Special Reports

X-ray Fluorescence Spectrometry of Asteroid Itokawa by Hayabusa

See allHide authors and affiliations

Science  02 Jun 2006:
Vol. 312, Issue 5778, pp. 1338-1341
DOI: 10.1126/science.1125731

This article has been retracted. Please see:

Abstract

X-ray fluorescence spectrometry of asteroid 25143 Itokawa was performed by the x-ray spectrometer onboard Hayabusa during the first touchdown on 19 November 2005. We selected those data observed during relatively enhanced solar activity and determined average elemental mass ratios of Mg/Si = 0.78 ± 0.09 and Al/Si = 0.07 ± 0.03. Our preliminary results suggest that Itokawa has a composition consistent with that of ordinary chondrites, but primitive achondrites cannot be ruled out. Among ordinary chondrites, LL- or L-chondrites appear to be more likely than H-chondrites. No substantial regional difference was found on the asteroid surface, indicating its homogeneity in composition.

Understanding the relation between asteroids and meteorites is a long-standing problem in asteroid science, and a relation between S-class asteroids and ordinary chondrites might be constrained with new evidence provided by the Hayabusa x-ray fluorescence (XRF) spectrometer. Near-Earth asteroid 25143 Itokawa is classified spectroscopically as an S (IV)–type asteroid, and ground-based observations suggest that Itokawa has an LL-chondrite composition [e.g., (1)]. Indeed, a likely correlation between ordinary chondrites and S (IV) class asteroids has been suggested [e.g., (2)], but direct evidence to support this is lacking. Understanding the material composition of Itokawa and its origin is a key scientific objective of the Hayabusa mission (3, 4). XRF spectrometry by the x-ray spectrometer (XRS) was conducted to obtain a major-elemental analysis of the asteroid′s surface in order to classify its rock type (or meteoritic class) and survey any regional variation (5, 6). Here we report preliminary results of the XRS observations during the first touchdown of Hayabusa on the surface of Itokawa, when sample collection was attempted.

XRF is a well-established technique for major-elemental analysis in the laboratory. In space, however, the excitation source is solar x-rays so the analysis is complicated by features in the solar spectrum and flux variations. As was shown in previous planetary missions (79), XRF can be used to determine the major-element composition of the uppermost several tens of micrometers of atmosphere-free planetary bodies such as the Moon, Mercury, and asteroids (5, 712).

The XRS is an advanced XRF spectrometer (5, 6) with a light-weight (1.5 kg) sensor unit based on a charge-coupled device (CCD) x-ray detector (13); this is the first time a CCD has been used for such a purpose on a planetary mission. The CCD has an energy resolution of 160 eV at 5.9 keV when cooled, which is much higher than that of the proportional counters used in previous planetary missions (14, 15). In addition, the XRS has a standard sample plate (SSP) for concurrently calibrating the XRF when it is excited by the Sun. The SSP is a glassy plate whose composition is intermediate between those of chondrites and basalts (table S1). By comparing x-ray spectra from the asteroid and from the SSP, quantitative elemental analysis can be achieved, although the intensities and spectral profiles of solar x-rays change over time. In-flight performance of the XRS was confirmed by observations of x-rays from the SSP (16), x-ray–emitting bodies such as Kepler′s supernova remnants (17), and x-rays from the far side of the Moon shortly before the Earth swing-by on 17 May 2004 (18).

After arrival at Itokawa on 12 September 2005, Hayabusa began scientific observations. Initially, the solar activity (19) was very faint and unfavorable for XRS observations. However, it became relatively brighter late in November 2005.

Under quiescent conditions, solar x-rays are produced in the solar coronal regions with typical temperatures of 1 × 106 to 4 × 106 K and have steeply decreasing spectra with energy. Under such conditions, K-α line spectra of lighter rock-forming elements such as Mg (1.25 keV), Al (1.49 keV), and Si (1.74 keV) are readily observed. Heavier elements such as Ca (3.69 keV) and Fe (6.40 keV) are detectable only during solar flares, when solar x-rays become more intense and relatively stronger at high energies. S (2.37 keV) is at the lower limit of detection, due to instrumental noise and the relatively low signal-to-background ratios obtained during that time. Therefore, our analysis is concentrated on three elements that play important roles in rock type classification and that serve as indicators of evolutional processes.

Hayabusa began its descent for the first touchdown early on 19 November 2005, staying 1.4 km earthward of Itokawa. The earthward direction was chosen because it allowed scientific instruments on Hayabusa to point toward the asteroid and maintain continuous communication with Earth through the high-gain antenna of the spacecraft. During the descent, the solar phase angle—that is, the angle between the Sun, Itokawa, and Hayabusa—was between 5° and 10°. This small phase angle is favorable for performing XRF spectrometry. At larger phase angles, shadows on the rough surface terrain can interfere with the elemental analysis [e.g., (20)].

Itokawa rotates once every 12.1 hours (21), and it rotated more than once by the time of touchdown. During descent, the XRS continued x-ray spectrometry of Itokawa, and the SSP simultaneously excited by the Sun, so that the entire asteroid surface was observed in a longitudinal direction due to its rotation.

The effective field of view of the XRS is 3.5° by 3.5° (5, 6), corresponding to a “footprint” size of 87 m by 87 m at an altitude of 1.4 km, which became proportionally smaller as Hayabusa descended. Consequently, the XRS observed local areas of Itokawa 550 m by 298 m by 244 m in size (22). XRS observations were sometimes affected by background radiation from space (x-rays and energetic particles) because the footprint was not centered on the asteroid, due to the rotation of the irregularly shaped asteroid and instability of the spacecraft′s attitude control that resulted from the failure of two of the three reaction wheels.

The XRS produced a set of x-ray energy spectra of Itokawa and the SSP, after onboard processing of data from each CCD (6), every 5 min. Much of the data could not be analyzed due to unfavorable solar activity, but six data sets with good signal-to-background ratios—observed during periods of relatively enhanced solar activity—were selected (Table 1). The names of footprint regions on the asteroid (23) are also listed in Table 1, including the touchdown site, Muses Sea. X-ray spectra of the SSP and of Itokawa are shown (Fig. 1, A and B, respectively). Line spectra due to the XRF of Mg, Al, and Si were best-fit with Gaussian profiles, superimposed on background continuum components (Fig. 1, A and B). The errors shown in Table 1 are mainly caused by the background extraction procedure and are almost equal to the uncertainties obtained by least-squares fitting.

Fig. 1.

X-ray spectra of the onboard standard sample (A) and asteroid Itokawa (B) were simultaneously observed by the XRS at 9:27 UTC on 19 November 2005. The observed spectra (OBS) are fitted by Gaussian profiles to K-α lines of major elements (Mg, Al, and Si) and by a background continuum component (CONT).

Table 1.

Elemental composition of Itokawa derived at six sites from the XRS observations on 19 November 2005. Errors are 2σ uncertainties. SSP, XRF, and Norm denote the standard sample plate, x-ray fluorescence, and normalized intensities, respectively.

TIME (UTC)Longitude (°) regionSSP XRFItokawa XRFNorm XRFItokawa composition (wt%)
Mg/SiAl/SiMg/SiAl/SiMg/SiAl/SiMg/SiAl/SiMgSi
01:19 40 1.16 0.49 2.01 0.15 1.73 0.31 0.76 0.06 15.1 19.8
“Neck” (Pencil boulder) ±0.07 ±0.08 ±0.05 ±0.08 ±0.15 ±0.17 ±0.08 ±0.03 ±2.3 ±3.0
09:27 285 1.11 0.58 1.97 0.18 1.77 0.31 0.78 0.07 15.2 19.4
Muses Sea ±0.06 ±0.06 ±0.07 ±0.06 ±0.11 ±0.12 ±0.07 ±0.03 ±2.0 ±2.5
12:10 5 0.88 0.47 1.63 0.16 1.85 0.34 0.82 0.08 15.5 19.0
“Head” region ±0.06 ±0.10 ±0.07 ±0.09 ±0.16 ±0.16 ±0.10 ±0.03 ±2.3 ±2.8
15:12 96 1.35 0.76 2.31 0.17 1.71 0.22 0.75 0.06 15.0 19.9
Tsukuba ±0.11 ±0.20 ±0.10 ±0.10 ±0.17 ±0.12 ±0.10 ±0.03 ±2.9 ±3.3
16:02 120 0.95 0.47 1.66 0.19 1.75 0.40 0.77 0.09 14.9 19.4
A rough terrain ±0.07 ±0.15 ±0.08 ±0.10 ±0.16 ±0.22 ±0.09 ±0.04 ±2.3 ±2.8
16:34 136 0.92 0.49 1.70 0.12 1.85 0.24 0.81 0.06 15.6 19.2
Yoshinodai ±0.07 ±0.08 ±0.07 ±0.07 ±0.14 ±0.14 ±0.08 ±0.03 ±2.0 ±2.8
Average 0.78 0.07 15.22 19.46
±0.09 ±0.03 ±2.3 ±2.8

The K-α line spectra of Mg, Al, and Si from the SSP are clearly distinguished (Fig. 1A). The continuum component at 1 to 2 keV is dominated by scattered solar x-rays. In contrast, the K-α line spectra of Mg and Si from Itokawa are clearly seen, but that of Al is relatively faint (Fig. 1B). The spectral profile with relatively higher Mg/Si and lower Al/Si indicates that Itokawa is more chondritic in composition than the SSP. The relatively large continuum strength at lower energy may be due to irradiation of cosmic x-rays and energetic particles, because at times one-third to one-half of the XRS′s field of view was occupied by background space.

Table 1 shows the intensity ratios of Mg/Si and Al/Si for the SSP and Itokawa at each observed time. The ratios of Mg/Si and Al/Si for the asteroid normalized by those of the SSP are also shown. A small variation of 10% is found among the sites. Based on the elemental-analysis method using the SSP (24), the elemental ratios of the asteroid surface normalized by those of the SSP are almost proportional to the normalized intensity ratios within the uncertainties of several percent. We then estimated the elemental ratios for each of the six sites and derived average compositions of Mg/Si = 0.78 ± 0.09 and Al/Si = 0.07 ± 0.03, respectively (Table 1). The derived elemental ratios of Mg/Si are 0.3 to 0.4 of the intensity ratios of Mg/Si, in good agreement with a previous numerical study of chondritic material (25).

The average elemental ratios of Mg/Si and Al/Si are plotted (Fig. 2) along with those for other stony and stony-iron meteorites (table S2), which are considered candidate materials of S-class asteroids (25, 26). Ordinary chondrite compositions, including those of LL-, L-, and H-chondrites, are likely for Itokawa, within 1σ uncertainties for both Mg/Si and Al/Si. Some primitive achondrites such as acapulcoites cannot be ruled out. IAB-winonaites, a kind of primitive achondrite, and the related silicated IAB iron meteorites, appear closest in composition, although the composition may be diverse.

Fig. 2.

(A) Elemental ratios of Itokawa (shaded area) plotted as Mg/Si versus Al/Si, together with typical compositions of stony and stony iron meteorites. (B) Enlargement of shaded area in (A). The dashed and dotted lines in (A) denote the best-fit values and their 2σ uncertainties, respectively.

We also estimated the absolute abundance of Si (and also Mg from Mg/Si) for each site (Table 1), under the assumptions that all minerals have a stoichiometric composition and there is no abundant water on the surface. Such information should take into account the instrumental characteristics including detection area, detection efficiency and field of view (or solid angle of the target), the intensity and spectral profile of the primary x-ray source, and the physical parameters such as fluorescence yield of each element. In this study, they are all given or canceled by using the compared method. Then we calculated the average abundances of Mg = 15.2 ± 2.3 wt. % and Si = 19.5 ± 2.8 wt. %, respectively. This estimate is valid under the assumption that the intensity of the continuum component at Si-Kα is dominated by solar-scattered x-rays.

The derived abundances of Mg and Si are also plotted (Fig. 3). LL- or L-chondritesappearmost likely within 1σ uncertainty, although H-chondrites are also possible within 2σ uncertainty. Some primitive achondrites remain possible candidates and cannot be ruled out. Analysis of 1-μm band absorption with near-infrared spectroscopy also suggests that olivine-rich LL-chondrites are preferred and other achondrites are possible although less likely (27). In contrast, analysis of the 2-μm band shows more olivine-enriched features than any known meteorite types, suggesting the possibility of a primitive achondrite whose parent body is probably an LL- or L-chondrite (28).

Fig. 3.

(A) Elemental composition of Itokawa (Si/Mg) plotted as in Fig. 2. (B) Enlargement of the shaded area in (A). Ordinary chondrites, acapulcoites, and IAB winonaites appear closest in composition, and Itokawa appears to be more like an LL- or L-chondrite in composition than an H-chondrite.

The target of the Near-Earth-Asteroid-Rendezvous (NEAR)–Shoemaker mission, 433 Eros, is also an S (IV)–class asteroid and is the only other asteroid analyzed by XRF spectrometry (8, 25). Its composition is best described by that of H-chondrites by XRF spectrometry, whereas L- or LL-chondrites are more consistent with γ-ray spectra and visible–to–near-infrared observations. But the conclusion that Eros is more like an ordinary chondrite is nevertheless consistent with the findings for Itokawa, although some melting cannot be ruled out (8, 25, 29). For both Eros and Itokawa, it is significant that the compositions derived by remote spacecraft observations in close proximity to the asteroid seem consistent with those found using Earth-based spectroscopy (1), which is applicable to a large number of asteroids.

As mentioned above, the normalized elemental ratios of Mg/Si and Al/Si are found to be almost constant within 10% among the sites. A homogeneous composition of Itokawa is likely when observed at a spatial resolution of several tens to a hundred meters. This homogeneity implies that Itokawa probably has a uniform rock type across the asteroid. It also shows that there is no evidence of either local evolution processes or that two bodies of different composition were involved if Itokawa has a contact binary origin (4). At a smaller scale, some regional variation of color and albedo has been found in visible–to–near-infrared wavelength data (22, 27). But the XRS results cannot currently constrain elemental ratios at that scale.

Supporting Online Material

www.sciencemag.org/cgi/content/full/312/5778/1338/DC1

Tables S1 and S2

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article