Binary central stars of PN discovered through photometric variability. I. What we know and what we would like to find out

TL;DR

This paper reviews the current understanding of binary central stars in planetary nebulae, highlighting biases in detection methods, unexpected period distributions, and proposing new surveys to improve knowledge of binary fractions.

Contribution

It critically assesses existing photometric variability surveys, identifies biases and discrepancies in binary period distributions, and introduces a new survey approach to better estimate binary fractions.

Findings

Short-period binary fraction is 10-15%.

Detected binaries mostly have periods less than 3 days.

Biases likely affect the observed period distribution.

Abstract

Shaping axi-symmetric planetary nebulae is easier if a companion interacts with a primary at the top of the asymptotic giant branch. To determine the impact of binarity on planetary nebula formation and shaping, we need to determine the central star of planetary nebula binary fraction and period distribution. The short-period binary fraction has been known to be 10-15% from a survey of ~100 central stars for photometric variability indicative of irradiation effects, ellipsoidal variability or eclipses. This survey technique is known to be biased against binaries with long periods and this fact is used to explain why the periods of all the binaries discovered by this survey are smaller than 3 days. In this paper we assess the status of knowledge of binary central stars discovered because of irradiation effects. We determine that, for average parameters, this technique should be biased against periods longer than 1-2 weeks, so it is surprising that no binaries were found with periods longer than 3 days. Even more puzzling is the fact that 9 out of 12 of the irradiated binaries, have periods smaller than one day, a fact that is starkly at odds with post-common envelope predictions. We suggest that either all common envelope models tend to overestimate post-common envelope periods or that this binary survey might have suffered from additional, unquantified biases. If the latter hypothesis is true, the currently-known short-period binary fraction is put in serious doubt. We also introduce a new survey for binary-related variability, which will enable us to better quantify biases and determine an independent value for the short period binary fraction.