The Astrophysical Journal, 857:30 (9pp), 2018 April 10 https://doi.org/10.3847/1538-4357/aab3d8
© 2018. The American Astronomical Society. All rights reserved.
Excitation Mechanism of O I Lines in Herbig Ae/Be Stars
Blesson Mathew1,2 , P. Manoj2 , Mayank Narang2, D. P. K. Banerjee3, Pratheeksha Nayak4, S. Muneer5, S. Vig4,
S. Pramod Kumar5, K. T. Paul1, and G. Maheswar5
1 Department of Physics, Christ University, Hosur Road, Bangalore 560029, India; blesson.mathew@christuniversity.in
2 Department of Astronomy and Astrophysics, Tata Institute of Fundamental Research, Homi Bhabha Road, Colaba, Mumbai 400005, India
3 Astronomy and Astrophysics Division, Physical Research Laboratory, Navrangapura, Ahmedabad 380 009, India
4 Indian Institute of Space Science and Technology (IIST), Trivandrum, India
5 Indian Institute of Astrophysics, Koramangala, Bangalore 560034, India
Received 2017 October 6; revised 2018 March 2; accepted 2018 March 2; published 2018 April 10
Abstract
We have investigated the role of a few prominent excitation mechanisms viz. collisional excitation, recombination,
continuum fluorescence, and Lyman beta fluorescence on the O Iline spectra in Herbig Ae/Be stars. The aim is to
understand which of them is the central mechanism that explains the observed O Iline strengths. The study is
based on an analysis of the observed optical spectra of 62 Herbig Ae/Be stars and near-infrared spectra of
17 Herbig Ae/Be stars. The strong correlation observed between the line fluxes of O Iλ8446 and O Iλ11287, as
well as a high positive correlation between the line strengths of O Iλ8446 and Hα suggest that Lyman beta
fluorescence is the dominant excitation mechanism for the formation of O Iemission lines in Herbig Ae/Be stars.
Furthermore, from an analysis of the emission line fluxes of O Iλλ7774, 8446, and comparing the line ratios with
those predicted by theoretical models, we assessed the contribution of collisional excitation in the formation of
O Iemission lines.
Key words: circumstellar matter – infrared: stars – stars: pre-main sequence – stars: variables: T Tauri,
Herbig Ae/Be – techniques: spectroscopic
1. Introduction various astrophysical objects. Prominent mechanisms discussed
Herbig Ae/Be (HAeBe) stars are intermediate mass for the formation of O Ilines are collisional excitation,
(2 M  M  8 M) pre-main-sequence stars with accretion recombination, continuum fluorescence, and Lyman beta
disks, the innermost regions of which also act as a reservoir for (Lyβ) fluorescence. For example, Grandi (1975b) showed
the production of major emission lines seen in the optical and that starlight continuum fluorescence is the favored excitation
infrared spectra (Herbig 1960; Hillenbrand et al. 1992; Waters mechanism for the O Iline in the Orion nebula, whereas
& Waelkens 1998). HAeBe stars were first discussed as a in Seyfert 1 galaxies it is excited by Lyβ fluorescence
distinct group of objects by Herbig (1960), who noted that they (Grandi 1980). In novae, Strittmatter et al. (1977) identify
were stars of spectral types A or B with emission lines, located Lyβ fluorescence as the dominant excitation mechanism; a
in an obscured region and often accompanied by a surrounding conclusion that has been supported by studies of several other
nebulosity. The present working definition of HAeBe stars novae (e.g., Ashok et al. 2006; Banerjee & Ashok 2012, and
includes, (a) pre-main-sequence stars of A−F spectral type, references therein). Lyβ fluorescence is identified as the
displaying emission lines in their spectra and (b) show a dominant contributor to the emission strength of the
significant IR excess due to hot or cool circumstellar dust shells O Iλ8446 line in classical Be (hereafter CBe) stars, whether
or a combination of both (The et al. 1994; Waters & it is isolated (Slettebak 1951; Mathew et al. 2012b) or part of an
Waelkens 1998; Vieira et al. 2003). There have been extensive X-ray binary system (Mathew et al. 2012a). Bhatia & Kastner
spectroscopic studies of HAeBe stars in the literature (e.g., (1995) and Kastner & Bhatia (1995) provided a theoretical
Hamann & Persson 1992; Hernández et al. 2004; Manoj framework of O Iexcitation and derived the expected line
et al. 2006); particularly important are the recent studies by the ratios of the prominent O Ilines, when collisional excitation
X-Shooter team (Mendigutía et al. 2011, 2012; Fairlamb and Lyβ fluorescence (referred to as photoexcitation by
et al. 2015, 2017). Most of these studies have been devoted to accidental resonance—PAR process—in Bhatia & Kastner
Hα line analysis, the most prominent emission feature seen in 1995) are the dominant excitation mechanisms. From a
the spectra of HAeBe stars (Finkenzeller & Mundt 1984; comparative analysis of the theoretical estimates with the
Hamann & Persson 1992). In the present study, we focus on the observed emission strengths of O Iλλ 7774, 8446, 11287, and
O Iemission lines in the optical and near-infrared (1–2.5 μm) 13165, Mathew et al. (2012b) demonstrated that Lyβ
spectra in HAeBe stars. fluorescence is the dominant excitation mechanism for the
O Iλ8446 is the most prominent O Iemission line seen in production of O Iλλ 8446, 11287 lines in CBe stars. CBe stars
the optical spectrum of HAeBe stars. This emission line results share similar spectral characteristics with HAeBe stars, such as
from the 3s3S0–3p3P transition and is seen as a triplet at high emission lines of Hα, O I, Fe II,and Ca IItriplet. It is worth
resolution, with wavelength values of 8446.25, 8446.36, and exploring whether both CBe and HAeBe stars share similar
8446.76Å. It is present in the spectra of a wide variety of excitation mechanisms for the formation of O Ilines. There
astrophysical sources, such as planetary nebulae, novae, and could be considerable difference between the O Iline forming
Seyfert galaxies. A number of studies have addressed the regions in both the stellar systems. CBe stars are found to have
question of excitation mechanisms of O Iemission lines in a circumstellar gaseous decretion disk wherein the O Iλ8446
1
The Astrophysical Journal, 857:30 (9pp), 2018 April 10 Mathew et al.
line is formed at a mean radial distance of ∼8 Rå, considering increase the sample size of the present study, we have included
Keplerian motion (Mathew et al. 2012b). However, the location the optical spectra of HAeBe stars from Manoj et al. (2006),
of the origin of the O Iλ8446 line in HAeBe is far from clear. which were observed with a similar observational setup. Thus
Most of the accretion related emission lines in HAeBe stars we have optical spectra for a total of 62 HAeBe stars and near-
(e.g., Hα, Paβ, Brγ) are thought to be formed in the IR spectra for 19 HAeBe stars. As a representative sample, we
magnetospheric accretion columns (Muzerolle et al. 2004). show Hα, O I λ λ7774, 8446 line profiles of V594 Cas, LkHα
This work is an attempt to bring more clarity to our 233, and MWC 297 in Figure 1.
understanding of the formation mechanisms of O Iemission The B, V, RC magnitudes, total extinction (AV), spectral type,
lines in HAeBe stars. and effective temperature (Teff) of 62 HAeBe stars are listed in
The paper is organized as follows. In Section 2, we present Table 2. The spectral type is converted to Teff using the
the optical and near-infrared (near-IR) spectroscopic observa- tabulated information in Pecaut & Mamajek (2013). We
tions carried out over a period of 3 years and describe the data compiled the photometric data from various sources in the
reduction techniques employed. We describe the methods and literature, whose references are given in Table 2. For some of
the python routines used for the spectral analysis and to the sources, R magnitudes are in the Johnson system, which are
estimate line flux in Section 3. The dominant excitation converted to the Cousins system following Bessell (1983). The
mechanism for the formation of O Ilines in HAeBe stars is color excess, E(B–V ), is calculated from the observed
evaluated in Section 4. The main results of the paper are (B–V ) colors and the intrinsic colors corresponding to each
summarized in Section 5. spectral type, from the table listed in Pecaut & Mamajek
(2013). Furthermore, we calculated AV from E(B–V )
2. Observations and Data Reduction considering a total-to-selective extinction value, RV=5. It
has been demonstrated from various studies (see Hernández
The optical spectroscopic observations were carried out et al. 2004) that RV=5 is the preferred value in the analysis of
using the Himalayan Faint Object Spectrograph Camera
6 HAeBe stars, suggesting grain growth in the disk of HAeBemounted on the 2 m Himalayan Chandra Telescope (HCT). stars (e.g., Gorti & Bhatt 1993; Manoj et al. 2006).
The spectroscopic observations were obtained with Grism 8 in
combination with 167l slit (1 92 wide and 11′ long), providing
an effective resolving power of ∼1050. The spectral coverage 3. Analysis
is from 5500 to 9000Å, which included the spectral lines 3.1. Classification Based on O Iline Profiles
relevant to this study, viz., Hα, O Iλ7774, and O Iλ8446.
After each on-source exposure, FeNe lamp spectra were From the observed spectra, we find that O Ilines, both in
obtained for wavelength calibration. We have followed the optical and infrared, are seen in emission as well as in
regular procedure of reducing the spectra after bias subtraction absorption. We adopted the classification scheme proposed by
and at- eld correction using the standard tasks in Image Felenbok et al. (1988), wherein Group I stars have bothfl fi
Reduction and Analysis Facility (IRAF).7 O Iλ8446 and O Iλ7774 in emission; Group II sources are
Near-IR spectra were obtained using the TIFR Near Infrared those with both lines in absorption; Group III is the case when
Spectrometer and Imager, mounted on the HCT. The spectra O Iλ8446 is in emission and O Iλ7774 in absorption. We
were obtained in Y and J passbands, at a resolving power of found 23 stars belonging to Group I, 16 in Group II, and 23 in
1200. The observations were performed in the dithered mode. Group III classes. A similar classification scheme is applied to
Argon lamp spectra taken after each on-source exposure are the infrared spectra. Although O Iemission is evident among
used for wavelength calibration. An appropriate telluric Group I stars, after subtracting the photospheric component, a
standard (of early A spectral type) is observed at the nearby net emission is seen in some of the Group II and Group III
airmass to the target object. The spectra of the target and the stars. For the current sample, we found net emission in
standard are reduced in a standard manner with the tasks in O I λλ7774, 8446 for 31 and 54 stars, respectively, whereas 17
IRAF. For telluric correction, we removed the hydrogen lines sources show net emission in O I λλ11287 and 13165.
from the telluric standard spectrum, which is then used to
divide the object spectrum. The resultant object spectrum is 3.2. Flux Measurement of O Iand Ha Emission Lines
multiplied with the blackbody corresponding to the spectral In this section, we describe the method we used to measure
type of the telluric standard in order to preserve the continuum the line fluxes of Hα, O Iλλ7774, 8446, 11287, and 13165
of the target spectrum. The log of optical and infrared lines from the wavelength calibrated optical and near-IR
spectroscopic observations is given in Table 1. spectra. The procedure can be summarized as (i) estimating the
The sample of HAeBe stars observed were drawn from a equivalent width (EW) of the lines of interest from a Gaussian
larger list of 142 HAeBe stars that we compiled from the profile fit using LMFIT routine in Python, (ii) removing the
literature (The et al. 1994; Manoj et al. 2006; Fairlamb contribution of Paschen P18 line from O Iλ8446, (iii)
et al. 2015). Given the location of the observatory and the accounting for photospheric absorption using synthetic spectra,
limiting magnitude of the spectrograph-telescope combination, (iv) estimation of continuum flux at the wavelength region
we were able to obtain the optical spectra of 56 HAeBe stars corresponding to Hα and O Ilines, and (v) the calculation of
and near-IR spectra of 19 HAeBe stars. The observations were extinction corrected line flux from the EW and the con-
carried out over a period of 3 years, from 2014 to 2017. To tinuum flux.
6 http://www.iiap.res.in/iao/hfosc.html
7 IRAF is distributed by the National Optical Astronomy Observatories, 3.2.1. Estimation of Line EW
which are operated by the Association of Universities for Research in
Astronomy, Inc., under cooperative agreement with the National Science We estimated the EW of O Iλλ7774, 8446, 11287, 13165,
Foundation. and Hα lines using the LMFIT module on the continuum
2
The Astrophysical Journal, 857:30 (9pp), 2018 April 10 Mathew et al.
Table 1
Log of Spectroscopic Observations
Object Date of Optical Date of Y band J band
Optical Observations Exp. Time (s) IR Observations Exp. Time (s) Exp. Time (s)
(1) (2) (3) (4) (5) (6)
51 Oph 2014 May 19 60 L L L
AB Aur 2017 Jan 21 40 2013 Dec 11 600 600
AS 442 2016 Aug 10 300 L L L
AS 443 2016 Aug 10 420 L L L
AS 505 2016 Aug 10 300 2016 Nov 21 600 600
BD+30 549 2016 Nov 22 400 L L L
BD+40 4124 2014 Jun 19 300 L L L
BD+65 1637 2016 Aug 10 300 L L L
CQ Tau 2016 Nov 23 300 2016 Nov 21 600 600
HBC 334 2017 Jan 03 1800 L L L
HBC 551 2014 Feb 25 1200 L L L
HD 141569 2014 Jun 19 300 L L L
HD 142666 2014 Feb 24 300 L L L
2014 May 19 300 L L L
HD 144432 2014 May 19 300 L L L
HD 145718 2016 May 15 300 L L L
HD 150193 2014 May 19 600 L L L
2014 Jun 19 300 L L L
HD 163296 2016 May 15 30 2016 May 15 120 120
HD 169142 2014 May 19 300 L L L
HD 190073 2014 Oct 02 60 L L L
HD 200775 2017 Jan 22 30 2016 Nov 20 120 120
L L 2017 Jan 22 320 320
HD 216629 2016 Aug 10 30 L L L
HD 245185 2017 Jan 03 600 L L L
HD 250550 2017 Jan 21 600 2017 Jan 22 600 600
HD 259431 2015 Dec 16 60 2016 Nov 21 600 480
HD 31648 2016 Nov 20 60 2016 Nov 20 180 240
HD 35187 2017 Jan 21 60 L L L
HD 35929 2017 Jan 21 120 2016 Nov 21 480 480
HD 36112 2014 Feb 24 180 2014 Feb 24 120 120
2015 Jan 27 90 2016 Nov 21 360 360
HD 37490 2017 Jan 21 20 2017 Jan 21 320 400
HD 37806 2016 Nov 22 30 L L L
HD 52721 2016 Nov 22 30 2016 Nov 21 180 180
HD 53367 2016 Nov 23 90 2016 Nov 21 180 180
HK Ori 2017 Jan 03 600 L L L
LkHa 167 2016 Sep 25 1200 L L L
LkHa 198 2015 Dec 16 1200 L L L
LkHa 224 2016 May 16 900 L L L
LkHa 233 2014 Oct 02 1800 L L L
LkHa 234 2016 Aug 10 600 L L L
LkHa 257 2016 Aug 10 900 L L L
MWC 1080 2014 Aug 17 180 2014 Aug 17 200 200
2014 Nov 02 180 L L L
MWC 297 2014 May 19 360 2014 Jun 19 120 80
2014 Jun 19 600 2014 Aug 17 200 120
PDS 174 2017 Jan 03 900 L L L
PX Vul 2014 Oct 01 600 L L L
SV Cep 2016 Aug 10 300 L L L
UX Ori 2017 Jan 03 60 L L L
UY Ori 2017 Jan 03 900 L L L
V1012 Ori 2017 Jan 03 900 L L L
V1366 Ori 2017 Jan 03 600 L L L
V376 Cas 2014 Oct 02 1800 L L L
V594 Cas 2014 Oct 01 120 2016 Nov 20 300 400
2014 Nov 03 180 L L L
V699 Mon 2016 Nov 23 600 L L L
VV Ser 2016 May 15 1200 L L L
VY Mon 2017 Jan 22 900 2017 Jan 22 500 600
WW Vul 2016 Aug 10 600 L L L
2016 May 15 600 L L L
3
The Astrophysical Journal, 857:30 (9pp), 2018 April 10 Mathew et al.
Table 1
(Continued)
Object Date of Optical Date of Y band J band
Optical Observations Exp. Time (s) IR Observations Exp. Time (s) Exp. Time (s)
(1) (2) (3) (4) (5) (6)
XY Per 2016 Nov 22 300 2017 Jan 22 400 600
Z CMa 2014 Feb 24 180 2014 Feb 24 120 40
subtracted, continuum normalized spectra. LMFIT, which is estimated from AV using the extinction curve of McClure
based on an Marquardt Levenberg nonlinear least squares (2009).
minimization algorithm, was used to fit Gaussians to the The continuum flux densities of O Ilines 7774 and 8446 are
profiles. estimated from Hα continuum flux using the relation,
3.2.2. Removal of Paschen Line (P18) Contribution from O Iλ8446 ⎛ ⎞
Fl,cont (
F (7774)
7774) = lc⎝⎜ ⎟⎠ ´ Fl,cont (Ha)For the spectral resolution of our observations, the line Flc (Ha)
profiles of O Iλ8446 and Paschen 18 (P18; 8437Å) are
blended (see Figure 1). We proposed a method in Mathew et al. ⎛Flc (8446) ⎞
(2012b) to deblend the P18 contribution from the net EW in the Fl,cont (8446) = ⎝⎜ ⎟ ´ F (Ha).F (Ha) ⎠ l,cont
study of CBe stars, which will be employed here as well. The lc
Paschen line strengths show a monotonic increase with
The ratio of continuum ux densities, Flc (7774) Flc (8446)wavelength and then display a trend of attening out around fl and , arefl Flc (Ha) Flc (Ha)
P17 and beyond (Briot 1981). Hence it is reasonable to obtain calculated using the synthetic spectra given in Munari et al.
the EW of P18 by linearly interpolating between the measured (2005). The line fluxes of O Iλ7774 and O Iλ8446 are
EW of P17 (8467Å) and P19 (8413Å) (see Mathew obtained by taking a product of the continuum flux density with
et al. 2012b). This value is subtracted from the combined the measured EW. Similarly, the continuum flux density in the
EW of O Iλ8446 and P18 to obtain the intrinsic EW of near-IR region is calculated from the extinction corrected J
O Iλ8446. magnitudes of HAeBe stars. Furthermore, the flux values of
O Iλ11287 and O Iλ13165 are calculated from the continuum
3.2.3. Accounting for Photospheric Absorption
flux densities and the measured EWs.
The EWs calculated from emission lines needs to be
corrected for the photospheric absorption. The strength of the
absorption component is estimated from the synthetic spectrum 4. Results and Discussion
corresponding to the spectral type of the central star from 4.1. Excitation Mechanisms for O IEmission
Munari et al. (2005), which are calculated from the SYNTHE
code (Kurucz 1993), using NOVER models as the input stellar The excitation mechanisms contributing to O Iemission that
atmospheres (Castelli et al. 1997). The EW of the underlying are discussed extensively in the literature are recombination,
absorption component for Hα, O Iλ7774, and O Iλ8446 is collisional excitation, continuum fluorescence, and Lyβ
estimated using the synthetic spectra corresponding to the fluorescence (Grandi 1975b, 1980; Strittmatter et al. 1977;
spectral type of the star. Since the synthetic spectra of Munari Ashok et al. 2006; Banerjee & Ashok 2012). In this section, we
et al. (2005) do not cover the infrared spectral region, we used assess which one of the above is the dominant mechanism for
NextGen (AGSS2009) theoretical spectra (Hauschildt et al. the production of O Ilines in HAeBe stars.
1999) for the analysis of O Iλλ11287, 13165 line profiles. The
EW of the photospheric absorption is subtracted from the EW
of the observed emission line to obtain the net EW. 4.1.1. Recombination
One of the possible formation mechanisms of permitted
3.2.4. Estimation of Line Fluxes O Iemission lines is through recombination followed by a
The EW of O Iemission lines, corrected for photospheric cascade from higher ionization states. However, the recombi-
absorption, needs to be multiplied with the underlying stellar nation process alone is not sufficient to explain the strength of
continuum ux density to obtain the line ux. We are taking O Ilines in systems such as the Orion nebula (Grandi 1975b).fl fl
the extinction corrected R-band ux density as a proxy for the If the recombination process is the dominant mechanism,fl
continuum ux density underlying Hα line. The method of then the emission strengths of λ7774 and λ8446 should followfl
calculating the continuum ux density at O Iemission lines the ratio of statistical weights, i.e., F(λ7774)/F(λ8446)=5/3fl
from Hα line is described below. The continuum ux density at (Strittmatter et al. 1977; Grandi 1980). So, if recombinationfl
Hα is given as operates in HAeBe stars, we should expect O Iλ7774 to be
stronger than O Iλ8446. The flux ratio of O Iλ7774 and
(-R0 ) λ8446 is shown as a function of F(λ8446) in Figure 2. For 77%Fn,cont (Ha) = Fn,0 ´ 10 2.5 , of HAeBe stars, the emission strength of O Iλ8446 is stronger
than O Iλ7774. Hence, recombination is not likely to be the
where F -23 -2 -1n,0 = 3.08 ´ 10 W m Hz and R0 is the extinc- dominant excitation mechanism for the production of O Ilines
tion corrected RC magnitude. The extinction in R-band, AR, is in HAeBe stars.
4
The Astrophysical Journal, 857:30 (9pp), 2018 April 10 Mathew et al.
Figure 1. Observed spectra of V594 Cas, LkHα 233, and MWC 297 (top to bottom). The line profiles of Hα, O Iλ7774, and O Iλ8446 are shown in each case (left to
right). Ca IItriplet (8498, 8542, 8662 Å) lines are seen in most spectra and appear to be blended with Paschen lines when they are in emission.
4.1.2. Collisional Excitation greater than those predicted for densities >1011 cm−3.
Bhatia & Kastner (1995) built a hybrid model to compute the Additionally, models for Hα emission in HBe stars also12 −3
collisionally excited level populations and line intensities of require densities of ne=2×10 cm for the line forming
neutral oxygen under optically thin conditions. The intensities region (see Patel et al. 2016, 2017). Although these studies do
of all possible allowed and forbidden O lines in ultraviolet, not discuss the O Iline forming region, the strong correlationI
between Hα and O Iline emission (see Section 4.1.4) indicates
visible, and infrared wavelength regions were calculated over a
that both lines are formed in the same region. Thus, our
range of densities and temperatures seen in astrophysical
( ) analysis suggests that collisional excitation may not be thesystems. Kastner & Bhatia 1995 estimated the expected
F (8446) prominent mechanism at densities >10
11 cm−3 seen in the line
values of for various temperature–density combinations
F (7774) forming regions of HAeBe stars.
for collisional excitation. In the magnetospheric accretion We have included O Iλ λ7774,8446 line measurements of a
models for HAeBe stars (e.g., Muzerolle et al. 2004), most of sample of HAeBe stars studied in Fairlamb et al. (2017). These
the emission lines observed in the visible and near-IR objects were observed with the X-shooter spectrograph mounted
wavelengths are formed in magnetospheric accretion columns. at the Very Large Telescope, Chile. Figure 3 shows that the
It is possible that O Ilines also form in these accretion inclusion of the sample of HAeBe stars from X-Shooter provides
columns. The typical accretion rates for HAeBe stars are in the more data in the lower flux regime of 7774 and 8446 lines.
range of 1.0×10−8–1.0×10−6M −1e yr with a median value Furthermore, O Iλ8446 flux values are more intense than the
of ∼2.0×10−7M yr−1e (e.g., Mendigutía et al. 2011, 2012). theoretical estimates corresponding to T=5000/10,000K and
The corresponding density of accretion columns is in the range ne=10
11 cm−3. This analysis strengthens the claim that
of 1011–1013 cm−3 for temperatures of 6000–10000 K, for collisional excitation is not the dominant excitation mechanism
typical parameters of magnetospheric accretion models (see for the production of O Iemission lines in HAeBe stars.
Hartmann et al. 1994; Muzerolle et al. 1998, 2001, 2004). We Further confirmation is obtained from the analysis of the
have taken theoretical O Iline flux ratio values corresponding infrared spectra of HAeBe stars. It has been proposed that
to these temperature–density combinations from Kastner & if collisional excitation is the dominant excitation mechanism,
Bhatia (1995). Observational data is shown in Figure 3 for 50 the EW of λ13165 should be greater than that of λ11287,
HAeBe stars, including the measurements of 30 sources from i.e., W(13165)/W(11287)…1 at T=10,000 and 20,000 K,
Fairlamb et al. (2017). The flux values of O Iλ7774 and λ8446 respectively, for ne=10
10 1012 cm−3– (Bhatia & Kastner
corresponding to a temperature of 5000 K and densities of 1010, 1995). For most of our sample of stars, we found that the
1011, and 1012 cm−3 are represented as dotted lines in Figure 3 emission line strength of O Iλ11287 is higher than that of
and T=10,000 K, ne=10
10, 1011, and 1012 cm−3 combina- λ13165 (see Figure 4), confirming that collisional excitation
tions are shown in dashed lines. Figure 3 shows that the does not play a major role in the formation of O Iemission
observed flux ratio for almost all the sources in our sample is lines in HAeBe stars.
5
The Astrophysical Journal, 857:30 (9pp), 2018 April 10 Mathew et al.
Table 2
List of Compiled Stellar Parameters for Analysis of Optical Lines
Source Sp. Type Ref. Sp. Type Teff (K) V B−V RC Ref. Photometry AV
(1) (2) (3) (4) (5) (6) (7) (8) (9)
51 Oph B9.5 IIIe 1 10400 4.78 0.03 4.75 1 0.4
AB Aur A1 1 9200 7.05 0.12 6.92 1 0.39
AS 442 B8Ve 14 12500 10.9 0.66 10.18 3 3.85
AS 443 B2 1 20600 11.35 0.66 10.78 1 4.35
AS 505 B5Vep 15 15700 10.85 0.43 10.66 4 2.93
BD+30 549 B8p 16 12500 10.56 0.35 10.42 4 2.3
BD+40 4124 B3 1 17000 10.69 0.78 9.92 1 4.79
BD+46 3471 A0 1 9700 10.13 0.4 9.8 1 2
BD+65 1637 B4 1 16700 10.18 0.39 9.79 1 2.78
BO Cep F4 1 6640 11.6 0.56 11.21 1 0.74
CQ Tau F3 1 6720 10.26 0.79 9.72 1 2.01
HBC 334 B3 1 17000 14.52 0.57 13.95 1 3.74
HBC 551 B8 1 12500 11.81 0.26 11.54 1 1.85
HD 141569 A0Ve 1 9700 7.1 0.1 7.03 1 0.5
HD 142666 A8Ve 1 7500 8.67 0.5 8.34 1 1.25
HD 144432 A9IVe 1 7440 8.17 0.36 7.92 1 0.53
HD 145718 A5Ve 8 8080 9.1 0.52 8.79 2 1.8
HD 150193 A2IVe 1 8840 8.64 0.49 8.28 1 2.08
HD 163296 A1Vep 1 9200 6.88 0.09 6.82 1 0.24
HD 169142 A5Ve 1 8080 8.15 0.28 7.95 1 0.6
HD 179218 A0IVe 1 9700 7.39 0.08 7.33 1 0.4
HD 190073 A2IVe 1 8840 7.73 0.13 7.7 1 0.28
HD 200775 B3 1 17000 7.37 0.41 7.01 1 2.94
HD 216629 B3IVe+A3 17 17000 9.32 0.45 9.11 4 3.14
HD 245185 A1 1 9200 9.94 0.1 9.88 1 0.29
HD 250550 B9 1 10700 9.54 0.07 9.41 1 0.7
HD 259431 B6 1 14500 8.73 0.27 8.36 1 2.05
HD 31648 A3Ve 1 8550 7.7 0.2 7.59 1 0.55
HD 35187 A2e+A7 1 8840 8.17 0.22 76.4 1 0.73
HD 35929 F2III 1 6810 8.13 0.42 7.87 1 0.23
HD 36112 A5IVe 1 8080 8.34 0.26 8.16 1 0.5
HD 37490 B2 5 20600 4.57 −0.11 4.59 5 0.5
HD 37806 A2Vpe 1 8840 7.95 0.04 7.89 1 −0.17
HD 38120 B9 1 10700 9.01 0.06 8.93 1 0.65
HD 52721 B1 5 26000 6.62 0.06 6.53 5 1.69
HD 53367 B0IV/Ve 9 31500 6.95 0.42 6.67 2 3.64
HK Ori A4+G1V 1 8270 11.71 0.56 11.2 1 2.1
IP Per A6 1 8000 10.47 0.33 10.24 1 0.8
LkHa 167 A2 6 8840 15.06 1.42 14.32 4 6.73
LkHa 198 B9 1 10700 14.18 0.95 13.31 1 5.1
LkHa 224 F9 1 6040 14.07 1.44 12.98 1 4.44
LkHa 233 A4 1 8270 13.56 0.84 12.91 1 3.5
LkHa 234 B7 1 14000 12.21 0.9 11.49 1 5.14
LkHa 257 B5 7 15700 13 0.3 12.72 4 2.28
MWC 1080 B0eq 1 31500 11.52 1.34 10.39 1 8.24
MWC 297 B1.5Ve 10 24800 12.03 2.24 10.18 2 12.46
PDS 174 B3e 11 17000 12.84 0.81 12.18 2 4.94
PX Vul F3 1 6720 11.54 0.83 11.12 1 2.21
R Cra A0 1 9700 12.2 1.09 11.03 1 5.45
SV Cep A0 1 9700 10.98 0.39 10.68 1 1.95
UX Ori A3 1 8550 10.4 0.33 10.13 1 1.2
UY Ori B9 12 10700 12.79 0.37 12.56 2 2.2
V1012 Ori A3e 13 8550 12.04 0.42 11.61 2 1.65
V1366 Ori A0 1 9700 9.89 0.16 9.8 1 0.8
V376 Cas B5e 1 15700 15.55 1.13 14.59 1 6.43
V594 Cas B8 1 12500 10.58 0.56 10.03 1 3.35
V699 Mon B6 1 14500 10.54 0.54 10.06 1 3.4
VV Ser B6 1 14500 11.92 0.93 11.11 1 5.35
VY Mon B8 1 12500 13.47 1.55 12.19 1 8.3
WW Vul A3 1 8550 10.74 0.44 10.45 1 1.75
XY Per A5 1 8080 9.21 0.49 8.86 1 1.65
Z CMa B0 IIIe 1 31500 9.47 1.27 8.63 1 7.89
References.(1) Manoj et al. (2006); (2) Fairlamb et al. (2015); (3) Mendigutía et al. (2012); (4) Zacharias et al. (2004); (5) Hillenbrand et al. (1992); (6) Cohen &
Kuhi (1979); (7) Liu et al. (2011); (8) Carmona et al. (2010); (9) Tjin A Djie et al. (2001); (10) Drew et al. (1997); (11) Gandolfi et al. (2008); (12) Vieira et al. (2003);
(13) Lee & Chen (2007); (14) Mora et al. (2001); (15) Garrison (1970); (16) McDonald et al. (2017); (17) Skiff (2014).
6
The Astrophysical Journal, 857:30 (9pp), 2018 April 10 Mathew et al.
Figure 4. Plot of ratio of EWs of O Iλ11287 to O Iλ13165 against effective
temperature of the star. It can be seen that in most cases EW(11287) > 1.
Figure 2. Log–Log plot of F(λ8446)/F( EW(13165)λ7774) vs. F(λ8446): the error bars
are indicated. In 77% cases, F(λ8446)>F(λ7774).
in the text. In essence, λ13165 is predicted to be much stronger
than λ11287 if continuum fluorescence is the dominant
excitation mechanism for the O Ilines (also see Strittmatter
et al. 1977). This prediction by Grandi was confirmed
observationally for the Orion nebula in the spectroscopic studies
by Lowe et al. (1977). Also, strong O Iemission lines at 7002,
7254, and 7990Ålines would be observed in the spectra
(Strittmatter et al. 1977; Grandi 1980). Apart from the Orion
nebula, another instance where the 13165 line is stronger than
the 11287 line is in the inner 10 arcsec sized nebula surrounding
P Cygni. Near-infrared 1−2.5 μm spectra by Smith & Hartigan
(2006) of this region gives a value of 2.55±0.57 for the ratio of
the 13165 and 11287 line strengths.
Our analysis shows that the emission strength of O Iλ11287
is greater than that of O Iλ13165 for our sample of HAeBe
stars (Figure 4). In addition, we do not see emission lines at
7002, 7254, and 7990Åin any of the object spectra. This
suggests that continuum fluorescence is unlikely to be the
dominant mechanism for the formation of O Iemission lines in
HAeBe stars.
4.1.4. Lyman b Fluorescence
Figure 3. Log–Log plot of F(8446) vs. F(7774): red dotted lines correspond to
theoretical flux values for T=5000 K and black dashed lines correspond to Lyβ fluorescence occurs because the 3d3D0 level of O Iis
T=10,000 K, for n =1010, 1011, and 1012 cm−3e (Kastner & Bhatia 1995). populated by Lyβ radiation, with subsequent cascades produ-
The sources from Fairlamb et al. (2017) are shown by purple triangles. cing the O Iλλ11287, 8446, and 1304 lines in emission
(Figure 5). This is due to the near coincidence in wavelength of
4.1.3. Continuum Fluorescence Lyman beta and the O Iresonance line 2p 3P2–3d 3D 0321 at
1025.77Å(Bowen 1947). Our analysis shows that the cascade
Continuum fluorescence was invoked as the excitation lines expected from Lyβ fluorescence, O Iλ8446 and
mechanism for the production of O Ilines in the spectra of O Iλ11287, are quite strong in the spectra of HAeBe stars.
planetary nebulae (Seaton 1968) and Orion Nebula Also, O Iλ7774 is less intense than O Iλ8446, suggesting that
(Grandi 1975b). Grandi, in his thesis (Grandi 1975a) and in a collisional excitation and recombination are relatively less
paper summarizing the thesis results (Grandi 1975b), showed important for O Iexcitation in HAeBe stars. Similarly, the
that the expected theoretical ratio of the line strengths of the lower emission strength of O Iλ13165 with respect to λ11287
13165/11287 line due to starlight excitation (equivalently rules out collisional excitation and continuum fluorescence as
continuum fluorescence) should be of the order of 10 or slightly the dominant mechanisms for the production of O Ilines.
more. These model predictions are summarized in Tables 7 Furthermore, O I λλ7002, 7254, and 7990 emission lines are
and 2 of Grandi (1975a, 1975b), respectively, and also described not present in the spectra of HAeBe stars. These lines are
7
The Astrophysical Journal, 857:30 (9pp), 2018 April 10 Mathew et al.
Figure 6. Log–log plot of F(8446) vs. F(Hα): the sample of HAeBe stars from
Figure 5. Pumping and the fluorescent transitions in O Icaused by the Lyman our studies and from Fairlamb et al. (2017) are shown in black circles. We
beta fluorescence process. plotted 97 sources including 57 from Fairlamb et al. (2017), of which 7 are
common between Fairlamb et al. (2017) and our sample.
generally seen in sources where O Ilines are excited by
continuum fluorescence. All these pieces of evidence strongly Lyβ fluorescence operates in HAeBe stars suggests that
suggest that Lyβ fluorescence is likely to be the dominant O Iemission lines are also formed in these accretion columns
excitation mechanism for the production of O Ilines in HAeBe in HAeBe stars.
stars. It is worth noting that the gas has to be optically thick in Hα
If Lyβ fluorescence is responsible for O Iemission in in order for O Iλ8446 to get excited by Lyβ fluorescence. From
HAeBe stars, then one would expect a correlation between the equations of level populations in statistical equilibrium,
Hα and O Iλ8446 line intensities. Lyβ photon results from the Grandi (1980) derived an optical depth in Hα (τHα) of 2000,
n=3−1 transition of the hydrogen atom and the Hα photon considering Lyβ fluorescence operating in active galaxies. We
results from the transition n=3−2. Thus the upper level of examined whether the line forming region in HAeBe stars are
both the transitions are the same. In other words, hydrogen optically thick in Hα. For the sample of HAeBe stars considered
atoms in the excited state of n=3 are responsible for both in this study, the median flux ratio F(Hα/λ8446)obs is 65.2.
lines. If these lines originate from the same gas component, one From the analysis of level populations, Strittmatter et al. (1977)
would expect their intensities to be correlated. If, in addition, derived a theoretical Hα to O Iλ8446 flux ratio of around 7500,
O Iλ8446 intensity is proportional to Lyβ intensity, then one under optically thin conditions. The optical depth of Hα can be
would expect a correlation between Hα and O Iλ8446. This is estimated from the ratio of theoretical to observed line flux ratio,
shown in Figure 6, where F(λ8446) is shown as a function of F i.e., τHα=7500/F(Hα/λ8446)obs=115. The sufficiently high
(Hα) for our sample and those from Fairlamb et al. (2017). A value of optical depth in Hα derived in the case of HAeBe stars
correlation is seen between the flux values of Hα and agrees with the assumption that the gas needs to be optically
O Iλ8446, suggesting the application of the Lyβ fluorescence thick thereby trapping the Lyman beta photons, paving way for
process. A linear fit to the distribution of points in Figure 6 Lyβ fluorescence.
gives a relation of the form F(λ8446)∝F(Hα)1.02±0.04, with a
Pearson’s correlation coefficient of 0.96. From the analysis, it is
clear that Hα emission is correlated with the emission strength 5. Conclusion
of O Iλ8446. This particularly has implications in under-
standing the region of formation of O Iemission lines in From an analysis of the observed optical spectra of 62
HAeBe stars. HAeBe stars and near-infrared spectra of 17 HAeBe stars, we
If Lyβ photons and O Iatoms do not coexist, Lyβ have shown that Lyβ fluorescence is likely to be the dominant
fluorescence would not have been possible since Lyβ gets excitation mechanism for the formation of O Iemission lines.
scattered off neutral hydrogen, resulting in the production of We ruled out recombination and continuum fluorescence as the
Lyα and Hα, before interacting with neutral oxygen atoms. possible excitation mechanisms because the emission strengths
This is the reason why Lyβ fluorescence does not operate in the of O Iλ8446 and λ11287 are much stronger than those of the
Orion nebula, where Lyβ photons are trapped in the inner adjacent O Ilines at λ7774 and λ13165, respectively. We
regions of the nebula, whereas the O Iis confined in the found that collisional excitation does not contribute substan-
exterior (Grandi 1975b). In the case of HAeBe stars, Hα is tially to O Iemission from the comparative analysis of the
thought to originate in the magnetospheric accretion columns, observed line flux values of λ7774 and λ8446 with those
which connect the inner disk to the central star. The fact that predicted by the theoretical models of Kastner & Bhatia (1995).
8
The Astrophysical Journal, 857:30 (9pp), 2018 April 10 Mathew et al.
We would like to thank the referee for comments that helped Grandi, S. A. 1975b, ApJ, 196, 465
to improve the quality of the manuscript. We would also to Grandi, S. A. 1980, ApJ, 238, 10
thank the staff at IAO, Hanle, and its remote control station at Hamann, F., & Persson, S. E. 1992, ApJS, 82, 247
Hartmann, L., Hewett, R., & Calvet, N. 1994, ApJ, 426, 669
CREST, Hosakote, for their help during the observation runs. Hauschildt, P. H., Allard, F., & Baron, E. 1999, ApJ, 512, 377
This research uses the SIMBAD astronomical database service Herbig, G. H. 1960, ApJS, 4, 337
operated at CDS, Strasbourg. This publication made use of data Hernández, J., Calvet, N., Briceño, C., et al. 2004, AJ, 127, 1682
from 2MASS, which is a joint project of University of Hillenbrand, L. A., Strom, S. E., Vrba, F. J., & Keene, J. 1992, ApJ, 397, 613
Massachusetts and the Infrared Processing and Analysis Kastner, S. O., & Bhatia, A. K. 1995, ApJ, 439, 346Kurucz, R. 1993, ATLAS9 Stellar Atmosphere Programs and 2 km/s Grid,
Centre/California Institute of Technology, funded by the Kurucz CD-ROM No. 13 (Cambridge, MA: Smithsonian Astrophysical
National Aeronautics and Space Administration and the Observatory)
National Science Foundation. Lee, H.-T., & Chen, W. P. 2007, ApJ, 657, 884
Liu, T., Zhang, H., Wu, Y., et al. 2011, ApJ, 734, 22
ORCID iDs Lowe, R. P., Moorhead, J. M., & Wehlau, W. H. 1977, ApJ, 214, 712
Manoj, P., Bhatt, H. C., Maheswar, G., & Muneer, S. 2006, ApJ, 653, 657
Blesson Mathew https://orcid.org/0000-0002-7254-191X Mathew, B., Banerjee, D. P. K., Naik, S., & Ashok, N. M. 2012a, MNRAS,
P. Manoj https://orcid.org/0000-0002-3530-304X 423, 2486
Mathew, B., Banerjee, D. P. K., Subramaniam, A., & Ashok, N. M. 2012b,
ApJ, 753, 13
References McClure, M. 2009, ApJ, 693, L81
McDonald, I., Zijlstra, A. A., & Watson, R. A. 2017, MNRAS, 471, 770
Ashok, N. M., Banerjee, D. P. K., Varricatt, W. P., & Kamath, U. S. 2006, Mendigutía, I., Calvet, N., Montesinos, B., et al. 2011, A&A, 535, A99
MNRAS, 368, 592 Mendigutía, I., Mora, A., Montesinos, B., et al. 2012, A&A, 543, A59
Banerjee, D. P. K., & Ashok, N. M. 2012, BASI, 40, 243 Mora, A., Merín, B., Solano, E., et al. 2001, A&A, 378, 116
Bessell, M. S. 1983, PASP, 95, 480 Munari, U., Sordo, R., Castelli, F., & Zwitter, T. 2005, A&A, 442, 1127
Bhatia, A. K., & Kastner, S. O. 1995, ApJS, 96, 325 Muzerolle, J., Calvet, N., & Hartmann, L. 1998, ApJ, 492, 743
Bowen, I. S. 1947, PASP, 59, 196 Muzerolle, J., Calvet, N., & Hartmann, L. 2001, ApJ, 550, 944
Briot, D. 1981, A&A, 103, 5 Muzerolle, J., D’Alessio, P., Calvet, N., & Hartmann, L. 2004, ApJ, 617, 406
Carmona, A., van den Ancker, M. E., Audard, M., et al. 2010, A&A, 517, A67 Patel, P., Sigut, T. A. A., & Landstreet, J. D. 2016, ApJ, 817, 29
Castelli, F., Gratton, R. G., & Kurucz, R. L. 1997, A&A, 318, 841 Patel, P., Sigut, T. A. A., & Landstreet, J. D. 2017, ApJ, 836, 214
Cohen, M., & Kuhi, L. V. 1979, ApJS, 41, 743 Pecaut, M. J., & Mamajek, E. E. 2013, ApJS, 208, 9
Drew, J. E., Busfield, G., Hoare, M. G., et al. 1997, MNRAS, 286, 538 Seaton, M. J. 1968, MNRAS, 139, 129
Fairlamb, J. R., Oudmaijer, R. D., Mendigutia, I., et al. 2015, MNRAS, Skiff, B. A. 2014, VizieR Online Data Catalog, 1, 2023
453, 976 Slettebak, A. 1951, ApJ, 113, 436
Fairlamb, J. R., Oudmaijer, R. D., Mendigutia, I., et al. 2017, MNRAS, Smith, N., & Hartigan, P. 2006, ApJ, 638, 1045
464, 4721 Strittmatter, P. A., Woolf, N. J., Thompson, R. I., et al. 1977, ApJ, 216, 23
Felenbok, P., Czarny, J., Catala, C., & Praderie, F. 1988, A&A, 201, 247 The, P. S., de Winter, D., & Perez, M. R. 1994, A&AS, 104, 315
Finkenzeller, U., & Mundt, R. 1984, A&AS, 55, 109 Tjin A Djie, H. R. E., van den Ancker, M. E., Blondel, P. F. C., et al. 2001,
Gandolfi, D., Alcalá, J. M., Leccia, S., et al. 2008, ApJ, 687, 1303 MNRAS, 325, 1441
Garrison, R. F. 1970, AJ, 75, 1001 Vieira, S. L. A., Corradi, W. J. B., Alencar, S. H. P., et al. 2003, AJ, 126, 2971
Gorti, U., & Bhatt, H. C. 1993, A&A, 270, 426 Waters, L. B. F. M., & Waelkens, C. 1998, ARA&A, 36, 233
Grandi, S. A. 1975a, PhD thesis, Univ. Arizona Zacharias, N., Monet, D. G., Levine, S. E., et al. 2004, BAAS, 36, 48
9