A Spectroscopic Classification Survey to Search for New ρ Puppis Stars
Courtney McGahee1
, Richard O. Gray1
, R. E. M. Griffin2
, Mariah Birchard1,3
, and Jared Day1
1
Department of Physics and Astronomy, Appalachian State University Boone, NC 28608, USA; mcgaheece@appstate.edu
2
Herzberg Astrophysics & Astronomy Research Centre, National Research Council 5071 West Saanich Road, Victoria, BC V9E 27, Canada
3
Gemini South Observatory, Alberto Arenas Carvajal, La Serena, Coquimbo, Chile
Received 2019 November 7; revised 2020 May 5; accepted 2020 May 19; published 2020 July 2
Abstract
The ρPuppis stars are mid-F-type stars that show peculiar chemical-abundance patterns similar to those of the Am
stars. Typically they exhibit overabundances of iron-peak elements such as Fe and Ni and s- and r-process
elements such as Sr and Eu, and underabundances of certain other elements including He, Ca, and Sc. It has been
proposed that these stars are evolving Am stars passing through the short-lived phase that occurs between the reestablishment
of convection and the consequent erasure of their chemical peculiarities. In this paper we suggest a
second hypothesis: these stars may have acquired their peculiar abundance patterns in a fashion similar to the
Barium stars, i.e., they may have gained the s-process element enhancements via mass transfer from a once
asymptotic giant-branch companion star, now turned white dwarf. This study will detail our efforts to investigate
the ρPuppis stars with regard to these two hypotheses, with a view to understanding these stars and their
significance in terms of stellar evolution. We have performed a spectral classification survey, and a detailed
chemical-abundance analysis of selected ρPuppis stars. This paper gives an overview of the ρPuppis stars and
describes the spectral classification survey which has increased the number of ρPuppis stars currently known from
6 to 49. A future paper will describe the chemical-abundance analysis which should provide insight into the nature
of these stars.
Unified Astronomy Thesaurus concepts: Stellar classification (1589); Chemically peculiar stars (226); Am stars (33);
Barium stars (135)
Supporting material: machine-readable tables
1. Introduction
The ρPuppis classification, established by Gray & Garrison
(1989), refers to stars that appear to show abundance patterns
similar to those of the metallic-line A-type (Am) stars, but
whose effective temperatures resemble those of the mid-F-type
stars. Like Am stars, ρPuppis stars appear to show overabundances
of iron-peak elements such as Fe and Ni and of sand
r-process elements such as Sr and Eu. However, they also
tend to show underabundances of certain other elements,
including Ca. It has been suggested that the process of diffusion
that is held to be responsible for the abundance anomalies in
Am stars cannot be maintained below a certain lower effectivetemperature
limit of F2 - =b y 0.22[ mag] (Smith 1973);
however, the effective temperatures of ρPuppis stars are cooler
than that limit and yet they still show Am-like abundance
patterns.
The most commonly accepted hypothesis to explain the
ρPuppis phenomenon is that these stars are in fact evolving Am
stars, caught in the short-lived phase between the re-establishment
of convection and the subsequent erasure of the chemicalabundance
peculiarities. This hypothesis was initially put
forward by Kurtz (1976) to explain the abundance anomalies
of the δ Delphini stars. It is suspected that this group may be
related to the ρPuppis stars (see Section 3.1). We suggest here a
different hypothesis to explain the abundance anomalies of the
ρPuppis stars, namely, that the ρPuppis stars are not related to
the Am stars, but instead to the Barium stars, a group that shares
certain abundance peculiarities with the ρPuppis stars. The
origin of the Ba-star abundance peculiarities is thought to be
mass transfer from an asymptotic giant-branch (AGB) companion
star, now turned white dwarf (McClure 1983). In this study
we report on our investigation of the ρPuppis stars with respect
to these two hypotheses, with a view to determining which
(if either) of those physical mechanisms is responsible for
producing the observed chemical peculiarities. To that end we
first conducted a spectroscopic classification survey which
increased the number of ρPuppis stars currently known from
6 to 49. We then carried out a chemical-abundance analysis of
selected ρPuppis and Am stars that allowed a more detailed
look at the chemical makeup of the atmospheres of the target
stars. This paper concentrates on the spectroscopic classification
survey, following a review of the Am and ρPuppis stars. The
chemical-abundance analysis will be described in a future paper.
2. Background of Am and ρPuppis Stars
2.1. The Am Stars
The standard explanation of the appearance of Am/Fm
abundance anomalies is the model of diffusion/chemical separation
(see Watson 1971; Michaud et al. 1983; Charbonneau &
Michaud 1991; Charbonneau 1993). In that model, both radiative
(gR) and gravitational (g) accelerations must be considered.
Elements for which gR>g will tend to rise in the atmosphere,
whereas those with gR<g will sink. In normal A- and early
F-type stars, rotationally driven meridional circulation overwhelms
chemical separation, but below a certain critical rotational velocity,
the gravitational settling of helium results in the disappearance of
the helium convection zone. In the absence of that zone, chemical
separation can occur all the way up to the base of the superficial
hydrogen convection zone. This mechanism, when combined with
a small mass-loss rate (∼10−15
Me yr−1
), successfully predicts
the Am/Fm abundance patterns and the existence and magnitude
of the rotational cut-off velocity (Charbonneau 1993). In addition,
The Astronomical Journal, 160:52 (8pp), 2020 July https://doi.org/10.3847/1538-3881/ab974c
© 2020. The American Astronomical Society. All rights reserved.
1
the disappearance of the helium convection zone suggests that
Am/Fm stars should not exhibit δ Scuti pulsations (which are
driven by instabilities in that zone), and indeed, with very few
exceptions, Am/Fm stars are not δ Scuti pulsators (Breger 1970).
It is well known that the incidence of Am/Fm anomalies
drops rapidly in the early F-type stars. The chemical separation
theory suggests that that behavior is related to the deepening of
the superficial hydrogen convection zone with lowering
effective temperature, Teff. As convective currents deepen,
mixing should prevent the gravitational settling of helium.
Therefore, radiative diffusion can only lead to the Amabundance
pattern above a certain minimum Teff. Furthermore,
there exists a critical rotational velocity Vcrit as a function of
Teff; a star that is rotating faster than Vcrit cannot maintain
stratification in its atmosphere, so only stars with a rotational
velocity of Vrot<Vcrit(Teff) will exhibit the Am-abundance
pattern (Smith 1973). Vcrit decreases with Teff until it
approaches 0 kms−1
at F2, establishing a lower limit for Teff
for Am/Fm stars. This finding agrees qualitatively with the
discovery that the number of stars with large metallicity (m1)
indices decreases dramatically for types later than F1
- =b y 0.21[( ) mag] (Strömgren 1963).
2.2. The ρPuppis Stars
Gray & Garrison (1989) contested the lower limit determined
for Teff for Am/Fm stars, reporting that a small subclass
of Fm stars exists that are cooler than spectral type F2. This
distinct group had only four members (ρ Pup, θ Gru,
HD 103877, and τ UMa) and was labeled ρPuppis” after the
brightest in the class. Since that time, Kurtz et al. (1995) and
Murphy et al. (2012) have classified HD40765 and
HD178875, respectively, as members of the ρPuppis class
to bring the total number of known ρPuppis stars to six. Stellar
parameters and spectral types for four of the six known
ρPuppis stars are reported in Table 1.
From the standpoint of spectral classification, the ρPuppis
stars appear to be quite similar to the Am/Fm stars in that they
share the same abundance anomalies, i.e., overabundances of
iron-peak and heavy metals (including Sr and Ba), and
underabundances of Ca and Sc. Furthermore, ρ Puppis stars
exhibit the anomalous luminosity effect, a feature commonly
found among Am stars having to do with wavelengthdependent
luminosity classification (see Section 3.3 for more
detail). The existence of the ρPuppis stars clearly poses a
theoretical problem since they are Am-like stars yet they are
cooler than the theoretical lower Teff boundary for the Amphenomenon
(metallicism) established by Smith (1973).
However, Gray & Garrison (1989, p. 314) investigated the
Balmer jump, via the c1 index, for each of the known ρPuppis
stars and found that they lie above the zero-age main sequence,
suggesting that they “may be unusually cool, evolved, Am
stars.”
Moreover, ρPuppis itself is known to undergo largeamplitude
δScuti pulsations (Mathias et al. 1997). It is
perplexing on two counts that ρPuppis exhibits δScuti
pulsations. The first is that this star presumably lacks the
driving mechanism (He ionization) for those pulsations to take
place, and yet they are present. The second is that, despite these
pulsations, we still see the Am-abundance anomalies that
should be erased owing to mixing induced by the pulsations.
The fact that ρPuppis is experiencing δScuti pulsations may
prove consistent with the theory that ρPuppis stars are simply
evolved Am stars, in that (1) it indicates that either the He
ionization zone remained intact or has been re-established in
order to drive those pulsations (Kurtz 1976), and (2) it was
predicted by Smith (1973) that Am stars that had cooled
enough to regain their He ionization zones and that had crossed
the critical rotational velocity barrier would be favored to
exhibit δScuti-type pulsations. Gray & Corbally (2009)
corroborated the suggestion of Gray & Garrison (1989), in
light of the points above, that ρPuppis stars may in fact be Am
stars in a short-lived phase between the re-establishment of
their convective zones and the consequent erasure of their
chemical peculiarities, which is in accord with the finding of
Kurtz (1976) that evolved Am stars can pulsate while
maintaining metallic-line anomalies.
2.3. The Barium Dwarf Stars
We now put forward a second hypothesis to explain the
existence of ρPuppis stars, and suggest that the ρPuppis stars
are related to another group of chemically peculiar F-type stars
known as the barium dwarf (Ba dwarf) stars. Barium stars, both
giants and dwarfs, show chemical peculiarities somewhat
similar to Am stars. It is believed that the barium stars gained
those peculiarities via mass transfer from an evolving former
AGB companion star, now turned white dwarf (McClure 1983;
North & Lanz 1991). Spectroscopically, the Ba dwarfs differ
from the Am stars in that all known Ba dwarfs tend only to
show enhancements of the s-process elements as opposed to an
overall metal enhancement as do the Am stars. Furthermore,
they have spectral types F4 and later. Nonetheless, there are
theoretical reasons to suspect that Ba dwarfs exist with earlier
spectral types and higher masses.
Observational evidence suggests that the Ba dwarfs and the
Ba giants are linked through their evolution. (1) The Ba giants
and the CH subgiants, the latter identified by North et al.
(2000), exhibit similar chemical-abundance patterns (Smith
et al. 1993). (2) The Ba dwarfs and Ba giants share similar
orbital characteristics (North et al. 2000; Escorza et al. 2019).
(3) White dwarf cooling times indicate that most Ba giants
Table 1
Stellar Parameters of Known ρPuppis Stars
ID Teff log(g) [Fe/H] Sp. Type Reference
ρPup 6691 3.23 0.36 kF2hF5mF5 II Gray et al. (2003)
HD103877 7200 L 0.66 kA7hF5mF5 (II) The current work
τUMa 6290 4.10 0.24 kA5hF0mF5 II Gray et al. (2003) a
HD178875 7300 3.40 0.05 kA9hF2mF3 D Abt (1984)
Notes. We report median values derived from the bibliographical catalog, PASTEL (Soubiran et al. 2016).
a
The spectral type determined in the current work is F6 IIIn. See Table 2 for additional notes.
2
The Astronomical Journal, 160:52 (8pp), 2020 July McGahee et al.
underwent mass-transfer contamination while still on the main
sequence (Böhm-Vitense et al. 2000). Those points imply that
the Ba dwarfs are the progenitors of the Ba giants (BöhmVitense
et al. 2000; North et al. 2000; Gray et al. 2011). Given
the masses of the known Ba giants, this hypothesis requires that
there should be a significant population of Ba dwarfs with
spectral types as early as A0; however, to date no Ba dwarf has
been detected with a spectral type earlier than F3. In fact, the
Ba giants are statistically more massive than the Ba dwarfs,
indicating that only the most massive of the known Ba dwarfs
will become Ba giants (North et al. 2000; Escorza et al.
2017, 2019). Therefore, it seems that we are failing to detect
the massive Ba dwarfs necessary to account for the known Ba
giant population. It has been proposed that these missing
massive Ba dwarfs may be masquerading as Am/Fm or Ap
stars (North et al. 2000; Gray et al. 2011; Escorza et al. 2019).
We now propose that it is possible, given their similar
abundance patterns, spectral types, and masses, that some of
the massive Ba dwarfs may be hiding among the ρPuppis
stars. In fact, the ρPuppis stars may comprise a subset of Ba
dwarfs that are metal-rich. de Castro et al. (2016) and Escorza
et al. (2017) report that the Ba dwarfs, on average, have
subsolar metallicities, yet the ages of the Ba stars reported by
de Castro et al. (2016; 60 Myr–10 Gyr) indicate that most of
these objects are younger that the Sun; there is no reason to
expect Ba dwarfs to be metal-poor. The ρPuppis stars, which
tend to be metal-rich (see Section 3.1) may be the missing link.
Another reason to consider our hypothesis will be discussed in
Section 4.
If the Ba dwarfs are, indeed, all members of binary systems
with white dwarf companions, and the ρPuppis stars are, in
fact, related to the Ba dwarfs, one would expect that the
ρPuppis stars should be found in single-lined spectroscopic
binary (SB1) systems. Escorza et al. (2019) determined that the
orbital periods of the white dwarf companions of Ba dwarfs
range from 1.2 to 23 yr. There is very little evidence in the
literature either for or against such orbital periods for the
known ρPuppis stars:
ρPuppis (single). Rainer et al. (2016) observed ρPuppis
with the High Accuracy Radial velocity Planet Searcher
(HARPS) spectrograph (Mayor et al. 2003) 360 times over a
period of slightly more than 4 yr (2008 December–2013
January), and found no evidence for binarity.
θGru (multiple). θGru is a well-known visual binary and
has recently been shown to form a physical triple with
HD218205 (Tokovinin et al. 2015). Horch et al. (2000, 2001,
2006) carried out speckle observations of the θGru system, but
only reported the separation and position angle of the known
visual companion. No radial velocity data that could shed light
on a possible SB1 nature for θGru appears to exist.
HD103877 (SB?). Grenier et al. (1999) found that this star
exhibits radial velocity variations, but the period of those
variations was not reported.
τUMa (SB1). Simbad lists this star as a spectroscopic
binary. Bretz (1961) identified τUMa as an SB1 and
determined orbital parameters, including a period of 1062.4
days (2.9 yr).
HD40765 (single?). Kurtz et al. (1995) finds that HD40765
is a δScuti pulsator with (1) nonsinusoidal light variability
with a period of 4.070hr, (2) a peak-to-peak B-magnitude
amplitude of 0.21mag, (3) a peak-to-peak V-magnitude
amplitude of 0.15mag, and (4) surface peak-to-peak radial
velocity variations of 14kms−1
. They go on to argue that it is
extremely unlikely that a hypothesis of binarity could explain
these results given the relationship between the amplitude and
incidence of δScuti pulsation. No long-term radial velocity
data appears to exist for HD40765.
HD178875 (SB?). Although Dommanget & Nys (1994) and
Catanzaro et al. (2011) proposed that HD178875 is a double
system, Murphy et al. (2012, p. 1421) argues that “there is no
evidence...to rule out chance alignment” and uses frequency
modulation techniques to indicate that it is, in fact, a single
star. Additionally, Lampens et al. (2018, p. 7) found small
(< 5 km s−1
), long-term (several months) radial velocity variations,
but because of spectral line profile variations, concluded that
the star is “a pulsator accompanied by long-term RV variability.”
Furthermore, Murphy et al. (2018) employed phase modulation
techniques to come to the conclusion that HD178875 is a single
star.
Because the evidence for the binary nature of the known
ρPuppis stars is either nonexistent or inconclusive, we suggest
that a better way to decide between the two hypotheses
regarding the origin of the abundance anomalies of the
ρPuppis stars is to conduct a detailed chemical-abundance
analysis. Barium dwarfs obtain their chemical abundances via
the s-process which causes enhancements of Sr, Ba, and other
s-process elements. The ρPuppis stars also show enhancements
of both Sr and Ba (as will be demonstrated), but in our
study we will pay close attention to the abundance of Eu, as it
is largely an r-process element, and is not normally enhanced in
Barium stars. If Eu is found to be consistently overabundant in
the ρPuppis stars, it will help to confirm the evolved Am-star
hypothesis, whereas if we find solar abundances of Eu in
ρPuppis stars it will point to the mass-transfer hypothesis.
3. A Spectral Classification Survey
3.1. Selection of ρPuppis Candidates
To date, only six stars are classified as ρPuppis types.
Therefore, our first task was an attempt to discover more
members of the class. To that end we followed two observing
strategies. The first involved taking a closer look at the stars
previously classified as δ Delphini stars, a now defunct
classification that was introduced by Bidelman (1965) and
whose history is described by Gray & Corbally (2009).
Because all of the ρPuppis stars identified by Gray & Garrison
(1989) were initially classified as δ Del stars, we chose
additional ρPuppis candidates from the stars classified as δ Del
in the Michigan HD reclassification project (Houk & Cowley
1973), volumes 4 (Houk & Smith-Moore 1988) and 5 (Houk &
Swift 1999), which cover declinations accessible from the Dark
Sky Observatory. We have obtained spectra of as many of
those as possible, at classification resolution.
That δ Del target list of only ∼100 stars proved to be too
short, and did not provide us with a sufficient number of
candidate stars . Upon classification, only six of the observed δ
Del targets proved to be ρPuppis stars. Therefore, we
employed a second strategy in which we selected roughly
800 stars from the Mermilliod photometric catalog (Mermilliod
et al. 1997) that used Strömgren photometry. Given that the
known ρPuppis stars stand apart from their normal F-type
dwarf counterparts in that they have higher metallicities and
Balmer jumps that are indicative of the early F-type giants
(Gray & Garrison 1989), we selected additional ρPuppis
3
The Astronomical Journal, 160:52 (8pp), 2020 July McGahee et al.
candidates by employing reddening-free versions of the
Strömgren metallicity and Balmer jump indices, defined by
= + -m m b y0.32 11 1[ ] ( ) ( )
= - -c c b y0.2 . 21 1[ ] ( ) ( )
Figure 1 shows a plot of the stars in the Mermilliod catalog
in the [m1], [c1] plane. It is evident in that figure that the known
ρPuppis stars are closely grouped, and possess higher [m1]
values than most of the stars with similar [c1] values. We
defined our sample in the [m1], [c1] plane by selecting a region
(indicated by the box) that contains the known ρPuppis stars.
The additional stars that are contained in that region were then
selected as ρPuppis candidates.
3.2. Classification-resolution Observations
Classification-resolution spectra of 631 of the ρPuppis
candidates were obtained with the Gray–Miller Cassegrain
spectrograph on the 0.8 m DFM Engineering telescope of
Appalachian State University’s Dark Sky Observatory (DSO),
located in the Blue Ridge mountains of North Carolina. During
the course of our observations, the detector was replaced once.
The two detectors that were employed for observations were:
(1) a back-illuminated 1024×1024 Tektronix CCD (used in
the MPP mode) with 24μm pixels and (2) a 1024×256 e2v
thinned and back-illuminated thermoelectrically cooled CCD
with 27μm pixels. The 600 and 1200 grooves/millimeter
gratings were used in conjunction with a 100 μm wide slit to
produce 3.6 and 1.8 Å/2 pixel resolutions (R=1300 and
2300) with spectral ranges of λλ3800–5600 and 3800–4600 Å
respectively. Exposure times were calculated so as to achieve a
signal-to-noise ratio (S/N) > 100. The observed spectra were
reduced using standard methods (including flat-fielding, dark
subtraction, and bias removal) in IRAF.4
The wavelength
calibration was determined using comparison spectra obtained
with an Fe–Ar hollow-cathode lamp. The reduced spectra were
rectified using an X-Window System program, xmk25, written
by R. Gray (Gray & Corbally 2009).
3.3. MK Spectral Classifications
The spectral classifications of the program stars observed at
DSO were carried out on the MK system (Morgan et al. 1943).
The current state of this system is described in Gray &
Corbally (2009).
In order to understand how the spectra of the ρPuppis
stars compare with those of the Am stars, we first detail the
spectral characteristics of Am stars. The spectrum of an
Am star is set apart from the spectrum of a normal A-type star
in that its metallic-line type appears to be cooler (later) than
the hydrogen-line type, and the Ca II K-line type appears to be
earlier than the hydrogen-line type. These discrepancies
are illustrated in Figure 2. An A- or early F-type star is
considered to be Am or Fm, respectively, if the difference
between the Ca II K-line type and the metallic-line type is at
least five subclasses. Because of these discrepancies between
classification features, the Am, Fm, and ρPuppis stars are
classified for each feature separately. The hydrogen-line type
of an Am/Fm star is most closely correlated with its effective
temperature.
In addition to the hydrogen-line, metallic-line, and Ca II
K-line discrepancies, many Am stars exhibit a discrepancy in
their luminosity classification criteria too, as mentioned in
Section 2.2. In Am stars exhibiting the anomalous luminosity
effect (Abt & Morgan 1976), the strength of the luminositysensitive
Fe II/Ti II blends in the λλ4395–4480 region
will often be comparable to the strength in a normal dwarf
star, while—simultaneously—the strength of the Fe II/Ti II
blends in the λλ4172–4179 region and the Sr II lines at
λ 4077 and λ 4216 can be comparable to the strength seen
in a giant star, and in some extreme cases even a supergiant.
In the case of a star exhibiting the anomalous luminosity
effect, it is common practice to determine the luminosity
class on the basis on the Fe II/Ti II blends in the λ 4172–4179
region, but to emphasize that this class may not reflect the true
luminosity of the star, that luminosity class so derived is
enclosed in parentheses. Figure 2 illustrates the anomalous
luminosity effect in the classical Am star 63Tau.
Figure 3 illustrates how the ρPuppis stars resemble the Am
stars spectroscopically. Notice that the Ca II K lines of both
ρPuppis stars are weak for their hydrogen-line types. Unlike
what is observed in Am stars, the metallic-line types of the
ρPuppis stars tend to be the same as their hydrogen-line types.
However, both are “late” as the hydrogen-line type necessary
for a ρPuppis classification is F3 or later (typically F5). Also
note that the appearance of the anomalous luminosity effect can
be extreme in ρPuppis stars. In particular, HD 103877 shows
line strengths in the λλ4395–4480 region that are clearly
depressed relative to those in the giant standard, 60UMa, while
the Fe II/Ti II blends in the λλ4172–4179 region are noticeably
stronger than those in the giant standard.
Because ρPuppis stars appear to have similarities to Am
stars, we identified stars as ρPuppis only if they (1) showed
Am-type discrepancies in their hydrogen-line, Ca II K-line,
and metallic-line types, (2) exhibited the anomalous luminosity
effect, and (3) had hydrogen-line types of F3 or later.
Our spectral classifications and the broader, corresponding
classification families for the ρPuppis candidates observed at
the DSO are presented in Table 2 for stars selected from the
the Mermilliod Catalog (Mermilliod sample), and in Table 3
for stars selected from the Michigan HD reclassification
project catalogs (Houk sample). Note that in Tables 2 and 3
Figure 1. Plot of the Mermilliod Catalog stars according to their uvby
photometric indices [m1] and [c1]. Four of the known ρPuppis stars are
indicated by the red dots. The black box indicates the boundaries of the chosen
[m1],[c1] region from which additional ρPuppis candidates were selected.
4
IRAF is distributed by the National Optical Astronomy Observatories,
which are operated by the Association of Universities for Research in
Astronomy, Inc., under cooperative agreement with the National Science
Foundation.
4
The Astronomical Journal, 160:52 (8pp), 2020 July McGahee et al.
the Am/Fm classification families have been further
divided using the term “extreme” to denote Ca II K- and
metallic-line discrepancies larger than six spectral subclasses,
the term “late” to denote metallic-line types of F3 or cooler,
and the term “hyper” to denote Am/Fm stars with hydrogenline
types earlier than F3 but that are both “extreme”
and “late.”
4. Discussion
Of the 631 ρPuppis candidates, 24% were identified as
normal A- and F-type stars and 52% were identified as either
Am, Fm, Ba Dwarf, or ρPuppis (see Table 4). Of the 45
candidates that were classified as ρPuppis, 43 are new
classifications. That significantly increases the number of
known ρPuppis stars.
Figure 2. Two Am stars, 63Tau and HD46614, are compared with two MK standards, 30LMi and βCas. Note that the Ca II K-line of the classic Am star 63 Tau is
closer to that of an A1.5 standard star, indicating an earlier temperature type than is indicated by the hydrogen lines, which are a good match with the A9 standard star
30 LMi. In addition, we see that the overall metallic-line spectrum for 63 Tau is somewhat stronger than the F2 standard. The metallic-line type is actually a good
match with an F3 standard, and is later than the temperature type indicated by the hydrogen lines. The anomalous luminosity effect can be observed by noting that the
Fe II/Ti II blends in the λλ4395–4480 region in 63 Tau are weaker than those in the same region in the giant standard stars; however, the Fe II/Ti II lines at
λλ4172–4179 are about same strength as those in the standard stars. The standard practice is to use the strength of the features in the λλ4172–4179 region; it results
in a luminosity classification of giant for the Am star. Adapted from a figure from Gray & Corbally (2009), with permission from Princeton University Press.
Figure 3. Two ρPuppis stars, ρPuppis itself and HD103877, are compared with two MK standards, 60UMa and α Per. Note that the luminosity classification for
HD1038777, as shown here, is based on the strong Sr II line at λ4077 and that the luminosity classification for this star depends on the wavelength region used.
Adapted from a figure in Gray & Corbally (2009), with permission from Princeton University Press.
5
The Astronomical Journal, 160:52 (8pp), 2020 July McGahee et al.
Interestingly, among the Am, Fm, and ρPuppis stars in our
sample there is a bimodal distribution according to hydrogenline
type, which is the dominant factor in determining whether or
not a ρPuppis-like star is considered to be a bona fide ρPuppis
member (see Figure 4). A peak in the number of stars at the
hydrogen-line type A9 is evident, and after that the number of
stars drops off sharply. The sharp disjunct peak at F5
corresponds to the ρPuppis stars. The rapid drop in the number
of Fm stars after F1 is in accord with the results of Smith (1973)
and Michaud et al. (2005), but the existence of the ρPuppis stars
as a well-separated group is new. The Hartigan dip test of
unimodality (Hartigan & Hartigan 1985) was performed for both
the Mermilliod and the Houk frequency distributions of Am, Fm,
and ρPuppis stars by hydrogen-line type; it yielded dip statistics
of 0.11379 and 0.09375 for sample sizes of N=272 and
N=48 for the Mermilliod and the Houk samples, respectively.
The corresponding probability that the distribution is at least
bimodal is over 99.9% for the Mermilliod sample and between
99.5% and 99.9% for the Houk sample, indicating that the
ρPuppis stars do indeed form a second population whose
origin may be physically distinct from that of the Am/Fm
stars. That result motivates us more strongly to explore the
mass-transfer (Ba dwarf) hypothesis put forward in Section 2.2
by carrying out a chemical-abundance analysis to be described in
a future paper.
Table 2
Spectral Classifications and Notes for Stars Selected from the Mermilliod Catalog (Mermilliod et al. 1997)
Classification Spectral
HD BD Mag Family Classification Notes
16 35.05164 8.092 Normal F5.5 III-IV G band is sl weak
154 33.04835 8.73 Am kA8hA8mF1 (III) Strong ALE; Ba just slightly enhanced
156 −18.06426 7.309 BaD F4 IV Sr Mild Ba Dwarf; sl metal weak for F4 IV
416 36.00004 8.87 Hyper Am kA7hA8mF5 (III) Ba Mg 4481 enhanced
848 67.00006 8.15 Composite A/F composite K-line A2; G band F5
1151 1.00025 8.43 Am kA7hA8.5mF2 (III) Ba Mg 4481 enhanced
1169 7.00023 7.588 Am kA8hA9.5mF2 (III) Mg 4481 enhanced but Ba 4554 normal
1601 48.00084 6.473 Normal F6 II-III
1607 21.00025 8.680 ρPup kA5hF5mF5 (III) ρPup
1616 9.00030 8.620 λBooa
F0 Vn kA6mA5 Possible mild λ Boo star; H lines F0 V, but clearly metal weak; sl rotationally
broadened
Note.
a
We point the reader to Murphy & Paunzen (2017) for a detailed description of the λBoo stars.
(This table is available in its entirety in machine-readable form.)
Table 3
Spectral Classifications and Notes for Stars Selected from the Michigan HD Reclassification Project (Houk; Houk & Cowley 1973)
Classification Spectral
HD BD Mag Family Classification Notes
728 Normal K0 IV Simbad lists as a variable δ Sct star; further investigation needed
3024 Am kA7hA9mF1 (III) EuSr
4630 Am kA7hA9mF1 (III) SiEuBa Sr 4077 enhanced but not Sr 4216; Mg 4481 enhanced
7119 Am kA5hA8mF1 (III) SiEuBa Sr 4077 enhanced but not Sr 4216
7133 ρPup KA6hF3mF5 (III) SiEuSrBa ρPup
8251 Late Fm kA8hF0mF4 (III) EuSr
15306 BaD F4 kF1mF2 V Sr Well-known Ba Dwarf
16641 Am kA5hA7mF0 (III) Ba
16932 Am kA4hA9mF0 (III) SrBa
18460 8.44 Hyper Am kA7hA7mF4 II Marginal ALE; H lines better match with lum class III
(This table is available in its entirety in machine-readable form.)
Table 4
Constituents of Classified Stars
Classification Number Percentage of
Family of Stars Classified Stars
Normal 154 24.3
Am 140 22.2
Fm 136 21.6
ρPuppis 44 7.1
Ba Dwarf 6 1.0
λBoo 2 0.3
SB2a
2 0.3
Compositeb
54 8.6
Peculiarc
42 6.8
Rapid Rotatord
51 7.9
Total 631 L
Notes. The broad classification “family” categories encountered in this work
are listed here. The percentage of stars that fell into each group are calculated
with respect to the entire candidate list including both the Mermilliod and the
Houk samples.
a
The spectra of these stars show double-lined features at classification
resolution (R=2300).
b
The Ca II K-line region of a composite spectrum tends to be dominated by the
early star, whereas the G-band region tends to be dominated by the late star.
c
The specific peculiarities of the individual spectra are indicated in the notes
section of the classification tables (Tables 2 and 3).
d
These spectra are marked by significant broadening of the spectral features.
6
The Astronomical Journal, 160:52 (8pp), 2020 July McGahee et al.
Abt (2017) argues against the hypothesis that ρPuppis stars
are Am-star descendants by pointing out thatin the Bright Star
Catalog (BSC; Hoffleit & Jaschek 1982),the number of
ρPuppis stars relative to the number of Am stars is too small
by a factor of roughly 100. This is determined by considering
the main-sequence and subgiant evolutionary lifetimes which
are nearly equal, and are intended to correspond to the Am stars
and ρPuppis stars respectively. It should be noted, however,
that the BSC is neither a volume-limited sample,nor have
many of its classifications been confirmed (Griffin et al. 2012),
so this estimate may not be accurate. A second issue with this
argument becomes clear when we consider the Am stars in the
volume-limited NStars sample (Gray et al. 2003, 2006).
Figure 5 plots 11 of the 13 Am stars that were identified in
the NStars project and for which Gaia parallaxes exist (Gaia
Collaboration et al. 2016, 2018). Evolutionary tracks from
Pietrinferni et al. (2004) for solar–metallicity stars between
1.5Me and 2.2Me are also plotted. It is evident that 9 of the 11
Am stars are in fact evolved in the sense that they no longer lie
near the zero age main sequence (ZAMS); indeed 3 of them are
near the hydrogen-exhaustion stage. It may be that many of the
Am stars identified in the BSC are also evolved,should
therefore be included with the ρPuppis stars when determining
the ratio of normal Am stars to evolved Am-type stars, and
would thereby reduce the aforementioned factor of 100.
Additionally, when we consider a timescale for the short-lived
evolutionary state in the evolved Am-star hypothesis, we must
remember that ρPuppis stars lie in a temperature range where
convection in the outer envelope is important and that it is the
convective mixing rather than the subgiant evolutionary
timescale that is relevant; it is convection that will erase the
ρPuppis chemical peculiarities from the atmosphere (see
Section 2.1). Since the convective timescale is much shorter
than the evolutionary timescale, this also helps to explain why
Figure 4. Am, Fm, and ρPuppis stars grouped according to the hydrogen-line (H-line) type.
Figure 5. 11 Am stars identified in the NStars project are compared to the solar–metallicity evolutionary tracks of Pietrinferni et al. (2004) for starsbetween 1.5Me
and 2.2Me. Uncertainties in MV and B−V are shown for the eight stars that have published B and V errors. It can be seen that only two of these stars appear to be true
zero-age main-sequence stars, six appear to be evolved main-sequence stars, and three appear to be nearing the termination of core hydrogen fusion.
7
The Astronomical Journal, 160:52 (8pp), 2020 July McGahee et al.
the number of ρPuppis stars is smaller than the number
predicted by Abt. As a case in point, the composite-spectrum
binary system, oLeo, supports the plausibility of a short-lived
phase of evolution between the re-establishment of convection
and the erasure of the chemical peculiarities for Am stars.The
primary component of o Leo may be an example of an
extremely late ρPuppis star given its Am-like abundance
anomalies, cool temperature(Teff=6100 K), and therefore late
spectral type (∼F8). Griffin (2002) examines various scenarios
and concludes that it is likely that o LeoA is experiencing the
short-lived phase of evolution described in the present work.
Furthermore, Griffin (2002) argues that the rarity of binaries
similar to o LeoA suggests that this phase is extremely rapid.
The next step in this investigation is to conduct a detailed
chemical-abundance analysis of the ρPuppis stars in order to
gain more insight into the chemical makeup of the atmospheres
of these stars. We hope it will help to answer the question of
whether these stars are in fact related to the Am stars or to the
Ba dwarfs.
C.E.M. acknowledges Sigma Xi for research travel support
through their Grants-In-Aid of Research program. R.O.G.,
C.E.M., and R.E.M.G. acknowledge support through NASA
grant NNX09AD64G under the GALEX Guest Investigator
program. R.E.M.G. is grateful for continuing Guest Worker
privileges at the DAO (National Research Council of Canada).
This work has made use of data from the European Space
Agency (ESA) mission Gaia (https://www.cosmos.esa.int/
gaia), processed by the Gaia Data Processing and Analysis
Consortium (DPAC,https://www.cosmos.esa.int/web/gaia/
dpac/consortium). Funding for the DPAC has been provided
by national institutions, in particular the institutions participating
in the Gaia Multilateral Agreement.
ORCID iDs
Courtney McGahee https://orcid.org/0000-0001-6967-1142
Richard O. Gray https://orcid.org/0000-0001-7588-0477
References
Abt, H. A. 1984, ApJ, 285, 247
Abt, H. A. 2017, PASP, 129, 044201
Abt, H. A., & Morgan, W. W. 1976, ApJ, 205, 446
Bidelman, W. P. 1965, VA, 8, 53
Böhm-Vitense, E., Carpenter, K., Robinson, R., Ake, T., & Brown, J. 2000,
ApJ, 533, 969
Breger, M. 1970, ApJ, 162, 597
Bretz, M. C. 1961, ApJ, 133, 139
Catanzaro, G., Ripepi, V., Bernabei, S., et al. 2011, MNRAS, 411, 1167
Charbonneau, P. 1993, in ASP Conf. Ser. 44, IAU Coll. 138: Peculiar versus
Normal Phenomena in A-type and Related Stars, ed. M. M. Dworetsky,
F. Castelli, & R. Faraggiana (San Francisco, CA: ASP), 474
Charbonneau, P., & Michaud, G. 1991, ApJ, 370, 693
de Castro, D. B., Pereira, C. B., Roig, F., et al. 2016, MNRAS, 459, 4299
Dommanget, J., & Nys, O. 1994, CoORB, 115, 1
Escorza, A., Boffin, H. M. J., Jorissen, A., et al. 2017, A&A, 608, A100
Escorza, A., Karinkuzhi, D., Jorissen, A., et al. 2019, A&A, 626, A128
Gaia Collaboration, Brown, A. G. A., Vallenari, A., et al. 2018, A&A, 616, A1
Gaia Collaboration, Prusti, T., de Bruijne, J. H. J., et al. 2016, A&A, 595, A1
Gray, R. O., & Corbally, C. J. 2009, Stellar Spectral Classification (Princeton,
NJ: Princeton Univ. Press)
Gray, R. O., Corbally, C. J., Garrison, R. F., et al. 2006, AJ, 132, 161
Gray, R. O., Corbally, C. J., Garrison, R. F., McFadden, M. T., &
Robinson, P. E. 2003, AJ, 126, 2048
Gray, R. O., & Garrison, R. F. 1989, ApJS, 69, 301
Gray, R. O., McGahee, C. E., Griffin, R. E. M., & Corbally, C. J. 2011, AJ,
141, 160
Grenier, S., Baylac, M. O., Rolland, L., et al. 1999, A&AS, 137, 451
Griffin, R. E. 2002, AJ, 123, 988
Griffin, R. E., Gray, R. O., & Corbally, C. J. 2012, A&A, 547, A8
Hartigan, J. A., & Hartigan, P. M. 1985, AnSta, 13, 70
Hoffleit, D., & Jaschek, C. 1982, The Bright Star Catalogue (4th revised ed.;
New Haven, CT: Yale Univ. Observatory)
Horch, E., Franz, O. G., & Ninkov, Z. 2000, AJ, 120, 2638
Horch, E., Ninkov, Z., & Franz, O. G. 2001, AJ, 121, 1583
Horch, E. P., Baptista, B. J., Veillette, D. R., & Franz, O. G. 2006, AJ,
131, 3008
Houk, N., & Cowley, A. 1973, in IAU Symp. 50, Spectral Classification and
Multicolour Photometry, ed. C. Fehrenbach & B. E. Westerlund
(Dordrecht: Reidel), 70
Houk, N., & Smith-Moore, M. 1988, Michigan Catalogue of Two-dimensional
Spectral Types for the HD Stars. Volume 4, Declinations −26°.0 to −12°.0
(Ann Arbor, MI: Univ. Michigan Press)
Houk, N., & Swift, C. 1999, Michigan Catalogue of Two-Dimensional Spectral
Types for the HD Stars, Vol. 5 (Ann Arbor, MI: Univ. Michigan Press)
Kurtz, D. W. 1976, ApJS, 32, 651
Kurtz, D. W., Garrison, R. F., Koen, C., Hofmann, G. F., & Viranna, N. B.
1995, MNRAS, 276, 199
Lampens, P., Frémat, Y., Vermeylen, L., et al. 2018, A&A, 610, A17
Mathias, P., Gillet, D., Aerts, C., & Breitfellner, M. G. 1997, A&A, 327, 1077
Mayor, M., Pepe, F., Queloz, D., et al. 2003, Msngr, 114, 20
McClure, R. D. 1983, ApJ, 268, 264
Mermilliod, J.-C., Mermilliod, M., & Hauck, B. 1997, A&AS, 124, 349
Michaud, G., Richer, J., & Richard, O. 2005, ApJ, 623, 442
Michaud, G., Tarasick, D., Charland, Y., & Pelletier, C. 1983, ApJ, 269, 239
Morgan, W. W., Keenan, P. C., & Kellman, E. 1943, An Atlas of Stellar
Spectra, with an Outline of Spectral Classification (Chicago, IL: Univ.
Chicago Press)
Murphy, S. J., Grigahcène, A., Niemczura, E., Kurtz, D. W., &
Uytterhoeven, K. 2012, MNRAS, 427, 1418
Murphy, S. J., Moe, M., Kurtz, D. W., et al. 2018, MNRAS, 474, 4322
Murphy, S. J., & Paunzen, E. 2017, MNRAS, 466, 546
North, P., Jorissen, A., & Mayor, M. 2000, in IAU Symp. 177, The Carbon Star
Phenomenon, ed. R. F. Wing (Dordrecht: Kluwer), 269
North, P., & Lanz, T. 1991, A&A, 251, 489
Pietrinferni, A., Cassisi, S., Salaris, M., & Castelli, F. 2004, ApJ, 612, 168
Rainer, M., Poretti, E., Mistò, A., et al. 2016, AJ, 152, 207
Smith, M. A. 1973, ApJS, 25, 277
Smith, V. V., Coleman, H., & Lambert, D. L. 1993, ApJ, 417, 287
Soubiran, C., Le Campion, J.-F., Brouillet, N., & Chemin, L. 2016, A&A,
591, A118
Strömgren, B. 1963, in Basic Astronomical Data: Stars and Stellar Systems, ed.
K. A. Strand (Chicago, IL: Univ. Chicago Press), 123
Tokovinin, A., Pribulla, T., & Fischer, D. 2015, AJ, 149, 8
Watson, W. D. 1971, A&A, 13, 263
8
The Astronomical Journal, 160:52 (8pp), 2020 July McGahee et al.