[Paper Review] On the discrepancy between Chandra and XMM temperature profiles for A1835

Maxim Markevitch|arXiv (Cornell University)|May 20, 2002

Galaxies: Formation, Evolution, Phenomena4 citations

TL;DR

This paper resolves a significant discrepancy between Chandra and XMM-Newton temperature profiles for galaxy cluster A1835 by identifying two key issues: a previously undetected background flare in the Chandra data that biased temperatures high at large radii, and unaccounted-for XMM-Newton point spread function (PSF) scattering that suppressed temperature gradients by contaminating outer annuli with flux from the bright core. After correcting for both effects, the Chandra and XMM profiles become consistent, highlighting critical calibration considerations for cluster X-ray analysis.

ABSTRACT

This short technical note addresses a large discrepancy between the temperature profiles for the galaxy cluster A1835 derived by Schmidt et al. (2001) using Chandra and by Majerowicz et al. (2002) using XMM. The causes of this discrepancy may be instructive for the Chandra and XMM cluster analyses in general. The observation used by Schmidt et al. was affected by a mild background flare that could not be identified by the usual technique. This flare biased upwards the measured temperatures at large radii. The remaining discrepancy appears to be due to the XMM PSF scattering that was not taken into account in the published analyses. While the XMM PSF is narrow, the surface brightness of a typical cluster also declines very steeply with radius. For the moderately distant, cooling flow cluster A1835, about 1/3 of the observed XMM brightness at any radius is due to the PSF scattering from the smaller radii. As a result, the contamination from the bright cool cluster center biases low the measured temperatures near the core, and in general, any temperature gradients are underestimated.

Motivation & Objective

• To investigate the origin of the large discrepancy between Chandra and XMM-Newton temperature profiles for the galaxy cluster A1835.
• To identify and correct systematic errors in the Chandra analysis, particularly undetected background flares affecting temperature measurements at large radii.
• To quantify the impact of XMM-Newton's point spread function (PSF) on temperature profile reconstruction, especially in clusters with steep brightness profiles.
• To improve the reliability of X-ray temperature profiles for galaxy clusters by addressing instrumental and data reduction artifacts.
• To provide methodological guidance for future Chandra and XMM-Newton cluster analyses to avoid similar discrepancies.

Proposed method

• Re-analyzed Chandra OBSID 495 using the 0.8–8 keV energy band and applied a soft Galactic background excess correction by fitting residual spectra in distant chip regions with a low-temperature MEKAL model.
• Identified a residual background flare in Chandra OBSID 495 by detecting a hard excess in the 2.5–7 keV band, consistent with a power-law model with photon index γ ≈ −0.1 and cutoff at 5 keV.
• Corrected the Chandra temperature profile by subtracting the spatially uniform flare component, normalized by solid angle, with a ±40% uncertainty in normalization.
• Modeled the XMM-Newton PSF using the on-axis MOS1 PSF (rc = 3.5′′, α = 1.36) and convolved a Chandra-derived brightness model with the PSF to estimate flux contributions from inner regions to each annulus.
• Used XSPEC simulations with EPIC responses to test whether PSF-smeared temperature profiles from Chandra data could reproduce the XMM-Newton results.
• Compared corrected Chandra profiles with XMM-Newton data, showing that both corrections reconcile the profiles.

Experimental results

Research questions

• RQ1 What causes the large discrepancy between the Chandra and XMM-Newton temperature profiles of A1835, particularly in the outer regions?
• RQ2 To what extent does an undetected background flare in Chandra OBSID 495 bias the temperature measurements at large radii?
• RQ3 How significant is the contribution of XMM-Newton PSF scattering from the bright core to the observed flux in outer annuli?
• RQ4 Can the XMM-Newton temperature profile be reproduced by degrading a corrected Chandra temperature profile through PSF smearing?
• RQ5 What instrumental and data reduction corrections are essential for accurate temperature profile measurements in X-ray cluster studies?

Key findings

• A previously undetected background flare in Chandra OBSID 495 caused a hard excess in the 2.5–7 keV band, contributing 30–40% of the nominal background at E = 3–5 keV, which biased temperature measurements upward at large radii.
• After correcting for the flare, the Chandra temperature profile from OBSID 495 became consistent with the profile from the unaffected OBSID 496, resolving the internal inconsistency in Chandra data.
• The XMM-Newton PSF contributes approximately 30% of the observed flux in each annulus from inner regions, with up to 10% or more from the central peak, due to the steep decline of the cluster brightness profile.
• XMM-Newton temperature profiles are systematically biased low in regions with temperature gradients because PSF scattering from the cool core contaminates outer annuli, suppressing the observed temperature rise.
• A simple XSPEC simulation showed that PSF-smeared temperature profiles from the corrected Chandra profile reproduce the XMM-Newton results: for the 2nd annulus, XMM-derived temperature was 6.4–6.8 keV vs. Chandra’s 8 keV; for the 3rd, 7.0–7.7 keV vs. 9 keV.
• The combined effect of Chandra flare correction and XMM PSF modeling reconciles the two profiles, indicating that both errors were responsible for the majority of the observed discrepancy.