Skip to main content

Full text of "IUE observations of the 1982-84 eclipse of Epsilon Aurigae"

See other formats


Thomas B. Ake 

Astronomy Operations , Computer Sciences Corporation 


We summarize the major characteristics in the ultraviolet of the 
198 2-84 eclipse of Eps Aur as observed with lUE by various workers . 
This star can be observed over the entire lUE wavelength range, from 
1200 to 3200 A, in low dispersion, allowing eclipse light curves to be 
obtained in broadband regions , but due to its steep spectral gradient 
and the sensitivities of lUE cameras , high resolution exposures 
adequately cover only the regions from 1700-1900 and 2400-3200 A. In 
many ways, the UV data confirms or expands upon interpretations of the 
system made from observations in other wavelength regions, but in other 
respects the system remains as enigmatic as before. 


The flux from an late-A, early-F supergiant like Eps Aur drops over 2 
orders of magnitude from 3200 to 1200 A. Thus it is necessary to take 
multiple exposures of different length to adequately cover all 
wavelengths. For Eps Aur in low dispersion, typically 2 exposures with 
each camera are needed, ranging from 7 sec to 20 minutes. In high 
dispersion, two LWR images with an exposure ratio of 4 are needed to 
optimize exposure levels in the continuum and at Mg II; for the SWF, a 
full shift exposure only extends to about 1700 A. Unfortunately this 
precludes studying astrophysically interesting high temperature lines 
such as C IV and Si IV if they are present. 

The characteristics in low dispersion can be summarized as follows : 

1« The eclipse light curve in the near UV generally follows that 
found in the optical region. During totality, the eclipse 
depth slowly increases up to third contact. The minor 
fluctuations in light seen optically are increasingly 
exaggerated in the UV from 3200 A down to about 1500 A. 
Shortward of this, the fluctuations become smaller in 
amplitude. The fluctuations occur predominantly prior to 

2. The eclipse depth is dependent upon wavelength, increasing 
somewhat in depth from 3200 A to 1600 A then becoming 
shallower such that at 1200 A it is only 0.2 mag. deep. 

Guest Obse.rver, International Ultraviolet E^cplorer 


3. Two extraordinary brightenings were seen in the UV near first 
and third contacts. The occurrance at first contact has 
caused various lUE observers to find different depths of 
eclipse because different reference spectra were used. The 
brightening seen optically at mid-totality is also visible in 
the UV, the degree being in the same ratio as the optical/UV 

4. The 2200 A depression characteristic of interstellar grain 
absorption is consistent with published values of E(B-V) 
-0.35. There is no detectable change in the dip during the 

5. Compared to other A-F supergiants, there is a UV excess 
shortward of 1400 A. The line spectrum, however, does not 
match that of a middle-B type star, and subtracting a hot 
continuum does not adequately explain the shallowness of the 
eclipse from 1400 to 1500 A. 

6. The only strong emission line seen is that of O I 1304. It 
fluctuates by a factor of 2 during the eclipse, but too 
little out-of-eclipse data is available to say if this 
behavior is eclipse dependent. 

The observations at high dispersion are more difficult to study because 
of the wealth of overlapping lines at lUE's resolution. Some early 
results are: 

1. P-Cyg profiles in the Mg II doublet were revealed in the 
central core as the continuum faded. The profile remained 
essentially unaltered throughout the eclipse, the absorption 
core being shifted -13 km/s. During ingress, the Mg II line 
wings dropped more rapidly than the continuum and during 
egress recovered more slowly. 

2. It is difficult to distinguish multiple velocity components 
in the UV, but some structure is reported. The radial 
velocity curve of the photospheric lines is nearly constant 
up to mid-totality, becomes more negative by about 45 km/s up 
to third contact, then returns to the velocity value expected 
from the orbital motion of the primary at the end of the 

3. The low-excitation lines of Fe II, Mn II and Cr II have 
stationary components that appear to be the tops of emission 
lines filling in the absorption cores of the stellar lines. 
At the constant velocity phases, they are appear on the 
redward side of the corresponding absorption lines. 



As would be expected from the pre-eclipse observations that the UV 
energy distribution is still mainly dominated by the primary, the 
eclipse data mimics in many ways the behavior seen in the optical 
region. The downward slope of the light curve during totality, the 
superimposed light fluctuations / and the radial velocity curve are 
consistent with previous observations? the Mg II emission has 
counterparts in other regions, such as in H "alpha, and is consistent 
with Ca II measurements. The lUE observations, however, do shed some 
light when interpreting this behavior. 

The UV fluctuations have been interpreted as being due to aperiodic 
Cephe id-like pulsation of the primary (Ake and Simon 1984 ) or as 
structure in the occulting body (holes or tunnels, Parthasarathy and 
Lambert 1983, Boehm et al 1984). We feel the fact that variations are 
enhanced with decreasing wavelength down to 1500 A favors the pulsation 
explanation. Schimdt and Parsons (1982) find that in Cepheids a 0.5 
mag. amplitude in V translates to amplitudes up to 5 mag. at 1600 A 
because of the extreme temperature sensitivity of the UV continuum and 
ionization edges in F supergiants. The strength of the I emission in 
Eps Aur is also consistent with the shock-induced I emission in 

The opacity of the occulting body is mainly continuous (Chapman et al 
1983) as most of the absorption lines in high dispersion do not change 
in depth nor do different lines appear during the eclipse. Some lines, 
however, are reported to show some structure taken to be evidence of 
multiple components (Ferluga and Hack 1984) and others seem to be 
filled in by emission peaks (Castelli et al 1982). Furthermore 
interstellar or circumstellar components are seen in low-excitation 
lines of Mg I, Mg II, Fe II, etc. 

The radial velocity curve in the UV derived from the photospheric lines 
is somewhat consistent with that reported in the past with other 
eclipses. The lines are found to be blueshifted after mid-totality, 
but prior to mid-totality no corresponding large redshift is seen (Ake 
and Simon 1985) . 

The constancy of the Mg II emission (Altner et al 1984) and deduced 
emission of other low-excitation lines is characteristic of the Zeta 
Aur systems where a hot secondary interacts with the wind from the 
cooler primary and excites the circumsystem material. In these 
systems, when the continuum from the hot star is reduced during an 
eclipse, the emission lines appear with redshifted peaks due to 
scattering of the hot star photons off the receding gas in the wind 
from the primary. In Eps Aur, the "emission peaks", which remain 
constant as the overlying absorption deepens during the eclipse, are 
found on the redward side of the corresponding absorption components 


much as in the Zeta Aur systems. 

Perhaps the most intriguing aspect of the lUE observations is the shape 
of the far UV energy distribution and the eclipse light curve since 
they provide new insight into the nature of the system. The 2200A 
depression does not change during the eclipse indicating that the 
occulting body is not composed of the types of grains typically found 
in the interstellar medium (Boehm et al 1984, Ake and Simon 1984), 
Moreover, the absence of additional line absorption implies that the 
occulting body is also devoid of a significant amount of gaseous 
material. Finally we note that the UV excess shortward of 1400 A, as 
reported by Hack and Sevelli (1979), is suggestive that a hot secondary 
has been detected, but it cannot be definitively stated to be that of a 
hot star (Parthasarathy and Lambert 1983, Ake and Simon 1984). The 
final test of the location of this added UV source will be observations 
at the predicted time of the next secondary eclipse. 


Ake, T.B. and Simon, T. 1984, Future of Ultraviolet Astronomy Based 
on Six Years of lUE Research, p. 361 

1985, in preparation 

Alter, B.M. , Chapman, R.C., Kondo, Y. and Stencel, R.E. 1984, Future 

of Ultraviolet Astronomy Based on Six Years of lUE Research, p. 365 
Boehm, C. , Ferluga, S. and Hack, M. 1984, Astron. Astrophys. , 130, 419 
Castelli , F. , Hoekstra, R. and Kondo, Y. 1982, Astron. Astrophys. 
Suppl. , 
50, 233 
Chapman, R. , Kondo, Y. and Stencel, R.E. 198 3, Ap. J. (Letters) , 269, 

Ferluga, S. and Hack, M. 1984, Proceedings of the Fourth European lUE 

Conference, p. 419 
Hack, M. and Sevelli, P. L. 1979, Astron. Astrophys., 75, 316 
Parthasarathy, M. and Lambert, D.L. 1983, Pub. Astr. Soc. Pac, 95,1012 
Schimdt, E.G. and Parsons, S.B. 1982, Ap. J. Suppl., 48, 185 


4800 5000 5200 5400 5600 





Figure 1. (Ake) Eclipse light curves for € Aur as measured by lUE's 
Fine Error Sensor (transformed to V magnitudes) and 
ultraviolet regions centered at 2850, 1900, 1500 and 
1260 A (converted to magnitudes on an energy scale 
where m^=0 is 3.64 x 10"-9 ergs/cm2/sec/A) . Dotted 
lines indicate unobserved dates due to € Aur*s 
proximity to the Sun.