The Astrophysical Journal, 786:97 (32pp), 2014 May 10 doi:10.1088/0004-637X/786/2/97 C 2014. The American Astronomical Society. All rights reserved. Printed in the U.S.A. AN OPTICAL SPECTROSCOPIC STUDY OF T TAURI STARS. I. PHOTOSPHERIC PROPERTIES Gregory J. Herczeg1,2,3,4 and Lynne A. Hillenbrand2 1 Kavli Institute for Astronomy and Astrophysics, Peking University, Yi He Yuan Lu 5, Haidian Qu, Beijing 100871, China 2 Caltech, MC105-24, 1200 East California Boulevard, Pasadena, CA 91125, USA 3 Max-Planck-Institut fur¨ extraterrestrische Physik, Postfach 1312, D-85741 Garching, Germany Received 2013 August 26; accepted 2014 March 4; published 2014 April 22 ABSTRACT Estimates of the mass and age of young stars from their location in the H-R diagram are limited by not only the typical observational uncertainties that apply to field stars, but also by large systematic uncertainties related to circumstellar phenomena. In this paper, we analyze flux-calibrated optical spectra to measure accurate spectral types and extinctions of 281 nearby T Tauri stars (TTSs). The primary advances in this paper are (1) the incorporation of a simplistic accretion continuum in optical spectral type and extinction measurements calculated over the full optical wavelength range and (2) the uniform analysis of a large sample of stars, many of which are well known and can serve as benchmarks. Comparisons between the non-accreting TTS photospheric templates and stellar photosphere models are used to derive conversions from spectral type to temperature. Differences between spectral types can be subtle and difficult to discern, especially when accounting for accretion and extinction. The spectral types measured here are mostly consistent with spectral types measured over the past decade. However, our new spectral types are one to two subclasses later than literature spectral types for the original members of the TW Hya Association (TWA) and are discrepant with literature values for some well-known members of the Taurus Molecular Cloud. Our extinction measurements are consistent with other optical extinction measurements but are typically 1 mag lower than near-IR measurements, likely the result of methodological differences and the presence of near-IR excesses in most CTTSs. As an illustration of the impact of accretion, spectral type, and extinction uncertainties on the H-R diagrams of young clusters, we find that the resulting luminosity spread of stars in the TWA is 15%–30%. The luminosity spread in the TWA and previously measured for binary stars in Taurus suggests that for a majority of stars, protostellar accretion rates are not large enough to significantly alter the subsequent evolution. Key words: stars: low-mass – stars: pre-main sequence Online-only material: color figures, machine-readable table 1. INTRODUCTION provide fascinating diagnostics of the stellar environment and star–disk interactions, they also pose significant problems for Classical T Tauri stars (CTTSs) are the adolescents of stellar assessing the stellar properties and evolution of the star/disk evolution. The star is near the end of its growth and almost system. How disk mass, structure and accretion rate change with fully formed, with a remnant disk and ongoing accretion. The age and mass requires accurate spectral typing and luminosity accretion/disk phase typically lasts ∼2–5 Myr, though some measurements (e.g., Furlan et al. 2006; Sicilia-Aguilar et al. stars take as long as 10 Myr before losing their disks and 2010; Oliveira et al. 2013; Andrews et al. 2013). While median emerging toward maturity. Strong magnetic activity leads to cluster ages provide an accurate relative age scale between pimply spots on the stellar surfaces. Some T Tauri stars (TTSs) regions (e.g., Naylor et al. 2009), age spreads within clusters are still hidden inside their disks, not yet ready to emerge. Manic may be real or could result from observational uncertainties mood swings change the appearance of the star and are often (e.g., Hartmann et al. 1998; Hillenbrand et al. 2008; Preibisch explained with stochastic accretion. Depression has been seen 2012). in light curves on timescales of days to years. Sometimes every The uncertainties in stellar parameters affect our interpreta- CTTS seems as uniquely precious as a snowflake. tion of stellar evolution. For example, Gullbring et al. (1998) TTS properties were systematically characterized in seminal found accretion rates an order of magnitude lower than those of papers by, e.g., Cohen & Kuhi (1979), Herbig & Bell (1988), Hartigan et al. (1995) and attributed much of this difference to Basri & Batalha (1990), Valenti et al. (1993), Hartigan et al. lower values of extinction. The Gullbring et al. (1998) accretion −8 −1 (1995), Kenyon & Hartmann (1995), Gullbring et al. (1998). rates of 10 M yr means that steady accretion in the CTTS In the last decade, dedicated optical and IR searches revealed phase accounts for a negligible amount of the final mass of a thousands of young stars, typically confirmed with spectral star. However, subsequent near-IR analyses have revised extinc- typing (e.g., Hillenbrand 1997; Briceno et al. 2002; Luhman tions upward (e.g., White & Ghez 2001; Fischer et al. 2011; 2004; Rebull et al. 2010). However, significant differences in Furlan et al. 2011). These higher extinctions would yield ac- −7 −1 extinction and accretion properties between different papers and cretion rates of 10 M yr , fast enough that steady accretion methods has led to confusion in the properties of even the closest over the ∼2–3 Myr CTTS phase would account for ∼20%–50% and best studied samples of young stars. of the final stellar mass, or more with the older ages measured Some of this confusion is exacerbated by stochastic and by Bell et al. (2013). The uncertainties in stellar properties intro- rotation variability of TTSs. While manic and depressive periods duce skepticism in our ability to use young stellar populations to test theories of star formation and pre-main-sequence evolution. For CTTSs, minimizing the uncertainties in spectral type, 4 Visiting Astronomer, LERMA, Observatoire de Paris, ENS, UPMC, UCP, CNRS, 61 avenue de l’Observatoire, F-75014 Paris, France. extinction, and accretion (often referred to as veiling of the 1 The Astrophysical Journal, 786:97 (32pp), 2014 May 10 Herczeg & Hillenbrand Table 1 Observation Setup and Log Telescope Dates Instrument Slit Blue Setup Red Setup Grating Wavelength Res. Grating Wavelength Res. Palomar 2008 Jan 18–21 DoubleSpec 1–4 B600 3000–5700 700 R316 6200–8700 500 Palomar 2008 Dec 28–30 DoubleSpec 4 B600 3000–5700 700 R316 6200–8700 500 Keck I 2006 Nov 23 LRIS 0. 7–1 B400 3000–5700 900 R400 5700–9400 1000 Keck I 2008 May 28 LRIS 1 B400 3000–5700 900 R400 5700–9400 1000 photosphere by accretion) requires fitting all three parameters 4 width slit. Our LRIS observations were obtained with the 0.7 simultaneously (e.g., Bertout et al. 1988; Basri & Bertout 1989; and 1.0 slits. Seeing during both Palomar runs typically varied Hartigan & Kenyon 2003). In recent years, such fits have from 2–4, though for a few hours the seeing reached ∼1. received increasing attention and have been applied to Hubble The seeing was ∼0.8 and ∼0.7 during our 2006 November Space Telescope (HST) photometry of the Orion Nebula Cluster and 2008 May Keck runs, respectively. Seeing was often worse (da Rio et al. 2010; Manara et al. 2012), broadband optical/ than these measurements for objects at high airmass. The po- near-IR spectra of two Orion Nebular Cluster stars (Manara sition angle of the slit was set to the parallactic angle for all et al. 2013a), and to near-IR spectroscopy (Fischer et al. 2011; observations of single stars to minimize slit loss. For binaries, McClure et al. 2013). the position angle may be misaligned with the parallactic angle. In this project, we analyze low-resolution optical blue–red These observations were timed to occur at low airmass or when spectra to determine the stellar and accretion properties of 281 the parallactic angle matched the binary position angle. of the nearest young stars in Taurus, Lupus, Ophiucus, the The images were overscan-subtracted and flat-fielded. Most TW Hya Association (TWA), and the MBM 12 Association. DBSP spectra were extracted using a 21 pixel (10 window This first paper focuses on spectral types and extinctions of centered on the source, followed by subtracting the sky as mea- our sample. The primary advances are the inclusion of blue sured nearby on the detector. Binaries with separations <5 were spectra to complement commonly used red optical spectra and extracted simultaneously by assuming a wavelength-dependent use of accretion estimates to calculate the effective temperatures point-spread function determined from an observation of a sin- and luminosities with a single, consistent approach for a large gle star observed close in time. The counts from one source are sample of stars. Discrepancies are found between our results and subtracted from the image, yielding a clean extraction of counts near-IR based extinction measurements. We then discuss how from the other source. In several cases, the counts are extracted these uncertainties affect the reliability of age measurements. on only half of the line spread function to further minimize con- This work was initially motivated to calculate accretion rates tamination from the nearby component. Each spectrum is then from the excess Balmer continuum emission, which will be corrected for light loss outside the slit and outside our extraction described in a second paper. A third paper in this series will window based on the measured seeing as a function of wave- discuss spectrophotometric variability within our sample. length and under the assumption that the point spread function is Gaussian.