9.1 Be Stars

In the field of hot-circumstellar environments one problem has continued to remain enigmatic; what happens when a star rotates at high speeds? The answer may lie in examining the fast rotating Be stars. Previous studies generally observe individual objects in selected passbands, however this approach is unable to give an overall picture of the Be phenomenon. To overcome this, this thesis has continued a multi-wavelength programme of photometric observations of a “representative” sample of Be stars. The sample is “representative” in that it contains several objects typical of each spectral and luminosity class for which the Be phenomenon occurs, it therefore does not reflect their space density, but instead aims to determine the average properties of each class. Previously, spectroscopic observations of the sample have been obtained and this data complements the new data presented in this thesis.

Chapter 1 covered the current status of Be star research. From the available literature it was able to conclude that the IR excess exhibited by a Be star is most likely caused by free-free and free-bound emission in a circumstellar disc. However the mechanism creating the disc is unclear. The viscous disc model has gained popular support as it is able to reproduce most of the attributes of observed Be star discs. However this model requires an angular momentum input to the disc and as yet this facet of the model has remained elusive.

Chapter 6 entertained the possibility that an optical excess existed in Be stars and could be separated from the intrinsic stellar flux. New optical Strömgren data of the representative sample was presented and reduced using the PL. A separation was attempted using the method outlined by Fabregat and Torrejon (1998). This method relies on the fact that the colour of the Strömgren filters is reddening free. Previous to the data analysis the physics of an optical excess was discussed, this showed that in the regime of an optically thick emission line and an optically thin continuum it is possible for the equivalent width to increase with circumstellar excess. However it became evident that whilst the method produced results agreeing with theory the errors on those results were too large to allow further work to be carried out and a more accurate method needed to be sort. The data was obtained during good conditions and was believed to be good quality.

In light of this, Chapter 7 presented a new method of separating the component parts of stellar flux from Be stars. The method entailed using observations of intrinsic B stars, an interstellar reddening law (Rieke and Lebofsky, 1985) and a derived equation which governs the circumstellar excess in the observed bands. using these components it was possible to derive the interstellar reddening and circumstellar excess of each object. The method was validated by implementing it on new JHK band IR data taken using the Carlos Sanchez Telescope (TCS). These data were also combined with existing and complementary representative sample data. Through this technique it has been possible to show that significant correlations between the infrared excess emission from the disc and emission from a range of lines (H, Br, Br11, and Br18) exist. The IR excess was also demonstrated to correlate with the stellar v sin(i) and sin(i). The correlations were derived using non-parametric tests (Spearman rank), which were employed because the form of the correlation, if any, to be obtain was unknown and in this way no pre-determined result was forced upon the data.

Chapter 8 returned to the Strömgren data sample of Chapter 6 and the new separation technique determined in Chapter 7 was implemented. The re-analysis showed that it was possible to produce smaller errors, although the errors were still large enough that it was not possible to progress further with the data. The reason for this is believed to be that the optical circumstellar component is too small to be separated out from the intrinsic stellar flux without very accurate photometric observations. Strömgren photometry therefore may not be an accurate and robust approach to studying Be stars.

To overcome this problem longer wavelength IR photometry (10 - 20m) has been approved and will be obtained in July 2002 using the new Michelle instrument on UKIRT. Michelle data is necessary to pin-down the density field of the disc. The IR excess is generated from free-free and free-bound processes, however it has been suggested that some Be stars also show a contribution from warm circumstellar dust. These observations will therefore enable the determination of whether the disc density is sufficient to allow dust to form if the discs cool to less than ~ 2000K.

The data and results presented in this thesis are paramount in knowing how to constrain theory, so that the physics of rotating stars may be understood. Conti  (1976) suggested that those B-type stars which show forbidden emission lines should be classified as B[e], following the notation for forbidden lines. Forbidden lines are those emission lines that are not normally observed under terrestrial conditions. They occur because in locations of low density the electrons, which would normally be collisionally excited to a higher energy state, are free to decay - exhibiting a forbidden line. Supergiant B[e] stars exhibit an IR continuum believed to be due to circumstellar discs. The method which creates this disc has caused much debate, however Porter (2002) has shown that the viscous disc model appears to be the best candidate. This is fundamental in linking the main sequence Be stars with supergiant B[e] stars - indeed does this indicate that B[e] stars are rotating at close to their critical velocities? It remains to discover the method by which angular momentum is passed from the stars to the disc - and whether it is the same in both cases.

It has been shown in this thesis that the velocity structure of Be star discs is correlated to the IR excess.

By studying the disc it is hoped to understand the reasons why it should be that ~20% (see Jaschek and Egret, 1982; Maeder et al., 1999, and references therein) of B stars develop into Be stars. Is it due to a natal effect? The accretion of matter at the equator during the early stellar stages often results in bipolar outflows from young stars. This process allows the excretion of matter and the venting of angular momentum. The time-scales over which the rotation and angular momentum are homogenized over the whole star may be slower than it takes to accelerate an equatorial belt to critical velocity.

If this is the case it could be possible to form an equatorial excretion disc. Once the belt had lost sufficient angular momentum to slow below its critical velocity the matter would be able to re-accrete. Unless the matter in the Be star disc becomes completely unbound from the star-disc system then the angular momentum conserving nature of the viscous disc model, the current most likely formalisation of the disc, ensures that the re-accretion of matter to the star will return all the disc’s angular momentum to the star. This would allow for the appearance and disappearance of the disc.