THE BINARY NATURE OF CH-LIKE STARS

Carbon stars are recognizable in the optical spectral region by the presence of absorption bands of carbon-bearing molecules (C₂, CN, CH, C₃, SiC, etc.) and are chemically characterized by having C/O > 1. The excess of carbon in such stars is the result of stellar nucleosynthesis in the star itself (intrinsic) or in a binary companion (extrinsic). The first classification system that subdivided the carbon stars by spectral characteristics was the R–N system established by Shane (1928). The R0–R3 stars showed relatively weak C₂ and CN bands, while the R5–R8 stars showed strong bands with a continuum that extended at least down to 3900 Å. A revision of the Morgan–Keenan (MK) system of classification was performed by Keenan (1993), who divided carbon stars into a C–R sequence, a C–N sequence, and a C–H sequence with subclasses according to temperature criteria. The updated sequences are modeled on the old R, N system with a separate category for the CH stars. Thus, stars with enhanced carbon features come in many spectral types: N-type stars, R-type stars, barium and CH stars, R CrB stars, and dwarf carbon stars (Wallerstein & Knapp 1998).

Yamashita (1975) called attention to early carbon stars (the C0–C3 classifications of the MK system, the Harvard R0–R3 classifications) that exhibit the typical spectrum of a CH star but do not have the large space motions that are characteristic of the old population. The CH-like stars are classified in the revised MK system like any other CH-stars, and spectral peculiarities are indicated by abundance indices (Keenan 1993). The N-type carbon stars and late R-type stars are intrinsic AGB stars, while barium and CH-stars are the result of mass transfer in a wide binary system. The evolutionary status of early R carbon stars and, in particular, CH-like stars is still uncertain. According to Zamora et al. (2009), a large fraction of R stars are misclassified stars, some are binaries, and only a small fraction ("true R stars") are single carbon-rich stars.

It is possible that CH-like stars are analogs of barium stars in the old disk population (Začs et al. 2000, 2005). Recently, abundances have been calculated for three CH-like stars in the framework of a larger project devoted to the abundance analysis of R-type carbon stars (Zamora et al. 2009). On the other hand, Mennessier et al. (1997) found that barium stars are an inhomogeneous group of stars. Five distinct classes have been selected, i.e., some halo stars and four groups belonging to the disk population: roughly, supergiants, giants and subgiants, "clump" giants, and dwarfs. The relationship is unclear between halo barium stars and CH-stars, the CH-like stars classified by Yamashita (1975), and the subtypes of disk barium stars selected by Mennessier et al. (1997). A comprehensive analysis is needed to understand the evolutionary status of different subgroups/fractions.

Here, the results of radial-velocity monitoring are presented for most of the CH-like stars classified by Yamashita (1975). CH-like stars are analyzed relative to three comparison stars: a single moderately metal-poor G8 III giant HIP 80843 (Jorissen et al. 2005); a barium star found in the old open cluster, NGC 2420 (Smith & Suntzeff 1987); and the CH-star HIP 22403 classified by Yamashita (1975). In addition, metallicity and s-process abundances are calculated using high-resolution spectra for three CH-like candidate stars (HIP 105212, HIP 62831, and HIP 53522). All of the results of the abundance analyses published to date and the kinematic data are summarized and discussed.

Radial-velocity monitoring of CH-like stars began in 2003 using the CORAVEL spectrometer of Vilnius University installed on the 1.65 m telescope at the Moletai Observatory (Lithuania). All of the classified CH-like stars accessible with the CORAVEL spectrometer are included in the program of observation. The CORAVEL spectrometer (Upgren et al. 2002) is based on principles of the photoelectric radial-velocity scanner developed by Griffin (1967) and operates by scanning a spectrum of star across the mask and obtaining an on-line cross-correlation velocity. The standard deviation of a single observation for late-type stars brighter than about the 11th magnitude is usually better than 0.8 km s⁻¹. The velocities have been standardized using observations of IAU radial-velocity standards (Udry et al. 1999) and are close to the system of velocities published by Nidever et al. (2002) and Marcy & Benitz (1989). The differences in zero-point were found to be 0.14 km s⁻¹ for F-G-K-type stars and 0.4 km s⁻¹ for M-type stars and exhibit rms scatters of 0.5 and 0.8 km s⁻¹, respectively.

Two of the CH-like stars, HIP 43042 (Začs et al. 2005) and HIP 99725, have been monitored by Roger Griffin with the photoelectric radial-velocity spectrometer at the coudé focus of the Cambridge 36 inch reflector. Fifty-one radial-velocity measurements were made for HIP 99725 with a standard deviation of one measurement of 0.5 km s⁻¹. To bring the zero-points of the Cambridge and Vilnius velocity systems into agreement, 0.8 km s⁻¹ has been subtracted from the Cambridge measurements.

In total, 10 stars from the list of 16 CH-like stars published by Yamashita (1975, see their Table 2) have been monitored for radial velocities over a period of about 10 years in this paper. In the case of BD +08°2654 (HIP 62827, HIP 62830, and HIP 62831), three optically nearby stars (components A,B,C) located within 40 arcsec of each other have been included in the program of observations because of contradictory published basic data and the inability to identify with certainty a priori this CH-like star (see the discussion in Section 3). Table 1 gives the basic data for program stars and summarizes the abundances from the literature calculated to date. The standard notations are adopted everywhere.⁴ According to the collected data, HIP 58786 and HIP 62831 do not appear to be enriched in s-process elements. Thus, the status of HIP 58786 and HIP 62831 as CH-like stars is rejected. On the other hand, HIP 62830 seems to be a single F6 dwarf which, in projection on the sky, merely lies close to the CH-like candidate star HIP 62827 = BD +08°2654A.

Table 1.  The Basic Data and Abundance Indices for CH-like Candidate Stars and Comparison Stars

HIP	Name	Spectral Type1	V	${M}_{\mathrm{bol}}{}^{2}$	${T}_{\mathrm{eff}}{}^{2}$	[M/H]a	[C/M]	[s/M]b	[hs/ls]c	References
2529	BD+29°95	R0; C3, 0 CH4	10.18	3.09	4200	...	...	...	...	...
43042	BD+75°348	R0; C3, 0 Ba4	9.54	−0.97	4760	−0.5	0.5	1.6	0.4	4
53522	BD+16°2188	R0; C1, 1; C3, 1 CH4	10.11	−1.89	4955	−0.2	...	1.0	0.6	present
53832	BD+41°2150	R0; C3, 1 CH3 Ba4	10.03	−0.63	4120	−0.77	0.46	1.26	0.72	6
58786	BD+71°600	R2; C2, 3 CH4	10.15	−0.98	4160	−0.29	0.53	−0.04	0.09	6
62827	BD+08°2654A	R0; C2, 1 CH5	9.31	−0.48	4235	...	...	...	...	...
62830	BD+08°2654C	F6 V	10.7	...	...	...	...	...	...	...
62831	BD+08°2654B	K5 III	9.20	...	4100	−0.6	...	−0.6	−0.6	present
80769	BD+19°3109	R2; C1, 1 CH4	10.29	−1.52	4600	...	...	...	...	...
85750	BD+02°3336	R2, C3, 1 CH4 Ba4	9.39	−3.32	3740	−0.48	0.47	1.11	0.68	6
99725	BD+57°2161	R0, C4, 1 CH3 Ba4	9.68	−1.53	4555	−0.1	0.5	1.6	0.4	5
102292	HD 197604	R4; R2; C4, 2 CH8; C3, 2 CH5	9.28	−1.15	4165	−0.9	1.75	1.8	1.0	3
105212	HD 202851	R2; C3, 1 CH4 Ba3	9.67	−1.63	4780	−0.71	...	1.5	0.2	present
		Comparison stars
80843	HD 148897	G8 III CN-2 CH-1 Fe-1	5.25	−2.0	4350	−1.0	...	0.0	−0.2	8
...	NGC 2420 250	Ba II star 7	11.14	...	4200	−0.6	0.8	1.2	0.2	7
22403	HD 30443	C4, 3 CH4: CN3: Ba4	8.82	−4.01	4115	...	...	...	...	...

Notes. [M/H], [s/M], and [hs/ls] values are calculated using abundances published in the papers cited in the last column.

^aM = $\langle$Ti, V, Cr, Fe, Ni$\rangle$. ^bs = $\langle$ls + hs$\rangle$. ^chs = $\langle$Ba, La, Ce, Nd, Sm$\rangle$, ls = $\langle$Sr, Y, Zr$\rangle$.

References. (1) Skiff (2009), (2) Bergeat et al. (2002), (3) Kipper et al. (1996), (4) Začs et al. (2000), (5) Začs et al. (2005), (6) Zamora et al. (2009), (7) Smith & Suntzeff (1987), (8) Jorissen et al. (2005).

Download table as:  ASCIITypeset image

High-resolution spectra for HIP 105212 and comparison stars (HIP 80843, NGC 2420 250, and HIP 22403) were observed with the coudé échelle spectrometer MAESTRO fed by the 2 m telescope at the observatory on the Terskol Peak in Northern Caucasus (altitude of 3100 m) equipped with a Wright Instruments CCD detector with a resolving power of ∼45,000. For HIP 105212, a total exposure of 6000 s was obtained on 2006 November 25. The spectrum covered a spectral region from about 3600 to 10,200 Å in 85 wavelength bands overlapping shortward of H_α with a signal-to-noise ratio (S/N) ∼ 100 in the red wavelength region. The spectra for two other CH-like stars, BD +8°2654 and HIP 53522, were extracted from the archive of observations achieved using the échelle spectrometer LYNX on the 6 m optical telescope of the Special Astrophysical Observatory of the Russian Academy of Sciences. The reddest star (component B) from the three close components was accepted as the CH-like star during the observations of BD +08°2654. These spectra covered the spectral region from about 5000 to 7200 Å and have a resolving power of R ∼ 30,000. The spectra were bias subtracted, flat-field corrected, and converted to one-dimensional spectra using the standard DECH20T package. The wavelength calibration was made using Th–Ar spectra obtained for each night. In addition, a spectrum of a hot and rapidly rotating star was observed to identify the telluric absorption lines. The equivalent widths of the absorption lines were measured by Gaussian fitting to the observed profiles. Three representative spectral regions are shown for the CH-like star HIP 105212 along with those for the three comparison stars in Figures 1–3. Some absorption lines identified in the spectrum of HIP 105212 are marked.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

3.1. Radial Velocities

The radial velocities along with the uncertainty of each observation, as a function of Julian Date for all program stars, are listed below (see Tables A1–A10 in the Appendix). To assess binarity, we have computed the ratio E/I (where E refers to the standard deviation of the velocities and I to the weighted mean uncertainty of the measurements) following Jasniewicz & Mayor (1988). For stars with peak-to-peak velocity variations lower than about 2–3 km s⁻¹, the detection of a binary motion is generally not statistically significant, since the corresponding E/I ratio is smaller than 2. The radial-velocity curves of the three stars (HIP 58786, HIP 62830, and HIP 102292) fulfilling this criterion are plotted in Figure 4, and the corresponding E/I ratios are given as labels in this figure. Among these, the first was shown in Section 2 (see Table 1) not be enriched with s-process elements and the second to be an F6V star. It is therefore not surprising that they do not show any binary signature, as it is very likely that they are not CH-like stars. HIP 102292 is thus the only CH-like star in our sample that does not show evidence of binarity based on its radial-velocity data. Nevertheless, it might be a binary seen nearly face-on. For the remaining stars of the sample, the accuracy of the observations and the length of the time monitoring are sufficient to search for orbital periods and calculate orbital parameters. Table 2 lists the orbital elements for those program stars for which orbits could be computed. The velocities, as a function of phase, resulting from these solutions are plotted in Figures 5–13, and the phases and residuals for the orbital curves are listed for the individual observations in the Appendix.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Table 2.  Orbital Elements with Standard Deviations for the Program Stars.

HIP	P		γ		K		e		ω		T	a sin i		f(m)		${\sigma }_{{\rm{O}}-{\rm{C}}}$
	(days)		(km s−1)		(km s−1)				(deg)		(JDa)	(Gm)		(M⊙)		(km s−1)
2529	725	(20)	20.82	(0.2)	2.4	(0.1)	0	⋯	0	⋯	55457	24	(1.2)	0.00104	(0.00013)	0.42
43042	1042	(5)	51.66	(0.08)	7.29	(0.13)	0.347	(0.018)	233.7	(2.9)	52669 (7)	97.9	(2.0)	0.0346	(0.0021)	0.40
53522	360.8	(0.2)	33.06	(1)	13.58	(1.5)	0.358	(0.06)	81	(3)	56156 (8)	62.9	(7.0)	0.0761	(0.030)	0.50
53832	321	(0.6)	−2.81	(0.2)	11.8	(0.1)	0.114	(0.012)	321	(9)	54182 (7)	51.7	(0.5)	0.0536	(0.0019)	0.50
62827	570.1	(0.5)	34.5	(0.11)	8.6	(0.2)	0	⋯	0	⋯	54449	67.2	(1.1)	0.0376	(0.0026)	0.48
80769	2129	(13)	−183.6	(0.1)	7.5	(0.2)	0.186	(0.016)	226	(56)	55711 (32)	217	(4.2)	0.0894	(0.0064)	0.47
85750	446.4	(0.6)	−27.23	(0.05)	10.6	(0.1)	0	⋯	0	⋯	55497	65.1	(0.5)	0.0552	(0.0016)	0.30
99725	541.1	(0.7)	−62.14	(0.05)	6.39	(0.07)	0.044	(0.011)	30	(17)	54311 (25)	47.5	(0.6)	0.0146	(0.0005)	0.30
105212	1295	(6)	20.53	(0.07)	8.5	(0.1)	0.055	(0.01)	327	(17)	56323 (67)	151	(1.6)	0.0829	(0.0026)	0.30
Non-CH-like system
62831	1963	(6)	−2.61	(0.09)	4.6	(0.1)	0.257	(0.029)	45	(6)	54567 (29)	121	(3.8)	0.0177	(0.0019)	0.64

Note. The Data for HIP 43042 have been Adopted from Začs et al. (2005).

^aJD = T + 2,400,000.

Download table as:  ASCIITypeset image

In the case of BD +08°2654A (HIP 62827), an orbit has already been published by McClure & Woodsworth (1990). The orbit calculated from our measurements is in good agreement with theirs, except for a difference of about one order of magnitude in the projected semimajor axis which seems to be a misprint in their Table 4 (McClure & Woodsworth 1990). The orbital elements estimated recently for two stars in common (HIP 53522, HIP 53832) by Jorissen et al. (2016) agree well with our calculations.

In addition, the radial velocities were measured using a sample of weak and symmetrical atomic lines in the high-resolution spectra of three stars. The direct and mirror profiles of each line were correlated, and the corresponding radial velocity was calculated from the relative shift of the line center from its rest wavelength. These values (see, for instance, Figure 12; open circles) agree well with the CORAVEL measurements.

3.2. Space Velocities

3.3. High-resolution Spectroscopy

The spectra of early-type carbon stars are relatively free of strong molecular lines in comparison with N-type carbon stars; however, many weak transitions of CN and C₂ are visible in the optical region of the spectrum and blend with the atomic lines. Inspection of a high-resolution spectrum shows that relative to the moderately metal-poor ([Fe] = −1.0) giant HIP 80843 (G8 III), the lines of the iron-peak elements are of similar strength (see Figures 2 and 3), while features due to C₂, CN, and neutron-capture elements are enhanced. For example, the neodymium, praseodymium, and cerium lines in the spectra of the CH-like star HIP 105212, and related spectroscopic binaries, are much stronger compared with those for HIP 80843, indicating a significant overabundance of neutron-capture elements. Typical R0–R3 stars show weak molecular bands, and their atomic-line spectra are equivalent to G9–K2 normal giants; however, the s-process elements usually are not enhanced (Zamora et al. 2009). The most prominent molecular features for HIP 105212 in the observed region are the C₂ Swan system bands (1, 0) at 4740 Å, (0, 0) at 5165 Å, and (0, 1) at 5635 Å. In addition, over the entire analyzed spectrum, bands of the CN red system are evident, significantly blending the atomic absorption lines. Thus, most of the atomic lines are blended with molecular lines. The final selection of clean atomic absorption lines for the abundance analysis was done using the synthesized atomic and molecular spectrum for a typical carbon star.

3.3.1. Atmospheric Parameters

Effective temperatures for most of the program stars are estimated by Bergeat et al. (2002) using the intrinsic spectral-energy distribution plus the atmospheric model technique (see Table 1). An excitation analysis of the Fe i lines for analyzed stars generally supports these values. The surface gravities ($\mathrm{log}g$) were determined from the Fe i/Fe ii ionization balance, and the microturbulent velocities by forcing the abundances of individual Fe i lines to be independent of the equivalent width. The final atmospheric parameters for the three program stars are displayed in Table 4.

Table 4.  Atmospheric Parameters for HIP 53522, HIP 62831, and HIP 105212

Star	HD/BD	Teff	$\mathrm{log}g$	ξt	[Fe/H]
		(K)	(cgs)	(km s−1)	(dex)
HIP 53522	BD +16°2188	4550	2.5	2.0	−0.2
HIP 62831	BD +08°2654B	4100	1.3	3.0	−0.4
HIP 105212	HD 202851	4800	2.1	1.5	−0.7

Download table as:  ASCIITypeset image

3.3.2. The Abundances

The abundances for selected iron-group and neutron-capture elements were computed using equivalent widths for three program stars with the main goal of estimating the metallicity and confirming the enhancement of s-process elements. A detailed abundance analysis for selected stars will be presented in the next paper. The standard LTE line analysis program WIDTH6, written by R. L. Kurucz, was used, which employs an input model atmosphere to compute the strength of a given atomic line formed in such an atmosphere. The model atmospheres were taken from the grid of Kurucz (1993). Oscillator strengths for the lines have been taken from a variety of sources compiled by E. Luck. A majority of the values are from high-precision laboratory measurements. Spectral lines stronger than 300 mÅ were not used for abundance calculations. The resulting relative abundances normalized with the solar-system abundances of Grevesse et al. (2007) are listed in Table 5. We used mean abundances of iron-group elements to quantify the metallicity, [M/H] = 1/5 ([Ti/H] + [V/H] + [Cr/H] + [Fe/H] + [Ni/H]), the abundance of Y to quantify ls, and a combination of the remaining s-process elements to quantify hs, [hs] = 1/4 ([Ba/Fe] + [La/Fe] + [Ce/Fe] + [Nd/Fe]; see Tables 1 and 5). One wavelength region around important atomic lines was calculated using the spectral synthesis code STARSP (Tsymbal 1996) based on Kurucz's code SYNTHE to illustrate the abundance results (Figure 14). The systematic errors produced in abundances by uncertainties in T_eff (±200 K), $\mathrm{log}g$ (±0.3 dex), and ξ_t (±0.5 km s⁻¹) would lead to errors of less than 0.3 dex, that is, a little worse than those for barium stars.

Zoom In Zoom Out Reset image size

Table 5.  Relative Abundances of the Iron-Group and n-capture Elements for Three Program Stars

Star	[Ti/H]	[V/H]	[Cr/H]	[Fe/H]	[Ni/H]	[Y/H]	[Ba/H]	[La/H]	[Ce/H]	[Nd/H]
HIP53522	−0.2 (12)	−0.4 (5)	−0.2 (5)	−0.2 (29)	−0.1 (6)	+0.4 (3)	+1.1 (1)	+1.1 (1)	...	+0.7 (2)
HIP62831	−0.7 (24)	−0.6 (18)	−0.4 (9)	−0.4 (71)	−0.8 (9)	−0.7 (4)	−1.8 (3)	...	−1.2 (3)	−0.9 (3)
HIP105212	−0.7 (10)	−0.7 (4)	−0.9 (3)	−0.7 (29)	−0.6 (3)	+0.6 (4)	+0.9 (3)	+0.7 (2)	+0.6 (2)	+0.9 (7)

Note. The number of lines used in the analysis is also given.

Download table as:  ASCIITypeset image

Large enhancements of s-process elements in the atmosphere of HIP 105212 relative to solar values are found; [s/M] ≃ +1.5 dex, on average. HIP 53522 displays a lower overabundance of neutron-capture elements, [s/M] ≃ +1.0 dex, on average. The fit between the observed and synthesized spectra confirms a moderate iron deficiency along with the overabundance of neutron-capture elements (see Figure 14). This suggests that material rich in carbon and neutron-capture elements has been added to the envelope of both stars. The heavy s-process elements are relatively more enhanced than the light s-process peak elements in the atmospheres of HIP 105212 and HIP 53522. However, for the reddest (coolest) star of the optical triplet BD +08°2654 (component B, HIP 62831), which is a spectroscopic binary (see Figure 9), no overabundance of neutron-capture elements has been found. Presumably, the carbon star is component A (HIP 62827) which, incidentally, is also a spectroscopic binary according to radial-velocity monitoring (see Figure 8). A high-resolution spectrum is needed to confirm the enhancement of carbon and neutron-capture elements in the atmosphere of HIP 62827.

Radial-velocity monitoring has confirmed with certainty the status of long-period binaries for all of the classified CH-like stars except two. The scatter of the radial velocity for the remaining two stars (HIP 102292 and HIP 58786) is approximately 3–4σ of the measurements, but regular variations have not been noted to date. However, the enhancement of s-process elements was not confirmed for HIP 58786 by Zamora et al. (2009). Thus, the status of CH-like star was rejected, and this star could be a single star. For HIP 102292, the orbital period could be very large or the orbit may be almost pole-on, and thus the amplitude of the velocity variations is quite low. Orbital periods calculated for the other nine program stars range from 321 to 2129 days. The eccentricities and mass functions are low on average, which lends support to the supposition that mass transfer might have taken place in the past and that the unseen companion is a low-mass star. The orbital eccentricities of the barium and CH-stars are, on average, lower than those estimated in a sample of normal G-K giants because of mass transfer in binary systems by different evolutionary channels as modeled by Han et al. (1995). Recently, more channels were introduced to explain some specific systems with a short period, large eccentricity, and a circumbinary disk (Abate et al. 2013; Dermine et al. 2013). However, we have not detected such systems in the sample of CH-like stars. Among the CH-like stars, none are found at short periods and large eccentricities on the $(e,\mathrm{log}P)$ diagram (see Figure 15). It is worth noting that neither pre-mass-transfer systems nor post-mass-transfer systems populate the avoidance zone in the bottom right corner of the (e–log P) plane (Mathieu 1994). The two exceptions, with P ∼ 3000–4000 days and 0.00 < e < 0.05, are due to the low accuracies of the eccentricities; high accuracies are difficult to obtain for such long-period systems, as discussed by Jorissen et al. (2016).

Zoom In Zoom Out Reset image size

The metallicities of CH-like stars are in the range from −0.2 to −0.9 dex (see Table 1): a typical value for the thin/thick-disk stellar population of the Galaxy. The abundances of neutron-capture elements are enhanced, [s/M] = +1.0 to +1.8 dex, in support of nucleosynthesis during double-shell burning. However, the luminosities of CH-like primary stars are too low to be on the asymptotic giant branch, in support of a mass transfer scenario from an AGB companion in the past. The [hs/ls] ratio in the atmospheres of the analyzed stars is clearly higher than that found for typical barium stars (Busso et al. 1999). The classical halo binaries, the CH-stars, are high-velocity objects with large [hs/ls] ratios, [hs/ls] ≃ +1.0 (see Figure 5 of Začs et al. 1998), because at their lower metallicities the heavy s-process elements have higher abundances than the light ones. In spite of a large spread of the [hs/ls] ratio, [hs/ls] ≃ 0.0 for most of the studied barium stars (see Figure 8 of Busso et al. 1999); however, all of the analyzed CH-like stars (except HIP 105212) display [hs/ls] ≥ 0.4.

Another difference is the high carbon abundance. Classical barium stars usually do not display in their spectra significant features of carbon-bearing molecules; as a rule, C/O < 0.6 (see, for example, Barbuy et al. 1992). However, no significant difference was found between spectra of the typical CH-like star HIP 105212 and the barium star in the old cluster NGC 2420; both display moderately strong C₂ and CH bands (see Figures 1 and 3). Unfortunately, spectroscopic estimations of oxygen abundances for carbon stars using high-resolution spectra in the optical region are problematic, and the calculated C/O ratios are uncertain. Spectra in the infrared region are needed. Zamora et al. (2009) were not able to calculate the oxygen abundances for three observed CH-like stars. Začs et al. (2005) employed one weak oxygen line and calculated a C/O ratio for the CH-like star HIP 99725: C/O = 2.24. Dominy (1984) calculated oxygen abundances using infrared spectra for a sample of 11 early-type R stars and found C/O ratios from 0.9 to 3.3. Isotopic ¹³C lines are enhanced in the spectrum of the CH-like star HIP 105212 (see Figure 3), similar to that observed for a number of barium stars (Barbuy et al. 1992). For two of the analyzed CH-like stars, HIP 53832 and HIP 85750, the ¹²C/¹³C ratio was found to range from 22 to 24 (Zamora et al. 2009). On the contrary, HIP 58786, for which the status of CH-like star was rejected, displays ¹²C/¹³C = 70. True early R stars have ¹²C/¹³C < 10.

For the program stars, the separation between halo and thin/thick-disk stellar populations has been performed. In spite of several stars with significant standard deviations, the results indicate that all of the analyzed stars, except one, are most likely disk objects, thin or thick. We conclude that some CH-like stars are apparently analogs of the classical CH and barium stars in the old disk population; a few of the stars are standard barium and CH-stars. HIP 80769 and HIP 2529 could be a standard CH-star and a barium star, respectively, according to their radial and space velocities. However, abundance analysis is needed to confirm s-process overabundances. The comparison star NGC 2420 250, classified as a barium star, instead seems to be a CH-like star, bearing in mind the abundances and the old disk population membership. For the remaining stars, more accurate space velocities, distances, and detailed abundances are needed to secure their subtypes. Thus, the evolutionary status of one fraction of early-type carbon stars, the so-called CH-like stars, is clarified. As a consequence, the number of confirmed true (single) R-type carbon stars is low in the Galaxy, which is in agreement with conclusion reached by Zamora et al. (2009).

We thank Roger Griffin for radial-velocity monitoring and calculation of the orbital elements for HIP 99725. We acknowledge support from the Research Council of the Lithuania under grant No. MIP-85/2012, from the Latvian Council of Science under grant No. 09.6190, and from the Universidad Nacional Autónoma de México by way of project IN103014 of the PAPIIT/DGAPA. This research has made use of the Simbad database operated at CDS, Strasbourg, France. We greatly thank the referee for his constructive and valuable discussions and comments which have improved this paper.