[Paper Review] Cosmic Microwave Background filters and the Dark-Flow measurement

F. Atrio‐Barandela, A. Kashlinsky+2 more|arXiv (Cornell University)|Nov 19, 2012

Cosmology and Gravitation Theories3 references3 citations

TL;DR

This paper defends the detection of a large-scale bulk flow—termed 'Dark Flow'—using Cosmic Microwave Background (CMB) data filtered via a specialized Wiener-type filter (KAB filter), which isolates kinematic Sunyaev-Zel'dovich (KSZ) dipole signals from galaxy clusters. The KAB filter effectively removes primary CMB anisotropies down to cosmic variance limits, enabling a 3.5–4σ detection of a coherent flow aligned with the CMB dipole, while flawed implementations of alternative filters fail due to poor monopole subtraction and incorrect cluster profile assumptions.

ABSTRACT

Recent measurements of large-scale peculiar velocities from the cumulative kinematic Sunyaev-Zel'dovich effect using WMAP data and X-ray selected clusters from ROSAT have identified a bulk flow of galaxy clusters at $\sim 600-1,000$ km s$^{-1}$ on scales of $\sim0.5-1$ Gpc, roughly aligned with the all-sky Cosmic Microwave Background (CMB) dipole. The flow is inferred from the detection of a residual dipole generated by CMB fluctuations exclusively in the direction of galaxy clusters, and measured within apertures containing zero monopole. Consistent with this interpretation, the amplitude of the dipole correlates with the X-ray luminosity of the clusters. To enable this measurement, the CMB data need to be filtered to remove the primary CMB, thereby increasing the data's signal-to-noise ratio. Filtering cannot imprint a signal with the mentioned properties at cluster positions; however, an inadequately designed filter can greatly suppress such a signal. We show here that recent studies that failed to detect a large-scale flow with various filters indeed adopted flawed implementations; when correctly implemented, these alternative filters lead to results that are in fact consistent with the Dark Flow signal. The discrepancies can be traced to the likely presence of residual dipoles caused by the thermal Sunyaev-Zel'dovich effect, to assumptions about cluster profiles incompatible with the data as well as failure to compute dipoles at the zero monopole aperture. PLANCK maps, with their large frequency coverage and a 217 GHz channel, will be instrumental to probe bulk flows, to remove spurious dipole signals and to help identify filtering schemes appropriate for this measurement.

Motivation & Objective

• To resolve discrepancies in bulk flow detection from WMAP CMB data using different filtering techniques.
• To demonstrate that alternative filters (OMCP, MH) fail to detect the Dark Flow signal due to improper design and calibration.
• To validate the KAB filter as the most effective method for isolating KSZ dipole signals from galaxy clusters by minimizing primary CMB contamination.
• To show that the residual dipole signal correlates with X-ray luminosity, confirming its physical origin in cluster peculiar velocities.
• To advocate for Planck frequency data and 217 GHz channel use to improve systematics control and confirm the flow.

Proposed method

• Application of a Wiener-type filter (KAB filter) to WMAP CMB maps to suppress primary CMB anisotropies while preserving KSZ-induced dipoles.
• Use of zero-monopole apertures to eliminate residual monopole contributions and isolate dipole signals at cluster positions.
• Comparison of dipole amplitudes and directions across multiple filters (KAB, OMCP, MH) to assess signal recovery and statistical significance.
• Computation of error bars using simulations, with correction for correlated WMAP Differential Assemblies (DAs) and proper dipole subtraction outside Galactic masks.
• Use of X-ray luminosity as a proxy to test signal correlation, confirming physical association with clusters.
• Evaluation of filter performance using angular power spectrum analysis and comparison of dipole amplitudes at different apertures.

Experimental results

Research questions

• RQ1 Do alternative filtering schemes such as those in OMCP and MH correctly recover the KSZ dipole signal from galaxy clusters in WMAP data?
• RQ2 Why do some studies fail to detect the Dark Flow signal despite using similar data, and is it due to filter design or statistical misestimation?
• RQ3 To what extent does the KAB filter outperform other filters in suppressing primary CMB while preserving the KSZ dipole?
• RQ4 How do assumptions about cluster electron density and pressure profiles affect the performance of filtering schemes?
• RQ5 Can Planck frequency data, especially the 217 GHz channel, improve systematics control and confirm the existence of the bulk flow?

Key findings

• The KAB filter successfully removes the primary CMB anisotropies down to the cosmic variance limit, enabling optimal signal-to-noise for detecting KSZ dipoles.
• Alternative filters (OMCP, MH) fail to detect the Dark Flow signal not due to absence of the signal, but due to flawed design, including poor monopole subtraction and incorrect cluster profile assumptions.
• The residual dipole measured with the KAB filter is strongly correlated with X-ray luminosity of clusters, confirming its physical origin in cluster peculiar velocities.
• The KAB filter achieves a 3.5–4σ detection of the bulk flow, while flawed implementations in other studies lead to artificially inflated error bars and false non-detections.
• The Planck Collaboration’s use of the ILC map and incorrect error estimation reduced the significance of the signal, but the underlying data support the KAB result.
• The 217 GHz Planck channel is critical for identifying and removing thermal SZ contamination and testing filtering systematics.