S. Darbon, J.-M. Perrin, J.-P. Sivan
Empirical relationships on the properties of the Extended Red Emission (ERE) are presented. They are based on published observational data and on new results obtained on reflection nebulae illuminated by cold stars. The plot of the width versus the central wavelength of the ERE band is in agreement with laboratory properties of the materials commonly proposed as the ERE carriers. But this is not the case for the plot of the ERE band width versus the effective temperature of the nebula illuminating star.
Key Words: ISM: general - ISM: reflection nebulae - ISM: dust,extinction - ISM: HII regions - ISM: planetary nebulae - Galaxies: individual: M82
The Extended Red Emission (ERE) is a continuous emission band observed in the
red part (6000-8000 Å) of the spectrum of various astrophysical objects :
reflection nebulae (Schmidt, Cohen & Margon 1980; Witt & Boroson 1990),
planetary nebulae (Furton & Witt 1992), galactic and extragalactic Hii
regions (Perrin & Sivan 1992; Sivan & Perrin 1993; Darbon, Perrin & Sivan
1998),
high-latitude galactic cirrus clouds (Szomoru & Guhathakurta 1998), the halo of
the
galaxy M82 (Perrin, Darbon & Sivan 1995) and the diffuse galactic medium
(Gordon,
Witt & Friedmann 1998). The ERE band is found to vary significantly both in
position and width from an object to an other. The diversity of objects where
the ERE is detected and the diversity of the observed band characteristics lead
us
to search for empirical relationships that should help identification of the ERE
carriers.
For all the available observations, we have plotted in Fig.
the width as a function of the central wavelength of the
ERE bands. For reflection nebulae, we have used the values of Witt & Boroson
(1990). For all the other objects, we have made measurements according to the
same definitions as Witt & Boroson : the central wavelength splits the band
luminosity
into two equal parts and the width is measured as the difference between the
wavelengths
of the first and third quartiles. Note that for planetary nebulae the values
derived from the published spectra of Furton & Witt (1992) are
underestimated.
The diagram in Fig. reveals a clear correlation between the
position of the maximum and the width of the band. We derived a correlation
coefficient r=0.68. This result confirms and reinforces the tendency
previously noted by Witt & Boroson (1990) from reflection nebulae only : the
correlation coefficient was r=0.52. The same effect
is observed in laboratory experiments : Hydrogenated Amorphous Carbon (HAC)
grains
(Furton & Witt 1993) and nanocrystals of silicon (Witt, Gordon & Furton 1998 ;
Ledoux et
al. 1998) exhibit luminescence bands whose width increases with the maximum
wavelength.
Also, it is found that the plotted values in Fig. are split
into two groups : reflection nebulae with smaller widths and bluer peaks and
Hii regions and planetary nebulae with larger widths and redder peaks
(the halo of M82 lies at the border of these two groups). This repartition might
be
related to the presence or absence of plasma within the nebula, in agreement
with laboratory results (Wagner & Lautenschlager 1986; Robertson & O'Reilly
1987)
As a mean, in Fig. , the reflection nebulae are illuminated by stars less energetic than the other nebulae. This repartition suggests that the characteristics of the ERE might depend on the spectral distribution of the exciting flux. This has lead us to study the variation of the ERE band width as a function of the effective temperature of the exciting stars. BVI photometric and spectrophotometric observations, conducted respectively by Witt & Schild (1985) and Witt & Boroson (1990), deal with reflection nebulae illuminated by stars with a spectral type earlier than A0 (i.e. Teff >= 10000 K) : most of these nebulae exhibit ERE. But laboratory results show that the ERE can be excited by low energy visible radiation. So, the observation of reflection nebulae illuminated by stars colder than A0 (i.e. with a spectral energy distribution peaking in the visible) should be useful in order to detect the presence or the absence of the ERE in their spectra. One such nebula, illuminated by an M giant star, has been observed by Witt & Rogers (1991). We have observed six additional nebulae illuminated by cold stars. They are listed in Table , together with spectral type and the effective temperature of the illuminating stars. We obtained low resolution spectra for these objects. Data reduction were conducted as previously described in Perrin, Darbon & Sivan (1995). None of these nebulae exhibits ERE in its spectrum.
In Fig. , we have plotted the ERE band width as a function of
the
effective temperature of the exciting star for the reflection nebulae and
Hii
regions of Fig. and for the nebulae of Table
. For
the nebulae without ERE, the width value is set to zero.
Clearly, no ERE appears for Teff =< 7000 K. On the contrary, for
Teff >= 10000 K, most of the plotted objects exhibit the ERE albeit few
of
them do not. This suggests a cut-off in effective temperature might exist
between 7000 and 10000 K (unfortunately no observation is available for this
range). This cut-off cannot be accounted for by extinction effects as
demonstrated by Fig. : the presence or the absence of the ERE
in a nebula does not appear to be correlated to the color excess of its
illuminating star.
The fact that no ERE is present for stars whose
effective temperature is smaller than 7000 K does not agree with current
laboratory data on various materials such as HACs and silicon nanocrystals.
These materials
exhibit red luminescence features when they are irradiated by photons of low
energy, namely of
energy smaller than 3 eV (see e.g. Sussmann & Ogden 1980, Lin & Feldman 1981,
Fang et al. 1988,
Wilson et al. 1993).
Further observations of reflection nebulae illuminated by cold stars would be
useful to confirm our finding. We cannot completely rule out that the absence of
ERE in
these nebulae could be simply due to the absence of any luminescent material.
This is probably the case for the nebulae without ERE belonging to the
left-hand
part of the
diagram in Fig.
(Teff >= 10000 K). This suggests that
the ERE
carriers would not be present everywhere in the interstellar medium.