MULTIPLICITY AMONG F-TYPE STARS

Stars of spectral type F provide the bridge from slowly rotating, Sun-like objects with large convective envelopes, toward younger and higher-mass objects, which are much less dominated by convection and consequently have equatorial rotational velocities that can easily exceed 100 km s⁻¹. This in turn causes dramatic changes in the spectra of these stars, many of which present broad and shallow absorption lines, which are all more or less mutually blended. The physics of the early-F stars is then more difficult to derive from their spectra; specifically, many of these "stars" actually consist of two or even more components. Thus, it happens that, although a significant fraction among the F-type stars are spectroscopic binaries, their double-lined character is often masked and difficult, if not impossible, to discover. A prominent example is the bright α Com, a near-equal-mass and equal-luminosity F-type binary at a distance of only 18 pc that was already investigated by Otto Struve in the 19th century (Struve 1875) and whose P ≃ 26 yr orbital period was possible to derive from visual observations, that is, as a result of its nearness. At only somewhat larger distances, the binary character of the perfectly edge-on α Com would have gone unnoticed for at least another hundred years, as its projected rotational velocity vsin i ≃ 20 km s⁻¹ also gives no spectroscopic indication of a double-lined system in the first place.

Today, visual and spectroscopic observations are to a great extent aided by other methods, such as precision astrometry, high-contrast direct imaging, or long-baseline interferometry, and much progress is particularly achieved for the nearest stars within 25 pc: the bright and fast-rotating F0IV α Hyi, for instance, was only uncovered as a P = 606 day astrometric binary from the space-borne Hipparcos observations in the late 1990s. Similarly, the nearby F6.5V star HR 2500 was first noted as an astrometric binary in the proper motion Δμ catalog of Wielen et al. (1999) and later confirmed by Makarov & Kaplan (2005) and Frankowski et al. (2007). The recent infrared direct imaging observations of HR 2500 by Ehrenreich et al. (2010) not only revealed this very companion, but also resolved it as a close Ba–Bb binary in the M/L dwarf mass regime. For another bright and nearby F star, α Cae, the observations of Ehrenreich et al. (2010) likewise demonstrated that its known, ρ ∼ 7 arcsec distant M-type companion also consists of two red dwarfs only 0.5 arcsec apart. Near-infrared 1.6 μm adaptive optics observations of the young and nearby F7V star υ Aqr by Lafrenière et al. (2007) describe a ρ = 6 arcsec distant companion that already showed up as a candidate in the Two Micron All Sky Survey (2MASS) catalog. Speckle measurements of the nearby F5V star 37 Cet by Tokovinin et al. (2010) resolved a visually 4.5 mag fainter secondary at ρ = 0.169 arcsec. Interferometric VLTI observations of the early-F star μ Vir by Richichi et al. (2009) have recently resulted in a suspiciously large angular diameter that leads to an implausibly low effective temperature, thereby suggesting that the star is either oblate or that it harbors a close companion. Similarly, the CHARA-based angular diameter for the slightly evolved F5 star ψ¹ Dra A reported by Boyajian et al. (2012) implies a grossly discrepant effective temperature of around 6000 K, but as the authors aptly point out, the cause for this may be the Ab companion that Toyota et al. (2009) only recently uncovered in their precision radial velocities of this star.

Very recently, we presented two nearby F-type stars, ξ Gem and λ Ara, as near-equal-mass binary candidates by combining their precise Hipparcos parallaxes with the spectroscopic signature from their pressure-dependent Mg Ib triplet lines. For both stars, it is the considerable rotational velocity that prevents a direct spectral resolution: while λ Ara rotates at vsin i = 15 km s⁻¹, the projected rotational velocity of ξ Gem amounts to vsin i = 67 km s⁻¹, which is already at the limit of this detection method (cf. Fuhrmann et al. 2011a, 2011b). For even higher velocities, as they usually occur for early-F stars and toward A-type stars, one must naturally expect that a much larger fraction of close companions is still awaiting detection.

The important point to appreciate is that even the nearest stars are actually very distant objects in terms of their ability to provide excellent hiding places not only for faint, low mass or degenerate objects, but also for bright, massive, and equal-luminosity companions. The situation is particularly severe for the stars in the Southern Hemisphere and it is perhaps best documented with the fast-rotating, second-magnitude star δ Vel (Argyle et al. 2002; Kellerer et al. 2007), which one can find at a declination of δ = −54^o, and which may in fact consist of three, four, five, or even six components. Although δ Vel is already of spectral type A, its projected rotational velocity vsin i ≃ 140 km s⁻¹ (Pribulla et al. 2011) is quite typical for other southern solar neighborhood early F-type stars, like α Hyi, ψ Vel, or η Sco, and therefore provides a very instructive case for the multiplicity uncertainties that exist even among the nearest stars.

Investigations on the multiplicity of stars have a long tradition and have necessarily resulted in an ever increasing fraction of binary or higher level systems with the course of time. The radial velocity studies by Abt & Levy (1976) and Duquennoy & Mayor (1991), for instance, present major efforts in that respect. Both studies focused on solar-type stars, but one of the major problems at that time certainly happened to be the absence of accurate stellar parallaxes required for an unbiased, volume-complete sample. With the Hipparcos astrometry, this situation is now considerably relaxed and Raghavan et al. (2010) have recently presented an all-sky multiplicity study of solar-type stars within 25 pc. Similar to Duquennoy & Mayor (1991), their sample commences at about F6, but reaches down to K3, for a total of 454 stars. A completely different approach is that of Eggleton & Tokovinin (2008). They consider two Hipparcos subsets with H_p ⩽ 4.0 mag and H_p ⩽ 6.0 mag, i.e., mostly early-type stars, but also out to much larger distances. Hence, and as Eggleton & Tokovinin (2008) demonstrate, the fraction of unknown companions, particularly for the larger H_p ⩽ 6.0 mag sample, is substantial.

Our own census on the stars within 25 pc and with M_V ⩽ 6.0 had previously been restricted to declinations north of δ = −15^o (Fuhrmann 2011). Recently, we began the complementary work in the Southern Hemisphere with the BESO high-resolution échelle spectrograph of the Universitätssternwarte Bochum near Cerro Armazones in Chile. The model atmosphere analyses discussed herein are all part of this project, which aims at a complete all-sky sample of the ∼500 solar-type stars within that volume. With the present work we report on a subset of these stars—with a major focus on those of spectral type F.

In Section 2 we briefly refer to the observations and analysis methods. The basic stellar parameters of the investigated objects are then presented in Section 3 and the results and the discussion on the stellar multiplicities follow in Section 4.

The high-resolution spectroscopic observations in the Southern Hemisphere have recently been described in Fuhrmann et al. (2011b). In brief, the employed BESO échelle spectrograph—a twin of the ESO FEROS instrument (Kaufer et al. 1999) in La Silla—produces high signal-to-noise ratio (S/N) spectra. These have an average resolution of R = 50, 000 and a wavelength coverage of λλ3620–8530 with a single exposure. We usually aim at two spectra for each star, but if necessary, a time-series of particular objects can immediately be initiated to follow, for instance, the periastron passages of binary systems or the non-radial, g-mode pulsations that one encounters for a substantial fraction among the early-F stars. The data reduction then includes the usual standard processing, with the resultant spectra being subject to grids of model atmospheres for solar-type dwarfs and subgiants at any metallicity and/or iron-to-α-element abundance mixtures.

The stellar effective temperatures are either derived from the Balmer line wings or refer to the LTE iron ionization equilibrium. The surface gravities mostly rely on Hipparcos astrometry, but also on the prominent Mg Ib triplet lines as well as the iron ionization equilibrium. Equivalent widths are measured from theoretical profile fits that include the macroturbulence and projected rotational velocities as well as the instrumental profile. The microturbulence velocity is set by the usual requirement that the elemental abundances may not depend on the equivalent widths. For a more general background on these procedures, their interdependencies, and their reliability, refer to Fuhrmann et al. (1993, 1997) and Fuhrmann (2011 and references therein).

Since essentially all of the previous analyses were done with the FOCES échelle spectrograph (Pfeiffer et al. 1998), there is a general interest in the comparison of those results with the BESO spectra for stars that are observable from both celestial hemispheres. Reanalyses of Procyon, β Vir, λ Ser, and η Ser, all with basically reassuring results, are therefore part of this work and we shall continue with these kinds of verifications in future contributions.

The single or single-lined stars in this work are set out in Table 1. Here, the given uncertainties are all meant as 2σ errors, although, and rather conservatively, we prefer to adopt fixed values of 0.1 dex, 0.2 km s⁻¹, 0.05 dex, and 0.05 mag for the surface gravity log g, microturbulence ξ_t, abundance ratio [Fe/Mg], and bolometric correction BC_V, respectively. In contrast to the previous work on the northern nearby stars with the FOCES spectrograph, we also make use of the Hipparcos parallaxes (van Leeuwen 2007) for the surface gravity determination. The only two exceptions in Table 1 are the very nearby Procyon (F5IV–V) and the subgiant ν Oct, where the adopted parallax is that of Gatewood & Han (2006) and Ramm et al. (2009), respectively. The masses (and ages) of the stars are either deduced from previously published evolutionary tracks in collaboration with Jan Bernkopf (e.g., Bernkopf et al. 2001) or refer to the recently updated grids of VandenBerg et al. (2006).

Table 1. Stellar Parameters of the Program Stars

Object	HR	V	Teff	log g	[Fe/H]	ξt	[Fe/Mg]	ζRT	v sin i	Mbol	BCV	Mass	Radius
	HD	(mag)	(K)	(cgs)	(dex)	(km s−1)	(dex)	(km s−1)	(km s−1)	(mag)	(mag)	(M☉)	(R☉)
θ Scl A	35	5.236	6395	4.25	−0.07	1.51	−0.03	5.9	1.0	3.52	−0.07	1.25	1.40
	739	0.005	80	0.10	0.07	0.20	0.05		1.0	0.05	0.05		0.05
ζ Tuc	77	4.219	5935	4.37	−0.24	1.04	−0.08	4.3	2.0	4.42	−0.13	0.98	1.07
	1581	0.005	60	0.10	0.06	0.20	0.05		1.0	0.05	0.05		0.03
β Hyi	98	2.801	5841	3.95	−0.15	1.16	−0.06	5.0	3.0	3.31	−0.13	1.11	1.85
	2151	0.005	60	0.10	0.06	0.20	0.05		0.7	0.05	0.05		0.06
ν Phe	370	4.966	6084	4.31	+0.15	1.14	+0.03	4.7	3.7	3.98	−0.09	1.17	1.25
	7570	0.005	80	0.10	0.07	0.20	0.05		0.6	0.05	0.05		0.05
χ Eri A	566	3.698	5135	3.42	−0.18	1.00	−0.07	3.8	2.0	2.16	−0.28	1.58	4.06
	11937	0.005	80	0.10	0.07	0.20	0.05		1.0	0.05	0.05		0.16
κ For A	695	5.200	5870	3.99	−0.06	1.22	−0.05	5.1	2.3	3.37	−0.12	1.13	1.78
	14802	0.010	70	0.10	0.06	0.20	0.05		0.8	0.06	0.05		0.07
ι Hor	810	5.399	6050	4.36	+0.16	1.08	+0.02	4.6	5.8	4.14	−0.09	1.16	1.18
	17051	0.005	70	0.10	0.07	0.20	0.05		0.6	0.05	0.05		0.04
τ1 Eri A	818	4.464	6340	4.28	+0.09	1.51	−0.01	5.8	25.0	3.63	−0.07	1.27	1.35
	17206	0.005	80	0.10	0.10	0.20	0.05		1.5	0.08	0.05		0.06
κ Ret A	1083	4.705	6557	4.10	−0.14	1.92	−0.05	7.0	14.8	2.96	−0.07	1.37	1.73
	22001	0.005	80	0.10	0.07	0.20	0.05		0.7	0.05	0.05		0.06
τ6 Eri	1173	4.214	6508	4.10	+0.05	1.74	−0.06	6.7	15.0	2.93	−0.06	1.44	1.78
	23754	0.005	80	0.10	0.06	0.20	0.05		0.7	0.05	0.05		0.06
Ret A	1355	4.435	4760	3.34	+0.32	0.92	+0.01	3.0	1.5	2.68	−0.45	1.09	3.71
	27442	0.005	90	0.10	0.10	0.20	0.05		1.0	0.05	0.05		0.17
ζ Dor A	1674	4.703	6104	4.40	−0.15	1.22	−0.05	4.9	15.2	4.27	−0.10	1.09	1.09
	33262	0.005	80	0.10	0.07	0.20	0.05		0.6	0.05	0.05		0.04
γ Lep A	1983	3.585	6255	4.27	−0.11	1.31	−0.03	5.5	7.9	3.75	−0.09	1.18	1.32
	38393	0.005	80	0.10	0.07	0.20	0.05		0.6	0.05	0.05		0.05
η Lep	2085	3.701	6872	4.15	−0.08	2.09	−0.07	7.9	16.9	2.79	−0.05	1.47	1.70
	40136	0.005	80	0.10	0.10	0.20	0.05		0.7	0.05	0.05		0.06
ν2 CMa	2429	3.959	4745	3.11	+0.20	0.99	+0.00	3.3	1.5	2.03	−0.45	1.19	5.05
	47205	0.005	90	0.10	0.10	0.20	0.05		1.0	0.05	0.05		0.23
	2500 A	5.916	6210	4.28	−0.23	1.14	−0.04	5.4	5.4	3.88	−0.10	1.09	1.26
	49095	0.005	80	0.10	0.07	0.20	0.05		0.6	0.06	0.05		0.05
Procyon*	2943	0.361	6470	3.96	−0.03	1.92	−0.07	6.8	2.7	2.58	−0.06	1.46	2.11
	61421	0.005	80	0.10	0.07	0.20	0.05		1.0	0.05	0.05		0.07
	2998 A	5.043	5405	3.74	+0.00	1.05	+0.00	3.7	2.5	2.88	−0.20	1.38	2.63
	62644	0.005	80	0.10	0.06	0.20	0.05		1.0	0.06	0.05		0.11
	3018 Aa	5.378	5856	4.24	−0.79	1.10	−0.44	3.6	1.0	4.30	−0.16	0.86	1.16
	63077	0.010	70	0.10	0.07	0.20	0.05		1.0	0.05	0.05		0.04
	3570	5.699	6296	4.28	−0.01	1.51	−0.03	5.6	10.6	3.71	−0.08	1.22	1.33
	76653	0.005	80	0.10	0.07	0.20	0.05		0.5	0.05	0.05		0.05
	3578	5.805	5910	4.15	−0.87	1.18	−0.42	4.1	1.0	4.03	−0.16	0.86	1.30
	76932	0.005	70	0.10	0.07	0.20	0.05		1.0	0.05	0.05		0.04
	3862	4.923	6085	4.28	−0.04	1.05	−0.02	4.8	5.2	3.94	−0.10	1.12	1.27
	84117	0.005	80	0.10	0.06	0.20	0.05		0.9	0.05	0.05		0.05
	4134	4.885	6133	3.97	−0.26	1.28	−0.11	5.5	8.7	3.09	−0.10	1.17	1.86
	91324	0.005	90	0.10	0.07	0.20	0.05		0.8	0.05	0.05		0.07
	4523 A	4.886	5684	4.45	−0.32	0.96	−0.21	3.2	1.5	4.90	−0.16	0.90	0.94
	102365	0.005	70	0.10	0.07	0.20	0.05		1.0	0.05	0.05		0.03
β Vir*	4540	3.596	6055	4.07	+0.11	1.36	−0.03	5.2	2.5	3.31	−0.09	1.25	1.72
	102870	0.005	70	0.10	0.06	0.20	0.05		1.0	0.05	0.05		0.06
	4600	5.140	6470	4.10	−0.18	1.63	−0.07	6.6	14.3	3.05	−0.08	1.33	1.70
	104731	0.005	80	0.10	0.07	0.20	0.05		0.8	0.05	0.05		0.06
	4979	4.847	5710	3.92	+0.17	1.12	+0.01	4.3	2.0	3.14	−0.13	1.34	2.09
	114613	0.005	70	0.10	0.06	0.20	0.05		1.0	0.05	0.05		0.07
	4989 A	4.898	6258	4.16	−0.28	1.40	−0.11	5.8	9.3	3.50	−0.10	1.14	1.47
	114837	0.005	80	0.10	0.07	0.20	0.05		0.7	0.05	0.05		0.05
61 Vir	5019	4.729	5570	4.43	−0.02	0.83	−0.06	2.7	2.0	4.90	−0.17	0.92	0.98
	115617	0.005	70	0.10	0.07	0.20	0.05		1.0	0.05	0.05		0.03
θ Cen+	5288	2.062	4766	2.48	−0.07	1.25	−0.02	4.4	1.0	0.37	−0.42	1.27	10.77
	123139	0.005	80	0.10	0.10	0.20	0.05		1.0	0.05	0.05		0.45
	5356 A	5.855	6040	4.38	−0.66	1.16	−0.19	4.6	1.0	4.43	−0.14	0.93	1.03
	125276	0.005	70	0.10	0.07	0.20	0.05		1.0	0.06	0.05		0.04
λ Ser*	5868	4.417	5920	4.18	−0.01	1.09	−0.05	4.5	2.0	3.89	−0.11	1.05	1.38
	141004	0.005	60	0.10	0.06	0.20	0.05		1.0	0.05	0.05		0.04
	6094	5.370	5848	4.49	+0.01	1.02	+0.03	3.9	2.0	4.71	−0.12	1.06	0.97
	147513	0.005	80	0.10	0.06	0.20	0.05		1.0	0.05	0.05		0.04
ζ TrA Aa	6098	4.898	6040	4.42	−0.08	1.14	+0.01	4.6	1.5	4.38	−0.10	1.08	1.06
	147584	0.005	80	0.10	0.06	0.20	0.05		1.0	0.05	0.05		0.04
ξ Oph A	6445	4.373	6625	4.16	−0.24	2.03	−0.05	7.2	20.1	3.10	−0.07	1.31	1.58
	156897	0.005	80	0.10	0.08	0.20	0.05		1.0	0.05	0.05		0.05
μ Ara	6585	5.125	5725	4.22	+0.27	1.01	+0.01	3.9	2.0	4.04	−0.13	1.14	1.37
	160691	0.005	70	0.10	0.06	0.20	0.05		1.0	0.05	0.05		0.05
58 Oph	6595	4.861	6295	4.22	−0.10	1.44	−0.06	5.6	12.3	3.54	−0.08	1.22	1.43
	160915	0.005	80	0.10	0.10	0.20	0.05		0.8	0.05	0.05		0.05
ι Pav A	6761	5.468	5895	4.25	−0.11	1.02	−0.04	4.1	2.5	4.12	−0.12	1.02	1.25
	165499	0.005	70	0.10	0.06	0.20	0.05		1.0	0.05	0.05		0.04
η Ser*	6869	3.240	4917	3.06	−0.20	1.14	−0.09	4.0	1.5	1.55	−0.35	1.46	5.87
	168723	0.005	80	0.10	0.07	0.20	0.05		1.0	0.05	0.05		0.24
ω Sgr A	7597	4.692	5485	3.63	−0.03	1.06	−0.02	4.5	1.5	2.43	−0.18	1.53	3.14
	188376	0.005	90	0.10	0.07	0.20	0.05		1.0	0.16	0.05		0.27
δ Pav	7665	3.550	5645	4.38	+0.39	0.92	+0.04	3.1	2.0	4.48	−0.14	1.17	1.16
	190248	0.005	80	0.10	0.08	0.20	0.05		1.0	0.05	0.05		0.04
ϕ2 Pav	7875	5.109	6039	3.92	−0.43	1.43	−0.18	5.3	5.3	3.03	−0.12	1.17	1.97
	196378	0.005	80	0.10	0.06	0.20	0.05		0.7	0.05	0.05		0.07
γ Pav	8181	4.212	6110	4.33	−0.67	1.26	−0.21	4.4	1.0	4.24	−0.14	0.95	1.10
	203608	0.005	80	0.10	0.07	0.20	0.05		1.0	0.05	0.05		0.04
ν Oct A	8254	3.738	4860	3.12	+0.18	1.04	−0.02	3.7	2.0	1.62	−0.39	1.61	5.81
	205478	0.005	80	0.10	0.08	0.20	0.05		1.0	0.05	0.05		0.24
τ PsA	8447	4.935	6335	4.25	+0.11	1.38	−0.02	5.8	13.9	3.56	−0.07	1.28	1.40
	210302	0.005	80	0.10	0.07	0.20	0.05		0.8	0.05	0.05		0.05
	8531 A	5.315	5685	4.00	−0.01	1.09	−0.02	3.9	1.7	3.60	−0.15	1.06	1.70
	212330	0.005	70	0.10	0.06	0.20	0.05		1.0	0.05	0.05		0.06
	8843	5.644	6205	4.35	−0.02	1.23	+0.00	5.3	7.9	4.00	−0.09	1.16	1.19
	219482	0.005	80	0.10	0.06	0.20	0.05		0.6	0.05	0.05		0.04

Notes. For each star the second row indicates the error estimates. Macroturbulence velocities ζ_RT are adopted from the relations given in Gray (1984, 1992). The bolometric corrections are taken from Alonso et al. (1995). Errors of log g, ξ_t, [Fe/Mg], and BC_V are generally assessed as 0.1 dex, 0.2 km s⁻¹, 0.05 dex, and 0.05 mag, respectively. All other errors are meant as 2σ errors. Uncertainties in the stellar masses are likely less than 10%—except for the giant θ Cen that may exceed this value. * = northern sample stars; + = effective temperature deduced from angular diameter.

Download table as:  ASCIITypeset images: 1 2

For all faster rotating stars with projected rotational velocities above 20 km s⁻¹, as well as for all double-lined spectroscopic binaries, dedicated discussions of their analyses follow hereafter.

χ Eri. The most remarkable aspect of this τ ≃ 2 Gyr old visual binary is its chromospheric and coronal activity. Thus, while the considerable X-ray luminosity L_x = 504.4 × 10²⁷ erg s⁻¹ (Hünsch et al. 1999) may in part (or mostly) originate from the Δm_V ≃ 7 mag fainter B component (as with the subgiant κ For that follows hereafter), the Hα line core of the subgiant primary χ Eri A is filled in at the 4% level and the Hipparcos H_p magnitude varies at ΔH_p = 0.04 mag. Compared to similar, but inactive, subgiants such as β Aql or HD 192699, the most plausible explanation for the photometric variability of χ Eri could be an unknown tertiary component or a former mass transfer scenario, or both.

κ For. This is nearly a twin of the closest, old thin-disk G-type subgiant β Hyi (cf. Table 1), with, however, a secondary that bears some surprises. The first clear-cut evidence for a companion around κ For arose from the Hipparcos data (cf. Wielen et al. 1999). At the same time, the remarkable ROSAT X-ray luminosity L_x = 394.4 × 10²⁷ erg s⁻¹ (Hünsch et al. 1999) also meant that an additional, tertiary component must be lurking in this system. Gontcharov & Kiyaeva (2002) first suggested the presence of a massive M = 1.1 M_☉ secondary (which they considered to be a white dwarf) from the astrometric orbit, and the precision radial velocities of Endl et al. (2002) also made clear that the primary could not be in charge of a close Ab component.

In fact, a direct comparison with the nearly coeval β Hyi at τ ≃ 6.5 Gyr, which possesses a negligible L_x = 1.4 × 10²⁷ erg s⁻¹, implies that essentially all of the observed X-rays of κ For must come from the similarly old secondary, that is, they can only originate from a short-period Ba–Bb subsystem. While the period of the outer A–B system of κ For should be around 21.5 yr (Gontcharov et al. 2000; Endl et al. 2002; Abt & Willmarth 2006), the secondary was not discovered until the year 2005 when Lafrenière et al. (2007), using near-infrared adaptive optics imaging, found this very component at ρ = 0.469 arcsec. Later, it was also confirmed at optical wavelengths (Tokovinin et al. 2010) with a photometric difference of about 5.0 mag and 4.3 mag in the y and Hα filters, respectively. Most recently, Tokovinin et al. (2012) concluded from a combination of the astrometric, spectroscopic, and visual orbits that the massive secondary should most likely be a close pair of M dwarfs.

For the spectroscopic modeling, we must assume that both the Ba and Bb components will not be as Doppler shifted as the radial velocity curve of Abt & Willmarth (2006, their Figure 4) would imply were this a single object. That is, we do not know "where" these components actually reside on the spectra, but on account of the 5.0 mag combined photometric contrast, the effect on the effective temperature from the Balmer line wings (which the M dwarfs tend to fill in) may only be at the 10–20 K level and the possible impact on the iron abundance may likewise only amount to Δ[Fe/H] ≃ +0.01 dex.

ι Hor. With a space velocity U/V/W = −21/−12/−2 km s⁻¹, Montes et al. (2001) have already found that ι Hor is a Hyades stream member, as we can also confirm from Figure 24 in Fuhrmann (2008). More recent measurements of the stellar acoustic oscillations by Vauclair et al. (2008) even led these authors to suggest that ι Hor shares the characteristics of the Hyades open cluster, in particular, its young age of about 600 Myr. With our Balmer-line-based effective temperature, T_eff = 6050 K, this age is, however, somewhat difficult to reconcile and instead points to an object that could be as old as τ ≃ 1.5 Gyr, suggesting a stream membership only.

As Vauclair et al. (2008) mention, their calculations are all slightly too luminous for the given Hipparcos parallax. Thus, one of their favored models with T_eff = 6189 K and M = 1.25 M_☉ would lead to R = 1.12 M_☉ and, hence, log g = 4.44. The surface gravity log g = 4.40 ± 0.01 as given by Vauclair et al. (2008) is evidently not consistent with this value and requires either a reduced mass M = 1.14 M_☉ or an effective temperature as low as T_eff = 6070 K (or some reduced combination of both these values). Certainly, a direct precision measurement of the angular diameter of ι Hor would be very helpful. In this context, we also refer to a more recent work by Metcalfe et al. (2010), who discovered a 1.6 yr magnetic activity cycle for ι Hor and who mentioned the possibility of p-mode frequency shifts, in particular, as the asteroseismic observations of ι Hor by Vauclair et al. (2008) took place during its magnetic minimum phase.

τ¹ Eri. While the first P = 958 day orbit of Abt & Levy (1976) was later disputed by Morbey & Griffin (1987), the issue found a continuation with the Hipparcos astrometry (Makarov & Kaplan 2005; Frankowski et al. 2007), and, in particular, the error in the revised Hipparcos parallax π = 70.32 ± 1.83 mas (van Leeuwen 2007) is now almost certain evidence for an astrometric binary.

With respect to our spectroscopy, there appears to be a slow drift in the radial velocity of τ¹ Eri with v_r = +28.1 km s⁻¹ in 2011 August and v_r = +30.2 km s⁻¹ in 2012 August. On account of the considerable projected rotational velocity, vsin i = 25 km s⁻¹, these radial velocities may, however, be uncertain at the 1 km s⁻¹ level. Since we find essentially no variability in the cross-correlation function, we conjecture that τ¹ Eri is then likely to be a single-lined binary. On account of the given rotation rate, we also note that the iron abundance given in Table 1 is based on only five Fe ii lines.

γ Lep. This F6V/K2V common-proper-motion pair has repeatedly been the subject of debate as to whether it belongs to the young Ursa Major Association (e.g., Roman 1949; Eggen 1960, 1992; Soderblom & Mayor 1993; King et al. 2003). At least two basic observations argue against this: first, the comparatively low X-ray luminosity L_x = 18.3 × 10²⁷ erg s⁻¹ (Hünsch et al. 1999), and second, the 18% Hα core intensity that we find for γ Lep A, which is typical for inactive field stars of this kind. The stellar parameters of γ Lep A in Table 1, which imply an age of τ ≃ 3 Gyr, concur with this assessment.

HR 2998. The radial velocity curve of Setiawan et al. (2004) for this P ≃ 380 day astrometric, visual, and spectroscopic binary implies that our single epoch spectrum is not suited to resolving the secondary. While Setiawan et al. (2004) provide a minimum mass of M_B ⩾ 0.33 M_☉ from the spectroscopic orbit, two astrometric solutions based on Hipparcos data by Goldin & Makarov (2006) lead to secondary masses of 0.49 M_☉ and 0.61 M_☉, each, however, with considerable uncertainties. Since the more massive of these two cases would result in a Δm_V ∼ 5.5 mag fainter secondary, one cannot at present completely exclude the weak impact of that component on the basic stellar parameters for the subgiant primary HR 2998 A (see Table 1).

HR 3018. This is a late-F, metal-poor, and low-mass Population II star. HR 3018, has a known common-proper-motion cool white dwarf companion (Kunkel et al. 1984; Sion et al. 2009) and the primary is itself an astrometric binary (Wielen et al. 1999; Makarov & Kaplan 2005; Frankowski et al. 2007). Most recently, Hartkopf et al. (2012) identified the Ab component as a red dwarf at ρ = 0.32 arcsec and ΔI = 3.0 mag.⁴ For the visual photometric contrast this approximately corresponds to ΔV ≃ 3.7 mag, with an uncertainty of about 0.3 mag. Thus, at V_Aa = 5.378 and V_Ab ≃ 9.1 the secondary must have a non-negligible effect on our spectra. In fact, the effect is directly discernible from the Hα core intensity, which is slightly filled in at the 2% level. Neglecting the Ab companion in a first approximation, our model atmosphere analysis results in T_eff/log g/[Fe/H]/[Fe/Mg] = 5821/4.21/−0.82/−0.44. However, the secondary, which we estimate to be a late-K dwarf at T_{eff, Ab} ∼ 4230 K and M_Ab ∼ 0.56 M_☉, fills in the Balmer line wings of its Aa primary. Thus, a composite analysis of the Balmer lines leads to an increase in the effective temperature of ΔT_eff(Hα)∼40 K and ΔT_eff(Hβ)∼30 K. The combined synthetic modeling also entails a slight increase in the surface gravity log g_Aa = 4.24 and iron abundance [Fe/H] = −0.79 and we will refer to these revised parameters in Table 1.

Note that as a Population II star, we will not consider HR 3018 Aa in the final section on the thin-disk multiplicity census, but also note that at M_Aa ≃ 0.86 M_☉ it would fall short of the relevant mass range anyway.

HR 3079. Kinematically, one may follow Eggen (1982) and consider this F5V/K3V visual binary a Hyades stream member. Our model atmosphere analysis of the primary, however, results in a slightly metal-poor star. At T_{eff, A} = 6440 K and log g_A = 4.32 and on account of a projected rotational velocity of vsin i_A = 43 km s⁻¹ we have only a few Fe i lines at our disposal and these lead to [Fe i/H] = −0.28. In this part of the H–R diagram, however, this value requires an upward correction of about Δ[Fe/H] ≃ +0.15 dex, as previously discussed in, e.g., Steffen (1985) and Fuhrmann et al. (1997) with the F5 standard Procyon. Hence, we adopt an iron abundance [Fe/H] ≃ −0.10 and this, in turn, leads to a stellar mass M_A ≃ 1.23 M_☉. We have no spectrum of the K-type secondary, but with the Hipparcos photometry and the given metallicity we expect an effective temperature T_{eff, B} ≃ 4640 K and a secondary mass M_B ≃ 0.74 M_☉.

At present, we consider a Hyades stream membership of HR 3079 as inconclusive. The issue may be resolved if the primary is itself a binary, as this may affect our effective temperature. In this respect, a dedicated observation of the secondary and an independent determination of its metal abundance would be very valuable.

α Cha. The Balmer line wings lead to an effective temperature T_eff = 6510 K and the Hipparcos parallax to a surface gravity log g = 3.95. The star has a considerable projected rotational velocity, vsin i = 31 km s⁻¹, and the few available Fe ii lines result in [Fe/H] = −0.09. Thus, except for the higher rotation rate, α Cha possesses stellar parameters very similar to the bright standard Procyon (see Table 1). Nevertheless, and unlike Procyon, there may be an oddity with the pressure-dependent Mg Ib triplet lines for α Cha. As we have demonstrated (see Fuhrmann et al. 2011a, 2011b) these lines can imply secondary components that are otherwise masked by stellar rotation, provided that there is independent information on the intrinsic stellar luminosity from, e.g., an accurate parallax. In the case of α Cha, a Δlog g ⩾ +0.15 dex higher surface gravity appears appropriate with reference to the Mg Ib lines. The effect depends, however, on the less well-constrained magnesium abundance, as the relevant weak Mg i absorption lines are all blended. An important concern is, therefore, the very accurate van Leeuwen (2007) Hipparcos parallax, π = 51.12 ± 0.12 mas, which is difficult to reconcile with the required luminous companion, unless it is a stellar twin. In this case, one would hope to see some profile variability in the cross-correlation function, yet our four spectra of α Cha offer no support in that respect and the putative companion remains only a possibility.

δ Vel. This star was already briefly mentioned in the Introduction and will be further discussed in the context of the stellar multiplicity in Section 4. Here we shall focus only on the solar-type δ Vel B with some general (theoretical) considerations.

This component possesses a fairly accurate Hipparcos photometry that leads to an absolute magnitude M_{V, B} = 3.50  ±  0.07, which is typical for an F-type star. Provided that this is a single object, one would need to know its metal abundance for a better characterization. In this respect, and in spite of the current uncertainties with a precise systemic velocity in addition to a lack of a parallax and proper motion that account for the P ≃ 142 yr visual orbit, it appears, from its space velocity, that δ Vel belongs to the Ursa Major Association. If so, we may assume [Fe/H] ≃ −0.06 to a first approximation (see Fuhrmann 2004), and since we know from the system primary that δ Vel is fairly young, the B component must be an unevolved main-sequence star. As such, and again, provided that we are dealing with a single object, one can infer the following stellar parameters: T_{eff, B} ≃ 6670 K, log g_B ≃ 4.31, R_B ≃ 1.33 R_☉, and a mass M_B ≃ 1.33 M_☉.

However, currently being separated only 0.5 arcsec from its 3.5 mag brighter primary, the fifth-magnitude star δ Vel B has never been observed spectroscopically. It could be a binary on its own, in which case many later, F- or G-type combinations are feasible. Thus, in spite of the fact that δ Vel B definitively belongs to the volume-complete sample, we can currently only speculate on its long-lived or short-lived status, its basic stellar parameters, and even worse, the number of its relevant solar-type components.

HR 3570. Very recently, Shaya & Olling (2011) proposed that this F-type star is a 15 distant, comoving companion to δ Vel (discussed just above). Kinematically, both objects appear to belong to the Ursa Major Association and δ Vel, in particular, must be young since its primary consists of a pair of A-type stars. A difficulty, however, arises with HR 3570, for which our model atmosphere analysis leads to a slightly evolved object at an age of about 2 Gyr. A companionship with δ Vel can then only be reached if HR 3570 is itself a binary, yet the observational evidence mostly argues against this: first, our few radial velocities are not significantly different from that of Nordström et al. (2004), second, there is a rather precise Hipparcos astrometry π = 41.40 ± 0.22 mas (van Leeuwen 2007), and third, the Mg Ib triplet lines also reveal no significant discrepancies with the astrometric surface gravity at log g = 4.28. The only oddity with HR 3570 occurs with its X-ray luminosity L_x = 214.3 × 10²⁷ erg s⁻¹ (Hünsch et al. 1999), which is unexpectedly high for a 2 Gyr old star, but it would be in keeping with a young Ursa Major Association member. In the end, the situation with HR 3570 remains inconclusive and requires intense Doppler monitoring as well as interferometric observations to solve the issue.

HR 3578. Like HR 3018 above, this is another late-F, metal-poor, and low-mass Population II star. There is some evidence from the radial velocities of Nordström et al. (2004), as well as from the slightly filled-in Hα core, that HR 3578 might have a faint companion. Note that HR 3578, like HR 3018, will also not be part of the final F star multiplicity statistics.

ψ Vel. For this P ≃ 34 yr F-type visual binary the high projected rotational velocity vsin i_A ∼ 160 km s⁻¹ of its primary renders a quantitative model atmosphere analysis very difficult. Moreover, the comparatively slowly rotating secondary vsin i_B ≃ 17 km s⁻¹ additionally fills in the primary lines, which in fact become very shallow. In this situation, the basic constraints for ψ Vel come from the Hipparcos photometry and astrometry, π = 54.6 ± 0.9 mas (Söderhjelm 1999), which provide M_{V, A} = 2.56 and M_{V, B} = 3.77 for the absolute visual magnitudes. The Balmer line wings then imply effective temperatures T_{eff, A} ≃ 7000 K and T_{eff, B} ≃ 6400 K, both, however, with uncertainties of at least 100 K. With a solar abundance, as approximately inferred from the central absorptions of the secondary, stellar masses M_A ≃ 1.56 M_☉ and M_B ≃ 1.26 M_☉ follow. From the mass ratio, visual orbit and the radial velocity of the blueshifted secondary we derive a systemic velocity γ = +6.9 ± 1.0 km s⁻¹.

HR 4600. We include this star in Table 1, but note that with reference to the revised van Leeuwen (2007) Hipparcos parallax, π = 39.49 ± 0.28 mas, HR 4600 has "left" the local 25 pc volume. Two aspects, nevertheless, deserve a brief mention. First, at a nominal age of only 2 or 2.5 Gyr, HR 4600 possesses remarkable kinematics, U/V/W = +63/−14/+15 km s⁻¹. This suggests some previous dynamical interaction, but could also mean that this F5V-type star is genuinely much older. Furthermore, HR 4600 has an entry in the proper motion Δμ catalog of Wielen et al. (1999) and may be an astrometric binary.

HR 4989. Only one observation of the 5.3 mag fainter visual companion exists at ρ = 2.7 arcsec in 1928. If not physical, it should easily appear in the 2MASS data, which is not the case, or, to be more concise, there is in fact a 16 arcsec distant suspicious object in this catalog, 2MASS J13141674-5906220, which is yet too faint to match the above 5.3 mag visual magnitude difference. It would be interesting to find out if either or both of these companions can be confirmed as physical.

θ Cen. This is one of the four giants of the southern star sample. For these kind of objects the Balmer line wings are generally very narrow and the effective temperature determination largely relies on the Hipparcos parallax and the T_eff/log g pair that follows from the iron ionization equilibrium. When applied to θ Cen this leads to T_eff/log g/[Fe/H] = 4850/2.74/−0.03 and a young 2 M_☉ star at τ ≃ 1 Gyr. This age is, however, difficult to reconcile with the fairly high space velocity U/V/W = −17/−46/−18 km s⁻¹ of θ Cen.

This is illustrated in the Toomre diagram in Figure 1, which references all nearby A- and F-type stars (N = 40) with masses above 1.7 M_☉, that is, young stars mostly at or below 1 Gyr. θ Cen, shown with a red circle, is the only star here with a peculiar space velocity above v_pec = 50 km s⁻¹ and it also possesses the lowest V component. The most plausible explanation for this is that θ Cen is not a young massive object ascending the red giant branch, but at M_V = 0.79 is a red clump star in its helium-core-burning stage of evolution. Further observational support for this is given by the low ¹²C/¹³C = 10 isotope ratio measured by Tomkin et al. (1976). Therefore, for an independent value of the effective temperature we make use of the interferometrically determined angular diameter θ_LD = 5.556 ± 0.005 mas by Richichi et al. (2009). This translates to T_eff = 4766 K and from the iron ionization equilibrium to a lower surface gravity log g = 2.48. In combination with the van Leeuwen (2007) Hipparcos parallax π = 55.45 ± 0.20 mas (which fixes the radius), the mass is then also considerably reduced to M ≃ 1.27 M_☉. Yet, and as for most giants, this value remains fairly uncertain: a small change to log g = 2.43 would, for instance, result in 1.13 M_☉.

Zoom In Zoom Out Reset image size

HR 5356. This is nearly a twin of the metal-poor thin-disk F star γ Pav, discussed further below. But while the latter, fourth-magnitude object has received much attention in the literature, there is, for instance, no entry in the PASTEL catalog (Soubiran et al. 2010) for HR 5356 A. From its basic stellar parameters listed in Table 1, its evolutionary tracks imply that the τ ≃ 7 Gyr old HR 5356 A originally arrived on the main sequence at T_eff ∼ 5850 K with plenty of time to slow down to its present vsin i = 1.0 km s⁻¹ and is now approaching its turnoff position.

Unlike γ Pav, HR 5356 is also known as a visual binary with only a few existing measurements, the most recent from the year 1936. Actually, Raghavan et al. (2010) and Tokovinin et al. (2010) could not recover the secondary, although this is not unexpected for the latter, speckle-based work, given that HR 5356 B is supposed to be more than seven magnitudes fainter. Thus, while there is also independent evidence for this companion from the Wielen et al. (1999) Δμ catalog, as well as from the parallax error of the improved van Leeuwen (2007) Hipparcos astrometry, a current direct observation of HR 5356 B would be important.

HR 5825. The considerable projected rotational velocity of this edge-on (Kalas et al. 2006) F5V star that we measure to vsin i = 76 ± 2 km s⁻¹ renders an accurate model atmosphere analysis rather difficult. Thus, we find only a handful of suitably unblended Fe i lines that lead to [Fe i/H] ≃ −0.13 ± 0.10. The effective temperature T_eff = 6540 K from the Balmer line wings then implies (cf. Steffen 1985; Fuhrmann et al. 1997) an upward correction Δ[Fe/H] ≃ +0.15 dex, i.e., toward an essentially solar iron abundance for HR 5825. From the space velocity, U/V/W = −5/−15/−3 km s⁻¹, we concur with López-Santiago et al. (2006) that HR 5825 is likely a member of the young (few 10⁸ yr) Hercules-Lyra Association (Fuhrmann 2004). This star should then also display an effective temperature of about T_eff ∼ 6700 K, whereas the above derived value T_eff = 6540 K and the Hipparcos-based surface gravity log g ≃ 4.25 lead to a slightly evolved star at τ ≃ 1.5 Gyr with a mass M ≃ 1.33 M_☉. While an as yet undetected secondary might be the cause of this discrepancy, existing radial velocity measurements remain inconclusive in this respect.

HR 6094. This is a probable member of the Ursa Major Association, although, as already pointed out by Soderblom et al. (1993), its lithium line λ6707 is comparatively weak. An explanation for this spectral peculiarity may be found with the accompanying overabundances of s-process elements on HR 6094, as discussed by Porto de Mello & da Silva (1997) and which these authors relate to wind accretion from its hot DA white dwarf common-proper-motion companion WD 1620-391.

This 11th magnitude and only 13 pc distant degenerate had gone unnoticed until 1968 (see Stephenson et al. 1968; Hiltner et al. 1968; Alexander & Lourens 1969), mostly on account of its low space velocity. We note here in passing that WD 1620-391 is currently the most distant directly observable thin-disk degenerate secondary of our nearby star sample.

ζ TrA. For this young, short-period binary of the Ursa Major Association, Skuljan et al. (2004) recently derived an improved spectroscopic orbit of P = 12.976 days. One year later, Jancart et al. (2005) published a tiny α = 2.71 ± 0.44 mas photocentric orbit from the Hipparcos intermediate astrometric data that discloses ζ TrA at a fairly low orbital inclination i = 1601 ± 247. When combined with our mass estimate for the primary M_Aa ≃ 1.08 M_☉ (cf. Table 1) and the Skuljan et al. (2004) spectroscopic orbit, a secondary mass M_Ab = 0.39^+0.09_−0.06 M_☉ follows, wherein we assume a 10% error for M_Aa. For a dwarf star, ζ TrA Ab would necessarily be too faint to affect our spectroscopy. Yet, one must assume that the secondary is a pre-main-sequence object and hence somewhat brighter than its mass would imply. Our three BESO spectra of ζ TrA, secured at very different orbital phases do, however, support the case of a single-lined binary.

ι Pav. There is no conclusive evidence for a companion of this astrometric binary from our spectroscopy.

γ CrA. The orbital elements for this equal-mass F-type binary with P ≃ 122 yr are given in Heintz (1986). At our single epoch observation the components were unresolved at ρ ≃ 1.3 arcsec and from the visual orbit we calculate a relative Doppler velocity Δv_r(A–B) = +4.5 km s⁻¹, meaning that our high-resolution spectrum is unresolved as well. However, on account of the fairly equal luminosity of both components, ΔH_p = 0.07 ± 0.03 mag, one can reasonably assume that the spectroscopic observations are photometrically unbiased. Thus, with the constraint of coeval objects and a slightly less evolved 10 K hotter secondary, the effective temperatures from the Balmer line wings result in T_{eff, A} ≃ 6090 K and T_{eff, B} ≃ 6100 K. The Hipparcos parallax then leads to absolute visual magnitudes, surface gravities, and stellar radii, M_{V, A} ≃ 3.73 and M_{V, B} ≃ 3.80, log g_A ≃ 4.17 and log g_B ≃ 4.19, R_A ≃ 1.47 R_☉ and R_B ≃ 1.42 R_☉, respectively. On the supposition of similar projected rotational velocities, vsin i_A ≃ vsin i_B, and assuming ξ_t ≃ 1.30 km s⁻¹ and ζ_RT = 5.20 km s⁻¹ for the micro- and macroturbulence velocities we find from the combined modeling of Fe ii lines, such as λ6432.684 in Figure 2, a slightly subsolar iron abundance [Fe/H] ≃ −0.07. The iron-to-magnesium abundance ratio [Fe/Mg] ≃ −0.04 then identifies γ CrA as a thin-disk object in line with its τ ≃ 5 Gyr age from evolutionary tracks that also lead to the stellar masses M_A ≃ 1.15 M_☉ and M_B ≃ 1.14 M_☉. Both the considerable age and the low X-ray luminosity L_x = 7.4 × 10²⁷ erg s⁻¹ (Hünsch et al. 1999) of γ CrA are inconsistent with the formerly assessed Hyades group membership by Eggen (1998).

Zoom In Zoom Out Reset image size

ω Sgr. This fourth-magnitude star is a single-lined spectroscopic binary with a velocity amplitude of no less than 20 km s⁻¹, but with as yet unknown orbital elements. The τ ≃ 2.5 Gyr old primary is a subgiant whose Hipparcos parallax, π = 42.03  ±  0.94 mas, was recently revised by van Leeuwen (2007) to π = 38.48 ± 2.66 mas. While a weighted mean would keep this star system within 25 pc, we find that the effective temperature from the Balmer line wings of ω Sgr A and its coupling through the iron ionization-equilibrium-based surface gravity favors the van Leeuwen (2007) parallax, in spite of its larger formal error.

Thus, although we do not consider this star a sample member, it would be desirable to improve on its parallax and to directly identify the secondary with high-contrast imaging observations. The latter appears feasible as the radial velocity monitoring by Murdoch et al. (1993), for instance, implies an orbital period of no less than three years.

HR 7631. The nature of this star in terms of its youth and kinematics was already realized by Jeffries & Jewell (1993). Compared to HR 5825, discussed above in this section, HR 7631 appears to be its slightly less massive sibling. Both stars are very likely members of the young Hercules-Lyra Association and HR 7631 also displays a projected rotational velocity very similar to HR 5825. At an effective temperature, which we provisionally assess as T_eff ∼ 6300 K, HR 7631 thereby becomes a remarkable object among the nearby stars: except for a few, very short period binaries in bound rotation, we know of no other late-F main-sequence star with a comparable rotational velocity. This is also manifest in the X-ray luminosity L_x = 799.9 × 10²⁷ erg s⁻¹ (Hünsch et al. 1999), but we also note an unusually strong filling-in of the Balmer line cores, which may even affect their line wings and, hence, our effective temperature determination. Assuming a solar metal enrichment (approximately in line with the few available iron lines), we adopt a provisional mass of M ∼ 1.22 M_☉ for HR 7631, but we also aim to perform a spectroscopic follow-up of this interesting object, as the literature suggests some radial velocity variability.

ψ Cap. This star is already listed among known or suspected radial velocity variable stars in the early work of Lunt (1919). As it turns out, ψ Cap is also similar to α Cha discussed above. We measure the same effective temperature, T_eff = 6510 K, from the Balmer line wings and a surface gravity log g = 4.21 from the Hipparcos parallax. At vsin i = 41 km s⁻¹ ψ Cap rotates somewhat faster than α Cha, and hence, all Fe ii lines become very difficult to measure. From the more numerous Fe i lines we get [Fe i/H] = −0.05, which in this part of the H-R diagram implies that ψ Cap should be enriched to [Fe/H] ≃ +0.10.

Similar to α Cha, it appears that the Mg Ib triplet lines of ψ Cap would better suit a higher surface gravity value. At log g = 4.21 and M_V = 3.30 the star is only about 0.4 mag above its main sequence at log g ≃ 4.3. A near-equal-luminosity Aa–Ab pair is therefore not possible and the most plausible combinations would instead be met at Δm_V ≃ 1.0–1.5 mag, with a T_{eff, Aa} ≃ 6600 K primary and a late F- or early G-type secondary. In fact, the Hipparcos parallax, π = 68.13 ± 0.27 mas, has a much larger uncertainty compared to α Cha and we also note small changes in the cross-correlation function on 11 spectra secured in the years 2010–2012. While this is only weak evidence for a secondary, it remains a possibility for future investigations.

γ Pav. Among the local thin-disk stars, γ Pav and HR 5356 (see above) are clearly the most metal-poor specimens. As with HR 5356 A, the fairly hot effective temperature of γ Pav allows us to constrain its age to τ ≃ 7 Gyr, in line with the recent detailed analysis by Deheuvels et al. (2010). At this age γ Pav and HR 5356 are among the oldest members of their population.

ν Oct. For this P ≃ 2.9 yr astrometric and spectroscopic binary Ramm et al. (2009) presented precision orbital elements and a revised Hipparcos parallax, π = 45.25 ± 0.25 mas. With our estimate for the primary mass M_A ≃ 1.61 M_☉ this leads to a secondary mass M_B ≃ 0.58 M_☉. Whether ν Oct B is then a seven magnitude fainter M dwarf, or instead, an even fainter white dwarf, is therefore not critical for our spectroscopic analysis of the evolved K-type primary.

HR 8531. As an 8 Gyr old subgiant of exactly solar metallicity, HR 8531 A—like 70 Vir (Bernkopf et al. 2001) and HR 7845 A (Fuhrmann 2008)—is a real key star for a differential and precise age dating of the local Population I. It is therefore of particular interest to constrain the characteristics of its unseen secondary (Wielen et al. 1999; Nordström et al. 2004; Makarov & Kaplan 2005; Frankowski et al. 2007; Tokovinin et al. 2010, 2012), as this will deliver a very precise dynamical subgiant mass.

We also note in passing that a comparison of the 2MASS and WISE all-sky catalogs implies that the 24 arcsec distant 2MASS J22245358-5748000 should be a common-proper-motion companion.

The majority of the stars presented in the previous section are of spectral type F and we shall continue hereafter with a few more special cases of this spectral class. Along with our previous work on the nearby stars (Fuhrmann 2011), we are thereby approaching a situation where we can already refer to a mostly volume-complete census of all F-type stars within 25 pc. This justifies a more general discussion of their multiplicities that cover the approximate mass range from 1.10 M_☉ to 1.70 M_☉.

The major problem with the F star multiplicities is certainly the transition to considerable rotational velocities that occurs at effective temperatures of about 6200 K and that in essence leads to the spectroscopic cocoon for as yet hidden companions. As exemplified in the Introduction, there is, however, a steady progress of observational techniques to overcome these shortcomings and to uncover their real multiplicity levels. Toward the higher mass A-type stars the situation is even worse as essentially all of them are fast rotators. We shall include these in our census, but since there are only a few dozen A stars within 25 pc, a solid discussion of their multiplicities is beyond the scope of the present work. Nevertheless, we commence this section with the bright and fast-rotating A star δ Vel (Argyle et al. 2002; Kellerer et al. 2007), as this represents an excellent textbook example of the observational hurdles on the stellar multiplicities.

Early observations of δ Vel in the 19th century had already found a 3 arcsec distant secondary as well as a common-proper-motion visual binary displaced at 1 arcmin. δ Vel was then known as a quadruple system for almost a hundred years when Tango et al. (1979) found another ρ = 0.63 arcsec close companion next to δ Vel A from intensity and speckle interferometry. The real surprise came, however, when Sebastian Otero—by visual observation—noticed a 0.3 mag drop of the brightness of δ Vel on 1997 July 1, i.e., the discovery of an eclipsing binary (Otero et al. 2000). As we know today, what he had found was actually another second-magnitude nearby A-type star hidden in the light of its primary and at that point δ Vel apparently consisted of no less than six components.

From the Hipparcos astrometry Argyle et al. (2002), however, had also noticed that δ Vel B was in a phase of rapidly closing-in, meaning that the Tango et al. (1979) speckle observations must have been concerned with this very component, whereas their former intensity interferometry most likely had dealt with the inner Aa–Ab eclipsing subsystem. Even worse, Kellerer et al. (2007) showed the distant δ Vel C–D visual binary to be physically unassociated. Within only a few years δ Vel was then degraded to a triple system.

In the current situation of the eccentric P ≃ 142 yr visual orbit δ Vel A–B has now become even closer than ρ ≃ 0.5 arcsec (see Kervella et al. 2009) and will reach a minimum at ρ ≃ 0.38 arcsec in 2013. In their detailed analysis of δ Vel, Pribulla et al. (2011) thus mention the unfortunate situation with the 3.5 mag fainter B component which, lacking any spectroscopic information, could be an ordinary F-type star or a later-type binary on its own. In the latter case, δ Vel could again become the former quadruple system, albeit of course with partly different participants. But even if not, Shaya & Olling (2011) most recently found the 15 distant, fifth-magnitude F-type star HR 3570 to be a very wide common-proper-motion companion to δ Vel. Thus, including this candidate, and depending on whether this is eventually a binary itself (see the discussion in the previous section), one can currently choose whether the multiplicity level of δ Vel is 3, 4, 5, or even 6.

Certainly, the most important point in terms of the involved mass budget with this concrete example is whether one would know of the 2.3 M_☉ δ Vel Ab component without its eclipsing property. The orbital period of the inner Aa–Ab δ Vel pair is remarkably long at P = 45.15 days and a tiny change of the inclination would have prevented this kind of observation. One should note that Grenier et al. (1999), from inspection of the spectroscopic cross-correlation function, had correctly suggested that δ Vel A could probably be a binary. But except for this single assessment, it appears that we are just lucky with our line of sight toward this system.

4.1. Multiplicities in the Presence of Stellar Rotation

Unless one can directly refer to long-baseline optical interferometry, the basic means to uncover short-period, non-eclipsing binaries is the spectroscopic Doppler response of the orbital motion. However, as soon as the components of a near-equal-luminosity binary system possess a considerable rotation, a direct spectroscopic resolution can easily become impossible. As we have demonstrated with the F stars α Com and λ Ara (Fuhrmann et al. 2011a), the issue can already occur at moderate projected rotational velocities of vsin i ≃ 15 km s⁻¹, or less, and then requires a comparison of appropriate luminosity tracers, such as the lines of the Mg Ib triplet, with the stellar parallax.

This situation becomes increasingly difficult toward higher mass, early-F stars with their intrinsically higher rotation rates, accompanied by a loss of sensitivity of the spectroscopic surface gravity tracers. Even more, we arrive here at the domain of the γ Dor-type stars, a class of objects considered to present non-radial, g-mode pulsations that are believed to originate from instabilities at the base of the convection zone and that reveal photometric periods between 0.3 and 3 days. In the spectra the non-radial pulsations produce variable, skewed line profiles (cf. Osaki 1971; Smith & McCall 1978; Aerts & Waelkens 1993; Krisciunas et al. 1995; Balona et al. 1996; De Cat et al. 2006) that often correlate with the observed photometry.

Many of the classified γ Dor stars are, however, also known as spectroscopic binaries (e.g., Mathias et al. 2004; Henry et al. 2007) and the question arises, whether the pulsations are actually tidally driven. For close and/or eccentric systems with periods of only a few days this may lead to considerable irregular oscillations or sinusoidal resonances that can affect the radial velocity curves and be visible in the photometry (Willems & Aerts 2002). A remarkable example with respect to the latter is the A-type binary HD 187091, a highly eccentric e = 0.83, P=41.8 day system, whose precision Kepler light curve was recently presented by Welsh et al. (2011). In the end, all γ Dor stars may be similarly structured systems and some of them only misclassified pulsators. De Cat et al. (2006, their Figure 10), for instance, mention in their spectroscopic time-series search for γ Dor variables the F3V star HD 35416, whose cross-correlation profiles reveal in only 1 out of 11 cases the unequivocal signature of a double-lined spectroscopic binary. Hence, the authors comment "if we did not have the first observation of HD 35416, we would have interpreted the correlation profile variations as being due to pulsation instead of binarity" (De Cat et al. 2006, p. 285).

To better assess what kind of spectroscopic absorption line profiles one may expect to find from purely orbital considerations, leaving non-radial pulsations aside, we proceed here with the construction of synthetic, near-equal or equal-luminosity F-type binaries, both as a function of their equatorial velocities and their phase-dependent Doppler shifts. These are displayed in Figures 3 and 4 for four different cases. In the left-hand panel of Figure 3, we show synthetic composite spectra of the Fe i line λ5393.174 for a main-sequence F-type binary at effective temperatures T_{eff, Aa} = 6900 K and T_{eff, Ab} = 6300 K, and stellar masses M_Aa = 1.40 M_☉ and M_Ab = 1.20 M_☉. The primary possesses a projected rotational velocity vsin i_Aa = 40 km s⁻¹, whereas the secondary rotates at only vsin i_Ab = 30 km s⁻¹. The sequence from top to bottom denotes different phases of the orbital motion in steps of Δv_{r, Aa} = 4 km s⁻¹ for the primary and corresponding values for the secondary. Evidently, as long as the relative Doppler velocity v_{r, Ab} − v_{r, Aa} remains below 20 km s⁻¹ it is difficult to identify the profiles to be of composite nature, given that real data always possess a certain noise level and in most cases weak, contributing absorption lines. Profile asymmetries and bumps then evolve in subpanels (d) to (f) toward larger Doppler displacements.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The same sequence, but with somewhat higher rotational velocities, vsin i_Aa = 50 km s⁻¹ and vsin i_Ab = 40 km s⁻¹, is then displayed in the right-hand panel of Figure 3. Here it becomes even more difficult to uncover the double-lined origin at least down to panel (d). Note also in both cases of Figure 3 that the Doppler velocities—as measured from the central line absorptions—reveal only small or insignificant changes. In the defining paper on the γ Dor variables, Kaye et al. (1999) had pointed out that "spectroscopic variations are also observed, and these manifest themselves both as low-amplitude radial velocity variations that cannot be attributed to duplicity effects" (Poss & Kremser 1967, p. 316). Evidently, the synthetic spectra in Figure 3 prove the opposite.

In the left-hand panel of Figure 4, we proceed with an example for a twin binary system with both components rotating at vsin i = 50 km s⁻¹, whereas the right-hand panel displays the 6900/6300 K pair from Figure 3, but this time with a larger contrast in the rotational velocities. Again, the Doppler velocities, derived from the central absorptions, reveal zero (left) or rather small (right) changes, and it is mostly the overall line broadening that varies. Real observations at different epochs may thus result in different pseudo-rotational velocities for a particular star. The vsin i = 50 km s⁻¹ twin in subpanel (d) of Figure 4, for instance, is approximately equivalent to a single object at vsin i = 55 km s⁻¹. Clearly, less Doppler-displaced, near-equal or equal-luminosity systems may then imaginably stay unnoticed and the observer may always interpret a binary as a single object.

If, however, asymmetries in the line profiles of a binary system become observable and the components even partly resolved, it is important to understand that the deepest absorption does not necessarily follow the primary. Thus, in the right-hand panel of Figure 4 the deepest absorption traces the secondary on account of its lower rotation rate, which renders a correct Doppler interpretation difficult.

A very slow rotation of a secondary is of course also conceivable and secondaries can likewise be rather faint objects. In both cases, it is then comparatively easy to follow the orbital motion of the primary. The real difficult systems to resolve and interpret remain cases like those displayed in Figures 3 and 4 and we shall proceed now with concrete examples of this kind.

Within the 25 pc solar neighborhood, it appears that there are only two known γ Dor stars and both are met at very southern declinations. The slightly brighter is the eponymous object itself, whereas its local sibling is the early-F star HR 2740 (Poretti et al. 1997). We come back to these two stars further below, but commence this investigation with another nearby early-F star, α Crv.

Two spectroscopic observations of α Crv were initially secured in 2010 April and are both set out in Figure 5. The 2010 April 3 spectrum shows very asymmetric line profiles that could be interpreted to result from a Δv_r ≃ +25 km s⁻¹ Doppler-displaced secondary. The second epoch observation, as of 2010 April 15, shows the putative secondary completely masked. These two very different spectra warranted further observations and these were taken in 2012 March and April. While these data were initially suggestive of a 10 day spectroscopic binary, it was then realized that α Crv changed its overall spectroscopic appearance within about 3 days and the star was then more intensely monitored. A corresponding sequence of these observations is given in Figure 6 and clearly demonstrates the considerable spectroscopic changes that occur on timescales of hours. From its space velocity α Crv is likely related to the Ursa Major Association, that is, rather young stars at ages of several 10⁸ Myr. For a short 3 day period and putatively double-lined spectroscopic binary one must then expect that both components are in bound rotation. With a provisional synthetic modeling for a pair at T_{eff, Aa} ≃ 6900 K and T_{eff, Ab} ≃ 6100 K, projected rotational velocities vsin i_Aa ≃ 17 km s⁻¹ and vsin i_Ab ≃ 12 km s⁻¹, and assuming rotational and orbital coplanarity, one should see α Crv at an intermediate orbital inclination i ∼ 55^o. However, the semi-amplitudes of the spectroscopic orbit must then exceed 70 km s⁻¹ and α Crv should show up as a well-resolved spectroscopic binary. This is not the case in Figure 6, and hence, the most plausible explanation of what we observe are instead non-radial, gravity-mode pulsations of the γ Dor-type.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Yet, the question remains what causes these stellar pulsations and whether one may eventually detect a driving agent in terms of a rather low-mass companion for α Crv. Since the non-radial pulsations and the projected equatorial rotation both possess certain terminal velocities, the superposition of their line profiles that we observe for α Crv might nevertheless follow a Doppler orbit and this could be traceable at the base of the cross-correlation functions. This is demonstrated in Figure 7, where the red circles in the upper panel follow the radial velocities of the cross-correlation peak at the above-mentioned cycle of about 3 days, while the yellow circles in the lower panel represent the radial velocities at the base of the cross-correlation function. From the latter, there is indeed some evidence for a companion with an orbital period of a few weeks. Although this certainly requires further confirmation on a longer timescale, we note that Duquennoy & Mayor (1991) had previously obtained a similar systemic velocity and velocity dispersion for α Crv and already assessed this star to be a spectroscopic binary.

Zoom In Zoom Out Reset image size

With the understanding that α Crv is a single-lined stellar source, we proceed with a model atmosphere analysis of this star. To this end, we select a spectrum with rather symmetric line profiles to keep the effect of the pulsations at a minimum. Yet, we consider the derived projected rotational velocity, vsin i ≃ 29 km s⁻¹, only as an approximate number and the same holds true for the other stellar parameters: T_eff ≃ 6820 K, log g ≃ 4.23, and [Fe/H] ≃ −0.09. With these values and the absolute visual magnitude M_V = 3.14, evolutionary tracks lead to a stellar mass M ≃ 1.39 M_☉.

The next star that we shall consider is HR 2740. As already mentioned in this section, this is thought to be a γ Dor-type variable (Poretti et al. 1997) and similar to α Crv we present in Figure 8 a sequence of cross-correlation functions of HR 2740 during a time span of 34 days. As one can see, there is also a considerable profile variability, although not as pronounced as with α Crv. However, there is also a basic difference with the line profiles of HR 2740 being considerably broader. Grenier et al. (1999) already conjectured that these may originate from a double-lined spectroscopic binary and Royer et al. (2002a) in particular noted that their two projected rotational velocity measurements of HR 2740 at vsin i = 45 km s⁻¹ and 56 km s⁻¹ were strikingly different.

Zoom In Zoom Out Reset image size

Similar to Figure 7 we have measured the radial velocities of HR 2740 at the peak and the base of the profile functions and show these values in Figure 9. The immediate impression is that the red and yellow circles correlate with each other and with a period of about 17 days. On closer inspection, however, this correlation is mostly restricted to the right half of this diagram, whereas in the left part the yellow circles show fairly constant values. The conclusion is then that these velocities may not be straightforward to interpret and the profiles in Figure 8 likewise appear to possess partly irregular patterns. On the other hand, if we go back to the right-hand panel of Figure 4, one can see the anti-correlation of peak and base profile velocities: the more the slowly rotating secondary proceeds to the red, the more the composite line profile retreats to the blue at its base. In other words, there is some evidence that HR 2740 may be an unresolved double-lined spectroscopic binary.

Zoom In Zoom Out Reset image size

To further investigate this possibility we consider in the spectrum of HR 2740 in Figure 10 a redshifted secondary. Synthetic composite Balmer line wings as well as suitably unblended Fe i lines lead here to effective temperatures T_{eff, Aa} ≃ 6900 K and T_{eff, Ab} ≃ 6300 K, corresponding to Δm_V ≃ 0.85 mag. The V = 4.48 mag, putatively single and slightly evolved HR 2740 then mutates to two main-sequence stars at V_Aa ≃ 4.88 mag and V_Ab ≃ 5.73 mag, and M_{V, Aa} ≃ 3.22 mag and M_{V, Ab} ≃ 4.08 mag from the constraint of the Hipparcos parallax. For the two Fe i lines λ5367.475 and λ5383.378 in Figure 10 we then arrive at Δv_r ≃ −11 km s⁻¹ for the blueshifted primary and Δv_r ≃ +31 km s⁻¹ for the redshifted secondary, and with projected rotational velocities vsin i_Aa ≃ 33 km s⁻¹ and vsin i_Ab ≃ 25 km s⁻¹. The given iron abundance, [Fe/H] = −0.33, would classify HR 2740 as a fairly metal-poor object; yet, this value is derived from two neutral iron lines, which are only trace species in these types of stars, and requires an upward correction of no less than Δ[Fe/H] ≃ +0.20 dex (see the discussion in Steffen 1985 and Fuhrmann et al. 1997 for the standard F5 star Procyon). Thus, if we adopt [Fe/H] = −0.10 for the iron abundance of this star, the evolutionary tracks in Figure 11 lead to stellar masses of M_Aa ≃ 1.38 M_☉ and M_Ab ≃ 1.18 M_☉ and a system age of about τ ≃ 1 Gyr.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The modeling in Figure 10 is, however, not the only possibility: instead of a redshifted secondary one may also consider a blueshifted secondary at Δv_r ≃ −16 km s⁻¹ and a redshifted primary at Δv_r ≃ +15 km s⁻¹, and with projected rotational velocities vsin i_Aa ≃ 46 km s⁻¹ and vsin i_Ab ≃ 35 km s⁻¹. The other parameters are fairly the same for they remain constrained from the individual magnitudes and from the Hipparcos parallax (cf. Figure 11). One may imagine that one could discriminate the different sets of rotational velocities from other spectra at different phases. However, and with reference to Figure 8, we may expect that at least the 6900 K hot primary is non-radially pulsating and there may be no unique solution from just a phase-dependent spectroscopy. Again, the time-dependent velocity distribution of the yellow circles in Figure 9 implies that a straight interpretation of the observed radial velocities for HR 2740 is eventually not possible, indeed, one cannot even state with certainty that there is a secondary at all.

This leads us to the third object, the prominent γ Dor (Cousins & Warren 1963, Balona et al. 1994a, 1994b), whose line profile and radial velocity variations are given in Figures 12 and 13, respectively. While the profile variations display a behavior rather similar to HR 2740, the red circles in Figure 13 are irregularly spread over the time span covered by our observations. The velocities from the base of the cross-correlation profiles (yellow circles) display a few outliers toward low v_r velocities. These may occur on a short, 0.1 day timescale as suggested for γ Dor in De Cat et al. (2006, their Figure 20) and would then be attributable to stellar pulsations. Orbital velocity changes on days or weeks are not obvious for γ Dor from our spectroscopic data, nor have they been reported in De Cat et al. (2006). However, and as we have shown in Figures 3 and 4, this does not mean that one can exclude those. On account of a fairly high projected rotational velocity, which in the case of γ Dor we measure to vsin i ≃ 55 km s⁻¹, and with at least one component being a non-radial pulsator any orbital motion could easily be masked.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

It is then important to realize that with reference to Figure 11 there is at least the principal possibility for a near-equal-mass binary for γ Dor. This is very similar to HR 2740, whereas α Crv is much closer to the main sequence and can hence only accommodate a rather low-mass companion. Treating γ Dor as a single object, we find approximate stellar parameters T_eff ≃ 6890 K, log g ≃ 4.11, and a solar metallicity. Thus, slightly different evolutionary tracks compared to those given in Figure 11 principally apply to γ Dor, yet they do not affect the main argument with a putative secondary.

We conclude at this point, noting that the spectroscopic observations of all three stars—α Crv, HR 2740, and γ Dor—display clear evidence for non-radial pulsations. We find some spectroscopic signatures for companions around α Crv and HR 2740, which in turn might drive these pulsations. We cannot exclude the possibility that this is also the case for γ Dor, but strictly said, we have no definite proof of any of these three stars harboring a companion.

In what follows we briefly present of few more early F-type stars that may be pulsators and/or binary candidates, but also at least one counter-example.

α Cae. From its kinematics, Eggen (1969) first related this object to the young Ursa Major Association. With an effective temperature T_{eff, A} ≃ 6950 K from the Balmer line wings, the F-type primary, however, resides 0.3 mag above its main-sequence position. A consequence could be that α Cae A, like its ρ ≃ 7 arcsec distant B component (Ehrenreich et al. 2010), is a binary on its own. Grenier et al. (1999) have already pointed to this possibility from the shape of the spectroscopic cross-correlation function and the seven observations of α Cae A that we can refer to could be in line with this conjecture.

As for most objects at this effective temperature, the considerable projected rotational velocity that—in our spectra—varies between vsin i ≃ 41 km s⁻¹ and 47 km s⁻¹ renders a direct detection of a putative Ab component difficult. The relative nearness of the primary to its main sequence, though, allows for the constraint that any plausible secondary must be Δm_V ⩾ 1.2 mag fainter and its effective temperature hence below 6100 K. Ursa Major Association stars of this kind are mostly restricted to projected rotational velocities of vsin i ⩽ 10 km s⁻¹ (Fuhrmann 2004). As a composite modeling of α Cae A implies, this component should, however, rotate at vsin i ⩾ 25 km s⁻¹ to hide in the light of its primary. One may not exclude this possibility, but at this point, a case of a single, non-radially pulsating α Cae A appears to be more plausible.

We should add that we do not detect any significant radial velocity variations at the base of the line profiles of α Cae A within a time span of one month. Given that the putative Aa and Ab components cannot have the same luminosity, much larger orbital periods should leave a trace in the parallax error of the Hipparcos data; at σ_π = 0.14 mas (van Leeuwen 2007) this is, however, among the most precise values for nearby stars. Taking, then, α Cae A as single, the Hipparcos astrometry provides a surface gravity log g = 4.20 and the investigation of the iron lines is approximately in line with a solar metal enrichment. Adopting, [Fe/H] ≃ −0.06, which we find as a mean value among other stars of the Ursa Major Association (Fuhrmann 2004), this would lead to a primary mass M_A ≃ 1.46 M_☉.

HR 4102. This star already appears in the work of Lunt (1919) among known or suspected radial velocity variable stars. Later, Evans et al. (1959) called this into question, however, and given the considerable projected rotational velocity of about 50 km s⁻¹, it appears that a rigorous Doppler monitoring of HR 4102 has never been pursued. Our one-month spectroscopic effort in that respect is set out in Figure 14 and could be interpreted in terms of a slope in the radial velocity. However, this assessment mostly depends on the first and the last observation in the lower panel of this figure and hence may not be very significant. The important point to understand is that HR 4102 shows line profile changes comparable to HR 2740 and γ Dor, which translate to projected rotational velocities in the range vsin i ≃ 46 km s⁻¹ to 56 km s⁻¹ and the data in the lower panel of Figure 14 may in part depend on the g-mode pulsations.

Zoom In Zoom Out Reset image size

If we take HR 4102 as single, the Balmer-line-based effective temperature T_eff ≃ 6760 K and the close-to-solar iron abundance, place this object at M ≃ 1.45 M_☉ and 0.5 mag above its main sequence. There is then principally sufficient room for an additional companion, yet a clear-cut decision in favor of this scenario is currently not possible from our data and may only succeed with interferometric methods.

η Cru. Many characteristics discussed with HR 4102 above are also met with η Cru. The star is slightly hotter, T_eff ≃ 6830 K, more luminous and massive at M ≃ 1.52 M_☉, has about a solar iron abundance and may be related to the Ursa Major Association. The projected rotational velocity is also similar to HR 4102, but varies on our spectra only between vsin i ≃ 50 km s⁻¹ and 54 km s⁻¹. We find no compelling evidence for a variable radial velocity during one month of observations (N = 12). Yet, we mention that η Cru lies 0.7 mag above its main sequence and could easily accommodate an F-type secondary in its spectrum. Apart from this possibility, it appears worthwhile to find out whether the ρ ∼ 16 arcsec distant 2MASS J12065422-6437033 could be a physical companion.

η Crv. This F2V-type star is classified as variable and very diverse projected rotational velocities such as vsin i = 92 km s⁻¹ (Royer et al. 2002b) and vsin i = 60 km s⁻¹ (Reiners & Schmitt 2003) have been reported in the more recent literature. Our measurements only confirm the latter work at vsin i ≃ 61 km s⁻¹. Also, the line profile functions are symmetric and do not show any variability, nor do we detect significant radial velocity changes during one month of observations (cf. Figure 15). We note that, kinematically, η Crv may be related to the Hyades moving group and its nearness to the main sequence excludes the possibility of a near-equal-mass binary. A long-term low-mass secondary also appears unlikely as this should show up in the Hipparcos astrometry. In other words, our observations give no hint for stellar pulsations or binarity for η Crv. A model atmosphere analysis leads to T_eff ≃ 6700 K from the Balmer line wings, and a super-solar Hyades-like iron abundance. If indeed single, η Crv should have a mass of M ≃ 1.47 M_☉.

Zoom In Zoom Out Reset image size

1 Cen. A comparison with η Crv reveals that 1 Cen has a similar effective temperature, T_eff ≃ 6660 K and a similar super-solar iron abundance, but is 0.2 mag more luminous and hence more massive at M ≃ 1.52 M_☉. The projected rotational velocity vsin i ≃ 65 km s⁻¹ of 1 Cen is also similar to η Crv, and from its kinematics 1 Cen is a certain Hyades moving group member. Differences occur for the X-ray luminosities L_x = 47.8 × 10²⁷ erg s⁻¹ (η Crv) versus L_x = 230.0 × 10²⁷ erg s⁻¹ (1 Cen) (Hünsch et al. 1999) and the cause for this may be found in the variable velocities for 1 Cen displayed in Figure 15.

In the literature, 1 Cen is known as a single-lined spectroscopic binary (Spencer Jones 1928) with an orbital period P = 9.9 days and a semi-amplitude K = 6.0 km s⁻¹. While we cannot confirm the latter value, the velocities in Figure 15 appear to support—if any—a short-term orbit. More importantly, however, one may not exclude that 1 Cen, being slightly more luminous than η Crv, can harbor a near-equal-mass F-type companion. If so, this could seriously affect the interpretation of the observed velocities, meaning that, for the time being, the question of a companion to 1 Cen is inconclusive.

β Tra. This second-magnitude F star possesses a fairly large projected rotational velocity vsin i ≃ 83 km s⁻¹ with varying, partly wiggly line profiles. For the basic stellar parameters we approximately find T_eff ≃ 6960 K, log g ≃ 4.02, and a solar metallicity. With these parameters, β Tra lies 0.7 mag above its main sequence with a mass of M ≃ 1.61 M_☉ and an age ∼1.4 Gyr, provided it is single. Interestingly, the kinematics of β Tra place it at the position of the young Hercules-Lyra Association (Fuhrmann 2008, Figure 24), a conflict that could immediately be solved with an equal-luminosity binary. On the basis of six rather constant radial velocities from our spectroscopy, we can of course only conjecture about this possibility.

We also note in passing that β Tra has a V = 13.2 common-proper-motion candidate, LTT 6333 = Gl 601 B, which, according to Caldwell et al. (1984), appears to be a degenerate. More recently, however, Stauffer et al. (2010) were unable to recover this object in the 2MASS catalog. It also appears difficult to assign this object a Hercules-Lyra Association membership.

4.2. F Star Multiplicity Census

The considerable interest in the stars of the solar neighborhood within recent years has unquestionably caused an enormous progress in almost every respect. Their census has become much clearer with the Hipparcos astrometric and photometric data, many spectroscopic orbits have been derived or improved, and high-angular resolution observations have led to discoveries of many close companions. In consequence, one might conjecture that there is by now a good understanding of the local inventory out to 25 pc and as far as the brightest stars of this volume are concerned.

However, the many examples discussed in this work leave no doubt that our understanding of the multiplicities among the nearby bright stars is, on the contrary, still fairly incomplete. This is in part due to the onset of stellar rotation, but also subject to a habit that only too often we content ourselves with mentioning the discovery of a new companion without further investigating whether this is eventually a multiple itself. Two good counter-examples are the low-mass pairs discovered by Ehrenreich et al. (2010) orbiting the nearby F-type stars α Cae and HR 2500 which we mentioned in the Introduction. Another famous example is the F-type multiple ζ Cnc (Griffin 2000) where the long-suspected fourth component was ultimately discovered to be a low-mass M-type binary by Hutchings et al. (2000).⁵ Yet another example could be the bright F0V multiple star ρ Gem, where the primary has often been considered a spectroscopic binary, but a convincing proof is still lacking and it may in fact be a single component. But what about its ρ ∼ 3 arcsec distant and visually 8 mag fainter secondary? We know almost nothing about it. And there is a 12 arcmin distant K2V common-proper-motion companion, BD +32 1561, that, in addition, has been "G" flagged by Hipparcos, meaning that it likely possesses a companion itself. We have this evidence, but no confirmation. If we find this supposed companion, can we ascertain that it has no further sub-companion of its own? In other words, there is a fourth-magnitude F-type star in the immediate solar neighborhood, presumed to be a triple system, but it may also consist of four, five, six, or more components. We simply do not know. Yet, the techniques for successful observations and a satisfactory answer are there, although ρ Gem is admittedly only one example in a long list.

But are the necessary and somewhat expensive observations worth all the effort? In our opinion, it would be important to know whether—eventually beginning with spectral type F—the common belief that stars form in binaries has to be replaced by an understanding that "stars form in binaries which tend to be binaries themselves," that is, as clumps of multiple systems. At the current stage, the observational evidence clearly points in that direction. Yet, we will only achieve a solid answer if not only the primaries but also the secondaries and proper motion companions receive the proper attention they deserve. Thus, the statistics that we present with this work, are not only "necessarily" incomplete, but incomplete at a level that prevents the solid conclusions one would like to draw on the stars in this important mass range between one and two solar masses.

With these concerns and restrictions in mind, we present in Figure 16 our current (2012 September) multiplicity census on the nearby stars as a function of their primary masses. The figure is a direct update of Figure 19 of the mostly northern sample in Fuhrmann (2011) and a census restricted to the Population I thin disk. Our observational strategy with this continuation on the southern stars in Figure 16 is to commence with the observation and analysis of the bright—mostly F-type—objects. In this manner, we can now refer to an essentially volume-complete, all-sky sample of some 150 objects in the mass range 1.1 M_☉ ⩽ M ⩽ 1.7 M_☉, with only some minor restrictions pending toward the lower mass bins.

Zoom In Zoom Out Reset image size

Thus, at the current stage of the project, we likely lack only four F-type stars above 1.2 M_☉, with incomplete observations and/or as yet provisional analyses (mostly based on single or less well-exposed spectra), but for the sake of completeness we include them here at least in terms of their currently assigned multiplicities and primary masses. The relevant objects, in order of their decreasing primary masses, are γ Tuc (N = 2, M ∼ 1.59 M_☉), κ Tuc (N = 4, M ∼ 1.38 M_☉), υ Aqr (N = 2, M ∼ 1.35 M_☉), and α For (N = 2+, M ∼ 1.22 M_☉), where the "+" for the latter signifies that it is likely a triple. The analyses and basic stellar parameters of these four stars will be added in a future contribution on this sample, as we progressively fill in the lower mass bins. As for stars of about 1.1 M_☉ masses, the bulk of these objects is already included in Figure 16, yet about another dozen, or so objects may ultimately contribute.

At the high-mass end at 1.7 M_☉ and above, it is clear from Figure 16 that we can only refer to a low number statistics. The last bin at 1.8 M_☉, in particular, comprises only 11 objects and just one discovery of a further companion would "dramatically" affect this bin. Thus, with our sample limit of 25 pc we cannot provide a meaningful A star multiplicity statistics and we refer here to more relevant projects such as the 75 pc Volume-complete A-Star survey of De Rosa et al. (2011). We note also that as a result of the given mass cut in Figure 16, we lose a few high-mass (evolved) F-type stars, such as α Hyi, ρ Pup, η Sco, or β Cas, which are all likely above 1.9 M_☉. On the other hand, the first A-type stars, like α Cir (Bruntt et al. 2008, 2009) or δ Cap (Malasan et al. 1989), appear already around 1.7 M_☉ in Figure 16, and of course, the given mass interval also allows for the inclusion of several evolved G- and K-type stars like β Hyi, κ For, or ν Oct (cf. Table 1). Most of the latter are former F stars and they may not much affect the multiplicity statistics by virtue of, e.g., merger descendants.

At this point we should add that a few F-type stars even fall short of the lower 1.1 M_☉ mass cut on account of their low metallicities. Thus γ Pav and HR 5356 A—both above effective temperatures of 6000 K—possess masses of only 0.95 M_☉ and 0.93 M_☉, respectively. A similar value is already met for the metal-rich K-type star α Cen B at T_eff = 5225 K, whereas the even more extreme HR 3430 B at T_eff = 5340 K and [Fe/H] = +0.32 exceeds these numbers at M ≃ 1.04 M_☉ (cf. Fuhrmann et al. 2011b). While even more metal-rich stars are rarely found in the solar neighborhood, some more metal-poor, Population II, late-F stars do exist. We already mentioned HR 3018 and HR 3578—both at M ≃ 0.86 M_☉—in the preceding section, but will not include them in the multiplicity statistics, which are only concerned with Population I stars.⁶

In the end one has to realize that on account of the various metallicities and/or evolutionary stages a number of different spectral types from A to K contribute to the given 1.1 M_☉ ⩽ M ⩽ 1.7 M_☉ mass range. Of primary interest is, of course, the stellar mass rather than the spectral type (although the mass is by no means an invariable characteristic as well). Yet, the bulk of the stars considered belong here to the spectral type F and this justifies the title of this contribution.

As the given sample size is low, and in particular so, as we need to work with subsets in terms of mass bins and multiplicity levels, we follow our previous work on the northern stars and present the data in Figure 16 in steps of 0.10 M_☉, but as a running mean of bin width 0.20 M_☉, with the number of stars per bin again indicated in the top of the figure. In addition, and for a better interpretation of the higher mass stars, we include in Figure 16 the information on the mean projected rotational velocities 〈vsin i〉 per mass bin with a light blue shading and with explicit numbers also given in the top of the figure. As with our previous work, we distinguish between single and binary stars, but subsume all higher level systems in a category dubbed "Triple+."

Similar to Figure 19 for the northern survey, the basic result of the enlarged sample in Figure 16 is a general increase toward non-single stars as a function of mass, with, however, considerable uncertainties in the highest mass bins, as indicated with the Poisson error bars. As one may expect from the blue shading, the highest mass bins will also be subject to the onset of stellar rotation, which decreases (increases) the binary (single-star) fraction. In contrast to this, the higher level systems are apparently less affected, which basically reflects the circumstance that the progress with tertiary components often arises from distant low-mass systems, as, for instance, with the triple α Cae (Ehrenreich et al. 2010) which we already mentioned in the Introduction. Thus, there is a steady increase of the multiple star fraction from about 10% for the lowest 0.90 M_☉ mass bin to more than 20% already achieved at 1.30 M_☉. At approximately this mass, and as already found for the northern sample, the crossover of preferentially single to binary stars occurs and the single-star fraction then further declines down to about a one-third percentage at the transition to the A-type stars.

However, and since we are here in the rotation-dominated mass regime, one may ask what the real multiplicity fractions of the single, binary, and higher level stars might be. A tentative answer to this question becomes apparent by referring to all candidate companions that we currently assess to be possible and that we individually discussed in the previous sections, as well as with the northern sample in Fuhrmann (2011, and references therein). The correspondingly revised distribution functions are set out in Figure 17, and bear a number of interesting prospects.

Zoom In Zoom Out Reset image size

In the first place, one can see in Figure 17 that the slope for the multiple systems is considerably steeper such that these potentially account for one-third of all objects at the transition to the A-type stars. At the same time, the crossover of the single versus binary star fractions already occurs below one solar mass and hereafter effectively diverges to a 60% (10%) level of binary (single) stars at the 1.60 M_☉ mass bin, beyond which rotation effects likely prevail. Thus, taken at face value, the immediate impression from Figure 17 is that the single-star fraction may already converge to zero at the 1.70 M_☉ mass bin, where the A-type stars set in.

But is a result like this conceivable? Since we are only dealing with candidates in Figure 17, we indeed encounter considerable uncertainties, and in particular so, the more we proceed to the higher mass stars with their enhanced rotation rates and the intrinsically smaller sample sizes. The suggested negligible single-star fraction for the A-type stars in Figure 17 is basically an extrapolation based on the slope deduced for the F-type stars. If, in this context, we consider the A-type stars of the solar neighborhood, it is well known that the brightest of these, Sirius, has a massive white dwarf companion. For others, like Alcor (Zimmerman et al. 2010; Mamajek et al. 2010) or ζ Vir (Hinkley et al. 2010), direct imaging observations have only recently uncovered additional M dwarf companions. The A stars Alderamin and Altair, on the other hand, are regarded as single. But can one be confident about this? The 7.7 pc nearby A star Fomalhaut, for instance, has a K4V companion, known as TW PsA, at an (almost) incredible angular distance of 196. And for Vega, perhaps the most-studied specimen of all nearby A-type stars: there is a flag in the Hipparcos catalog that says "suspected non-single." Actually, one may not even be sure with this extremely prominent object in terms of its multiplicity.

In fact, the situation with the local A-type stars is anything but straightforward. On account of their usually high rotation rates, one must generally appreciate that we are facing very oblate entities, causing, in part, weird effects on their observed radii, effective temperatures, luminosities, masses, convection, evolutionary state, etc. (see, e.g., Aufdenberg et al. 2006; Peterson et al. 2006; Monnier et al. 2007; Zhao et al. 2009). And, of course, the situation becomes imaginably worse if pairs of these peculiar sources contribute simultaneously.

As a simple, concrete, and memorable example, we conclude these considerations with the nearby, second-magnitude A-type star α² Lib, which is actually a multiple system, with additional, distant companions (Caballero 2010). As we can see, we meet here the odd circumstance where the primary managed to hide in the light of its secondary for more than 100 years of optical spectroscopy after its discovery as a spectroscopic binary with a large radial velocity variability by Vesto M. Slipher in 1904. In fact, Slipher (1904) mentioned in those days that the "appearance and behavior of the hydrogen line Hγ suggest that both components are bright." This was confirmed through lunar occultation observations of α² Lib several decades later in 1966 May and led Poss & Kremser (1967) to suggest that the components were separated at only 0.01 arcsec and with a blue photometric difference of 0.4 mag and an orbital period "estimated to be in the range of 20 days."

Yet, a spectroscopic confirmation of Slipher's (1904) suggestion has—to our knowledge—never been reported and the secondary was instead deemed to be "invisible." But, as stated above, the remarkable characteristic of α² Lib is that it is not the secondary, but the primary, that has escaped unequivocal spectroscopic detection. This, as one can see in Figure 18, is mostly a consequence of the very different projected rotational velocities of vsin i ∼ 170 km s⁻¹ for the Aa primary and vsin i ∼ 80 km s⁻¹ for the Ab secondary. The key point to understand is that the already shallow lines of the primary are additionally filled in with continuum light from the secondary. The primary thereby becomes difficult to discern and it was in fact only the extreme orbital velocity and its impact on the Hγ line cores that led Slipher (1904) to suggest that α² Lib should be a near-equal-luminosity system.

Zoom In Zoom Out Reset image size

The question then arises: would we know about the binary character of α² Lib in, e.g., a putative 2 year orbit and with its components still as close as 0.1 arcsec? One may doubt we would. And one may ask how many near-equal-luminosity fast-rotating F- and A-type binaries in the solar neighborhood—often called "variable stars"—are still awaiting detection.

Nearby bright A-type stars, like δ Vel and α² Lib, thus remind us of our incomplete knowledge of their multiplicities and it appears then that the intrinsic luminosity as well as the high stellar rotation rates for the bulk of the A-type stars render them still much unexplored stellar objects. Certainly, there is currently rapid observational progress in this field and there are many more bona fide single objects among the nearby early F- and A-type stars to uncover as binary or higher level systems by means of modern analyses techniques. Dedicated and serious efforts in this direction have a bright future ahead and may likely provide satisfactory answers on the local stellar inventory within just one or two decades.

This work has made extensive use of the Catalogue of Nearby Stars (CNS4) of the Astronomisches Rechen-Institut at Heidelberg, Germany, the ESA Hipparcos catalog, NASA's ADS bibliographic service, and the CDS SIMBAD database, operated at Strasbourg, France. K.F. acknowledges support from the DFG grant FU 198/10-1. We are grateful to L.-S. Buda, T. Dembsky, H. Drass, R. Lemke, and M. Ramolla, who were helpful with the observations for a subset of the sample. This publication is supported as a project of the Nordrhein-Westfälische Akademie der Wissenschaften und der Künste in the framework of the academy program by the Federal Republic of Germany and the state Nordrhein-Westfalen.