3. Gould Belt or Gould Disk ?
Following ideas outlined in the literature (see e.g. Blaauw 1991, Pöppel 1997), let us assume the geometry of the GB projected onto the galactic plane as a rim outlined by the solid thick curve in Fig. 1. In this projection the GB is an ellipsoid with semi-major and minor axes of 500 and 340 pc respectively. The center of the GB is located 200 pc from the Sun towards l = . This off-center position of the Sun with respect to the entire structure is of major importance in the following. Since some parts of the GB are closer to the Sun than others, we expect a strong galactic longitude dependence for our ability to detect GB members.
Let us assume a G0-G5 main sequence star as a representative element of the RasTyc sample (mainly consisting of young ( 1Gyr) F, G, and a few K (M) stars). Assuming no interstellar absorption, the maximum distance out to which such a star can be detected in the Tycho catalog is D = 150 to 200 pc. We thus define DTyc = 175 pc as the mean optical horizon of the RasTyc sample.
According to current stellar X-ray population models (Favata et al. 1992, Micela et al. 1993, Guillout et al. 1996a) the stellar content of soft X-ray survey is a mixture of stars with X-ray luminosity LX in the range erg s-1. The youngest stars are more X-ray luminous and therefore preferentially detected. An adopted age of the GB between 30 to 80 million years implies X-ray luminosities in the range erg s-1 for GB members.
We now discuss the detectability of the GB in the RasTyc sample at low (Sect. 3.1) and then high (Sect. 3.3) sensitivity. We wish to discuss both thresholds in order to introduce in a logical and understandable fashion the proposed, favourable Gould Disk scenario (Sect. 3.2). The comparison of observations at these two sensitivities allows us to draw constraints on the geometry of the GB.
3.1. Low sensitivity (Sthr = 0.10 cts s- 1)
Let us start with the discussion of the 2 140 RasTyc stars detected in the RASS above a PSPC count rate threshold of Sthr = 0.10 cts s- 1. We adopt the following notation: GBl refers to regions inside the GB strip of the quadrant Ql centered on galactic longitude l.
In Fig. 1 we also plot the maximum distance DX (our "X-ray horizon" around the Sun) up to which we can detect stars with an X-ray luminosity of log(LX) = 29.5, 30.0, and 30.5 above Sthr = 0.10 cts s- 1. Using an absorbing column of = cm-2 and an X-ray temperature of = K for the emitted spectrum, we obtain radii of DX = 50, 90 and 160 pc, respectively. Note that this choice of PSPC threshold implies that the X-ray horizons are closer than the optical horizon set by the magnitude limit of the Tycho catalog. From Fig. 1 it is apparent that our low sensitivity horizon hardly reaches the GB feature in any direction except towards Q330, i.e., the near side of the GB. Even here we would not expect to find any significant stellar density enhancement (with respect to mean galactic plane) because only the few very X-ray brightest GB members could be detected at this threshold.
3.1.2. Stellar surface density
In the upper panel of Fig. 4 (see color plate) we plot the sky positions of the RasTyc stars brighter than 0.10 cts s-1 detected at low galactic latitude ( ). The color coding refers to the stellar number density per square degree. Since the precise location of the GB is somewhat uncertain in the literature, we took a wide strip with an inclination i = and ascending node = for the GB (see e.g. Comerón et al. 1994, Guillout et al. 1998a). It is obvious from Fig. 4 that GB330 (the direction where the GB structure is expected to come closest to the Sun) shows a significant stellar density enhancement. It seems to extend to the adjacent quadrant GB240, but it does not show up symmetrically at GB60. Apart from the Hyades feature unrelated to the GB, GB150 does not display any obvious enhancement in stellar density.
3.1.3. Distance distribution
In order to distinguish the stellar population belonging to the GB structure from the ambient galactic plane component we compare the distributions in distance D and X-ray luminosity LX (see next subsection) for stars inside and outside GB regions. We analyze the data independently in the four 90 degrees wide quadrants defined above, in order to account for a possible direction dependence in our data, but at the same time retain sufficiently large sample sizes. For 54% of the 2 140 active stars detected in the galactic plane above 0.10 cts s-1 we derived accurate distances from RasHip or RasTyc parallaxes (relative error 30%). Since RasHip invariably provides higher precision optical data than the corresponding data from RasTyc we chose RasHip data whenever available. Nevertheless the two color Tycho magnitudes represent independent and reliable photometric informations. For stars with 0.3 we calculated photometric distances Dphot. Dphot was computed assuming a standard (B-V) - Mv calibration for main sequence stars. This is not a critical point since we are mainly dealing with F-G type stars which reach the ZAMS in a relatively short time scale. In the middle panel in Fig. 4 we plot the histograms for the distance distributions (normalized to stars per square degree) in each of the four quadrants defined above. The red histogram refers to stars located in the GB region, the blue histogram refers to stars outside the GB region (but within of the galactic plane). Since the analyzed areas are confined towards the galactic plane, we take them to be representative for a typical galactic plane distribution. The black histogram identically shown in all four panels is constructed as the mean of the four blue quadrant histograms and represents the typical galactic plane distribution.
Despite of poor statistics, the GB330 D histogram (red) clearly exhibits an excess of X-ray emitting stars with respect to mean galactic plane histogram (black). The GB240 D distribution shows a less pronounced but similar feature and thus confirms our previous suspicion based on visual inspection (Sect. 3.1.2). Surprisingly and interestingly, in both quadrants the stars responsible for this excess are distributed from the solar vicinity up to distances of 180 pc! On the other hand, the histograms referring to regions within the GB (red) or the galactic plane (blue or black) are practically indistinguishable towards GB60 and GB150. The large deviation in GB150 is the Hyades cluster, which clearly shows up between 40 to 50 pc; the RASS data on optically selected Hyades members have been presented by Stern et al. (1995).
3.1.4. X-ray luminosity distribution
In the lower panel of Fig. 4 we display for each quadrant the X-ray luminosity distribution calculated using the previously computed distances and a flux conversion of 110-11 erg/PSPC count. Independent of quadrant, the galactic plane region LX histograms show a broad distribution from LX = 1028 to 1031 erg s-1, as expected for a mixture of stars mostly younger than a billion years. On the other hand, stars accounting for the enhancement towards GB330 and GB240 are preferentially detected at high X-ray luminosities, thus favouring the interpretation that these stars are on average younger than those detected anywhere else in the galactic plane. Galactic plane and GB regions are again indistinguishable from the X-ray luminosity point of view towards GB60. Stars from the Hyades cluster are responsible for the peak around LX erg s-1 while the significance of other features is too low at this threshold.
3.2. The Gould Disk scenario
The detection of a significant surface density enhancement in the GB330 quadrant above a PSPC count-rate level as high as Sth = 0.10 cts s- 1 is surprising. Most intriguing, however, is the radial distribution of this excess population which is essentially continuous along the line of sight in quadrants GB330 and (to a lesser extend) in GB240. These features are not expected from our initial assumption that GB members are located at the outer rim of the GB, where prominent SFRs and OB associations are located. The X-ray luminosity distribution, skewed towards higher X-ray luminosities, strengthens the interpretation that these stars tend to be younger than those representative for the galactic plane population at large.
Now, let us alternatively imagine a slightly modified scenario in which the Gould Belt is not a belt but rather a disk or even better a disrupted disk composed of young stars with a smaller inner radius, as indicated in Fig. 1 with the grey shaded area. In this case, our low sensitivity expectations would be similar towards GB60 and GB150 but drastically different towards GB330 and GB240, where the GB should show up prominently and marginally, respectively. This scenario would also explain the wide radial distance distribution of GB excess stars. Thus, at low semsitivity, the picture of a Gould Disk can reconcile expectations and observations in all four quadrants much better than the original Gould Belt scenario.
3.3. High sensitivity (Sthr = 0.03 cts s- 1)
In order to test the modified GB picture and to check its consistency with our data we now discuss the 6 348 stars detected as X-ray sources brighter than Sthr = 0.03 cts s- 1, the deepest threshold allowed to avoid biases due to RASS exposure time (see Guillout et al. 1998b). In order to allow the reader a clearer conception of the geometry and the biases at that count rate threshold, we illustrate in Fig. 2 the proper "X-ray horizon" now reaching out to 95, 165 and 300 pc for log(LX) = 29.5, 30.0, and 30.5 respectively.
The basic expectations regarding surface density enhancements are similar to the case of lower sensitivity : the GB should clearly show up towards GB330 and GB240, but should remain undetected towards its far side GB150. In principle Sthr = 0.03 cts s- 1 should be sufficient to detect GB members towards GB60 in case of the disk scenario. However, in contrast to the low sensitivity case, DTyc sets the horizon; thus the sample is now optically limited implying that GB members can only be marginally detected even in case of a disk due to the Tycho threshold. We thus note that it might be difficult to discriminate between the Gould Belt or Gould Disk picture by considering the expected surface density excesses alone. However, a clear discrimination between these two scenarios is expected to arise from the distance distribution along the line of sight towards the two quadrants at the near side of the GB.
3.3.2. Stellar surface density
In the upper panel of Fig. 5 (see color plate) we plot - similarly to Fig. 4 - the sky positions of the RasTyc stars brighter than 0.03 cts s-1. Now the density enhancement towards GB240 is much more prominent than in the case of a higher threshold as expected from the general geometry of the GB. Apart from well known features like the Hyades cluster and the Taurus Auriga SFR, GB and GB do not display any obvious stellar density enhancements. Other young clusters and SFRs like IC2602, IC2391, Orion, Chamaeleon and the Ophiuchus-Scorpius-Lupus-Centaurus complex do obviously contribute to the enhancement towards GB and GB, but cannot account for the whole feature, which is spatially much more extended. As a test we constructed stellar surface density, distance and X-ray luminosity distributions excluding areas with a radius of 2 to around all known young galactic clusters (IC2391, IC2602,...) and SFRs (Ophiuchus, Scorpius, Lupus,...) but found the enhancement still present (albeit slightly reduced), thus confirming that clusters and SFRs (in particular Oph-Sco-Cen) located near the outer boundary of the GB cannot totally account for the observed stellar density enhancement . On the other hand, GB and GB display some clustering which is not associated with identified young clusters and which could be interpreted as residuals of 'old' clusters much more dispersed now. However, a deeper analysis is needed to confirm their spatial coherence.
3.3.3. Distance distribution
In the middle panel of Fig. 5 we show the distance distributions towards the four quadrants inside and outside the GB region (as in Fig. 4). Our conclusions derived at low sensitivity are fully confirmed at this deeper threshold; the D distribution again shows a 30 - 50% higher density extending from 30 pc to about 180 pc from the Sun towards GB and as far as 300 pc towards GB. GB and GB now show extremely small deviations from galactic plane D distribution. No detection of GB towards GB is possible since the GB is too far to be detected even in case of the disk scenario. The Hyades stars give a prominent signal at this threshold.
Interestingly, the comparison of distance distributions towards GB at low and high sensitivities shows that the enhancement does not extend further than about 180 pc noted already for the lower sensitivity case, despite the fact that the X-ray horizon moves to larger distances because of the higher sensitivity. In other words, a threshold of 0.10 cts s-1 was already sufficient to reach the true outer boundary of the disk, and a deeper sampling of distances does not find evidence for any additional excess beyond the assumed outer edge of the GB structure . The situation is, however, different towards GB, where Sthr =0.03 cts s- 1 is still insufficient to detect the true outer boundary of the disk. These features are naturally explained by the GB geometry and the off center position of the Sun with respect to the structure.
3.3.4. X-ray luminosity distribution
Again, our conclusions drawn from lower sensitivity are confirmed. Stars accounting for the enhancement towards GB and GB show values in the range of erg s-1, in agreement with what we expect for 30 to 80 million years old stars belonging to the GB . The Hyades cluster now shows up even more prominently and accounts for the enhancement between to 30.0 while the Pleiades and Taurus-Auriga SFR responsible for the small enhancement around 130 and 160 pc can explain the excess above .
© European Southern Observatory (ESO) 1998
Online publication: August 6, 1998