[Paper Review] The Optical Gravitational Lensing Experiment. The OGLE-III Catalog of Variable Stars. IV. Long-Period Variables in the Large Magellanic Cloud

I. Soszyński, A. Udalski+7 more|arXiv (Cornell University)|Oct 7, 2009

Astronomical Observations and Instrumentation21 citations

TL;DR

This paper presents the OGLE-III Catalog of Long-Period Variables (LPVs) in the Large Magellanic Cloud, identifying 91,995 stars including 79,200 OSARGs, 11,128 SRVs, and 1,667 Miras using 8 years of multi-epoch I-band photometry. The key contribution is the discovery of an additional period-luminosity sequence between Wood's sequences C and C′, populated by SRVs, and evidence of distinct amplitude distributions for carbon-rich vs. oxygen-rich stars, enhancing understanding of pulsation modes and AGB evolution.

ABSTRACT

The fourth part of the OGLE-III Catalog of Variable Stars presents 91 995 long-period variables (LPVs) in the Large Magellanic Cloud (LMC). This sample consists of 79 200 OGLE Small Amplitude Red Giants (OSARGs), 11 128 semiregular variables (SRVs) and 1667 Mira stars. The catalog data include basic photometric and astrometric properties of these stars, long-term multi-epoch VI photometry and finding charts. We describe the methods used for the identification and classification of LPVs. The distribution of I-band amplitudes for carbon-rich stars shows two maxima, corresponding to Miras and SRVs. Such a distinction between Miras and SRVs is not obvious for oxygen-rich stars. We notice additional period-luminosity sequence located between Wood's sequences C and C' and populated by SRVs.

Motivation & Objective

• To identify and classify long-period variable stars (LPVs) in the Large Magellanic Cloud (LMC) using deep, long-baseline photometry.
• To improve the understanding of pulsation modes and period-luminosity relations in red giants, particularly in the context of AGB and post-AGB evolution.
• To investigate the distinction between Mira stars, semiregular variables (SRVs), and OGLE Small Amplitude Red Giants (OSARGs) based on light curve morphology and amplitude distributions.
• To identify rare or extreme objects such as R Coronae Borealis (R CrB) star candidates among C-rich giants.
• To explore the existence and nature of additional period-luminosity sequences beyond those identified by Wood et al. (1999).

Proposed method

• Utilized 8 years of multi-epoch I-band photometry from the OGLE-III survey (2001–2009), with additional OGLE-II data extending the baseline to 13 seasons for central fields.
• Applied automated period-finding algorithms to detect pulsation periods, using phase-coherent folding of light curves to classify variability types.
• Classified stars into Mira, SRV, and OSARG categories based on light curve morphology, amplitude thresholds (e.g., V-band >2.5 mag for Miras), and period-luminosity (PL) sequences.
• Constructed Wesenheit index diagrams (W_JK) to derive extinction-independent luminosities and identify PL sequences, including a newly identified ridge between sequences C and C′.
• Used color-magnitude and period-amplitude diagrams to distinguish between carbon-rich and oxygen-rich stars, noting bimodal amplitude distributions in C-rich stars.
• Identified candidate R CrB stars via irregular, deep brightness declines in light curves, particularly among C-rich giants, and flagged them for spectroscopic follow-up.

Experimental results

Research questions

• RQ1 What is the distribution of long-period variables in the Large Magellanic Cloud, and how do their light curves and periods relate to their spectral types and pulsation modes?
• RQ2 Do additional period-luminosity sequences exist beyond those identified by Wood et al. (1999), and what do they reveal about pulsation behavior in red giants?
• RQ3 How do the amplitude distributions of carbon-rich and oxygen-rich long-period variables differ, and what does this imply for pulsation mode identification?
• RQ4 Can R Coronae Borealis star candidates be reliably identified from photometric light curves alone, and what is their frequency among C-rich giants?
• RQ5 What is the significance of the newly identified PL sequence between Wood's sequences C and C′, and does it represent a different pulsation mode or a distinct stellar population?

Key findings

• The catalog contains 91,995 long-period variables in the LMC, comprising 79,200 OSARGs, 11,128 SRVs, and 1,667 Mira stars, significantly expanding the known sample of LPVs.
• A new period-luminosity sequence was identified between Wood's sequences C and C′, populated by SRVs with smaller amplitudes, suggesting a distinct pulsation or evolutionary population.
• Carbon-rich stars exhibit a bimodal amplitude distribution in the I-band, with two distinct peaks corresponding to Miras and SRVs, while oxygen-rich stars show no such clear separation.
• The light curves of 10 very long-period Miras (P > 1000 days) with large amplitudes (AI ≈ 4 mag) show high stability, suggesting they are likely oxygen-rich, possibly with initial masses ≥4 M☉ on the AGB.
• Several dozen R CrB star candidates were identified among C-rich giants, with irregular, deep declines in brightness, though spectroscopic confirmation is required.
• Very red (J-K > 2 mag) and faint (I > 18 mag) Miras are heavily obscured by circumstellar dust, consistent with C-rich AGB stars, and most lie on sequence C in the W_JK diagram.