The fraction of Compton-thick sources in an INTEGRAL complete AGN sample

TL;DR

This study analyzes the distribution of absorption in a complete AGN sample from INTEGRAL, revealing that the true fraction of Compton-thick AGN is higher than observed at greater distances due to selection biases.

Contribution

It provides a corrected estimate of the Compton-thick AGN fraction accounting for selection biases and redshift effects, aligning with optically selected sample results.

Findings

The fraction of absorbed AGN is 43%, with 7% being Compton-thick.

Selection bias causes underestimation of Compton-thick AGN at higher redshifts.

The true Compton-thick fraction at low redshift exceeds 24%.

Abstract

We study the NH distribution in a complete sample of 88 AGN selected in the 20-40 keV band from INTEGRAL/IBIS observations. We find that the fraction of absorbed (NH > 10^{22} cm^2) sources is 43% while the Compton thick AGN comprise 7% of the sample. While these estimates are fully compatible with previous soft gamma-ray surveys, they would appear to be in contrast with results reported by Risaliti et al. (1999) using an optically selected sample. This apparent difference can be explained as being due to a selection bias caused by the reduction in high energy flux in Compton thick objects rendering them invisible at our sensitivity limit. Taking this into account we estimate that the fraction of highly absorbed sources is actually in close agreement with the optically selected sample. Furthermore we show that the measured fraction of absorbed sources in our sample decreases from 80% to about 20-30% as a function of redshift with all Compton thick AGN having z < 0.015. If we limit our analysis to this distance and compare only the type 2 objects in our sample with the Risaliti et al. objects below this redshift value, we find a perfect match to their NH distribution. We conclude that in the low redshift bin we are seeing almost the entire AGN population, from unabsorbed to at least mildly Compton thick objects, while in the total sample we lose the heavily absorbed 'counterparts' of distant and therefore dim sources with little or no absorption. Taking therefore this low z bin as the only one able to provide the 'true' distribution of absorption in type 1 and 2 AGN, we estimate the fraction of Compton thick objects to be >24%.