1 Introduction

It seems of importance to check possible relations between identified atomic and molecular features and diffuse interstellar bands (DIBs) which remain unidentified since 1922. It has already been suggested that all possible interstellar absorptions (extinction, atomic and molecular lines, diffuse bands) change in unison i.e. their strengths vary together from cloud to cloud (Kreowski et al. 1992). This fact may be important as the observations of well-identified spectral features can help us to determine physical conditions in individual clouds and thus relate the observed variations of DIB strengths to physical parameters such as temperature or density. This may be very helpful for the task of identifying DIBs - the longest standing unsolved problem in all of spectroscopy.

Figure 1: Figure demonstrates the lack of Doppler splitting in atomic potassium interstellar lines in the spectra of the programme stars (excluding HD152236). The spectra are from Terskol observatory (R=120 000 coudé-echelle spectrometer, Musaev et al. 1999). Note that the set of targets splits into the subsets of objects in which the line is either strong or weak. Very recent paper (Welty & Hobbs 2001) based on the spectra from Ultra High Resolution Facility confirms this result.

Figure 2: The major diffuse interstellar bands observed in the spectra of our five selected targets using the McDonald Observatory echelle spectrometer. HD 152236 was not observed, but we consider it as "zeta''-type object (see text).

Douglas (1977) proposed linear carbon molecules as possible carriers of DIBs. Bare carbon chains, being homonuclear species, do not create rotational transitions observable at radio wavelengths and thus only their electronic and/or vibrational spectral features can be compared with those observed. The latter may cover the spectral range from vacuum UV until far infrared. It seems thus of importance to estimate the abundances of simplest carbon molecules in interstellar clouds. They can be building blocks for many observed (due to radio rotational transitions) interstellar molecules which are often based on carbon skeletons (chains up to 11 atoms long).

   

Table 1: The list of targets with the $\it {A}$¹ $\Pi _{\rm u}$ - $\it {X}$¹ $\Sigma _{\rm g}$⁺ band of C₃ molecule observed (upper part - published data, lower part -our data). Column headings: HD - HD number; SpL - spectral and luminosity class; V - visual apparent magnitude; E_B-V - reddening; $v \cdot$sini - rotational velocity; K I - equivalent width of K I 4044.14 line; 5780 and 5797 - equivalent widths of 5780 and 5797 DIBs respectively; C₂ - column density of C₂ molecule, based on the average of Q(2) and Q(4) lines of the (2,0) Philips system, oscillator strengths from Federman et al. (1994); C₃ - column density based on the average of Q(4), Q(6), Q(8), Q(10), Q(12) lines of $\it {A}$¹ $\Pi _{\rm u}$ - $\it {X}$¹ $\Sigma _{\rm g}$⁺ transition; all of the same oscillator strength $\it {f}$ = 0.0073 (Roueff et al. 2002); C₂/C₃ - the ratio of average column densities. All EW measurements are in mÅ.

HD	SpL	V	EB-V	$v \cdot$sini	K I (4044.14 Å)	5780	5797	C2	C3	C2/C3
24398	B1I	2.96	0.34	59	-	98 $\pm$ 4	57 $\pm$ 1	3.95(12) $\pm$ .45a	1.74(11) $\pm$ .6e	22.7 $\pm$ 10.4
149757	O9V	2.60	0.29	379	-	70 $\pm$ 2	29 $\pm$ 2	4.25(12) $\pm$ .85b	1.96(11) $\pm$ .6e	21.7 $\pm$ 11
179406	B3V	5.36	0.30	170	-	143 $\pm$ 2	79 $\pm$ 2	1.92(13) $\pm$ .22c	2.02(11) $\pm$ .3e	95 $\pm$ 25
210121	B3Vj	7.83	0.40	<20	-	57	-	1.6(13)d	6.8(11) $\pm$ 2.6f	23.5 $\pm$ 12
143275	B0.3V	2.30	0.19	200	-	79 $\pm$ 4	14 $\pm$ 1.0	-	$\le$0.30(11)	-
144217	B0.5V	2.62	0.17	130	0.19 $\pm$ 0.03	161 $\pm$ 4	15.3 $\pm$ 0.5	$\le$1.7(12)g	$\le$0.42(11)	-
147165	B1III	2.89	0.32	53	0.22 $\pm$ 0.03	243 $\pm$ 2	26.0 $\pm$ 3.0	-	$\le$0.29(11)	-
148184	B2V	4.40	0.44	118	1.23 $\pm$ 0.10	104 $\pm$ 3	48 $\pm$ 1.8	7.15(12) $\pm$ .15h	3.4(11) $\pm$ 0.5	21 $\pm$ 3.5
149757	O9V	2.60	0.30	379	0.8 $\pm$ 0.04	70 $\pm$ 2	29 $\pm$ 1.5	4.25(12) $\pm$ .85b	2.2(11) $\pm$ 0.2	19.3 $\pm$ 5.6
152236	B1I	4.80	0.65	60	0.5 $\pm$ 0.04	-	-	3.06(12) $\pm$ .45i	2.6(11) $\pm$ 0.2	11.8 $\pm$ 2.6

^a Chaffee et al. (1980); ^b Danks & Lambert (1983); ^c Federman et al. (1994); ^d Gredel et al. (1992); ^e Maier et al. (2001); ^f Roueff et al. (2001); ^g Lambert et al. (1995); ^h van Discoeck & de Zeeuw (1984); ⁱ van Dishoeck & Black (1989).

The gas-phase optical spectra of linear carbon chains are known for C₂, C₃, C₄ and C₅ (e.g. Motylewski et al. 1999). The first pure carbon molecule, the two-atom homonuclear C₂, was discovered by means of near infrared spectroscopy in 1977 by Souza & Lutz in the spectrum of the opaque cloud obscuring the star Cyg OB2 No. 12. The same authors failed to find the Phillips band (2-0) near 8760 Å in the spectrum of HD 149757 because of the weakness of the features, seemingly correlated with the reddening.

Figure 3: C3 band observed in the spectra of our six targets: relatively strong towards "zeta'' objects and below the level of detection in "sigma'' ones (bottom panel). The neighbour KI interstellar lines are also presented but in another scale as they are deeper than C3 features. The "synthetic'' spectrum of C3shows only the wavelengths of the subsequent transitions. It does not carry any physical information.

The latter, Phillips (2-0) band of C₂ was found in the spectrum of HD 149757 by Hobbs & Campbell (1982) and confirmed in higher S/N spectra by Danks & Lambert (1983). Both teams estimated the C₂ column densities to be of the order $10^{13}~{\rm cm}^{-2}$ in the case of this, well known object characterized by E_B-V close to 0.3. The estimate of van Dishoeck & Black (1986) based on (3-0) Philips band around 7720 Å gave a very similar result. The estimates given for another targets by Danks & Lambert (1983) proved that the ratio of the C₂ column density and E_B-V is rather similar in cases of other reddened stars. Also Crawford (1990) found similar (relative to E_B-V) C₂ column densities towards Sco OB1 stars. The largest existing survey by van Dishoeck & Black (1989) supports also the above mentioned results. The estimates based on the HST spectra which contain the Mulliken system at 2313 Å are below those based on the infrared spectra by a factor of 1.5-2 (Lambert et al. 1995). The extensive survey of C₂ abundances, based on the Phillips 2-0 band (published by van Dishoeck & Black 1989) contains 18 objects. The most recent compilation of Federman et al. (1994) was able to give estimates of the C₂ column densities towards 32 reddened stars plus a couple of upper limits. Not less important seems the fact that vacuum-UV Mulliken band of C₂ has not been detected in the HST spectra of HD's: 144217, 143018 and 144470 despite a substantial reddening and the presence of reasonably strong diffuse interstellar bands observed towards them (Westerlund & Kreowski 1988).

The next member of the possible family of carbon molecules, C₃, is much more difficult to be observed. It was discovered by Hinkle et al. (1988) in the infrared spectrum of the circumstellar shell of the star IRC +10216. This spectral range is, however, not covered with observations of interstellar, translucent clouds due to relatively low opacity. The possible discovery of this molecule (its absorption band $\it {A}$¹  $\Pi _{\rm u}$ - $\it {X}$¹  $\Sigma _{\rm g}$⁺ near 4052 Å) was described by Haffner & Meyer (1995). It was based on several spectra of the heavily reddened star HD 147889 in which the possible C₃ feature appeared as a very weak one. The detailed structure of this band was found towards four nearby reddened stars in July 2000 (Maier et al. 2001; Roueff et al. 2002). The summary of already existing observations gives Table 1. The data in which DIBs have been measured are McDonald R=60 000 spectra (Kreowski & Sneden 1993), except the case of HD 210121. The table contains also C₂column densities, calculated as averages of published equivalent widths of two strong transitions Q(2) and Q(4) with the oscillator strengths given by Federman et al. (1994). The column densities of the C₃ species were calculated in a similar way using Q(4), Q(6), Q(8), Q(10), Q(12) lines of $\it {A}$¹ $\Pi _{\rm u}$ - $\it {X}$¹ $\Sigma _{\rm g}$⁺ transition; all of the same oscillator strength $\it {f}$ = 0.0073 (Roueff et al. 2002).

 

 

Table 2: Equivalent widths (mÅ) of C₃ lines measured in program stars spectra. Wavelengths are adapted from Gausset et al. (1965).

$\lambda$(air)	Line	HD 149757	HD 148184	HD 152236
4049.963	R(16)	0.07 $\pm$ 0.03	0.26 $\pm$ 0.08	-
4050.081	R(14)	0.12 $\pm$ 0.04	0.31 $\pm$ 0.07	0.13 $\pm$ 0.05
4050.206	R(12)	0.13 $\pm$ 0.05	0.12 $\pm$ 0.06	0.11 $\pm$ 0.02
4050.337	R(10)	0.12 $\pm$ 0.04	0.16 $\pm$ 0.04	0.22 $\pm$ 0.04
4050.495	R(8)	0.13 $\pm$ 0.03	0.19 $\pm$ 0.06	-
4050.67	R(6)	0.173 $\pm$ 0.04	0.26 $\pm$ 0.06	0.23 $\pm$ 0.03
4050.866	R(4)	0.102 $\pm$ 0.05	0.31 $\pm$ 0.07	0.21 $\pm$ 0.04
4051.069	R(2)	0.16 $\pm$ 0.04	0.32 $\pm$ 0.05	0.23 $\pm$ 0.04
4051.519	Q(4)	0.25 $\pm$ 0.03	0.28 $\pm$ 0.04	0.26 $\pm$ 0.04
4051.59	Q(6)	0.27 $\pm$ 0.04	0.35 $\pm$ 0.05	0.28 $\pm$ 0.04
4051.681	Q(8)	0.23 $\pm$ 0.04	0.41 $\pm$ 0.06	0.29 $\pm$ 0.02
4051.793	Q(10)	0.21 $\pm$ 0.04	0.39 $\pm$ 0.04	0.31 $\pm$ 0.05
4051.929	Q(12)	0.23 $\pm$ 0.04	0.40 $\pm$ 0.04	0.27 $\pm$ 0.05
4052.473	Q(18)	0.128 $\pm$ 0.05	-	-
4052.698	Q(20)	0.16 $\pm$ 0.03	0.28 $\pm$ 0.05	0.14 $\pm$ 0.04
4052.792	P(8)	0.13 $\pm$ 0.04	0.28 $\pm$ 0.05	0.24 $\pm$ 0.04
4053.591	P(12)	0.13 $\pm$ 0.03	-	-
4053.794	Q(28)	0.08 $\pm$ 0.03	0.19 $\pm$ 0.05	-

Figure 4: Equivalent widths of the subsequent transitions inside the C2 (left panels) and C3 (right panels) bands for individual targets. Panels of each molecule are shown in the same wavelength and intensity scale for better clearness. Note the different relative abundances of the considered species. The measurements of C3 features in the spectrum of HD 210121 are less certain because of the relatively low resolution of the spectra applied. EW data are taken from: a Federman et al. (1994); b Danks & Lambert (1983); c Chaffee et al. (1980); d van Dishoeck & Black (1989); e van Discoeck & de Zeeuw (1984); f Gredel et al. (1992).

However, the existing data on C₃ do not allow to compare the abundances of this species in different environments. All the existing observations concern "zeta'' type clouds i.e. the objects in which the narrow DIBs and the spectral features of simple molecules are relatively strong. The lack of C₂ in certain ("sigma'' type, where molecular and narrow DIBs are typically very weak) objects may suggest that one should expect C₃ to be very weak as well. However, if the Douglas hypothesis is correct, the abundances of longer carbon chains may start growing starting from a certain length. Our observations were made in order to make the simplest test of this possibility i.e. to estimate the abundances of the C₃ bare carbon chain in "sigma'' and "zeta'' environments as defined by Kreowski & Sneden (1995).