2Gallot et. al.
time of impact showed that the target, the 160m asteroid Dimor-
phos, resembles a rubble-pile. With the NASA-DART mission, im-
pacting an asteroid to deflect its trajectory is not science fiction
anymore. The Impact creates an important material ejection ob-
served seconds after the impact, but the brightness increase is still
measurable two weeks after ?. This means that materials escape at
very low velocity, and that the direct impact is not the only reason
for ejection, seismic waves propagating from the impact all around
the asteroid are also responsible for the material ejection (Tancredi
et al. 2022).
A few experimental works study impacts on grains in uncon-
fined (Yasui et al. 2015) or confined (van den Wildenberg et al.
2013; Mart´
ınez et al. 2021) media. The micro-gravity on asteroids
confers mechanical properties to granular media that are not easy
to reproduce on earth (Altshuler et al. 2014; Villalobos et al. 2022).
However, these experiments are fundamental to validate numerical
models, in particular Discrete Element Models (DEM) (Sch¨
opfer
et al. 2009; Wang & Mora 2009) for the complex physics of granu-
lar mechanics (Duran 2012).
Inhomogeneity of grain packing (Liu et al. 1995; Jaeger et al.
1996), together with material relaxation (Alexander 1998), explain
most of the complexity in granular media. The contacts between the
grains form a network that reorganizes under stresses (Mueth et al.
1998b; Howell et al. 1999; Cambau et al. 2013). Because of this
reorganization, most of the numerical and laboratory experiments
on grains begin with the preparation of the material. The quasi-
static problem of stress distribution has been addressed by Janssen
model (Nedderman et al. 1992). The observation of the chain force
provides some insight to discuss the model limitation due to corre-
lation length, microscopic features, reorganization, and hysteresis
(Ovarlez et al. 2003; Ovarlez & Cl´
ement 2005a; de Gennes 1999).
A spectacular consequence of reorganization is a jamming
transition from fluid to solid state (Cates et al. 1999; Liu & Nagel
1998; van Hecke 2009). Wave propagation in grains is an amazing
probing tool for mechanical parameters in granular media (Som-
fai et al. 2005; Jacob et al. 2008; Silbert et al. 2005), but its un-
derstanding is still challenging (Luding 2005), because of a va-
riety of phenomena such as nonlinear propagation (Zhang et al.
2020), nonlinear constitutive equations (Renaud et al. 2013; God-
dard 1990; Trarieux et al. 2014), wave dispersion (Chrzaszcz 2016;
Cheng et al. 2020), multiple scattering (Jia 2004; Tell et al. 2020;
Langlois & Jia 2015; Trujillo et al. 2011; Brunet 2006; Page et al.
1996), or path-dependent propagation (Hua & Van Gorder 2019;
Owens & Daniels 2011).
Effective Medium Theory (EMT) (Walton 1987) predicts a
scaling of the coherent wave speeds with pressure between p1/6
for Hertzian contact, or p1/3considering non-Hertzian contact or
variation in the coordination number C(Goddard 1990). Discrete
Element Models (DEM) and experimental observations confirmed
these numbers (see Jia et al. 2021, for a non-exhaustive review).
The contact between two grains can be described by Hertz-like
models; then, EMT stipulates that the macroscopic response of a
medium is the sum of an averaged grain-grain contact (Ovarlez &
Cl´
ement 2005b; Kocharyan & Karanjgaokar 2022). This strong hy-
pothesis of linearity explain why EMT fail to explains many ob-
servations where the scaling law exponent is shown to depend on
the pressure range (Makse et al. 2004), stress history (Cheng et al.
2020), wave macroscopic amplitude (Wichtmann & Triantafyllidis
?See NASA News: https://www.nasa.gov/feature/nasa-dart-imagery-
shows-changed-orbit-of-target-asteroid
2004), and local amplitude around force chains (Owens & Daniels
2011).
In this work we propose an experimental study of laboratory
scaled asteroid impacts. We use granular media as a model aster-
oid. There are two fundamental differences between a real asteroid
and our experiment: the gravity conditions and the impact velocity.
Self gravity induces a pressure distribution inside an asteroid that is
not well known (see different estimates by: Cheng 2004; Sharma
2013; Zhang et al. 2018); but it certainly presents a pressure gradi-
ent with increasing values towards the interior of the body. For this
reason, the granular media is confined and the impacts are realized
for different confining pressure steps. As regard to the low velocity
of our impactors, we did study neither the crater geometry nor the
energy transfer that would depend on the impact velocity. Instead
we were interested in the wave propagation outside of the impact
zone.
In the present experimental work, we face the whole complex-
ity of the quasi-static and dynamic mechanical response of granular
media. In Section 2 we present the experimental setup, the charac-
teristic of the materials and the devices used in the experiments.
This is why our main parameter is confining pressure of the granu-
lar media. In Section 3 we present the quasi-static response in glass
beads. Then, the results of the impact-generated and shaker-born
seismic waves as a function of the confining pressure are presented
in Section 4.
2 EXPERIMENTAL SETUP
The experimental approach focuses on the propagation of short
waves generated by perturbations due to impacts or vibration on
the surface of a box containing a confined granular media. The box
is a cube of side L= 50 cm (internal distance between the lat-
eral walls, Fig. 1). The walls are made of 14 mm thick transparent
acrylic. The cube rests on a moving platform with a sliding top lid.
A circular opening of 16 cm in diameter allows the direct impact
of the projectile or the contact of the shaker with the material. The
inner top lid is stationary as it is welded to the hydraulic press struc-
ture. The box is uplifted by the hydraulic jack (Enerpac RC106 with
a 15-cm stroke), compressing the material. The hydraulic press has
been designed for a 10 tonnes maximum load.
Experiments are performed using three different granular ma-
terials: glass beads (artificial), sand, and gravel (both natural). Size
distributions are shown in Fig. 2, while angularity, sphericity, den-
sity, and volume fraction are described in Table 1. The granular ma-
terial, stored in a 100-liter barrel, is positioned over the hydraulic
press structure using an electric winch. A plug at the bottom of the
barrel is removed releasing the material, filling up the box through
its upper aperture. By the end of this process, the accelerometers
and pressure sensors inside the box are completely covered (see be-
low for a description of the location of these devices). The barrel is
weighted in order to have 195±1kg of grains inside the box. After
discharge, the pile needs to be manually even. Material preparation
consists of a series of five compression-relaxation cycles from 0
to 5 tonnes. This procedure rearranges the grains on the top of the
pile, flattening the surface.
The coordinate system (−→
x,−→
y,−→
z) has its origin at the impact
point as shown in Fig. 1. The perturbations are generated along the
x-direction. An array of 3-axis accelerometers (Analog Devices,
ADXL327, ±2g sensitivity 0.42 V/g) embedded in the granular
media registers the vibrations. The accelerometers were located in a
vertical array, at a horizontal distance of y0= 9 cm from the impact