Several low-resolution UV-spectra of BN Ori have been obtained
with the International Ultraviolet Explorer (IUE) satellite in 1984 and
1986. Details of the observations are given in Table 5 (click here).
Before each exposure the brightness of BN Ori was measured by the
Fine Error Sensor (FES) of IUE in the Field Camera mode (Barylak 1989).
In all cases the resulting 
magnitude was around 968, which is
close to the results of the ground-based photometry in the periods around the
IUE observations (Tables 3 (click here) & 12 (click here)).
The low-resolution UV-spectra of BN Ori, HR 5999 and several
other Herbig Ae stars show strong emission lines of OI, CII, SiIV
and CIV (Tjin A Djie et al. 1982; Brown et al. 1986) shortward 1600Å.
The high intrinsic radiation flux in these lines can probably only be
supplied by accretion (Tjin A Djie & Blondel 1997)
and it has been possible to derive information on the accretion parameters
of A and F stars from the UV-continuum over the full wavelength range
(1150- 3200Å) of IUE (Blondel & Tjin A Djie 1997). After the choice of a
stellar spectral type and luminosity-class and addition of the observed
visual magnitude, the distance and the foreground colour excess, we have
matched the observed UV-continuum with standard star spectra from the IUE
spectral atlasses for the photospheric and boundary-layer components and a
optically-thick model for the accretion-disc (UV3C, Blondel & Tjin A Djie 1994, 1997).
Table 7:
Parameters
used in the UV3C-model calculations with the low-resolution UV-spectra.
For the B9I ab component we used 
= 10300K (Schmidt-Kaler 1982).
The value given by {
is the interstellar (foreground) E(B-V), while the
value is the total excess
The results of the parameter adjustments required for the matching of the
low-resolution UV-spectra of BN Ori, HR 5999, BF Ori and FU Ori are
collected in Table 7 (click here).
The masses of the stars (
) were estimated from the stellar radii
(
derived from the model calculation) and the photospheric 
(Schmidt-Kaler 1982) with the use of the evolutionary tracks of
Palla & Stahler (1993).
With this mass and radius, the width of the boundary layer (
)
and its effective temperature, the corresponding mass-accretion rate (
)
was derived.
Since the mass of FU Ori cannot be derived from the evolutionary tracks,
we have assumed that its precursor was a 1.0
T Tauri star.
In general the stellar spectral type and luminosity-class appear
to be strongly constrained (to 1 or 2 subtypes) by the observed UV-continuum.
For BN Ori we estimated a distance of 
400pc from the data of Artyukhina
(1959) and a foreground E(B-V) = 007 from the NaI D interstellar
components.
Figure 8:
Low-resolution UV-spectrum of BN Ori after correction for the foreground
extinction (grey line) with the accretion model calculation (black line).
The used model parameters are listed in Table 7 (click here), and the "W"
indicates the Walraven W-band flux (Table 2 (click here)) after
correction for the foreground extinction.
The low S/N in the observed spectrum (2000Å- 2400 Å) is due to the
low sensitivity of the LWP camera in this range
Figure 9:
SED of BN Ori. The fluxes derived from the photometry of Sect. 2 (click here)
are indicated by squares. The (*) points correspond to the fluxes
calculated with the UV3C-model (Table 7 (click here)), obtained by matching
the low-resolution UV-spectrum (Fig. 8 (click here)). The K- and L-band
fluxes (open star) show the effect of removing the disc contribution in the
model-calculation. Circles indicate the extiction-corrected model results
The calculated spectrum and the SED corresponding to the
parameters of BN Ori in Table 7 (click here) are given in Figs. 8 (click here)
and 9 (click here). These results show that the UV3C-model gives a good
agreement with the observations from the far-UV up to the L-band in the NIR.
In general the model (with an optically thick disc up to 40
)
gives higher values than observed for the fluxes in the K- and L-band.
The disagreement could not be reduced by taking the disc optically thin
but when we limit the outer radius of the disc to 2.0- 2.5
the NIRE
almost disappears (Fig. 9 (click here)). This suggest that the disc
has been largely dissipated during the outburst of 1947.
For the parameters in Table 7 (click here) the E(B-V) is 012, which with
an observed 
045 corresponds with 
= 033. This
value is slightly higher then the 030 for spectral type F0III given by
Schmidt-Kaler (1982) but not unacceptable as BN Ori may be not exactly F0III.
Both the small circumstellar colour excess (005) and the negligeable NIRE
indicate that BN Ori has almost no circumstellar dust left.
Table 8:
Identification, EW (Å) and FW (
) of lines in the high-resolution
UV-spectra of BN Ori, HR 5999 and FU Ori.
See Table 10 (click here) for the used symbols
In addition to the continua of the UV-spectra of BN Ori, we have also
information on the individual lines from the high-resolution LWP image
obtained with IUE.
In spite of its long exposure (13
) the S/N ratio is
very low in the less sensitive part of the LWP camera (shortward 2400Å)
but the region longward 2560Å is better exposed. Since the three low
resolution UV-spectra of BN Ori show no differences (although they were
taken at different times) we assume that the high-resolution spectrum of
Nov.13, 1986 is representative for at least the period Jan'84- Nov'86.
The FES-magnitude (
) during this observation
agrees with a simultaneous ground-based photometric observation from DASA.
From the comparison of the high-resolution UV-spectra of BN Ori, FU Ori
and HR 5999 we can draw the following conclusions:
30%) of the EWs of the lines in
the FU Ori spectrum. Table 8 (click here), in which the EWs and FWs of the
three stars are listed, shows that the EWs and FWs of the BN Ori lines are smaller
than those of the corresponding lines in the spectrum of FU Ori, and both
are smaller than the EWs and FWs of HR 5999.
Since the spectral types of the three stars are not too different we expect
that the column densities in the shells will also decrease from HR 5999
through smaller values for FU Ori to almost negligible values for BN Ori.
from its heliocentric position (determined from the NaI D\
interstellar lines) and a circumstellar one (with EW = 0.09Å) with a
velocity of -53
with respect to the interstellar one (Fig. 11 (click here)).
Similar to the NaI D lines the continuum in the neighbourhood suggests that
there exists a third, broad (FW = 
400
) and shallow component,
centered at the location of the two narrow components.
Because of its width this component could be formed either at the stellar
photosphere or near the inner accretion-disc.
The MgI line of FU Ori also appears to be double. Both
components have an EW of about 0.26Å and the circumstellar component has a
velocity of 
-67
with respect to the interstellar component.
The EW of the circumstellar component is larger than the corresponding one
for BN Ori, while the foreground interstellar components of FU Ori and BN Ori
are of comparable strength.
The MgI components in the UV-spectra of HR 5999 are not resolved
because of the large width of the circumstellar component. Its EW is about
10 times larger than the upper limit of the EW of the corresponding component
for BN Ori and 3- 4 times the corresponding EW for FU Ori.
of 
200
and a very weak shell component in the MgII
profiles.
Figure 10:
Observed MgII h&k profiles of BN Ori (LWP 9512) and the residual profiles after
normalisation to the 21 Vul (LWR 9037) profiles. Top-axis give the velocity
(
) relative to the wavelengths of the MgII h&k lines
The result is given in Fig. 10 (click here), which shows that the shell
component of MgII has a P-Cygni profile with outflow-velocities up to
-250
. We have made a rough estimate of the MgIIk emission-component
(EW 
15Å) by assuming that the intrinsic emission (without the
blue absorption trough) is symmetric.
With the continuum flux near 2790Å from the low-resolution UV-spectra of
BN Ori (2.5 
), the distance to radius ratio from the UV3C-model
(4.35 
) and the total extinction correction (a flux factor of 2.08) we
derive an extinction-free surface flux of 8.8 
for the
emission-component of the MgIIk line of BN Ori.
In the same way we derive from the corresponding profile of HR 5999 in the
phase of maximum visual brightness (Blondel et al. 1989) an extinction-free
surface flux of 6.45 
(
= 68) and 1.2 
(
= 76).
The ratio of the MgIIk emission fluxes of HR 5999 and BN Ori is therefore
close to 70- 140. This is higher than the ratio of the H
fluxes of
the two stars (Sect. 4 (click here)).
Figure 11:
MgI 2852Å components of BN Ori
Table 9:
Comparion of emission fluxes (
)
The high-resolution MgII profiles of FU Ori have been published by
Ewald et al. (1986). The profiles have a P-Cygni shape, but the noise in the
neighbouring continuum does not permit to determine an outflow velocity.
The extinction-corrected surface flux in the symmetrised MgIIk
emission-profile is 8.6 
and the corresponding flux of
the MgIIh emission is 7.1 
. This total MgII emission
flux of FU Ori is close to the value obtained by Ewald et al.
With the use of the observed MgII profile of the F0I b star
Car (Praderie et al. 1980), which has a 
= 0 
, we
corrected the profile of FU Ori for the photospheric contribution.
The extinction-corrected surface-flux of the symmetrised MgIIk-profile
then becomes 1.67 
. Table 9 (click here) gives a survey of the
emission fluxes of the 3 stars.
Although the H
and MgII profiles show variations on a short
time-scale (
1
) the time variations in the integrated emission
fluxes are small. This gives us a justification to use Table 9 (click here)
in the following discussion, in spite of the fact that the H
and
MgII fluxes were not observed simultaneously.