3 Results

3.1 Survey of asteroids

Our spectroscopic survey reveals that more than 65% of the 34 observed asteroids show the presence of absorption bands due to aqueous alteration products, in particular of the 0.7 $\mu$m band, which is the most characteristic feature of the hydrated materials in the visible.

The depth of these bands varies between 2% and 6% with respect to the continuum.

We considered only the absorption features deeper than the peak-to-peak scatter (that is $\le$ 0.02) in the spectrum, which, from previous experience, seems to be a better indicator of the spectrum quality than the calculated signal to noise ratio (Vilas & Smith 1985).

The repeatability of the 0.7 $\mu$m absorption band in 19 Fortuna and 51 Nemausa, which were both observed twice on different observing runs, is a good indicator of the quality of data reduction.

Moreover the location and the extension of aqueous altered absorption characteristics do not match any atmospheric absorption band or solar analog feature.

The intense telluric water absorption beginning near 0.9 $\mu$m coupled with a drop in responsivity of the CCD detectors have affected the identification of 0.8-0.9 $\mu$m features in the asteroid spectra, which we have clearly identified only on 1 Ceres.

Some spectra present spurious features due to an incomplete removal of telluric H₂O at 7300 and 8200 Å and/or to the atmospheric O₂A and O₂B bands at 7619 and 6882 Å respectively, but they do not influence the identification of aqueous altered bands.

Of six investigated objects located outside the "aqueous alteration zone'', four have shown the presence of hydration features (Fig. 4). We think that more observations could help to understand the efficiency zone of the aqueous alteration process.

  

Figure 4: Number of the observed objects as function of the semimajor axis. The black part represents the hydrated asteroids (only those which have clear and well identified hydration absorption features)

Our results are consistent with those obtained by Barucci et al. (1998), who found that more than 65% of their observed asteroids are hydrated. They also observed hydration features on asteroids located closer to the Sun than 2.6 AU.

Our data also confirm the existence of a relationship between the albedo of the objects and the aqueous alteration process (Fig. 5): the percentage of the observed hydrated asteroids grows as albedo increases. This relation may be explained with the progressive leaching of iron from silicates as the aqueous alteration proceeds. Leached iron (iron is the most important opaque phase in the visible range associated to aqueous alteration process) would be enveloped into magnetite and iron sulfide grains, so less material would be available to absorb the incoming sunlight and this would cause the increasing of the albedo (Vilas 1994).

  

Figure 5: Number of the observed objects as function of the geometric albedo. The black part represents the hydrated asteroids (only those which have clear and well identified hydration absorption features)

All the 3 G-class observed asteroids (1 Ceres, 19 Fortuna, 130 Elektra) have features attributed to aqueous altered materials, confirming the fact that G-type objects seem to be the most aqueous altered asteroids. In fact aqueous alteration sequence seems to begin with P class objects (the least altered) and to increase through F $\rightarrow$ B $\rightarrow$ C $\rightarrow$ G asteroids (Vilas 1994). Two of the G-type observed objects, 19 Fortuna and 130 Elektra show a well defined 0.7 $\mu$m absorption band, while 1 Ceres has the 0.8-0.9 $\mu$m band and two weak absorption bands at 0.6 and 0.67 $\mu$m, as observed by Vilas & McFadden (1992) and Sawyer (1991).

Ceres has not the 0.7 $\mu$m band, that seems to be a spectral characteristic of G class objects, but, owing to its significant size, it cannot be considered typical of any asteroidal class.

  

Table 2: Identification of absorption bands on the observed asteroids. For each band we indicate the central wavelength position, the depth and extension. We have identified only those bands whose depth is greater that the peak-to-peak scatter of the spectra due to noise. With the ? we have indicated those bands that seem to be present but are too weak with respect to the noise, so we believe that more observations are necessary to confirm them

3.2 Comparison between CM2 chondrites and hydrated asteroids

Finally, we have compared the spectra of the observed hydrated asteroids with those of several CM2 carbonaceous chondrite meteorites.

The spectra of the CM2 meteorites have been obtained from literature (Vilas et al. 1994) and have laboratory origin.

They reveal features probably due to aqueous altered materials on their surfaces. This investigation is important because the origin of meteorites is not well known yet.

Many factors affect the comparison between meteorites and asteroids (Pieters & McFadden 1994):

• Physical preparation of a meteorite for measurement in the laboratory may not accurately reproduce the physical form of material on an asteroid surface.
• A significant limiting factor that affects the mineralogical interpretation of meteorite and asteroid spectra includes the signal to noise ratio (so the quality of the spectra), the spectral range of the measurement and the spectral resolution. In fact signal to noise ratio of dark asteroids spectra is generally quite lower than that of meteorites spectra obtained in the laboratory environment (100 or less for asteroids, about 3000 for meteorites). Available spectra of asteroids are limited by the sensitivity of detectors and the faintness of the signal from the object combined with the atmospheric and instrumental noise contribution.
• There are significant dimensional differences, so meteorites and asteroids may have experienced different thermal and chemical processes.
• Many asteroids have not contributed to our meteorite collection and may not correspond to any meteorite class. Moreover, asteroidal observations allow only the study of their surface properties, while meteorites may be derived also from the inner parts of the parent bodies.
• Space weathering influences the surface composition, and may affect asteroid and meteorites in different manner. This is a result of studies especially on lunar meteorites, and seems to produce lower albedo, weaker absorptions and a red-sloped continuum.
• Meteorites may be deeply altered from their pristine interspace condition by terrestrial weathering (Britt et al. 1992).

  

Table 3: Values of the Chi-Square computed by the comparison of each hydrated asteroid with the 7 CM2 chondrites. The maximum likelihood between the asteroid and the meteorites is obtained for that CM2 which gives the least Chi-Square value (represented in boldface)

We compared our hydrated asteroids(those which have the 0.7 $\mu$m band) with 7 CM2 carbonaceous chondrites (see Fig. 6),

  

Figure 6: Reflectance spectra of 7 CM2 carbonaceous chondrite meteorites (Vilas et al. 1993). The spectra were normalized to 5500 Å and a linear continuum has been removed. They were offset by 0.15 in reflectance for clarity

whose spectra were treated as those of the asteroids, that is they were divided by a linear continuum defined by a linear least squares fit to the spectral data points.

We used the Chi-Square Fitting method to measure the agreement between each hydrated asteroid and the 7 CM2. The Chi-Square is defined in the following way:

$$\chi^2 = \sum_{i=1}^{N}\frac{(ya_i - yc_i)^2} {\sigma_{i}^{2}}$$

where (x_i, ya_i) are the mean data points (x = wavelength; y = reflectance) of the asteroid (i.e. a mean of the asteroidal signal, affected by noise); (x_i, yc_i) are the data points of CM2 chondrites (these spectra were obtained in laboratory, so they have pratically no noise); N = number of degree of freedom, i.e. the number of points in which we have divided the wavelength range; $\sigma_i$ is the standard deviation of the asteroidal data points.
We have computed the Chi-Square on 200 points (N = 200).

As we have not a set of observations for each asteroid, but a single spectrum per asteroid, we have assumed $\sigma_i$ = costant = 0.02, which is about the mean peak-to-peak variation of the asteroidal signal.

The maximun likelihood estimate between each asteroid and the 7 "models'' represented by the 7 CM2 meteorites is obtained when the Chi-Square assumes the least value.

The results of this quantitative comparison are summarized in Table 3. The best meteoritical analog of all the hydrated asteroids is LEW90500 CM2: the other 6 CM2 have a deeper and wider 0.7 $\mu$m absorption band than that of the asteroids.

In Figs. 7 and  8 we report some examples of the comparison between hydrated asteroids and LEW90500 CM2.

  

Figure 7: Comparison between the asteroids 19 Fortuna, 41 Daphne, 51 Nemausa, 70 Panopaea, 105 Artemis and 130 Elektra with the carbonaceous chondrites LEW90500. The spectra are offset by 0.3 in reflectance for clarity

  

Figure 8: Comparison between the asteroids 137 Meliboea, 144 Vibilia, 146 Lucina, 211 Isolda, 410 Chloris and 776 Berbericia with the carbonaceous chondrites LEW90500. The spectra are offset by 0.3 in reflectance for clarity

The differences in depth of the band and in its extension may depend on the degree of aqueous alteration and on the presence of different amount of Fe in silicates crystal lattices. Moreover, laboratory experiments show that the reflectance of a mineral mixture is nonlinear and is a function of viewing geometry and properties of the particles such as single scattering albedo (efficiency of an average particle to scatter and not absorb light), porosity, diameters and mass fractions (Burbine et al. 1996). So small differences in composition or in particle sizes and properties are sufficient to produce a different spectral response.

The good match between several observed hydrated asteroids and CM2 meteorites, in particular LEW90500, resulting from our analysis, is a valid confirmation that aqueous altered asteroids could be the parents of CM2 carbonaceous chondrite meteorites.