Astron. Astrophys. 323, 853-875 (1997)

5. Results from individual source observations

5.1. Accreting candidates

5.1.1. VES 625 = LS I +61 235 (Group 1)

This Be/X-ray system was discovered by Motch et al. (1991b) as a result of a preliminary investigation of the ROSAT Galactic Plane Survey. It actually stands out from Table 5 as exhibiting one of the largest / . Optical observations revealed only moderate Balmer emission (EW (H ) = -7,-10 Å) with a distinct absorption core on occasions (Coe et al. 1993). Follow-up ROSAT pointed PSPC observations by Hellier (1994) showed that the source was pulsating with a 1412 s period thus being the slowest Be/neutron star known. Together with the nearby unusual ultrasoft 8.7 s pulsar (Israel et al. 1994), RX J0146.9+6121 = LS I +61 235 contributes to the Uhuru source 4U 0142+61. The energy distribution of RX J0146.9+6121 is hard with a power law energy index of -0.6 0.7 (Hellier 1994) in agreement with the very positive HR1 and HR2 measured during the ROSAT all-sky survey.

Medium resolution blue spectroscopy obtained at OHP on 1992 October 23 and 1995 January 30 (see Fig. 4) indicates a spectral type significantly earlier than the B5III type proposed by Slettebak (1985). The relative strength of the Si III 4552 / He I 4387 suggests a giant luminosity class. For this class, the C III + O II blend at 4650 rather indicates an early B1 type. Therefore, although our spectral range may not be ideal for spectral classification, we believe that the star is hotter than previously thought and may be classified as a B1IIIe type star. The revised distance of LS I +61 235 (d 2.9 kpc) is more consistent with the distance of the open cluster in which it is located (d = 2.5 kpc; Tapia et al. 1991) than that derived from a B5III type (d 1.3 kpc).

Fig. 4. Rectified blue medium resolution spectrum of VES 625 = LS I +61 235 obtained with the OHP 1.93 m and CARELEC spectrograph on 1995 January 30. Exposure time was 10 min

The Balmer emission lines exhibit dramatic V/R changes on time scales of two years or less as shown for the H line in Fig. 5. Our red H spectrum obtained on 1992 December 15 (see Fig. 6) is mirror-symmetric of that plotted in Coe et al (1993) and obtained on 1991 August 28, in the sense that the 'violet' component is much stronger than the 'red' one. The maximum time for the H V/R reversal is of the order of 1.5 yr.

Fig. 5. Rectified blue medium resolution spectrum of VES 625 = LS I +61 235 obtained with the OHP 1.93 m and CARELEC spectrograph on 1992 October 23 and 1995 January 30 showing the large V/R changes having occurred in the 2 years interval

Fig. 6. Red medium resolution spectrum of VES 625 = LS I +61 235 obtained with the OHP 1.93 m and CARELEC spectrograph on 1992 December 15. Exposure time was 30 min

5.1.2. BSD 24- 491 (Group 1)

Visual inspection of the POSS O and E plates and B and I band CCD images failed to reveal any plausible alternative active corona counterpart within the error circles of the ROSAT survey and pointing positions. The star BSD 24- 491 = LS V +44 17 = VES 826 is referenced as a B0 star with indication for H emission (Seyfert & Popper 1941; Coyne & McConnell 1983). Low resolution spectrum obtained on 1991 November 16 (see Fig. 7) shows clear evidence for H emission with an EW = -2.0 0.2 Å .

Fig. 7. Low resolution spectrum of BSD 24- 491 obtained on 1991 November 16 UT at OHP. The FWHM resolution is 15 Å . Relatively weak H emission is clearly detected

A blue high resolution spectrum obtained on 1991 November 27 confirms the B0 type listed in SIMBAD for this star (see Fig. 8). There are no good luminosity sensitive lines in our wavelength range. However, the EW of He I 4471 rather suggests a dwarf or giant star (Didelon 1982). The photometry reported by Bigay (1963); V = 10.78, B-V = 0.61, U-B = -0.36 indicates a rather high interstellar absorption E(B-V) = 0.90 in agreement with the depth of the 4430 and 6284 interstellar bands. An additional medium resolution spectrum obtained at OHP in 1995 January reveals an increase of the Balmer emission compared to our observation in 1991 with strong double peaked H and iron/He I lines in emission.

Fig. 8. Rectified blue medium resolution spectrum of BSD 24- 491 obtained on 1991 November 27 at OHP

The source was detected again during a dedicated pointed PSPC observation with a count rate 2.5 times higher than during survey observations. Our pointed observation indicates that the source is hard (HR2 = 0.55 0.07) and maybe variable (see Fig. 9). The position of the X-ray source derived from the pointed observation is only away from the GSC position of the Be star. From all these evidences we conclude that BSD 24- 491 is very likely to be a new Be/X-ray system.

Fig. 9. The pointed observation 0.5-2.0 keV X-ray light curve of BSD 24- 491. Integration time is 150 s

5.1.3. LS 992 (Group 1)

LS 992 is one of the X-ray brightest and hardest new Be/X-ray candidates in the ROSAT all-sky survey. Close inspection of B and I images (see Fig. 10) failed to reveal any alternative optical counterpart to the X-ray source. Optical spectroscopy of object B, the second brightest in the error circle, suggests a late type star with H and H in absorption, leaving the B star as the most likely optical counterpart of RX J0812.4-3114. Our blue medium resolution spectrum of LS 992 is unfortunately somewhat too noisy to derive a very accurate spectral type. However, the Si III/IV and He I lines indicate a B0-1 spectral type and a luminosity class V to III (see Fig. 11). The red spectra show the presence of H emission with a central absorption core (see Fig. 12) similar to that observed from LS 1698 (see below). The total H equivalent width does not vary much over the one week interval between the two spectroscopic observations (EW(H ) = -4.6 0.1; -4.9 0.1 on 1992 April 18 and 25 respectively). However, the V/R ratio may have slightly evolved.

Fig. 10. B and I CCD images of the field of LS 992 obtained on 1992 April 18 with EFOSC2 and the ESO-MPI 2.2 m telescope. Both images have 2 min exposure time. North is at top and East to the left. Each frame is wide. On the B image we plot the ROSAT survey 90% confidence error circle and show the position of LS 992. On the I band image we show the position of the alternative optical candidate "B" investigated spectroscopically and apparently unrelated to the X-ray source

Fig. 11. Rectified blue medium resolution spectrum of LS 992 obtained with the ESO-MPI 2.2 m + EFOSC2 on 1992 April 18. Exposure time was 15 min

Fig. 12. Red medium resolution spectrum of LS 992 showing the H emission. The upper spectrum was exposed 20 min on 1992 April 18 and the lower one 5 min on April 25. Both spectra were obtained with the ESO-MPI 2.2 m telescope

Optical photoelectric photometry by Reed (1990) gives V = 12.42; B-V = 0.41 and U-B = -0.69 in agreement with the B0-1 V-IIIe spectral type proposed. Assuming (B-V)₀ = -0.28 yields E(B-V) = 0.69, consistent with that derived from the depth of the 4430 and 6284 interstellar bands which imply E(B-V) = 0.7-1.0 and E(B-V) = 0.8 0.2 respectively.

Although LS 992 is the second brightest X-ray binary candidate from the survey data, the statistics are still not good enough to really constrain the shape of the X-ray energy distribution. The spectrum is undoubtedly much harder (HR2 = +0.74 0.09) than normal OB star emission. Fitting simple power law models and fixing the column density at its interstellar value ( = 5 10²¹ cm^-2) implies a photon index in the range of +0.5 to -1.5 (95% confidence level), well within the range observed from accreting neutron stars in massive X-ray binaries (White et al. 1983). The hard X-ray light curve plotted on Fig. 13 shows some evidence for variability.

Fig. 13. Survey light curve of RX J0812.4-3114, the source associated with LS 992. Each data point corresponds to one satellite scan (10 to 32 s every 96 min) integrated between 0.4 and 2.4 keV. Time 0 is JD 2448192.5759

A long follow-up pointed PSPC observation on 1992 November 20 failed to recover the source. A 3 upper limit of 3 10^-3 cts s^-1 can be set to the count rate of RX J0812.4-3114, i.e. about 100 times fainter than during survey observations. Contemporaneous optical spectroscopy obtained at LNA, Brazil, on November 28 shows H emission with an EW = -4.6 Å , similar to that observed 7 months before at La Silla.

No catalogued X-ray source is found at the position of LS 992. The large long term flux variability, similar to that exhibited by many Be/X-ray transients and the X-ray hardness strongly suggest that LS 992 is a new accreting Be/X-ray system.

5.1.4. LS 1698 (Group 1)

Comparison of the I and B band CCD images failed to reveal any other bright or very red star in the error circle which could be an alternative active corona identification (see Fig. 14). In the red, H is clearly seen in emission with a strong absorption core (see Fig. 15). The same H profile was again observed in February 1994. Our medium resolution blue spectrum shown in Fig. 16 indicates a B0 V-IIIe type star. The strength of the 4430 and 6284 interstellar bands suggests a rather large interstellar absorption E(B-V) = 0.75 0.25. Using the GSC V magnitude and taking into account the reddening estimated from interstellar bands lowers the distance of LS 1698 from 18 kpc to 5 kpc.

Fig. 14. B and I CCD images of the field of LS 1698 obtained on 1992 April 17 with EFOSC2 and the ESO-MPI 2.2 m telescope. Both images have 2 min exposure time. North is at top and East to the left. Each frame is wide. On the B image we plot the ROSAT pointing 90% confidence error circle and show the position of LS 1698

Fig. 15. Rectified red medium resolution spectrum of LS 1698 showing the H emission. This spectrum (15 min exposure time) is the sum of two spectra obtained on 1992 April 15 and 17 with the ESO-MPI 2.2 m telescope

Fig. 16. Rectified blue medium resolution spectrum of LS 1698 obtained with the ESO-MPI 2.2 m + EFOSC2 on 1992 April 17. Exposure time was 10 min

The source was again detected during a dedicated pointed observation with a count rate about 20 times lower. The survey energy distribution characterized by HR2 is hard and HR1 is consistent with the large interstellar absorption. From these evidences, we conclude that LS 1698 is most probably a new Be/X-ray system.

We note that LS 1698 is the likely optical counterpart of the hard X-ray transient source 4U1036-56/3A 1036-565. Its position lies well inside the error box of the Uhuru source and close to the Ariel V position and the ROSAT survey source is the only one in the Uhuru and Ariel error boxes. The former identification of 3A 1036-565 with the bright Be star HD 91188 (Buckley et al. 1985) is thus probably not valid since the star lies quite outside the error box and is furthermore not detected in the ROSAT all-sky survey.

The source was first observed by the Uhuru satellite (Forman et al. 1978) and OSO 7 (Markert et al. 1979). Ariel V observed a flare in November 1974 (Warwick et al. 1981). At maximum flare the flux was 2.4 10^-10 erg cm^-2 s^-1 (2-10 keV) while the mean Uhuru flux was 1.0 10^-10 erg cm^-2 s^-1 (2-10 keV). During the survey, ROSAT detected a flux about 10 times less than during the years 1970-1976.

5.1.5. LS 5039 (Group 1)

From our B and I band images of the field of LS 5039 we could select 3 alternative candidate stars designated B,C and D on Fig. 17. Low resolution spectroscopic observations of these stars failed to reveal spectral signatures of any known class of galactic or extragalactic X-ray sources. All three candidates exhibit H in absorption. Object B is an early M star. However, it lacks the Balmer emission which is nearly always found in X-ray active M stars (see e.g. Fleming et al. 1988; Motch et al. 1996b). Object C and D are too faint to be easily classified. Therefore, LS 5039 constitutes the most likely optical counterpart of the X-ray source.

Fig. 17. B and I CCD images of the field of LS 5039 obtained on 1992 April 19 with EFOSC2 and the ESO-MPI 2.2 m telescope. Both images have 2 min exposure time. North is at top and East to the left. Each frame is wide. On the B image we plot t he ROSAT survey 90% confidence error circle and show the position of LS 5039. On the I band image we show the position of the alternative optical candidates investigated spectroscopically and apparently unrelated to the X-ray source

The medium resolution blue spectrum of LS 5039 shown on Fig. 18 is typical of an O7V star with well marked He II 4686 absorption and some evidence for weak N III 4634-4642 emission, i.e. an O7V((f)) classification. No H emission is present in our medium resolution red spectrum.

Fig. 18. Rectified blue medium resolution spectrum of LS 5039 obtained with the ESO-MPI 2.2 m + EFOSC2 on 1992 April 19. Exposure time was 5 min

The optical photometry reported by Drilling (1975) gives V = 11.23, B-V = 0.94, U-B = -0.16. A (B-V)₀ = -0.32 implies E(B-V) = 1.26, a value slightly larger than that derived from the 4430 and 6284 interstellar band which both point at E(B-V) = 0.8 0.2.

During survey observations the X-ray source was one of the hardest of the OB/X-ray correlation list. The source was detected again during a subsequent PSPC pointing. The X-ray position derived from this observation is completely consistent with the survey position. The count rate is close to that of the survey but owing to the longer exposure time the slightly improved statistics allows to confirm the hardness of the source, clearly compatible with an accreting neutron star or black hole seen through the large intervening column density.

In spite of the slight increase of the bolometric correction with respect to the automatic process which assumed a B0 III type, the updated / ratio remains high (10^-5), exceeding by an order of magnitude at least that expected from normal O stars. We conclude that LS 5039 is a likely massive X-ray binary in which the compact object may accrete from the O star high velocity wind.

5.1.6. HD 161103 (Group 3)

HD 161103 is a relatively bright and variable (V = 8.4-8.7) B2 V-IIIe star. Spectroscopic observations on 1994 February 13 show prominent Balmer emission with EW(H ) = -32 Å . The GSC star nearest () to the B star (GSC 0683600952; V 12.4) but located slightly outside the 90% confidence survey error circle was also observed at this occasion and failed to reveal the signature characteristic of an active corona.

The source was detected again during a short dedicated PSPC pointing with a count rate about half of that recorded during the survey. The position resulting from the pointed observation is now only away from HD 161103 and the energy distribution is clearly hard (HR2 = +0.47 0.21) although we still have too few photons to really characterize the energy distribution. The / ratio is close to the end of the observed distribution for O stars (Sciortino et al. 1990). However, this is still 50 times more than for the B1-B2 type stars measured by Cassinelli et al. (1994). The corresponding un-absorbed 0.1-2.4 keV X-ray luminosities are close to 10³² erg s^-1.

Our conclusion is that the evidences in favour of an accreting component are strong but not firm enough in order to definitively claim that HD 161103 is a new Be/X-ray system. However, this object is a good candidate and would certainly deserve follow-up X-ray observations in a harder X-ray band.

5.1.7. SAO 49725 (Group 3)

SAO 49725 is a relatively bright B0e star (V = 9.23) suffering a reddening E(B-V) 0.67. The survey source was recovered in a subsequent pointing with a count rate of about twice of that given by the EXSAS analysis for the survey. Both survey and pointed positions encompass SAO 49725 (see Fig. 19).

Fig. 19. B and I CCD images of the field of SAO 49725 obtained on 1992 May 27 with the CCD camera at the OHP 1.2 m telescope. Both images have 10 min exposure time. North is at top and East to the left. Each frame is 3.54 3.54 wide. On the I image we plot the ROSAT survey and pointed 90% confidence error circles. On the B band image we show the position of the other possible optical candidates which were spectroscopically investigated

None of the red excess candidates located in or close to the X-ray error circles and marked on Fig. 19 display any signature of chromospheric activity. Objects A, D and C are late G-F stars. Object B is a M star without any detectable Balmer or Ca II emission and is therefore an unlikely optical counterpart. Our blue medium resolution spectrum (see Fig. 20) displays, in addition to H , O II, C III and weak He II lines which indicate a B0.5 type in agreement with former determinations. The relative strength of the He II 4686 and He I 4713 and that of the Si II lines suggests a dwarf or giant luminosity class.

Fig. 20. Blue medium resolution spectrum of SAO 49725 obtained on 1992 October 22 at OHP with the 1.93 m telescope and the CARELEC spectrograph. Exposure time is 20 min

Fig. 21 shows the prominent H (EW = -30 1 Å) and H (EW = -2.0 0.1 Å) emission lines. The Balmer continuum is also seen in emission revealing the presence of a large circumstellar envelope.

Fig. 21. Low resolution spectrum of SAO 49725 obtained on 1991 June 8 at OHP with the 1.93 m telescope and the CARELEC spectrograph. The 60% flux difference between this spectrum and that shown in Fig. 20 is due to inaccurate flux calibration. Exposure time is 2 min

The pointing data show that the X-ray source is relatively hard. With the updated optical and X-ray data, the / ratio remains close to 3 . This is rather strong evidence, although not fully compelling, for an accreting compact object around SAO 49725.

5.2. Uncertain cases

HD 38087 (Group 1): HD 38087 is a B5V star (Schild & Chaffee 1971) located in the Orion OB1 association. The distance resulting from the photometry is 470 pc in agreement with that derived by Warren & Hesser (1978) for the belt subregion ( 440 pc). However, the star is associated with a reflection nebulosity and may have an anomalous reddening law ( = /E(B-V) = 5.3-5.7; Whittet & van Breda 1980, Cardelli et al. 1989) possibly caused by unusually large grains (Snow & Witt 1989). These authors derive a total = 4.0 0.5 cm^-2 whereas with an observed E(B-V) = 0.29 the expected should be 2.3 cm^-2.

The position of the source was covered by two ROSAT pointings and it was detected on each occasion with a mean count rate similar to that observed during survey observations. Survey and pointing spectra have rather soft HR2s usually not observed in OB/X-ray systems. The spectrum accumulated from WG900386 can be represented by a two temperature thin thermal Raymond Smith spectrum (kT1 = 0.28 keV; kT2 = 3.1 keV; = 3 10²¹ cm^-2) typical for active coronae. Assuming a distance of 470 pc the un-absorbed 0.1-2.4 keV X-ray luminosity is 7 10³¹ erg s^-1 (two temperature plasma), close to that computed by the automatic process. Such a high X-ray luminosity is sometimes observed during flaring states in the most active T Tauri stars (see e.g. Montmerle et al. 1983). With an estimated cluster age of 5 10⁶ yr for the OB1 b1 group (Warren & Hesser 1978) and considering the softness of the source, the X-ray emission could still originate from a very active late type star physically linked to HD 38037. The emission line star LkHA 291 is apparently located well outside the ROSAT error circle as seen from the finding chart in Herbig and Kuhi (1963).

HD 53339 (Group 1): This B3V star belongs to the Canis Major OB1 association located at a distance of 1150 pc (Claria 1974). It consists of a close pair separated by 1.1 and with a magnitude difference of . If the pair is physically linked, the faint component is a late B star with little intrinsic X-ray emission. The Einstein IPC did not detect X-ray emission from the region with an upper limit of 5 10^-3 cts s^-1 probably consistent with the PSPC count rate of 2.3 10^-2 cts s^-1. Hardness ratios are marred by large errors and compatible with both an active corona or a hard accreting source. At the time of writing this paper the AO-3 pointing data covering HD 53339 were not available from the archive.

Our optical observations failed to reveal any alternative bright optical counterpart. However, several faint stars in the error circle (see Fig. 22) would deserve further investigation before settling the case. The reddish object B is a late K-M star without H emission. With an estimated un-absorbed 0.1-2.4 keV X-ray luminosity of 6.5 10³¹ erg s^-1 (computed using the results of the EXSAS analysis of the survey), HD 53339 clearly deserves further optical and X-ray monitoring.

Fig. 22. B and I CCD images of the field of HD 53339 obtained on 1992 April 22 with EFOSC2 and the ESO-MPI 2.2 m telescope. Both images have 5 min exposure time. North is at top and East to the left. Each frame is wide. On the B image we plot th e ROSAT survey 90% confidence error circle and show the position of a late K-M star which we have investigated spectroscopically and which does not exhibit H emission

HR 2875 (Group 1): HR 2875 is a bright (V = 5.41) B5Vp star also detected by the ROSAT Wide Field Camera (Pounds et al. 1993). The PSPC energy distribution is extremely soft and therefore the standard process computing the / ratio, which assumed a 10⁷ K thin thermal emission, largely overestimated the X-ray flux in that particular case. No photon is detected above 350 eV in the ROSAT survey data and the overall PSPC spectrum looks like that of an isolated hot white dwarf. Considering the very low probability of a positional coincidence with such a bright star, the hypothetical white dwarf should be physically linked to the B5 star and at a distance of 190 pc the observed count rate is compatible with this hypothesis (see e.g. Barstow et al. 1994). The expected 'normal' harder X-ray emission from the early type star should give a count rate 200 times smaller than observed and virtually be undetectable in our data.

LS IV -12 70 (Group 1): A source located at a position consistent with that found in the survey was re-detected during the PSPC pointing WG 500040. However, the pointed position is incompatible with the B star and a red excess object now lies in the middle of the revised error circle.

HD 38023 (Group 2): This position was not optically investigated. The B4V star HD 38023 is located in Orion inside a reflection nebula. The X-ray source was detected again during a long PSPC pointing with a count rate similar to that recorded during the survey. HD 38023 now falls within the 90% confidence radius of the pointed position. It appears as the brightest spot of a patch of several nearby sources. RX J0542.3-0807 is variable between 0.01 and 0.04 cts s^-1 on a time scale of two days and displays a soft energy distribution. Fixing the interstellar absorption to the value derived from the optical colours ( = 3.7 10²¹ cm^-2) a thin thermal Raymond-Smith spectrum fit yields kT = 0.85 0.12 keV indicating that the PSPC energy distribution is compatible with that expected from an active corona. The corresponding mean un-absorbed 0.1-2.4 keV X-ray luminosity of 1.7 10³¹ erg s^-1 is consistent with the luminosities derived by running the automatic process on survey data. Considering the X-ray luminosity and softness of the source the odds are more in favour of a young late type star physically related to HD 38023 rather than for a genuine accreting binary.

LS V +29 14 (Group 2): This position was not optically investigated. Comparison of POSS O and E plates reveals the presence close to the survey X-ray position of at least two faint red candidates which could be Me stars.

LS 5038 (Group 2): The field of LS 5038 was covered by a long PSPC pointed observation. A weak source (1.8 10^-3 cts s^-1, i.e. about 10 times fainter than the survey count rate) is detected at a position probably compatible with the survey position. The source detected during the pointing is now farther away () from LS 5038 than was the survey position ruling out an association with the early type star. The pointed position is closer () to the bright (V = 8.0) K0IV star HD 169651, however, the late type star is still outside the 90% confidence radius. Our optical spectroscopic observations failed to identify any active corona in the close neighbourhood of the early type star and revealed that LS 5038 is a Be star (EW(H ) = -4.4 0.3 Å) exhibiting a central H absorption core similar to that seen in LS 1698 and LS 992. The 6284 interstellar band indicates E(B-V) 0.85. Our low resolution spectroscopic data do not allow accurate spectral type determination. We cannot completely rule out the unlikely possibility that we have witnessed a hard X-ray flare from an accreting object around LS 5038 during the survey observation.

HD 36262 (Group 3): HD 36262 is another early type star in Orion displaying an X-ray excess. With a B3V spectral type and an un-absorbed X-ray luminosity of 3 10³¹ erg s^-1 (using EXSAS results), the most likely explanation of the X-ray excess is the presence of an extremely active companion star. The hardness ratios are compatible with a rather soft stellar source.

5.3. Likely non-accreting sources

5.3.1. LS III +46 11 (Group 1)

This OB star is a probable member of the Berkeley 90 cluster (Sanduleak 1974) and located within the HII region Sharpless 115 (Harten & Felli 1980). LS III +46 11 is well placed within the ROSAT survey error circle. Spectroscopic observations of all nearby objects marked on Fig. 23 failed to reveal any convincing alternative optical counterpart leaving the OB star as the most likely identification of RX J2035.2+4651.

Fig. 23. B CCD image of the field of LS III +46 11 obtained on 1992 May 27 with the RCA3 CCD camera and the 1.2 m telescope at OHP. We show the 90% confidence ROSAT survey error circle and the position of the alternative candidate stars for which we obtained optical spectroscopy. North is at top and East to the left. The field is 3.54 3.54 wide

Medium resolution ( 3900-4400 Å ; Fig. 24 and 6300-6700 Å ; Fig. 25) and low resolution ( 3800-7100 Å ; not shown) optical spectroscopy indicate a very hot giant star. The He II 4541 / He I 4471 and the He II 4025 / He II 4200 line ratios both indicate an O3-O5 spectral type (see Fig. 24). The probable presence of N IV 4058 emission also visible in our low resolution spectrum and possible N V 4606-4620 absorption suggest a class III luminosity. Broad H emission is conspicuous and indicative of the high luminosity of the star (see Fig. 25). We thus propose an O3-O5III(f)e spectral type for LS III +46 11.

Fig. 24. Rectified blue medium resolution spectrum of LS III +46 11 obtained with the OHP 1.93 m telescope and the CARELEC spectrograph. The spectrum shown here is the mean of 4 spectra collected in the time interval 1991 November 20-24. Total exposure time is 100 min

Fig. 25. Red medium resolution spectrum of LS III +46 11 obtained with the OHP 1.93 m telescope and the CARELEC spectrograph on 1992 December 15. A broad H component (FWHM 60 Å) is visible and indicative of the high stellar luminosity

In a subsequent dedicated pointed observation LS III +46 11 was again detected with a count rate consistent with that recorded during the survey (0.046 cts s^-1). The nearby O6 star LS III +46 12 was also detected with a count rate of 0.016 0.004 cts s^-1.

In this particular case, the assumption of a default B0 III spectral type by the automatic SIMBAD/ROSAT correlation process led to a clear overestimation of the / ratio. Using an O3 III spectral type now gives / 2.7 10^-6. With this spectral type the star is at 1.9 kpc, in agreement with the distance of the Be 90 cluster and the 0.1-2.4 keV un-absorbed luminosity is of the order of 5 10³³ erg s^-1. We also note that the ratio of the PSPC count rate from LS III +46 11 over that of LS III +46 12 is similar to the ratio of their bolometric luminosities. This implies that most of the ionizing power in S 115 is actually provided by LS III +46 11 rather than by LS III +46 12. Although extremely rare, such high X-ray to bolometric ratios and soft X-ray luminosities have been reported for a few O type stars (Sciortino et al. 1990). We therefore conclude that there is probably no need for an accreting component in order to explain the X-ray emission of LS + 46 11. However, considering the extreme value of the X-ray luminosity this star would certainly deserve further detailed investigation.

5.3.2. Other sources

HD 59364 (Group 1): RX J0728.6-2629 is probably identified with a red Me star (object C in Fig. 26) which exhibits strong Ca II H&K emission and Balmer emission.

Fig. 26. B and I CCD images of the field of HD 59364 obtained on 1992 April 18 with EFOSC2 and the ESO-MPI 2.2 m telescope. Both images have 2 min exposure time. North is at top and East to the left. Each frame is wide. On the B image we plot t he ROSAT survey 90% confidence error circle and show the position of the objects investigated spectroscopically. We identify the X-ray source with the red Me star C

HD 67785 (Group 1): RX J0807.2-5053 is most likely identified with object "B" (see Fig. 27). This late K-M star displays Ca II H&K emission consistent with the X-ray flux.

Fig. 27. U and I CCD images of the field of HD 67785 obtained on 1992 April 17 with EFOSC2 and the ESO-MPI 2.2 m telescope. Both images have 30 s exposure time. North is at top and East to the left. Each frame is wide. On the U image we show th e position of HD 67785. On the I image we plot the ROSAT survey 90% confidence error circle and show the position of object B which is the actual optical counterpart of RX J0807.2-5053

HD 165424 (Group 1): Follow-up pointed PSPC observations give an improved position only away from the V = 9.0 K0 star HD 315203 and the former proposed B4II identification HD 165424 is now outside the ROSAT error circle. Systematic attitude errors are unlikely to be large for this pointing since we detect the X-ray emission of the symbiotic star He 3 -1591 only away from the optical position. We conclude that RX J1806.8-2606 should rather be identified with the late type star HD 315203.

BD + 60 282 (Group 2): RX J0136.7+6125 is probably identified with the V = 10.7 star GSC 0403100953, away from the ROSAT survey position. Optical spectroscopy shows that GSC 0403100953 is a late type star displaying strong Ca II H&K emission consistent with the measured X-ray flux.

LS I +61 298 (Group 2): RX J0234.4+6147 is probably identified with the V = 14.1 red excess object (GSC 0404700465) located away from the survey position. Red medium resolution spectroscopy reveals the presence of narrow H emission superposed on a deep absorption profile indicating that GSC 0404700465 is a late F or G type active star. The source was detected again during an AO3 pointing with a count rate similar to that of the survey observation. The pointed position is closer to the B star () but still compatible with the emission line candidate star ().

LS 3122 (Group 2): The Me star (object C in Fig. 28) is the likely counterpart of RX J1335.5-6211. The star is located well inside the 90% confidence error circle and medium resolution red spectroscopy revealed strong H emission (EW 6.0 Å).

Fig. 28. B and I CCD images of the field of LS 3122 obtained on 1992 April 17 with EFOSC2 and the ESO-MPI 2.2 m telescope. Both images have 2 min exposure time. North is at top and East to the left. Each frame is wide. On the B image we plot t he ROSAT survey 90% confidence error circle and show the position of LS 3122. On the I band image we show the position of object C, the Me star optical counterpart of RX J1335.5-6211

SS 73 49 (Group 2): The emission line star Wray 15-1400 = SS 73 49 was discovered by Wray (1966) and is identified with GSC 0784600316. Using objective prism low resolution spectra Sanduleak & Stephenson (1973) classify the object as a Be type star. Our medium (Fig. 29) and low resolution (Fig. 30) spectra show strong Balmer and Ca II H&K emission. We classify SS 73 49 as a strong line T Tauri star rather than Be. Ca II H&K emission is unusual in Be stars and there are evidences for Mg b and TiO 6187-6215 bands in absorption suggesting a late K star underlying continuum. The Ca II H&K flux is consistent with that expected from the PSPC count rate for active coronae. The survey hardness ratios are also compatible with those normally observed from young active stars. Interactive analysis of the survey photons yields a position slightly closer to the active star at the edge of the 95% confidence error circle. We conclude that RX J1559.2-4157 is optically identified with the T Tauri star SS 73 49.

Fig. 29. Flux calibrated blue medium resolution spectrum of SS 73 49, the proposed optical counterpart of RX J1559.2-4157. The spectrum was obtained on 1992 April 20 with EFOSC2 and the ESO-MPI 2.2 m telescope. Exposure time is 10 min

Fig. 30. Flux calibrated low resolution spectrum of SS 73 49, the proposed optical counterpart of RX J1559.2-4157. The spectrum was obtained on 1992 April 20 with EFOSC2 and the ESO-MPI 2.2 m telescope. Exposure time is 5 min

LS IV -05 35 (Group 2): RX J1900.7-0503 is most probably identified with the V = 12.2 star GSC 0513601711 located within the 90% confidence error circle. Optical spectroscopy reveals noticeable Ca II H&K emission from GSC 0513601711.

© European Southern Observatory (ESO) 1997

Online publication: May 26, 1998

helpdesk.link@springer.de