An investigation of the "10 keV feature" in the spectra of Accretion Powered X-ray Pulsars with NuSTAR

TL;DR

This study systematically investigates the presence of the mysterious 10 keV spectral feature in accreting X-ray pulsars using NuSTAR data, revealing it in a subset of observations and emphasizing careful analysis to avoid false detections.

Contribution

It provides the first comprehensive systematic analysis of the 10 keV feature across multiple pulsars with NuSTAR, clarifying its occurrence and potential analysis pitfalls.

Findings

Detected the 10 keV feature in 16 out of 58 observations across 11 sources.

The feature is modeled as a Gaussian absorption line around 10 keV.

Incorrect analysis of flare data can lead to false detection of the feature.

Abstract

Some of the accreting X-ray pulsars are reported to exhibit a peculiar spectral feature at $\sim$10 keV, known as the "10 keV feature". The feature has been characterized as either an emission line or an absorption line, and its origin is unknown. It has been found in multiple observations of the same source by different observatories, but not all the observations of any particular source consistently showed the presence of it. In this work, we have carried out a systematic investigation for the presence of the "10 keV feature" using data from NuSTAR, a low background spectroscopic observatory having uninterrupted wide band coverage on either side of 10 keV. We performed a systematic spectral analysis on 58 archival NuSTAR observations of 30 bright X-ray pulsars. The 3$-$79 keV spectral continua of these selected sources were fitted with a model chosen on the basis of its fitting quality in 3$-$15 keV and model simplicity, and then inspected for the presence of the "10 keV feature". Our analysis indicates the presence of such a feature in 16 out of 58 the NuSTAR observations of 11 different sources and is fitted with a Gaussian absorption model centered around 10 keV. Our analysis also suggests that such a feature could be wrongly detected if flare data is not analyzed separately from persistent emission.