Content uploaded by Mark J. Cintala
Author content
All content in this area was uploaded by Mark J. Cintala
Content may be subject to copyright.
NASA TM-98-201792
Capture of Hypervelocity Particles With
Low-Density Aerogel
Friedrich Hörz, Mark J. Cintala, Michael E. Zolensky, Ronald B. Bernhard, William E. Davidson,
Gerald Haynes, Thomas H. See, Peter Tsou, Donald E. Brownlee
April 1998
Since its founding, NASA has been dedicated to
the advancement of aeronautics and space
science. The NASA Scientific and Technical
Information (STI) Program Office plays a key
part in helping NASA maintain this important
role.
The NASA STI Program Office is operated by
Langley Research Center, the lead center for
NASAs scientific and technical information.
The NASA STI Program Office provides access
to the NASA STI Database, the largest collection
of aeronautical and space science STI in the
world. The Program Office is also NASAs
institutional mechanism for disseminating the
results of its research and development activities.
These results are published by NASA in the
NASA STI Report Series, which includes the
following report types:
TECHNICAL PUBLICATION. Reports of
completed research or a major significant
phase of research that present the results of
NASA programs and include extensive data
or theoretical analysis. Includes compilations
of significant scientific and technical data
and information deemed to be of continuing
reference value. NASAs counterpart of peer-
reviewed formal professional papers but has
less stringent limitations on manuscript
length and extent of graphic presentations.
TECHNICAL MEMORANDUM. Scientific
and technical findings that are preliminary or
of specialized interest, e.g., quick release
reports, working papers, and bibliographies
that contain minimal annotation. Does not
contain extensive analysis.
CONTRACTOR REPORT. Scientific and
technical findings by NASA-sponsored
contractors and grantees.
The NASA STI Program Office . . . in Profile
CONFERENCE PUBLICATION. Collected
papers from scientific and technical confer-
ences, symposia, seminars, or other meetings
sponsored or cosponsored by NASA.
SPECIAL PUBLICATION. Scientific,
technical, or historical information from
NASA programs, projects, and mission, often
concerned with subjects having substantial
public interest.
TECHNICAL TRANSLATION. English-
language translations of foreign scientific
and technical material pertinent to NASAs
mission.
Specialized services that complement the STI
Program Offices diverse offerings include
creating custom thesauri, building customized
databases, organizing and publishing research
results . . . even providing videos.
For more information about the NASA STI
Program Office, see the following:
Access the NASA STI Program Home Page
at http://www.sti.nasa.gov
E-mail your question via the Internet to
help@sti.nasa.gov
Fax your question to the NASA Access Help
Desk at (301) 621-0134
Telephone the NASA Access Help Desk at
(301) 621-0390
Write to:
NASA Access Help Desk
NASA Center for AeroSpace Information
800 Elkridge Landing Road
Linthicum Heights, MD 21090-2934
NASA/TM98-201792
Capture of Hypervelocity Particles With
Low-Density Aerogel
Friedrich Hörz, Mark J. Cintala, and Michael E. Zolensky
NASA Johnson Space Center
Ronald B. Bernhard, William E. Davidson, Gerald Haynes, and Thomas H. See
Lockheed-Martin Space Mission Systems and Services
Peter Tsou
Jet Propulsion Laboratory
Donald E. Brownlee
University of Washington
National Aeronautics and
Space Administration
Lyndon B. Johnson Space Center
Houston, Texas
April 1998
iv
Available from:
NASA Center for AeroSpace Information National Technical Information Service
800 Elkridge Landing Road 5285 Port Royal Road
Linthicum Heights, MD 21090-2934 Springfield, VA 22161
Price Code: A17 Price Code: A10
v
Contents
Summary ............................................................................................................................... vii
1 Introduction ........................................................................................................................... 1
2 ExperimentalObjectives ........................................................................................................ 4
3 Experimental Procedures and Limitations ............................................................................ 4
3.1 Light-Gas Gun Operations .................................................................................................... 4
3.2 Aerogel Materials, Target Manufacture, and Particle Recovery .......................................... 6
3.3 Projectile Materials and Properties ....................................................................................... 8
3.3.1 Aluminum, Metal .................................................................................................................. 9
3.3.2 Aluminum, Oxide ................................................................................................................. 9
3.3.3 Soda-Lime Glass ................................................................................................................... 13
3.3.4 Olivine .................................................................................................................................. 13
3.3.5 Pyrrhotite .............................................................................................................................. 13
3.3.6 Allende .................................................................................................................................. 17
3.3.7 Pampa-A ............................................................................................................................... 17
3.3.8 Powdered Soda-Lime Glass .................................................................................................. 17
4 Impact Experiments at Normal Incidence............................................................................. 20
5 Experiments at Oblique Impact Angles ................................................................................ 36
6 Experiments With Particles of Low Cohesion ..................................................................... 41
6.1 Cocoa Powder....................................................................................................................... 41
6.2 Clusters of Collisionally Produced Fragments ..................................................................... 45
7 Discussion ............................................................................................................................ 52
8 Conclusions .......................................................................................................................... 55
9 References ............................................................................................................................ 56
Figures
1a-c Target assembly for aerogel experiments ............................................................................. 7
2a SEM photographs of projectiles available for present study ............................................... 10
2b SEM photographs of aluminum oxide .................................................................................. 11
2c SEM photographs of soda-lime glass ................................................................................... 12
2d SEM photographs of powdered olivine ................................................................................ 14
2e SEM photographs of powdered pyrrhotite ........................................................................... 15
2f SEM photographs of powdered Allende meteorite .............................................................. 16
2g SEM photographs of Pampa-A meteorite ............................................................................. 18
vi
Contents
(continued)
Figures (continued)
2h SEM photographs of powdered soda-lime glass ................................................................... 19
3 Plot outlining the experiment matrix of the initial series of tests ........................................... 20
4a Optical photographs of penetration tracks of spherical soda-lime glass projectiles .............. 22
4b Penetration tracks of powdered Allende projectiles ............................................................. 23
4c Penetration tracks of powdered Pampa-A projectiles .......................................................... 24
5 Bifurcated tracks caused by a 50-µm Pampa-A projectile ................................................... 25
6 Track length plotted as a function of aerogel density ........................................................... 26
7 Track length in aerogels of different density produced by soda-lime glass spheres ............. 27
8a-c Typical populations of penetration tracks from Allende and Pampa-A projectiles ............... 29
9a-b Evidence of projectile fragmentation during launch in the light-gas gun ............................... 31
10 Summary of recovered residues sizes for good tracks ......................................................... 32
11a SEM photographs of typical residue of soda-lime glass sphere ............................................ 33
11b SEM photographs of typical Allende meteorite residue ........................................................ 34
11c SEM photographs of typical Pampa-A particle residue ........................................................ 35
12a-b Examples of aerogel penetration tracks produced at variable impact angles ....................... 37
13 Plot of track length versus impact angle ............................................................................... 39
14a-e Track length as a function of impact velocity and impact angle ........................................... 40
15 Summary plot of average track lengths as a function of impact angle and velocity ............. 41
16a-e Photographs of penetration features produced by mixture of cocoa powder and
aluminum spheres ................................................................................................................. 43
17a-e Detailed views of cocoa track ..............................................................................................44
18a-c Typical damage patterns on polished copper witness-plates ................................................ 47
18d-f More widely dispersed fragment impacts due to increased standoff distance ..................... 48
19a-c View of impacts into aerogel made by clustered impactors ................................................. 49
20a-b Side views of compound penetration tracks ......................................................................... 50
21a-c Optical and SEM images of damage by fragmented and molten projectile ......................... 51
22 Summary of past and present experiments ........................................................................... 54
vii
Summary
Impact experiments into low-density (0.01-0.05 g/cm
3
) SiO
2
aerogel targets have been conducted using a
variety of spherical projectiles, including 2024 aluminum, aluminum oxide, and soda-lime glass, irregular
particles of the Allende and Pampa-A meteorites obtained by crushing, and single-crystal fragments of olivine
and pyrrhotite. All projectiles were between 43 and 54 µm in diameter, and impact velocities ranged from
3 to 7 km/s. Our objective was to evaluate the performance of low-density aerogels for inclusion in space-
borne flight instruments whose aim is to capture natural and man-made hypervelocity particles in the least
destructive fashion, hopefully without melting. Such capture media were recently exposed on space station
Mir and are under development for Stardust, a Discovery-class mission. These captured-particle residues will
be returned to Earth for detailed mineralogic, chemical and isotopic analyses to characterize their respective
source areas and formative processes. Previous studies demonstrated the basic utility of aerogel in accom-
plishing these capture objectives. In an attempt to simulate actual flight-conditions with improved fidelity,
the current experiments employ lower-density aerogels, as well as smaller projectiles than most previous
work.
Our experiments show that definition of the initial impact conditions for such small impactors (<100 µm
diameter) poses a substantial experimental challenge when using light-gas guns and associated shot-gunning
methods for particle acceleration. Individual projectiles, including precision microspheres of glass and
aluminum, will inherently vary in size and mass proportionate with the mesh-openings of the sieves em-
ployed. In addition, all projectiles manufactured by the crushing of coarser materials, whether man-made or
natural, will contain microcracks and other flaws that cause individual particles to fragment during launch,
thereby inhibiting any control of the exact projectile size, mass, and shape. As a result of these experimental
limitations, past and present experiments are characterized by poor reproducibility and large data scatter.
Consequently, there is currently no reliable experimental basis to utilize penetration tracks in aerogels for the
reconstruction of the initial mass or encounter velocity of man-made orbital debris or natural cosmic-dust
particles as already noted and emphasized by Burchell and Thomson (1996). Improved control of the initial
size, mass, and shape of the projectile(s) is needed to yield more reproducible experiments and a more
reliable basis to understand the deceleration of very small hypervelocity impactors by targets of very low
density and exceptionally high porosity.
Nevertheless, we found that the density of aerogel affects the length of the resulting penetration track. For
soda-lime glass spheres of 50 µm diameter at 6 km/s, the track length (l) normalized to projectile diameter
(Dp) is l/Dp = 60-80, 120-150, and 400-450 using aerogels of 0.1, 0.05 and 0.01g/cm
3
density, respectively.
The relationship of track length to impact velocity is complex, as was previously observed for porous Al
2
O
3
(Werle et al., 1981) and Styrofoam (Tsou, 1990) targets. Our experiments showed that 50-µm glass spheres
penetrate deeper at 5 km/s than at 6 km/s. In addition, we conducted experiments at oblique impact angles
(30
°, 45°, and 60° from normal incidence) while holding the density of the aerogel constant (0.02g/cm
3
) and
varying the impact velocity from 3 to 6 km/s. The resulting penetration tracks faithfully recorded the angle of
incidence relative to the collector surface, as earlier reported by Thomson (1995), yet the absolute track
length was unaffected by impact angle. This behavior is unlike that found with projectile penetration into
higher-density targets, where the absolute crater depth depends strongly on the impact angle. These observa-
tions lend strong support for the model of Anderson and Ahrens (1994), which suggests that most deceleration
of hypervelocity particles in highly porous, low-density targets is governed by viscous drag rather than by
classical shock and cratering processes.
viii
In addition, we simulated the impact of fluffy natural particles, presumably very friable and porous, for which
there is some concern about their efficiency of capture. Two approaches were used. One employed com-
pressed cocoa powder, which was loaded into the sabot cavity, that did not disperse sufficiently during free
flight resulting in small clumps (on the scale of 100 µm) impacting the aerogel target. These clumps produced
characteristic short and stubby tracks that were lined with remnants of cocoa powder. The second approach
employed a 4-µm-thick aluminum bumper foil that was easily penetrated by 50-µm glass spheres, yet was
sufficiently thick to disrupt each projectile, thus generating a cloud of projectile fragments that could be
intercepted with the aerogel target downrange. These aerogel targets were located at variable standoff
distances (L) from the bumper foil such that either tightly clustered or widely dispersed fragment clouds
would impinge on the aerogel. The resulting impact features faithfully record the detailed mass distribution
of these fragment clouds, as evidenced by tightly clustered or (arbitrarily) dispersed penetration tracks of
widely variable depth. Projectile residues were found at the termini of many of these subsidiary tracks.
These experiments show aerogels efficiency in capturing analyzable residue of hypervelocity projectiles,
including fluffy particles at impact velocities of ~6km/s, the applicable encounter velocity for the Stardust
spacecraft with Comet Wild 2. In addition, differentiation between metallic and oxidized aluminum-rich
impactors in low Earth orbit, the major objective of the orbital debris collector on Mir, should be readily
accomplished at the somewhat higher mean encounter velocities (9-10 km/s) that are typical for man-made
debris in low Earth orbit. Finally, the direction of impact relative to the collector surface can be deduced
from the inclination of individual penetration tracks. However, current empirical or theoretical understanding
of aerogel penetration processes is insufficient to reconstruct projectile mass and impact velocity from the
measured dimensions of penetration tracks and/or the mass of the residue. If required by scientific need,
other concepts are available to independently determine projectile mass and detailed trajectory elements on a
particle by particle basis in free flight, prior to impact and capture by aerogel (e.g., Auer and Bun, 1994).
1
1. Introduction
This report summarizes laboratory impact experiments aimed at evaluating the suitability of SiO
2
aerogel for
the collection of hypervelocity particles on space-exposed flight instruments. As recently summarized by
Tsou (1995), such aerogels constitute a unique class of low-density, highly porous materials (Fricke, 1988;
Hrubesh and Poco, 1990) that are capable of decelerating high-velocity particles without substantial melting,
or other modifications of their component materials. This is in stark contrast to high-velocity impact into
dense targets that result in complete melting, if not vaporization, of the impactor during the formation of a
hypervelocity impact crater. Consequently, such aerogels appear to be suitable for the capture of natural
cosmic-dust particles and/or man-made debris in such a manner that their residues may be returned to Earth
for a wide variety of detailed mineralogic, chemical, and isotopic analyses. Thus, aerogel capture is a promis-
ing approach for detailed investigation of the natural and man-made particle environment in space, and the
asteroidal, cometary, or man-made sources of those particles (e.g., Tsou et al., 1988; CDCF, 1990; Zolensky,
1994; Brownlee et al., 1994).
Specifically, this report summarizes recent experiments conducted in the Experimental Impact Laboratory at
the Johnson Space Center (JSC) in support of two flight instruments. One of these instruments, the orbital
debris collector (ODC), was recently returned following an ~18-month exposure on the space station Mir,
while the other will be on board the Stardust spacecraft, a Discovery-class mission to comet Wild 2 (Stardust
experiment).
ODCs primary purpose was to evaluate the man-made particle environment in Mirs high-inclination orbit
(51
°) for comparison with instruments exposed by the Shuttle and the Long-Duration Exposure Facility
(LDEF), both of which represent modest inclination orbits (28
°). LDEF revealed numerous types of natural
and man-made particles, including a common subgroup that is composed entirely of aluminum (Al), either
metallic or oxidized, and that must derive from man-made sources in highly elliptical orbits (Bernhard et al.,
1996). While ODC indiscriminately trapped all types of particles in low Earth orbit (LEO), its major objec-
tive is to assess the relative frequency of Al particles, and specifically, the relative roles of metallic versus
oxidized species. ODC was deployed on the Mir complex on March 23, 1996, and was retrieved in October
1997. It exposed an array of 72 individual SiO
2
aerogel samples, each 10 x 10 cm in surface area and
~12 mm thick, all of 0.02 g/cm
3
bulk density.
The Stardust spacecraft is currently scheduled for launch in 1999 and will fly through the coma of Comet
Wild 2 in 2003 at a distance of about 100 km and a velocity of approximately 6 km/s relative to the comet
nucleus. This is an extraordinarily low-encounter velocity, well within laboratory impact-simulation capabili-
ties, and in a velocity regime for which the successful recovery of unmolten projectile residue has been
demonstrated by a number of experimental studies (e.g., Tsou et al., 1988; Barrett et al., 1992; Mendez,
1995). After comet passage, the Stardust spacecraft will return to Earth orbit in 2007. The aerogel collectors
will be housed inside an atmospheric reentry capsule that will be parachuted for recovery on land. Commen-
surate with its significance as the first extraterrestrial sample-return mission since Apollo, optimization of the
Stardust aerogel collector medium constitutes a high-priority development effort, both experimental and
theoretical. In situ measurements and observations of comet Halley dust by the Giotto Spacecraft (Kissel
2
etal., 1986; Grewing et al., 1986; Jessberger et al., 1989) suggest that a wide range of particle sizes may be
encountered, as well as a range of particle compositions and physical properties.
The basic objectives for the development of suitable aerogel-collector materials for the Mir and Stardust
instruments relate to a detailed understanding of the deceleration process and the associated thermal environ-
ment that controls the yield of unmelted residue mass. How will these deceleration processes affect the
mineralogic, chemical, and isotopic characteristics of the initial impactors and how much mass will be lost in
the process? Recovery of hypervelocity projectiles varying widely in size, mass, and chemical and physical
properties is a prerequisite for any successful particle collector. Consequently, the potentially nondestructive
deceleration of highly diverse natural and man-made particles in space by (ideally) a single target medium
constitutes a most challenging development task.
This report summarizes a variety of pilot-type impact experiments recently conducted at JSC to arrive at a
general overview of the performance of SiO
2
aerogel (Hrubesh and Poco, 1990), or simply aerogel through-
out this report for the sake of brevity. The objective was to identify and define first-order issues for addi-
tional development, if needed. As a consequence, a variety of topics were addressed that had been recognized
as potentially important in previous aerogel studies, or on the basis of general cratering studies in dense-target
media. Clearly, some of the present work was guided by earlier studies using aerogel targets (e.g., Tsou et al.,
1988, 1995; Barrett et al., 1992; Mendez, 1995; Thomson, 1995; Burchell and Thomson, 1996). In addition,
non-SiO
2
, low-density, highly porous materials have been used previously, and have provided important
general insights into the capture performance of low-density target media such as porous Al
2
O
3
(Werle et al.,
1981), Styrofoam (Tsou, 1990), and other foamed polymers (Maag et al., 1992). Cumulatively, these experi-
ments demonstrated that the recovery of unmolten projectiles and/or fragments is feasible in highly porous
materials at velocities as high as 7km/s. Given the relatively small number of past experiments, combined
with a wide range of projectile sizes, materials, and target-media and associated densities, it is difficult to
synthesize the existing work into a systematic database that could guide, with confidence, the design and
interpretation of the Mir and/or Stardust flight collectors; a similar conclusion was reached by Tsou (1995),
and Burchell and Thomson (1996).
The present experiments were conceived to complement these previous efforts by employing smaller projec-
tiles, typically 50 µm in diameter, than previous studies (typically >100 µm, if not >1000 µm). Experimenta-
tion with such small projectiles is desirable for the realistic, dimensional scaling of projectile size and physi-
cal target structure. Aerogels are composed of irregular chains and clusters of SiO
4
-tetrahedra that are
typically 4-6 nm thick and 20-30 nm long (e.g., Fricke, 1988). The smaller the particle to be captured, the
larger the relative size of these SiO
4
clusters and the more pronounced potential shock effects and associated
heating of the projectile. Therefore, it is important to evaluate the utility of aerogel with projectiles that
approach or duplicate the size(s) of particles in space. We also employed aerogels of lower density, 0.01 to
0.05 g/cm
3
, compared to most of the previous work at >0.05g/cm
3
; materials <0.05 g/cm
3
are now readily
available, yet were somewhat experimental and of low transparency in the past (see also Tsou, 1995). Fur-
thermore, we experimented over a wider range in velocity (3-7 km/s) than most of the previous studies, and
we investigated the effects of oblique impact. Some experiments simulated for the first time the impact of
3
projectiles of low-cohesive strength. The overall objective was to conduct experiments with a high degree of
fidelity relative to space-exposed aerogel collectors, and to vary some initial impact parameters (e.g., velocity
and aerogel density) in a more systematic fashion than was the case in most of the earlier studies. The
expected result was an improved empirical and theoretical understanding of the deceleration process in aerogel
and of associated effects regarding possible particle modification(s).
The most comprehensive set of experiments with low-density targets that is available employed Al projectiles
and Styrofoam targets (Tsou, 1990). These data were utilized by Anderson and Ahrens (1994) to arrive at a
first-order model of deceleration processes in highly foamed, low-density media, which describes the capture
of hypervelocity particles in such materials as the result of shock processes and those governed by classical
continuum mechanics, including viscous drag and ablation. While the Anderson and Ahrens (1994) model is
not readily transferred to SiO
2
without additional model assumptions, it is obvious that projectile size as well
as bulk target density and internal structure will play crucial roles. It is for these reasons that we experi-
mented with very small impactors and employed aerogels of variable density, duplicating those exposed on
Mir and contemplated for Stardust.
Unfortunately, the experiments described below must be considered somewhat qualitative because they are
afflicted with some experimental uncertainties, as detailed throughout this report. Most of our experimental
data are characterized by substantial, if not intolerable, scatter, akin to that emphasized recently in the most
systematic aerogel studies to date by Thomson (1995) and Burchell and Thomson (1996). Evidence suggests
that this scatter is largely due to experimental difficulties in duplicating given sets of initial impact conditions,
notably constant size, mass, or shape of very small, individual projectiles. Such projectiles can only be
launched as an ensemble of particles in light-gas guns, and therefore can be characterized only by a range of
(sieved) sizes without good control of mass or shape on a particle-by-particle basis. To complicate matters,
many projectile materials tend to break apart during light-gas gun acceleration, producing a wide spectrum of
impactor sizes and masses. In addition, reproducibility of exact aerogel density may be difficult to accom-
plish. As a result, variations in target density may be a contributing factor in producing large variances from
experiment to experiment. Having recognized some of these experimental difficulties, we hope to eliminate
some of them in the near future to improve on the reproducibility of individual experiments.
This document represents an informal progress report. Emphasis is on photographic documentation and some
first-order dimensional analysis of penetration tracks and projectile residues. As stated above, a wide variety
of independent topics will be addressed in this report. Some might be woefully incomplete in the readers
view, yet they are consistent with pilot-type experimentation and associated evaluations. The report is in
approximate, chronological sequence, as we wish to guide the reader through some of the experimental
difficulties in approximately the time-sequence in which they were encountered and recognized. Also note
that the existence of projectile residue at the end of individual penetration tracks was only determined, in
many cases, via binocular microscope. Few residues were physically recovered, and even fewer were
analyzed in detail using scanning electron microscope (SEM) and associated energy dispersive X-ray spec-
troscopy (EDS) or transmission electron microscope methods; such detailed analyses were not within the
scope or resources of the current effort.
4
2. Experimental Objectives
The experiments conducted to date fall into two categories along the following major objectives:
(1) Qualitatively demonstrate that the collection of unmolten projectile residues of silicates is feasible at
impact velocities applicable to Stardust (~6.1 km/s), and explore the effects of aerogel density on the
efficiency of capture and the degradation of initial particle properties. In addition, determine how well
particles can be captured that are of relatively low cohesion and high porosity, and which are thought to
be relatively abundant among cometary dust grains. Establish the survivability of metallic- and oxidized-
aluminum impactors in aerogel at impact velocities of 7 km/s in support of ODC, and compare the
residues of brittle silicates, ductile metal, and highly refractory oxide impactors. In short, this objective
related to the empirical recovery and mineralogic-petrographic characterization of analyzable residues of
a diversity of projectiles.
(2) Collect a suitable experimental database
over a wide range of aerogel density, impact velocity, and
projectile materials
for the development of theoretical models and a more general understanding of the
physical processes that govern particle deceleration and associated heating in aerogel targets. These
highly parametric tests may be qualified as calibration experiments as they relate the initial impact
conditions to the dimensional characterization of the resulting impact features and associated projectile
residues.
Most experiments documented in this report relate to both objectives, which is why implications for the
capture and calibration objectives are not separated rigorously throughout this report. In this report, Sections
1 and 2 provide background, while Section 3 details experimental procedures, including target preparation
and projectile selection. Sections 4 through 6 describe the experimental results. Section 4 addresses the
length of penetration tracks at normal projectile incidence and as a function of aerogel density, impact veloc-
ity, and projectile materials; it also introduces some of the limitations of launching a collection of small
particles with light-gas guns. Section 5 details the length of penetration tracks as a function of impact angle
and velocity at otherwise constant impact conditions, while Section 6 describes impacts by projectiles of low-
density/low-cohesive strength. Section 7 provides a general discussion of current and past results.
3. Experimental Procedures and Limitations
3.1 Light-Gas Gun Operations
Impact experiments were carried out with a small-caliber (5 mm), light-gas gun that was used as a high-
velocity, small-scale version of a shotgun. Ensembles of 50-100 projectiles (50 µm in diameter) were acceler-
ated via a four-piece, serrated sabot. Before launch, the projectiles reside in a 0.5-mm-diameter, 2-to 3-mm-
long cylindrical cavity at the front of the 5-mm-diameter, 4- to 6-mm-long sabot. Upon exiting the muzzle,
the petals of the sabot and most of the projectiles radially separate, being stopped by collision with a massive
steel aperture about 5 m downrange. A concentric hole in the aperture permits only the central cluster of
projectiles to proceed downrange to impact the target, which is about 4 m further downrange from the aper-
ture. This latter flight section contains additional mechanical apertures, such that a relatively small number of
projectiles
typically 2-5 and occasionally as many as 10encounter the target which, in this case, was
5
monolithic aerogel squares of variable thickness measuring roughly 5 cm x 5 cm. The majority of hydrogen
driver-gas used to accelerate the sabot is blocked with a mechanical flapper valve. Carbon-rich vapors and
particulates from a variety of sources, including the sabot, high-pressure diaphragm, pump-tube piston, and
launch-tube abrasion, are largely trapped by additional mechanical apertures. As a result, the delicate aerogel
surfaces remain undisturbed and sufficiently clean to permit direct transfer to an SEM for imaging purposes
and associated chemical analysis by EDS.
The front face of the aerogel target is monitored by a photodiode that records the light flashes of individual
projectile impacts (at times t
1
, t
2
, t
3
,t
n
) via digital-storage oscilloscopes (Tektronix 2232). The entire
velocity-monitoring system is initiated (T
0
) when a laser beam about 2 mm in front of the gun muzzle is
occulted by the sabot slug extruding from the launch tube. Four additional laser/photodiode stations monitor
the sabot pieces in free flight (providing times T
1
, T
2
, T
3
, T
4
), and an additional diode station monitors the
light flashes of the sabot impacts (T
5
). Generally, velocities deduced from times T
1-5
agree with each other to
<0.5%, and with the impact-flash data (t
x
) to <1%, suggesting no measurable deceleration of the impactors at
our typical chamber pressures of 1-2 x 10
-3
torr. There is no capability to monitor the X/Y location of specific
projectiles at the target front face, and specific impact features cannot be associated with a specific impact
flash.
Two general types of oscilloscope records characterize the present experiments. The first and ideal type
consists of sharp, clearly resolvable peaks, attesting to the arrival of a small number of projectiles, typically
fewer than five. The arrival times of all projectiles are generally spread over ~30 µs, leading to <2% variance
in velocity among individual impactors. The aerogel targets associated with such good records generally
display a small number of well-defined tracks. The second type of oscilloscope record consists of a relatively
broad hump of long duration, commonly >100 µs, that may or may not contain discrete spikes indicative of
individual impactors. Invariably, the aerogel targets associated with such oscilloscope traces display numer-
ous aerogel tracks of highly variable length. As a consequence, such records indicate the presence of a large
number of very small projectiles that display a substantial spread in velocity that is as high as 20%, but more
typically about 10%. Such prolonged and broad oscilloscope records, combined with the physical evidence in
the aerogel targets, attest to pervasive fragmentation of projectiles during launch. Such records are the rule
when launching powdered, natural materials and may even occur for some metallic aluminum projectiles at
velocities above 6km/s.
This breakup of projectiles takes place during actual light-gas gun launch where g-forces >10
6
g apply.
Breakup seems unavoidable for projectiles manufactured by mechanical crushing of coarse raw stock, be-
cause such particles are invariably weakened by microcracks or deeper fractures. We expended considerable
effort attempting to eliminate or decrease this projectile fragmentation by varying launch parameters such as
the sabot material and length, sabot-cavity dimensions, number of projectiles, high-pressure diaphragm burst
pressure, piston velocity, and initial driver-gas pressure without measurable success. In addition, we experi-
mented with a number of fine-grained powders (<1 µm) that were mixed in variable weight ratios with the
projectiles in an attempt to provide a cushion between neighboring grains; unfortunately, some powder would
invariably cling to individual projectiles causing compound tracks unsuitable for calibration purposes. From
these tests, we found that commercially available cocoa powder worked best (see Section 6). Shots were also
6
made with pure cocoa powder such that isolated clods impacted the aerogel targets, thus simulating impacts by
relatively porous, fluffy particles of low-cohesive strength.
3.2 Aerogel Materials, Target Manufacture, and Particle Recovery
All aerogel materials used in this study were manufactured at the Jet Propulsion Laboratory, Pasadena,
California (Hrubesh and Poco, 1990; Tsou, 1995). Typically, the aerogel was cast as tiles about 10cm x
10cm square and about 1 cm thick. The bulk density of aerogel is manipulated via the stoichiometric ratio of
SiO
2
and solvent during aerogel manufacture. For our experiments, we determined the specific density of
each tile by measuring and weighing each sample with an accuracy of no better than 10%, as tile thickness is
difficult to measure and, on occasion, uneven. There is no detailed knowledge of possible density variations
within a given aerogel tile, although some of our experiments suggest modest internal variations (as detailed
below). An initial batch (Shipment I) of aerogel tiles consisted of 2.5-cm-thick monoliths ranging in density
from 0.01 to 0.05 g/cm
3
. These materials were used for the experiments assessing aerogel density, impact
velocity, and a variety of projectile materials, including powdered chondrite meteorites Allende and Pampa-A.
A second shipment (Shipment II) of aerogel consisted of material of nominally 0.02 g/cm
3
density; these
materials were largely used for the oblique impacts and part of the cocoa shots. A third shipment (Shipment
III), also of nominal density of 0.02g/cm
3
, was consumed during the experiments addressing the effects of
collisionally disaggregated projectiles. As detailed below, distinction between these three shipments seems
warranted.
Actual targets were typically 5 cm x 5 cm square after cutting the original tiles into four quadrants with a wire
saw, which employed diamond-studded copper wire about 150 µm in thickness. This method allows reliable,
rapid, and precise cutting without any lubricant or coolant, thus minimizing contamination. The sawed
surfaces are rough, however, and of such decreased transparency that optical identification and measurement
of penetration tracks becomes impractical.
Each target was placed into a dedicated, clear-plastic specimen box with the hinged lid removed (Fig. 1).
Some spring-loaded, thin cardboard was first dropped into the empty box and the aerogel tile was placed on
top of this platform, such that it modestly protruded above the open face of the specimen box. The plastic
boxes were then placed into a cutout in the center of a Plexiglas disc, which constituted the interface with the
remaining impact chamber hardware. The box is held in place by the protruding hinge knobs and the knob
associated with the latching mechanism of the plastic boxes. This allows a spring-loaded frame, also made of
cardboard, to push against the protruding aerogel until it is flush with the box opening. The spring-loaded
frame and its tensioner screws (in corners) were mounted to a Plexiglas hold-down frame (Fig. 1) which, in
turn, is mounted (two slotted, black screws) to the Plexiglas disc at a controlled distance. The entire device
(Fig. 1b) is loaded into the guns impact chamber. Following an experiment, the plastic box is removed from
the target-holder disc and the lid of the plastic box is reattached; at this stage the target box becomes the long-
term storage container for the aerogel target. Handling, viewing, and first-order optical inspection via binocu-
lar microscope are readily performed without removing the aerogel from the container.
7
Figure 1a-c. Target assembly for aerogel experiments. Note the spring-loaded (lower) platform and (upper) frame in
Fig. 1b; this arrangement permitted the aerogel sample to be firmly positioned so that the surface was flush with the
opening of the plastic sample container. The overall arrangement provided for convenient and highly reproducible
positioning of the aerogel target relative to gun axis. The large Plexiglas disk served as the interface with a target-
mounting fixture inside the impact chamber.
8
Similar mounting procedures were used when vertically mounted slabs of aerogel were required (Fig. 1c),
which was commonly the case when track length was expected to be deeper than the tiles original thickness.
For such experiments, the 5-cm-x-5-cm tiles were cut in half, and the slabets were mounted on edge, thus
resulting in targets ~2.5 cm thick. In such cases, two (Fig. 1c), three, or four such aerogel slabs were placed
side-by-side inside the plastic box and held in place by multiply folded, springy cardboard.
These procedures were developed to permit fairly precise and reproducible cutting of large aerogel speci-
mens, and to provide for easy manipulation of the delicate materials during a number of experimental steps,
including reproducible mounting and orientation of the target relative to the gun axis and projectile path.
At the scale of the current experiments, many tracks measure 10 mm to 20 mm in length, and 1mm to 3mm
in maximum width. As a result, it is possible to take a sharp razor blade, position it across the entrance hole
of an experimental track, and split the entire target specimen along the defect represented by the track itself.
While the track is not perfectly halved by this procedure, its interior is generally well exposed and amenable
to detailed morphologic studies, including SEM analysis, without requiring additional sample preparation
steps. Using this track-splitting method, any captured particle or its residue will reside at or very close to a
freshly created, generally very transparent fracture surface. Consequently, the particle can be readily spotted
under a binocular microscope and removed from the aerogel with a single-hair bristle.
We are confident that this dry recovery method will also work for large tracks in space-exposed collectors, yet
there must be a size cutoff below which the splitting of individual, small tracks by razor blade will fail. This
cutoff must be determined empirically for aerogels exposed on Mir and Stardust, as such a dry method of
recovering projectile residue is extremely attractive from a contamination point of view, as well as being
fairly rapid.
3.3 Projectile Materials and Properties
As previously mentioned, one of the most troublesome difficulties we encountered during experimentation is
the breakup of some projectiles during launch. Projectiles, including our costly, commercially available, so-
called precision spheres, vary in size, mass, and even shape. To minimize such effects, we routinely sieve
(multiple times) the commercially acquired materials as well as those generated within the laboratory via
grinding; sieves with mesh openings of 43 µm and 54 µm are used for this purpose. The inability to control
and reproduce exactly the mass of what we refer to as 50-µm particles represents a serious handicap in
establishing the precise, initial impact conditions on a particle-by-particle basis. Projectile radius may vary
by factors of 1.25, implying mass variations of up to a factor of 2. In addition, detailed particle shape remains
uncontrolled during sieving, yet it will substantially affect penetration processes (e.g., Hohler and Stilp, 1987;
Orphal et al., 1990). To date, we have not conducted quantitative analysis of particle-size distributions and
shape factors for our present projectile materials, as we are in the process of attempting to acquire superior
materials to those presently in use. Regardless, we either suspect (e.g., Burchell and Thompson, 1996; see
below) or know (e.g., Barrett et al., 1992) that uncertainties in the initial impactor conditions can greatly
affect the experimental results (i.e., track length, recovered projectile mass, etc.).
9
The rationale for actual selection of projectile species is given below, together with some photodocumenta-tion
of typical projectile sizes and shapes. In addition, some brief evaluations concerning the ease of projectile
fragmentation at 6 km/s are given, loosely defined as event rate per individual gun firing. A nominal event
implies a straight, slender track that contains clearly discernible residue at the terminus; as discussed below,
they are usually differentiated from possible background tracks that formed by projectile fragments and/or the
occasional gun debris.
3.3.1 Aluminum, Metal (Fig. 2a)
The rationale to experiment with metallic aluminum relates to the Mir objective of characterizing structurally
disintegrated metallic aluminum. It is also a metal that is well-characterized in terms of physical properties
for modeling purposes. Among a variety of commercial products, we selected Al
2024
spheres on the basis of
uniform size and relatively uniform, smooth surface. Nevertheless, some ellipsoids are present, along with
(rare) compound particles that consist of numerous smaller particles welded together.
In general, aluminum can be launched at velocities below 5 km/s without difficulty, while some modest
projectile breakup may occur above 5 km/s, becoming much more common at 6km/s. During a typical
experiment, the collection of projectiles will consist of 50 to 80 projectiles loaded into the sabot, which
typically yields 2 to 5 projectiles on target.
3.3.2 Aluminum, Oxide (Al
2
O
3
; Fig. 2b)
Precision ruby spheres, 130 µm in diameter were selected and acquired in direct support of ODC. Clearly,
this material is of great theoretical interest, as its high melting temperature and compressive strength render it
highly suitable to study penetration processes of unfragmented impactors that may not have suffered exten-
sive ablation. Among the diverse projectile materials tested, Al
2
O
3
is the most refractory and should be the
least modified during penetration.
Considerable effort was made to obtain suitable 50 µm Al
2
O
3
spheres; unfortunately, this effort was unsuc-
cessful. In addition, we evaluated diverse grinding materials, but these were disqualified due to extreme
variation in particle shape; powdered corundum (Fig. 2b) was deemed unsuitable for this reason. The use of
solid-fuel rocket exhaust products as projectiles was explored, but most of those particles >20µm in size are
typically compound, sintered, highly heterogeneous aggregates.
High quality (dimensions and purity) Al
2
O
3
spheres 130 µm in diameter are readily launched, typically
yielding more than two events per experiment on target. When testing powdered corundum, we observed
projectile breakup, demonstrating that crushing via pestle and mortar generates penetrative cracks in most
materials, causing individual particles to fail during launch.
10
Figure 2a. SEM photographs of the projectiles available for the present study, as well as that of Barrett et al. (1992). Note
the differences in size and shape between these carefully sieved materials, especially those that were manufactured by
crushing: (a) Metallic aluminum, series 2024, (b) aluminum oxide, (c) soda-lime glass, (d) olivine, (e) pyrrhotite, (f) Allende
meteorite, (g) Pampa-A meteorite, and (h) powdered soda-lime glass.
11
Figure 2b. SEM photographs of aluminum oxide.
12
Figure 2c. SEM photographs of soda-lime glass.
13
3.3.3 Soda-Lime Glass (Fig. 2c)
The anticipated comparisons in the penetration behavior of metals, oxides, and silicates necessitated a spheri-
cal silicate. We obtained precision spheres of soda-lime glass (Fig. 2c); even such materials contain a fair
number of ellipsoids and non-spherical particles, or spheres that display internal flaws and accretionary
microspheres. For the purposes of some select calibration shots, such flawed spheres were culled under the
optical microscope. Soda lime is the best spherical silicate we were able to find.
Soda-lime glass spheres are easily launched (more than two impacts per experiment) over the full range of
velocity (3-7 km/s), yet occasional breakup of individual particles, most likely flawed by vesicles or
microcracks, is unavoidable. Clearly, this projectile species represents the best silicate to be launched, as our
other silicates consisted of powders crushed from coarse raw stock.
3.3.4 Olivine (Fig. 2d)
Olivine ranks among the most abundant rock-forming minerals in extraterrestrial materials. Polycrystalline
dunite (99% olivine, from Twin Sisters, Washington) was ground in a mortar and repeatedly sieved. As
illustrated in Fig. 2d, particle shapes and masses vary considerably. Among all natural materials used, this
material may be viewed as an endmember in terms of physical toughness and resistance to fragmentation and
comminution (Cintala and Hörz, 1992).
Olivine fractures noticeably during launch at all impact velocities; up to 6 km/s the total yield of nominal
particles averages better than one event per experiment, second best for all silicates in our projectile
inventory.
3.3.5 Pyrrhotite (Fig. 2e)
A number of single-crystal pyrrhotites were generated via grinding and, despite repeated sievings, the particle
shapes and masses vary considerably (Fig. 2e). This material represents abundant Fe-S-phases in comic-dust
and meteorite specimens. It seemed particularly suited to reveal potential, selective loss of relatively volatile
S in the captured residues, thereby illuminating details of the thermal environment during aerogel penetration
and th-e potential loss of elements by selective volatilization. At 6 km/s, pyrrhotite breaks up more readily
than olivine, yielding, on average, one nominal event per experiment.
14
Figure 2d. SEM photographs of powdered olivine.
15
Figure 2e. SEM photographs of powdered pyrrhotite.
16
Figure 2f. SEM photographs of powdered Allende meteorite.
17
3.3.6 Allende (Fig. 2f)
Chunks of the Allende meteorite, a carbonaceous chondrite, were ground in the laboratory and, despite
repeated sieving, the particle size and shape varied considerably (Fig. 2f). In addition, the mineralogical
mode seemed to vary from grain to grain, consistent with the modal abundance and grain-size distribution of
component minerals, chondrules, and matrix. This Allende powder must be considered a high-fidelity miner-
alogical analogue to prospective natural particles; as meteorites go, Allende constitutes a moderately tough
rock.
Powdered Allende samples are difficult to launch at all velocities, and most mass loaded into the sabot cavity
breaks up; an average of all experiments yields less than one nominal impact per experiment.
3.3.7 Pampa-A (Fig. 2g)
Pampa-A is another ordinary chondrite that was ground into projectile material; it is noticeably more compact
and tougher than Allende. The idea was to attempt to generate projectiles that would not fragment during
launch. As illustrated (Fig. 2g), particle shape and size varied considerably.
Powdered Pampa-A is similar to Allende in launch characteristics, typically yielding less than one nominal
impact per experiment.
3.3.8 Powdered Soda-Lime Glass (Fig. 2h)
We have not yet experimented with this material, but we are considering it and the powdered corundum of
Fig. 2b to assess the effects of projectile shape on penetration depth. These two powders will be paired with
observations from spherical impactors to illuminate the effects of particle shape.
18
Figure 2g. SEM photographs of Pampa-A meteorite.
19
Figure 2h. SEM photographs of powdered soda-lime glass.
20
4. Impact Experiments at Normal Incidence
This section describes the experimental series aimed at evaluating the effects of aerogel density on track
length, and especially on mass loss of the projectile during hypervelocity capture. The experiments of
Barrett et al. (1991) that used aerogel densities of 0.02
and 0.04 g/cm
3
suggested superior residue recovery
compared to the 0.12-g/cm
3
case. Thomson (1995) also employed relatively high-density aerogels (0.09 and
0.12 g/cm
3
), recovering, at best, about 10% of the initial impactor mass. The present tests concentrate on
aerogels with densities below 0.050 g/cm
3
, with the majority of experiments conducted at 0.02 g/cm
3
. All
aerogel specimens for this series were exceptionally clear and homogeneous materials (Shipment I), with
nominal densities of 0.050, 0.040, 0.035, 0.020, and 0.010 g/cm
3
. Figure 3 illustrates the experiment matrix
for this particular series; note the extensive experimentation with aerogel of 0.02 g/cm
3
density (in part from
Shipment II), the density of the exposed aerogel collectors on Mir.
3.0
3.5
4.0
4.5
5.0
5.5
6.0
6.5
7.0
01020304050
Velocity (km/s)
Allende (50 µm)
Pampa A (50 µm)
Soda-Lime Glass (50 µm)
Aluminum, Metal (50 µm)
Al O (130 µm)
32
Aerogel Density (mg/cm
)
3
Figure 3. Plot outlining the experiment matrix of the initial series of tests, the primary purpose of which was aimed at an
improved understanding of the effects of aerogel density and impact velocity on absolute penetration-depth and mass of
the recovered residue.
Figure 4a presents typical penetration tracks as a function of aerogel density produced by soda-lime glass
spheres at velocities near 6 km/s. Note the systematic dependence of track length on aerogel density. The
morphologies of such tracks were extensively described by Barrett et al. (1992), Mendez (1994), and
Thomson (1995), and are similar to those observed in porous alumina (Werle et al., 1981) or Styrofoam
21
(Tsou, 1990). There appears to be nothing unusual about tracks produced in very low-density aerogels
compared to published observations, except for the distinctly stepped tracks in the 0.01 g/cm
3
case.
The latter tracks repeatedly change their diameters with depth into the target in rather abrupt fashion, a
seemingly unique and previously unreported phenomenon for which we have no ready explanation. Figures
4b and 4c illustrate typical tracks for impacts by powdered Allende and Pampa-A projectiles, respectively.
Note that Figures 4a-4c are reproduced to approximately the same scale and that the natural impactors gener-
ate somewhat shorter tracks than the glass spheres.
We also observed bifurcation, if not multiple splitting of some tracks as previously reported as starburst
tracks by Thomson (1995); good examples are illustrated in Figure 5, using Pampa-A projectiles. Obviously,
such tracks reflect breakup of projectiles, typically in the early stages of penetration, but occasionally at late
stages as well. Such split tracks are fairly common for the powdered meteorite projectiles, but they are absent
in the case of soda-lime glass, metallic aluminum, and aluminum-oxide spheres. Track splitting seems to
suggest projectile failure along microcracks that were introduced either during projectile manufacture or
launch. Invariably, targets containing split tracks also display a substantial population of relatively short
tracks, indicative of projectile fragmentation during launch.
Some minor tracks are visible even in the case of soda-lime glass spheres (e.g., the 0.02 g/cm
3
case in
Figure 4a). Such small tracks seem unavoidable and result from particulate contaminants such as gun-barrel
abrasion products and particles shed from the bursting of the high-pressure diaphragm or the high-pressure
piston; they could also represent ejecta generated during sabot impact. Such contaminants can occur over a
wide range of sizes. Debris much smaller and, on rare occasions, much larger than the nominal impactors is
readily spotted by absolute track length. Contaminants producing tracks of expected nominal dimensions, not
to be confused with the nominal projectile impact, are invariably associated with a detectable population of
small debris tracks. Nevertheless, optical or chemical characterization and identification of the impactor
residue is necessary to positively differentiate between tracks resulting from nominal projectiles and those
produced by undesirable contaminants.
The optical photographs presented in Figure 4 do not portray the target surfaces in great detail. Nevertheless,
it seems obvious that all tracks in <0.05 g/cm
3
aerogel lack the substantial spall zones that are so prominent in
0.09 g/cm
3
aerogel, as detailed by Thompson (1995) and Burchell and Thomson (1996). Indeed, most shots
into 0.01 and 0.02 g/cm
3
aerogel display a rather crisp and small entrance hole without any spallation fea-
tures. Immediately below the surface, the penetration track rapidly grows into a bulbous shape, from which it
gradually tapers off to form the typical slender, carrot-shaped track. Also note a few modestly curved tracks
and those of apparently oblique orientation relative to the aerogel surface for the powdered meteorites; in
comparison, the tracks formed by regular spheres tend to be straight. We ascribe the curvature and deviation
from normal-incidence angle of the Allende and Pampa-A tracks to irregular projectile shapes that result in
uneven drag forces during deceleration.
22
Figure 4a. Optical photographs of penetration tracks created by 50-µm impactors in aerogel targets of different density.
(a) Spherical soda-lime glass projectiles, (b) powdered Allende projectiles, and (c) powdered Pampa-A projectiles.
23
Figure 4b. Penetration tracks of powdered Allende projectiles.
24
Figure 4c. Penetration tracks of powdered Pampa-A projectiles.
25
Figure 5. Bifurcated tracks caused by a 50-µm Pampa-A projectile impacting at about 6.0 km/s. Note the presence of a
third track, in the top view, that is partly obscured behind the track in the foreground. Dark residue is visible at the end o f
most tracks, which tend to curve gently as the projectile decelerates.
The track lengths at a constant 6 km/snormalized to projectile diameterare plotted as a function of aerogel
density in Figure 6a, while absolute track lengths of the silicate impactors only are detailed in Figure 6b. There
is substantial variation among the five projectile materials employed. Not surprisingly, Al
2
O
3
seems to pen-
etrate the deepest, as it is not only the highest density impactor, but also the one characterized by the highest
melting point and compressive strength. Soda-lime glass penetrates with almost the same ease as Al
2
O
3
,
followed by metallic aluminum, and the powdered meteorites, with Pampa-A penetrating modestly deeper than
Allende. The relatively poor penetration power of metallic aluminum is somewhat surprising, yet it is the most
malleable material employed in this study, as well as that with the lowest melting point, so it most likely suffers
the most deformation and/or ablation during capture.
26
0
100
200
300
400
500
0 1020304050
0
5
10
15
20
25
Absolute Track Length (l; mm)
p
Scaled Track Length (l / D )
Velocity = 6 km/s
Allende
Pampa A
Soda-Lime Glass
Projectile
Diameter (D ) = 50 µm
p
0
100
200
300
400
500
0 1020304050
Aluminum, Metal (50 µm)
Allende (50 µm)
Pampa A (50 µm)
Soda-Lime Glass (50 µm)
Al O (130 µm)
32
Velocity = 6 km / s
Various Projectile Types
p
Scaled Track Length (l / D )
(A)
(B)
Aerogel Density (mg/cm )
3
Figure 6. Track length plotted as a function of aerogel density; projectile velocity remained constant near 6 km/s while
the projectile types were varied. (a) Track length (l) normalized to the nominal projectile radius (D
p
= 50 µm) and
(b) absolute track length for the 50-µm silicate impactors only. Note the apparently good reproducibility of track length
for any given initial condition (see text).
27
Clearly, all materials display the same trends, albeit to various degrees, of having increasingly deeper
penetration with decreasing aerogel density. There is tentative evidence that a particularly significant change
in penetration capabilities occurs between 0.01 and 0.02 g/cm
3
, as detailed via Figure 6a. Based on the
experimental difficulties and uncertainties afflicting these experiments, more rigorous interpretations of
Figure 6 do not appear warranted at this stage.
The velocity dependence of penetration depth in aerogels of variable densities was investigated with 50-µm
soda-lime glass spheres. The results for 5 and 6 km/s (±0.1 km/s) are illustrated in Figure 7. Obviously, the
5 km/s impactors penetrate deeper than those at 6 km/s under all experimental conditions, a result that appears
consistent with the penetration experiments using porous alumina by Werle et al. (1981) and low-density
polymer foams (Tsou, 1990). The latter studies report an optimum penetration depth in these highly porous
targets at relatively modest velocities (<5 km/s).
0
100
200
300
400
500
0 1020304050
0
5
10
15
20
25
Aerogel Density (mg / cm )
3
6 km / s
5 km / s
Soda-Lime Glass Projectile
Diameter (D ) = 50 µm
p
Absolute Track Length (l; mm)
p
Scaled Track Length (l / D )
Figure 7. Track length in aerogels of different density produced by 50-µm soda-lime glass spheres at 5 and 6 km/s; note
that some of the data result from duplicate experiments (see text).
28
A major reason for presenting the above dimensional data is to illuminate the reproducibility of aerogel pen-
etration tracks under presumably controlled laboratory conditions. The reproducibility is poor, at present, for
the various experimental reasons outlined above (see also Thomson, 1995, and Burchell and Thomson, 1996).
Substantial differences in track length seem unavoidable when experimenting with very small projectiles that
are sized by sieving. Referring only to the soda-lime glass shots (e.g., in Figure 6b), the present data tend to
group in much tighter clusters than those of Thomson (1995) and Burchell and Thomson (1996). In particular,
we emphasize that the 0.01 and 0.02 g/cm
3
data each refer to two repeat experiments, attesting to good
reproducibility of track length from experiment to experiment. Good reproducibility of track length is portrayed
in Figure 6b as well, which includes the Allende and Pampa-A projectiles, with the Allende data at 0.02 g/cm
3
being the composite of two experiments. While there are not very many individual experiments that produced
multiple tracks with the Al
2
O
3
and metallic aluminum projectiles, the trends displayed in Figure 6a also seem
sufficiently systematic to suggest much less data scatter than that reported by Thomson (1995) and Burchell
and Thomson (1996).
The apparent tighter clustering of data points in Figure 6 relative to those of Thomson (1995) or Burchell and
Thomson (1996) is real for the unbroken projectiles (glass, Al, Al
2
O
3
), yet it is an artifact for the meteorites.
The meteorite data are the result of subjective classifications of an observable track population into good and
bad tracks, the latter unplotted. Figure 8 (unlike Figures 4b and 4c) illustrates typical target scenes employing
Allende (Figures 8a and 8b) and Pampa-A (Fig. 8c) projectiles. First, note the very large number of tracks
per any individual experiment, clearly the first sign and telltale evidence that projectile fragmentation during
launch may have occurred. It seems self-evident that the large number of small tracks must be assigned to
small-fragment impacts, and only some of the largest tracks may be assigned to nominal impactors near
50 µm in size. Figure 8a contains two tracks that are about 1/3 the length of the longest nominal track, and
we consider them to be the result of projectile fragments. Figure 8b illustrates the common and more ambigu-
ous case, in which a number of tracks are on the order of the longest track, consistent with expected mass
variations of the projectiles. Recovery of projectile residues identified them as meteorite fragments; we
consider them to be nominal and plotted those data in Figure 6. Somewhat shorter tracks are also present in
Figure 8, yet their interpretation was ambiguous, so they are not plotted in Figure 6. Figure 8c illustrates a
single nominal track (towards top) that terminates outside the image and a second, major track (towards
bottom) that terminates close to the images edge. Although of different length, both tracks may represent
nominal impactors, given the variation in mass of the sieved projectile ensemble. Both of these long tracks
are readily contrasted with a population of smaller tracks of variable lengths, all of which were assigned to
fragment origins.
On the basis of these examples, recognition and subsequent elimination of unsuitable tracks can be a difficult
judgment call that varies from target to target. Fortunately, in the majority of cases there are a small number
(2 to 5) of the longest penetrations, all within 10% to 20% in absolute length, that contrast markedly with a
fair number of tracks that are typically less than half the length of the longest tracks. Nevertheless, classifica-
tion as a nominal track and elimination of the remaining population remains subjective in a significant
number of cases. In addition, there is no independent evidence or guarantee that any nominal track is indeed
the product of a nominal meteorite projectile; the latter assignment rests, in large part, on comparative data
with unbroken glass projectiles.
29
Figure 8a-c. Typical populations of penetration tracks in 0.02-g/cm
3
aerogel resulting from the impact of 50-µm
particles of powdered meteorite. Samples shown in photographs (a) and (b) are tracks resulting from Allende
projectiles with velocities of 6.3 and 6.4 km/s, respectively, while (c) shows tracks from Pampa-A particles (6.7 km/s).
While qualitative, these photographs serve to illustrate that projectile breakup is more severe at higher velocities and
accelerations than at lower velocities. Some tracks exhibit slight curvature toward the end, suggesting the presence of
irregular grain shapes.
30
When severe, launch-induced projectile fragmentation occurs, track classification is much more difficult, and
none of the tracks can be considered representative of nominal impactors. Such is the case at high impactor
velocities, which are invariably characterized by a very large population of small tracks. The positive recog-
nition of an unsuccessful experiment on the basis of an anomalous oscilloscope record and an unusually large
population of short tracks justify rejection of such experiments. Therefore, the meteorite data illustrated in
Figures 6 and 7 refer only to the few, longest tracks for any given target that also contained relatively massive
impactor residues.
One way of conveying these experimental and interpretational difficulties is illustrated in Figure 9. Figure 9a
plots track length against impact velocity for seven experiments using Allende projectiles and aerogel of
constant density (0.02 g/cm
3
). The observed population of tracks was classified into (a) tracks >5 mm in
length, (b) tracks between 3 and 5 mm, and (c) tracks <3 mm long. Figure 9a portrays all tracks >5mm and a
select set of tracks between 3-5 mm long; tracks <3 mm were not measured. The obvious point of this plot
relates to the enormous scatter of data points if all tracks were plotted. Obviously, any reported data scatter
(e.g., Burchell and Thomson, 1996) strongly depends on the somewhat arbitrary and highly interpretative
selection of a cut-off value for short tracks. Also note that the longest tracks in the 3.4 and 5.3 km/s experi-
ments seem to be short relative to the general trends in the data, especially those at neighboring velocities.
It is such experiments that we would reject in their entirety as unsuitable due to possible projectile frag-
mentation. Also note that the longest tracks in Figure 9a define some general, qualitative trend of increasing
track length with decreasing velocity. However, we do not have enough confidence in the data to justify
further quantification. We are simply not certain about the actual impactor mass responsible for these track
populations.
The latter point is further emphasized in Figure 9b, which plots track length against the diameter of the
projectile residue (D
r
). The latter was measured in situ, via a binocular microscope, and is normalized to the
nominal 50-µm-diameter projectile (D
p
); the precision of the D
r
measurements is low, about ±20% for large
(>30 µm) residues and ±50% for smaller fragments. While this plot shows that there is some gross correla-
tion of track length and relative residue size, it also shows that individual tracks cannot be interpreted with
confidence in an internally consistent manner. Grossly similar residues (e.g., at D
r
/D
p
= 0.5) may be found at
the termini of tracks that differ vastly in length. This poor correlation may reflect, in part, a significant
spectrum of impact velocities among severely fragmented projectiles, and/or may result from additional
projectile fragmentation and track splitting during penetration. On average, the high-velocity experiments
tend to produce the smaller residues, which, in turn, represents further evidence for progressively severe
projectile breakup as impact velocity and associated accelerations increase.
For the time being we are forced to select, in some defensible and prudent fashion, a population of tracks
(Figures 6 and 7) or residue sizes (Figure 10) that best represent specific initial impact conditions. Figure 10
summarizes the relative residue size as a function of aerogel density for all 50-µm projectiles at a constant
impact velocity of 6 km/s. As was the case in Figures 6 and 7, we plotted all data points for the Al and glass
shots, but only select data for the Allende and Pampa-A powders reflecting only the largest residues. Clearly,
all projectile materials have undergone some mass loss, yet one may recover substantial residues (>50% of
the projectiles initial size) at 6 km/s using 0.02 g/cm
3
aerogel collectors. Under all conditions simulated,
aluminum suffers the least mass loss, and soda-lime glass is superior to the powdered meteorites. At face
31
0
2
4
6
8
10
12
14
2.5 3.5 4.5 5.5 6.5 7.5
0.0
0.2
0.4
0.6
0.8
02468
10 12 14
Projectile Velocity (km/s)
Track Length (l; mm)
Track Length (l; mm)
Recovered / Original
Projectile Diameter (D
r
/ D
p
)
3.02
3.40
4.10
5.30
6.50
6.80
6.88
Projectile Velocity (km/s)
(A)
(B)
Figure 9a-b. Evidence of projectile fragmentation during launch in the light-gas gun. (a) Variation of track length in
individual 0.02-g/cm
3
aerogel experiments by 50-µm Allende projectiles. The apparent cut-off of tracks under 3 mm in
length is arbitrary; the plot does not portray the correct frequency distribution of tracks 3-5 mm in length. The longest
tracks in each experiment were measured and suggest some inverse dependence of impact velocity and track length. The
3.4-km/s and 5.3-km/s experiments are believed to illustrate severe projectile breakup, in that no large projectile reached
the target. (b) Absolute track length plotted as a function of the diameter of the recovered residues (D
r
) normalized to
projectile size (D
p
, 50 µm) for the same experiments depicted in (a). There is a general positive correlation between D
r
and l, yet the data scatter is so substantial that quantitative information cannot be extracted with confidence.
32
0.0
0.2
0.4
0.6
0.8
1.0
02040
50
Recovered / Original
Projectile Diameter (D / D )
3010
Nominal Velocity = 6 km/s
Projectile Diameter = 50 µm
Soda-Lime Glass Pampa AllendeAluminum
Aerogel Density (mg/cm )
3
r
p
Figure 10. Summary of recovered residues sizes for good (see text) tracks associated with soda-lime glass, aluminum
spheres, Allende, and Pampa-A projectiles, all of which were 50 µm in diameter and impacted at a velocity of 6 km/s. All
projectile types experienced some mass loss during aerogel capture, with aluminum being preserved the best, followed by
glass. Note that the Allende and Pampa-A data are afflicted with substantial uncertainties regarding the exact pre-impact
size of the projectile.
value, recovery of all silicates seems to be best at aerogel densities between 0.02 and 0.04g/cm
3
; aerogels
of higher and lower density seem to exhibit inferior recovery properties, but this suggestion must be verified
by additional experiments. In general, the recovery efficiencies illustrated in Figure 10 are approximately a
factor of two higher than those found for 80-µm glass spheres encountering higher density aerogel
(0.09 g/cm
3
, Thomson, 1995; Burchell and Thomson, 1996).
Figure 11 shows typical large residues recovered from the present experiments. Detailed mineralogical/
geochemical investigations are presently being conducted to determine potential modifications of the par-
ticles phases and textures during the capture process. It is clear that substantial mass loss occurs and that
irregular particles will be substantially rounded during aerogel penetration. Note also that molten aerogel
adheres to the surfaces of the residues and may even invade the latter along microcracks as detailed by Barrett
et al. (1992) or Mendez (1995). Typically, the front faces of the recovered projectiles are associated with
massive plugs of compressed and molten aerogel. Such plugs may play a role in the deceleration process by
providing some cushioning effect for the projectile.
33
Figure 11a. SEM photographs of typical, largest residues collected at 6 km/s using 0.02-g/cm
3
aerogel. These particles
have not been otherwise characterized for possible modifications of texture, phase, or bulk composition. (a) Soda-lime
glass sphere, (b) Allende particle, and (c) Pampa-A particle. Note that the meteorite particles compared to those shown in
Figure 2 are substantially rounded, and that all residues are partly draped by molten aerogel.
34
Figure 11b. SEM photographs of Allende particle residues collected at 6 km/s.
35
Figure 11c. SEM photographs of Pampa-A particle residues collected at 6 km/s.
36
Summarizing this series of experiments that mainly addressed the effects of aerogel density, impact velocity,
and projectile physical properties, we found that the length of penetration tracks strongly depends on aerogel
density and, in complex ways, on impact velocity and projectile properties. The data currently in hand do not
allow much more detailed conclusions, as they are afflicted with substantial experimental uncertainties
regarding exact initial impact parameters, most notably projectile size. Calibration experiments with very
small impactors turned out to be unexpectedly challenging when employing light-gas guns and associated
shot-gunning methods to accelerate an aggregate of multiple projectiles. Clearly, the subjective differentia-
tion into good or bad penetration tracks, as practiced in this work, is an unsatisfactory procedure. The
substantial scatter of results reported by Thomson (1995) and Burchell and Thomson (1996) are most likely
due to similar experimental uncertainties, as was the case in our previous experiments as reported by Barrett
et al. (1992). On the other hand, the present experiments employing unbroken, spherical projectiles (e.g.,
Figure 6 or 7) demonstrate that track length and other experimental effects may be reproduced with good
precision. Clearly, initial projectile mass and shape must be better controlled and reproduced than we were
able to accomplish in this study. This is a substantial challenge in itself considering the small size (<100 µm)
and the large numbers (50 to 100 per experiment) of projectiles needed for systematic studies, that must also
include a variety of projectile materials.
5. Experiments at Oblique Impact Angles
The aerogel collectors exposed on the Mir station intercepted natural and man-made particles at essentially
unconstrained impact angles of 0° to 90° from the target surface at encounter velocities of ~10 km/s for man-
made debris and approaching 20 km/s for natural dust particles (e.g., Zook, 1993). In contrast, the Stardust
collectors will be pointed normal relative to freshly released particles from Comet Wild 2, while the encoun-
ter velocity is expected to be a modest 6.2 km/s based on current mission profiles. Interpretation of the Mir
collectors demands that the relationships of penetration track length, impact velocity, and impact angle be
known. In addition, Tsou (1990) suggested that inclination angle and absolute track length could be used to
reconstruct the trajectory of individual particles.
Based on these considerations, the objective of our second test series was to evaluate track length as a func-
tion of impact angle and velocity. A single projectile type (soda-lime glass) was used, as well as a constant
aerogel density (0.02 g/cm
3
). These tests employed the so-called slabbed target configuration (Fig. 1c),
since the initial tiles (Shipment II) were too thin (10 to 12 mm) to accommodate the expected track lengths.
Wedges of 60°, 45°, and 30° were attached to the target-support plate, onto which the entire target assembly is
mounted. In addition, we began to suspect that aerogel density may vary among Shipment II samples. There-
fore, we exposed the same aerogel specimen to multiple gun-firings (
<4 times), with each exposure typically
occurring at a different impact angle. The inclination of individual tracks relative to the target surface is
sufficiently distinct (Fig. 12) to permit assignment of individual tracks to any specific experiment at some
given impact angle. We also repeatedly exposed individual targets at constant impact angle with variable
impact velocities; in this case, all tracks (typically fewer than five) were measured for length and their X-Y
positions were marked on the cardboard hold-down frame before the next experiment, each experiment using
different-colored pens.
37
Figure 12a-b. Examples of aerogel penetration tracks produced at variable impact angles. Target (a) was exposed
twice, while target (b) was used in three experiments, as can be seen from the variable inclinations of the penetration
tracks. Note the overall reproducibility of track length among individual and different experiments in any given aerogel
specimen.
38
The need for multiple exposure of the same aerogel specimen to ensure constant target density was deter-
mined during some normal incidence experiments using Shipment II targets that did not reproduce the (glass)
data illustrated in Figures 8 and 9. Some tracks were shorter than those produced earlier, and we suspected
variable target density to be the cause. Examples of such track variations from tile-to-tile are illustrated in
Figure 13, which shows track lengths in a single (Shipment II) target at 90° and 45°, and a repeat test at 45°
into a second (Shipment II) aerogel sample; both targets possessed a nominal density of 0.02 g/cm
3
. Note the
substantial overlap of experiments 1864 and 1865, and the difference with duplicate experiment 1866. Also
note the good agreement of shots 1864 and 1865 with equivalent experiments in Shipment I tiles (Figure 7),
all yielding tracks ~12-14 mm long. Clearly, experiment 1866 systematically yielded shorter tracks and we
suspect that target density was higher than the nominal 0.02 g/cm
3
. In detail, tracks within a given aerogel
tile that are either on the short or long side are often juxtaposed, possibly indicating density variations
within a specific specimen, but the present data are insufficient to conclusively demonstrate such intra-tile
variations.
Figure 12 illustrates representative results of the oblique-impact tests . Note the paucity of short tracks
considering that the targets were exposed to multiple gun-firings (i.e., twice [Fig. 12a at 90° and 45°] and
three times [Fig. 12b at 90°, 45°, and 30°], respectively). Figure 12a is somewhat unusual and illustrates
three vertical tracks of identical depth and two inclined tracks, also of essentially identical length; the
projectile mass/energy and target density seem exceptionally constant, yielding such reproducible results
(Figure 12a). Figure 12b is more typical of the present aerogel experiments and does exhibit noticeable
scatter in track length for any specific angle of incidence, presumably associated with impactors of variable
mass. In most cases, the reproducibility of track length in similar experiments is not as good for the Shipment
II tiles as it was for the experiments that employed Shipment I aerogel tiles (Figure 6 and 7).
Returning to the major objective of this second series of experiments, Figure 14 summarizes all observations
of track length as a function of impact angle and velocity for 0.02 g/cm
3
aerogel and 50-µm spherical soda-
lime glass projectiles. Figures 14a-14d illustrate track lengths at nominal velocities of 3, 4, 5, and 6 km/s,
respectively, while Figure 15 shows averaged data for easy comparison. In general, the data are disappoint-
ingly and surprisingly uniform, as there is not a clear relationship of track length with impact angle or veloc-
ity. The 6.2-km/s, 60° impact-angle experiment appears to be a particularly uncooperative dataset that we are
at a loss to explain.
Nevertheless, what is apparent from Figures 14 and 15 is that all tracks have similar lengths and other mor-
phologic attributes regardless of impact angle. The total mass displaced and damaged by oblique impacts
appears to be insensitive to impact angle, unlike cratering events in dense media that predominantly reflect
only the vertical velocity component (e.g., Christiansen et al., 1993). We interpret the data in Figures 14 and
15 to imply that processes other than pure shock phenomena play the dominant role in the penetration of low-
density, porous media, an observation in support of the continuum mechanics model of Anderson and Ahrens
(1994). This model largely relies on a bulk-viscosity term, drag forces, and ablative processes for the decel-
eration of hypervelocity impactors in porous media of exceptionally low bulk-density.
39
7 8 9 10 11 12 13 14
Track Length (l; mm)
90
O
45
O
45
O
Impact Angle
1866
1864
1865
Shots 1864 & 1865 utilized
the same aerogel tile.
Figure 13. Plot of track length versus impact angle showing the reproducibility of track lengths per experiment using
50-µm aluminum spheres and 0.02-g/cm
3
aerogel at impact angles of 90° and 45°. Individual aerogel tiles were utilized
in multiple experiments; shots 1864 (V = 5.03 km/s) and 1865 (V = 5.13 km/s) used the same tile, while a second slab was
used for experiment 1866 at V = 4.98 km/s. Note that the tracks in 1866 are systematically shallower than those of 1865,
which is essentially a duplicate experiment (see text).
The experiments at inclined impact angles corroborate the findings of Tsou (1990) and Thomson (1995) who
found that the track orientation faithfully records an initial impact angle. Our data, however, do not support
the suggestion by Tsou (1990) that complete trajectory data (i.e., direction and velocity) may be extracted
from the inclination and length of penetration tracks in low-density foams. In an earlier work (Hörz et al.,
1992), we disagreed with this view and continue to maintain that the encounter velocity is not readily ob-
tained from absolute track length, a conclusion also reached by Thomson (1995), and Burchell and Thomson
(1996).
40
4
6
8
10
12
14
20 30 40 50 60 70 80 90 100
10
12
14
16
18
20
10
12
14
16
18
20
10
12
14
16
18
20
10
12
14
16
18
20
3.0 km/s
4.0 km/s
5.0 km/s
6.0 km/s
6.2 km/s
Impact Angle
Track Length (l; mm)
(A)
(B)
(C)
(D)
(E)
Target a
Target b
Average
Target a
Target b
Average
Target a
Target b
Average
Target a
Target b
Average
Target a
Target b
Average
Figure 14a-e. Track length as a function of impact velocity and impact angle for 50-mm soda-lime glass spheres and 0.02-
g/cm
3
aerogel. The solid lines connect the shortest and longest track per experiment, while the dashed line represents the
average of all tracks for a given set of impact conditions. The 6.2-km/s experiments appear to be anomalous; all tracks at
90° occurred in one aerogel tile, while those at 60° occurred in another; the latter tile was reimpacted at 60° and the short
tracks were confirmed; we suspect that this specific tile possessed a higher density than the nominal value of 0.02 g/cm
3
.
41
4
6
8
10
12
20 30 40 50 60 70 80 90 100
14
16
18
20
Impact Angle
Track Length (l; mm)
3.0 km/s
4.0 km/s
5.0 km/s
6.0 km/s
6.2 km/s
Figure 15. Summary plot of the average track lengths, depicted in Figure 14, as a function of impact angle and velocity.
Note the fairly constant track length at most impact angles, as well as at any given velocity. These represent the best
calibration data we currently have, as they were generated with unbroken soda-lime glass spheres (see text).
6. Experiments With Particles of Low Cohesion
Particles of low-cohesive strength, typically with low density and some porosity, constitute an important
subclass of natural cosmic-dust particles (e.g., Love et al., 1993). Because their physical properties substan-
tially differ from those of dense silicates and rock, their cratering and penetration behavior may not readily
be extrapolated from impact experiments using dense, non-porous projectiles. As a consequence, dedicated
impact simulations are required to delineate the penetration behavior and capture efficiency of porous materi-
als of low-cohesive strength. However, hypervelocity accelerators, including light-gas guns, cannot readily
launch materials of low-cohesive strength. The following section describes two new approaches that pro-
duced impactors of low-cohesive strength traveling at 6 km/s; although such impactors may not be well
characterized, they will be helpful in answering some first-order questions about penetration and retention of
fluffy cosmic-dust particles in aerogel.
6.1 Cocoa Powder
A number of fine-grained (<1 µm), flour-like materials were mixed with nominal projectiles in an attempt to
eliminate projectile fragmentation during light-gas gun acceleration. The idea was to provide some cushion-
ing between neighboring projectiles inside the sabot cavity. None of these tests resulted in a noticeable
improvement of unbroken, nominal projectiles. Somewhat fortuitously, however, our tests with commercial
cocoa powder resulted in various clumps of the powder impinging upon the aerogel target. Such clumps
made crater-like tracks on dimensional scales of a few millimeters. Unfortunately, we had no control on the
42
clump size, so variable track sizes resulted. The velocity detectors at the target site (t
1-n
) measured arrival
times of the early clumps at a velocity identical to that of the sabot impact(s) and consistent with expected
projectile velocity, based on empirical gun parameters. However, there was an extra long tail to the arrival
times of the impactors, with some events occurring as late as 300 µs after the first signals, corresponding to a
decrease in impact velocity of around 20%. It follows that projectile velocity is essentially uncontrolled for
these tests; the values given are those for the first arrivals.
Figure 16 illustrates representative examples from experiments that employed a 1:1 (weight) mixture of cocoa
and 50-µm aluminum projectiles. The presence of cocoa clumps is readily apparent from the characteristi-
cally stubby nature of some tracks that are distinctly light brown in color. Figure 16a illustrates two experi-
ments, one at 90°, the other at 45°, while Figure 16b shows only one experiment at 90°. Note the stubby
nature of the tracks, none resembling their slender counterparts seen in Figure 4. Frequently, a relatively
slender, long track emerges from the stubby and bulbous cavities as detailed in Figures 16c and 16d, contain-
ing nominal aluminum-projectile residue at their termini. Such tracks were absent in a control shot that
launched only cocoa powder. This suggests that the compound tracks visible in Figure 16 are the result of
nominal impactors that had variable amounts of cocoa powder clinging to the projectile. While long com-
pared to the pure cocoa-powder control shot, such composite tracks are substantially shorter than those
associated with pristine aluminum spheres (e.g., Figure 6a). Unfortunately, tracks without evidence of some
cocoa powder adhering to the projectile were absent. Therefore, the idea of cushioning the nominal projec-
tiles by some fine-grained powder was abandoned as it did not yield impactors of well-controlled mass.
Nevertheless, the cocoa experiments relate to the penetration of aerogel by low-density, porous, weakly
cohesive particles.
Figure 16e shows the interior of a track produced by impact of one of these cocoa-powder clumps; the aerogel
target was sectioned via a razor blade. Much of the dark material adhering to the walls is brown, remnant
cocoa powder. Figure 17 shows detailed SEM studies of a single cocoa track produced at 6 km/s. Figure 17a
reflects standard optical photography of the unprocessed, pristine track at some modest depth inside the
(1-cm-thick) aerogel slab. In contrast, the SEM image of Figure 17b was taken after careful splitting of the
slab and presents a detailed view of the track interior. Comparison of Figure 17a and 17b indicates that
additional breakage and damage in the vicinity of the track occurred during splitting of the aerogel via the
razor blade method. The track interior is substantially blocky and, surprisingly, there is no evidence of
abrasion or fine-scale erosion. Figure 17c shows the entrance hole for the track, taken prior to splitting, and
its irregular outline suggests an irregular cross-section of the impactor. Some radial cracks exist, as well as
poorly developed concentric fractures, none of them sufficient to initiate spallation of the aerogel surface.
Figures 17d and 17e illustrate progressively higher magnifications of the pristine track interior, viewed
through the entrance hole. Again, the largely blocky nature of the track walls is evident, yet some areas seem
to be crushed more intensely than others (e.g., lower right-hand corner in Figure 17e). As before, there is no
evidence of striation or of melting phenomena. Abundant, fine-grained debris populates the freshly created,
brown surfaces. Some of this fine debris, and an occasionally larger particle, such as the pear-shaped object
in the lower left-hand corner of Figure 17e, display prominent charging effects and are interpreted as rem-
nants of cocoa powder. These cocoa residues are the objects of ongoing chemical analyses.
43
Figure 16a-e. Optical photographs of penetration features in 0.02-g/cm
3
aerogel produced by a 1:1 (weight) mixture of
cocoa powder and 50-µm aluminum spheres. (a) Typical tracks resulting at 5.1 km/s and impact angles of 90° and 45°,
respectively. (b) Additional tracks from the 90° at 5.1 km/s experiment of (a). Note the bulbous, stubby nature of the
short tracks by pure-cocoa clumps, as opposed to the longer, composite tracks resulting from nominal aluminum
projectiles embedded into a cocoa matrix or that had substantial clumps of cocoa adhereing to their surfaces. Figures (c)
and (d) show typical 45° impacts at 6.0 and 4.7 km/s, respectively, suggesting different amounts of cocoa associated with
the aluminum spheres. (e) Enlarged view of the track in (c).
44
Figure 17a-e. Detailed views of a cocoa track produced at 6 km/s and 90°. (a) Optical image of entire track, (b) detailed
track interior, (c, d, e) various enlargements of the track entrance hole and interior. Most of the fine particles are cocoa
powder (see text).
45
These experiments with cocoa clods are interesting from a phenomenology point of view, because individual
clods may be viewed as analogs of the physical properties of fairly friable, fluffy, and porous natural particles.
These impact tests are significant as they demonstrate, for the first time, that residues of such poorly cohe-
sive, and generally low-density aggregates may be successfully trapped in aerogel collectors at velocities as
high as 6 km/s. The composite, long tracks produced by cocoa-coated, nominal impactors suggest that
relatively large and dense components of natural aggregate particles may also penetrate in differential fashion
relative to the fluffy matrix, making deep, subsidiary tracks that emanate from otherwise stubby and bulbous
aerogel cavities. In addition, there is considerable interest in the exobiology community regarding the ability
of organic compounds to survive impact. If warranted, additional and more detailed experiments could be
conducted with fine-grained organic matrices other than cocoa. It seems possible to mix silicates of all sizes
(and most likely compositions/structures) with cocoa or other organic materials of more direct interest to
exobiology and to evaluate their survivability and/or modification(s) during hypervelocity aerogel capture.
6.2 Clusters of Collisionally Produced Fragments
Our second approach in producing particles of low-cohesive strength is a spin-off from penetration experi-
ments in thin targets, commonly conducted in the development of collisional shields (e.g., Anderson, 1993,
1995). Such experiments typically employ massive witness plates behind a penetrated bumper-foil (e.g., Hörz
et al., 1994), or high-speed photography or X-ray shadowgraphs (e.g., Piekutowski, 1993) to monitor the
nature and evolution of the resulting debris clouds. Accordingly, fragment-size distribution and radial disper-
sion angle of the resulting debris plumes depend on bumper thickness and impact velocity for any given
projectile. Therefore, it is possible to generate a desired population of fragments from almost any projectile
by judicious selection of bumper thickness and impact velocity. In addition, by varying the standoff distance
of the witness plate from the penetrated bumper, one may intercept the debris cloud at various stages of radial
dispersion; this permits some manipulation of the spatial density of fragments at the location of the witness
plate. These considerations were used to expose aerogel, in lieu of witness plates, to clouds of projectile
fragments that ranged from barely fractured projectiles to widely dispersed plumes. Obviously, such debris
plumes possess no cohesive strength, and they represent conservative end-member conditions in evaluating
the capture of fluffy, highly friable, natural particles.
For these experiments, we selected 4-µm-thick aluminum (1100 series) as a bumper foil; this thickness readily
disrupts 50-µm soda-lime glass projectiles at velocities above 5.5 km/s and is sufficiently thin to ensure that
the debris plumes are totally dominated by projectile debris (Hörz et al., 1994). All experiments in this test
series employed aerogel with a nominal density of 0.02g/cm
3
from Shipments II and III. Standoff distance
was varied between 2 and 15 mm from the bumper foil.
We conducted a series of precursor experiments, exposing polished-copper witness-plates, to pin down
specific conditions for the aerogel shots. Such solid-Cu plates are not only less costly than aerogel, but they
also reflect the size distribution of projectile fragments much better than do the brittle silicate aerogels. This
is especially true for tightly clustered fragment clouds that generate many overlapping, yet clearly resolvable
craters in copper, but which typically produce only a single, very large entrance hole in the aerogel. Some of
46
these precursor experiments are shown in Figure 18. Figure 18a illustrates two barely fragmented projectiles
at 6.0 km/s and a standoff distance (L) of 4 mm. Note the presence of few very small fragments and the
survival of a massive, central core-fragment. Detailed inspection of the interior of the major crater reveals the
presence of subsidiary depressions that are separated by ridges and delicate septa, suggesting that this core
was an internally fractured object of distinctly heterogeneous mass distribution. Progressive projectile
disruption and dispersion as a function of impact velocity is illustrated in Figures 18b and 18c at 6.2 and 6.4
km/s, respectively; L was 2 mm for both shots. Also note the presence of projectile melt, which is manifested
by radial (Figure 18a) and increasingly concentric (Figure 18b) stringers and filaments, many of a distinctly
beaded nature. As detailed by Gwynn et al. (1996), such features are erosive gouges and depressions of
negative relief, rather than depositional features. Figures 18d-18f illustrate the effects of increased standoff
distance; note the more widely dispersed fragment impacts. Cumulatively, Figure 18 demonstrates that it is
possible to design fragment plumes that range from very tight clusters to widely dispersed clouds by judicious
selection of standoff distance and impact velocity. Also note that some debris plumes include a sizable
fraction of melt, as evidenced by the highly two-dimensional stringers and beaded filaments that are identical
to those from Cu witness-plates as described by Gwynn et al. (1996).
Select experiments that employ aerogel witness-plates are illustrated in Figure 19. Figure 19a and 19b shows
features that are essentially equivalents to those in Figure 18a (at 6.03 km/s). All three impact events shown
in Figure 19a and 19b are from the same experiment (V = 5.92 km/s and L = 4 mm), and demonstrate the
reproducibility of tightly clustered objects using the collisional fragmentation method. Modest dispersal of a
dominant, central mass is indicated by the relatively large entrance hole. Figure 19c illustrates increased
dispersion of such a mass at L = 15 mm, yet otherwise nearly identical impact conditions (V = 5.94 km/s).
Note that some of the large holes visible in Figure 19c (and Figures 19a and b) are complex and compound,
suggesting that some of the more massive fragments were themselves internally fragmented, yet not fully
disaggregated and dispersed. Figures 20a and 20b present cross-sections of the events illustrate in Figure 19a
and 19c, respectively. These cross-sections corroborate the above interpretations of internally fragmented
aggregate impactors, as numerous small penetration tracks emanate from what may be perceived as a
single, major penetration hole (and impacting mass?). These cross-sections enhance the sense of compound
impactors, with individual mass elements penetrating differentially. Clearly, variable track lengths mandate a
somewhat heterogeneous size distribution among the larger fragments. Note the initially, fairly bulbous
penetrations in Figure 20 that are somewhat reminiscent of the previously described cocoa tracks, possibly
reflecting a very low bulk density of the impacting mass. Binocular microscope inspection reveals that most
of the large tracks exhibit projectile residues at their termini.
The erosive nature of impact by molten projectile materials is detailed in Figure 21, which represents views
of one experiment (V = 6.6 km/s, L = 15 mm) at different magnifications. Note the distribution of fine,
beaded melt stringers. Most of the melt, however, appears to be concentrated in a horseshoe-shaped loop,
most likely the incomplete or distorted equivalent of the ring-shaped melt deposits described by Gwynn et al.
(1996). Undoubtedly, the erosive nature of small and large melt particles, including delicate filaments of
melt, is obvious in these images. The melts themselves represent low-strength impactors also capable of
generating substantial damage and penetration of aerogel targets.
47
Figure 18a-c. Typical damage patterns on polished copper witness-plates made by a 50-µm soda-lime glass sphere after
it was fragmented by passing through a 4-µm-thick Al
1100
foil. The distance between the foil and witness plate (the
standoff distance, (L) is given for each experiment. The images are arranged with increasing velocity. Note that the size
distribution of the fragments depends on the impact velocity and that the degree of fragment dispersion can be controlled
by varying L. Many of the radial and circular lineaments represent stringers of projectile melt; see the detailed study at
larger experimental scales by Gwynn et al. (1996).
48
Figure 18d-f. More widely dispersed fragment impacts due to increased standoff distance.
49
Figure 19a-c. Plan view of impacts into aerogel made by clustered impactors similar to those illustrated in Figure 18, all
at about 5.9 km/s. Impacts in (a) and (b) are from the same experiment, illustrating good reproducibility of the fragment
clusters; (c) is a duplicate experiment at L = 15 mm, again showing increased geometric dispersion of individual
fragments. Note that the tracks are compound features, suggesting the presence of internally disrupted fragment masses
that disperse modestly as standoff distance increases (see text).
50
Figure 20a-b. (a) Side views of two compound penetration tracks at 5.9 km/s with a standoff distance of 4 mm,
corresponding to Figure 18a; (b) impact velocity of 5.9 km/s and standoff of 15 mm, corresponding to Figure 19c. Note
the presence of numerous subsidiary tracks caused by apparently coherent fragments that penetrate deeper (see text).
51
Figure 21a-c. (a) Optical and (b, c) SEM images of damage in 0.02-g/cm
3
aerogel by a severely fragmented and
substantially molten projectile (see text).
52
In summarizing this section on aggregate particles of low-cohesive strength, we suggest that Figures 16-21
demonstrate successful production and hypervelocity impact of such particles. However, neither the mass nor
actual encounter velocity can be documented with precision. As a result, these experiments remain highly
qualitative. Nevertheless, they seem to demonstrate that such particle impacts are recognizable in aerogel
targets. They produce compound penetration features characterized by stubby, bulbous penetration cavities
from which any number of subsidiary, slender tracks of variable depth may emanate, the latter reflecting
discrete, individual fragments. The presence of cocoa powder in some tracks and impactor residue at the
termini suggest that natural aggregate particles of low bulk density and low-cohesive strength can be collected
successfully with aerogel, at least at encounter velocities of up to 6 km/s.
7. Discussion
We have conducted a variety of experimental impacts into aerogels of variable density (0.01to 0.05 g/cm
3
),
all of which were somewhat lower than the majority of previous studies (e.g., Barrett et al. [1992], Mendez
[1995], Thomson [1995], and Burchell and Thomson [1996]). In addition, we conducted most of our experi-
ments with projectiles (50 mm) that are smaller than those employed in past efforts (>100 mm, commonly
>500 mm).
One of our major findings is an increased appreciation of the experimental difficulties that are involved when
attempting to launch ensembles of small projectiles using a shotgun method. Most natural materials, which
must be crushed for proper sizing, simply fail during the acceleration to typical light-gas gun velocities. It is
not possible to grind coarse, raw stock without generating microcracks and other defects along which indi-
vidual particles fail during launch. Nevertheless, a few intact particles might make it to the target in some
cases. Unfortunately, the size, mass, and shape of any given particle will not be well-defined, because sieving
operations are poorly suited to ensuring good reproducibility of these critical parameters, especially for
particles <100 µm in diameter. These uncertainties in projectile properties necessitate considerable interpre-
tative judgment on a track-by-track basis to differentiate between nominal impact events that are consistent
with initial impact conditions and those that are not. Dedicated efforts are needed to reduce, and hopefully
eliminate, the current uncertainties in projectile mass and shape for shot-gunned impactors of a few tens of
microns in diameter.
With these reservations in mind, the calibration work presently available for aerogel targets is summarized in
Figure 22, albeit on the basis of select data. Figure 22a includes only the deepest penetration tracks, either
from the present experiments or as reported in the literature for specific impact conditions. The general
agreement among a wide diversity of investigations, employing a wide variety of projectile materials, is
noteworthy. There is a general trend of increasing track length with decreasing aerogel density, yet this trend
is largely controlled by the present study. Note that the data illustrated in Figure 22a reflects a variety of
particle velocities ranging from 5 to 7 km/s, and that there is no obvious trend regarding velocity dependence.
Furthermore, all non-spherical materials that were crushed in the laboratory generated shorter penetration
tracks than did the spherical impactors. This dichotomy possibly relates to modest yet efficient mass loss by
increased chipping and rounding of irregular corners during light-gas gun launch and especially during early
penetration of the aerogel targets.
53
Figure 22b presents a summary of existing work related to the maximum fraction of the initial projectile
diameter preserved, either physically retrieved or measured in situ at the terminus of individual tracks. Again,
these are high-graded data, depicting the maximum residue diameter (D
r
), normalized to the nominal projec-
tile diameter (D
p
). As mentioned above, the spherical projectiles systematically yielded more residues than
the crushed powders, with the latter barely amounting to 10% of the original mass in most cases. Once again,
these data are difficult to interpret in an internally self-consistent manner.
On the basis of the preceding summary and in agreement with Thomson (1995) and Burchell and Thomson
(1995), we conclude that the detailed penetration behavior of aerogel is presently poorly understood, as is the
actual efficiency of particle capture. Current uncertainties result from poorly constrained experiments and
will not improve until initial impact conditions are better defined. This is most readily accomplished by
employing larger projectiles of more narrowly constrained mass and shape. This requires correspondingly
larger and costlier aerogel targets and it has the additional disadvantage of moving the absolute scale of the
experiments into a realm where its applicability to the capture of very small cosmic-dust particles, typically a
few tens of microns in size, becomes questionable. Ideally, simulations with impactors much larger than
those expected in space demands the use of aerogels of correspondingly dimensioned internal structures (e.g.,
Schmidt et al., 1994). None of these developments seems easy, and all would require dedicated efforts.
The experiments employing oblique impact angles confirm the results of Tsou (1990) and Thomson (1995)
that the orientation of individual tracks faithfully records the angle of incidence and direction of projectile. In
addition, the absolute track lengths observed in these experiments show that track length is not dependent on
impact angle. This result alone suggests that high-velocity penetration of aerogel is unlike cratering in dense
targets (e.g., Gault, 1973; Christiansen et al., 1993), but that it is strongly controlled by viscous drag and
associated ablation as postulated by Anderson and Ahrens (1994). We do not understand the apparent weak
dependence (if any) of track length on absolute impact velocity as suggested by the present data.
High-velocity impact of friable, low-density, porous particles into aerogel targets has been simulated at
5 to 7 km/s for the first time. One approach employed clods of fine-grained powder (cocoa), the other used
collisionally generated clouds of glass fragments and melts. These experiments seem significant since a
substantial fraction of cometary particles and interplanetary dust may be of low density and/or modest cohe-
sion. The present experiments demonstrate that low-density, highly porous and fluffy aggregate particles
form readily recognizable penetration tracks, and that residues do reside within such impact features, at least
at experimental velocities around 6.5 km/s.
54
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0 20 40 60 80 100
Recovered / Original
Projectile Diameter (D
r
/ D
p
)
0
100
200
300
400
500
0 20 40 60 80 100 120
Scaled Track Length (l / D
p
)
Aerogel Density (mg/cm
)
3
Soda-Lime Glass; 6 km/s
Allende; 6 km/s
Pampa; 6 km/s
Soda-Lime Glass; 5 km/s
Olivine; 5 km/s
(Barrett et al., 1992)
Olivine; 5.3 km/s (Burchell & Thomson, 1996)
Glass; 5.3 km/s (Burchell & Thomson; 1996)
Olivine; 6 km/s(Barrett et al., 1992)
Aluminum
Al O
23
(a)
(b)
Figure 22. Summary of past and present experiments that employed projectiles smaller than 200µm and aerogels of
various densities. (a) Track length normalized to projectile diameter (l/D
p
) and (b) relative size of recovered residue
(D
r
/D
p
). Note that all data are high-graded and refer to either the longest tracks or largest residues for a given set of
initial impact conditions. We ascribe most of the observed scatter in these figures to experimental difficulties in the form
of either poorly characterized initial impactor size or poorly known aerogel density. Additional experiments are needed
to try to resolve some of these ambiguities.
55
8. Conclusions
In support of ongoing flight instruments on board the Mir complex and of the Discovery Class Stardust
sample return mission to Comet Wild 2, we have conducted a wide variety of impact experiments into aerogel
targets. Most of these must be viewed as pilot-type tests to establish an improved understanding of the
performance of low-density SiO
2
-aerogel as a medium with which to trap hypervelocity particles. Aimed at
providing high-fidelity tests of potential flight collectors, we experimented with unusually small projectiles
(50 µm in diameter), and with aerogels of very low density (<0.05 g/cm
3
), thereby complementing previous
studies at typically larger projectile scales and higher aerogel densities. While we set out to relate initial
impact conditions to the morphology of the observed penetration tracks and the mass of the recovered projec-
tile, we encountered a number of experimental problems that prevented us from establishing the desired
quantitative relationships.
Nevertheless, we fully concur with previous studies that aerogel is an excellent capture medium that will
permit the collection of unmelted silicates at encounter velocities of 6 to 7 km/s. We were able to recover
analyzable projectile residue from experiments with encounter velocities as high as 7 km/s, including metallic
aluminum, aluminum oxide, powdered meteorites, powdered minerals, glass fragments, and even fine-grained
powders. Consequently, we expect little difficulty differentiating between metallic and aluminum-oxide
impactors among the particles trapped by the ODC collectors on board Mir, or in recovering cometary par-
ticles of a wide variety of physical and chemical properties via the upcoming Stardust mission.
However, on the basis of the current experiments, it could prove impractical to relate dimensional measure-
ments of individual penetration tracks and associated projectile residues in space-exposed aerogels to detailed
initial impact conditions, such as the original particle size, mass, and/or velocity. While improved impact
simulations might provide more quantitative relationships than those presented here, it must be emphasized
that a variety of concepts currently exist (e.g., Auer and Bun, 1994; Tuzzolino, 1994; and others) to determine
these properties independently, including complete trajectory measurements on a particle-by-particle basis via
active instruments. If required by scientific need, such active detectors can be readily combined with any
passive aerogel collector.
Regardless, aerogel is a highly suitable, if not the most outstanding, capture medium over a wide range of
initial impact conditions, including impact velocities as high as 7 km/s, possibly higher. Aerogel should not
be faulted for yielding limited dynamic particle information. Its primary purpose and function is to trap the
residue(s) of hypervelocity impactors, either natural or man-made. It does so extremely well and without
peer.
56
9. References
Anderson, C.E., ed. (1993). Proc. Hypervel. Impact Symp. 1992, Int. J. Impact Engin., 14, 1-4, pp. 891.
Anderson, C.E., ed. (1995). Proc. Hypervel. Impact Symp. 1994, Int. J. Impact Engin., 17, 1-6, pp. 947.
Anderson, W., and Ahrens, T.J. (1994). Physics of interplanetary dust capture via impact into organic foam.
J. Geophys. Res. E 99, pp. 2063-2071.
Auer, S. and Bun, O.V. (1994). Highly transparent and rugged sensor for velocity determinations of cosmic-
dust particles. In Workshop on Particle Capture, Recovery and Velocity/Trajectory Measurement Tech-
nologies (M.E. Zolensky, ed.), Lunar and Planetary Institute, Houston, TX, LPI Technical Report 94-05,
pp. 25-29.
Barrett, R.A.; Zolensky, M.E.; Hörz, F.; Lindstrom, D.; and Gibson, E.K. (1992). Suitability of SiO
2
aerogel
as a capture medium for interplanetary dust. Proc. 22
nd
Lunar Planet. Sci. Conf., pp. 203-212.
Bernhard, R.P.; Hörz, F.; and Kessler, D.J. (1996). Orbital debris impacts on the trailing edge of the Long-
Duration Exposure Facility (LDEF). In Proc. 1
st
International Workshop on Space Debris, Moscow, 1995,
to be published by Space Research Institute/National Academy of Sciences, Moscow (in press).
Brownlee, D.E.; Hörz, F.; Hrubesh, L.; McDonnell, J.A.M.; Tsou, P.; and Williams, J. (1994). Eureka!
Aerogel capture of Meteoroids in space. Lunar Planetary Sci. XXIV, 183-184 (abstract).
Burchell, M.J. and Thomson, R. (1996). Intact hypervelocity capture in aerogel in the laboratory. In Shock
Compression of Condensed Matter-1995, Schmidt, S.C. and Tao, W.C., eds., AIP Conf. Proc. 370, Part 2,
pp. 1155-1158.
CDCF (1990), Cosmic Dust Collection Facility: Scientific Objectives and Programmatic Relations, CDCF
Committee Report (F. Hörz, ed.), NASA TM 102160, p. 29.
Christiansen, E.L.; Cykowski, E.; and Ortega, J. (1993). Highly oblique impacts into thick and thin targets.
Int. J. Impact Engin. 14, pp. 157-168.
Cintala, M. J. and Hörz, F. (1992). An experimental evaluation of mineral specific comminution, Meteoritics
27, pp. 395-403.
Fricke, L. (1988). Aerogels, Scientific American 258, pp. 92-97.
Gault, D.E. (1973). Displaced mass, depth, diameter, and effects of oblique trajectories for impact craters
formed in dense, crystalline rocks. The Moon 3, pp. 32-44.
Grewing, M.; Praderie, F.; and Reinhard, R., eds. (1986). Exploration of Halleys Comet, Springer Verlag,
Berlin, pp. 984.
57
Gwynn, D.W.; Hörz, F.; Bernhard, R.P.; and See, T.H. (1996). The dispersion of soda-lime glass projectiles
following penetration of thin aluminum membranes. Proc. HVIS 1996 Conf., Int. J. Impact Engin. Special
Issue (in print).
Hohler, V. and Stilp, A.V. (1987). Hypervelocity impact of rod projectiles from L/D from 1 to 32. Int. J.
Impact Engin. 5, pp. 323-332.
Hörz, F.; Cintala, M.J.; and Zolensky, M.E. (1992). Hypervelocity penetration tracks in very low-density,
porous targets. In Hypervelocity Impacts in Space (J.A.M. McDonnell, ed.) Univ. of Kent at Canterbury,
pp. 19-23.
Hörz, F.; Cintala, M.J.; Bernhard, R.P.; and See, T.H. (1994). Dimensionally scaled penetration experiments:
Aluminum targets and glass projectiles 50 µm to 3.175 mm in diameter. Int. J. Impact Engin. 15,
pp. 257-280.
Hrubesh, L.W. and Poco, J.F. (1990). Development of low-density silica aerogel as a capture medium for
hypervelocity particles. Lawrence Livermore National Laboratory Report UCLR-CR 105858 SUM, p. 12.
Jessberger, E.K.; Rahe, J.; and Kissel, J. (1989). The Composition of Comets. In Origin and Evolution of
Planetary and Satellite Atmospheres (Atreya, S.K.; Pollack, B.P.; and Matthews, M.S., eds.), The
University of Arizona Press, Tucson, AZ, pp. 167-191.
Kissel J., et al. (1986). Composition of comet Halley dust particles from Giotto observations. Nature 321,
pp. 280-282.
Love, S.G.; Joswiak, D.J.; and Brownlee, D.E. (1993). Densities of 5-15 µm interplanetary dust particles,
Lunar Planet. Sci. XXIV, 901-902 (abstract).
Maag, C. and Kelly-Linder, W. (1992). Results of Space Shuttle intact particle capture experiments. In
Hypervelocity Impacts in Space (J.A.M. McDonnell, ed.), University of Canterbury, pp. 186-190.
Mendez, D.J. (1995). Physical Characterization of SiO
2
Aerogel: Phase I Final Report. Lockheed Missiles
and Space Co., Contractor Report LMSC-P0166, pp. 95.
Orphal, D.L.; Anderson, C.E.; and Franzen, F.F. (1990). Impact calculations for L/D < 1 penetrators. 12
th
Int.
Symp. on Ballistics, San Antonio, TX.
Piekutowski, A. (1993). Characteristics of debris clouds produced by hypervelocity impact of aluminum
spheres with thin aluminum plates. Int. J. Impact. Engin. 14, pp. 573-586.
Schmidt, R.M.; Housen, K.R.; Piekutowski, A.J.; and Poormon, K.L. (1994). Cadmium simulations of orbital
debris shield performance to scaled velocities of 18 km/s. J. Spacecraft and Rockets 31, pp. 866-877.
58
Thomson, R. (1995). An Investigation of the Intact Capture of Hypervelocity Particles in Aerogel. Unpub-
lished MS Thesis, Univ. of Kent, Canterbury, UK, pp. 125.
Tsou, P. (1990). Intact capture of hypervelocity projectiles. J. Int. Impact Engin. 10, pp. 615-627.
Tsou, P. (1995). Silica aerogel captures cosmic dust intact. J. Non-Crystalline Solids 186, pp. 415-427.
Tsou, P.; Brownlee, D.E.; Lawrence, M.R.; Hrubesh, L.; and Albee, A.L. (1988). Intact capture of
hypervelocity micrometeoroid analogues. Lunar Planet. Conf. XIX, 1205-1206 (abstract).
Tuzzolino, A.J. (1994). PVDF Flux/Mass/Velocity/Trajectory Systems and their applications in space. In
Workshop on Particle Capture, Recovery and Velocity/Trajectory Measurement Technologies, Lunar and
Planetary Institute, Houston, TX, LPI Technical Report 94-05, pp. 89-93.
Werle, V.; Fechtig, H.; and Schneider, E. (1981). Impact Accretion Experiments. Proc. Lunar Planet. Sci.
Conf. 12
th
, pp. 1641-1647.
Zolensky, M.E., ed. (1994). Workshop on Particle Capture, Recovery and Velocity/Trajectory Measurement
Technologies, Lunar and Planetary Institute, Houston, TX, LPI Technical Report 94-05, p. 102.
Zook, H.A. (1991). Deriving the velocity distribution of meteoroids from the measured meteoroid impact
directionality on the various LDEF surfaces. In LDEF - 69 Months in Space, First Post Retrieval
Symposium, A.S. Levine, ed., NASA Conference Publication 3134, pp. 569-581.
59
REPORT DOCUMENTATION PAGE
Form Approved
OMB No. 0704-0188
Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and maintaining the
data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information, including suggestions for reducing
this burden, to Wwashington Headquarters Services, Directorate for information Operations and Reports, 1215 Jefferson Davis Highway, Sutie1 1204, Arlington, VA 22202-4302, and to the Office of
Management and Budget, Paperwork Reduction Project (0704-0188), Washington, DC 20503.
1. AGENCY USE ONLY
(
Leave Blank
)
2. REPORT DATE 3. REPORT TYPE AND DATES COVERED
April 1998 NASA Technical Memorandum
4. TITLE AND SUBTITLE 5. FUNDING NUMBERS
Capture of Hypervelocity Particles With Low-Density Aerogel
6. AUTHOR(S)
Friedrich Horz, Mark J. Cintala, Michael E. Zolensky, Ronald P. Bernhard*, William E.
Davidson*, Gerald Haynes*, Thomas H. See*, Peter Tsou**, Donald E. Brownlee***
7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) 8. PERFORMING ORGANIZATION
REPORT NUMBERS
Lyndon B. Johnson Space Center
Houston, Texas 77058
S-838
9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES) 10. SPONSORING/MONITORING
AGENCY REPORT NUMBER
National Aeronautics and Space Administration
Washington, D.C. 20546-0001
TM-98-207192
11. SUPPLEMENTARY NOTES
* Lockheed-Martin Space Mission Systems and Services; ** Jet Propulsion Laboratory; ***Dept. of Astronomy/University of
Washington
12a. DISTRIBUTION/AVAILABILITY STATEMENT 12b. DISTRIBUTION CODE
Unclassified/unlimited
Available from the NASA Center for AeroSpace Information (CASI)
800 Elkridge Landing Rd
Linthicum Heights, MD 21090-2934 (301) 621-0390 Subject Category: 88
13. ABSTRACT
(
Maximum 200 words
)
Recent impact experiments conducted at Johnson Space Center supported a space-exposed flight instrument called the orbital debris
collector (ODC) to see whether SiO2 aerogel performed adequately as a collector to capture cosmic dust particles and/or manmade
debris, or whether additional development is needed. The first ODC was flown aboard the Mir for 18 months, while the second will
be flown aboard a spacecraft (Stardust, to be launched in 1999) that will encounter the comet Wild 2 and return to Earth. Aerogels
are highly porous materials that decelerate high-velocity particles without substantial melting or modifications to the particles’
components; in other denser materials, these particles would melt or vaporize upon impact.
The experimental data in this report must be considered somewhat qualitative because they are characterized by substantial, if not
intolerable, scatter, possibly due to experimental difficulties in duplicating given sets of initial impact conditions. Therefore, this
report is a chronological guide of the experimenters’ attempts, difficulties, progress, and evaluations for future tests.
14. SUBJECT TERMS 15. NUMBER OF
PAGES
16. PRICE CODE
aerogel; space debris; cosmic dust; cometary atmospheres; deceleration; hypervelocity
impact; impact melts
6 0
17. SECURITY CLASSIFICATION
OF REPORT
18. SECURITY CLASSIFICATION
OF THIS PAGE
19. SECURITY CLASSIFICATION
OF ABSTRACT
20. LIMITATION OF ABSTRACT
Unclassified Unclassified Unclassified None
Standrard Form 298 (Rev Feb 89) (MS Word Mar 97)
Prescribed by ANSI Std. 239-18
298-102
NSN 7540-01-280-5500
