Carrier (1991) Physical Propeortis of The Lunar Surace - Lunar Sourcebook-Chapter09
Carrier (1991) Physical Propeortis of The Lunar Surace - Lunar Sourcebook-Chapter09
Carrier (1991) Physical Propeortis of The Lunar Surace - Lunar Sourcebook-Chapter09
PHYSICAL PROPERTIES OF
THE LUNAR SURFACE
W. David Carrier III, Gary R. Olhoeft, and Wendell Mendell
This chapter discusses the physical properties of the The physical properties of lunar soils and rocks
lunar surface, determined at depths varying from a have been measured in situ by robots and astronauts,
few micrometers to a few tens of meters. The in the laboratory on returned samples, and by remote
properties considered include geotechnical properties sensing from the Earth’s surface and from lunar orbit.
(those properties of a planetary surface needed to The results of these measurements are presented in
evaluate engineering problems, including the the following sections.
mechanical properties of soil and rock), electrical and
electromagnetic properties, and the reflection and
9.1. GEOTECHNICAL PROPERTIES
emission of radiation.
The lunar surface material (regolith or “lunar soil”) is “I’m at the foot of the ladder. The LM [Lunar
a complex mixture of five basic particle types: Module] footpads are only depressed in the
crystalline rock fragments, mineral fragments, surface about 1 or 2 inches, although the
breccias, agglutinates, and glasses. The relative surface appears to be very, very fine-grained, as
proportion of each particle type varies from place to you get close to it, it’s almost like a powder;
place and is dependent on the mineralogy of the down there, it’s very fine ... I’m going to step off
source rocks and the geologic processes that the rocks the LM now. That’s one small step for (a) man.
have undergone (see Chapters 4 and 7 for more One giant leap for mankind. As the—The surface
details). is fine and powdery. I can—I can pick it up
The lunar regolith is distinctly different from loosely with my toe. It does adhere in fine layers
terrestrial materials. For one thing, its mineral like powdered charcoal to the sole and sides of
composition is limited. Fewer than a hundred my boots. I only go in a small fraction of an inch.
minerals have been found on the Moon, compared to Maybe an eighth of an inch, but I can see the
several thousand on Earth (see Chapter 5). footprints of my boots and the treads in the fine
Furthermore, continuous meteoroid impact has sandy particles.”
converted the lunar surface materials into a well-
graded silty sand that has reached a “steady state” in Neil A. Armstrong Tranquillity Base (Apollo 11),
thickness, particle size distribution, and other July 20, 1969
properties at most locations on the Moon. Hence, the
limited particle types and sizes, the absence of a lunar Prior to the Luna, Surveyor, and Apollo soft landings
atmosphere, and the lack of water and organic on the Moon, various theories had been proposed
material in the regolith constrain the physical regarding the nature of the lunar surface. One theory
properties of the lunar surface material to relatively was that the lunar craters were volcanic in origin and
narrow, well-defined ranges. The primary factors that that the surface was mostly lava and hence very hard.
affect the physical properties are bulk density, relative The other popular theory was that the craters had
density, and temperature. been formed by meteoroid impact
475
476 Lunar Sourcebook
and that the surface was a regolith or soil with diverse in case it disappeared into the soft dust at touch-
particle sizes. In short, the lunar surface might be down (Cooper, 1969, 1970). Just a few months before
either very hard or very soft. The “hard soil” advocates the flight of Apollo 11, it was seriously suggested that,
believed that the ultrahigh vacuum conditions on the when Astronaut Armstrong took his first steps on the
Moon would produce “cold welding” between the soil lunar surface, the soil would jump onto his spacesuit
particles. This phenomenon is known to occur if two because of the electrostatic attraction. The resulting
carefully cleaned surfaces (e.g., of aluminum) are coating of soil covering his suit was expected to be so
brought together in a vacuum. The lack of water and thick that he would not be able to see and might not
organic molecules in the lunar vacuum would allow be able to move. Then, if he were able to stagger back
the surfaces to bond atomically or weld. Experiments into the LM, there was concern that the highly
had also shown that the mineral mica could be reduced soil would burst into flames when the cabin
cleaved and pressed back together in a vacuum. was repressurized with pure oxygen. Even after
Thus, the lunar surface was thought to be porous and several successful Apollo missions to the Moon, some
yet rock-hard, like pumice. At the other extreme, it scientists were still convinced that the lunar soil was
was believed that the combination of low gravity, like fresh snow and that an astronaut could push a
vacuum, and electrostatic charges on the soil rod into it to almost any desired depth.
particles would produce a “fairy-castle” structure, We now know that the lunar surface is neither so
with a consistency similar to sifted flour. soft, nor so hard, as was once imagined. We also
Because of the long lead time required, the design of know that the astronauts can land a spacecraft far
the Apollo Lunar Module (LM) landing gear had to be more gently than the designed 3 m/sec. In fact, the
completed in the mid-1960s, before there were any landing struts were stroked only slightly on all of the
direct tactile data available about the lunar surface. six Apollo landings. Furthermore, we know that the
Because of the different theories that were being hotly ranges of geotechnical properties of lunar materials
debated at that time, the engineers chose a landing are less than those that occur in surficial materials on
gear that could accommodate a wide range of Earth. This is because of the following factors:
conditions. To accommodate the “soft” extreme of 1. The familiar terrestrial geologic processes of
possible lunar surface properties, the LM landing pad chemical weathering, running water, wind, and
was sized so that it could sink as much as 60 cm into glaciation are absent on the Moon. These processes
the soil and would have a bearing pressure on the tend to produce well-sorted sediments with uniform
Moon of only approximately 4.6 kPa. This value is grain sizes. The primary lunar soil-forming process is
only about one-fifth of the stress that a typical person meteoroid impact, which produces a heterogeneous
exerts when standing on one foot on Earth. To and well-graded (poorly-sorted) soil.
accommodate the “hard” extreme, a shock absorber 2. The three main constituents most likely to
system was built into the LM landing struts; this produce unusual or “problem” soils on Earth are
system would allow a safe landing on concrete at a absent on the Moon; there is no water, and there are
vertical velocity of up to 3 m/sec, which is equivalent no clay minerals or organic materials.
to free-fall from a height of 2.9 m in lunar gravity. 3. As discussed in Chapters 5 and 7, the variety of
This shock absorber consisted of a honeycomb minerals in lunar soil is much less than that found on
material that would crush as the landing struts were Earth. In fact, many soil particles are simply
forced into it during touchdown, thereby dissipating fragments of rocks, minerals, and glass stuck together
the energy of impact. Unlike the normal hydraulic with glass (agglutinates).
shock absorber system in an aircraft, the landing gear As a result, the geotechnical properties of lunar soil
on the LM only had to work once and could therefore tend to fall in a fairly narrow range, and the most
use the crushable design, which was both simpler significant variable is the relative density, defined in
and lighter. section 9.1.5. To a certain extent, relative density also
The landing pads of the pre-Apollo Surveyor controls the other physical properties discussed in
spacecraft were sized for a bearing pressure of 2.2 this chapter.
kPa, and when these unmanned probes were To set the stage for discussion of the effects of
successfully soft-landed on the Moon during 1966- relative density, other properties will be described
1968, the design of the Apollo LM pads was confirmed first: particle size distribution, particle shapes,
as conservative. In spite of this evidence, some specific gravity, bulk density, and porosity. Relative
scientists were still concerned about the bearing density will then be described. After that, the
capacity of the lunar surface and recommended that influence of these properties on compressibility, shear
the astronauts have snowshoes and that the LM be strength, permeability, bearing capacity, slope
equipped with a long radio antenna stability, and trafficability will be considered.
Physical Properties of the Lunar Surface 477
TABLE 9.1. Weight distribution in size fractions of representative scooped surface soils
(from Morris et al., 1983; data emphasize coarser fractions).
Thus, particle shape has a significant influence on Roundness. Roundness is defined as the ratio of the
bulk physical properties. Although shape is difficult to average of the radii of the corners of the particle image
quantify, a number of measurements have been made, to the radius of the maximum inscribed circle.
and the results are summarized in Table 9.2. The Mahmood et al. (1974b) measured roundness on
various shape parameters are discussed in the silhouettes of the same 1136 particles for which they
following sections. had measured elongation. The weighted average
Elongation. Elongation is defined as the ratio of the roundness values varied from 0.19 to 0.26, corres-
major to intermediate axes of the particle, or length to ponding to subangular particle shapes. They also
width. Particles with values of the ratio <1.3 are measured roundness in direct light on 641 particles
considered equant, and particles whose ratio is >1.3 from the same three lunar samples and found that the
are elongate. weighted average value varied from 0.20 to 0.25,
Mahmood et al. (1974b) measured elongations of corresponding to angular particles. More details are
1136 particles from three lunar soil samples given in Table A9.4.
(14163,148, 14259,3, and 15601,82). Individual Volume coefficient. Volume coefficient is defined as
particle sizes ranged from 2300 to <44 µm. The the volume of a particle divided by the cube of the
weighted average elongation ranged from 1.31 to 1.39. diameter of the circle that encloses the same area as
More details are given in Table A9.2. the particle profile. Heywood (1971) determined the
Heywood (1971) measured elongation on 30 volume coefficient of 6755 particles taken from one
particles with a nominal size of 700 µm, all taken from Apollo 12 sample (12057,72); particle sizes ranged
one soil sample (12057,72). His values ranged from from 733 to 60 µm, which represents roughly the
1.32 to 1.38. Heywood also measured six other shape coarser half of a typical lunar soil sample. The results
coefficients for these same particles: flatness, area are plotted in Fig. 9.2, where it can be seen that the
ratio, volume coefficient, rugosity coefficient, specific volume coefficient varies from 0.24 to 0.37, with an
circularity of profile, and specific circularity including average value of about 0.3; this value corresponds
the effect of rugosity. More details are given in Table approximately to a prolate spheroid with a major-to-
A9.2. minor axis ratio of 3 to 1.
Aspect ratio. In geotechnical studies, aspect ratio is The volume coefficient for a sphere is greater than
inversely related to elongation; it is defined as the 0.52, well above the measured values shown in Fig.
ratio of the minor axis to the major axis of an ellipse 9.2. Heywood found that the glass spherules represent
fitted to the particle by a least-squares approximation. only about 1 in 500 particles in the coarse fraction of
Görz et al. (1971, 1972) measured aspect ratios on this Apollo 12 sample, and hence their effect on
2066 particles from seven lunar samples; particle geotechnical properties is negligible. [Immediately
sizes ranged from 1.25 to 30 µm. Values of the aspect after the Apollo 11 mission, some scientists had
ratios varied from 1.0 (equant) to <0.1 (very elongate), speculated that the “slipperiness” of the lunar soil
with most values falling in the range 0.4 to 0.7 reported by the astronauts was caused by the
(slightly-to-moderately elongated). Görz et al. (1972) presence of a large proportion of such glass beads.
also measured the shape complexity factor for most of This was even commented upon in the September
these same particles. More details are given in Table 1969 issue of Scientific American. The observed
A9.3. amounts of such beads in most lunar soils
480 Lunar Sourcebook
(<1%) are too low to produce any such mechanical measured the shapes of monomineralic silicate
effects.] mineral fragments in a lunar soil breccia (10065), in
Fourier analysis. Particle shape can also be lunar drill-core soils (15003,6059), and in brecciated
quantified from the two-dimensional image of a achondrites (meteorites). Results show that the
particle. Commonly, a two-dimensional image is the general shapes of lunar soil particles and those of
only “sample” available for shape analysis when clasts in meteoritic breccias are similar (Fig. A9.1).
working with thin sections of grain mounts. The The shapes of these particles are different from those
outline of a two-dimensional image is digitized using a that have undergone abrasion in water and wind on
polar coordinate system with the center of the image Earth.
as the origin. The measurements are expressed in the Specific surface area. Specific surface area is
form of a Fourier expansion series (Ehrlich and defined as the surface area of a particle divided by its
Weinberg, 1970). Briefly, an image is expressed as mass, usually expressed in square meters per gram.
It is a measure of both the size and the shape of the
particle. For a spherical particle, the specific surface
area (SSA) is inversely proportional to the diameter,
where θ = polar angle; Ro = average radius, n = and is given by
harmonic order, Rn = harmonic amplitude, and φ=
phase angle. The shape of the grain thus becomes
represented as the sum of the radii, expressed as a
function of polar angles and phase angles, ideally for where d is the diameter of the sphere in micrometers;
every point on the outline of the image. In essence, G is the specific gravity; and ρw is the density of
each harmonic may be viewed as a system of flowers water, 1 g/cm3.
with two petals (harmonic 2), three petals (harmonic For example, uniform, solid spheres with a specific
3), four petals (harmonic 4), . . . n petals (harmonic n). gravity of 3.1 (typical for lunar soils, see section
The amplitude of each harmonic is the contribution of 9.1.3) would have the following specific surface areas:
that harmonic to the shape (in this case the outline of
the image) of the particle. Comparison between
samples is achieved by amplitude spectral analysis
and/or by analyzing the amplitude frequency
distribution at selected harmonics. Kordesh et al.
(1982) and Kordesh (1983a,b)
Physical Properties of the Lunar Surface 481
A “soil” consisting of spheres with the same In order to determine the specific gravity of a soil, a
submillimeter particle size distribution as lunar soil portion of the material is first weighed, and then
would have a SSA of about 0.065 m2/g. This SSA immersed in a fluid to measure the volume that it
value corresponds to an equivalent diameter of displaces. Various fluids can be used, including
approximately 30 µm, which is less than the average water, air, or helium. The specific gravity of lunar
particle size by weight of lunar soils (~65 µm). Because soils, breccias, and individual rock fragments has
of the reciprocal relationship shown above, the smaller been measured by various investigators, and the
particles in a soil tend to contribute most to the bulk results are summarized in Table 9.3. Values for lunar
SSA value. soils range from 2.3 to >3.2; we recommend a value
By comparison, terrestrial clay minerals have much of 3.1 for general scientific and engineering analyses
higher SSA values, owing to their very small size and of lunar soils.
platy shape: The average specific gravity of a given lunar soil is
related to the relative proportions of different particle
types; i.e., basalts, mineral fragments, breccias,
agglutinates, and glasses (Fig. 7.1). However, the
interpretation of the specific gravity is complicated by
the porosity of the particles. As illustrated in Fig. 9.3,
the porosity may be divided into three categories: (1)
intergranular porosity, or the volume of space
between individual particles;
The results of 17 measurements of the SSA of lunar (2) intragranular porosity, or the volume of reentrant
soils are tabulated in Table A9.5 (Cadenhead et al., surfaces on the exterior of the particles; and
1977). All these measurements were performed on the (3) subgranular porosity, or the volume of enclosed
submillimeter soil fraction by means of nitrogen gas voids within the interior of particles.
adsorption. The SSA values range from 0.02 to 0.78 When the soil particles are immersed in a fluid, the
m2/g, with a typical value of 0.5 m2/g, which intergranular and intragranular space is filled, but
corresponds to an equivalent spherical diameter of 3.9 not the subgranular space. Thus, the measured
µm. Hence, the SSA of lunar soil is much less than specific gravity is not solely an index of particle
that of terrestrial clay minerals, and yet it is mineralogy, but also includes the effect of enclosed
significantly greater than can be accounted for by voids. By suspending the soil particles in a density
small particle sizes alone. Rather, the relatively large gradient, produced by varying the proportions of a
SSA of lunar soils is indicative of the extremely mixture of methylene iodine and dimethyl formamide,
irregular, reentrant particle shapes. Duke et al. (1970b) found the following values of
These considerations suggest in turn the definition specific gravity:
for a new shape parameter that is independent of the
particle size, which is called the equivalent surface
Agglutinate and glass particles 1.0 to >3.32
area ratio (ESAR)
Basalt particles >3.32
Breccia particles 2.9 to 3.10
where ds = equivalent diameter of spheres with the The enclosed voids in a lunar soil particle with a
same particle size distribution as the soil being tested specific gravity of 1.0 occupy two-thirds of the total
and de = equivalent spherical diameter of the soil, as volume of the particle. Thus, the average specific
calculated from the specific surface area. gravity of the particles would be even greater if there
For a typical lunar soil, the equivalent surface area were no enclosed voids. For example, if the lunar soil
ratio would be nearly 8; that is, the soil would have were ground into a fine powder (in which the
eight times as much surface area as would an resulting particles were smaller than the enclosed
assemblage of spheres with the equivalent particle voids), these voids would be destroyed, and the
size distribution. specific gravity would be increased. The actual
subgranular porosity of individual lunar soil particles
9.1.3. Specific Gravity is only poorly known, and additional measurements
of subgranular porosity are needed.
The specific gravity, G, of a soil particle is defined as
the ratio of its mass to the mass of an equal volume The intragranular porosity has a strong effect on
the bulk density of the lunar soil, whereas the
of water at 4°C. Many terrestrial soils have a specific
gravity of 2.7; that is, the density of the individual intergranular porosity affects both the bulk density
and the relative density. These relations will be
particles is 2.7 g/cm3, or 2.7 times that of water (1
g/cm3). discussed in more detail in the following sections.
482 Lunar Sourcebook
9.1.4. Bulk Density and Porosity 1965) in an effort to estimate lower-bound bearing
capacities. Halajian (1964) also assumed a very low
The bulk density, ρ, of soil is defined as the mass of density of 0.4 g/cm3, but believed that the strength of
the material contained within a given volume, usually the lunar surface was similar to that of pumice.
expressed in grams per cubic centimer. The porosity, Robotic measurements of bulk density on the
n, is defined as the volume of void space between the lunar surface. When Surveyor 1 landed on the Moon
particles divided by the total volume. Bulk density, in June 1966, a much higher density of 1.5 g/ cm3
porosity, and specific gravity are interrelated as was deduced by Christensen et al. (1967), using
records of the interaction between the lunar soil and
the spacecraft footpads, combined with analysis of the
television images, to determine the particle size
distribution. Shortly thereafter, the first in situ
where G = specific gravity (including subgranular measurement of lunar soil density was made by the
porosity; section 9.1.3); ρw = density of water = 1 U.S.S.R. probe, Luna 13, using a gamma-ray device.
g/cm3; and n = porosity, expressed as a decimal The calibration curve for this device was double-
(combining both inter- and intragranular porosity). valued, and the data obtained were consistent with
It is convenient in geotechnical engineering to also two density values: 0.8 and 2.1 g/cm3; Cherkasov et
define another parameter, the void ratio, which is al. (1968) chose the lesser value. Based on the results
equal to the volume of void space between the from the soil mechanics surface sampler experiments
particles divided by the volume of the “solid” particles on Surveyor 3 and 7, Scott and Roberson (1967,
(again, including the subgranular porosity). Void 1968a,b) confirmed the Surveyor 1 value of 1.5 g/cm3,
ratio, e, and porosity are interrelated as and Scott (1968) suggested that the Soviet
investigators had chosen the wrong portion of their
calibration curve. Just before the Apollo 11 landing,
Jaffe (1969) reevaluated the Surveyor data and
proposed that the bulk density was 1.1 g/cm3 at the
surface and increased linearly to 1.6 g/cm3 at a depth
The in situ bulk density of lunar soil is a funda- of 5 cm.
mental property. It influences bearing capacity, slope Later, during the early 1970s, the U.S.S.R.
stability, seismic velocity, thermal conductivity, unmanned roving vehicles Lunokhod 1 and Lunokhod
electrical resistivity, and the depth of penetration of 2 traversed a total of 47 km on the lunar surface.
ionizing radiation. Consequently, considerable effort These vehicles performed approximately 1000 cone-
has been expended over the years in obtaining vane penetrometer tests to depths of 10 cm (see
estimates of this important parameter. section 9.1.7). Leonovich et al. (1974a, 1975) correlated
Prior to the soft landings on the Moon, remote- the lunar surface penetration resistance with
sensing techniques were used to infer the bulk measurements on returned lunar samples from Luna
density of the lunar soil. These techniques included 16 and Luna 20, and they deduced an average
passive measurements of optical, infrared, and surficial bulk density of about 1.5 g/cm3.
microwave emissivity and active measurements of Inferred values of bulk density based on
radar reflectivity. With the Surveyor and Luna correlations with simulated lunar soil. The in situ
unmanned landings, direct measurements were bulk density of lunar soil has also been estimated from
possible at discrete points. In addition, correlations analyses of astronaut bootprints, vehicle tracks,
with laboratory tests on simulated lunar soil boulder tracks, and penetration resistance. The
permitted extrapolation over wider areas. Finally, results are summarized in Table 9.4. Mitchell et al.
beginning with Apollo, core tube samples of lunar soil (1974) found that the astronaut bootprints indicated a
were returned that permitted unambiguous density of approximately 1.45 to 1.59 g/cm3,
measurements of the in situ bulk density. At present, representative of the top 15 cm of the lunar surface in
the best estimate for the average bulk density of the the intercrater areas. The soils on crater rims were
top 15 cm of lunar soil is 1.50 ± 0.05 g/cm3, and of found to be slightly less dense: 1.34 to 1.57 g/cm3.
the top 60 cm, 1.66 ± 0.05 g/cm3 (Mitchell et al., Analysis of vehicle tracks made by the Modular
1974). Equipment Transporter (Apollo 14) and the Lunar
A summary of estimated lunar bulk densities is Roving Vehicle (Apollo 15, 16, and 17) indicated values
presented in Table 9.4, together with recommended of 1.40 to 1.56 g/cm3, also representative of the top 15
typical values. The various estimating and measuring cm. The tracks left by boulders that had rolled
techniques are described in the following sections. downslope at the Apollo 17 site indicated a density of
Early inferred values of bulk density based on 1.38 to 1.68 g/cm3,
remote sensing. As shown in Table 9.4, a very low
density of 0.3 g/cm3 was assumed by Jaffe (1964,
484 Lunar Sourcebook
representative of the top 300 to 400 cm of the regolith core tube samples have been recovered, from depths
(Mitchell et al., 1974). Costes et al. (1971) analyzed the of up to 3 m. Density data for these samples are
resistance of the soil to penetration by the flagpole, the summarized in Tables A9.6 and A9.7, respectively.
Solar Wind Composition Experiment staff, and other The drive core tube sampler used on the Apollo 11
tools at the Apollo 11 and 12 sites; they deduced mission is shown in Fig. 9.4, and a photograph of
upper limits for the densities of 1.81 to 1.92 g/cm3 Astronaut Aldrin hammering one of the core tubes
and 1.80 to 1.84 g/cm3, respectively. Finally, Mitchell into the lunar surface is shown in Fig. 9.5. Two core
et al. (1974) found that penetrometer measurements tube samples were recovered on Apollo 11, and
made by Lunokhod 1 and on the Apollo 14, 15, and 16 densities measured on the returned samples were.
missions indicated densities of 1.58 to 1.76 g/cm3, 1.59 and 1.71 g/cm3 (see Table A9.6). The Apollo 11
representative of the top 60 cm. drive core tube utilized a reverse-flare bit, as indicated
In each of the above analyses, it was necessary to schematically in Fig. 9.6. Thus, if the surface soil were
assume that the lunar soil behaves approximately the initially loose, the shape of the bit would tend to
same as a simulant of crushed basaltic lava with a compress the soil as it entered the core tube, thereby
similar particle size distribution, after correction for leading to an overestimation of the in situ bulk
the effect of gravity. These correlations were developed density. On the other hand, if the soil were initially
by normalizing with respect to relative density (see dense, the shearing action of the bit would tend to
section 9.1.5). During the early Apollo. missions, when loosen the soil, leading to an underestimation of the in
very little data were available, it was necessary to situ density. As a result of this ambiguity, Scott et al.
assume that both the specific gravity and maximum (1971) could only conclude that the value of the in situ
and minimum porosity values for lunar soil were the density was somewhere between 0.75 and >1.75
same as for the simulant. Later, it became clear that g/cm3. A photograph of one of the Apollo 11 core tube
this was not the case. Hence, these interpretive samples as returned to the Lunar Receiving
methods are better estimators of relative density than Laboratory is shown in Fig. 9.7.
of bulk density, and will be discussed in more detail in The drive core tube bit used on the Apollo 12 and 14
section 9.1.5. missions did not have a reverse flare, as shown in Fig.
Measurements of bulk density on returned core 9.6. However, the wall thickness of the bit still caused
samples. Beginning with the Apollo 11 mission, core significant disturbance to the soil sample. The in situ
tube samples of lunar soil were collected and returned density at the Apollo 12 site was initially estimated to
from all of the Apollo sites and three of the Luna sites. be from 1.6 to 2 g/cm3 (Scott et al., 1971). Core tube
Such cores are important because they provide a more simulations performed later by Houston and Mitchell
or less continuous section into the uppermost lunar (1971) and Carrier et al. (1971) yielded additional
regolith, to depths up to 3 m. estimates of 1.55 to 1.9 g/cm3 and 1.7 to 1.9 g/cm3,
Two different types of coring tools were used to respectively.
sample the regolith: Drive core tubes were hollow tubes Vinogradov (1971) estimated a value of 1.2 g/cm3 for
hammered vertically into the regolith by an astronaut a 101-g rotary drill sample returned by Luna 16 (Fig.
to depths of less than a meter. Rotary drill core tubes 9.8), and Carrier et al. (1972a) estimated in situ
were drilled into the surface to depths of several densities of 1.45 to 1.6 g/cm3 for the Apollo 14 core
meters. tube samples.
Altogether, nearly 16 kg of drive core tube materials The shapes of the drive core tube bits used on the
have been recovered, using core tubes driven to depths Apollo 11, 12, and 14 missions not only affected the
of approximately 70 cm into the lunar surface. In measurements of the in situ soil density, they also
addition, more than 4 kg of rotary drill complicated the interpretation of the relationships
Fig. 9.4. Core tube sampler used to obtain cores of lunar soil during the Apollo 11 mission. Separate tubes,
each 2.8 cm in diameter (outside diameter) and 39.9 cm long, could be connected into a single tube that was
progressively hammered into the lunar surface by an astronaut. (NASA Photo S69-31856.)
486 Lunar Sourcebook
Fig. 9.6. Comparative cross-sections with dimensions (in centimeters) of the tips of the core tubes and drill bits
used to collect core samples of lunar soil on the various Apollo missions.
Physical Properties of the Lunar Surface 487
Fig. 9.7. Photograph of Apollo 11 core tube sample 10005 (S/N 2007), immediately after opening of the tube in
the Lunar Receiving Laboratory at the NASA Johnson Space Center in 1969. The core was opened by extruding
the liner, which consisted of two hemicylindrical, thin aluminum sleeves held together by teflon shrink tubing.
The teflon was cut off and the upper half of the aluminum sleeve lifted off without disturbing the core inside.
The lunar soil material densely fills the core tube, but some cracking and disintegration are visible in the
middle of the portion of the core shown. The white plug is a teflon stopper inserted into the core tube. Metal bar
at top gives scales in inches and centimeters. (NASA Photo S-69-45048.)
Fig. 9.8. Photograph of the U.S.S.R Luna 16 drill core sample collected and returned by a robot lander in 1970.
The returned core, about 35 cm long and collected in a flexible tube, was extruded out into a metal tray for
examination. This picture shows a section of the core about 40 cm long (after being shaken into the tray); the
material in the tray is about 3 cm across. Upper part of the core is at the top (left); coarse rock fragments can
be seen in the lower part of the picture (bottom; right). (Photograph from Vinogradov, 1971.)
488 Lunar Sourcebook
between density and depth. The problem arises from sample disturbance in some of the drill stems
the fact that a depth of, e.g., 50 cm in an early Apollo (tubes). Depth estimates were developed by Carrier
core tube sample does not correspond to the same (1974) and are summarized in Fig. A9.3.
depth below the lunar surface. In general, the sample
A completely new drive core tube was used on the
lengths in the core tube are less than the depths to
Apollo 15, 16, and 17 missions (Fig 9.9). Because of
which the cores were driven. These discrepancies arise
its greater diameter, each tube recovered more than
from the process by which the core tubes penetrated
four times as much sample as did the earlier Apollo
the soil as they were hammered in. As the core tube
core tubes. Furthermore, the much thinner wall (see
penetrates the surface, a portion of the soil enters the
Fig. 9.6) greatly reduced sample disturbance during
tube and a portion is pushed outward; the portion
collection. As a result, samples collected with this new
that enters the tube may be either compressed or
tube yielded very accurate measurements of the in
loosened. As the depth increases, the portion entering
situ bulk density. Sampling with a double core tube at
the tube progressively decreases until no more sample
Station 9A at the Apollo 15 site is shown in a series of
is recovered; the tube is now plugged, and the deeper
photographs in Fig. 9.10. The neat, round hole left in
soil is simply pushed aside. Detailed laboratory
the lunar surface after the core tube was withdrawn is
simulations were performed with the Apollo 11, 12,
shown in Fig. 9.10d.
and 14 core tubes, and approximate depth
Based on samples collected in these new core tubes,
calibrations were developed by Carrier et al. (1971,
the in situ bulk density of the lunar soil at the Apollo
1972a). The results are summarized in Fig. A9.2.
15 site was estimated to vary from 1.36 to 1.85 g/cm3
The same types of laboratory simulations were used
(Carrier et al., 1972a). A plot of density vs. depth for
to evaluate the depth relationships for the rotary drill
each of the individual core tubes is shown in Fig.
cores used on the Apollo 15, 16, and 17 missions. In
9.11; the bands indicate the small uncertainty
some of these returned core samples, the
associated with measurements of the returned bulk
interpretation was complicated by obvious
density. The in situ bulk density at the Apollo 16 site
was estimated to vary from 1.40 to
Fig. 9.9. Photograph of the core tube sampler used on the Apollo 15-17 missions. The unit consists of two
single tubes, each 4 cm in (inside) diameter and 42 cm long, that can be connected to form a double tube. The
“rammer jammer” was used to insert and seat the sample retainer before the core was pulled from the ground.
The larger diameter and thinner tube walls on this sampler made possible the recovery of larger and less
disturbed samples than had been collected with core samplers used on earlier Apollo missions.
Physical Properties of the Lunar Surface 489
Fig. 9.10. Sequential photographs showing collection of a deep core sample of lunar soil on the Apollo 15
mission (Station 9A), using the new double core tube sampler. The core tube barrel is 4.4 cm in diameter. The
tripod-like device at the left is a gnomon that measures the sun angle and local vertical and also contains a
comparative color chart. (a) Core tube in position, ready for penetration (NASA Photo AS-15-82-11160); (b) core
tube pushed to a depth of about 22 cm (NASA Photo AS-15-82-11161); (c) core tube driven to a final depth of
68 cm after approximately 50 hammer blows (NASA Photo AS-15-82-11162); (d) open hole remaining in lunar
surface material after withdrawal of the core tube (NASA Photo AS-15-82-11163). Note the numerous well-
molded footprints of the astronaut’s boots in the lunar surface surrounding the hole.
490 Lunar Sourcebook
For this relation, the density is given by Thus, at z = 0, the surface density is 1.30 g/cm3,
and as z increases indefinitely, the maximum density
approaches 1.92 g/cm3, which appears to be a
494 Lunar Sourcebook
Fig. 9.16. Calculated in situ bulk density in the lunar soil layer as a function of depth below the surface,
derived from data shown in Fig. 9.15. Three calculated density-depth relations are presented: linear (solid line),
power-law (heavy dashed line), and hyperbolic (light dashed line). Boxes at the top of the plot show
recommended near-surface bulk density values for intercrater areas.
reasonable upper limit. The integrated mass depth for (1975a,b) reanalyzed lunar microwave emission data
this density relationship is then given by and concluded that the soil layer may be 10 to 30 m
thick over a large portion of the Moon. This should be
considered a tentative prediction until more data are
obtained (see also Fig. 4.21).
Fig. 9.17. Diagrams showing packing of ideal uniform spheres into two extreme configurations, together with
the resulting values of porosity, void ratio, bulk density, and relative density for each. (a) Face-centered cubic
packing, the loosest possible arrangement of uniform spheres. (b) Hexagonal close packing, the densest possible
arrangement of uniform spheres.
be drilled into the lunar soil (Langseth et al., 1972). Table 9.7. In those cases where the specific gravity is
Each auger tube had a smooth interior, with a helical also known, the corresponding maximum and
structure on the outside of the tube to do the actual minimum void ratios have also been calculated.
drilling. A long probe containing sensors and The Apollo 11 minimum and maximum densities
electronics was then to be inserted into each hollow reported by Costes et al. (1970a,b) were determined as
stem for measurements of thermal conductivity and part of a study of penetration resistance. Cremers and
temperature gradients in the surrounding soil. his colleagues (Cremers, 1972; Cremers and Birkebak,
Each hollow stem consisted of six 0.5-m-long 1971; Cremers and Hsia, 1973; Cremers et al., 1970)
sections for ease of assembly by the astronaut. found only minimum densities for Apollo 11-15
Because of scientific requirements, the stem was samples as part of their investigation of thermal
made entirely of a low-thermal-conductivity material, conductivity. The densities determined by Jaffe
boron-fiberglass; consequently, the sections were (1971a) were for a sample returned inside the scoop of
connected by a press-fit joint, rather than being the Surveyor 3 spacecraft and were part of a study on
screwed together. As a result, the helical auger on the penetration resistance. Carrier et al. (1973a,b) made
exterior of the stem was discontinuous at each joint. measurements of relative density on one Apollo 15
When the Apollo 15 astronauts attempted to drill and two Apollo 14 samples. The densities of a Luna
the hollow stems into the lunar surface, they were 16 sample were determined by Gromov et al. (1972) in
only able to reach a depth of approximately 1.5 m connection with penetrometer, compressibility, and
with each stem. As the drill cuttings rode up the shear strength tests. Later, Leonovich et al. (1974a,
helical auger on the outside of the tube, the soil 1975) also reported similar test results for samples
particles encountered the joints and could travel no from the Luna 16 and 20 missions.
further upward. They had no place to go except into The maximum and minimum void ratios for the
the surrounding soil. Because of the high relative Apollo 11, 14, and 15 samples shown in Table 9.7 are
density of the soil, not all the soil particles in the drill plotted in Fig. 9.18. Also shown for comparison are
cuttings could be pressed into the wall of the boring, the results for uniform spheres and for a basaltic
and the stem became bound to such an extent that simulant with a particle size distribution similar to
the safety clutch in the drill powerhead slipped. (In that of the lunar soils. The minimum void ratio of the
drillers’ terms, the auger flight could not be cleared simulant is 0.45, corresponding to an intergranular
and the bore stem became stuck in the ground.) So porosity of 31%. However, the minimum void ratios of
much heat was generated by the friction between the natural lunar soils are significantly higher. The value
fiberglass stem and the soil that one of the joints is for the Apollo 15 sample is 0.71, or 0.26 greater than
suspected to have collapsed. that of the simulant. This difference is a function of
Prior to the Apollo 15 mission, all the drilling tests the intragranular void ratio, which represents
had been made with lightly compacted soils, and this approximately one-third of the total minimum void
kind of binding failure had never been experienced. ratio of the Apollo 15 sample and corresponds to an
Immediately after the mission, tests were run with intragranular porosity of about 21%.
heavily compacted basaltic lunar soil simulant, and Similar relations are seen in lunar soils from other
the same problems developed that had occurred on missions. The minimum void ratio of the two Apollo
the Moon. The heat flow stems were promptly 14 samples is approximately 0.46 greater than that of
redesigned with titanium inserts so that the joints the simulant. This value represents approximately
could be screwed together to form a continuous auger. one-half of the total minimum void ratio of the Apollo
The inclusion of a high thermal conductivity metal in 14 samples, corresponding to an intragranular
the hollow stem was, of course, a compromise porosity of about 32%. The intragranular porosity of
between science and engineering requirements the Apollo 11 sample appears to be similar to that of
(Crouch, 1971). With the new design, full penetration the Apollo 15 sample, but the slope of the line in Fig.
depth was achieved on the Apollo 16 and 17 missions. 9.18 suggests that the maximum void ratio (minimum
A summary of estimated lunar soil relative density) is in error.
densities is presented in Table 9.6, together with Measurement of the minimum void ratio of a lunar
typical average values. The various estimating and soil can thus be used to estimate its intragranular
measuring techniques are described in the following porosity. The resulting porosity values of 21% to 32%
sections. are large compared to typical terrestrial granular soils,
Laboratory measurements of minimum and and this result is another indication of the
maximum density. Values for minimum and significance of the irregular, reentrant shapes of lunar
maximum density of lunar soils are presented in soil particles. The higher values for Apollo 14
Physical Properties of the Lunar Surface 497
TABLE 9.7. Measured minimum and maximum densities of lunar soils (after Carrier et al, 1973a,b).
498 Lunar Sourcebook
TABLE 9.8. Summary of statistical analyses of relative density based on astronaut bootprint depth
(depth range 0-15 cm) (after Houston et al., 1972; Mitchell et al., 1973a,b, 1974).
Mitchell et al. (1974) compared measurements of compared to the real lunar surface, Aldrin replied that
penetration resistance on the Lunokhod 1 and Apollo the sand was too yielding. While walking on the Moon,
14-16 missions, using correlations with lunar soil he had noticed that although the lunar soil was soft
simulants. From these measurements, they deduced at the surface, there was a firmer stratum at a
relative densities that varied from 63% to 95%, with shallow depth.)
an average value of 83-84%. This large and relatively sudden change in relative
Leonovich et al. (1974a, 1975) analyzed the results density with increasing depth has been attributed to
of approximately 1000 cone penetrometer tests made the effects of continuing small meteoroid impacts,
by Lunokhod 1 and 2. They also presented the results which evidently generate a loose, stirred-up surface
of penetration measurements made on lunar soil but at the same time shake and densify the
returned by the Luna 16 and 20 missions. By underlying soil (Carrier et al., 1973a,b).
comparing these measurements with the estimated Best estimates of relative density vs. depth for the
values of minimum and maximum density in Table intercrater areas have been made by Mitchell et al.
9.7, the average relative density is calculated to be (1974) and Houston et al. (1974) as follows:
70%. (In section 9.1.7, a somewhat lower value of
relative density is deduced, based on a different
analysis of the same data.) This value is representative
of the top 5 cm of the lunar surface, over a combined
traverse distance of 47 km. The frequency distribution
of relative density based on the cone penetrometer
tests on the Lunokhod 1 and 2 missions is presented
in Fig. A9.6.
Best estimates of relative density. The in situ
relative density of lunar soil has been estimated by a
variety of methods, as described in the preceding
sections. Despite differences in the values obtained These values are presented graphically in Fig. 9.20.
from different data and by different methods, the
following trends are apparent: (1) The relative density
of lunar soil tends to be low on the rims of fresh 9.1.6. Compressibility
craters, on slopes, and virtually everywhere within the The preceding sections have discussed the
top few centimeters of the surface. (2) However, in geotechnical index properties of lunar soil, e.g.,
intercrater areas, the relative density is exception-ally density and porosity. The next three sections
high at depths of only 5-10 cm. (Shortly after the describe important engineering properties: com-
Apollo 11 mission, astronaut Edwin Aldrin walked and pressibility, shear strength, and permeability.
ran on a simulated lunar surface at the Johnson Compressibility describes the volume change, or
Space Center in Houston. The test facility consisted of densification, that occurs when a confining stress is
a circular sand track, approximately 15 cm deep, 1 m applied to soil. At low stress or low initial density,
wide, and 30 m in diameter; a special marionette rig compression of the soil results from particle
supported five-sixths of Aldrin’s weight, including his slippage and reorientation. At high stress or high
spacesuit. When asked how the test track initial density, particle deformation and breakage at
Physical Properties of the Lunar Surface 501
Fig. 9.23. Densification of lunar soil caused by self-weight compression in lunar gravity. The plot shows in situ
bulk density (top horizontal axis) and relative density (bottom horizontal axis) as a function of depth below the
surface (vertical axis). Self-compression of material with an initial relative density of 0% (light dashed curve, left)
produces densities at depth that are much lower than those actually observed. Postdepositional vibration of the
material (e.g., by seismic waves associated with meteorite impacts) could produce a material with an initial
relative density of 90% or more (heavy dashed curve, right). The fact that lunar soil samples have relative
densities of 60–100% (Fig. 9.20) indicates that such a vibration/compaction process has taken place. The
steepness of both curves is related to the fact that the effects of self-compaction are limited because of the low
lunar gravity.
it was found that lunar soil is slightly more com- a range of extreme practical interest. Furthermore, the
pressible than the simulant, regardless of whether the stress produced by self-weight is not sufficient to
two soils are compared at the same void ratio or at the squeeze the soil to such high relative densities at
same relative density. Evidently, the irregular fragile shallow depth. At some time after deposition, the in
particles in lunar soil, such as agglutinates, which situ lunar soil has either been exposed to higher
have no equivalent in the basaltic simulant, crush stresses, or compacted by vibration, or a combination
under relatively low confining stress. Thus, the of the two processes.
intragranular and subgranular porosities also If lunar soil were sifted gently into place, with an
influence the compressibility of lunar soil. initial relative density of 0%, the density profile
The compressibility curves for soils of different developed by self-weight would be that shown in Fig.
initial relative densities (Fig. 9.22) do not converge 9.23. This profile is obtained by integrating the
until the stress exceeds 100 kPa, which corresponds, compressibility curve for DRi = 0% in Fig. 9.22. Even
on the Moon, to a depth of more than 30 m below the though the compression index of loose soil is much
lunar surface. Thus, these oedometer data cover higher than that of dense soil, the densification as a
function of depth is small because of the low lunar
504 Lunar Sourcebook
gravity. At a depth of 50 cm, the calculated bulk varying from 0 to 0.013 in lunar soils, with a typical
density produced by self-weight would be 1.30 g/ cm3 value of 0.003. The data are presented in Fig. 9.24
and the relative density would be 32%, both of which and tabulated in Table A9.10.
are significantly less than the measured in situ values Maximum past pressure. If a soil has been exposed
shown in Figs. 9.16 and 9.19. In fact, the actual in to a higher stress than is presently imposed on it by
situ soil would had to have been compressed to a self-weight, this higher stress is referred to as the
pressure greater than 100 kPa, or more than 100 maximum past pressure, and the soil is called over-
times its present stress, to account for the measured consolidated. In such a case, the soil’s currently-
density. On the other hand, if the soil had been measured compressibility would correspond to the
vibrated after being deposited (i.e., by meteoroid recompression index. In contrast, if the maximum
impacts), the relative density might be 90% or more, past pressure is equal to the present stress, the soil is
and the density profile would be virtually uniform, as called normally consolidated, and its compressibility
shown in Fig. 9.23. corresponds to the compression index.
To explain the extreme changes in relative density As mentioned above, it is still not clear whether the
that occur in the top 30 cm of lunar soil, there must lunar soil is normally consolidated or over-
be a mechanism that both stirs up the surface and consolidated. From a practical point of view, this
densifies the immediately underlying soil. Both of uncertainty does not matter; the discussion in section
these effects are presumably caused by repeated 9.1.9 shows that the predicted settlement of a typical
meteoroid impacts. However, we do not know yet structural footing in lunar soil would be small in
whether such impacts have a greater tendency to either case.
vibrate the soil or to stress it; the observed increases However, future experiments in the laboratory or on
in relative density with depth could reflect either or the lunar surface could resolve the issue. In their
both processes. vacuum oedometer tests of lunar sample 12001,1 19,
Recompression index. When the applied stress is Carrier et al. (1972b, 1973c) observed gas pressure
decreased after compression in an oedometer test, the bursts whenever the stress exceeded a previous
soil sample swells slightly and rebounds elastically. maximum value (Fig. 9.25). The released gas consisted
When the same stress is applied again, the soil returns of 60% hydrogen, 35% helium, and 4.6% amu = 28,
to approximately the same void ratio as in the previous consistent with the composition of the solar wind. The
compression with a small hysteresis. The slope of the gas bursts were attributed to the breakage of fragile
rebound-reload curve is called the recompression particles such as agglutinates (section 7.1.3), which
index, Cr, and it is always much less than the contain surface-correlated solar-wind gases. Thus,
compression index. Carrier et al. (1972b, 1973c) and vacuum oedometer tests on undisturbed returned core
Jaffe (1973) measured values of Cr tube samples could be
Fig. 9.25. Experimental data showing bursts of released gas pressure from lunar soil sample 12001,119 during compression (vacuum
oedometer) and shear experiments (after Carrier et al., 1973c). Data are plotted as functions of elapsed time (horizontal axis) for applied
vertical stress (top), N2 equivalent in the released gas (center), and partial gas pressures for different observed masses (bottom). The peaks of
gas release are closely correlated with periods of increasing stress, indicating that the gas is being released from individual soil particles as
they are fractured and crushed. Note that gas pressures read from these graphs must be divided by the factors shown (e.g., the highest value
shown for total gas pressure from the direct shear test is 16 × 10-8 torr).
506 Lunar Sourcebook
TABLE 9.11. Estimates of lunar soil cohesion and friction angle (after Mitchell et al., 1972d, 1974).
508 Lunar Sourcebook
no moving parts. The depth of penetration was able to insert a “rammer-jammer” tool down the open
estimated by the astronaut calling out the exposed hole to a depth of more than 2 m. Analysis indicated
stripes on the rod, and the required force was the cohesion would have to be 1.3 kPa for a friction
estimated on the basis of measurements made with a angle of 46.5° to keep such an unsupported hole open
spacesuited subject flying in an aircraft under 1/6-g (Mitchell et al., 1972b).
conditions. Although approximate, these tests again Apollo 17. At the Apollo 17 site, a thin, cylindrical
confirmed the Surveyor Model (Mitchell et al., 1971). Neutron Flux Probe was inserted into the open drill
Apollo 15 and 16. A more sophisticated Self- core hole, which was even deeper (3 m) than the
Recording Penetrometer (SRP) was operated by the similar hole at the Apollo 16 site. In this case, it was
astronauts on the Apollo 15 and 16 missions. A total found that the cohesion would have to be 1.1 to 1.8
of 17 tests were performed to depths of 74 cm. Details kPa for a wide range of friction angles (Mitchell et al.,
are given immediately below. 1973a).
In addition, when the deep rotary drill core was Cone penetrometer measurements. Cone
successfully removed from the lunar surface during penetrometer tests were made by astronauts on the
the Apollo 16 mission, Astronaut Charles Duke was Apollo 14, 15, and 16 missions and from the
Physical Properties of the Lunar Surface 509
Fig. 9.26. Calculated in situ shear strengths of lunar soil (vertical axis), plotted as a function of normal stress
(bottom horizontal axis) and corresponding depth below the lunar surface (top horizontal axis) (after Mitchell et
al., 1972d, 1974). Envelopes of shear strength are shown for Surveyor model data (oblique solid ruling), for
Apollo and Lunokhod results (oblique dashed ruling), and for a basaltic simulant at relative densities of 30%
and 90% (dashed lines).
Fig. 9.28. Cone penetrometer resistance data measured for in situ lunar surface material, plotted as
penetration depth (vertical axis) vs. penetration resistance (horizontal axis). Data were obtained from the
Lunokhod 1 and 2 automated rovers and by astronauts on several Apollo missions. Inset (upper right) shows
diagrammatically the shape of the cone penetrometer and the forces involved (after Mitchell et al., 1973b, 1974).
Lunokhod 1 and 2 robotic roving vehicles (Fig. 9.28). Typical cone penetrometer test results are shown in
These measurements are the most important source Fig. 9.29 (Leonovich et al., 1971, 1972).
of in situ shear strength data for the lunar soil. The base of the Lunokhod cone was level with the
The Lunokhod roving vehicle carried a combination soil surface when the tip depth was 4.4 cm, as shown
cone penetrometer and vane shear device. The two by the horizontal dashed line in Fig. 9.29. Using
perpendicular blades of the vane were 7 cm long and bearing capacity theory, Mitchell et al. (1972d)
4.4 cm high. The cone had an apex angle of 60° with a analyzed this case for curves 1, 2, and 3, with the
base diameter of 5 cm and a height of 4.4 cm. results shown in Fig. 9.30. This bearing capacity
Physical Properties of the Lunar Surface 511
theory was specifically developed for evaluating lunar the latter three unknowns were assumed by Mitchell
cone penetrometer tests and is expressed as follows et al. (1972d); cohesion was then calculated for
(Mitchell and Durgunoglu, 1973; Durgunoglu and selected values of friction angle (Fig. 9.30). The
Mitchell, 1975) calculated cohesion decreases as the friction angle
increases: For example, on horizontal ground (curve
1), when φ= 25°, c = 1.8 kPa; and when φ = 45°, c =
0.34 kPa. The complete range of values is listed in
where q = cone penetration resistance; c = cohesion; Table 9.11.
ρ= density of lunar soil; gm = acceleration of lunar The measured shear strength of the basaltic
gravity; B = penetrometer base width or diameter; simulant (Fig. 9.27) is also plotted on Fig. 9.30 as the
Nc, Νγq = bearing capacity factors (functions of dashed curve, in which c and 4) both increase as the
friction angle, cone roughness, and coefficient of relative density increases. Where the measured
lateral stress); and ξc, ξγq = shape factors (functions of (dashed) curve crosses the theoretical curves (solid), a
width, length, and depth). unique solution can be obtained: For example, on
There are five unknowns in this one equation: horizontal ground, φ = 42° and c = 0.5 kPa. Further-
cohesion, friction angle, soil density, cone roughness, more, the corresponding relative density (DR) is about
and coefficient of lateral stress. Reasonable values of 65%.
512 Lunar Sourcebook
In later papers, Leonovich et al. (1974a, 1975) the base diameters varied from 1.28 cm to 2.86 cm.
presented statistical results for approximately 1000 The plate was 2.54 cm wide and 12.7 cm long. The
cone penetrometer tests from Lunokhod 1 and 2, all astronaut could attach the plate or one of the cones to
at the same tip depth of 4.4 cm. A similar bearing the shaft as needed to perform a test. The record of
capacity analysis to that above, performed on their each penetration test was scribed on a gold-plated,
data, yields the following average values: φ = 40°, c = stainless-steel cylinder contained in the upper
0.40 kPa, and DR = 59%. Frequency distributions for housing assembly (Fig. 9.32). After the test, this
these data are presented in Fig. A9.7. This relative assembly was disconnected from the rest of the SRP
density is somewhat less than the value calculated in and returned to Earth, while the cones and plate were
section 9.1.5 and Fig. A9.6 using the same Lunokhod discarded. The recording cylinder from the Apollo 16
data. SRP is shown in Fig. 9.33.
The Apollo Self-Recording Penetrometer (SRP) (Fig. A total of 17 cone and plate tests were performed on
9.31) was equipped with three cones and one flat the Apollo 15 and 16 missions (Mitchell et al., 1974;
plate. Each cone had an apex angle of 30°, and Mitchell and Houston, 1974). With the smallest
Physical Properties of the Lunar Surface 513
Fig. 9.31. Photograph and explanatory diagram of the Self-Recording Penetrometer (SRP) used on the Apollo 15
and 16 missions. When used, the SRP is held vertically, with the pad (left) placed flat on the soil surface. The
astronaut holds the upper housing assembly (right) and pushes the cone into the surface. The depth of cone
penetration is recorded by a scriber on a metal cylinder in the upper housing assembly (right). Three cones were
available for testing, with diameters ranging from 1.28 to 2.86 cm (only the 1.28 and 2.03 were actually used,
plus the 2.54 × 12.7-cm plate).
Fig. 9.32. Close-up view of upper housing assembly of Apollo Self-Recording Penetrometer (SRP) (see Fig. 9.31).
Scale bar below housing is in inches. (NASA Photo S-71-39600.)
514 Lunar Sourcebook
cone, the astronaut was able to push the SRP to a that the soil below 20 cm had a high shear strength
maximum depth of 74 cm. The SRP test results from and high relative density. Consequently, we have
the Apollo 16 site are summarized in Fig. 9.34. Station derived recommended “typical” values of shear
4 was located on the slope of Stone Mountain, and the strength parameters by combining the data in Figs.
penetration resistance at that station indicates that 9.20 and 9.27. These values are presented in Fig. 9.35
the soil is softer there than at Station 10, which was and Table 9.12. For example, the average friction
located in an intercrater area near the Apollo Lunar angle in the top 60 cm is 49°, and the average
Surface Experimental Package (ALSEP). Detailed SRP cohesion is 1.6 kPa.
penetration curves obtained at the Apollo 15 and 16 Laboratory measurements of shear strength. The
sites are presented in Figs. A9.8 and A9.9, first laboratory measurements of lunar soil shear
respectively. strength were made in 1969 in the Lunar Receiving
The SRP results were analyzed in a manner similar Laboratory at the NASA Manned Spacecraft Center
to the treatment of the Lunokhod cone data described (later renamed the Johnson Space Center) as part of
above. Only eight SRP tests penetrated below 20 cm; the preliminary examination of Apollo 11 returned
of these, the lowest measured values of friction angle lunar samples. The tests consisted of pushing a
and cohesion are 41° and 0.45 kPa, respectively, standard hand-penetrometer (flat base) into several
found at Station 4. These values remain essentially hundred grams of compacted soil. The soil had been
constant to a depth of more than 60 cm at the same sieved to remove particles coarser than 1 mm and was
station. kept in a dry nitrogen atmosphere. The minimum
In an intercrater region (Station 10, Apollo 16 site), density of the soil was 1.36 g/cm3 and the maximum
at a depth of 20 cm, the average friction angle and density was 1.80 g/cm3 (see Table 9.7). The results of
cohesion are 46° and 0.95 kPa, respectively, based on these penetration tests were reported in Costes et al.
only four tests. These values correspond to an average (1969, 1970a,b) and Costes and Mitchell (1970). The
relative density of 75% for the upper 20 cm, which is test data are tabulated in Table A9.1 1.
consistent with the data shown in Fig. 9.20. To reach Reanalysis of the data for the bearing test performed
greater depths with the probe, the astronaut generally on this Apollo 11 soil at the maximum density,
had to apply a force that exceeded the capacity of the utilizing the bearing capacity theory de-
recording mechanism, indicating
Physical Properties of the Lunar Surface 515
Fig. 9.35. Calculated average (“recommended”) values of shear strength parameters (friction angle and
cohesion) (horizontal axes) in the lunar surface material as a function of depth below the surface (vertical axis)
(see also Table 9.12). Both parameters, particularly the cohesion, increase with depth.
516 Lunar Sourcebook
scribed above (Durgunoglu and Mitchell, 1975) show confining stresses applied to the samples were one to
that for a friction angle in the range from 42° to 38°, two orders of magnitude greater than the in situ
the cohesion would be 0.25 to 0.85 kPa. lunar stresses.
Jaffe (1971a) performed similar penetrometer The last point appears to be especially significant.
experiments on a 6.5-g soil sample contained in the Under the low stresses present near the lunar
Surveyor 3 scoop returned in October 1969 by the surface, irregular and reentrant soil particles tend to
Apollo 12 astronauts. The test data are tabulated in interlock, producing an usually high shear strength.
Table A9.12. Even when the Lunokhod cone penetrometer was
Carrier et al. (1972b, 1973c) performed three direct pushed into the soil with a tip pressure of 50 kPa,
shear tests in vacuum on a 200-g sample of Apollo 12 the average stress on the failure surface was only
soil. For a range of cohesion from 0-0.7 kPa, the about 5 kPa. For the Apollo SRP, the average applied
measured friction angle was 28°-35°. These values stress was even less because of the greater depth of
were significantly less than the shear strength of a penetration by the probe. On the other hand, when a
basaltic simulant. This difference was attributed to returned soil sample is tested at laboratory stresses
the crushing of fragile particles such as agglutinates (30-70 kPa in the case of the vacuum direct shear
and breccias, which are unique to the lunar soil. A tests), the fragile particles tend to crush, resulting in
similar difference was observed during measurements a lower measured shear strength.
of compressibility (section 9.1.6). Gas-pressure As a result, Carrier et al. (1972b, 1973c), Mitchell et
bursts, releasing a gas of solar-wind composition, also al. (1974), and Leonovich et al. (1974a) have all
occurred during shear tests, as they had during suggested that a curved Mohr-Coulomb equation
compression tests (see Fig. 9.25). Complete test data would more appropriately describe the shear strength
are tabulated in Table A9.13. of lunar soil.
Jaffe (1973) also performed miniature direct shear Instead of the normal Mohr-Coulomb equation, τ =
tests on 1.3 g of the soil returned with the Surveyor 3 c+σ tan φ, Carrier (1984) has proposed that a better
scoop. The tests were performed in air, using soil shear strength equation might be of the form
samples with five initial bulk densities, ranging from
0.99 to 1.87 g/cm3. The measured values of cohesion τ = aσb
varied from 0.1 to 3.1 kPa and the friction angle from
13° to 56°. Jaffe concluded that the test results were The apparent friction angle is then given by
consistent with an in situ bulk density of about 1.7 g/
cm3 (Table 9.4). He also concluded that exposure to φ = tan–1 (abσb–1)
air did not significantly alter the shear strength
and the apparent cohesion is given by
properties of lunar soil. The test results are tabulated
in Table A9.14; most of the reported values of friction
c = a (1–b)σb
angle are significantly lower than those determined in
other experiments.
Scott (1987, and personal communication, 1988) Carrier has proposed preliminary values of a = 1.83
performed miniature triaxial shear tests on a 1.1-g and b= 0.73, when the stresses are expressed in
sample of the soil returned with the Surveyor 3 scoop. kilopascals. For example, at σ = 5 kPa, the shear
The test results are tabulated in Table A9.15. strength, τ, would be 5.9 kPa, φ would be 41°, and c
Leonovich et al. (1974a, 1975) performed direct would be 1.60 kPa, all of which are consistent with
shear and coulomb device tests on Luna 16 and 20 the data in Fig. 9.26. In contrast, at σ = 100 kPa, τ
samples and reported a cohesion of 3.9 to 5.9 kPa and
would be 53 kPa, φ would be 21°, and c would be 14
a friction angle of 20°-25°. Additional laboratory
kPa, which are close to the laboratory measurements
measurements of shear strength have been made by
of Leonovich et al. (1974a, 1975).
Soviet scientists, and they are reported (in Russian) in
For the present, the recommended shear strength
Leonovich et al. (1974b), Vedenin et al. (1974),
parameters shown in Fig. 9.35 and Table 9.12 are
Douchowskoy et al. (1974, 1979), and Gromov et al.
sufficiently accurate for the design of near-surface
(1979).
lunar structures and construction activities.
Lunar soil shear strength: Some speculation. All
However, at later stages in the human development
the laboratory tests on returned lunar soil samples
of the Moon, heavy structures or deep excavations
suffer from the following limitations: (1) disturbance:
will likely be constructed. At that point, a more
the samples were sieved, remolded, and recompacted
complete understanding of the shear strength of
prior to testing; (2) size: the samples were small (to
lunar soil under higher stress conditions will be
very small) by terrestrial testing standards; and (3)
required.
stress: unavoidably, the
Regolith seismic properties. P-wave velocities (α)
measured in regolith at the Apollo landing sites are
very low and uniform and range from 92 to
Physical Properties of the Lunar Surface 517
114 m/sec, with an average of 105 m/sec (Watkins in the soil was established at about 0.6, and the tests
and Kovach, 1973). The low velocities are interpreted were run at room temperature under very low vacuum
as being related to brecciation and high porosity. The conditions, such that the mean free path length of the
seismic Q (αQ/103; loss of energy by absorption) of molecules was much greater than the pore
lunar regolith is ~200 at a velocity of 105 m/sec; for dimensions. The measured results for gas diffusivity
comparison, terrestrial analogues such as were 7.7, 2.3, and 1.8 cm2/sec for He, Ar, and Kr,
unconsolidated volcanic ash have seismic Qs of 1.5- respectively.
3.0 at an α = 400 m/sec (Watkins and Kovach 1973). For these relatively nonreactive gases, the diffusiv-
ity is proportional to (T/M)1/2, where T is the absolute
9.1.8 Permeability and Diffusivity temperature and M is the molecular weight of the gas.
The “sticking time” of the gas molecules, or heat of
The coefficients of permeability and diffusivity are used absorption, becomes significant if the gas is more
to evaluate the movement of a fluid (gas or liquid) reactive or the temperature is reduced.
through a porous solid medium. No direct
measurements of these parameters have been made on 9.1.9. Bearing Capacity
lunar soil. Approximate values are presented in the
following sections. The remaining three subsections of 9.1 discuss im-
Permeability. The coefficient of permeability defines portant practical consequences of the geotechnical
the quantity of flow of a fluid through a porous index and engineering properties of the lunar regolith:
medium in response to a pressure gradient. bearing capacity, slope stability, and trafficability.
In the simplest case of steady-state flow of an Bearing capacity describes the ability of a soil to
incompressible, nonreactive liquid, Darcy’s Law states support an applied load, such as a vehicle, a
that structure, or even an astronaut. Usually, the topic of
bearing capacity is divided into two categories:
ultimate bearing capacity and allowable bearing
capacity; each of these are then subdivided further
into static and dynamic quantities.
where Q = flow rate (m3/sec); K = absolute Ultimate bearing capacity defines the maximum
permeability (m2); ρ= density of the liquid (kg/m3); gm possible load that can be applied without causing
gross failure, such as the overturning of a structure.
= acceleration due to lunar gravity (m/sec2); µ =
For a 1-m footing on the lunar surface, the ultimate
viscosity of liquid (N-sec/m2); i = hydraulic gradient
bearing capacity is approximately 6000 kPa.
(m/m); and A = cross-sectional area normal to the
Allowable bearing capacity defines a lesser load
flow (m2). For a given arrangement of particles, the
that can be applied without exceeding a given amount
coefficient of absolute (or specific) permeability is
of settlement. A settlement limit is usually imposed
controlled by the size and shape of the pore spaces,
either for structural or operational requirements. For
and it is independent of the properties of the liquid.
example, if the settlement of an experiment package
The transient flow of rocket-engine exhaust into the
placed on the lunar surface must be less than 1 cm,
lunar surface material was studied extensively in
with a confidence level of 95%, and the width of the
order to estimate the amount of erosion that could
package is less than 0.5 m, then its allowable bearing
occur when the accumulated gas pressure in the
pressure is just 2 kPa.
pores exceeded the weight of the overlying soil under
These topics are discussed in more detail in the
different operating conditions (cf. Scott and Ko, 1968).
following sections.
Based on a test firing of the Surveyor 5 vernier engine
Ultimate bearing capacity: Static. The static
on the lunar surface, the permeability of the soil to a
ultimate bearing capacity, quit, can be estimated on
depth of about 25 cm was deduced to be 1-7 × 10-12
the basis of plasticity theory and is given by
m2 (Choate et al., 1968). This value was considered to
(Durgunoglu and Mitchell, 1975)
be consistent with the fine-grained nature of the soil.
Diffusivity. The gas diffusivity defines the molec-
ular flow of gas through a porous medium in response
to a concentration gradient. The diffusivity depends
on the gas composition, the pressure and where ρ = density of soil; gm = 1.62 m/sec2 (acceler-
temperature, and the particle size and shape ation of gravity on the Moon); B = footing width; c =
distributions of the soil. Martin et al. (1973) measured cohesion of soil; Nc, Νγq = bearing capacity factors,
the diffusivity of noble gases—He, Ar, and Kr— which are primarily dependent on the friction angle,
through a basaltic lunar soil simulant. The void ratio φ, of the soil; and ξc, ξγq = shape factors (see section
9.1.7 for additional details).
518 Lunar Sourcebook
Fig. 9.36. Ultimate bearing capacity data for lunar surface material. Data are plotted in terms of the applied
stress (horizontal axis) produced by an object on the lunar surface and the footing width (essentially the
footpad size) of the object. Data (filled squares) are given for an astronaut standing on one boot, for an Apollo
Lunar Roving Vehicle (LRV) wheel, and for the footpads of the robotic Surveyor spacecraft and the manned
Lunar Module. The envelope of ultimate bearing capacity (right side, vertical ruling) lies at much higher applied
stresses, indicating that the lunar surface can support virtually any type of conceivable spacecraft or
structure.
The ultimate bearing capacity is therefore con- core tube into the lunar surface material to almost
trolled by the soil density, its shear strength, and the any depth. In fact, the practical limit of the Apollo
size of the footing. Using the in situ bulk density 15-17 thin-walled core tubes (see Figs. 9.6 and 9.10)
estimates in section 9.1.4 and the in situ shear was only about 70 cm, and it typically required about
strength estimates in section 9.1.7, the ultimate 50 hammer blows to reach this depth. An analysis by
bearing capacity vs. footing width can be calculated Carrier et al. (1971) showed that if energy losses were
(Fig. 9.36). As an example, during the Apollo 11 neglected, the number of hammer blows required to
landing, the Lunar Module (LM) footpad was nearly 1 reach a given depth would increase with the square
m in diameter, and the ultimate bearing capacity of of the depth; if energy losses were included, then
the lunar soil was approximately 3000 to 11,000 kPa. there would be a depth beyond which no amount of
Because the stress actually applied by the footpad hammering would drive a rod or core tube.
was only about 5 kPa, the factor of safety was 600 to The same analysis (Carrier et al., 1971) concluded
2200. Furthermore, for larger footings, the ultimate that it would be impossible for the Apollo 17
bearing capacity is roughly proportional to the width. astronauts to drive the long, cylindrical Lunar
That means that the ultimate load (stress × area) for a Neutron Probe Experiment (Woolum et al., 1973) into
circular or square footing is proportional to the cube the soil, as was originally planned. The probe had a
of its width. Consequently, the ultimate bearing diameter of about 2 cm, and it had to be inserted to a
capacity of the lunar surface is more than sufficient to depth of 1-2 m. As a result of this analysis, it was
support virtually any conceivable structure. decided that the Apollo 17 astronauts would attempt
Ultimate bearing capacity: Dynamic. The to drop the probe down the hole opened up by the
dynamic ultimate bearing capacity defines the drill to collect the deep core sample. Earlier studies
maximum resistance to impact loading. This dynamic (Mitchell et al., 1972a) had predicted that the hole
capacity is always greater than the static capacity would remain open after the drill stem was
because of the inertial resistance of the soil. withdrawn, and this effect had been demonstrated by
Even after the experience from several manned the Apollo 16 astronauts (Mitchell et al., 1972b).
lunar landings, many scientists and engineers still When this operation was attempted during the Apollo
believed that an astronaut could hammer a rod or 17 mission, the probe fell into the drill core
Physical Properties of the Lunar Surface 519
hole without difficulty. The success of the controls the design of a foundation. If a 95%
predictions served to confirm the shear strength confidence level is required for acceptable settlement
parameters of the lunar soil (section 9.1.7 and depth, then Fig. 9.37 indicates that a modulus value
Tables 9.11-9.12). of k = 2 kPa/cm should be used
Allowable bearing capacity: Static. The
allowable bearing capacity is controlled by the
compressibility of the soil and the acceptable
amount of settlement for a given structure. For a The modulus of subgrade reaction (k) presented in Fig.
load applied directly on the lunar surface, and a 9.37 is sufficiently accurate to design many structures
footing width of less than about 0.5 m, our best and experiment packages for emplacement on the lunar
source of information about the behavior of the surface. However, for highly settlement-sensitive
lunar soil is the astronaut bootprint data discussed strucures, such as telescopes, the lunar surface
in section 9.1.5 and Fig. A9.4. With a known load compressibility from point to point is too variable for a
and depth of bootprint, the modulus of subgrade general approach. Founding such a structure directly
reaction, k, can be calculated statistically, as shown on the surface could lead to unacceptably high
in Fig. 9.37. Then the allowable bearing capacity, differential settlements. These structures will more
qall, is given by than likely have to be supported by spread footings set
at a depth of at least 30 cm, and perhaps 1 m beneath
the actual surface, in order to be below the depth of
where dacc = the amount of acceptable settlement. diurnal temperature fluctuations. The settlement d of
For example, the average value of k = 8 kPa/cm such a footing is calculated on the basis of elasticity
for a range of lunar locations, including the rims of theory and is given by
fresh craters. If the acceptable settlement is 1 cm,
then
Fig. 9.37. Data for calculating the maximum allowable bearing capacity for lunar surface material. The plot
shows a histogram of values of the modulus of subgrade reaction, determined from measurements of 776
astronaut bootprints made on all Apollo missions (data after Mitchell et al., 1974). Inset (upper left) shows a
schematic cross-section of a footprint and the forces involved. The maximum allowable bearing capacity for any
required minimum subsidence can be calculated from the determined values of the modulus (see text).
520 Lunar Sourcebook
where C j = compressibility of soil layer j (either density of lunar soil has been found to be
compression index, C c , or recompression uniformly very high below a depth of 30 cm,
index, C r ); e j = initial void ratio of soil layer; the differential settlement between footings
∆H j = thickness of soil layer; σ ij = initial should be very small.
vertical stress in soil layer (caused by If the lunar soil is in fact over-consolidated (a
overburden stress); and ∆σ j = increment of question that remains to be answered; see sec-
vertical stress in soil layer (related to tion 9.1.6), the calculations of settlement for
foundation load; estimated from elasticity the same footings would be based instead on
theory). the recompression index, C r , which is
A given soil profile can be subdivided into as approximately 0.003 (Table 9.9). In this case,
many layers, n, as desired. The settlement of where the lunar soil is over-consolidated, the
each layer is calculated, and the results are settlements would be only 1/17 of those shown
then summed to yield the total settlement of in Fig. 9.38; such settlements are practically
the foundation, d. negligible.
Typical settlement estimates for isolated Allowable bearing capacity: Dynamic. In
spread footings are shown in Fig. 9.38. The the case of a slowly cyclical load, such as a
calculations assume that the footings are rotating telescope, it is important to avoid
square and are placed at a depth of 1 m. The differential settlements that produce rocking
initial void ratio and vertical stress motions. The usual terrestrial practice is to
distributions are based on the bulk density require that the dynamic component of load be
profile described in section 9.1.4. It is also no greater than one-fourth of the static load.
assumed that the soil is normally consolidated In addition, if the structure contains rotating
and has a compression index, Cc, of 0.05 (Table machinery, it is very important that
9.9). As shown in Fig. 9.38, the calculated resonances be avoided. Because of the low
settlement varies from a low of 0.7 cm for an lunar gravity, the resonant frequency of a soil-
applied stress of 10 kPa on a 1-m footing to supported structure will be less than that for a
11.6 cm for an applied stress of 50 kPa on a 5- similar structure on Earth. Careful design will
m footing. The footing sizes and applied be required in order to eliminate annoying, or
stresses shown in Fig. 9.38 cover a wide range even detrimental, vibrations and settlements
of possible conditions, and the calculated (cf. Richart et al., 1970).
settlement values can be easily accommodated
in the design of the structure. Because the
relative
Physical Properties of the Lunar Surface 521
9.1.10. Slope Stability depth (i.e., the depth up to which no slope failure will
occur) of an excavation in an intercrater area can be
In the past 100 years, literally dozens of numerical calculated by combining values for the in situ density
methods have been developed to evaluate the stability (section 9.1.4) with the in situ shear strength (section
of a soil slope, i.e., its ability to stand without 9.1.7). Using a factor of safety of 1.5, which is more
support. The most common methods are based on than adequate for design purposes, the calculations
limit equilibrium analysis of circular potential slip show that a vertical cut could be made in lunar soil to
surfaces. This approach is described in detail in a depth of about 3 m (Fig. 9.39), and a slope of 60°
numerous references (cf. Taylor, 1937, 1948; Bishop, could be maintained to a depth of about 10 m.
1955; Lambe and Whitman, 1969). In order to construct an embankment, the soil must
On the Moon, the absence of water greatly first be excavated, then transported, placed, spread,
simplifies the analysis of slope stability. The factor of and compacted. As discussed above (section 9.1.5),
safety, F.S., against slope failure can be reduced to the in situ lunar soil is very dense, with a greater
the following expression density than could be produced with mechanical
compaction equipment. The processes of handling and
manipulating the lunar soil would loosen it
considerably, and it would not then be possible to
where π = density of soil; c = cohesion of soil; gm = compact the soil back to its original, undisturbed
acceleration of gravity on the Moon; h = height of density. As a result, the density of the soil in a
slope; and N = stability number, which is a function of compacted embankment would be lower than its
the friction angle, φ, of the soil and the slope angle, β. original density, and the maximum possible slope
Stability of constructed slopes. A constructed angle of the embankment would be less than that of
slope could be either an excavation, a compacted an excavation in undisturbed soil. Assuming a
embankment, or a dumped pile (Fig. 9.39). The safe compacted relative density of 65–75%, a 10-m-
high slope could be constructed at an angle of about layer to move with almost no frictional resistance.
45°, or one horizontal increment to one vertical Finally, Florensky et al. (1978) analyzed the X-ray
increment (Fig. 9.39). fluorescence spectroscopic data obtained from the
If lunar soil were simply dumped in a pile, it would Lunokhod 2 roving vehicle. They concluded that
attain a relative density of about 30–40% and the lateral movement of soil was occurring on slopes
flatter than the angle of repose. While some of this
factor of safety (i.e., the “angle of repose”) would be
movement could be attributed to meteoroid impacts,
1.0. The pile could be raised to a height of 10 m at an they proposed that a portion was caused by some
angle of nearly 40°. kind of soil creep of “problematic origin.”
Stability of natural slopes. We really do not know Clearly, many questions regarding the stability of
very much about the stability of natural lunar slopes. natural lunar slopes remain unresolved.
In the early days, prior to any actual lunar landings,
9.1.11. Trafficability
some scientists thought that the hills and mountains
must be stronger than the mare basins, simply Trafficability is defined as the capacity of a soil to
because the hills had slopes and the maria were support a vehicle and to provide sufficient traction
relatively flat. However, the very limited cone for movement. Before actual landings on the Moon,
penetrometer data obtained by both human and it was believed that trafficability on the lunar surface
robotic missions have established that the soil on the would be poor and that the movement of vehicles
slopes is actually somewhat weaker than the soil in would be very difficult. Consequently, all sorts of
the flatter intercrater areas, at least to a depth of 70 strange-looking walking machines were developed.
cm. From the experience of the Apollo and Lunokhod
There is also considerable evidence, chiefly from missions, we now know that almost any vehicle with
orbital photography of the Moon, that slope failures round wheels will perform satisfactorily on the lunar
have occurred on the lunar surface. In some cases, surface, provided the ground contact pressure is no
the resulting talus or landslide material has covered greater than about 7–10 kPa.
large areas and has traveled many kilometers. On As a wheeled vehicle travels on a level lunar
Earth, slope failures are usually caused by fluctua- surface, the energy consumed in overcoming the
tions in the groundwater table, by erosion from combined effect of surface roughness and rolling
running water, and occasionally by tectonic activity. resistance (soil compaction) is roughly equivalent to
On the Moon, the triggering mechanism is presumed the energy expended in continuously climbing a
to be the seismic vibrations produced by meteoroid smooth, rigid, 1½° slope. For the Apollo Lunar
impact, but explanations of how the talus has been Roving Vehicle (LRV), which carried two astronauts,
able to move such long distances are very speculative. the combined surface roughness and rolling resist-
Houston et al. (1973) attacked the question of lunar ance accounted for only 15% of the total energy
slope stability by analyzing the stability of lunar expended. The remainder was lost to mechanical
slopes under seismic conditions. They calculated that linkages, inertia, instrumentation, etc.
the cumulative downslope movement produced by The primary limitations on the trafficability of the
more than 3.5 b.y. of meteoroid impacts was lunar soil are speed and slope climbing. The normal
insignificant except on very steep slopes. They cruise speed of the LRV was 6–7 km/hr. This speed
concluded that meteoroid impact could not account was constrained by the irregularity of the cratered
for the observed lunar slope degradation or “erosion.” surface, coupled with the low lunar gravity. The
On the other hand, Carrier et al. (1973c) suggested speed of a future lunar vehicle can be increased only
that outgassing during the shearing of lunar soil (see by modifying its dynamic response (e.g., by using
Fig. 9.25) could produce a fluidized flow condition, larger diameter wheels, an increased wheel base,
thereby resulting in large run-out distances for the greater mass, or a softer suspension system), and/or
transported material. Scott (1987) suggested an by constructing permanent roads.
alternate mechanism based on numerical analyses by The steepest slope that the LRV could climb was
Campbell and Brennen (1985). He proposed that an 19°–23°. This value could be improved by adding
intact soil layer could move at high velocity, variable-length grousers (cleats) to the wheels.
suspended by the extremely dynamic motion of a few Additional details on trafficability are provided in the
particles bouncing between the bottom of the soil following sections.
layer and the underlying stable stratum. The Lunar vehicles. To date, three types of wheeled
bouncing soil particles would provide an equivalent vehicles have been operated on the lunar surface:
pressure that would support the overlying layer in a the U.S.S.R. Lunokhod, the Apollo 14 Modular
pseudofluidized condition, even in the absence of an Equipment Transporter (MET), and the Apollo Lunar
atmosphere, thus allowing the Roving Vehicle (LRV) (Figs. 9.40–9.42, respectively).
The
Physical Properties of the Lunar Surface 523
unmanned Lunokhod had a mass of 756 kg and rode Apollo 15. Figure 9.43 shows this two-wheeled,
on eight rigid, spoked wheels (Petrov, 1972; Alexan- pneumatic-tired vehicle undergoing trafficability
drov et al., 1972). The rim of each wheel was testing.
constructed of metal mesh, reinforced with three Performance characteristics of the LRV. The
hoops connected by transverse cleats. The diameter performance characteristics of the LRV are summar-
of the wheel was about 51 cm and the width about ized in Table 9.13; vehicle dimensions are presented
20 cm. The MET was an unpowered two-wheeled, in Fig. 9.44. The payload mass of the LRV was 490
rickshaw-type vehicle used by the Apollo 14 astro- kg, more than twice the mass of the empty vehicle;
nauts to carry tools and samples on their geology the bulk of the payload was the two astronauts
traverses. Its loaded mass was about 75 kg. The themselves. Each of the wheels was individually
wheels were pneumatic tires, each with a diameter of powered by an electric motor coupled to a harmonic-
about 40 cm and a width of about 10 cm. The LRV, drive gear reduction unit. In case of a motor or gear
or Rover, was a four-wheeled, battery-powered failure, the wheel involved could be disengaged and
vehicle driven by the astronauts on the Apollo 15, allowed to roll free. Wheel-slip on the lunar surface
16, and 17 missions. Its gross mass was 708 kg, was measured to be between only 2–3%, which
more than half of which consisted of the two allowed for reasonably accurate navigation by dead-
astronauts and their life support systems. The reckoning. The maximum speed on a smooth, level
wheels were flexible and consisted of woven piano surface was about 13 km/hr, but because it was
wire, with chevron-shaped titanium treads. The necessary to slow down and turn to avoid small,
wheel diameter was about 82 cm, and the width was hummocky craters and other obstacles, the cruise
about 23 cm. speed was limited to 6–7 km/hr. Hard turns of the
In addition, a fourth vehicle, the “Lunar Motorcy- LRV at speeds above 5 km/hr resulted in skidding.
cle,” was designed at NASA’s Manned Spacecraft The maximum negotiable slope was 19°–23°, and
Center during the Apollo program to serve as a such slope values were approached at
backup in case the LRV was not ready in time for
Physical Properties of the Lunar Surface 525
the Apollo 16 and 17 sites. The measured energy where W = wheel load (N); A = wheel footprint area
consumption of the LRV during the Apollo 15, 16, and (cm2); L = wheel chord length of ground contact (cm);
17 missions is shown in Fig. 9.45. The average b = wheel width of ground contact (cm); kc = cohesive
“mileage” was 35–56 W-hr/km or, normalized to the modulus of soil deformation [N/(cm)n+1]; kφ, =
gross mass, about 0.050–0.080 W-hr/km/kg. frictional modulus of soil deformation [N/(cm)n+2]; k =
Trafficability parameters. The energy consumed kc/b + kφ, = soil consistency [N/(cm)n+2]; n =exponent
by a wheeled vehicle operating on the lunar surface of soil deformation; cb = coefficient of soil/wheel
can be divided into three components: soil cohesion (N/cm2); φb = soil/wheel friction angle (deg);
compaction, roughness, and elevation changes. K = coefficient of soil slip (cm); and s = wheel slip
Soil compaction can be estimated from empirical (dimensionless).
equations developed by Bekker (1969) Based on detailed wheel-soil interaction studies of
34 simulants, Costes et al. (1972) concluded that
variations in the trafficability soil parameters had
little influence on the energy consumption of the
LRV. For planning the LRV traverses, they recom-
mended the trafficability soil parameters shown in
Table 9.14, which were based on soil mechanics data
obtained from the Surveyor missions. For the LRV
characteristics, the ratio of soil compaction resis-
tance to wheel load (Rc/W) is about 0.016. In other
words, the rolling resistance for the LRV is equivalent
to climbing a rigid slope of only 1°. In terms of energy
consumption, this amounts to only 0.0073 W-
hr/km/kg.
Physical Properties of the Lunar Surface 527
Fig. 9.44. Schematic three-way diagram of the Apollo Lunar Roving Vehicle (LRV) and its dimensions (after
Costes et al., 1972). Upper diagrams show entire LRV; lower diagrams show details of wheel and wheel
assembly.
528 Lunar Sourcebook
Fig. 9.45. Diagrams of cumulative energy consumption of the Apollo Lunar Roving Vehicle (LRV) as a function
of distance traversed during the Apollo 15 (top), Apollo 16 (center), and Apollo 17 (bottom) missions. Durations
of each Extravehicular Activity (EVA) are given at the top of each diagram. The agreement between actual
measurements of battery levels (points) and the theoretical consumptions predicted from a soil model for each
landing site (solid lines) is excellent. Average “mileage” for the LRV was about 35–56 W-hr/km altogether, or
about 0.05–0.08 W-hr/km for each kilogram of mass transported.
Surface roughness is usually described by a power could be attained if improved roads are constructed
spectral density curve, as shown in Fig. 9.46. For this on the lunar surface; a value of practically zero could
type of analysis to apply, the wheel-chassis suspen- be used for a slow-moving unmanned vehicle.
sion system must be sufficiently “soft” to keep the The component of energy consumption lost during
vehicle on the surface in the low lunar gravity. All four elevation change depends on the traverse. Over
wheels of the LRV can (and did) leave the ground at distances in excess of 100 km, the average lunar
high speeds. For vehicle trafficability estimates, the slope is typically less than 2°. In the case of the Apollo
energy loss caused by roughness over a given distance 15–17 LRV traverses, the net elevation change was
is proportional to the speed (Costes et al., 1972). zero, because they were all round trips, beginning and
During the LRV traverses on Apollo 15–17, this ending at the Lunar Module. As shown in Fig. 9.47,
component of energy consumption amounted to about the up and down slopes on these traverses were
0.0027 W-hr/km/kg, which is equivalent to climbing essentially symmetrical.
a smooth slope of 0.4°. This value is probably a fairly Altogether, the LRV energy consumption caused by
reasonable estimate for designing future manned all lunar surface characteristics amounted to only
vehicles, even for travel in rougher areas, because the about 0.01 W-hr/km/kg, or about 15% of the total
speed in such regions will necessarily be reduced. A “mileage” (Table 9.13). The remainder is the sum of
lower energy consumption energy losses in the traction drive system, the energy
Physical Properties of the Lunar Surface 529
TABLE 9.14. Trafficability soil parameters (after Cosies the vicinity of some craters. In fact, the unmanned
et al., 1972; Mitchell et al., 1973a, 1974). vehicle had great difficulty in proceeding in these
areas.
Clearly, future lunar surface vehicle design and
operation will need to take into account such soft-soil
areas and other impediments. For adverse conditions,
a more flexible (i.e., variably-geared) vehicle or
multiple vehicles may be needed. There is an old rule
for terrestrial four-wheel-drive vehicles that is equally
applicable to the Moon: Use two-wheel drive until you
get stuck, then use four-wheel drive to get unstuck.
The corollary to this rule is that if you use four-wheel
drive continuously, you will eventually get very stuck.
That is when a winch and another vehicle come in
handy.
Fig. 9.47. Distribution of observed slope angles on the lunar surface, using data obtained from the three
Apollo 15 Extravehicular Activities (EVAs) employing the Lunar Roving Vehicle (LRV) (after Costes et al., 1972).
The slopes encountered on this mission were typically <2°. All EVA traverses began and ended at the same
point, the Lunar Module.
this situation partially compensates on the Moon for and T = temperature (Kelvin), and E0 and E1 are
the inability to skip signals off an ionosphere for long activation energies (Table 9.15). Because these lunar
distances as is done on Earth. The low conductivity materials have very low conductivities, dielectric
and low loss are also responsible for the fact that relaxation effects and displacement currents dominate
lunar materials are readily chargeable and will remain at very low frequencies. Olhoeft et al. (1973a) and
electrically charged for long periods of time. The large Schwerer et al. (1974) provide the best reviews of
photo-induced change in electrical conductivity at electrical conductivity of lunar rocks.
lunar sunrise and sunset can charge surface soil The electrical conductivities measured on lunar
particles to the point that they will levitate and move samples (Schwerer et al., 1972, 1973, 1974; Olhoeft et
(Alvarez, 1977a; De and Criswell, 1977; Criswell and al., 1973a,b, 1974b) and related materials (Housley
De, 1977; and Morin, 1972; Piwinskii and Duba, 1975; Duba et
al., 1976, 1979; Huebner et al., 1978, 1979; Schock et
al., 1980) are key parameters in using the data
532 Lunar Sourcebook
* Conductivity.
†Activation energy.
Data for sample 12002 are from Olhoeft et al. (1973); the rest are from Schwerer et al. (1974).
obtained by deep electromagnetic induction sounding of (1986) provide the best reviews of the problems of
the Moon to derive a selenotherm, i.e., a temperature interpreting electrical conductivity profiles from the
profile within the Moon (Dyal and Parkin, 1971; electromagnetic sounding data in order to further
Schubert et al., 1973; Sonett and Duba, 1975; Dyal et infer a selenotherm. Figure 9.49 illustrates a typical
al., 1976; Wiskerchen and Sonett, 1977; Dyal and Daily, set of such interpretations from Hood (1986).
1979; Vanyan et al., 1979; Hobbs, 1980; Hood et al.,
1982a,b). Sonett (1982) and Hood
9.2.2. Photoconductivity
Alvarez (1975, 1977b) measured the effects of
ultraviolet (UV), visible, and infrared (IR) irradiation
on the electrical conductivity of lunar samples,
including lunar soils (Figs. 9.50 and 9.51). Relative to
their surface electrical conductivity in darkness, the
soils show about a 101 increase in conductivity in the
IR and a 106 increase in the UV. The latter increase is
comparable to that produced by an 800°C increase in
temperature. These large changes in conductivity
with irradiation can produce large movements of
electrical charge across the solar terminator on the
lunar surface.
Fig. 9.50. Changes ίn surface electrical conductivity (vertical axis) of lunar soil 63501,52 produced as the
result of irradiation by ultraviolet (UV; the sum of 235 nm + 365 nm wavelength irradiation), visible (VIS), and
infrared (IR) radiation (after Alvarez, 1977b). Conductivity rises rapidly at the start of irradiation (time = 0 on
horizontal axis), then drops rapidly when the irradiation ceases (“OFF”). The effect is greater at higher
temperatures (295 K; upper curves) than at lower ones (100 K; lower curves). Note that the two highest curves
were obtained with all three sources on (VIS, IR, UV).
Ι
Physical Properties of the Lunar Surface 535
Fig. 9.51. Changes in surface electrical conductivity (vertίcal axis) of lunar soil 15041,21 produced as the
result of irradiation by ultraviolet (UV; the sum of 235 nm + 365 nm wavelength irradiatίon), visίble (VIS), and
infrared (IR) radiation (after Alvarez, 1977b). Conductivity rises rapidly at the start of irradiation (time = 0 on
horizontal axis), then drops rapίdly when the irradiation ceases (“OFF”). The effect ίs greater at higher
temperatures (295 K; upper curves) that at lower ones (100 K; lower curves). As in Fίg. 9.50, the two highest
curves were obtained with all three sources on (VIS, IR, UV).
Ι
536 Lunar Sourcebook
The large electrical conductivity change with visible effects, but not the charge accumulations and
and UV irradiation (discussed above), combined with discharge hazard. More detailed discussions of these
the very low electrical conductivίty and dίelectric effects are found in Criswell (1972, 1973, 1974),
losses of lunar materials, can produce an extremely Fenner et al. (1973), Gold and Williams (1973, 1974),
efficient electrostatic charging mechanism between Arrhenίus (1974), Alvarez (1974c, 1975, 1977α),
opposite sides of the lunar terminator. Across this Cameron (1974, 1978), Freeman and Ibrahim (1974),
moving boundary, charging of lunar soil particles Gold (1974), Goldstein (1974), McCoy and Criswell
could be sufficient to levitate them above the surface, (1974b), Williams (1974), Rennilson and Criswell
producing a “dust storm” of particles that would (1974), Reasoner (1975), Criswell and De (1977), and
follow the solar terminator around the Moon. De and Criswell (1977). Similar effects have also been
McDonnell and Flavill (1977) and McDonnell (1979) suggested to exist on the surface of Mercury (Ip,
have calculated that such electrostatic levitation 1986).
would extend to 10 m above the lunar surface. Another hypothesis to explain lunar transient
However, if the dust could reach higher altitudes, phenomena is that they are associated with electro-
such a phenomenon could explain some observations dynamic effects generated during rock fracturing
of the Apollo astronauts, who reported and drew (Zito, 1989). For other hypotheses, see section 11.4.2.
pictures of “streamers” and bands of corona/zodiacal
light (the results of light scattering by dust?) 9.2.4. Dielectric Permittivity
extending several kilometers above the lunar surface Dielectric permittivity is a measure of the ability of a
while approaching orbital sunrise (McCoy, 1976a,b). material to keep electrical charges physically
The Apollo 17 Lunar Ejecta and Meteorites (LEAM) separated by a distance, a process called electrical
experiment (Berg et al., 1973) found increased particle polarization. The relative dielectric permittivity is the
counts during the passage of the terminator over the ratio of the permittivity of a material to the
instrument (Berg et al., 1974), and the experimenters permittivity of vacuum. The speed of electromagrietic
noted that “ . . . all of the events recorded by the propagation in a material is the speed of light in
sensors during the terminator passages are vacuum divided by the square root of the permittivity.
essentially surface microparticles carrying a high Available measurements on lunar sample dίelectric
electrostatic charge.” Other experimenters observed properties as a function of bulk density, frequency,
that “the particle event rate increases whenever the and TiO2 + FeO content are summarized in Table
terminator passes over the instrument. This increase Α9.16. The same data are plotted in Figs. 9.52–9.55
starts some 40 hours before sunrise and ends about in terms of relative dielectric permittivity and loss
30 hours after it” (Rhee et al., 1977). Such charged, tangent vs. density. Data are presented for all lunar
levitating soil particles, moving across the lunar sample types in Table Α9.16 and Fig. 9.52, for soils
surface, could create visibility problems during only (Fig. 9.53), for measurements at 450 MHz only
passage through the terminator, and the particles (Fig. 9.54), and for the Apollo 15–17 samples only
would also readily adhere to surfaces, creating (Fig. 9.55). In Figs. 9.56–9.59 the data are plotted as
coatings. Because the electrical conductivίty is lowest a function of (TiO2 + FeO) content for the same data
during lunar night, the soil will have the highest groups.
electrostatic charge-ability at night. Nighttime surface In each of the permittivity plots in Figs. 9.52-9.59,
activities that disturb the lunar soil could generate the solid curve represents the rigorous Bruggeman-
dust that would thickly coat exposed surfaces. Upon Hanai-Sen (BHS) mixing formula (Sen et al., 1981),
lunar sunrise, the resultant photo-induced increase in given as
electrical conductivίty would cause most of the
coatings to slough off, leaving only a thin residual
coating behind.
During lunar night, the low conductivity of the soil
will also create significant electrical charging hazards where ρ = bulk density in g/cm3; k = the complex
between mobile objects on the lunar surface, relative dielectric permittivity at the density, p; i =
producing the well-known wintertime “spark” square root of –1; and [9.8 + i0.1] is the complex
electrical discharge when the charged objects meet. relative dίelectric permittivity of a hypothetical zero-
The possible presence of a stabilizing lunar duricrust porosity lunar material with ρ = 3.5 g/cm3. The
(Greenwood et al., 1971; Heiken and Carrier, 1971; dashed curve in the same plot is the simpler formula
Carrier and Heiken, 1972) (see section 9.1.4) would proposed by Olhoeft and Strangway (1975), after
tend to reduce the levitation regression analysis on the entire larger dataset of
Table Α9.16
Physίcal Propertίes of the Lunar Surface 537
Fig. 9.52. Relative dielectric permittivity (upper vertical axis) and loss tangent (lower right vertical axis) for
lunar materίals as functions of bulk density (horizontal axis) and %(TiO2 + FeO). Data points represent all
samples from Table Α9.16. Both permittίvity and loss tangent increase with increasing density. The permittivity
plot (upper) shows two theoretical fits to the data poίnts: the Bruggeman-Hanai-Sen (BHS) mixing formula (Sen
et al., 1981) (solid line; see text for details) and a least-squares power-law fit (Olhoeft and Strangway, 1975)
(dashed line). The loss tangent plot (lower) shows three theoretical fits to the data: the equivalent BHS formula
(solid line), the two-dimensional regression to the power-law formula (given in upper left) (dashed line), and
three-dimensional regressίon (light dashed lines) of variance as a function of density and selected values of
%(TiO2 + FeO) (light dashed curves for 0%, 10%, 20%, and 30%).
538 Lunar Sourcebook
Fig. 9.53. Relative dielectric permittivity (upper vertίcal axis) and loss tangent (lower right vertίcal axis) for
lunar materίals as functions of bulk density (horizontal axίs) and %(TiO2 + FeO). In contrast to Fig. 9.52, data
points represent only soil samples from Table Α9.16. Both permittivity and loss tangent increase with
increasίng density. Theoretical fitted curves and their derivation are the same as described in Fig. 9.52.
Ι
Physical Properties of the Lunar Surface 539
Fig. 9.54. Relative dielectric permittivity (upper vertical axis) and loss tangent (lower right vertical axis) for
lunar materials as functίons of bulk densίty (horizontal axis) and %(TiO2 + FeO). In contrast to Fig. 9.52, data
points represent only measurements obtained at 450 MHz (from Table 19.16). Both permittivity and loss
tangent increase with increasing density. Theoretical fitted curves and their derivation are the same as
described in Fig. 9.52.
540 Lunar Sourcebook
Fig. 9.55. Relative dielectric permittivity (upper vertical axis) and loss tangent (lower right vertίcal axis) for
lunar materials as functions of bulk density (horizontal axis) and %(TiO2 + FeO). In contrast to Fig. 9.52, data
points represent only soil samples from the Apollo 15–17 missions (from Table A9.16). Both permittivity and
loss tangent increase with increasing density. Theoretical fitted curves and their derίvatίon are the same as
described in Fig. 9.52.
Ί
Physical Propertίes of the Lunar Surface 541
Fig. 9.56. Relative dielectric permittivity (upper vertical axis) and loss tangent (lower right vertical axis) of lunar
materials as functions of %(TiO2 + FeO) (horιzontal axis) and bulk density. Data points represent all samples
from Table 19.16. The relative permittivity (upper) is apparently independent of chemistry, but the loss tangent
increases with increasing %(TiO2 + FeO). Two theoretical fits to the loss-tangent data (lower) are shown: the
two-dimensional least-squares power-law fit (Olhoeft and Strangway, 1975) (dashed line) and the same three-
dimensional regression as in Fig. 9.52 (light dashed lines), plotted for four specific density values (0.8, 1.6, 2.4,
and 3.2 g/cm3).
Ι
542 Lunar Sourcebook
Fig. 9.57. Relative dielectric permittivity (upper vertical axis) and loss tangent (lower right vertical axis) for
lunar materials as functions of %(TiO2 + FeO) (horizontal axis) and bulk density. In contrast to Fig. 9.56, data
points represent only soil samples from Table Α9.16. The relative permittivity (upper) is apparently independent
of chemistry, but the loss tangent increases with increasing %(TiO2 + FeO). Theoretical fitted curves and their
derivation are the same as described in Fig. 9.56.
Ι
Physical Properties of the Lunar Surface 543
Fig. 9.58. Relative dielectric permittivity (upper vertical axis) and loss tangent (lower right vertical axis) for lunar
materials as functίons of %(TiO2 + FeO) (horizontal axis) and bulk density. In contrast to Fig. 9.56, data points
represent only measurements at 450 MHz (from Table Α9.16). The relative permittivity (upper) is apparently
independent of chemistry, but the loss tangent increases with increasing %(TiO2 + FeO). Theoretical fitted curves
and their derivation are the same as described in Fig. 9.56.
Ι
544 Lunar Sourcebook
Fig. 9.59. Relative dielectric permittivity (upper vertical axis) and loss tangent (lower right vertical axis) for lunar
materials as functίons of %(TiO2 + FeO) (horizontal axis) and bulk density. In contrast to Fig. 9.56, data points
represent only measurements on Apollo 15–17 samples (from Table A9.16). The relatίve permittivity (upper) is
apparently independent of chemistry, but the loss tangent increases with increasing %(TiO2 + FeO). Theoretical
fitted curves and their derivation are the same as described in Fig. 9.56.
Physical Properties of the Lunar Surface 545
(The parameters are the same as those above.) The The data in Figs. 9.60 and 9.61 (after Strangway
value of the regression for each subset of the data is and Olhoeft, 1977) illustrate the frequency depen-
shown in the upper left corner of each plot. dence of the complex dielectric permittivity (shown as
In the plots of loss tangent as a function of density relative dielectric permittivity) (closed circles) and the
(Figs. 9.52–9.55), the solid lines shown are derived loss tangent (open circles). Figure 9.60 shows data for
from the same BHS formula as above for the soil sample 14163, reduced to a density of 1 g/cm3, at
permittivity. The heavy dashed line is the two- several different frequencies: <1 MHz (Strangway et
dimensional regression to the formula (shown at the al., 1972), 450 MHz (Gold et al., 1972), 9375 MHz
top of each loss tangent plot) (Bassett and Shackelford, 1972), and at 3.6 × 107 MHz
in the IR (Perry et al., 1972). Figure 9.61 shows the
loss tangent = 10(0.440ρ – 2.943) for all samples frequency dependence of rock sample 14310 at
similar frequencies: <1 MHz (Chung et al., 1972), 450
The light dashed fine is a three-dimensional regres- MHz (Gold et al., 1972), 9375 MHz (Bassett and
sion to the formula (shown at the bottom of each loss Shackelford, 1972), and in the IR (Perry et al., 1972).
tangent plot) In both samples, the dielectric permittivity (dotted
loss tangent = 10(0.038(%TiO2 + %FeO) + 0.312ρ – 3.260) line) shows little or no frequency dependence at
frequencies below the IR, whereas the loss tangent
for all samples, plotted at 0, 10, 20, and 30%(TiO2 + (dashed line) is strongly frequency dependent. The
FeO). apparent minimum seen in the loss tangent for the
The plots of relative permittivity as a function of soil sample near 10 MHz (Fig. 9.60) would suggest
(TiO2 + FeO) contents (Figs. 9.56–9.59) indicate that that frequency to be the optimum choice for a deep
the permittivity shows no dependence on the (TiO2 + penetrating radar (Olhoeft, 1984) or for long distance
Fe)) content. The light dashed lines in the plots of loss subhorizon radio communication through soils.
tangent as a function of (TiO2 + FeO) (Figs. 9.56–9.59) However, for the rock sample (Fig. 9.61), the
are the same as those in Figs. 9.52–9.55, but they are appearance of a strong dielectric relaxation near 0.1
plotted at densities of 0.8, 1.6, 2.4 and 3.2 g/cm3; the MHz moves the equivalent minimum in loss tangent
heavy dashed line in these plots is the two- up to 500 MHz for rocks. However, few rock samples
dimensional regression of the formula (shown at the exhibited such a relaxation effect.
top of the loss tangent plot) The temperature dependence of the dielectric
properties of lunar soil sample 15301,38 are
loss tangent = 10(0.045(%Τi02 + %FeO) – 2.754) illustrated in Figs. 9.62–9.65 (Olhoeft et al., 1974b).
The dielectric permittivity of the soil (Fig. 9.62) is
for all samples. Figures 9.56–9.59 show a variety of independent of temperature at frequencies above 1
presentations: all data from Table A9.16 (Fig. 9.56), MHz, and both the permittivity (Fig. 9.64) and loss
the soil data only (Fig. 9.57), the 450 MHz data only tangent (Fig. 9.65) are relatively independent of
(Fig. 9.58), and the Apollo 15–17 data only (Fig. 9.59). temperature below 200°C. Solid lines in Figs. 9.62–
There are detectable differences between the Apollo 9.65 represent data for the model equation
11–14 samples and the Apollo 15–17 samples, and
these differences reflect different procedures in the
quarantine and curation of the sample sets after
collection (see section Α2.3). The Apollo 11–14 lunar
samples were quarantined to protect the Earth from
possible lunar life forms, and the procedures used where ko = 6700; k1 = 3.0; ω = frequency in
produced some contamination of the samples by the
radians/sec; τo = 2.56 × 10–12 sec; Eo = 2.5 eV; and α =
Earth’s atmosphere before measurement. The Apollo
the Cole-Cole frequency distribution parameter.
15–17 samples collected after the quarantine
The frequency distribution parameter (α) was also
procedures had been dropped were more carefully
found to be temperature-dependent, with the form
protected against the outside environment. As a
result, the Apollo 15–17 samples showed slightly
lower loss tangents and conductivitdes, because they
had only minimal exposure to the effects of moisture
(Olhoeft et al., 1975). The loss tangent is given as
The data in Figs. 9.52–9.59 show clearly that the
relative dielectric permittivity is a function of bulk
density but not of chemistry, and the loss tangent is a
strong function of both density and chemistry.
546 Lunar Sourcebook
Fig. 9.60. Relative dielectric permittivity (left vertical axis) and loss tangent (right vertical axis) for a lunar soil
sample (14163 fines) as a function of frequency (bottom horizontal axis) or wavelength (top horizontal axis)
(after Strangway and Olhoeft, 1977). Data are normalized to a soil density of 1 gm/cm3. The diagram
summarizes measurements made by several workers over a range of frequencies: <1 MHz (Strangway et al.,
1972); 450 MHz (crosses labeled G; Gold et al., 1972); 9375 MHz (crosses labeled B; Bassett and Shackleford,
1972); and 3.6 × 107 MHz (in the infrared) (crosses labeled P; Perry et al., 1972). The value marked Μ is an
overall mean permittivity for all lunar samples at 107 Hz.
Fig. 9.61. Relative dielectric permittivity (left vertical axis) and loss tangent (right vertical axis) of a lunar rock
(crystalline rock 14310) as a function of frequency (bottom horizontal axis) or wavelength (top horizontal axis)
(after Strangway and Olhoeft, 1977). The diagram summarizes measurements made by several workers over a
range of frequencies: <1 MHz (circles labeled C; Chung et al., 1972); 450 MHz (crosses labeled G; Gold et al.,
1972); 9375 MHz (crosses labeled B; Bassett and Shackleford, 1972); and permittivity at >3.6 × 107 MHz (in the
infrared) (curve labeled P; Perry et al., 1972).
Physical Properties of the Lunar Surface 547
Fig. 9.62. Relative dielectric permittivity (vertical axis) of lunar soil sample 15301,38 as a function of frequency
(horizontal axis) at several different temperatures (after Olhoeft et al., 1974b). Theoretical curves (solid lines)
fitted to the data poίnts are derived from the model equation discussed in the text. The data indicate that the
permittivity of the soil sample is largely independent of temperature at frequencies >106 Hz.
Fig. 9.63. Loss tangent (vertical axis) of lunar soil sample 15301,38 as a function of frequency (horizontal axis)
at several different temperatures (after Olhoeft et al., 1974b). Theoretical curves (solid lines) fitted to the data
points are derived from the model equation discussed in the text. On this log-log plot, the loss tangent shows an
approximately linear decrease with increasing frequency for temperatures between 25°C and 700°C.
Ι
548 Lunar Sourcebook
Fig. 9.64. Relative dielectric permittivίty (vertical axis) of lunar soil sample 15301,38 as a functίon of
temperature (horίzontal axis) at several different frequencies (after Olhoeft et al., 1974b). Theoretical curves
(solid lines) fitted to the data points are derived from the model equation discussed in the text. The data indicate
that the permittivity of the soil sample is largely independent of both temperature and frequency at
temperatures <400°C, or essentially over the whole range of lunar surface temperatures.
Fig. 9.65. Loss tangent (vertical axis) of lunar soil sample 15301,38 as a function of temperature (horizontal
axis) at several different frequencies (after Olhoeft et al., 1974b). Theoretical curves (solid lines) fitted to the
data points are derived from the model equation discussed in the text. The loss tangent of the soil sample is
largely independent of temperature below about 200°C, or essentially over the whole range of lunar surface
temperatures.
Physical Properties of the Lunar Surface 549
where σDC = DC electrical conductivity discussed increase with further stress (at a rate of about 0.5%/
above, and εo = vacuum permittivity = 8.854185 × bar), up to the maximum applied stress of 2 bar.
10–12 Farad/m. In summary, the relative dielectric permittivity of
The temperature dependence of the dielectric lunar rocks and soils is strongly dependent upon
properties of lunar rock is shown in Figs. 9.66–9.71 density and independent of chemistry; above 1 MHz, it
for several different samples: 12002,85 (Olhoeft et al., is also independent of frequency and temperature over
1973b), 14310,75 (Chung et al., 1972), and 70215,14 the lunar surface temperature range. The loss tangent
(Alvarez, 1974a,b,c). The temperature dependence of is strongly dependent upon density, chemistry,
the permittivity (Fig. 9.66) decreases with increasing temperature, and frequency, with a minimum loss
frequency, but the temperature dependence of the near 10 MHz in soils. The best reviews of these
loss tangent (Fig. 9.67) depends strongly on both properties and the related research are found in
temperature and frequency. Further studies of the Olhoeft and Strangway (1975), Strangway and Olhoeft
effects of frequency and temperature on the dielectric (1977), and Gary and Keihm (1978).
permittivity and loss tangent are found in Chung et al. These data have been used to estimate the
(1970, 1971, 1972), Katsube and Collett (1971, electromagnetic properties of the lunar surface
1973b), Chung (1972), Strangway et αl. (1972, 1975), materials to significant depths. Olhoeft and Strangway
Chung and Westphal (1973), Alvarez (1973a, 1974b), (1975) combined the density profiles of Carrier et al.
Οlhοeft et al. (1973a, 1974a, 1975), and Bussey (1972a, 1973b), Houston et al. (1974), and Mitchell et
(1979). al. (1972d) with the formula for permittivity as a
Frisillo et al. (1975) added another variable to these function of density to produce a profile of electrical
investigations by studying the effects of frequency, properties with depth in the lunar surface to depths of
temperature and vertical confining stress on the 100 m. Using more recent statistical data and
electromagnetic properties of lunar soils. The loss regressions, the formula for density vs. depth in the
tangent showed no dependence upon stress. However, regolith (see Fig. 9.16)
the relative dielectric permittivity increased rapidly
from 6% to 20% between 0.04 and 0.4 bar stress. This
rapid increase was followed by a linear
Fig. 9.66. Relative dielectric permittivity (vertical axis) of lunar rock sample 12002,85 as a function of frequency
(horizontal axis) at several temperatures (after Olhoeft et al., 1973a). The solid lines are smooth curves fitted to
the data. Curves derived from the model equation described in the text are more complicated and are not shown
here. At prevailing lunar surface temperatures (<200°C), the permittivity of the rock is virtually independent of
frequency for frequencies >103 Hz.
Ι
550 Lunar Sourcebook
Fig. 9.67. Loss tangent (vertical axis) of lunar rock sample 12002,85 as a function of frequency (horizontal
axis) at several temperatures (after Olhoeft et al., 1973a). The solid lines are smooth curves fitted to the data.
Curves derived from the model equation described in the text are more complicated and are not shown here.
The measured loss tangent varies significantly with frequency over the full range of frequencies and
temperatures measured, including temperatures corresponding to the lunar surface.
Fig. 9.68. Relative dielectric permittivity (vertical axis) of lunar rock sample 14310 as a function of frequency
(horizontal axis) at several temperatures. The solid lines are smooth curves fitted to the data. Curves derived
from the model equation described in the text (see Chung et al., 1972) are more complicated and are not shown
here. In contrast to rock 12002,85 (Fig. 9.66), the permittivity of rock 14310 varies significantly with frequency
at all temperatures.
Physical Properties of the Lunar Surface 551
Fig. 9.69. Loss tangent (vertical axis) of lunar rock sample 14310 as a function of frequency (horizontal axis) at
several temperatures. The solid lines are smooth curves fitted to the data. Curves derived from the model
equation described in the text (see Chung et al., 1972) are more complicated and are not shown here. In
contrast to rock 12002,85, the measured loss tangent varies with frequency over the full range of frequencies
and temperatures measured, including temperatures corresponding to the luanr surface. However, the value
remains close to 0.01 for a wide range in both temperature and frequency.
Fig. 9.70. Relative dielectric permittivity (vertical axis) of lunar rock sample 70215,14 as a function of
frequency (horizontal axis) at several temperatures (after Alvarez, 1974b). The solid lines are smooth curves
fitted to the data. The permittivity of this rock sample varies slightly with frequency at all temperatures, but the
variation is lowest at lower temperatures.
Ι
552 Lunar Sourcebook
Fig. 9.71. Loss tangent (vertical axis) of lunar rock sample 70215,14 (basaltic lava) as a function of frequency
(horizontal axis) at several temperatures (after Alvarez, 1974c). The solid lines are smooth curves fitted to the
data. In contrast to other rock samples, the measured loss tangent varies with frequency at higher
temperatures, but is relatively independent of frequency (and close to 0.1) at the lοw temperatures
corresponding to the lunar surface (lower curve).
where z is depth in centimeters, may be combined mitted and received by Earth-based radar systems
with the formula for relative dielectric permittivity vs. (DeWitt and Stodola, 1949; Evans and Pettengίll,
density 1963; Low and Davidson, 1965; Hagfors and Evans,
1968; Evans, 1969; Pettengίll, 1978; Thompson, 1979;
k = 1.919ρ Ostro, 1983), by orbiting systems (Brown, 1972;
Phillips et al., 1973a,b; Tyler and Howard, 1973;
to produce a profile for relative dielectric permittivity Porcello et al., 1974; Elachί et al., 1976; May, 1976;
vs. depth, and a similar depth profile for the loss Peeples et al., 1978; Sharpton and Head, 1982;
tangent. These combined formulas should be valid to Sharpton, 1985), or by surface instruments (Brown,
regolith depths of at least tens of meters. 1967; Brown et al., 1967; Muhleman et al., 1968a,b,
1969; Simmons et al., 1972, 1973; Kroupenio, 1973;
Kroupenio et al., 1975; Rossiter, 1977).
9.2.5. Electromagnetic Sounding Α third method involves the passive Earth-based
Electromagnetic sounding of the Moon can be detection of microwave radiation associated with
accomplished in three different ways. The first natural lunar thermal emissions (Krotikov and
technique is passive and uses the interaction between Troitsky, 1963; Gary et al., 1965; Hagfors, 1970;
the Moon and the solar wind. This interaction Muhleman, 1972; Keίhm and Langseth, 1973, 1975b;
produces low-frequency electromagnetic waves, which Schloerb et al., 1976). Table 9.16 summarizes these
diffuse into the Moon, inducing eddy currents. These dielectric sounding results, excluding the Earth-
induced currents in the Moon then perturb the based measurements. Figure 9.72 summarizes the
existing magnetic fields around the Moon. These results of Earth-based observations. These data are
perturbations may be monitored with magnetometers in excellent agreement with the laboratory measure-
on the lunar surface or in orbit around it, as a ments on lunar samples discussed earlier.
function of frequency, to produce an electrical In each of these methods, electromagnetic scat-
conductivity profile of the Moon. This concept was tering is the largest loss mechanism, and this process
discussed earlier with regard to Fig. 9.49 (section causes the single largest uncertainty in
9.2.1), and it is best reviewed by Sonett (1982) and interpretation. Electromagnetic scattering occurs
Hood (1986). when the wavelength of the electromagnetic energy
The second method of electromagnetic sounding being propagated in a material approaches the same
uses the artificial propagation of high-frequency size as the scale of heterogeneities in the material’s
electromagnetic energy. This energy may be trans- electrical properties. Such scattering may occur
Ι
Physical Properties of the Lunar Surface 553
Ι
554 Lunar Sourcebook
TABLE 9.18. Hagfors RMS slope for the Moon (after Pettengill, 1978).
within an otherwise uniform volume of material as At the simplest level, the whole Moon scatters
volume scattering produced by appropriate-sized electromagnetic energy by reflection (Evans and
heterogeneities within the material. (The same effect Pettengill, 1963). Table 9.17 lists the total radar cross-
on the scale of visible light waves prevents us from section of the Moon as a function of wave-length, after
seeing through snow, although ice may be clear.) Pettengill (1978). The total radar cross-section is the
Scattering may also occur at the boundary between area of an isotropic scatterer (identical in all
two different materials, such as the soil-vacuum directions), perpendicular to the transmitted energy,
interface at the Moon’s surface or the regolith-bedrock that would yield the observed radar echo amplitude if
interface below the surface. Α similar effect prevents it were located in place of the Moon. Table 9.18 lists
seeing through the surface of scratched ice or frosted the Hagfors root-mean-square (RMS) slope quantity
glass. Fundamental references on the nature of the (Hagfors, 1970), which is one measure of the angular
various scattering processes are Evans and Pettengill dependence of scattering. Increasing Hagfors RMS
(1963), Salomonovich and Losovskii (1963), Hagfors slope values indicate increasing quasispecular
(1964, 1967, 1968), Burns (1969), Tyler et al. (1971), (quasismooth) scattering. Small slopes correspond to
Pollack and Whitehill (1972), Parker and Tyler (1973), smooth surfaces (like that of still water). Higher
Pettengill et al. (1974), England (1975), Fisher and slopes indicate increasing roughness (like choppy or
Staelin (1977), Strangway and Οlhοeft (1977), wavy water).
Ishimaru (1978a,b), Olhoeft et al. (1979), Tyler (1979), Several key quantities can be measured in radar
Fung (1982), Simpson and Tyler (1982), Ulaby et al. returns from the Moon. These include the wave-length
(1982), Moore (1983), and Elachi et al. (1984). dependence of reflections, the changes in polarization
of the radar beam, the Doppler shift of
Physical Properties of the Lunar Surface 555
the original frequency, and the angular dependence of grain size distribution (slope -4), and scattering from a
the return energy amplitude. By analyzing these two-parameter model with stratified layers (long
quantities, radar scattering measurements may be wavelength scattering; based upon the results of
interpreted to determine the variations in surface Fisher, 1977). Figures 9.74 and 9.75 (after Strangway
topographic relief, hillslope angles, surface undula- and Olhoeft, 1977) illustrate two more measures of
tions, and boulder populations (roughness). These scattering for the Moon. Figure 9.74 illustrates
analyses and their interpretations are discussed in scattering as a function of delay time at various
the above references and also in Low and Davidson wavelengths (after Hagfors, 1970). Figure 9.75
(1965), Thompson et al. (1970, 1974), Zisk (1970, illustrates scattering from different regions of the
1972a,b), Hagfors and Campbell (1973), Linsky (1973), Moon in terms of Hagfors RMS slope angle and shows
Thompson (1974, 1979, 1986), Ulich et al. (1974), Zisk comparisons with similar measurements of Mercury,
et al. (1974), Schaber et al. (1975), Gary and Keihm Venus, and Mars.
(1978), Schaber et al. (1986), and Thompson and The earlier Earth-based radar investigations of the
Saunders (1986). Moon were used to help locate the Surveyor and
The results of some of these analyses are shown in Apollo landing sites. Similar Earth-based investiga-
Figs. 9.73–9.75. Figure 9.73 (after Gary and Keihm, tions of Mars were used to help select the landing
1978) shows an effective specific loss tangent derived sites for the Viking landers. Radar methods, both
from both laboratory studies on lunar samples Earth-based and from the Pioneer Venus orbiting
(similar to results shown in Figs. 9.60 and 9.61) and spacecraft, have produced the only topographic maps
Earth-based measurements, including the effects of of the atmosphere-obscured surface of Venus.
scattering. Their model (Fig. 9.73) incorporates several [Because of the strong absorption of radar waves by
factors: an intrinsic loss mechanism (slope –0.25), the water present in virtually all terrestrial surficial
Rayleigh-type scattering from the soil deposits, the only comparable terrestrial location
Fig. 9.73. Plot of effective specific loss tangent (vertical axis) as a function of wavelength (horizontal axis) for
lunar surface material, based on the results of active radar experiments (after Gary and Keihm, 1978). The loss
tangent appears to decrease with increasing wavelength in a nonuniform manner. The heavy solid line passing
through most of the experimental points represents a model consisting of the three components shown as solid
lines: (A) intrinsic loss (slope –0.25); (B) Rayleigh scattering from particle grains (slope –4); (C) scatterίng from
stratified subsurface layers (peak at long wavelengths; after Fisher, 1977).
Ί
556 Lunar Sourcebook
Fig. 9.74. Data on scattering from the Moon using Earth-based active radar observations at different
wavelengths. Data are presented as a function of relative returned power (vertical axis) and delay time
(horizontal axis) for different wavelengths (individual curves). Scattering by the lunar surface increases with
increasing wavelength, causing the values of returned relative power (plotted here) to decrease. The decline of
relative power with delay time for a given wavelength is called limb-darkening, i.e., the center of the Moon (the
subradar point) is brighter than the edges. Data from Hagfors (1970) and Strangway and Olhoeft (1977).
where these methods can be applied is in mapping In summary, a variety of methods have used
the subsurface of the extremely arid Western Desert electromagnetic energy to probe both the interior and
in Egypt, where scattering, rather than water the surface of the Moon. Low-frequency elec-
absorption, is the dominant loss mechanism (Schaber tromagnetic induction sounding, using the solar wind
et al., 1986).] Figure 9.76 illustrates a topographic and magnetometers, has been interpreted to produce
contour map, obtained from radar studies, of a a conductivity vs. depth profile for the lunar interior.
portion of the lunar surface. Table 9.19 lists some of Laboratory measurements of conductivity vs.
the available images and maps derived from radar temperature on lunar samples have then been
observations, two of which are shown in Figs. 10.8 combined with these profiles to infer a profile of
and 10.9. temperature vs. depth (selenotherm). In addition to
Physical Properties of the Lunar Surface 557
Ι
558 Lunar Sourcebook
TABLE 9.19. Available maps and images of the similar regardless of whether the incoming illumina-
Moon derived from radar observations. tion is dominated by UV, visible, or IR wavelengths.
Visible light corresponds to electromagnetic
radiation ranging in wavelength from about 0.4 µm
(violet) to about 0.76 µm (red). For wavelengths longer
than about 3 µm, however, lunar thermal emission
begins to make up a significant proportion of the
radiation received from the illuminated Moon, and at
wavelengths longer than about 4 µm, this thermal
radiation dominates. Other factors, including the
optical transmission characteristics of the Earth’s
atmosphere, the onset of thermal emission, and the
spectral response of common telescope detector
systems combine to set an effective upper limit of
about 2.5 µm for the wavelengths that can be
the deep probing of the Moon, high-frequency detected in modern lunar reflectance observations. Α
electromagnetic propagation sounding has been used lower limit of approximately 0.3 µm, between the
from Earth-based, orbital, and surface radar systems visible and UV, is defined for similar reasons.
to map the topography, stratigraphy, roughness, and In considering the physics of moonlight, the Moon
homogeneity of the upper 100 m of the lunar surface can be regarded as a sphere that is illuminated from a
(see section 9.3.2 below). specific direction in space. The intensity of the
radiation measured by an observer depends on what
fraction of the incoming sunlight is reflected by an
9.3. REFLECTION AND EMISSION OF element of surface area (albedo) and how the reflected
RADIATION FROM THE MOON light is distributed into various directions (photometric
Much of what we know about the lunar surface function) (Minnaert, 1961). Since most solar system
comes from remote observations using optical or radio objects appear to be single points of light in the sky,
astronomical methods, either from Earth or from the definitions in planetary photometry usually
spacecraft orbiting above the Moon. In general, the describe any planet as if it were a sphere with uniform
astronomical instrumentation used in these studies is surface properties, and these definitions are adequate
designed to detect various forms of electromagnetic for most purposes. However, the Moon’s nearness
radiation, which differ significantly in their allows us to observe its surface at relatively high
wavelengths. Different kinds of detectors and telescope spatial resolution, and the standard definitions of
combinations are used, depending on the wavelength observational astronomy sometimes require
of the radiation. modifιcations for the Moon.
The lunar radiation measured by these devices may Photometric function. The photometric function of
have several different origins: (1) insolation (sunlight) a surface is a mathematical expression that relates
reflected from the lunar surface, (2) manmade the intensity of the radiation scattered into any one
radiation reflected from the surface (e.g., radar), (3) direction to the intensity of the incident radiation,
thermal emission from the Moon itself, and (4) X-rays which arrives from a different direction. The
and gamma rays induced in the lunar surface parameters in the function vαry somewhat with
material by solar or galactic radiation. In this section wavelength, but the differences are small. Optical
we will discuss only the first three, which are related astronomers can usually discuss the scattering of
to the optical and dielectric properties of the surface. sunlight from a planetary surface without reference to
More information on the fourth type is provided in a specific wavelength.
Chapter 10. The directions of the incident beam and the
scattered beam are described in terms of the angles
9.3.1. Optical Astronomy they make with the surface normal, which is a vector
Galileo made the first recorded telescopic obser- perpendicular to the local surface. Α third angle is
vations of the Moon, looking at it through a crude required to fully specify the geometrical relationships;
telescope that sensed visible-light radiation. Modern in lunar photometry this is taken to be the phase
astronomy has expanded the use of telescopes with angle between the incident and scattered (outgoing)
lenses and/or mirrors to detect “light” (e.g., IR and beams. The incoming and outgoing beams define the
UV wavelengths) that may not be visible to the phase plane, which contains the phase angle. Full
human eye. However, the properties of the lunar Moon corresponds to (almost) zero phase angle; at
surface that determine how it reflects sunlight are new Moon, the phase angle is 180°.
Physical Properties of the Lunar Surface 559
(At a truly zero phase angle, the Earth lies exactly number of locations, the report is not widely available,
between the sun and the Moon, and we have a lunar and the method has not yet been widely applied.
eclipse.) Hapke (1963) developed the first theoretical model
A hypothetical surface that reflects incident light that adequately accounted for the lunar photometric
uniformly into all directions is said to be Lambertian. function. A later reformulation of the theory (Hapke,
Astronomers were puzzled for many years by the fact 1981, 1984) is widely accepted as the definitive model
that the Moon is decidedly non-Lambertian in its for reflectance from planetary surfaces. Hapke (1963)
reflectance. The lunar surface is highly bases his formulation on the two-stream solution of
backscatterίng, meaning that it preferentially reflects the radiative transfer equation, applied to a medium
incoming light back in the direction of illumination. composed of widely dispersed particles externally
For this reason, the apparent brightness of the lunar illuminated by a beam of radiation. He corrected the
surface increases as the phase angle decreases obtained solution to account for actual contiguity of
toward zero, and for phase angles of less than 5° the the particles on a real surface and the resultant
brightening is dramatic. This surge in surface shadowing of deeper particles by particles in the upper
brightness near zero phase angle is called the layers. He also utilized an observation by Irvine (1966)
oppositίon effect (Gehrels et al., 1964) or heiligenschein that photons penetrating the medium to illuminate a
(Wildey and Pohn, 1969). This effect is also visible to particle can always escape unattenuated if they
observers on the lunar surface. Astronauts working happen to scatter back along the path of incidence.
on the Moon experience a loss of contrast and detail These unattenuated backscattered photons produce
as they look at the surface with their backs to the the opposition effect, and the range of near-zero phase
sun, and ground that slopes away from the observer angles over which the brightening occurs is a measure
is just as bright as slopes facing an observer. of the ratio of the mean spacing between particles of
Minnaert (1961) provided an extensive review of the the lunar surface layer to the mean free path of
early observational work that defined the general photons within the individual solid particles
scattering behavior of the lunar surface (Fedorets, themselves.
1952; van Diggelen, 1959). Phase curves for various The bidirectional reflectance is the ratio of the
features were determined by measuring their radiance (or luminance) received by a detector viewing
brightness on photographs of the Moon at different a surface from a specific direction e (the emission
phases. Gehrels et al. (1964) measured such phase angle) to the radiance (or luminance) from the source,
curves by using photomultipliers in the focal plane of assumed to be illuminating the surface from a
a telescope and thus documented the reality of the direction i (incidence angle). Hapke’s expression for the
opposition effect. Wildey and Pohn (1964) also bidirectional reflectance function (Hapke, 1981,
produced phase curves using photoelectric pho- equation 16) is
tometry, and they discounted the apparent results,
obtained from early photographic work, that not all
lunar features peaked in brightness at zero phase.
Because Earth-based measurement of the Moon’s
brightness cannot be made exactly at zero phase,
measurements of the phase curve for nonzero angles
must be extrapolated to the maximum value. The (where µO = cos i, the direction cosine for the incident
heiligenschein surge at lοw phase angles makes that beam; µ = cos e, the direction cosine for the emerging
extrapolation difficult. However, photography from the beam; g = the phase angle; w = the single scattering
orbiting Apollo Command Module has captured the albedo for a particle; P(g) = the particle phase
zero phase point in some instances. Photographic function normalized so that its surface integral over a
measurements of the brightness surge between phase sphere equals 4π; Η(µ) = [1+2µ]/[l+2µ(1–w)1/2]; and
angles of 1.5° and 0° (Pohn et al., 1969; Wildey and B(g) = the backscatter function for the opposition
Pohn, 1969) show a variability, ranging from albedos effect = Bo{1–tan|g|[3–exp(–h/tan|g|)] [1–exp (–h/tan
of 7% at the Apollo 11 site to 19% at a site in |g|)]/2h} for |g| ≤ π/2, and B(g) = 0 for |g| > π/2. Bo
highland terrain. is an empirical factor ≈ exp(-w2/2); h is a parameter
Saari and Shorthill (1967), also using a photomul- depending on particle spacing and has a value
tiplier tube, scanned the image of the Moon in the approximately equal to 0.4 for lunar soils.
focal plane of a telescope and produced digitized This bidirectional reflectance function is used
images of the illuminated Moon at successive phases widely by astronomers and other scientists for remote
throughout a lunation. In principle, this dataset can sensing of planetary surfaces from space-
be used to produce phase curves for any visible lunar
feature. Although this has been done for a limited
Ι
560 Lunar Sourcebook
craft. For such observations, the resolution element Finally, the reflectance of a surface element viewed
on the planetary surface can range from tens of normally (i.e., along the surface normal) at zero phase
meters to several kilometers across. The terrain may angle relative to the ideal white diffusing plate is called
contain craters, mounds, and other topographic the normal albedo (Minnaert, 1961, p. 218). When the
features that cause the direction of the local surface angles i = e = 0 are substituted into Hapke’s formula,
normal to vary within a single resolution element. the result is the normal albedo. The value of the
When this happens, local angles of incidence and normal albedo for an area of lunar surface will depend
emission will also vary. When i > π/2, an area will lie on the local chemical and mineralogical composition,
in shadow; when e > π/2, an area will be on the particle size, packing density, etc. In general, crater
backside of a slope and invisible to the observer. In ray systems are the brightest features on the Moon,
addition, one feature may block the view of another. and highland areas are brighter than maria.
All these conditions associated with macroscopic For lunar surface materials, the brightness of any
roughness lead to errors in the predictions of the area, observed from the Earth at full Moon (zero
model. The errors are particularly large when phase), yields a value that is virtually that of the
astronomers observe the limb regions of a planet or normal albedo for that area, irrespective of the
when spacecraft sensors point to surface regions lying orientation of the surface normal to the direction of
near the horizon. illumination. However, the formal definition of normal
Hapke (1984) derived an expression for correcting albedo requires the surface element to be illuminated
any photometric function for the effects of roughness. normally.
The correction works quite well, as he demonstrated. Maps of the Moon showing the normal albedo of
Application of the correction is not difficult, but the surface features have been produced by Pohn and
explanation is too lengthy to reproduce here. The Wildey (1970) and Wildey (1978). The earlier map is
interested reader will find it discussed in detail in the based on a carefully calibrated full Moon photograph
original paper. (phase angle = 1.5°) that was digitized by a scanning
Albedo. The spherical albedo or Bond albedo is the microdensitometer. The resulting contours of equal
fraction of sunlight reflected by a planet in all film density were converted to radiance using
directions. For the Moon, the actual value will depend photoelectric observations of a network of lunar
on which side is illuminated. The values discussed surface features, taken at the same time as the
here refer to the Earth-facing hemisphere, or nearside, photographic exposure. However, as discussed earlier,
which contains most of the dark maria. no photograph of the Moon from Earth can exhibit
The Bond albedo of a planet is difficult to measure, exact zero phase, and these studies therefore do not
and astronomers historically have derived it from rigorously provide the true normal albedo of lunar
measurements of a related quantity, the geometrical features.
albedo (p). The geometrical albedo is the ratio of the In the later work, Wildey (1978) digitized a series of
radiance from a planet at zero phase (e.g., full Moon) photographs taken over a range of phase angles
to that of a perfectly diffusing planar surface (Lambert between 2.5° and 4.0°. By fitting a parametric function
reflector) located at the same distance from the sun to the data as a function of phase, he was able to
and normal to the incident radiation (Kaula, 1968, p. extrapolate to develop a lunar “image” of ideal normal
260). The geometrical albedo is related to the albedo as well as a corollary image representing the
spherical (Bond) albedo (A) by the simple relationship magnitude of the heiligenschein effect across the
A = pq, where q, the phase integral, is given by nearside of the Moon.
The mapping of lunar surface brightness during a
lunation (Saari and Shorthill, 1967) included an image
taken at a phase angle of –2°, which is, in essence, a
map of normal albedo. The data were collected using
a photomultiplier to scan the lunar image in the focal
The phase angle g is the angle between the sun and
plane of the telescope.
the Earth as seen from the planet, and the phase
Dollfus and Bowell (1971) studied measurements of
function φ describes the variation of brightness of a
normal albedo in 9 different reports and derived a
planet over the range of possible phase angles. The
best estimate of this quantity for 67 lunar regions; the
value of the phase integral can be obtained by
albedo for these regions ranges from 5.7% to 18.4%.
calculating the area under the phase curve, the plot of
Single scattering albedo. Modern formulations of
φ(g). For the Moon the function φ can be derived from reflectance from planetary surfaces are based on the
the Hapke photometric function. Values quoted in the concepts of radiative transfer theory (e.g., Hapke,
literature for the phase integral are derived from
Earth-based measurements of moonlight.
Physical Properties of the Lunar Surface 561
1981). The fundamental transport equation invokes a that contains the incoming and reflected light rays.
hypothetical volume element within the surface that, These quantities are called the degree of polarization
when illuminated by a pencil of radiation from an and the azimuth of polarization, respectively.
arbitrary direction, absorbs a fraction of it and Although the lunar surface is far from smooth at the
scatters the remainder in various directions relative to dimensions (~0.4–1.0 µm) equivalent to optical
the incident direction. The fraction absorbed wavelengths, moonlight nevertheless is partially
(absorption coefficient), the fraction scattered polarized. The degree of polarization (P) is given by the
(scattering coefficient), and the distribution function relation
for the scattered component are all parameters in the
solution to the transport equation. The single
scattering albedo is the ratio of the scattering
coefficient to the sum of the scattering and absorption where I1 is the intensity observed normal to the phase
coefficients. plane, and I2 is the intensity parallel to it (Dollfus,
In the Mie theory, scattering and absorbing 1961). If P is positive the azimuth is assumed to be
particles are ideal dielectric spheres. In such a 180°, while a negative value of P implies an azimuth of
situation, the single scattering albedo is related to the 90°. These simple definitions suffice for lunar
physical properties of the material and the size of the polarimetry because Lyot (1929) discovered that the
particles relative to the wavelength of the radiation. polarization of moonlight is coincident with the phase
The derivation of the basic radiative transport plane near full Moon, disappears around phase 23°,
equation further assumes that the scattering centers and then grows in magnitude but rotated 90°,
in the medium are well separated so that they interact perpendicular to the phase plane. Similarly, Gehrels et
independently with the radiation field. Although the al. (1964) failed to find any gradual rotation of the
more complicated structure of a real planetary surface plane of polarization; the polarization vector either lies
violates these assumptions, the Mie theory works well within or normal to the phase plane. They also
within the limits of various approximations that must observed that the positive maximum in polarization of
be applied to the real problems. reflected light from the lunar surface was highest in
Unfortunately, the considerable unreality of these the UV (0.36 µm), intermediate in the green (0.54 µm),
same approximations blurs the physical meaning of and lowest in the near IR (0.94 µm).
the single scattering albedo values that are derived Dollfus and Bowell (1971) reviewed and discussed
from applying the theory to actual reflectance lunar polarization observations, presented new data,
measurements. For example, values obtained by and elucidated the details of the polarization
fitting Hapke’s (1981) function to phase curves for mechanism. For example, at a given wavelength, the
lunar regions can be used only for intercomparison negative branch of the curve, corresponding to phases
with other lunar or planetary data or with laboratory
of 0 ≤ g <
~ 23°, has approximately the same form for all
measurements on lunar soils and analogue materials.
Polarization. Sunlight is unpolarized. In other lunar terrains. Similarly, the minimum value of the
words, the oscillations of the electric and magnetic negative branch, Pmin, remains constant for all lunar
vectors of the light waves in sunlight are not regions, but it decreases from –10‰ (–10 parts per
preferentially aligned in any particular direction. If thousand) at a wavelength λ = 0.325 µm to –12‰ at λ
sunlight is transmitted through a polarizer, its = 1.050 µm. The average value for all wavelengths is
intensity after transmission does not depend on the Pmin = 11.2 ± 0.3‰ and g(Pmin) = 9.6 ± 0.6°.
rotation angle of the polarizer. On the positive branch of the polarization curve,
If sunlight is reflected from a smooth surface in corresponding to phases > ~ 23°, the maximum value of
some direction other than back along the surface polarization (Pm) varies inversely with the albedo of the
normal, the reflected light becomes at least partially lunar terrain. The highest value noted for Pm is 350‰
plane polarized. (This is a familiar phenomenon for a dark mare region in Oceanus Procellarum at λ =
associated with Fresnel reflection, and it explains why 0.325 µm. However, regions of high albedo, such as
polaroid sunglasses are effective in reducing glare.) highlands or young rayed craters, can display Pm
The polarization of the reflected light is specified by values as low as 30‰, a value measured at
two quantities: the percentage of the intensity Aristarchus at λ = 0.6 µm. The phase angle of
attributable to the polarized component and the angle maximum polarization, g(Pm), increases with Pm, but
that the direction of the electric vector makes with the the correlation is independent of wavelength. This
normal to the phase plane, i.e., the plane relation can be expressed by the equation
562 Lunar Sourcebook
where g(Pm) ίs expressed in degrees and Pm in incoming energy would be scattered by the rest of the
thousandths. Dollfus and Bowell (1971) note that Moon and would travel in other directions, according
uniform behavior of polarization across the entίre to the laws of specular reflection.
observable lunar dίsk provides strong evidence that Most of the returned energy is observed to come
the properties of the microstructure of the upper from an area surrounding the subradar point.
lunar surface are remarkably uniform all across the However, definite radar echoes can be detected from
Moon. regions out to the lunar limb, implying that the lunar
A fascίnating correlation between the polarίzation surface is not entirely smooth at radar wavelengths.
maximum (Pm) and the normal albedo (λ) has also For these reasons, radar reflection from the Moon is
been dίscovered. The relation is given by commonly described as “quasispecular,” because the
strong central reflection only approximates the
expected behavior of an ideally smooth spherical
mirror.
A similar relationship between geometric albedo (q) Radar backscattering. The brightness (cross-
and the slope of the polarization curve at the inversion section) of the subradar point is a measure of the
point has also been documented in later work (Zellner dielectric constant of the lunar surface, while the
et al., 1977). This latter correlation has proved to be shape of the trailing echo (scattering function) is a
particularly important in the determination of albedos measure of the surface roughness and heterogeneity
of asteroids and other small bodies in the solar of the surface on the scale of the wavelength being
system. Based on both lunar sample measurements used. The radar echo from the zone immediately
and lunar observations (Bowell et al., 1973), the best- around the subradar point is approximately three
fit relation for the Moon is orders of magnitude stronger than the echo from the
annular region at the lunar limb (Thompson, 1979).
Radar systems have the capability to transmit
energy with a single mode of polarization and then to
where p is the geometric albedo for a lunar region and detect and analyze the returned energy in both the
h is the inversion slope given in percent per degree. originally transmitted mode of polarίzation and the
This relation is also consistent with one derived by orthogonal mode. Commonly the transmitted signal is
Veverka and Noland (1973) from lunar data published circularly polarized to minimize the effects of Faraday
by Dollfus and Bowell (1971) and Dollfus et al. (1971). rotation during the signal’s outward passage through
the Earth’s ionosphere. The process of reflection from
9.3.2. Radar Astronomy an ideal dielectric sphere reverses the direction of
circular polarίzation. When lunar radar observations
Powerful radar installations built for military are made, the energy returned with the direction of
purposes during World War II were subsequently polarization to be expected from an ideal sphere is
utilized in the postwar period to detect echoes called polarized; the energy received in the other
reflected from the Moon. The first successful attempts direction of polarization is called depolarized. A
in 1946 produced only marginal returned signals nonspecular component is also present in the
(DeWitt and Stodola, 1949), but rapid improvements in polarized return; this component is qualitatively
the technology soon made possίble accurate ranging attributed to flat surfaces that happen to be tilted
to the Moon, measurements of the Doppler shifts toward the radar. The depolarized return is attributed
produced by lunar motions, the detection of Faraday to the presence of surface rocks and surface
rotation of the polarίzation vector of the signals in the roughness. Radar energy can penetrate a distance of
Earth’s ionosphere, the determination of the gross 10 to 50 wavelengths into the extremely dry lunar
radar scattering law for the lunar surface, and regolith, allowing subsurface rocks, possibly as deep
estimates of the Moon’s radar cross-section (Evans, as several meters, to contribute to the radar
1962). reflections returned to Earth.
The transmitted radio signal can be pictured as a Although the strengths of the polarized and
plane wave traveling from the antenna and striking depolarized returns (echoes) are quite different near
the Moon. If the Moon behaves like a smooth dielectric the subradar point, the falloff in the strengths of the
sphere, the plane wave will strike it first at the point two types of returns toward the limb regions exhibits
closest to the antenna—the subradar point. The only somewhat similar behavior (Hagfors, 1967). Near the
energy reflected back to the antenna would come from subradar point, the polarized echoes are quite strong,
a small zone surrounding the subradar point, and they decrease sharply and nonunίformly with
corresponding to the area of the first Fresnel zone of increasing angle of incidence from the subradar point
the wavefront. The remainder of the (i = 0°) outward to about i = 35°. From about
Physical Properties of the Lunar Surface 563
i = 35° outward, the slope of the angular dependence where σ is the radar cross-section of the Moon.
becomes more gradual, varyίng as cos3/2i out to The radar cross-section σ has the dimensions of an
approximately i = 80° (Evans and Hagfors, 1966). area and is defined as the geometrical cross-section of
Beyond i = 80° to the limb (0 = 90°), the dependence is a metallic sphere whose radar echo has the same
proportional to cos i. In contrast, the strength of the strength as the Moon. We can relate this quantity to
depolarized echo varίes uniformly as cos i from the the physical parameters of the Moon by considering
subradar point outward. the Moon as an absorbing dielectric sphere of radius
Although the lunar radar backscattering law is a. Following a derivation presented by Bohren and
similar in form at all wavelengths, the fraction of Huffman (1983, p. 120), we can write
power in the diffusely scattered component varies
considerably, being much greater at higher
frequencies (Hagfors, 1970). As a result, the lunar
disk appears more uniformly bright when illuminated where ρ is the Fresnel reflection coefficient of the
by the shorter radar wavelengths. Pettengill dielectric material (Evans and Hagfors, 1964) and g is
(discussion following Hagfors, 1970) points out that a a gain factor equal to unity for our ideal sphere.
plot comparing the lunar scattering laws is misleading Because the Moon is not ideally smooth, some of the
when the plots are normalized to unity at the returned signal comes from reflections outside the
subradar point. Rather, it is more meaningful to scale small region surrounding the subradar point. (This is
the areas of the curves to be proportional to the total the diffuse component of the polarized return.)
radar cross-sections at the corresponding Because the lunar surface is not perfectly absorbing,
wavelengths. some of the returned echo comes from scattering
The form of the scattering law that is most widely within the surface. (This is the depolarized
used for studying the Moon and the inner planets is component.) For these reasons, g is larger than unity
(Ostro, 1983) by some unknown amount.
The product gρ can be derived directly from
observational data. If the gain factor can be
independently calculated, then the dielectric constant
where σo is radar cross-section per unit area, φ is the (ε) of the lunar surface can be obtained from the
angle of incidence, ρ is the Fresnel normal-incidence relation
power reflection, and βrms is the root-mean-square
surface slope at wavelength-dependent scales. The
theoretical basis for this model can be found in
Hagfors (1964, 1970).
Radar cross-section. To understand the radar Attempts to calculate (Rea et al., 1964, 1965) or
signal returned from the Moon, we use the basic radar estimate (Evans and Hagfors, 1964) the gain factor
equation that says have resulted in estimates of the dielectric constant, ε,
Received Power = given by Hagfors (1970). Further complications can be
Transmitted Power Density at the Moon × introduced into these calculations by assuming that
Backscatter Cross-Section of the Moon × the surface consists of layers with differing dielectric
Geometrical Attenuation on Return Path × properties or that the dielectric properties vary
Effective Antenna Aperture smoothly with depth in some way (Hagfors, 1970),
both of which assumptions are more apt to be more
For a transmitter of power Pt, the power density at the
realistic representations of the actual lunar surface.
distance of the Moon, R, is given by ΡtG/4πR2. The
Bistatic radar observations. When spacecraft
denominator of this expression is just the geometrical
were placed into orbit about the Moon, it became
attenuation due to the familiar inverse-square law,
possible to carry out bistatic radar experiments in
and the antenna gain G can be related to the effective
which the transmitting and receiving antennas are in
antenna aperture A by
different places. In the lunar experiments, antennas
on Earth could detect both the direct radio or radar
transmissions from the spacecraft and the nearly
where λ is the wavelength of observation. Combining simultaneous reflections of the signals from the lunar
these terms, we get an expressίon for received echo surface toward Earth. A small Doppler shift in the
power Ρr (Ostro, 1983) reflection wavelength relative to the transmissiοn
wavelength allows the two signals to be differentiated.
The orbital motion of the radar-
564 Lunar Sourcebook
sounding spacecraft causes the point of specular were observed. No successful models for the structure
reflection to move across the lunar surface along the and properties of the anomalous regions could be
ground track. Careful measurement of the intensity developed; the authors speculated that their data
and polarization of the returned signals yields implied the existence of some frequency dispersive
information on the reflectivity of the lunar material, mechanism, such as layering in a high-loss dielectric
possible subsurface layering, and surface slopes medium.
(Tyler, 1968b). In highland regions the longer wavelength (116 cm)
The effective surface resolution element in these data were fitted successfully to a model using a
experiments corresponds to the first Fresnel zone uniform dielectric interface with ε = 2.8, a lower value
associated with the quasispecular reflection. The size for dielectric constant than that found in the maria.
of this zone depends on the altitude of the spacecraft Observations at the shorter wavelength (13 cm) in the
as well as on the surface slope distribution. For Apennine Mountains and in the central highland
measurements made using Explorer 35 in 1967, the regions also yielded consistent results, but the data
surface resolution was approximately 100 km (Tyler, from units in the Palus Somni area and the rim of
1968a). For the Apollo 14 and 15 experiments, the Mare Crisium could not be fitted with simple models
resolution elements varied from 10 to 40 km, at λ = 13 cm.
depending on the nature of terrain overflown (Tyler Since the polarization of returned radio signals can
and Howard, 1973). be determined, the echoes from the lunar surface also
During a pass over Oceanus Procellarum, the contain information about slopes and surface
transmitted signal from Explorer 35 was maintained roughness. The complexity of both the surface and
with its linear polarization lying approximately in the the physical mechanisms that alter the polarization of
plane of incidence. A Brewster null in the reflected the signal permit only some general statements with
signal was observed at an angle of incidence of 60° ± any degree of confidence. In general, slopes in the
1°, a result corresponding to a dielectric constant of highlands are higher than slopes in the maria. At λ =
3.0 ± 0.2 at a wavelength of 2.2 m (Tyler, 1968a). This 13 cm, highland rms (root-mean-square) slopes range
measurement was made in mare material located from 6° to 8°, while mare rms slopes range from 2° to
north-northwest of Hansteen Crater, and the result 4° (Tyler and Howard, 1973). Although values derived
corresponds to the dielectric properties of the surface from the 116-cm observations agree with the 13-cm
material averaged over a depth of roughly λ/8 (25 cm). data in the highlands, the mare values are only half
The reflectivity along groundtracks in Oceanus as large as those derived from the 13-cm data.
Procellarum showed quite reproducible enhancements Apparently, small-scale roughness is larger in the
of approximately 30% in echo intensity for mare maria.
regions compared to highland areas (Tyler, 1968b). Little work has yet been done on the lunar surface
These differences in reflectivity can be explained itself to verify the surface structures inferred from
either by variations in the packing fraction of the radar scattering behavior. A more technical discus-
regolith or by variations in its depth. The latter sion of these issues is found in Tyler (1979).
explanation appears more physically realistic if one Apollo Lunar Sounder Experiment (ALSE). The
also takes into account the results of Earth-based ALSE was a three-channel, synthetic-aperture, orbital
observations. radar experiment operated during three orbits of the
The Apollo 14 and 15 missions provided the Apollo 17 mission to infer subsurface electrical
opportunity to continue bistatic radar investigations conductivity structure, to profile lunar surface
at wavelengths of 13 cm and 116 cm (Tyler and topographic variations, and to image the surface along
Howard, 1973). The signal-to-noise ratio and the the spacecraft ground track (Phillips et al., 1973a,b).
surface resolution of these measurements were both In this experiment, a continuous series of short
an order of magnitude better than previous exper- bursts of radar energy at three frequencies (5, 15, and
iments of the same type. Although the extent of the 150 MHz) was directed from an instrument on the
datasets obtained was limited, the investigators found spacecraft toward the lunar surface, and the
regions of the lunar surface whose physical properties reflections, received by the same instrument, were
could not easily be explained by simple models. measured in terms of intensity, time of arrival, and
For λ = 13 cm, mare surfaces could be modeled phase. The very low electrical loss characteristics of
well using a value of the dielectric constant ε = 3.1, the lunar regolith permitted detection of echoes from
with few deviations observed. At λ = 116 cm, much of subsurface features as well as from the surface itself.
the mare could also be modeled simply with a similar Reduction of the ALSE data has been a very
dielectric constant, but in this case, considerable complex process, and analysis of the information has
deviations from the simple characterization been hampered by two fundamental problems
(Sharpton et al., 1984). First, the relatively weak
Physical Properties of the Lunar Surface 565
signals returned from subsurface reflectors are Thus far, reflecting subsurface layers have been
obscured by strong surface clutter. Second, time detected with confidence in both Mare Serenitatis and
delays associated with signals from subsurface Mare Crisium. The placement and extent of the
features cannot be unambiguously distinguished from reflectors beneath Mare Serenitatis have been used to
those associated with surface features lying some infer the geological history of the basin (see Fig. 4.31)
distance from the ground track. These problems have (Maxwell, 1978; Sharpton and Head, 1982),
been overcome to some extent in the data from flat illustrating the potential of the information in the
mare regions by correlating reflections from adjacent ALSE dataset.
ground tracks. This correlation allows discrimination Range data from the ALSE was used (W. E. Brown et
between subsurface layers and long linear surface al., 1974) to produce continuous elevation profiles
features, which coincidentally parallel the orbital path around the Moon. These profiles, which extended
and produce reflections at a distance from it (Peeples along nearly two complete lunar orbits, were
et al., 1978). constructed from sampling at intervals of about
Fig. 9.77. Temperature profiles across Aristarchus Crater, a large central-peak crater emplaced in a mare
region. Temperatures were determined from Earth-based thermal infrared measurements. Profiles originate at
the center of the crater and traverse the crater in the directions indicated (southwest, south-southwest, and
south-southeast). The higher pre-sunrise temperatures of the center indicate that much of the crater interior
consists of large blocks of exposed bedrock. The nighttime temperatures fall significantly along the crater floor
and are low but somewhat variable across the skirt of continuous ejecta around the crater. The ejecta deposits
that occur beyond the limit of near-crater continuous ejecta and as far out as three crater radii (60 km) from
the rim have smooth and uniform temperature profiles, unlike the irregular mare surface that is exposed
beyond the ejecta. From Schultz and Mendell (1978).
566 Lunar Sourcebook
45 m along the orbital ground tracks. Elachi et al. hot spots and the interiors of fresh craters, where
(1976) later used these elevation profiles to derive large blocks are exposed (Mendell, 1976). In a
diameter/depth ratios for small craters along the flight pulverized regolith, composed of small particles and
path and to study the topography of a few large having a large ratio of surface area to particle mass,
craters. They also published very accurate (better than absorbed heat is lost more rapidly than from large
±25 m in elevation) profiles of Mare Crisium and Mare blocks or from outcrops with a small ratio of surface
Serenitatis. area to mass. Winter (1970) showed that rocks
9.3.3. Thermal Infrared Astronomy smaller than 30 cm tend to have thermal losses just
like soil, but rocks larger than 10 m retain heat as if
The emission of thermal IR (λ = 3–15 µm) radiation they were solid outcrop.
from the darkened Moon provides information on heat The most significant result from the Apollo 17 IR
retention as the lunar surface cools during the two- radiometer is that its data can be used to provide
week-long lunar night. An IR scanning radiometer was estimates of minimum large-block sizes and abun-
operated in the orbiting command module during the dances in the centers of young craters. In
Apollo 17 mission. This instrument collected data on Aristarchus Crater, for example, the temperature in
lunar surface emissions in the wavelength range from the crater center just a few hours before sunrise is as
1.2 µm to about 70 µm. The Apollo 17 instrument was high as 120–130 K, whereas the crater walls and the
capable of resolution on the scale of a 2.2-km- surrounding plains are about 100 K (Fig. 9.77;
diameter circle on the surface immediately beneath Mendell, 1976; Schultz and Mendell, 1978). The
the spacecraft. However, since the collecting mirror on thermal retention in Aristarchus is too great to be
the instrument rotated at right angles to the flight modeled, even by 3-m blocks, and the thermal
path, the resolvable area degenerated into an ellipse behavior is more consistent with a model in which
toward the lateral horizons (Mendell and Low, 1975). about 11% to 16% of the crater interior is outcrop.
The useful scanned area was within a band of ±6° of Similar extreme temperature differences were
the command module’s orbital latitude. Compared to recorded in the central peak area of Copernicus
IR observations made from Earth, which have lunar Crater. In addition, some smaller fresh craters have
surface resolutions no better than 20 km and also heat-retentive rims (e.g., Bessarion A and B Craters;
suffer from atmospheric interference, the Apollo 17 Mendell and Low, 1975). Both these central and
data are an order of magnitude better (Low and marginal thermal anomalies disappear in older
Mendell, 1973). craters, and this loss of large blocky exposures or
The lunar nighttime surface temperature falls from outcrops that retain heat can be used as another
about 100 K at the antisolar meridian to about 90 K at measure of crater age (see Chapters 4 and 10).
the sunrise terminator. Superimposed on this general Thermal IR mapping may be a practical method for
nighttime cooling, however, are many “hot spots” that locating lunar outcrops of large blocks or solid rock.
cool more slowly and therefore retain more of their These data would supplement the information on
heat during the lunar night. There is a strong mineralogic composition that would be obtained from
correlation between the locations of these shorter wavelength near-IR spectroscopy (Chapter
10).
Physίcal Propertίes of the Lunar Surface 567
APPENDIX
SUPPLEMENTARY DATA ΟΝ LUNAR PHYSICAL PROPERTIES
(Tables Α9.1–Α9.16; Figs. Α9.1–Α9.9)
TABLE A9.1. Particle size distributions of Apollo 17 soils (after McKay et al., 1974), with graphical
size parameters (after Folk and Ward, 1957); (1) <1-cm data, (2) <1-mm data only.
568 Lunar Sourcebook
TABLE Α9.3. Particle shapes: Aspect ratio (after Görz et al., 1971, 1972).
TABLE Α9.11. Laboratory hand-penetrometer measurements on lunar soil samples from Apollo
11 (sample no. 10084) (after Costes et al., 1969, 1970a,b; Costes and Mitchell, 1970).
TABLE A9.13. Vacuum direct shear test data (Apollo 12, sample no. 12001,119)
(after Carrier et al, 1972b, 1973c).
TABLE A9.14. Miniature direct shear test data on returned lunar soil from
Apollo 12 (Surveyor 3 scoop sample) (after Jaffe, 1973).
580 Lunar Sourcebook
TABLE Α9.15. Miniature triaxial shear test data on returned lunar soil from
Apollo 12 (Surveyor 3 scoop sample) (sample no. 12029,8)
(after Scott, 1987, personal communication, 1988).
TABLE Α9.16. Summary of dielectric properties of lunar samples.
TABLE Α9.16. (continued).
TABLE Α9.16. (continued).
TABLE Α9.16. (continued).
Physical Properties of the Lunar Surface 585
Fig. A9.1. Results of Fourier analyses of measurements of the shapes of individual silicate particles in lunar
soil samples: (a) impregnated soil sample 15003,6059 from Apollo 15 drill core; (b) lunar regolith breccia
10065 from the Apollo 11 site (data from Kordesh et al., 1982; Kordesh, 1983a,b). The data represent results of
two-dimensional Fourier analyses on individual soil particles, treating the two-dimensίonal outline of each
particle as a polar-coordinate plot of radius and angle. In these plots, mean harmonic amplitude (i.e., the
contributίon of each harmonic to the particle shape) (vertical axis) is plotted as a function of harmonίc number
(horizontal axis). The shapes of lunar soil particles can generally be represented by harmonics of number <6,
similar to particles in meteoritic breccias, but distinctly different from partίcles altered by terrestrial media
such as wind and water.
586 Lunar Sourcebook
Fig. Α92. Depth relatίons for Apollo core tube samples, showing the divergence between measured sample
depth in the core tube (horizontal axis) and the actual depth to which the core was driven in undisturbed soil
(vertical axis) (after Carrier et al., 1971). If the core tube operated ideally, the data should fall along a line with a
45° slope. The divergence from this ideal line is caused by the fact that, as the tube is driven downward, it
gradually becomes plugged with material from shallow depths, until the deeper material can no longer enter. As
a result, the material present at a given depth in the core tube was actually collected from a shallower depth
than the depth to which the core tube was driven. (a) Apollo 11; (b) Apollo 12; (c) Apollo 14; (d) Apollo 15. “S/N”
refers to serial numbers of core tubes; all four-digit numbers in (c) and (d) are also serial numbers.
Physical Properties of the Lunar Surface 587
Fig. A9.3. Depth relations for Apollo rotary drill core samples, showing the divergence between measured
sample depth in the core tube (horizontal axis) and the actual depth to which the core was driven (vertical
axis). Data after Carrier (1974) (see also Allton and Waltz, 1980 for details). Because these drill cores
operated more ideally than the hammered-in core tubes (Fig. A9.2), the data fall closely along the ideal 45°
line. (a) Apollo 15; (b) Apollo 16: asterisk (*) indiates zone where sample moved during return to Earth,
leaving a 5-cm void in the bottom of core 60006 and disturbed material along the length of core 60005 (see
Allton and Waltz, 1980 for an alternate explanation); (c) Apollo 17: asterisk (*) indicates zone where sample
moved during return to Earth.
588 Lunar Sourcebook
Fig. Α9.4. Histograms showing values of relative density of lunar soil determined from measurements of
astronaut bootprints in the lunar surface soil (after Houston et al., 1972; Mitchell et al., 1973a,h, 1974). Data
include combined measurements from both intercrater areas and crater rims at the (a) Apollo 11, (b) Apollo 12,
(c) Apollo 14, (d) Apollo 15, (e) Apollo 16, and (f) Apollo 17 landing sites. At all sites, the lunar soils exhibit high
relative densities, mostly in the range 60–80%.
Physical Properties of the Lunar Surface 589
Fig. Α9.4. (continued) (e) Apollo 16, and (f) Apollo 17 landing sites. At all sites, the lunar soils exhibit high
relative densities, mostly in the range 60–80%.
Fig. Α9.7. Histograms showing ranges of calculated bearing characteristics of lunar soil, derived from cone
penetrometer measurements carried out by the Lunokhod 1 and 2 robotic roving vehicles on their surface
traverses (after Leonovich et al., 1974a). (a) Force on cone penetrometer required to drive the cone to a tip depth
of 4.4 cm; (b) friction angle, calculated from bearing capacίty theory and data from a basaltic sίmulant of lunar
soil; (c) cohesion, calculated from bearing capacity theory and basaltic simulant studies; (d) relative density,
calculated from data on a basaltic simulant. Note that the range of relative densities calculated in (d), generally
50–70%, is somewhat less than that (60–80%) obtained by other calculations using the same Lunokhod data
(Fig. Α9.6).
Physical Properties of the Lunar Surface 591
Fig. Α9.8. Results of individual penetration measurements made in lunar soil at the Apollo 15 landing site,
using the Self-Recording Penetrometer (SRP) (after Mitchell et al., 1972a,d, and unpublished data). Each plot (a–f)
represents a single penetration experiment carried out ίn a different location. Data are plotted as penetration
depth to the base of the cone (vertical axis) as a function of the stress applied by the astronaut (horizontal axis).
The variety of results from individual penetration experiments (a–f) shows that a significant range of soil
behavior exίsts at different locations, even within the single Apollo 15 landing site.
592 Lunar Sourcebook
Fig. Α9.9. Results of individual penetration measurements made in lunar soil at the Apollo 16 landing site,
using the Self-Recording Penetrometer (SRP) (after Mitchell et al., 1972b, 1973b). Each plot (a–j) represents a
single penetration experiment carried out in a different location. Data are plotted as penetration depth to the
base of the cone (vertical axis) as a function of the stress applied by the astronaut (horizontal axis). The variety
of results from individual penetration experiments (a–j) shows that a significant range of soil behavior exists at
different locations, even within the single Apollo 16 landing site.
Physical Properties of the Lunar Surface 593