A New Method for Deriving Composition of S-type Asteroids from Noisy and Incomplete Near-infrared Spectra

Abstract

The surface composition of S-type asteroids can be determined using band parameters extracted from their near-infrared (NIR) spectra (0.7-2.50 µm) along with spectral calibrations derived from laboratory samples. In the past, these empirical equations have been obtained by combining NIR spectra of meteorite samples with information about their composition and mineral abundance. For these equations to give accurate results, the characteristics of the laboratory spectra they are derived from should be similar to those of asteroid spectral data (i.e., similar signal-to-noise ratio (S/N) and 2 Sanchez et al. 2020wavelength range). Here we present new spectral calibrations that can be used to determine the mineral composition of ordinary chondrite-like S-type asteroids. Contrary to previous work, the S/N of the ordinary chondrite spectra used in this study has been decreased to recreate the S/N typically observed among asteroid spectra, allowing us to obtain more realistic results. In addition, the new equations have been derived for five wavelength ranges encompassed between 0.7 and 2.50 µm, making it possible to determine the composition of asteroids with incomplete data. The new spectral calibrations were tested using band parameters measured from the NIR spectrum of asteroid (25143) Itokawa, and comparing the results with laboratory measurements of the returned samples. We found that the spectrally derived olivine and pyroxene chemistry, which are given by the molar contents of fayalite (Fa) and ferrosilite (Fs), are in excellent agreement with the mean values measured from the samples (Fa 28.6±1.1 and Fs 23.1±2.2 ), with a maximum difference of 0.6 mol% for Fa and 1.4 mol% for Fs.

Physical Characterization of Near-Earth Asteroid (52768) 1998 OR2: Evidence of Shock Darkening/Impact Melt

“…The Band I parameters were measured 12 times using a third-order polynomial fit and sampling different ranges of points within the band. The Band II parameters were measured 13 times using the 2.45 μm upper limit as described in Sanchez et al (2020), to account for the noisy data and missing points at the longest wavelength. The band area is calculated as the area between the continuum and the data curve.…”

“…While band depths are not diagnostic of surface mineralogy, it is interesting to note that 1998 OR2ʼs absorption bands are subdued (4.5% and 4.0% for Bands I and II, respectively) compared to most S-complex asteroids (typically ∼15%-25% for Band I and ∼5%-10% for Band II; e.g., Sanchez et al 2012). The spectrally derived olivine-to-pyroxene abundance ratio, ol/ (ol+px), and the molar percent fayalite (Fa) and ferrosilite (Fs) were estimated using the appropriate equations from Dunn et al (2010) and Sanchez et al (2020). The summary of the band parameters and the estimated mineralogy is shown in Table 3.…”

“…The wavelength range we used for these calculations was 0.45-2.45 μm. Previous studies (e.g., Cloutis et al 1986;Gaffey et al 1993;Dunn et al 2010;Sanchez et al 2020) have discussed that the BAR can be particularly sensitive to the wavelength range used. Sanchez et al (2020) shows that BAR measurements for H-chondrites can vary greatly if the incorrect wavelength range is used for the estimation.…”

“…Previous studies (e.g., Cloutis et al 1986;Gaffey et al 1993;Dunn et al 2010;Sanchez et al 2020) have discussed that the BAR can be particularly sensitive to the wavelength range used. Sanchez et al (2020) shows that BAR measurements for H-chondrites can vary greatly if the incorrect wavelength range is used for the estimation. Although we use the best wavelength range (0.45-2.45 μm) for this measurement and follow methods laid out by previous studies, it is still possible that the BAR uncertainty could be larger than suggested by the variance in our repeated measurements.…”

Physical Characterization of Metal-rich Near-Earth Asteroids 6178 (1986 DA) and 2016 ED85

“…The spectra of these two asteroids are very similar, with 2016 ED85 showing a redder slope than 1986 DA. Spectral band parameters, including the band center and band depth of the 0.9 µm pyroxene band, were measured from the spectra of 1986 DA and 2016 ED85 using a Python code described in Sanchez et al (2020). The Band I center was measured after dividing out the linear continuum and corresponds to the position of the minimum reflectance value obtained by fitting a polynomial over the bottom of the absorption band.…”