Brought to you by:

HIGH-RESOLUTION H-BAND SPECTROSCOPY OF Be STARS WITH SDSS-III/APOGEE. I. NEW Be STARS, LINE IDENTIFICATIONS, AND LINE PROFILES

, , , , , , , , , , , , , , , , , , , , , , , , , , , , and

Published 2014 December 3 • © 2015. The American Astronomical Society. All rights reserved.
, , Citation S. Drew Chojnowski et al 2015 AJ 149 7 DOI 10.1088/0004-6256/149/1/7

Download Article PDF

Article metrics

2657 Total downloads

Share this article

Author e-mails

drewski@virginia.edu

Author affiliations

1 Department of Astronomy, University of Virginia, P.O. Box 400325, Charlottesville, VA 22904-4325, USA; drewski@virginia.edu

2 Apache Point Observatory and New Mexico State University, P.O. Box 59, Sunspot, NM, 88349-0059, USA

3 Department of Physics, Austin College, 900 N. Grand Ave., Sherman, TX 75090, USA

4 Department of Physics & Astronomy, The University of Oklahoma, 440 W. Brooks St., Norman, OK 73019, USA

5 Department of Astronomy, The University of Texas at Austin, 2515 Speedway, Stop C1400 Austin, Texas 78712-1205, USA

6 Department of Astronomy, University of Florida, 211 Bryant Space Science Center, Gainesville, FL 32611-2055, USA

7 ELTE Gothard Astrophysical Observatory, 1H-9704 Szombathely, Szent Imre herceg st. 112, Hungary

8 Department of Astronomy, Indiana University, Bloomington, IN 47405, USA

9 Department of Astronomy, University of Michigan, 830 Dennison, 500 Church St., Ann Arbor, MI 48109-1042, USA

10 Department of Astronomy and Astrophysics, The Pennsylvania State University, University Park, PA 16802, USA

11 Institute for Gravitation and the Cosmos, The Pennsylvania State University, University Park, PA 16802, USA

12 Department of Physics & Astronomy, Johns Hopkins University, Bloomberg Center for Physics and Astronomy, Room 366, 3400 N. Charles Street, Baltimore, MD 21218, USA

13 Department of Physics and Astronomy, Texas Christian University, Box 298840, Fort Worth, TX 76129, USA

ORCID iDs

Dates

  1. Received 2014 April 2
  2. Accepted 2014 July 25
  3. Published
Keywords

circumstellar matter; infrared: stars; line: identification; line: profiles; stars: early-type; stars: emission-line, Be

Journal RSS
Create or edit your corridor alerts

What are corridors?opens in new tab

1538-3881/149/1/7

ABSTRACT

The Apache Point Observatory Galactic Evolution Experiment (APOGEE) has amassed the largest ever collection of multi-epoch, high-resolution (R∼22,500), H-band spectra for B-type emission line (Be) stars. These stars were targeted by APOGEE as telluric standard stars and subsequently identified via visual inspection as Be stars based on H i Brackett series emission or shell absorption in addition to otherwise smooth continua and occasionally non-hydrogen emission features. The 128/238 APOGEE Be stars for which emission had never previously been reported serve to increase the total number of known Be stars by ∼6%. Because the H band is relatively unexplored compared to other wavelength regimes, we focus here on identification of theH-band lines and analysis of the emission peak velocity separations (${\Delta}{{v}_{{\rm p}}}$) and emission peak intensity ratios (V/R) of the usually double-peaked H i and non-hydrogen emission lines. H i Br11 emission is found to preferentially form in the circumstellar disks at an average distance of ∼2.2 stellar radii. Increasing ${\Delta}{{v}_{{\rm p}}}$ toward the weaker Br12–Br20 lines suggests these lines are formed interior to Br11. By contrast, the observed IR Fe iiemission lines present evidence of having significantly larger formation radii; distinctive phase lags between IR Fe ii and H i Brackett emission lines further supports that these species arise from different radii in Be disks. Several emission lines have been identified for the first time including C i 16895, a prominent feature in the spectra for almost a fifth of the sample and, as inferred from relatively large ${\Delta}{{v}_{{\rm p}}}$ compared to the Br11–Br20, a tracer of the inner regions of Be disks. Emission lines at 15760 Å and 16781 Å remain unidentified, but usually appear along with and always have similar line profile morphology to Fe ii 16878. Unlike the typical metallic lines observed for Be stars in the optical, the H-band metallic lines, such as Fe ii 16878, never exhibit any evidence of shell absorption, even when the H i lines are clearly shell-dominated. The first known example of a quasi-triple-peaked Br11 line profile is reported for HD 253659, one of several stars exhibiting intra- and/or extra-species V/R and radial velocity variation within individual spectra. Br11 profiles are presented for all discussed stars, as are full APOGEE spectra for a portion of the sample.

Export citation and abstract BibTeX RIS

Previous article in issue
Next article in issue

1. INTRODUCTION

Since the first observational description (Struve 1931) of the characteristic double-peaked emission lines of classical Be stars, a wealth of research has demonstrated that the emission lines are formed in geometrically thin, equatorial circumstellar disks fed by gas ejected from the surfaces of rapidly rotating B stars (Porter & Rivinius 2003; Rivinius et al. 2013b). Rapid rotation is certainly involved in the formation of these disks, but a comprehensive model of Be disk formation has yet to be created and efforts toward one are complicated by factors including the lack of examples of critically rotating Be stars, star-specific peculiarities, and the requirement of an "on/off" switch to the Be phenomenon. For an uncertain but non-negligible fraction of Be stars, the disks are transient, appearing in one epoch but not another (McSwain et al. 2009; Wisniewski et al. 2010). A variable rotation speed ($v\;{\rm sin} \;i$) that occasionally reaches or exceeds the critical breakup limit is an attractive concept for such a phenomenon (Rivinius et al. 2013c) that needs to be explored for a sample of transient Be stars. Non-radial pulsation and turbulence due to small-scale magnetic fields remain the most likely mechanisms, along with rapid rotation, responsible for the creation of the disks (Rivinius et al. 2013b). When the disks are present, they appear to undergo Keplerian rotation (Meilland et al. 2007; Wheelwright et al. 2012) and many of the observational signatures are consistent with those predicted by the viscous decretion disk model (Lee et al. 1991; Carciofi et al. 2009; Carciofi 2011).

Multi-wavelength studies of Be disks are particularly valuable for diagnosing their structure because emission at different wavelengths originates from different physical locations within the disks (Carciofi 2011). However, unlike in the optical wavelength regime where they have been extensively studied at high spectral and temporal resolution, only a limited number of Be star surveys have been performed at near-infrared (NIR) wavelengths, and these have typically utilized low spectral resolution (Steele & Clark 2001) and small sample sizes (Murdoch et al. 1994; Clark & Steele 2000; Mennickent et al. 2009; Granada et al. 2010). Detailed NIR spectroscopic studies of individual Be stars are more common (e.g., Hony et al. 2000; Mathew et al. 2012) and have been used to better diagnose the gas distribution within Be disks, including the structure of one-armed density waves (Wisniewski et al. 2007; Štefl et al. 2009; Carciofi et al. 2009).

The Apache Point Observatory Galactic Evolution Experiment (APOGEE; Majewski 2012) is actively providing the first ever bulk view of the high-resolution H-band properties and variability of Be stars. APOGEE is one of four surveys comprising the Sloan Digital Sky Survey III (SDSS-III; Eisenstein et al. 2011). While the primary goal of the APOGEE survey is to measure the dynamical and chemical history of the Milky Way Galaxy using high-resolution H-band spectroscopic observations of 105 red giant branch (RGB) stars, APOGEE devotes 35 fibers per 300-fiber pointing to observe hot stars as telluric standards. This, in addition to the surveys large sky coverage and multi-epoch observing strategy, has made APOGEE ideal for serendipitous Be discoveries and high-resolution NIR time series data of Be stars.

Here, we present the first catalog of APOGEE Be (ABE) stars. An overview of the APOGEE survey and APOGEE data is provided in Section 2, the Be sample is described in Section 3, and the identifications of observed metallic emission lines are discussed in Section 4. Sections 5 and 6 focus on quantitative and comparative analysis of emission double-peak separation (${\Delta}{{v}_{{\rm p}}}$) and double-peak intensity ratios (V/R). Commentary on the more unusual or exceptional Be stars within the ABE sample is interspersed throughout, and an atlas of Br11 profiles is provided in Section 7. The Appendix includes supplemental figures displaying full APOGEE spectra for stars with strong emission features, as well as an expanded stellar data table. Future work will focus on the observed spectral variability of sources with multi-epoch APOGEE data as well as follow-up optical spectroscopy for a subset of the sample.

2. APOGEE OVERVIEW

2.1. APOGEE Instrument and Observations

The APOGEE instrument is a 300 fiber, R∼22,500 spectrograph (Wilson et al. 2010) attached to the SDSS 2.5 m telescope (Gunn et al. 2006) at Apache Point Observatory. APOGEE records a vacuum wavelength range of 15145–16955 Å via an arrangement of three Teledyne H2RG 2048 × 2048 detectors. The detector layout consists of "blue," "green," and "red" detectors which cover 15145–15808 Å, 15858–16433 Å, and 16474–16955 Å respectively, resulting in coverage gaps between 15808–15858 Å and 16433–16474 Å. Dispersion varies with wavelength, but the central dispersions of the blue, green, and red detectors are 0.326 Å pix−1, 0.283 Å pix−1, and 0.236 Å pix−1 respectively. As with the original SDSS spectrograph (Smee et al. 2013), APOGEE fibers are plugged into custom, pre-drilled aluminum plates which are loaded into the telescopeʼs focal plane and which can cover 3° diameter areas of sky. Each fiber has a 2'' field of view.

The APOGEE survey uses the Two Micron All Sky Survey (2MASS; Skrutskie et al. 2006) as a source catalog and focuses on observations of known or photometrically likely RGB stars for its main science objective (230/300 fibers per plate). For calibration purposes, blank sky (35/300 fibers) and blue telluric standards (35/300 fibers) are also observed. Typical exposure times are 1 hour, and the number of repeat observations per field is approximately equal to the number of 1 hour observations needed to reach a combined signal-to-noise ratio (S/N) per raw pixel of 100 for stars at the field-specific H magnitude limit. The bright limit for science targets is always H = 7.0, while the faint limit is variable and can be H = 11.0 (1 hour visit), H = 12.2 (3 1 hour visits), H = 12.8 (6 1 hour visits), H = 13.3 (12 1 hour visits), or H = 13.8 (24 1 hour visits). Cohorts of RGB targets with similar H magnitudes are exchanged in and out of the observing sequence as S/N ∼ 100 is reached.

2.2. APOGEE Telluric Standard Stars

Hot O- and B-type (OB) stars are ideal candidates for telluric standard stars (TSSʼs) in the H band because the associated spectra are relatively featureless (Meyer et al. 1998; Steele & Clark 2001). Selection of the TSS for each APOGEE field is based on H magnitude and non-reddening-corrected color rather than on intrinsic spectral properties (Zasowski et al. 2013), such that the TSS for a given field are simply the apparent bluest available stars. Thus, APOGEE makes no distinction between "normal" and emission-line stars other than to prevent from selection as TSS any stars that are reddened with respect to other stars (e.g., dusty B[e] stars) in the 3° fields.

Unlike the RGB science targets, TSS are restricted to $5.5\leqslant H\leqslant 11.0$ and are therefore always expected to reach ${\rm s}/N\;\geqslant 100$ in the typical hour exposures. In addition, the TSS for each APOGEE field are generally "locked-in," meaning that they are observed every time their respective fields are observed rather than being traded in and out of the sequence as are the RGB stars. A more comprehensive description of TSS selection is presented in Zasowski et al. (2013).

2.3. Apache Point Observatory Galactic Evolution Experiment Spectra

There are several details worth noting about the APOGEE spectra. Vacuum wavelengths given in angstroms (Å) are used in APOGEE data and throughout this paper. Some of the spectra were recorded during APOGEE commissioning, prior to the instrument having achieved optimal focus. The resolution of red detector data in APOGEE spectra taken prior to MJD < 55804 (September 2011) is R∼16,000, while the resolution is R∼22,500 for all blue and green detector data regardless of date and for all post-55804 red detector data. The raw data is processed by an automated reduction pipeline (Nidever, in preparation) that extracts the spectra, performs flat-field and wavelength calibration, and performs sky and telluric corrections. APOGEEʼs reduction pipeline is designed to use sky and TSS exposures (see Section 2.2) to remove airglow and telluric absorption lines from the high-resolution spectra. Because the airglow removal process has not yet been perfected, residuals from partially subtracted airglow lines remain in the final reduced data products.

Since the APOGEE survey focuses on chemical abundance and radial velocity (RV) analysis, flux standard stars are not observed and therefore the spectra are not flux-calibrated. All spectra displayed in this paper were continuum normalized using the CONTINUUM task in IRAF by fitting low-order splines to sections of blank continuum adjacent to H i Brackett lines, separately for each detector. Quoted emission line intensities and V/R intensity ratios refer to intensity relative to normalized continuum level (${{F}_{\lambda }}/{{F}_{{\rm c}}}$). Due to the proximity of the Br12 and Br14 lines to the coverage limits of the green detector (21 and 27 Å respectively), it was at times difficult to achieve a continuum fit that did not result in obviously incorrect intensity levels for those with respect to the other Brackett lines. The tendency of the full-width at continuum level for Brackett series emission lines to well exceed 1000 km s−1 was a further complication in salvaging the Br12 and Br14 lines, which are are de-emphasized from analysis for these reasons.

Figure 1 displays examples of APOGEE spectra for two newly-discovered Be stars, demonstrating the three-detector arrangement and associated coverage gaps. The ABE star IDs and modified Julian dates (MJD) of observation are provided above or below "red detector" continuum level, and commonly observed emission lines (see Section 3.5) are labeled with blue text and arrows. Examples of airglow residuals are noted with red text and arrows. The right-hand panels show Br11 line profiles from the same spectra on a velocity scale, with the line profile features of interest labeled. In most cases, Br11 is the strongest hydrogen line covered as well as the hydrogen line least likely to be affected by airglow or telluric contaminants should those be a significant issue. In subsequent figures, narrow contaminants (airglow and hot/cold/bad pixels) have been carefully trimmed from the spectra so as to avoid distraction from the features of astrophysical importance.

Figure 1.

Figure 1. Two examples of APOGEE spectra of Be stars. The "red," "green," and "blue" detectors and the gaps between detectors are labeled in the left panels and examples of residuals from the airglow and telluric removal process airglow residuals are labeled in the upper left panel (red arrows and text). The APOGEE Be star designations (ABE-184 and ABE-045) are shown along with the modified Julian date (MJD) of observation, and the Brackett series lines are labeled (blue arrows and text). Right-hand panels show the Br11 profiles on a velocity scale: ABE-184 exhibits a typical Be star line profile, where the central depression is a consequence of the disk geometry (e.g., Huang 1972), while the Brackett lines for ABE-045 are dominated by shell absorption resulting from occultation of the star by the disk in a nearly or exactly edge-on system (e.g., Rivinius et al. 2006).

Download figure:

Standard image High-resolution image

2.4. Public Availability of the Spectra

Both proprietary and publicly available spectra are used and displayed in this paper. The publicly available spectra were included in SDSS data release 10 (DR10: pertains to APOGEE data taken prior to MJD = 56112), and the full data set will be made publicly available in SDSS data release 12 (DR12: scheduled for 2014 December). Shortly after DR12, we intend to convert the ABE star spectra to the format accepted by the Be Star Spectra Database (BeSS; Neiner et al. 2011) and deposit them there, ensuring convenient public access. More details on DR10-released APOGEE data can be found on the SDSS-III website (https://www.sdss3.org/dr10/irspec/).

3. THE ABE SAMPLE

3.1. Sample Description

The sample at hand consists of 238 Be stars that have been observed by APOGEE a total of 1082 times. Of the 238 ABE stars, 202 were identified through periodic visual inspection of APOGEE spectra and 36 were targeted intentionally to expand the subset of previously known Be stars.

We measured the velocity separations (${\Delta}{{v}_{{\rm p}}}$) of the violet (V) and red (R) emission peaks of all lines with well-defined peaks in all spectra of sufficient quality (typically ${\rm s}/N\gt 50$; dependent on emission strength) using the SPLOT feature of IRAF. Measurements pertaining to emission peaks coincident in wavelength position with strong airglow lines or diffuse interstellar bands (DIBs) (see Section 4.1) were thrown out. No attempt was made to remove underlying photospheric absorption prior to ${\Delta}{{v}_{{\rm p}}}$ measurement.

Table 1 provides the ABE identifiers, star names, 2MASS H magnitudes, literature spectral types and references where available (see Section 3.2), and the mean ${\Delta}{{v}_{{\rm p}}}$ for the Br11 line from all APOGEE spectra for each source. Star names beginning with "J" are 2MASS designations, and newly identified Be stars are indicated by bold font for the ABE ID. If a ${\Delta}{{v}_{{\rm p}}}$ measurement for the Br11 line could not be made in any of the available spectra despite evidence of Br11 emission or shell absorption, one of the following abbreviations is provided in place of a ${\Delta}{{v}_{{\rm p}}}$ value: "w" weak emission-peaks not discernible; "sp" single-peaked emission; "sh" shell absorption without resolved adjacent emission peaks; "as" severe asymmetry in emission peak heights such that only one peak is discernible (not the same as single-peaked); "bl" V peak of Br11 is severely blended with Fe ii 16792 (ABE-013); "tc" spectra are heavily contaminated by telluric features (ABE-058).

Table 1. List of ABE Stars

ABE2MASSStar2MASSLit.Ref. ${\Delta}{{v}_{{\rm p}}}$
IDDesignationName H Spectral Br11
   (mag.)Type (km s−1)
00120212485+3722482VES 2039.108B0.5Ve74208
00220151525+3654562HD 2285769.888Ae29152
00320162816+3703229HR 77576.548B6IIIe22305
00420234596+3830033HD 2292216.734B0.2IIIe66115
00520184170+3759106Hen 3-18769.699OB6271
00619124025-0627316HD 1794058.084B2Ve68294
007 19104149-0542581HD 1789209.236B8II/III58181
008 18000176-2323071 10.699  233
009 20461437+5039005TYC 3586-282-19.192(B8)C175
01020450869+5033004BD+50 31889.311(B)C138
011 20535693+5005293TYC 3583-670-19.698  252
012 20554731+5040274 10.766  115
01318574179-0419113EM* CDS 103810.428OB24bl
01421300088+4529390HD 2048606.931B5.5Ve60326
01518123846-2708292HD 1666299.155B5nne4170
01618185069-1227145HD 1681357.406B8Ve59238
01718122758-1546123BD-15 48639.695Be24376
018 18194798-1724130 10.976  sp
019 23533653+5649116HD 2239248.177B1.5:III:n25296
02018432516-0339100SS 41210.528OB:e33217
021 04220085+5430434HD 2329408.625(B9)C291
022 04254177+5615294TYC 3727-1849-19.438  495
023 03282223+4507560BD+44 709s10.546OB14386
024 05113282+2408029TYC 1846-17-19.596(A3)C158
025 05452088+2909281HD 2470429.165O9.5:15346
02605485364+2908100HD 387088.015B3:e:psh8324
027 05453713+3007253TYC 2405-1358-19.825  489
02806123985+3258216HD 425298.090B9V57297
029 06165595+3414299HD 2541689.169  301
030 06185921+3413502HD 436818.806(A2)39333
03106273614+1815476HD 2574739.378B5e5133
03222252246+5642384SS 45310.200Be:33508
033 23181131+5550356HD 2402499.250(B8)C205
03423521212+6710073MWC 10858.787(B3Ve)54551
03506334350-3202486Hen 3-149.650B24241
03606283925-3222165HR 23646.113B3Ve69456
037 05590290+3101488HD 402549.210(B8)39234
038 20032620+2242411HD 34558910.595(A3)52224
03918000839-2356576SS 3389.647B8e:33259
04017375482-2457569HD 1598456.773B3IIIe4599
041 18382765-1014211 10.640  141
04218464650-1021523BD-10 479910.094(B)C181
04317441414-2727284HD 1610047.880B9IVe41290
04417432344-2715411HD 3161797.673Be:35247
045 01542524+5651061TYC 3692-1234-110.321  344
046 02135183+5354525HD 135449.055B0.5IIIn27294
047 05390426+3758359HD 372667.309B8V57250
048 21375002+4258309BD+42 41628.916(A0)39333
04921103095+4741321HR 81076.360B6IV22197
05019584848+2305215HD 34543910.629B2Vpe771153
051 19571220+2158541HD 3454399.869(A0)52373
052 18052711-2921573TYC 6854-2016-110.009(B8)61405
053 18023404-3027418HD 3170269.467(B9)52216
054 05064093+2230389HD 328116.523B8.5V57315
055 06591343+0427179HD 518939.077B9V58344
056 00302445+6010093HIP 23829.710B6III48as(sh)
057 02534457+6443068TYC 4056-415-19.293(B5)61sh
05818072725-2506165HD 1655176.967B0Iae45tc
059 18144388-2137597HD 1671139.030(B9IV:)61415
060 07134058+3806302HD 552008.297(A0)39sh
06104361726+5230217HD 289428.159Ash44sh
062 03030934+6654223TYC 4060-96-18.398  456
063 06351041+0634180HD 2601539.417B8III42w
064 18451050-0545120TYC 5126-2325-110.733  291
065 18434859-0608188HD 1730759.358B9IV58235
066 18402120-0455127TYC 5121-940-110.299  187
06703292627+4656162HR 10475.902B7Vne51133
068 17224642-2533273HD 1571748.488A0IV45208
069 18274975-1104312 10.165  238
07018251991-0918163BD-09 47249.550(A0IV)Cw
07118424368-1000273TYC 5696-503-110.454Be33138
07218043735+0155085HD 1651746.224B0IIIn22w
073 22583631+5528116BD+54 28879.538(A0)61193
074 19233702+3859363HD 1825508.818B8V4766
07503464087+3217247HD 234786.486B3IVe10913
07604432066+4754385VES 82810.896  301
077 04423114+3830469 10.455  162
078 21351726+5647589TYC 3975-1585-110.101B812400
079 07081479-2325541HD 545518.775B1.5II45524
080 20282074+4526025BD+44 34759.451  sh
081 00104514+5801058HD 6287.521(B9)39205
082 05540108+1241033BD+12 93810.172  250
083 05453721+1311210HD 2472999.987(A0)62119
08401042742+5756263HD 2366118.978(A)C265
08501183306+5822304HD 2366898.853B1.5V:epsh8321
086 01355734+5809128TYC 3683-1262-19.840  209
08707331124-1136421HD 602609.232B3Ve58129
08801055296+6558158HD 63436.843B5Vn:e20115
089 00573323+6709339TYC 4029-428-19.595  342
090 00501808+6710377BD+66 648.588(B9)Csh
091 05564423+1601018HD 399849.140(A2)39115
092 03250006+5029394TYC 3320-1906-19.425B79417
093 00080292+7332356TYC 4306-1125-19.115B8V37337
09400331160+5140069HD 2322148.855(B8)C92
095 06272285+0824429HD 453968.907(A2)39w
096 06300018+0817453HD 458288.506(B8?)Cw
09705164674+3022455HD 341938.498(A0e)39203
098 23152849+6416002HD 2195237.218B5V36301
09906073989+2751354HD 416398.599B6Vne:38578
100 05341240+4516410HD 364678.167B9III37300
101 05124298+4754272BD+47 11089.581(A0)39326
102 05003547+3552170TYC 2400-1784-110.396  273
103 17494627-2249517 10.473  <61
104 18411366-0247380BD-02 46988.853(B)Csh?
105 20451060+5112379HD 2353508.652B0.5IV8502
106 02250591+5515032TYC 3690-1236-110.581  480
107 21563126+5041249TYC 3617-2074-110.112  233
108 06185755+2323286HD 2548428.631  185
109 06231994+2506057HD 2561379.734(A2)39106
11004365908+5217135HD 290357.930B9.5Ve64w
11118581515-0528567AS 3329.639Be43w
112 04321707+4816572HD 285437.725(A0)39w
113 04363913+4104368HD 290967.323B8IV57333
114 04460607+4705516TYC 3347-1615-110.713  sh?
115 20185676+3745319 11.326  139
116 17331509-1922379HD 1590328.725B9IV45387
117 17515926-3029411HD 1623458.291(B8)39128
118 17534729-2945087HD 3165739.850(B9)52115
11917531191-2857284AS 2519.961B24259
120 17521395-2744257HD 3164759.225(B9)39282
121 17525570-2218434TYC 6262-3203-19.292  166
12218194176-1058093TYC 5681-507-110.027(A5)62179
123 18220126-1048042TYC 5681-151-110.614  177
124 18161427-2906365HD 1674019.322B4II/III41627
125 18221389-1307360TYC 5689-54-110.273  <60
126 18245968-1406408HD 1694188.995B9.5III45326
127 18205460-1243598HD 1685668.710B9III45114
12818404465-0758241HIP 915918.825B8Ve49276
129 18385819-0827466GSC 05692-0054010.451B717267
130 18405017-0741018GSC 05692-0039910.508B717423
13118395898-0733138BD-07 46479.642B517143
132 18355878-0744307BD-07 46308.963B917w
13317500331+482339188 Her6.913B6IIInpsh53285
134 18295996-0908375 10.761  198
13518432970-0919127HD 1730107.179O9.7Ia73sh?
136 18064578-2821496HD 1653657.024B7.5III41232
137 18042703-2228572NGC 6531 F19511.130  <25
13819150144+0948272HD 1801267.565B2IV68365
13919164642+1058468HD 1805877.578Apsh18sh
14020234436+3728351HR 78076.228B2Vne63382
141 21061887+2824477HD 2010368.996B6/8Vn38247
142 19270008+1632172HD 1830357.844A0V70320
143 20040584+3009117HD 33337810.150(A0)52313
144 20005874+3113497HD 1898477.122B7V16260
145 17530194-2219531TYC 6262-1413-19.926  sh?
146 06072002+2640558HD 416007.132B9IV57w
147 19424993+4239003HD 1864858.484B9V34123
148 20002133+2135515HD 3455069.683(B8)52sh?
149 03434449+3143092IRASF03406+313310.780  sp
150 18412551-0534033 10.902  388
151 18404500-0740458TYC 5692-1370-110.799B717137
15207290132-0832539SS 12010.733B8e:35461
153 17221970-2833450 10.940  sh
154 05394249+2215279TYC 1310-2084-19.969(B8)61360
15501590196+5725521TYC 3692-1671-110.611  819
156 23380341+5556420HD 2221858.343(A2)39w
15719450599+1617091HD 1866377.937(B9e)39359
15821520306+5853123AS 4789.792  87
15922275192+6300090MWC 10628.804B5:e3265
16021365704+6811073HD 2061357.811B3V21316
161 21551055+5326166TYC 3968-1354-110.574OB-14422
162 06063872+2754038BD+27 9819.960(B8)62219
163 22465987+5345241HD 2158378.104(A0)39357
16422202269+5151395HD 2120447.702B1:V:nnep8191
16522245295+5207583HD 2126668.681B5.5e59397
166 06460565-0558109TYC 4812-2496-19.971  308
16706495552-0530472HD 497877.836B1Ve68300
16806570947-0832309HD 514778.223B3Ve58628
169 06024105+2202482HD 408978.000(B9)39255
170 06014161+2224036HR 21166.400B8V26153
171 06081219+2156586TYC 1326-1188-110.257(A2)62362
172 05212545+1601440HD 349068.641(B9V)61142
173 05175643+1519211TYC 1283-1360-110.617  sh
174 05240938+1633455HD 352697.552A0V76448
175 05214456+1709194TYC 1300-652-110.731  w
176 05133229+3806348HD 336569.215B52551
177 05194374+3820304HD 2808499.654(B3)52522
178 06450343-0034140CoRoT 10276253611.651B1V75sp
17906450928-0115205EM* RJHA 5110.563B5Ib75243
18006422978+0053582EM* RJHA 4010.613B3Ib75183
18105355408-0537423HD 371157.204B7.5V58238
182 06123495+4145095TYC 2934-118-110.240  376
183 07364229-1303499HD 609939.131B2II45348
184 05361555+3257145HD 2451749.790(B3)62268
185 05580029+2437017HD 401327.564B91985
186 07075533+0143105HD 541679.672B5/7Ib:58328
18706121659+2005097HD 2532148.917B1.5:V:nn8220
18806121386+2000034HD 25321510.444  298
18907370572+1654153HD 608487.071O9.5IVe67305
190 06110671+1810591HD 2529049.089B9V11w
19118084894-1858344HD 1658547.903B9e59274
192 18103823-1910006HD 1662918.303B3II45272
19318011841-2721498SS 33910.713B8e:33250
194 18173492-1842282HD 3130629.676  327
19519120326+0237212HD 1793436.613B8III68266
19605010612+4134002HD 27724110.836B813w
19718095327-2302251HD 1660559.662(B9)62221
198 18091443-2246378HD 1658948.183B3IV/V45158
199 17554711-2142367TYC 6262-371-19.301  164
200 18235550-1547477Lan 6719.976  280
201 18282453-1642195BD-16 4888p9.825  138
202 18283909-1512088TYC 6266-143-110.487  w
A0100201742+6227498MWC 58.025B0.5IV7122
A0200320285+6709401HD 27897.519B3:Vne23320
A0304042164+5319447MWC 807.165B1Vnnpe8as
A0405251782+2936535HD 353477.948B2:nne38268
A0505445623+2127384HD 381918.380B1:V:ne:8246
A0605503228+1801349HD 390187.898(B9)39w
A0706183944+2300285HD 437037.475B1IV:p?8w
A0806330559+1656553HD 462647.522B5Vne31409
A0906333223+0820080HD 2595977.779B1Vnne28289
A1006525305-1000270HD 504248.971B8e35178
A1106545882-0342013HD 508917.880B0.5Ve68215
A1206561908-0348254HD 511938.039B1.5IVe68358
A1306574289+1754071HD 513547.183B3Vn40257
A1407093697-1605467HD 547869.051B1.5Ib:45256
A1507133410-0204390HD 556068.704B0.5Vnnep69237
A1619525141+2214226HD 3451228.963B3Ve46233
A1720024644+2151160HD 1901508.186B6IV-Ve40241
A1821082962+4715254HD 2015228.022B0V16 
A1921250244+4427063MWC 6407.206B1.5V:nnep8205
A2021291483+4420173HD 2047227.643B1.5IV:np25358
A2122013820+5010046MWC 6498.701B3e1109
A2222060834+4954088AS 4839.631B1.5V:nne:8362
A2318211606-1301256MWC 9227.396unclB[e]55sp
A2405181018+3739003HD 343027.534(B8)39163
A2505231490+3742536HD 2809999.582(B3)52214
A2605254477+3538499HD 353457.851B1Vep7sp
A2705280968+3516540EM* CDS 4968.669OB24w?
A2805595354+2505190HD 2500288.083B2:V:nep32288
A2905530609+2626435HD 393407.579B3Ve59197
A3005533110+2544321HD 2487537.361B2:Vnne23279
A3105535984+2625212HD 394787.691B2Ve59242
A3206072661+2205477IGR J06074+220510.189B0.5Ve71380
A3306074953+1839264EM* LkHA 2089.834A7e65sp?
A3406065436+1902040HD 2517267.644B1V:e8159
A3506184553+1516522MWC 1377.840sgB[e]5657
A3606135416+1631049HD 2536598.327B0.5:V:nne8312
Q0120240654+3829332HD 2292397.093B0.2III66 
Q0220194162+3811060HD 2289329.386B35 
Q0320213867+3725151BD+36 40327.566O8.5V66 
Q0420213589+3721395Hen 3-188510.855A0V30 
Q0506382991+0042351HD 2919469.321B92 
Q0618071193-2516305HD 31517710.022   
Q0719582996+2033006HD 35098910.534B7IIIn50 
Q0820012170+2217258HD 3454759.484B04 
Q0923204452+6111404EM* CDS 14597.471O6.5(f)(n)p72 
Q1000350607+6258585EM* CDS 5310.369OB-e:14 
Q1100310135+5539101HD 2322089.519B3:e3 
Q1200255124+7148258HD 20837.041O9.5III34 
Q1301243585+5812454EM* CDS 14410.499OB-e:14 
Q1404310304+4146289HD 27641410.148(B8)52 
Q1504390489+4115001HD 293328.246B3ne1 
Q1605085056+4144262HD 329618.970B213 
Q1704360336+3640031HD 2800067.747A0Ibe:33 
Q1806321639+0110289HD 2888059.611B52 
Q1906594264-1109265HD 521599.793B3Vne69 
Q2005204307+3726192HD 346566.634O7.5(f)II73 
Q2107213463-0553498HD 575396.834B3IV58 
Q2206154017+0603582HR 22316.336B6Ve69 
Q2322061730+6355026EM* CDS 129910.236OB24 

References. (1) Merrill et al. (1942); (2) Cannon & Mayall (1949); (3) Merrill & Burwell (1949); (4) Popper (1950); (5) Miller & Merrill (1951); (6) Nassau & Harris (1952); (7) Morgan et al. (1953); (8) Morgan et al. (1955); (9) Heckmann et al. (1956); (10) Hiltner (1956); (11) Duflot et al. (1958); (12) Alknis (1958); (13) McCuskey (1959); (14) Hardorp et al. (1959); (15) Bouigue et al. (1961); (16) Fehrenbach et al. (1962); (17) Roslund (1963); (18) Feast & Thackeray (1963); (19) McCuskey (1967); (20) Schmidt-Kaler (1967); (21) Racine (1968); (22) Lesh (1968); (23) Guetter (1968); (24) Wackerling (1970); (25) Walborn (1971); (26) Cowley (1972); (27) Lesh & Aizenman (1973); (28) Turner (1976); (29) Henize (1976); (30) Voroshilov et al. (1976); (31) Davis (1977); (32) Christy (1977); (33) Stephenson & Sanduleak (1977); (34) Hill & Lynas-Gray (1977);(35) Stephenson & Sanduleak (1977); (36) Roman (1978); (37) Bartaya (1979); (38) Clausen & Jensen (1979); (39) Ochsenbein (1980); (40) Jaschek & Jaschek (1993); (41) Houk (1982); (42) Voroshilov et al. (1985); (43) Bopp (1988); (44) Bidelman (1988); (45) Houk & Smith-Moore (1988); (46) Radoslavova (1989); (47) Sato & Kuji (1990); (48) Turner et al. (1992); (49) Grillo et al. (1992); (50) Turner (1993); (51) Garrison & Gray (1994); (52) Nesterov et al. (1995); (53) Abt & Morrell (1995); (54) Kohoutek & Wehmeyer (1997); (55) Lamers et al. (1998); (56) Esteban & Fernandez (1998); (57) Grenier et al. (1999); (58) Houk & Swift (1999); (59) Yudin (2001); (60) Chauville et al. (2001); (61) Kharchenko (2001); (62) Fabricius et al. (2002); (63) Abt et al. (2002); (64) Miroshnichenko et al. (2003); (65) Hernández et al. (2004); (66) Negueruela (2004); (67) Negueruela et al. (2004);(68) Frémat et al. (2006); (69) Levenhagen & Leister (2006); (70) Uzpen et al. (2007); (71) Reig et al. (2010); (72) Walborn et al. (2010); (73) Sota et al. (2011); (74) Mathew & Subramaniam (2011); (75) Sebastian et al. (2012); (76) Chargeishvili et al. (2013); (77) Eikenberry et al. (2014).

Download table as:  ASCIITypeset images: Typeset image b c d

The ABE identifiers were assigned to avoid the use of sometimes lengthy survey identifiers which are the only star names available. Three groups of ABE stars are distinguished from one another by ABE ID as follows:

  • (1)  
    ABE-001–ABE-202 refer to Be stars that were quasi-randomly targeted by APOGEE as TSS and subsequently identified as Be stars through visual inspection of the wavelength region encompassing Br11 and Fe ii 16878. To account for sources only producing emission lines in certain epochs, which was frequently the case, it was necessary to examine all > 70,000 individual spectra for all > 17,000 telluric stars.
  • (2)  
    ABE-A01–ABE-A36 refer to Be stars that we targeted intentionally via internal proposals for APOGEE observations of ancillary (hence the "A" prefix of the ABE IDs) science targets falling within a subset of pre-planned APOGEE fields. Most of the intentionally targeted Be stars are early-type (B3 and hotter) classical Be stars, showing stronger than average H-band emission in the APOGEE spectra, but two stars classified as B[e] in the literature were observed (ABE-A23 and ABE-A35) as was a reported Herbig Ae star (ABE-A33).
  • (3)  
    ABE-Q01–ABE-Q23 refer to stars which (a) had existing "emission line star" classifications in the literature, (b) appeared to be hot OBA stars in the APOGEE spectra, but (c) did not produce any discernible emission in any of the associated APOGEE data (all have multi-epoch data), or in others words, were H-band quiescent (hence the "Q" of the ABE IDs) during the observations.

3.2. Literature Spectral Types

The He i 17007 line, analogous to optical He i in terms of utility as an effective temperature (Teff) diagnostic for OB stars (Blum et al. 1997; Meyer et al. 1998) and the only non-hydrogen stellar absorption feature expected to be present for B-type stars (the earliest-O stars exhibit He ii 15723, 16923 absorption), is not covered by APOGEE spectra. Therefore, detailed spectral classification of OB stars is not possible with these data and the literature was perused for the existing spectral classifications included in Table 1.

The Catalogue of Stellar Spectral Classification (CSSC; Skiff 2013) was the primary resource used for locating spectral type information, but some of the original sources of spectral types in the CSSC (primarily those pre-dating 1940) could not be tracked down. In those cases, the spectral types are enclosed in parentheses and the provided reference is "C." Other second-hand spectral types, culled from modern compilations of historical data, are also enclosed in parentheses. The spectral types not enclosed in parentheses are therefore those that could be linked directly to the paper or catalog where the spectral type was determined or estimated.

3.3. New Be Star Discoveries

A total of 128 Be stars have been identified as Be stars for the first time via Brackett series emission in APOGEE spectra. According to the BeSS (Neiner et al. 2011), which maintains a comprehensive database of classical Be and main sequence B[e] stars, 2070 Be stars are cataloged in the Milky Way and Magellanic clouds combined. The 128 new Be stars presented in this work therefore represent a $\sim 6.2\%$ increase in the number of known Be stars.

The positions of all 238 ABE stars are shown in Figure 2, along with the Be star entries included in the BeSS database (Neiner et al. 2011). Although APOGEE observes a large number of fields in the Galactic Halo, the majority (90%) of Be stars observed during the survey reside along the plane of the Milky Way (at Galactic latitudes, $\mid b\mid \lt 10$), similar to the trend seen in BeSS.

Figure 2.

Figure 2. This adaptation of Neiner et al. (2011) Figure 1 shows the RA and decl. positions of all the BeSS entries as black dots, known Be stars observed by APOGEE as squares (blue), and new Be stars discovered in the APOGEE survey as triangles (red). The Galactic Center and Magellanic Clouds are labeled.

Download figure:

Standard image High-resolution image

Stars included in the ABE sample were generally required to exhibit evidence of emission or shell absorption in at least the H i Br11 line. The exceptions to this rule are ABE-111, ABE-196, and ABE-A06; these stars appear clearly to have emission from Fe ii 16878 (see Section 4.2) despite very weak or no emission in the Brackett lines. Figure 3 shows examples of new and previously known Be stars representing the most borderline cases included the ABE sample. For many of these stars, the Br11 emission is sufficiently weak that double-peaks are not discernible. Rather, the photospheric Br11 absorption wings appear filled in with emission, creating "shoulders" on the line profiles (e.g., ABE-112) and making them easily distinguishable from purely photospheric lines profiles. Fe ii emission is also apparent for a number of these stars, despite weakness of the H i emission.

Figure 3.

Figure 3. Br11 region for some new and previously known Be stars showing very weak evidence of circumstellar emission.

Download figure:

Standard image High-resolution image

The large number of new Be stars identified by APOGEE is due in large part to the high-resolution, high-S/N spectra which permit identification of very weak disk signatures (e.g., Figure 3) that might be overlooked in lower-resolution spectra or narrow-band photometry. Repeated observations of most of the stars (> 1 observation for 93% of sample) can provide confirmation of very weak disk signatures and also reveals transient Be disks, where Brackett series emission either fades away or appears unexpectedly from epoch to epoch (J. P. Wisniewski 2015, in preparation). Among the reasons for 86/128 newly identified ABE stars having been classified in the literature as normal O-, B-, or A-type stars is that the stars did not possess CS disks at the time the spectral types were determined or estimated.

3.4. ABE-144 and ABE-170: the Brightest New Be Stars

The brightest newly identified Be star among the ABE sample is ABE-170, a.k.a. HR 2116 (V=6.36), and the second brightest is ABE-144, a.k.a. HD 189847 (V = 6.92). The lack of a prior indicator of emission lines for ABE-144 may be due to few or no historical spectroscopic observations of the star beyond Fehrenbach et al. (1962). It is unclear whether or not Balmer series emission would have been noticed in that study. ABE-170 has been observed spectroscopically in more recent studies including Abt et al. (2002) and Strom et al. (2005), but the spectra used in both of those papers were limited in wavelength coverage to the region encompassing He i 4471 and Mg ii 4481, such that emission at Hα or Hβ would not have been recognized if present. Among the possible reasons for the Be nature of these stars not having been previously recognized is that ABE-144 and ABE-170 were normal B stars during past observations (similar to, e.g., Bjorkman & Miroshnichenko 2000).

3.5. Observed Emission Lines

The emission lines detected in APOGEEʼs wavelength range are listed in Table 2. For each line, Table 2 provides the (1) line identification, (2) laboratory rest wavelength, (3) observed wavelength (see concluding paragraph of this section), (4) the difference between laboratory and observed wavelengths, (5) lower level energy Ei , (6) upper level energy Ek , (7) transition strength expressed as log(${{g}_{i}}{{f}_{ik}}$), (8) for metallic lines only, the number of confident and possible detections, (9) the number of stars for which ${\Delta}{{v}_{{\rm p}}}$ was measured, (10) the range of ${\Delta}{{v}_{{\rm p}}}$ measurements, (11) the average of all ${\Delta}{{v}_{{\rm p}}}$ measurements, (12) and other transitions possibly contributing to the observed emission line profiles.

Table 2. Observed Emission Lines and Summary of ${\Delta}{{v}_{{\rm p}}}$ Measurements

(1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)
Atom or ${{\lambda }_{{\rm v}ac}}$ ${{\lambda }_{{\rm v}ac}}$ Diff. lab–obs   N ${\Delta}{{v}_{{\rm p}}}$ ${\Delta}{{v}_{{\rm p}}}$ ${\Delta}{{v}_{{\rm p}}}$ Other
Ionlabobs a   Ei Ek  detectionsNrangemeanpossible
 (Å)(Å)(Å)(eV)(eV)log(gf)yes (maybe)stars(km s−1)(km s−1)contribution
H i (Br20)15195.99615195.9320.06412.74913.564−1.487⋯4868–566314⋯
H i (Br19)15264.708⋯⋯12.74913.561−1.414⋯⋯⋯⋯DIB 15271
Fe I15298.74015298.5280.2125.3096.1190.65063188–283228⋯
H i (Br18)15345.98215345.987−0.00512.74913.556−1.337⋯5767–533276⋯
H i (Br17)15443.13915443.187−0.04812.74913.551−1.255⋯7366–537266⋯
H i (Br16)15560.69915560.6970.00212.74913.545−1.166⋯4765–436266⋯
N I15586.54515586.591−0.04612.12612.922−0.0231152–5252⋯
H i (Br15)15704.95215705.015−0.06312.74913.538−1.071⋯7665–552282⋯
Mg I15753.291blend⋯5.9326.7190.1407⋯⋯⋯Mg i 15745
$\lambda 15760$ ⋯15760.161⋯⋯⋯⋯36 (13)1127–304161Mg i
Mg I15770.14915770.943−0.7945.9336.7190.4119 (1)2338–375356 $\lambda 15760$
H i (Br14)15884.88015884.8750.00512.74913.529−0.967⋯9161–522259⋯
Si I15892.771blend⋯5.0825.862−0.0076⋯⋯⋯⋯
Si I15964.42215963.2291.1935.9846.7610.1987 (1)3321–383345⋯
C I16009.270blend⋯9.63110.4060.2347 (3)⋯⋯⋯⋯
C I16026.080blend⋯9.63110.4050.2225 (5)⋯⋯⋯⋯
H i (Br13)16113.71416113.766−0.05212.74913.518−0.852⋯9661–517249⋯
H i (Br12)16411.67416411.763−0.08912.74913.504−0.725⋯9559–524252⋯
Ca II16565.59016565.973−0.3839.2359.9830.36863208–333291C i 16564
Ca II16654.430blend⋯9.2409.9840.6261 (2)⋯⋯⋯⋯
Si I16685.341blend⋯5.9846.727−0.1171 (2)⋯⋯⋯⋯
Mg II16764.80016764.922−0.12212.08312.8220.4812 (2)134–3434⋯
$\lambda 16781$ ⋯16781.115⋯⋯⋯⋯36 (13)1328–309157⋯
Fe II16791.76216791.953−0.1915.4846.222−2.3258 (6)148–4848⋯
Mg II16804.520blend⋯12.08512.8220.7372 (2)⋯⋯⋯Mg ii 16804
H i (Br11)16811.111⋯⋯12.74913.486−0.582⋯19457–1153282⋯
Fe II16877.80816877.822−0.0145.4846.219−1.25676 (33)2624–23282⋯
C I16895.03116894.8980.1339.0039.7360.53443 (19)1431–539243⋯
Si II16911.43016911.646−0.21612.14712.8800.3501 (1)141–4141⋯

a The emission peak midpoint corrected by the emission peak midpoint of Br11.

Download table as:  ASCIITypeset image

Attempts to identify all non-hydrogen (metallic) lines, described in Section 4, made use primarily of Peter van Hoofʼs Atomic Line List v2.05b16 14 (PLL from here on) and to a lesser extent the NIST Atomic Database (Kramida et al. 2013) and the Kurucz line list. 15 The identities of metallic emission lines at 15760 and 16781 Å remain ambiguous due to few transitions around the correct wavelengths having available transition probability data needed for confident identification. These lines, referred to as $\lambda 15760$ and $\lambda 16781$ throughout this paper, are discussed in more detail in Section 4.5. Since forbidden line emission was present for only one source, ABE-A23 a.k.a. MWC 922, the central star of the Red Square Nebula (Tuthill & Lloyd 2007), Table 2 is limited to the permitted (E1) transitions observed for Be stars. The H-band spectrum of MWC 922 is sufficiently different from the rest of the sample and sufficiently more complex that an in-depth analysis is currently being pursued separately (Whelan, in preparation).

The observed wavelengths as well as the differences between laboratory and observed wavelengths, provided in columns (4) and (5) of Table 2, pertain to the average position of double emission peaks for each line plus a correction factor based on the Doppler shift found for the Br11 line. Br11 is the strongest line covered for these stars and provides the most reliable peak position measurements, so correction to rest frame was done simply by adding to the observed wavelength of each line the difference between Br11 emission peak midpoint and Br11 rest wavelength.

4. NON-HYDROGEN LINE IDENTIFICATION

4.1. Diffuse Interstellar Bands

The DIB at 15271 Å, discovered by Geballe et al. (2011), is present in most of the ABE spectra and in numerous APOGEE spectra (Zasowski et al. 2014). Because DIB 15271 usually falls on or near the Br19 R emission peak, Br19 peak separation measurements are omitted from this paper. Examples of DIB 15271 absorption (marked with red dotted lines) in spectra for four active Be stars and two currently emission-less stars are displayed in Figure 4. Other DIBs (15615, 15651, 15671 Å) discussed in Geballe et al. (2011) are present for most objects with DIB 15271, as are other possible DIBs at ∼15314 and ∼16154 Å. Of the spectra shown in Figure 4, DIB 15314 appears most prominently in the spectrum for ABE-137.

Figure 4.

Figure 4. Spectra for six stars with visible DIB 15271 absorption around or on the Br19 line. No correction for radial velocity has been applied to the spectra. The dotted lines (red) mark DIB 15271 and another likely DIB at ∼15314 that appears in numerous APOGEE spectra.

Download figure:

Standard image High-resolution image

4.2. Fe ii

The Fe ii 16878 line appears in emission for between 32%–46% (upper limit includes weak or ambiguous detections) of the 238 active Be stars discussed here, making it the most frequently observed metallic feature in APOGEEʼs coverage of the H band. For stars with very strong Fe ii 16878, the much weaker Fe ii 16792 also appears in emission but is usually blended with Br11. As for Fe ii 16878, proximity of the feature to C i 16895 often leads to a blend of the two lines, especially since C i emission is always broad compared to Fe ii (see Section 5.2).

Examples of stars with emission from one or both H-band Fe ii lines are presented in Figure 5. The left panel demonstrates the wide range of H i strength corresponding to Fe ii detections. As is seen quite clearly for the lowermost stars (ABE-A06, ABE-111) in the left panel of Figure 5, Fe ii emission may be present even when there is no perceptible emission from Brackett series lines, contrary to the finding of Steele & Clark (2001). ABE-A06 has been a Be-shell star at various epochs (see BeSS spectra), but in the APOGEE data exhibits only very weak filling of the Br11 photospheric absorption wings in addition to the weak Fe ii 16878 emission that, for ABE-A06, persists in four spectra sparsely covering 0.77 years. Less evidence is available for H i emission in the case of ABE-111, despite the Fe ii feature appearing in all six APOGEE covering 2.29 years. Though not shown in Figure 5, ABE-196 also lacks convincing evidence of H i emission and yet exhibits Fe ii 16878 emission in all 13 APOGEE spectra covering 3.02 years. Line profile variability in the Brackett lines is observed for all three stars and is likely due to varying degrees of emission filling, but lack of knowledge of the true photospheric absorption profiles prevents us from confidently claiming H i emission is present.

Figure 5.

Figure 5. Be star spectra with combinations of emission from $\lambda 16781$, Fe ii 16792 and 16878 (blue), and C i 16895 over a spectrum of Brackett series emission strength. The left and right panels show the same wavelength and intensity ranges. A dotted line separates the unclassified B[e] star ABE-A23 (MWC 922) from the other stars; ABE-A23 is unique among this sample (see Section 3.5) in being the only source to show forbidden line emission (mostly [Fe ii]).

Download figure:

Standard image High-resolution image

The right panel of Figure 5 focuses on some of the more extreme Be stars in this sample, starting with the obvious outlier ABE-A23, an unclassified B[e] star (Lamers et al. 1998) for which the exceptionally strong Fe ii and [Fe ii] emission lines reported by Rudy et al. (1992) dominate the APOGEE spectra. In contrast to ABE-A23, where the emission lines all appear truly single-peaked, ABE-137 is likely a classical Be star viewed at an inclination, i, of nearly or exactly zero. The Brackett series lines for ABE-137 show some peak structure even though the peaks are not resolved. The Fe ii 16878 line is very narrow and pointed but double peaks are resolved in the C i 16895 line, suggesting the presence of a circumstellar disk. ABE-A35 exhibits strong Fe ii emission, and is the only source of the ABE sample for which the ${\Delta}{{v}_{{\rm p}}}$ of Fe ii 16792 could be measured. The average Fe ii peak separations from five spectra for ABE-A35 are in good agreement: ${\Delta}{{v}_{{\rm p}}}$(Fe ii 16878) = $52.6\pm 2.53$km s−1, ${\Delta}{{v}_{{\rm p}}}$(Fe ii 16792) = $48.3\pm 4.01\;$ km s−1. The lower resolution of red detector data from APOGEE commissioning data is likely a factor in the single-peaked appearance of the Fe ii lines for ABE-015.

4.3. Fe ii Profiles as a Function of Inclination

In past studies, optical Fe ii lines have been used (Hanuschik 1996) to establish a strict definition of the shell, or edge-on, class of Be stars. Photospheric Fe ii absorption lines are usually observed at greatest strength for A−F supergiants (Gray & Corbally 2009), so if the central depression of an Fe ii emission line for a Be star extends below undisturbed, adjacent continuum level, the implication is that the the disk is viewed at sufficiently large inclination that our line of sight passes through an appreciable volume of cool gas in the inner, equatorial disk. It is a well-known fact that Fe ii and Ti ii shell lines are among the strongest metallic features present in the spectra of edge-on Be stars.

In contrast to the observed behavior of optical Fe ii lines, stars with obvious shell absorption in the Brackett series lines exhibit no evidence of shell absorption in the H-band Fe ii lines nor in any of the covered metallic lines, such that the Fe ii line profile shapes for pole-on Be stars differ from those of edge-on Be stars only in line width. This fact is demonstrated in Figure 6, where the upper panel compares Br11 and Fe ii profiles for five stars viewed over a range of inclination angles. As can be seen, the Fe ii profiles are pure emission regardless of the what form the H i profiles take. The lower right panel of Figure 6 presents additional examples of H i-shell stars with Fe ii emission, while the lower left panel (as well as the edge-on example in the upper panel) highlights ABE-035, the most extreme shell star within this sample in terms of Brackett series shell depth.

Figure 6.

Figure 6. (Top panel) An assortment of observed spectra, showing the variety of metallic and hydrogen line profiles observed in the sample as a function of inclination angle. (Left, bottom) A portion of a spectrum for ABE-035, highlighting the immunity of metallic emission lines to shell absorption, a fact that is observed in all shell absorption sources (examples shown on right, bottom).

Download figure:

Standard image High-resolution image

4.4. C i and Other Neutral Lines

An emission line at 16895 Å is identified for the first time as C i 16895.031 and is observed for between 18–26% of the 238 Be stars. Figure 5 displays 11 examples of stars with C i emission and the strongest detections will be discussed in Section 4.4.1. A C i 16895 absorption line is present in numerous APOGEE spectra of A−F stars, but the line is typically not present for OB stars unless in emission. Except in the case of very narrow-lined Be stars (e.g., ABE-015, ABE-040 in Figure 5), the R peak of the C i 16895 emission profile is frequently compromised by a strong airglow line around ∼16904 Å.

Prior mentions in the literature of NIR C i emission include Groh et al. (2007) and Štefl et al. (2009), where several C i emission lines were detected around 10700 Å in Be star spectra. Spectra showing C i 16895 emission have been included in a number of papers, but the line is usually either confused and/or blended with Fe ii 16878, or not identified at all. Ashok & Banerjee (2000) noticed the C i 16895 line in a subset of low-resolution Be star spectra and realized that it was probably not Fe ii 16878 due to the measured wavelength of the line (∼16893Å, or ∼15Å from the Fe ii wavelength). Kendall et al. (2003) presented medium resolution H-band spectra of three young stellar object (YSO) candidates, one of which, IRAS 17441–2910, was found to be a very strong emission line source. A plot of the spectrum shows single-peaked Br11 and Fe ii 16878 emission and strong double-peaked C i 16895 emission, but the authors did not comment on the latter.

NIR emission from C i is not limited to classical Be stars. C i 16895 emission was present in a high-resolution spectrum shown by Kraus et al. (2012) for the Herbig B[e] star V921 Scorpii, and the C i emission line is also present in APOGEE spectra for both B[e] stars observed by APOGEE to date (see right-hand panel of Figure 5). Even luminous blue variable stars display evidence of C i emission lines (Groh et al. 2007), suggesting that NIR C i emission is ubiquitous across a wide range of evolutionary states.

4.4.1. C i-strong Be Stars

Abnormally strong C i 16895 emission is accompanied by weaker, similarly profiled emission lines from neutral and singly-ionized species in the spectra for at least five ABE stars. Figure 7 displays full APOGEE spectra for ABE-A15, ABE-188, and ABE-084, ABE-031, and ABE-004, the best examples of this marked deviation from the typical H-band emission line content for Be stars. The C i 16895 emission is blended with Fe ii 16878 for ABE-A15 and possibly also for ABE-188. Two other C i lines at 16009.27 Å and 16026.08 Å are blended in emission for these stars, leading us to refer to the group as "C i-strong" Be stars.

Figure 7.

Figure 7. Spectra of five Be stars (ABE-A15, ABE-188, ABE-084, ABE-031, ABE-004) with strong C i 16895 emission and many weak, double-peaked metallic emission features. The spectrum of a strong-metal-lined A star (HD 163271) is included to demonstrate that the additional emission lines for these four Be stars correspond to absorption lines for cooler stars. The small lines (blue) above the A star spectrum mark the positions of numerous Fe i lines with log(${{g}_{i}}{{f}_{ik}}$) $\gt $ −3.

Download figure:

Standard image High-resolution image

Most of the metallic emission features for the C i-strong Be stars correspond to strong absorption lines for late-A and cooler stars. An APOGEE spectrum for HD 163271, which is either a single metallic-line A star (A2/A3m) or the superposition (A2/A3+F0) of an A star with an F star (Houk & Smith-Moore 1988; Renson et al. 1991), is provided in Figure 7 to demonstrate the typical line content for A−F stars. Small blue line segments indicate the numerous strong Fe i lines covered, with Fe i 15299 being the strongest and appearing in emission for the Be stars. It is likely that emission from other Fe i lines is involved in much of the blending in the C i-strong Be star spectra.

Detections of resolved emission peaks for the S i lines labeled in Figure 7 are unavailable, but the lines may contribute to the weak bumps and blending around Br17. The transition probability measures for these S i lines suggests S i 15426 should be the strongest of them and the A star spectrum appears to confirm this. Since the Br17 profiles for the C i-strong Be stars do not appear distorted by significant underlying emission from other lines however, it is unclear whether the S i lines are actually observed as emission features.

The strong emission lines redward of Br15 are due partly to several Mg i lines, with the strongest contributions being Mg i 15753.291 and Mg i 15770.149. The Mg i lines are also seen weakly in emission and unblended for ABE-149; all of the emission lines are single-peaked for that source, including H i, Fe ii 16878, C i 16895, and the Mg i lines (see the Appendix). Above the ABE-A15 spectrum in Figure 7 is a small panel that zooms in on the Mg i blend for ABE-A15, demonstrating that emission from $\lambda 15760$ is also a major contributing factor in the blend. Black arrows in the small panel point out the sharp $\lambda 15760$ peaks that mimic the sharp $\lambda 16781$ peaks. ABE-004 similarly has the $\lambda 15760$ and $\lambda 16781$ lines clearly in emission.

As for the line around 15964 Å, PLL suggests two possible identities: Cl i 15964.11 and Si i 15964.4218. Since other covered Cl i lines are expected to be stronger than Cl i 15964.11 are covered but do not appear in emission (e.g., Cl i 15524.70), Si i seems the more likely to cause the 15964 Å emission. The line blended with Br14 (most noticeable for ABE-084 and ABE-031) is suspected to be Si i 15892.7713, the next strongest Si i line covered after Si i 15964.4218.

The weak double-peaked line around ∼16565 Å is possibly Ca ii 16565.59, but the ambiguous detection of Ca ii 16654.43 calls the Ca ii identification into question since the latter line should be stronger. On the other hand, the position of Ca ii 16654.43 corresponds to a strong telluric band which is poorlycorrected and may cause the ambiguity. Emission from the Ca ii triplet (8498, 8542, 8662 Å) is observed for some Be stars (Hiltner 1947; Polidan & Peters 1976), so H-band Ca ii emission would not be terribly unexpected. A C i line at 16564.13 Å probably does not contribute since similar C i lines, covered and expected to be stronger than C i 16564, fail to appear.

The cause of the strong C i 16895 in addition to other weaker emission lines for the C i-strong Be stars remains unknown. Based on the available examples however, such as ABE-031 where the weak emission features persist in 12 spectra covering 1.2 years, the phenomenon appears to be permanent rather than a particular stage of short- or medium-term intrinsic variability.

4.5.  $\lambda 15760$ and $\lambda 16781$

The $\lambda 15760$ and $\lambda 16781$ emission lines discussed by Steele & Clark (2001) are present for between 15–21%. As is demonstrated in Figure 8, these lines always appear together with matching intensity and V/R orientation. In the available examples where peak separations were measurable for $\lambda 15760$ and $\lambda 16781$, those values are nearly identical as well (see Section 5.2). Fe ii 16878 is usually detected in unison with $\lambda 15760$ and $\lambda 16781$, but this is not a strict rule. Non-detection of Fe ii 16878 is accompanied by detections $\lambda 15760$ and $\lambda 16781$ for ABE-180, ABE-A05, and ABE-005, the three lower-most stars represented in Figure 8.

Figure 8.

Figure 8. Identifications of $\lambda 15760$ and $\lambda 16781$ are uncertain; however, these lines are never detected separately and in most cases Fe ii 16878 emission is also present. The three lines always share a common V/R orientation, but the Fe ii intensity varies with respect to $\lambda 15760$ and $\lambda 16781$. Small absorptions around the $\lambda 15760$ line are telluric correction artifacts.

Download figure:

Standard image High-resolution image

The $\lambda 15760$ line has been identified as Fe ii in several past papers (Steele & Clark 2001; Smith 2001; Kraus et al. 2012). In a study of η Carinae, Hamann et al. (1994) was apparently the first to note proximity of $\lambda 15760$ to an Fe ii transition. The authors of that paper appended a question mark to the Fe ii identification listed in an emission line table, but it seems that over the years the question mark was forgotten. PLL lists an Fe ii line at 15761.78 Å, but no indication of expected transition strength is available. NIST provides wavelength for different Fe ii transitions, at 15759.720 and 15760.563 Å, with the lower energy levels again more than doubling those of Fe ii 16878 (5.5 eV versus 13.4 eV) and again lacking transition strength indication. A firm identification for this emission line remains elusive.

Whereas Steele & Clark (2001) restricted the possible identifications for $\lambda 16781$ to [Fe ii] 16773 and Fe ii 16792, the much higher-resolution APOGEE spectra rule out both of those lines as possibilities (see ABE-A23 spectrum in Figure 5). PLL includes several He i lines around 16780 Å, but considering that an absorption line is never seen at this wavelength for normal OB stars, $\lambda 16781$ is probably not He i. Also listed in PLL is an O i multiplet at 16781.7 Å, lacking transition probability data and being quickly ruled out by non-detection of other O i lines covered and expected to be stronger. Through similar argument, other lines listed in PLL and NIST around $\lambda 16781$ are readily ruled out as possibilities.

Whatever the identities of $\lambda 15760$ and $\lambda 16781$, the features behave similarly to Fe ii in being present as emission lines or not present at all: no corresponding absorption for features are seen for APOGEE-observed stars of any type. Reliable spectral types have been reported for 24 ABE stars with $\lambda 15760$ and $\lambda 16781$ detections and 20/24 are B3 or hotter, so it is possible that $\lambda 15760$ and $\lambda 16781$ are relatively high-ionization lines. One possible example of $\lambda 16781$ being detected despite absence of $\lambda 15760$ is ABE-A36, a peculiar star discussed in Section 6.1. However, the bump in the V wing of Br11 for ABE-A36 is not sufficiently convincing to cause us to doubt that $\lambda 15760$ and $\lambda 16781$ should always be expected to appear simultaneously.

4.6. N i

The expected strongest and second strongest N i lines covered are seen in emission for ABE-A35. Figure 9 shows a portion of a spectrum for ABE-A35 encompassing the N i 15586.545 and N i 15687.160 lines as well as Br16, Br15 and $\lambda 15760$. Neither of the N i lines are detected for any other objects beyond ABE-A35, but they are present in all five APOGEE spectra for ABE-A35. Although [Fe ii] 15586.550 is coincident in position with the stronger N i line, it is far from the strongest [Fe ii] feature covered. The lack of detection in the ABE-A35 spectra of the stronger [Fe ii] lines rules those lines out as possibilities. Forbidden line emission in the optical was noted as early as 1976 for ABE-A35 (Allen & Swings 1976), but in the H-band the only clues suggesting the B[e] nature of this object are the abnormally strong H i, Fe ii, and Fe ii-like emission lines.

Figure 9.

Figure 9. Emission from N i is seen only for ABE-A35, a supergiant B[e] star (Esteban & Fernandez 1998; Oksala et al. 2013). Both lines, N i 15587 and N i 15687, are partially blended with H i emission wings.

Download figure:

Standard image High-resolution image

4.7. Mg ii and Si ii

The lowest-energy Mg ii and Si ii lines covered in APOGEE data appear clearly in emission for ABE-A26 and are not confidently detected for any of the other stars. ABE-004, ABE-A05, and ABE-A29 possibly show exceedingly weak contributions from these lines, but blending renders the situation ambiguous in all cases aside from ABE-A26.

Figure 10 displays a spectrum for ABE-A26 over differing wavelength ranges: the upper panel shows the full spectrum, while the middle and lower panels focus on the weak emission lines around Br11. Identification of the line blueward of $\lambda 16781$ as Mg ii 16764.80 requires that the stronger of three lines comprising this Mg ii multiplet also be detected, and indeed the lower panel of Figure 10 shows that Mg ii 16804.52 is visible in the V wing of Br11 at apparently the correct intensity relative to Mg ii 16764.80. Based on the intensities of these lines, the third Mg ii line of the multiplet (16803.67 Å) is not expected to appear and would overlap with Mg ii 16804 anyway. The weak emission line redward of Fe ii 16878 in the middle panel of Figure 10 is identified as Si ii 16911.430, the strongest Si ii line covered and a line with very similar energy levels to the Mg ii lines (see Table 2).

Figure 10.

Figure 10. Mg ii and Si ii emission lines in a spectrum of ABE-A26, the only star for which these lines are detected. While the full spectrum is presented in the upper panel, the lower two panels highlight the Br11 region and the weak metallic lines therein. The single-peaked Br11 line profile is displayed in the inset panel of the larger middle panel for comparison to the double-peaked profiles of the metallic lines. As expected from the detection of Mg ii 16765, the stronger line of this multiplet, Mg ii 16804, appears blended with Br11 in the lowermost panel.

Download figure:

Standard image High-resolution image

In addition to detection of the relatively high-ionization Mg ii and Si ii lines, the combination of single-peaked Brackett series lines and double-peaked metallic lines is unique to ABE-A26 among this sample. The double-peaked lines indicate that at least some of the circumstellar gas is organized in a disk. It is possible that ABE-A26 was observed by APOGEE during of after an outburst such that substantial Brackett series emitting gas is in the polar regions, leading to the single-peaked Brackett lines.

5. PEAK SEPARATIONS

5.1. Stars with Abnormally Large ${\Delta}{{v}_{{\rm p}}}$

Optical spectroscopy revealed that two stars, ABE-050 (HD 345439) and ABE-075 (HD 23478), with extremely large Brackett emission widths and double-peak separations are not classical Be stars (Eikenberry et al. 2014). Rather, these stars are analogues to the prototype magnetic emission B star σ Orionis E first described as "helium-rich" by Greenstein & Wallerstein (1958) and subsequently providing the first application (Townsend et al. 2005) of the Rigidly Rotating Magnetosphere model of Townsend & Owocki (2005). Large ${\Delta}{{v}_{{\rm p}}}$ for the Brackett series emission was a clue suggestive of a non-classical nature for these stars, but confirmation lay in the fact that both stars exhibit H i emission well beyond the projected $v\;{\rm sin} \;i$ values (in these cases, a factor of two ore more beyond the projected $v\;{\rm sin} \;i$). For classical Be stars with Keplerian disks, the velocity separations of emission peaks do not exceed $2\;v\;{\rm sin} \;i$ (Dachs et al. 1992).

Figure 11 compares the Br11 profiles of ABE-050 and ABE-075 to the stars with the next largest peak separations. Arrows indicate the measured peak separations and for ABE-050 and ABE-075, the inner sets of arrows indicate the $v\;{\rm sin} \;i$ values from Eikenberry et al. (2014). As there are only a handful of magnetic B emission stars known to exist, it seems more likely that the ABE-155, ABE-168, ABE-124, and ABE-099 are weak-disked, edge-on classical Be stars rather than additional σ Orionis E types. Either way, optical follow-up spectroscopy is required for proper diagnosis.

Figure 11.

Figure 11. Br11 line profiles for the ABE stars with the largest peak separations are shown. The ${\Delta}{{v}_{{\rm p}}}$ is listed and marked with arrows (red) for each source, while the $v\;{\rm sin} \;i$ measurements for ABE-050 and ABE-075 are given and indicated with arrows (blue) interior to the ${\Delta}{{v}_{{\rm p}}}$ arrows. Whereas ABE-050 and ABE-075 are confirmed σ Orionis E type stars, the other four stars remain to be investigated further.

Download figure:

Standard image High-resolution image

5.2. Line-by-line ${\Delta}{{v}_{{\rm p}}}$ Comparison

Figure 12 plots the peak separations for Br11 versus the peak separations for Br12–Br18, Br20, $\lambda 16781$, $\lambda 15760$, Fe ii 16878, Fe ii 16792, C i 16895, Si i 15964, and Mg i 15770. Each point represents the average peak separation for a line, from all spectra for a given star in which the Br11 ${\Delta}{{v}_{{\rm p}}}$ was measured in addition to the ${\Delta}{{v}_{{\rm p}}}$ of the line represented on the y-axis. In the upper nine panels, plus signs (red) correspond to stars for which the Br20 peak separation was measured and therefore to stars with strong or particularly sharp-peaked emission. The gaps between ∼400–500 km s−1 in the Br11 versus Br18 and Br11 versus Br17 panels are due to strong airglow lines impacting the emission peaks at large line width. High velocity gaps in the Br11 versus Br12 and Br11 versus Br14 panels are caused by either the Br12 V peak or the Br14 R peak falling too close to gaps between detectors. For Br13 and Br15, telluric absorption contamination is more likely for large line width. Grey lines indicate 1-to-1 relationships between the lines plotted in each panel.

Figure 12.

Figure 12. Peak separation of Br11 is compared to the peak separations for the other Brackett series lines as well as the most routinely detected metallic emission lines. Symbol meanings are described in Section 5.2.

Download figure:

Standard image High-resolution image

The effect of increasing ${\Delta}{{v}_{{\rm p}}}$ toward weaker H i lines is well-known for the Balmer (Hanuschik et al. 1988) and Paschen (Andrillat et al. 1990) lines, and the Brackett series lines are not an exception. Some stars (primarily the narrow-lined variety with Br11 ${\Delta}{{v}_{{\rm p}}}\lt 200$ km s−1) show very little or no variation among Brackett series ${\Delta}{{v}_{{\rm p}}}$ but no convincing examples are found of decreasing ${\Delta}{{v}_{{\rm p}}}$ from Br11 toward Br20. We interpret the increasing peak separations toward weaker lines as kinematic in nature, such that the weaker Brackett lines (Br12–20) are simply formed closer to the rapidly-rotating central stars than e.g., Br11. The Br11–20 lines never take the form of winebottle-type profiles frequently observed in the optically thick Hα line, where the effect of non-coherent scattering can produce inflections in the emission profile and effectively reduce the observed peak separation (Hummel & Dachs 1992). Section 4 of Hanuschik et al. (1996) provides a summary of the line broadening factors that contribute to Be star emission line profile shapes.

Based on the lower three panels of Figure 12, when the ${\Delta}{{v}_{{\rm p}}}$ of $\lambda 15760$ and $\lambda 16781$ are measured simultaneously, very similar values are found. The ${\Delta}{{v}_{{\rm p}}}$ for these lines are usually slightly smaller than the Br11 ${\Delta}{{v}_{{\rm p}}}$ but can be slightly larger as well. Excluding ABE-A26, for which Br11 is single-peaked, Fe ii 16878 is a strictly small-${\Delta}{{v}_{{\rm p}}}$ line relative to Br11 with all ${\Delta}{{v}_{{\rm p}}}$ measurements less than 140 km s−1. The C i 16895 Si i 15964, and Mg i 15770 lines share nearly identical ${\Delta}{{v}_{{\rm p}}}$ in the available examples (ABE-084, ABE-188, ABE-A15; see Figure 7), and all of the ${\Delta}{{v}_{{\rm p}}}$ measurements for C i 16895 exceed the Br11 ${\Delta}{{v}_{{\rm p}}}$.

5.3. Line-emitting Disk Radii

For Keplerian rotation in a gaseous disk, the orbital velocity decreases according to ${{r}^{-1/2}}$, where r is the radial distance from star to disk. Given knowledge of the stellar rotational velocity, $v\;{\rm sin} \;i$, the peak separation of an emission line can be used to calculate the approximate outer radius in the disk, rd, at which that line is preferentially formed (Smak 1969; Huang 1972; Smak 1981; Horne & Marsh 1986). Many authors (e.g., Hanuschik 1987; Hanuschik et al. 1988; Andrillat et al. 1990; Dachs et al. 1992; Slettebak et al. 1992) have used this relation (Huangʼs law) to study the geometry of Be disks by estimating the individual line-emitting radii for H i and metallic lines in the optical region. In units of stellar radii, ${{R}_{*}}$, rd is calculated via Huangʼs Law (Huang 1972) as

Equation (1)

where the Equation is squared due to the assumption of a circular orbit. The resulting outer disk radii measurements for the subset of stars with available $v\;{\rm sin} \;i$ from the literature and Br11 or metallic line ${\Delta}{{v}_{{\rm p}}}$ measurements are listed in Table 3. In estimating rd, the average ${\Delta}{{v}_{{\rm p}}}$ measured from all spectra for each star (the number of spectra used for each star is indicated in the "# Obs" column) have been used. The relation in Equation (1) may not necessarily hold for cases of emission lines with asymmetric peak intensities or for shell profiles so these instances have been noted in Table 3. In particular, large rd estimates (${{r}_{{\rm d}}}\gt 5{{r}_{*}}$) correspond to asymmetric and shell profiles.

Table 3. Line-emitting Disk Radius Estimates

ABELit.Ref.Atom $\#$ MeanMean
ID $v\;{\rm sin} \;i$  orspectra ${\Delta}{{v}_{{\rm p}}}$ rd
 (km s−1) Ion (km s−1)(${{R}_{*}}$)
001 a 2669Br1142086.56
    $\lambda 15760$ 42464.67
    $\lambda 16781$ 42614.16
0032255Br11103052.18
0062317Br1132942.47
0142303Br1173261.99
016 a 2504Br1112384.43
   Fe ii 16878111120.41
026 b 2303Br11123242.02
0282424Br11122972.66
036 b 3074Br1134561.81
046 b 2352Br1112942.56
0491202Br11231971.48
   Fe ii 1687876712.73
0671205Br1181333.24
   Fe ii 1687884134.32
0851828Br1143211.29
09816611Br1113011.21
133 b 2864Br1132854.02
1382437Br1123651.77
1403284Br1143822.95
1652744Br1133971.91
1671607Br1133001.14
1701305Br1161532.90
   Fe ii 1687866416.76
   C i 1689521692.36
1892566Br1153052.81
1912424Br1112743.11
195 b 1487Br1132661.24
A0834311Br1134092.82
A112207Br1132154.19
A122157Br1113581.44
A13 a b 3064Br1162575.68
    $\lambda 15760$ 42486.07
    $\lambda 16781$ 42506.00
A15 a 3507Br1132378.76
   Fe ii 1687832329.07
   C i 1689533454.12
    $\lambda 15760$ 32339.05
    $\lambda 16781$ 32339.04
A17 b 3001Br1132416.19
   Fe ii 16878310234.32
A202104Br1133581.37
A29 a 2224Br1131975.09
   Fe ii 16878312113.42
    $\lambda 15760$ 31449.55
    $\lambda 16781$ 31449.52
A3226010Br1133801.87

References. (1) Uesugi & Fukuda (1970); (2) Uesugi & Fukuda (1982); (3) Halbedel (1996); (4) Yudin (2001); (5) Abt et al. (2002); (6) Frémat et al. (2005); (7) Frémat et al. (2006); (8) Huang & Gies (2006); (9) Bhavya et al. (2007); (10) Reig et al. (2010); (11) Huang et al. (2010).

a Asymmetric emission peak intensities. b Shell stars.

Download table as:  ASCIITypeset image

Table A1. List of ABE stars

Identifiers Magnitudes Spectral type Line detection or ${\Delta}{{v}_{{\rm p}}}$ [km s−1]
 
ABE2MASSHDOther  V H   Ref. Br11 16811 $\lambda 15760$ $\lambda 16781$ Fe ii 16792Fe ii 16878C i 16895
001 $20212485+3722482$ ...VES 203 12.099.108 B0.5Ve74 208246261-w-
002 $20151525+3654562$ 228576AS 394 11.419.888 Ae29 152w86-85-
003 $20162816+3703229$ 192987HR 7757 6.466.548 B6IIIe22 305---w?w
004 $20234596+3830033$ 229221MWC 344 9.226.734 B0.2IIIe66 115131105-w?191
005 $20184170+3759106$ ...Hen 3-1876 11.379.699 OB6 271ww---

References. (1) Merrill et al. (1942); (2) Cannon & Mayall (1949); (3) Merrill & Burwell (1949); (4) Popper (1950); (5) Miller & Merrill (1951); (6) Nassau & Harris (1952); (7) Morgan et al. (1953); (8) Morgan et al. (1955); (9) Heckmann et al. (1956); (10) Hiltner (1956); (11) Duflot et al. (1958); (12) Alknis (1958); (13) McCuskey (1959); (14) Hardorp et al. (1959); (15) Bouigue et al. (1961); (16) Fehrenbach et al. (1962); (17) Roslund (1963); (18) Feast & Thackeray (1963); (19) McCuskey (1967); (20) Schmidt-Kaler (1967); (21) Racine (1968); (22) Lesh (1968); (23) Guetter (1968); (24) Wackerling (1970); (25) Walborn (1971); (26) Cowley (1972); (27) Lesh & Aizenman (1973); (28) Turner (1976); (29) Henize (1976); (30) Voroshilov et al. (1976); (31) Davis (1977); (32) Christy (1977); (33) Stephenson & Sanduleak (1977); (34) Hill & Lynas-Gray (1977);(35) Stephenson & Sanduleak (1977); (36) Roman (1978); (37) Bartaya (1979); (38) Clausen & Jensen (1979); (39) Ochsenbein (1980); (40) Jaschek & Jaschek (1993); (41) Houk (1982); (42) Voroshilov et al. (1985); (43) Bopp (1988); (44) Bidelman (1988); (45) Houk & Smith-Moore (1988); (46) Radoslavova (1989); (47) Sato & Kuji (1990); (48) Turner et al. (1992); (49) Grillo et al. (1992); (50) Turner (1993); (51) Garrison & Gray (1994); (52) Nesterov et al. (1995); (53) Abt & Morrell (1995); (54) Kohoutek & Wehmeyer (1997); (55) Lamers et al. (1998); (56) Esteban & Fernandez (1998); (57) Grenier et al. (1999); (58) Houk & Swift (1999); (59) Yudin (2001); (60) Chauville et al. (2001); (61) Kharchenko (2001); (62) Fabricius et al. (2002); (63) Abt et al. (2002); (64) Miroshnichenko et al. (2003); (65) Hernández et al. (2004); (66) Negueruela (2004); (67) Negueruela et al. (2004);(68) Frémat et al. (2006); (69) Levenhagen & Leister (2006); (70) Uzpen et al. (2007); (71) Reig et al. (2010); (72) Walborn et al. (2010); (73) Sota et al. (2011); (74) Mathew & Subramaniam (2011); (75) Sebastian et al. (2012); (76) Chargeishvili et al. (2013); (77) Eikenberry et al. (2014).

Only a portion of this table is shown here to demonstrate its form and content. Machine-readable and Virtual Observatory (VO) versions of the full table are available.

Download table as:  Machine-readable (MRT)Virtual Observatory (VOT)Typeset image

Taking the average of the Br11 rd estimates for the 19 non-shell stars with roughly symmetric emission peaks, we find an average Br11 formation outer radius and associated standard deviation of

Equation (2)

Hanuschik et al. (1988) found an average Hα-emitting radius of ∼20 ${{R}_{*}}$, while Slettebak et al. (1992) found an average of ∼19 ${{R}_{*}}$. It is important to note, however, that results from interferometry confirm that application of Huangʼs law to the double peaks of winebottle-type profiles (appearing for optically thick lines like Hα) leads to artificially large disk radii estimates (Hummel & Dachs 1992). Interferometric studies typically produce radii estimates of less then 10 ${{R}_{*}}$over the optical and JHK bands (see disk radius measurements and papers referenced in Table 2 of (see disk radius measurements and papers referenced in Table 2 of Rivinius et al. 2013b), more similar to what is found here from the Brackett lines.

For Hγ and Fe ii 6516 Å, Slettebak et al. (1992) found emitting radii of $\sim 7.4$ ${{R}_{*}}$ and $\sim 3.9$ ${{R}_{*}}$ respectively. In a study of optical Fe ii emission lines for Be stars, Arias et al. (2006) found that, on average, the optical Fe ii lines are formed at an outer disk radius of 2.0 stellar radii. Therefore, the Br11-emitting outer radius is roughly coincident with the optical Fe ii-emitting outer radius and well inside the Hα-emitting radius.

Given the sparsity and potential wide-range of quality of $v\;{\rm sin} \;i$ information for our sample, disk radii are only estimated for the Br11 and metallic lines. However, it follows from Figure 12 that the Br12–Br20 formation outer radii are interior to that of Br11. Andrillat et al. (1990) and Slettebak et al. (1992) found a correlation between formation location of individual optical lines and the upper energy levels (Ek ) of the lines. Weaker lines with higher Ek were generally found to have larger ${\Delta}{{v}_{{\rm p}}}$ and hence smaller rd. This is also the case for the Brackett series lines, where Ek increases slightly from Br11 to Br20.

A trend toward large rd is evident for the Fe ii 16878 line with respect to Br11. The average of the five available rd estimates for Fe ii is ∼19 ${{R}_{*}}$, almost ten times the disk radius where Br11 is preferentially formed. A consequence of the widely varying formation radii between Br11 and Fe ii 16878 is discussed in the following section.

6. SINGLE-EPOCH VARIATION IN V/R AND RADIAL VELOCITY

As outlined in Okazaki (1991), long-term V/R variability for Be stars often entails shifts in RV of entire emission line profiles toward whichever peak is stronger at the time and differences in V/R orientation between lines with different formation loci, such that V/R is necessarily constant from atomic species to atomic species or from line to line. These effects are believed to be caused by perturbations with the disks that give rise to one-armed global density waves that slowly precess through the disk with periods averaging 7 years (Rivinius et al. 2013b).

Recent papers discussing the well known Be-shell star ζ Tau provided an example of V/R phase lags between Balmer and Brackett series lines and also between individual Brackett series lines. Wisniewski et al. (2007) hypothesized that the optical/NIR phase lag in V/R could be understood in terms of differing preferential formation radii and of the global density perturbation within the disk taking the form of a spiral arm. Štefl et al. (2009) and Carciofi et al. (2009) subsequently showed this to be the case.

6.1. H i versus H i V/R Phase Lags

Evidence of V/R phase lags within the Brackett series lines is present in the spectra for the ABE stars represented in Figure 13. Each Brackett line is displayed individually on a velocity scale in Figure 13 and, with the exception of ABE-A36 (discussed below), the V/R of Br11 for each spectrum is printed in left-most Br11 panels while the differences between the V/R of Br11 and the V/R of Br12–Br20 are printed in the Br12–Br20 panels. For ABE-A29, the V/R orientations progress from V $\lt $ R at Br11 to V $\simeq $ R at Br17. The Br19 profile is contaminated by DIB 15271 absorption, but the Br18 and Br20 profiles have V $\simeq $ R similar to Br17. For ABE-A31 and ABE-181 the opposite progression takes place as the R peak increases in dominance from Br11 to Br20. Although the Br11 profiles for ABE-A36, ABE-A31, and ABE-181 are contaminated by underlying metallic emission ($\lambda 16781$ and/or Fe ii 16878), the V/R phase lag is nonetheless plainly visible from comparison of the Br12 or Br13 profiles to Br20.

Figure 13.

Figure 13. Four examples of variation in V/R phase across the Brackett series lines. The Br11 profile for ABE-A36 has a quasi-triple-peaked morphology which gradually becomes a single-peaked morphology at Br20. For ABE-A29, ABE-A31, and ABE-181, the V/R ratio of Br11 is provided in the leftmost panels and subsequent panels provide the difference between the Br12–Br20 V/R ratios and the Br11 V/R ratio. Gradual changes in V/R orientation are seen among Brackett series lines for these stars: blue text for the differences means increasing V/R ratio from Br11 to Br20 (ABE-A29) while red text means decreasing V/R ratio from Br11 to Br20. DIB 15271 absorption is evident on the Br19 line for all four stars.

Download figure:

Standard image High-resolution image

Of the 238 stars comprising the ABE sample, ABE-A36 is the only available example of quasi-triple-peaked (qtp) Brackett series lines (see upper row of Figure 13), where the "quasi" implies ambiguity as to whether there is actually a true third emission peak or whether instead the central absorption is split into multiple components. The V emission peak of Br11 for ABE-A36 is slightly higher than the R peak due to possible blending with $\lambda 16781$, and the possible third peak appears between the dominant outer emission peaks with lesser intensity than those peaks. Br12 is similar in profile to Br11, but from Br12–Br17 the profiles gradually assume a "flat-topped" morphology with the apparent blue central absorption still weakly visible in contrast to the apparent red central absorption having all but disappeared. At Br18, evidence of the third (middle) emission peak emerges again, this time as the dominant peak since the outer V and R peaks have all but disappeared. Br19 is directly impacted by DIB 15271 but otherwise appears similar to Br18. Finally, the Br20 profile is smooth and rounded with only subtle traces of the blue central absorption and middle "emission peak" (the outer emission peaks are no longer visible).

Štefl et al. (2009) pointed out that although qtp Hα profiles occur at certain times during the V/R cycle of ζ Tau, optically thin lines likes O i 8446 and the Brackett series never exhibited any evidence of qtp. It is therefore unusual that qtp profiles are observed in the Brackett lines for ABE-A36. We can report no additional examples.

6.2. H i versus Metallic V/R and RV Phase Lags

The five examples shown in Figure 14 represent the first known examples of disagreement between the V/R orientations of Brackett series versus metallic lines (Fe ii 16878, $\lambda 15760$, and $\lambda 16781$). ABE-097, ABE-181, and ABE-002 have V $\lt $ R for H i and V $\gt $ R for metallic lines, while ABE-016 and ABE-013 have V $\gt $ R for H i and V $\lt $ R for metallic lines. Due to the contaminated Br11 profiles for ABE-013, ABE-002, and ABE-181, where the V peak height has been increased by underlying Fe ii 16792 emission, the left-hand panels of Figure 14 are extended to encompass not only $\lambda 15760$, but also Br15–Br17 to show the typical H i V/R orientation for each star.

Figure 14.

Figure 14. H i vs. metallic V/R orientation mismatches are evident in the APOGEE spectra of ABE-097, ABE-181, ABE-002, ABE-016, and ABE-013. The emission wings for ABE-181 and ABE-002 are also clearly extended in the direction of the weaker emission peak for each line, and the metallic emission profiles for ABE-013 appear to be offset in radial velocity from the Brackett lines.

Download figure:

Standard image High-resolution image

Although the lack of available stellar absorption lines means that precise stellar RV determination is not possible, the spectrum for ABE-013 in Figure 14 has been corrected to rest frame based on the average positions of the deep absorptions in the Br12–Br20 lines, while the other spectra lack the deep H i absorptions and therefore were corrected for Doppler shift based on average emission peak shift for the Brackett lines. The result that emerges for ABE-013 is that the Brackett series absorptions do not coincide in RV not with the central depressions in the metallic emission lines as is normally true (see Figure 8), but instead the Brackett series absorptions coincide in RV with the R emission peaks of the metallic lines. More specifically, the Fe ii 16878, $\lambda 15760$, and $\lambda 16781$ profiles are shifted in RV with respect to H i absorption by ∼50 km s−1 in the direction of stronger H i peaks as expected from Okazaki (1991).

ABE-002 and ABE-181 exhibit clear evidence of V/R-related RV shifts in the emission profiles, though of a slightly different variety from that of ABE-013. Metallic and H i emission wings for both stars are conspicuously enhanced on the side of the line profiles opposite the stronger emission peak for ABE-002 and ABE-181, with the H i wings being enhanced on the blue side and the metallic wings enhanced on the red side. The enhanced blue wings suggest that significantly more emission is being formed in the inner regions of the approaching side of the disk, and the steep declines in intensity, from stronger emission peak to narrower emission base (R side of H i for ABE-002 and ABE-181), imply cavities in the inner regions of the receding sides of the disks and relatively increased emission coming from the outer regions of the disks. We interpret these line profiles to suggest more tightly wound spiral patterns to the density oscillation in the disks of these stars versus ζ Tau.

7. BR11 LINE PROFILES

Br11 line profiles from the highest-quality-available spectrum of each ABE star are displayed in Figure 15. In Figure 15, the Br11 profiles of 165 ABE stars are qualitatively sorted by profile type, going from single-peaked and narrow double-peaked profiles to deep shell profiles. According to the models of Hummel & Dachs (1992) and Hummel & Vrancken (2000), the major line profile shape differences for Be stars are an effect of the inclination angles (i) at which the circumstellar disks are observed. Hanuschik et al. (1996) used high-resolution Hα and optical Fe ii profiles to devise a Be sub-classification scheme based on the notion of i dictating to a large extent line profile morphology. Silaj et al. (2010) later showed that line shape in an optically-thick line like Hα is not dictated solely by i and that very different profile shapes may be observed at fixed i, but no such investigations of the Brackett series lines have been done.

Figure 15.

Download figure:

Standard image High-resolution image
Figure 15.

Download figure:

Standard image High-resolution image
Figure 15.

Download figure:

Standard image High-resolution image
Figure 15.

Figure 15. Br11 line profiles sorted approximately according to inclination angle, from pole-on to edge-on. ABE identifiers, observation MJDs, 2MASS H magnitudes, and literature spectral types (where available) are printed in each panel. The average Br11 profile of the quiescent Be stars (ABE-Q01–ABE-Q23) is displayed as a dotted line (blue), and vertical dotted lines (gray) indicate emission peak midpoints or estimated line centers if emission peaks are not present.

Download figure:

Standard image High-resolution image
Figure 16.

Figure 16. Br11 line profiles for stars with weak or ambiguous emission profile type, as well as for the σ Ori E type stars ABE-075 and ABE-050. The panels are sorted by Br11 peak separation. Meanings are otherwise the same as in Figure 15.

Download figure:

Standard image High-resolution image
Figure 17.

Figure 17. Br11 line profiles for stars with weak or ambiguous emission profile type and lack of discernible emission peaks, followed by the sample of previously known emission stars that produced little or no emission in the APOGEE spectra. The panels are mostly sorted by ABE identifier (followed by ABE-A23). Note that the telluric correction is problematic for the plug-plate on which ABE-058 was observed; Br11 is clearly in emission, but the profile is badly contaminated by telluric absorption features. Meanings are otherwise the same as in Figure 15.

Download figure:

Standard image High-resolution image
Figure A1.

Figure A1. APOGEE spectra for Be stars with strong Brackett series emission. The emission lines for ABE-A35, ABE-A21, and ABE-015 are double-peaked, whereas the other stars have single-peaked emission.

Download figure:

Standard image High-resolution image
Figure A2.

Figure A2. APOGEE spectra for a selection of 20 Be stars with asymmetric Brackett series emission. Note the striking similarity between the spectra of ABE-001, ABE-A28, and ABE-A30 (the latter is a V/R reflection of the former two).

Download figure:

Standard image High-resolution image
Figure A3.

Figure A3. APOGEE spectra for a selection of 18 Be-shell stars. Broad photospheric absorption wings are clearly visible in the Brackett lines for the upper nine stars, while the lower nine stars exhibit mostly smooth continua and shell features with adjacent emission.

Download figure:

Standard image High-resolution image

In sorting the Br11 profiles of the ABE stars according to expected i, we relied largely on the models of optically thin lines from Hummel & Dachs (1992). The most readily classified Br11 profiles correspond to $i\sim {{0}^{\circ }}$ (pole-on), where single-peaked or narrow double-peaked emission is expected, and $i\sim {{90}^{\circ }}$ (edge-on), where deep shell absorption with a sharp core (with or without adjacent emission) is expected. The situation is far more ambiguous for profiles corresponding to intermediate i, but a general trend of increasing central depression depth and overall line width with increasing i is apparent. Line profiles that could not be satisfactorily sorted by i, due to weakness or ambiguity of the disk features, are shown in Figures 16 and 17. Figure 16 profiles are sorted by Br11 peak separation, and Figure 17 profiles are sorted by ABE identifier.

8. CONCLUSIONS

SDSS-III/APOGEE has serendipitously provided the first high-resolution view of the H-band properties of a large number of Be stars, the majority of which are targeted quasi-randomly by the survey as telluric calibrators. Although significant progress has been made toward understanding Be stars over the past few decades via high-resolution optical, interferometric, and spectropolarimetric studies (Rivinius et al. 2013b), any fully explanatory model of the classical Be phenomenon will need to account for the multi-wavelength properties of these stars. Multi-wavelength studies of statistically-significant samples of Be stars are critical yet have historically been few and far between, though the limited exceptions (Clark & Steele 2000; Steele & Clark 2001) have been highly valuable. Due to simultaneous coverage in the H-band of numerous H i lines that are minimally affected by underlying photospheric absorption in comparison to the Balmer series lines, the H-band is particularly promising in terms of utility toward V/R variability and general Be disk studies. Despite the H-band covering only a limited number of metallic emission lines, we have shown that the Fe ii and Fe ii-like ($\lambda 15760$ and $\lambda 16781$) lines are highly interesting in the context of V/R variability and phase lags between various atomic species.

In the first of a series of papers exploring the H-band properties of Be stars, we have identified the non-hydrogen emission line content of the ABE star spectra, analyzed the kinematic properties of the metallic and H i features, and discussed the more exceptional Be stars within the sample as well as those deviating from the typical emission line content. Further investigation of the identities of emission lines at 15760 Å and 16781 Å is needed, but may require updated atomic line lists. Since little is known about most of the ABE stars themselves, including spectral type and rotation speed, optical follow-up study of these stars is also needed in order to develop a better understanding of H-band properties as they relate to known stellar parameters.

Funding for SDSS-III has been provided by the Alfred P. Sloan Foundation, the Participating Institutions, the National Science Foundation, and the U.S. Department of Energy Office of Science. The SDSS-III web site is http://www.sdss3.org/. SDSS-III is managed by the Astrophysical Research Consortium for the Participating Institutions of the SDSS-III Collaboration including the University of Arizona, the Brazilian Participation Group, Brookhaven National Laboratory, Carnegie Mellon University, University of Florida, the French Participation Group, the German Participation Group, Harvard University, the Instituto de Astrofisica de Canarias, the Michigan State/Notre Dame/JINA Participation Group, Johns Hopkins University, Lawrence Berkeley National Laboratory, Max Planck Institute for Astrophysics, Max Planck Institute for Extraterrestrial Physics, New Mexico State University, New York University, Ohio State University, Pennsylvania State University, University of Portsmouth, Princeton University, the Spanish Participation Group, University of Tokyo, University of Utah, Vanderbilt University, University of Virginia, University of Washington, and Yale University. J. P. W. acknowledges support from NSF-AST 1412110. We thank the anonymous referee and Kevin Covey, both of whom provided feedback that substantially improved the paper. The first author additionally thanks his mother for proofreading drafts of the paper.

Appendix

Full APOGEE spectra for stars with strong Brackett series features are displayed in Figures A1−A3. Shown above each spectrum are ABE identifiers, observation MJDs, and literature spectral types for each star. Newly identified Be stars are indicated with an asterisk before the ABE identifiers. Small line segments mark the positions of the most prominent DIBs present for some of the stars and arrows mark the positions of the most frequently detected emission lines.

Footnotes

Please wait… references are loading.
10.1088/0004-6256/149/1/7