of approximately $280~{\rm km\,s^{-1}}$, while turbulent velocity dispersions remain below about $150~{\rm km\,s^{-1}}$ in all regions.

The scientific motivation is direct: deep Chandra imaging reveals a large-scale spiral surface-brightness structure extending to roughly 600 kpc, interpreted as the relic of an off-axis minor merger (Watson et al., 31 Oct 2025). If that spiral is produced by gas sloshing, coherent velocity gradients should be detectable even in a cluster whose globally averaged kinematics are quiescent. The Resolve measurements provide exactly this test, connecting the morphological evidence for sloshing to a measured velocity field for the first time in A2029.

Observations and spatial-spectral mixing methodology

The data set comprises three PV pointings from January 2024 (two central observations totaling ~42 ks and one northern offset pointing, N1, with 90.6 ks) and one GO central observation from July 2025 with 118.3 ks. Spectra were extracted using high-primary-grade events only, excluding the calibration pixel and pixel 27 due to gain variations. Six extraction regions were defined: center, northeast (NE), northwest (NW), southeast (SE), southwest (SW), and far northeast (FNE). Spectra were fitted above 2 keV with TBabs$\times$bapec models in XSPEC, minimizing the C-statistic, with non-X-ray background modeled from the archival night-Earth database.

Because PSF spillover between regions is significant, each observed spectrum was modeled as a linear combination of emission from all six source regions weighted by region-to-region ancillary response files generated with xaarfgen from a Chandra image; 88 ARFs encode all relevant source–extraction pairs. All spectra were fitted simultaneously via XSPEC's multiple-response functionality. This treatment is essential: without it, thermodynamic parameters would be biased by photons scattered from neighboring regions, though the paper shows the bulk-velocity pattern survives even without SSM correction.

Detection of a coherent velocity gradient

The headline measurement is summarized as follows:

Region	$kT$ (keV)	$v_{\rm bulk}$ (km s$^{-1}$)	$\sigma_v$ (km s$^{-1}$)
FNE	$6.52$	$280~{\rm km\,s^{-1}}$0	$280~{\rm km\,s^{-1}}$1
NE	$280~{\rm km\,s^{-1}}$2	$280~{\rm km\,s^{-1}}$3	$280~{\rm km\,s^{-1}}$4
NW	$280~{\rm km\,s^{-1}}$5	$280~{\rm km\,s^{-1}}$6	$280~{\rm km\,s^{-1}}$7
Center	$280~{\rm km\,s^{-1}}$8	$280~{\rm km\,s^{-1}}$9	$150~{\rm km\,s^{-1}}$0
SE	$150~{\rm km\,s^{-1}}$1	$150~{\rm km\,s^{-1}}$2	$150~{\rm km\,s^{-1}}$3
SW	$150~{\rm km\,s^{-1}}$4	$150~{\rm km\,s^{-1}}$5	$150~{\rm km\,s^{-1}}$6

The inverse-variance weighted mean velocities are $150~{\rm km\,s^{-1}}$7 for the northern regions and $150~{\rm km\,s^{-1}}$8 for the southern regions, giving $150~{\rm km\,s^{-1}}$9 at $\times$0 significance. Even after adding the Resolve energy-scale systematic uncertainty (~15 km s$\times$1 at 6 keV) in quadrature, the significance remains $\times$2. The sign reversal across the core is difficult to attribute to a uniform calibration offset, and it is recovered in independent fits without SSM correction, confirming that it is not an artifact of the mixing analysis.

The implication is immediate: A2029, often treated as a dynamically relaxed archetype, is not kinematically featureless. The gradient amplitude corresponds to bulk Mach numbers of only $\times$3–0.15 against sound speeds of 1300–1500 km s$\times$4, indicating gentle, clearly subsonic motion rather than merger- or shock-driven flow.

Consistency with sloshing simulations

The observed pattern aligns naturally with idealized simulations of off-axis minor mergers, in which displacement of the cool core excites sloshing motions producing spiral cold fronts and organized flows. Synthetic microcalorimeter observations show that projected velocity gradients depend on viewing angle relative to the sloshing plane. The modest observed amplitudes—one-sided projected velocities of order 100–150 km s$\times$5—are consistent with a true sloshing speed of a few hundred km s$\times$6 viewed at an inclination of several tens of degrees from the plane of the sky; the authors illustrate that $\times$7 implies $\times$8–30°, while noting this estimate is illustrative and tighter constraints require tailored simulations.

This interpretation dovetails with the deep Chandra analysis, which reproduces the spiral, southeastern splash feature, and possible northwestern shock with a 1:10 mass-ratio off-axis merger, corresponding to an epoch about 0.2 Gyr after second core passage. It also parallels the XRISM result in Centaurus, where bulk flows of 130–310 km s$\times$9 were attributed to sloshing alongside low dispersion; the difference between the two clusters is plausibly viewing geometry, with Centaurus seen closer to edge-on.

Non-thermal pressure support

Despite the detected ordered motions, the velocity dispersion remains highly subsonic everywhere, implying three-dimensional turbulent Mach numbers of $kT$0–0.2. Including both turbulent and bulk contributions, the total non-thermal pressure fraction stays below about 2.5% in all regions, with the largest values in NW (~2.5%) and Center (~2%). A forward-modeling comparison with TNG-Cluster indicates that projection effects cannot fully reconcile A2029 with simulated clusters; the discrepancy reflects genuinely low turbulent and bulk amplitudes.

Two consequences follow directly. First, hydrostatic mass bias from gas motions in the A2029 core is limited to a few percent, consistent with field-averaged XRISM results and with independent hydrostatic-mass analyses. Second, sloshing can produce spectroscopically detectable velocity structure without substantially compromising the hydrostatic approximation—a useful calibration point for how "relaxed" clusters should be treated in mass measurements. The authors caution that the NE region's near-zero best-fit dispersion has non-negligible systematic uncertainty due to lower counts and PSF sensitivity, and that the non-thermal fractions assume isotropic random motions.

Heating–cooling balance

Using the standard turbulent heating prescription $kT$1 with $kT$2 and the measured dispersions, compared against radiative cooling rates computed from the X-COP density profile, the paper finds $kT$3 at most (Center and NW), $kT$4 in FNE, below 0.1 in SE and SW, and effectively zero in NE. Turbulent dissipation alone therefore appears insufficient to offset radiative cooling throughout the core. This conclusion is robust to excluding the Fe XXV resonance line, although that test shifts inferred dispersions substantially in some regions (e.g., Center from 131 to 83 km s$kT$5, NW from 123 to 29 km s$kT$6), indicating genuine sensitivity of $kT$7 to line modeling.

The result depends strongly on the assumed driving scale. If AGN-driven motions on the scale of radio/X-ray bubbles (~10–20 kpc) dominate the central region rather than the emission-measure-weighted effective scales of 70–196 kpc, the local heating rate rises closer to the cooling rate. The paper also notes tension—but not contradiction—with buoyancy-driven heating models that adopt higher field-averaged dispersions and smaller injection scales. More broadly, the resolved decomposition separates sloshing-driven bulk advection from unresolved dispersion, leaving open whether mixing and sloshing-driven transport of high-entropy gas contribute meaningfully to thermal balance over the full cooling volume.

Limitations and open questions

Several caveats bear on the quantitative results. The velocity dispersion is only marginally constrained in some regions—the NE solution sits near the fit boundary under PV-only statistics—and $kT$8 shows measurable dependence on the treatment of the Fe XXV resonance line and on PSF mixing assumptions, whereas the bulk-velocity pattern does not. Metal abundances derived here run systematically lower than Chandra values (ratio $kT$9), consistent with known cross-calibration trends but introducing some model dependence in thermal parameters. The inclination of the sloshing plane, and hence the true three-dimensional flow speed, remains unconstrained without tailored simulations matched to the specific geometry. Finally, whether AGN-driven turbulence on bubble scales rescues the heating budget in the innermost core cannot be settled with the present data.

Conclusion

Combining PV and GO Resolve observations with a full spatial-spectral mixing treatment, this work establishes a coherent, highly significant ($v_{\rm bulk}$0; $v_{\rm bulk}$1 including systematics) north–south bulk-velocity gradient of up to ~280 km s$v_{\rm bulk}$2 across the A2029 core, providing direct spectroscopic confirmation that the Chandra spiral traces sloshing-induced gas motion. At the same time, subsonic dispersions below ~150 km s$v_{\rm bulk}$3 limit non-thermal pressure support to under ~2.5%, and turbulent dissipation alone falls short of balancing radiative cooling at the assumed driving scales. A2029 thus emerges as a well-defined case in which ordered sloshing coexists with an otherwise hydrostatic atmosphere, and the remaining open questions—inclination of the sloshing plane, the dominant turbulence driving scale, and the role of mixing in the thermal balance—define the specific follow-up measurements required.