Abstract– Keystones removed from the Stardust cometary collector show varying degrees of visible fluorescence when exposed to UV light, with the brightest fluorescence associated with the space-exposed surface. We investigated the spatial characteristics of this phenomenon further by using fluorescence microscopy, confocal Raman microscopy, and synchrotron Fourier transform infrared (FTIR) spectromicroscopy. Twenty-four keystones, extracted from the Stardust cometary collector, were analyzed. Fluorescence measurements show two distributions with different excitation characteristics, indicating the presence of at least two distinct fluorophores. The first distribution is confined to within about 10 μm of the space-exposed surface, whereas the second distribution is much broader with a maximum that is typically about 30–50 μm below the surface. Confocal Raman measurements did not reveal any changes associated with the surface; however, only features associated with aliphatic hydrocarbons were strong enough to be observed. FTIR measurements, on the other hand, show two distinct distributions at the space-exposed surface: (1) a narrow, surface-confined distribution originating from −O3SiH groups and (2) a broader, sub-surface distribution originating from −O2SiH2 groups. These functional groups were not observed in keystones extracted from the cometary flight spare or from the Stardust interstellar collector, indicating that they may result at least partially from cometary exposure. The presence of O3SiH and O2SiH2 groups at the comet-exposed surface suggests that the enhanced surface fluorescence is caused by defects in the O-Si-O network and not by organic contamination.
NASA’s Stardust mission was the first mission to return solid extraterrestrial material to Earth from beyond the orbit of the Moon. These samples, which were collected from two separate sources (the coma of comet 81P/Wild 2 and the interstellar dust stream), provide unique insights into the chemistry and formation of the early solar system. Because the relative velocities of the extraterrestrial dust with the Stardust spacecraft were greater than 6 km s−1, the main challenge was to capture the dust particles with minimal alteration of their original shape and/or chemical composition. To collect the cometary and interstellar dust without damaging them, the Stardust mission used aerogel, a porous, low-density silicate glass (Brownlee et al. 2006).
Aerogel for the Stardust mission was made in batches from a single aerogel precursor (tetraethyl-orthosilicate) with processes designed to produce a density gradient with the lowest density at the front surface (Tsou et al. 2003; Sandford et al. 2010). The goal of this density gradient was to allow small particles to be decelerated as gently as possible, while still providing stopping power for grains as large as about 200 μm. Cells from 10 different aerogel production batches were placed in the cometary collector tray and cells from 19 different batches were placed in the interstellar collector tray (Sandford et al. 2010). The unused cells, including those in the flight spare trays, were collected and stored by the Stardust curator at the Johnson Space Center in a nitrogen-purged environment.
Although aerogel was crucial to the success of the mission, it is not without shortcomings. In particular, the manufacturing process involved the use of several organic solvents that left significant concentrations of carbon bonded to the aerogel primarily in the form of Si-CH3 (Sandford et al. 2010). This residual carbon has two consequences for measurements of extraterrestrial carbon: (1) it increases the background and therefore reduces the sensitivity of measurements and (2) it provides a reservoir of organics that could be modified during an impact. Indeed, there is evidence that some of the carbon indigenous to the flight aerogel was modified by the high-impact collection; however, in most cases, the signatures of returned cometary organics were readily distinguished from contamination of this form (Sandford et al. 2006, 2010; Bajt et al. 2009).
One intriguing aspect of the Stardust flight aerogel is the varying degree of visible fluorescence when exposed to ultraviolet (UV) light. The space-exposed surfaces of the aerogel tiles show the most fluorescence, and there is concern that this phenomenon arises from the exposure of the surface to an organic contaminant and/or due to radiation exposure changing some of the residual carbon into more complex fluorescence forms. Sandford et al. (2010) examined the entire cometary collector tray with a UV source and CCD camera and found that the fluorescence was not uniformly distributed across the collector tray, suggesting that the source of fluorescence was not a contaminant to which the entire tray had been exposed after tray assembly. Rather, these measurements of fluorescence intensity showed a strong correlation with aerogel production batch number. Further analysis of the aerogel tiles suggested that the fluorescence is not caused by organics, but instead by partially polymerized silica tetrahedral species having one and two nonbridging oxygen species. The number of these defect sites might be expected to increase in aerogel exposed to ionizing radiation, and could explain the increased fluorescence at the space-exposed surface, but the origin of the variable production of Si-O defects in different aerogel batches remains unclear.
Here, we investigate the fluorescence properties of space-exposed aerogel in more detail using the spatial resolution provided by fluorescence microscopy, confocal Raman microscopy, and synchrotron-based FTIR spectromicroscopy. Keystones, or wedge-shaped volumes of aerogel, were machined from the aerogel collector tiles using glass needles controlled by automated micromanipulators (Westphal et al. 2004) and mounted on small “pickleforks.” This process enables individual tracks to be removed from the collector tile for analysis, while preserving the trajectory information of the tracks and associated terminal particles. Another advantage of this technique is that the original space-exposed surface is left intact and unmodified. As shown in Table 1, we examined 24 keystones extracted from the cometary collector: Seven of the keystones contained Tracks 156–160, 162, and 165, and the remaining 17 keystones were blanks (i.e., sections of aerogel that did not contain obvious particle tracks, but still have a space-exposed surface). The keystones were extracted from numerous tiles distributed throughout the cometary collector (see Fig. S1). Additionally, we measured five keystones from the cometary flight spare, which has been stored in a nitrogen environment on Earth. These keystones were extracted from tiles that were composed of aerogel produced in the same batches as the keystones extracted from the cometary collector. We also examined several keystones from the Stardust interstellar collector, which was mounted on the opposite side of the cometary collector and experienced the same flight conditions (except for the exposure to the comet) as the cometary collector.
Table 1. Measured keystones.
Keystone depth × width (μm)
Track length (μm)
C2005, 5, 0
1700 × 1700
C2012, 11, 0
1600 × 1600
C2017, 3, 0
1700 × 1700
C2027, 8, 0
1600 × 1600
C2035, 1, 0
1600 × 1600
C2054, 38, 0
1700 × 2200
C2061, 6, 0
1700 × 1700
C2065, 3, 0
1700 × 1700
C2067, 3, 0
1700 × 1700
C2078, 3, 0
1700 × 1700
C2081, 2, 0
1700 × 1700
C2092, 9, 0
1700 × 1700
C2097, 2, 0
1700 × 1700
C2103, 20, 0
1700 × 1700
C2121, 1, 0
1700 × 1700
C2035, 4, 155
1100 × 1300
C2061, 8, 158
C2062, 1, 160
1300 × 1300
C2062, 2, 162
∼3000 × 3000
C2078, 4, 156
1300 × 1300
C2112, 2, 157
1100 × 1300
C2117, 1, 159
1600 × 1600
Cometary flight spare
E2321A, 1, 0
1600 × 1600
E2353E, 1, 0
1600 × 1600
E2353E, 2, 0
1600 × 1600
E2372C, 1, 0
1600 × 1600
E2372C, 2, 0
1600 × 1600
I1003, 1, 40
600 × 600
I1004, 3, 21
900 × 900
I1017, 4, 0
1000 × 1000
I1031, 1, 23
800 × 800
I1043, 1, 30
800 × 800
I1093, 2, 26
∼650 × ∼650
Fluorescence images were taken using three optical microscopes. A Zeiss Lumina microscope with an Hg lamp produced images that allowed us to identify the presence of fluorescence signals in a large number of keystones near the space-exposed surface. Two filters were used to collect the fluorescence: DAPI (blue) at 460 ± 10 nm, which is designed to pass 4′-6 Diamidino-2-Phenylindole fluorescence and FITC (green) at 540 ± 30 nm, which is designed to pass fluorescein isothiocyanate. The Hg lamp contains strong excitation lines extending into the UV, which appear necessary to excite the surface fluorophores.
A second fluorescence microscope, Zeiss Axioimager M1 with an Xe lamp, failed to excite the surface fluorophores even though DAPI and FITC filters were also used. The optics chain was tested using fluorescent plastics and DAPI and FITC signals were successfully obtained on these test materials. The Xe spectrum does not contain strong features in the UV, but instead has a broadband spectrum strongest near 450 nm (optical blue) and some excitation lines in the IR. Therefore, it appears that the aerogel surface fluorophores were only successfully excited by UV photons.
A third microscope, a Zeiss 50 UV/VIS Meta laser scanning confocal microscope, was also able to successfully view the surface fluorescence feature in confocal mode. Excitation photons at 458, 488, 364, and 351 nm were able to successfully produce the fluorescence and were used to restrict the position of the signal to within a few microns of the surface of the aerogel. A bandpass filter between 505 and 550 nm was used (essentially FITC). Because the point spread function (PSF) at the surface of aerogel is unknown, no PSF deconvolution was attempted.
Aerogel keystones were measured on two different confocal Raman microscopes: a Witec alpha 300R and a Thermo-Scientific DXR. The Witec alpha 300R microscope was equipped with a 532 nm laser and delivered 10.8 mW on the sample with a 50 × (NA = 0.75) objective and 20.8 mW on the sample with a 20× objective (NA = 0.40). The microscope was used primarily with a low-resolution grating that allowed simultaneous collection of the range 0–3500 cm−1 at about 3 cm−1 spectral resolution, but a high-resolution grating was also used that allowed a collection range of 1500–2500 cm−1 at about 1 cm−1 spectral resolution. The Thermo-Scientific DXR microscope was equipped with 532 nm and 780 nm lasers, and delivered 10 and 14 mW, respectively, on the sample with a 50× objective (NA = 0.50). The microscope was used with a grating that allowed a collection range from 100 to 3300 cm−1 at 5–9 cm−1 spectral resolution. For both microscopes, 532 nm excitation led to fluorescence of the sample; however, this fluorescence photobleached considerably after 2–3 mins. Linear profiles were taken by moving the stage in 5 μm steps.
Infrared transmission spectra were acquired with a synchrotron-based FTIR on Beamline 1.4.3 at the Advanced Light Source (ALS), Lawrence Berkeley National Laboratory (LBNL). The synchrotron source produces a diffraction-limited spot size, which is approximately 3 μm in the region of the C-H stretching frequencies (approximately 3000 cm−1). Spectra were typically averaged 16–128 times and collected in the 650–4000 cm−1 range with 4 cm−1 resolution using a KBr beamsplitter and a mercury cadmium telluride (MCT-A) detector. Because infrared radiation is non-ionizing and the peak and average powers on the sample are low, infrared microscopy is truly a noninvasive technique. The only risk of sample damage, therefore, is sample handling.
The keystones were mounted with the goal of orienting the space-exposed surface parallel to the optical axis of the microscope; however, because of limitations in the mounting scheme, some of the keystones were rotated as much as ±10° from the ideal orientation. Depth profile measurements of the aerogel composition were made by moving the stage perpendicular to the space-exposed surface of the keystone in 3 μm increments. Although the spot size of the infrared beam is about 3 μm in the focal plane, the lateral resolution of the profile measurements is worse because the aerogel keystone thickness (200–300 μm) is much larger than the focal depth and the infrared absorption measurements sample the entire thickness of the keystone. Indeed, with a 32× objective (NA = 0.65) and with the focus at the center of the aerogel, the spot size at the top and bottom of the aerogel can be as large as 260 μm, but with the central portion of the “out of focus” beam being obscured by the secondary mirror in the Schwarzschild objective design. We estimated the instrumental response by measuring the step-edge profile of the Si-O stretching mode at approximately 1100 cm−1, with the assumption that this functional group is relatively constant throughout the aerogel. The full-width, half-maximum (FWHM) of the first derivative of the profile is comparable to the resolution of the system (Russ 2002), and although it varied slightly from keystone to keystone, the FWHM was found to be approximately 30–50 μm. We have not attempted to deconvolve the instrumental response from the measured profiles, however, because of possible systematic errors in focus depth and orientation of the keystone with respect to the optical axis. As a consequence, the depth profiles from the infrared measurements cannot be quantitatively compared to the fluorescence measurements. The depth profiles do give qualitative information about the distribution of the functional groups within an aerogel keystone and because the keystones were all approximately the same thickness, they can be compared with each other.
Figure 1 shows fluorescent images and profiles of three keystones from the same aerogel production batch (E235). Keystones C2005, 5 (Fig. 1a) and C2078, 3 (Fig. 1b) were extracted from the cometary collector tray, whereas keystone E235-3E, 1, 0 (Fig. 1c) was extracted from the cometary flight spare aerogel blocks. Fluorescence is clearly associated with the aerogel in all the measured keystones (see Figs. S2–S4), but the intensity profiles of the fluorescence vary considerably, even among keystones from the same production batch. Many keystones extracted from the cometary collector, such as C2005, 5, exhibit strong fluorescence at the space-exposed surface, whereas other keystones, such as C2078, 3, have a more uniform fluorescence profile. Careful examination of these profiles, however, reveals that all keystones extracted from the cometary collector consist of at least two distributions: a narrow distribution that is confined to within about 10 μm of the space-exposed surface, and a second distribution that is much broader with a maximum that is typically about 30–50 μm below the space-exposed surface. The relative intensities of these distributions vary considerably from keystone to keystone, as does the width of the subsurface distribution. Moreover, the relative intensities of these distributions depend on which filter set is used to collect the fluorescence, suggesting that the two distributions originate from different fluorophores. All keystones extracted from the same aerogel production batches on the cometary flight spare, on the other hand, have much more uniform fluorescence profiles.
The fluorescence intensity from the measured keystones does not appear to correlate with the aerogel production batch, in contrast to the measurements of Sandford et al. (2010). Those measurements, however, were collected in an entirely different orientation with a much reduced spatial resolution compared with the measurements presented here. Specifically, Sandford et al. (2010) collected fluorescence from whole tiles still within the Stardust collector, thereby sampling the space-exposed surface face on, whereas the measurements presented here are of individual keystones measured in profile, and therefore sample fluorescence changes perpendicular to the space-exposed surface. The fluorescence of a given tile in Sandford et al. (2010) represents the combined fluorescence of a large number of keystone equivalent areas, and it is likely that the small scale variations seen in our measurements are averaged out by sampling a larger area. Thus, the two sets of measurements do not necessarily contradict each other.
The bulk keystone also shows a significant fluorescent intensity, which is not correlated with the surface profiles, or with the FTIR traces. The bulk fluorescent activity is not necessarily generated by the same fluorophores as the surface fluorophores, but simply has a visible intensity in the same optical range as the fluorescent filters used to view the surface fluorophores. This then provides a large background, which is not a function of the space-exposed surface. While the intensity of the surface profiles is much greater than the bulk fluorescence (a factor of several), the volume of the surface layer is many orders of magnitude less than the volume of the entire aerogel tile. Therefore, in a plan view, such surface profiles would present a contrast in Sandford’s measurement of the order of one part per thousand—clearly not a significant measurement in the plan geometry.
Figure 2 shows the fluorescence intensity profile obtained with a confocal fluorescent Zeiss 50 UV/VIS Meta microscope with a 5 μm thick confocal slice. Because of the aerogel surface roughness and varying density, the PSF at the aerogel surface is difficult to measure and is likely to change from keystone to keystone. Therefore, we did not attempt to deconvolve the PSF from the experimental profiles. For comparison of the fluorescence signal to the aerogel surface, we also obtained a non-confocal optical image. The surface of the aerogel is difficult to constrain due to the interference pattern appearing at the surface. If the aerogel was a knife-edge object, then the surface should appear at the minimum intensity of the interference pattern. With this assumption, the fluorescence signal is within a couple microns of the surface, as seen in Fig. 2. The fluorescence signal appears to have an experimental rise distance of 10 μm measured from 10 to 90% intensity approaching the aerogel surface from the non-aerogel side of the image. The decay of the signal is several tens of microns longer within the aerogel and the signal is asymmetric. However, one may also see that the unfiltered optical interference pattern is asymmetric as well, and we cannot determine whether the fluorophores are exactly constrained to the aerogel surface or lie within a few microns of it. However, we can definitely state that the fluorophores are not located more than about 10 μm below the surface for this signal.
The infrared spectra of aerogel are dominated by the intense Si-O stretch vibration at about 1200–1050 cm−1. Additional features are attributed to structural Si-OH groups (3745 cm−1), absorbed H2O (3700–3200 cm−1), aliphatic C-H groups (2990–2700 cm−1), and C=O groups (approximately 1700 cm−1) (Sandford et al. 2006, 2010; Rotundi et al. 2008; Bajt et al. 2009). The relative intensities of these features vary considerably from keystone to keystone and appear not to be correlated with aerogel production batch or to location within the cometary collector. Rather, keystones located in the same shipment/storage container appear to have similar features, suggesting that volatile contaminants are absorbed by the aerogel during shipping and/or storage. Although this situation is not ideal, the contamination appears to be uniformly distributed throughout a keystone, and is thus not a large hindrance to the analysis of intensity profiles across an individual keystone.
Figure 3 shows representative spectra from keystone C2005, 5 obtained on the space-exposed surface and the interior of the keystone. In addition to the typical aerogel features, the spectrum near the space-exposed surface has features at 2261 and 2200 cm−1. These peaks can be assigned to silicon hydride stretching motions, corresponding to O3SiH (2261 cm−1, henceforth labeled −SiH) and O2SiH2 (2200 cm−1, henceforth labeled −SiH2), and are further supported by peaks at 886 cm−1 and 916 cm−1, which correspond to the −SiH bending mode and the −SiH2 scissor mode, respectively (Lucovscky 1979; Tsu et al. 1989).
Figure 4 shows the spatial distribution of several of these key functional groups in keystone C2005, 5. Because aerogel is composed primarily of Si-O bonds, the Si-O profile (Fig. 4a) is expected to give a good estimate of the column density of the aerogel, whereas the aliphatic C-H stretches (Figs. 4b and 4c) give an estimate of the amount and complexity of the organics within the aerogel. It is readily apparent from Fig. 4c that the aliphatic CH2 concentration is higher at the space-exposed surface. Moreover, the CH3 (Fig 4b) concentration at the edge increases at a slower rate than the Si-O concentration, indicating that there is less CH3 at the edge. This result suggests that the organics near the space-exposed surface are more complex and have longer chain-lengths than in the interior aerogel.
The most striking result from Fig. 4, however, is the spatial distribution of the −SiH (Fig. 4d) and −SiH2 (Fig. 4e) groups, which are shown to be present only near the space-exposed surface. In particular, the −SiH group distribution appears to be quite narrow (FWHM approximately 55 μm), symmetric, and concentrated at the space-exposed edge. The −SiH2 group distribution is broader, asymmetric, and has a maximum below the aerogel surface before falling to the noise floor at about 600 μm into the keystone. Neither the −SiH nor the −SiH2 group could be detected at any of the other edges of the keystone.
All 24 keystones extracted from the cometary collector had detectable quantities of the −SiH and −SiH2 groups at the space-exposed surface. For keystones containing cometary tracks (see Fig. S5), the −SiH and −SiH2 distributions appear to be mostly unaffected near the track. For the −SiH2 group, there is some increased intensity at the edges of the track and decreased intensity inside the track, but these changes correlate well with the aerogel density, as measured by the O-Si-O band at 810 cm−1.
No evidence for the −SiH and −SiH2 groups was found on keystones extracted from the cometary flight spare (Fig. 5), which suggests that these surface modifications have a space origin. Analysis of more than 25 keystones from the interstellar tray as part of the Interstellar Preliminary Examination (Westphal et al. 2008; Bechtel et al. unpublished data) also failed to show these functional groups localized at the surface. Many of these measurements were performed prior to this work, but the six keystones listed in Table 1 were remeasured in a manner identical to the cometary keystones to confirm the lack of −SiH and −SiH2 groups. For most of the interstellar keystones, measurements were performed on the picokeystones (Westphal et al. 2004), which are only approximately 70 μm thick. The pathlength of these keystones was thus smaller by a factor of 4–5 when compared with the cometary keystones, which are typically about 300 μm thick. Despite this reduction in signal, we still expect the −SiH and −SiH2 groups to be detectable on the interstellar picokeystones if they were present in the same concentration. Furthermore, line scans of several interstellar keystones, including I1004, 3, 21; I1017, 4, 0; and 0 I1031, 1, 23 were performed on the main keystone (approximately 300 μm thick). They did not show any evidence of −SiH and −SiH2 groups. The keystones I1004, 3, 21, and I1017, 4, 0 did, however, show a feature at 2250 cm−1 that is uniformly distributed throughout the keystone. This feature is assignable to a C≡N stretch and is a signature of contamination from the cyanoacrylate used in mounting the interstellar keystones in Si3N4 windows (Bechtel et al. unpublished data).
The intensity and distribution of the −SiH group (narrow distribution, closer to the surface) within each keystone extracted from the cometary collector are relatively consistent: each has a similar maximum peak area (approximately 0.2 absorbance units) and the distributions are peaked at the space-exposed surface with a FWHM approximately 40–80 μm. The only exception is keystone C2054, 38 which has a much stronger maximum (approximately 0.4 absorbance units). The intensity and distribution of the −SiH2 groups, on the other hand, vary considerably, although all keystones extracted from the cometary collector have −SiH2 sub-surface maxima. In many keystones, such as C2005, 5 (Fig. 5, top trace), the −SiH2 group is more intense than the −SiH group, whereas in other keystones such as C2078, 3 (Fig. 5, middle trace), the −SiH2 peak is less intense. In the keystones with −SiH2 groups that are less intense, the distribution also appears to be much broader and penetrates deeper into the keystone. The fluctuations appear to have no correlations with sample batch, location within the collector, or shipment/storage container, and as a consequence suggest that they were caused by a heterogeneous source. Alternatively, fluctuations may be caused by the heterogeneity of the aerogel itself, where some physical property of the aerogel (density, surface roughness, hydrophobicity, etc.) makes it more predisposed to surface modifications.
The survey of the 24 cometary keystones also showed variability in the distribution of other functional groups when compared with the Si-O group. For many keystones, the Si-OH (structural OH, approximately 3745 cm−1), CH2, and CH3 profiles closely match the Si-O profiles. In other keystones, such as C2005, 5, however, the CH2 concentration is higher at the space-exposed surface, while the Si-OH and CH3 concentration is lower. With a few exceptions, the Si-OH and CH3 concentration profiles generally match each other, but are particularly low at the space-exposed surface with keystones having intense −SiH2 groups.
To obtain linear profiles of the −SiH and −SiH2 groups with better spatial resolution than with the FTIR measurements, confocal Raman measurements were performed on keystones C2054, 38, 1; C2092, 9; and C2121, 1. However, despite several hours of integration time and different excitation wavelengths (532 and 780 nm), features associated with the −SiH and −SiH2 groups could not be detected. Indeed, the only detectable features in the Raman spectra were attributed to C-H stretching and bending at about 2900 and 1450 cm−1, respectively. Profile measurements of these groups revealed little changes across the keystone.
It may not be surprising that −SiH and −SiH2 groups could not be detected with the Raman confocal microscope because spontaneous Raman cross-sections are typically much smaller than infrared absorption cross-sections. Moreover, the confocal nature of the Raman microscope probes only a small sample volume, whereas FTIR probes the entire keystone thickness. In the FTIR measurements, the −SiH peaks were typically 100 times weaker than the C-H stretching groups, whereas the signal-to-noise ratio of the C-H stretching group was only approximately 50 in the Raman measurements. Therefore, unless the Raman cross-sections of the −SiH and −SiH2 groups relative to the C-H stretching groups are significantly larger than those in FTIR, the −SiH and −SiH2 groups will be undetectable in Raman.
Origin of Surface −SiH and −SiH2 Groups
FTIR analyses of keystones from the cometary flight spare show no indications of −SiH and −SiH2 groups at the surface or anywhere within the aerogel keystones, strongly suggesting that the formation of these functional groups on the cometary flight aerogel occurred during space exposure. The exposure to intense ionizing radiation and solar wind are two candidates for the modification of the aerogel surface. Aerogel in the interstellar tray was also exposed to these sources, but −SiH and −SiH2 groups could not be detected on these keystones. Thus, exposure to the comet appears to be a likely source for the formation of the −SiH and −SiH2 groups. There are a couple of other differences between the interstellar and cometary tray, however, that may also account for the lack of −SiH and −SiH2 groups in the interstellar tray. Although both trays were exposed to ionizing radiation during flight, the space-exposure of the two trays was highly asymmetric. The longest space-exposures were during the two interstellar collection periods (total of approximately 229 days), at which point the interstellar tray was largely facing away from the sun, while the cometary tray was exposed directly to it. Thus, the cometary tray received larger doses of ionizing radiation compared with the interstellar tray, which may account for the observed differences. A second difference between the two trays was the aerogel density: the interstellar aerogel was designed to capture grains at higher velocities and therefore had a lower density at the aerogel surface (target approximately 2 mg mL−1) than the cometary aerogel (target 5–7 mg mL−1) (Tsou et al. 2003). Although individual cell densities are not well characterized, it is known that the aerogel density varies even among cells within the cometary tray. Some densities of aerogels may be more predisposed to surface modifications than others, and this may be a possible source of the heterogeneity we observe.
Regardless of the source, the distribution of the −SiH and −SiH2 groups is particularly intriguing. It is not readily apparent why −SiH would only be at the surface and why −SiH2 would penetrate deeper into the aerogel. The subsurface maximum of the −SiH2 group distribution eliminates the possibility that the surface modifications arise from a contaminant coating the surface, and instead implies an implantation or a two-stage diffusion mechanism. One possibility is that during comet exposure, the surface aerogel was reduced to form −SiH2 with a decreasing concentration into the aerogel. Then, an alternative source after comet exposure (e.g., solar wind) could have partially re-oxidized the surface to −SiH while leaving the interior and less-exposed aerogel as −SiH2.
The large penetration depth of the −SiH2 into the aerogel is also quite puzzling. For solar wind at about 600 km s−1, the penetration depth into aerogel is expected to be only about 4 μm for protons and about 10 μm for heavier ions like Fe (Ziegler et al. 2010). Implantation of ions at the lower encounter velocity (approximately 6 km s−1) of the Stardust spacecraft with the comet coma would therefore result in submicron penetration depths. Thus, it appears that the appearance of −SiH2 several hundred microns into the aerogel was caused by collisions with heavier particles or occurred through a diffusion process. The lack of defined tracks associated with the −SiH2 groups and the relatively uniform distributions on the keystone-size scale, however, seems to preclude the possibility of collisions with heavier particles. It should be noted that measurements of keystones nearly a year apart produced qualitatively similar results in terms of −SiH and −SiH2 penetration depth and intensity; thus, any diffusion process that created these functional groups has stopped or slowed considerably. Because diffusion rates are a strong function of temperature, and the temperature in the space environment is generally much lower than the Earthbound storage facility for the Stardust aerogels, this may imply that such a diffusion mechanism may be catalyzed by ionizing radiation present only in the space environment, or only during heating of the aerogel due to direct exposure to sunlight.
Ionizing photonic radiation such as solar UV cannot produce the −SiH2 profile by itself because such a process would produce a profile that was most intense at the surface, and then exponentially decaying as a function of depth in the keystone in accordance with Beer’s law. Therefore, while photonic energy may catalyze the reaction, there must also be a mass transport mechanism in play.
FTIR and Fluorescence Correlations
Fluorescence was prevalent in all keystones examined with the fluorescence microscope, even in the flight spare keystones that were never exposed to space or ionizing radiation. Enhanced surface fluorescence was also seen in keystones extracted from the interstellar tray. Because the flight spare and interstellar keystones exhibit no signs of −SiH or −SiH2, these functional groups cannot be the only source of fluorescence within the aerogel keystones. There are some qualitative correlations between the two data sets, however, that suggest that the enhanced surface fluorescence is at least related to these two functional groups: (1) both measurements show two distributions near the space-exposed surface of the aerogel, one of which is narrow and confined to the surface (−SiH), whereas the other distribution peaks just below the surface and is much broader (−SiH2) and (2) the subsurface distribution in both data sets varies in intensity and width in a similar fashion, such that for keystones that exhibit strong surface fluorescence, there is also a corresponding strong and narrow distribution of −SiH2.
These correlations indicate that the surface fluorescence is arising from structural changes within the aerogel, as suggested by Sandford et al. (2010). Nishikawa et al. (1996) have shown that γ-irradiated silica glasses exhibit photoluminescence that results from the generation of non-bridging oxygen hole centers formed by the disruption of the Si-O-Si bonds, and 29Si NMR measurements of flight spare aerogel show that high fluorescence aerogel has increased Q3 and Q2 species that are partially polymerized silica tetrahedral species having one and two non-bridging oxygen species, respectively. Clearly, areas with more −SiH and −SiH2 groups have more disruption of the Si-O tetrahedral network and it follows that these areas may be more fluorescent. The difference between the widths of the fluorescent and FTIR profiles makes the argument less convincing. However, one should not forget, as discussed above, that the large numerical aperture of the infrared objective and the thickness of the aerogel keystone distort the infrared profile measurements, making them broader than they actually are. Ayers and Hunt (1997) have also demonstrated that treatment of silica aerogels with microwave-energized reducing gases, such as hydrogen and ammonia, induces permanent, visible photoluminescence in the material. Although they did not measure FTIR spectra of their aerogel monoliths, it is conceivable that these gases formed −SiH and −SiH2 groups in the aerogel.
An alternative explanation for the fluorescence is increased organic content near the surface. For many keystones with a strong surface fluorescence, there appears to be increased CH2 functional groups at the keystone surface relative to the interior of the aerogel. However, we observe no correlation between the absolute CH2/CH3 ratio (or the CH2 and CH3 absorbance values) and the fluorescence intensity of the keystone. Moreover, the lack of aromatic or olefinic C-H stretches above 3000 cm−1 in the FTIR measurements indicates that the surface does not contain organics with conjugated double bonds, making the C-H stretches less likely to be associated with UV fluorescence.
The space-exposed surface of aerogel from the Stardust cometary tray appears to have been modified by space-exposure and possibly from the comet coma. In particular, FTIR measurements demonstrate that comet-exposed aerogel has −SiH and −SiH2 functional groups at the space-exposed surface. These groups are not detectable in aerogel from the cometary flight spare or from the interstellar collector. Both the −SiH and −SiH2 functional groups are linked to strong surface fluorescence, which is likely a result of the disruption of the Si-O tetrahedral bonding network in aerogel. The formation mechanism(s) of the −SiH and −SiH2 functional groups remain(s) unknown and the penetration depth of the −SiH2 distribution is particularly puzzling. Although these surface modifications are unlikely to have affected the capture of cometary material and their interpretation, the presence of the −SiH and −SiH2 functional groups at the surface does have potential implications for the chemical environment of the cometary coma.
Acknowledgments— We thank Scott Sanford for valuable criticism of the manuscript. We also thank Steve Ruzin and Denise Schichnes at the U.C. Berkeley Biological Imaging Facility. The Advanced Light Source is supported by the Director, Office of Science, Office of Basic Energy Sciences, of the U.S. Department of Energy under Contract No. DE-AC02-05CH11231.