[Paper Review] Pile-up correction for the Swift-XRT observations in WT mode

T. Mineo, P. Romano+13 more|arXiv (Cornell University)|Jan 29, 2007

Astronomical Observations and Instrumentation3 citations

TL;DR

This paper presents three robust, complementary methods to correct for pile-up in Swift-XRT's Windowed Timing (WT) mode, where high X-ray count rates (>300 c s⁻¹) cause photon pile-up that distorts flux and spectral measurements. Using data from GRB 060124 and ground calibration, all three methods—radial profile analysis, spectral index stability, and grade distribution correction—consistently identify a 4-pixel central exclusion radius as optimal for accurate source characterization at rates >400 c s⁻¹.

ABSTRACT

The detector at the focal plane of the Swift X-ray Telescope (XRT) supports four readout modes, automatically changed on board, to cover the dynamical range of fluxes and rapid variability expected from GRB afterglows. The Windowed Timing (WT) mode is used for sources with flux higher than a few mCrab and is obtained by compressing 10 rows into a single row, and then reading out only the central 200 columns of the CCD. Point sources with a rate above ~300 c/s produce severe pile-up in the central region of the Point Spread Function. This paper presents three methods to correct the effects of the pile-up in WT mode. On ground calibration results and data from the very bright GRB 060124 are used to define and test these methods.

Motivation & Objective

• To address severe pile-up effects in Swift-XRT's Windowed Timing (WT) mode, which distorts flux and spectral measurements at high count rates.
• To develop and validate multiple independent methods for detecting and correcting pile-up in real-time and post-observation data analysis.
• To determine the optimal central exclusion region size in the PSF to restore accurate spectral and flux measurements for bright X-ray sources.
• To validate the methods using high-rate observations of GRB 060124 and ground calibration data from the Panter laboratory.
• To ensure consistency across different diagnostic approaches—spatial, spectral, and grade-based—to improve reliability in pile-up correction.

Proposed method

• The first method uses radial intensity profile comparison: the observed PSF is compared with the on-ground calibrated PSF; regions with lower counts than expected are flagged as pile-up affected.
• The second method evaluates spectral stability: the photon index is measured in concentric annular regions with increasing central holes; pile-up is corrected when the index stabilizes.
• The third method analyzes the event grade distribution: pile-up distorts grades (deficit at grade 0, excess at higher grades); correction is achieved when the deviation becomes negligible.
• All three methods are applied to five count rate intervals (100–200, 200–300, 300–400, >400 c s⁻¹) to assess consistency across flux levels.
• The methods are tested on Swift-XRT data from GRB 060124, where the XRT observed in WT mode during a high-rate burst, and on ground calibration data from the Panter laboratory.
• The exclusion region is defined as a circular central area of 4 pixels radius, confirmed by all three methods at rates >400 c s⁻¹.

Experimental results

Research questions

• RQ1 What is the optimal size of the central exclusion region in the PSF to correct for pile-up in Swift-XRT WT mode at high count rates?
• RQ2 How do different diagnostic indicators—radial intensity profile, spectral index stability, and grade distribution—compare in detecting pile-up effects?
• RQ3 To what extent does pile-up distort flux and spectral measurements in Swift-XRT WT mode, and can this be quantitatively corrected?
• RQ4 Can ground calibration data reliably validate pile-up correction methods for real high-flux transient events like GRB 060124?
• RQ5 Do the three proposed methods yield consistent results across varying count rate regimes?

Key findings

• All three pile-up correction methods—radial profile, spectral index stability, and grade distribution—yield consistent results, identifying a 4-pixel central exclusion radius as optimal for count rates >400 c s⁻¹.
• Pile-up in WT mode becomes significant at source intensities above ~250 c s⁻¹, as confirmed by ground calibration at the Panter laboratory.
• At rates >400 c s⁻¹, the radial intensity profile shows a deficit in the central region compared to the expected PSF, indicating severe pile-up.
• The spectral index stabilizes when a central exclusion of 4 pixels is applied, confirming that pile-up effects are corrected at this radius.
• The grade distribution shows a clear deficit at grade 0 and excess at higher grades due to pile-up, which diminishes when a 4-pixel hole is excluded.
• The consistency across all three methods confirms the reliability of the 4-pixel exclusion radius for accurate flux and spectral analysis in high-rate WT mode observations.