2 Observations

The observations have been made by one of us (JK) at the McDonald Observatory in Texas during 1993 using the Cassegrain echelle spectrograph fed with the 2.1-m telescope. A typical spectrum is divided into 26 orders covering the range from $\sim$ 5600 to $\sim$ 7000 Å (one order consists of 1200 pixels in the wavelength scale). The instrument described by McCarthy et al. ([1993]) consists of refractive collimator and camera optics, a 23.2 $\ell$/mm echelle grating with a blaze angle of $65^{\circ}$, a prism cross-disperser and a Reticon 400$\times$1200 CCD with 27$\times$27 $\mu$m pixels. This configuration gives the effective resolution R=64 000 and spectral dispersion of 0.050 Å /pixel at 5790 Å. The S/N ratio for spectra of individual stars is typically $\sim$500 but

 

 

Table 1: Basic stellar data

HD	Name	Sp/L	V	E(B-V)	vsini	cloud type	5780/5797	$\Delta D_{1}$	$\Delta D_{2}$	shift
					km s-1			Å	Å	pixels
23180	o Per	B1 III	3.82	0.26	85	$\zeta$	0.6	0.18	0.15	-16
24912	$\xi$ Per	O7.5 III	4.02	0.29	216	$\sigma$	1.7	0.28	0.23	-14
141637		B3 V	4.64	0.16	300	$\sigma$	2.5	0.26	0.20	+7
142096		B2 V	5.03	0.20	197	$\sigma$	3.0	0.17	0.11	+7
142114		B2 V	4.59	0.14	308	$\sigma$	2.5	0.17	0.11	+9
142184		B2 Vne	5.42	0.16	349	$\sigma$	1.6	0.32	0.30	+9
144217	$\beta^1$ Sco	B0.5 V	2.62	0.17	130	$\sigma$	2.7	0.15	0.12	+7
144470	$\omega^1$ Sco	B1 V	3.96	0.19	142	$\sigma$	1.8	0.18	0.17	+7
145502	$\nu^1$ Sco	B3 V	4.01	0.25	199	$\sigma$	1.9	0.20	0.15	+7
148184		B2 IVpe	4.42	0.49	134	$\zeta$	0.9	0.28	0.17	+8
149757	$\zeta$ Oph	O9.5 V	2.56	0.29	379	$\zeta$	0.8	0.21	0.19	+11
179406		B3 V	5.34	0.32	150	$\zeta$	0.7	0.32	0.30	+15
184915	$\kappa$ Aql	B0.5 III	4.95	0.22	259	$\sigma$	1.9	0.25	0.19	+15
$\star$210839		O6Iab:	5.06	0.52	285	$\sigma$/$\zeta$	1.0	0.75	0.73	+1

is not constant inside a spectrum, even inside one order. Some of the features are easier to detect while situated in a "better'' segment of an investigated spectrum. To obtain high S/N for the final spectra of individual stars the multiple exposures of many program stars were obtained. This signal-to-noise ratio in any spectrum of an individual object is not high enough for detection of very weak features and thus the method of improvement the signal-to-noise is desirable. The detailed methods of data reduction for obtaining spectra of individual stars are described by Kreowski & Sneden ([1993]). Out of the $\sim$80 objects observed at McDonald Observatory during 1993 we selected 14 targets listed in Table 1.

Spectrum of each selected star is characterized by a very narrow and symmetric absorption lines D₁ and D₂ of the interstellar sodium. The widths in Å  of interstellar sodium lines measured on the level of half continuum for our program stars are presented in Table 1. These features allow us to shift the wavelength scale of the spectra to the rest wavelength scale for interstellar clouds, and thus to determine the rest wavelengths of the observed DIBs (the last column of Table 1 presents shifts in pixels for interstellar sodium lines in spectra of our stars). Narrow and symmetrical profiles of sodium lines allow us to suppose that our targets are probably obscured by single clouds (see widths of Na D lines in spectra of our program stars listed in Table 1 in Cols. $\Delta D_{1}$and $\Delta D_{2}$). It is obvious that towards our targets, except HD 210839 we do not observe any detectable Doppler splitting. Thus the latter cannot be expected inside DIB profiles which are broader. Additionally the chosen targets are fast rotators and thus a spectrum of each star is free of weak stellar features containing only strong ones as well as interstellar features and telluric lines.

Figure 1 shows the spectra of a program star (HD 23180) and the divisor (Spica) before division and the resultant spectrum after the cancellation of atmospheric features in the Na D spectral region (thick line). To cancel the telluric lines the spectrum of each star was divided by divisor spectrum (Spica in most of cases) in the whole range of this survey.

Figure 1: The cancellation of $\rm H_{2}O$ contamination in the region of interstellar sidium Na D lines in the spectrum of HD 23180. The lowest (A) spectrum was divided by Spica spectrum (B). The spectrum of the HD 23180 at the top of the panel contains no telluric lines only interstellar features (clearly seen DIB at 5910.40 Å)

The selected objects belong to two distinct types coded as "sigma'' and "zeta'' in accordance to the strength ratio of the central depths of the two strong DIBs: 5780 and 5797 (see 5780/5797 column in the Table 1). All our targets rotate with high velocities. The only one star with smaller rotational velocity listed in Table 1 (HD 23180) is a typical "zeta'' type target. Observational material allows us to select only a couple of each cloud type. These samples can hardly be much extended as a great majority of bright OB stars is observed through several interstellar clouds each. The merging procedure is the averaging of the overlapping wavelength ranges of two neighbour echelle orders. The points are weighted - the weight is proportional to the distance from the edge where the S/N is lowest. Before the merging procedure (described by Kreowski & Sneden [1993]) the telluric line removal has been accomplished. This procedure and continuum placement at the ends of echelle orders is somewhat arbitrary at the $\pm$1% level. This uncertainty does not affect the detection of very weak features which are typically "narrow'', but it compromises the ability to detect the broad and shallow features which stretch over more than one echelle order. As the arbitrarily chosen continuum level in the case of spectra of individual stars (used for achieving spectra of "sigma'' or "zeta'' clouds) makes measurements of weak features difficult (see Kreowski & Sneden [1993]) we measured equivilent widths of new features in the spectrum of HD 210839.