Total Reflection Spectrophotometry and "l'hermogravimetric Analysis of Simulated Martian Surface Materials

m~Rus 4, 43-61 (1965)

Total Reflection Spectrophotometry and "l'hermogravimetric
Analysis of Simulated Martian Surface Materials
C A R L SAGAN 1
Harvard University and Smithsonian Astrophysical Observatory, Cambridge, Massachusetts

AND

J E A N P. P H A N E U F AND M I C H A E L I H N A T
Research and Advanced Development Division, Avco Corporation, Lowell, Massachusetts

Received December 1, 1964

The ultraviolet, visible, a n d n e a r infrared reflection spectra of the Martian bright
areas have been compared with the corresponding laboratory reflectivities, measured
with an integrating sphere, of a variety of minerals containing ferric oxides and sili-
cates, as solids and in pulverized form. Except in the ultraviolet, where the effects of
the Martian blue haze are prominent, pulverized limonite, a ferric oxide polyhydrate,
matches the shape and amplitude of the Martian Russell-Bond albedo within experi-
mental and observational error. Further observational tests of this identification a r e
outlined. If water-rich limonite is a primary constituent of the Martian bright areas,
conditions in the earlier history of Mars were probably much more equable than
contemporary conditions, and the origin and evolution of life on primitive M a r s
becomes easier to understand.

1. INTRODUCTION with the corresponding parameters of the
M a r t i a n bright areas. Sharonov stresses
T h e bright areas of the planet Mars,
t h a t the astronomical observations refer to
often described as deserts, have r e m a r k a b l e
M a r t i a n areas several hundreds of kilome-
photometric, eolorimetric, and polarimetric
ters across; accordingly, he averages the
properties. Their visible reflectivities are
laboratory measurements of a large number
low, their colors are ruddy, and their
of samples of a given mineral t y p e before
polarization phase curves show a striking
comparing with Mars. B u t since the present
negative branch for phase angles 4 ) ~ 23 °.
physical conditions and earlier history of
All three observables point to ferric oxides
the planets E a r t h and M a r s are dissimilar,
as a p r i m a r y constituent of the bright
there is no guarantee t h a t a terrestrial
areas.
averaging has M a r t i a n relevance. While
The geochemically most a b u n d a n t min-
m a n y categories of minerals were found to
erals with strong blue-absorbing ehromato-
m a t c h the M a r t i a n reflectivities, only one
phores are the iron oxides (e.g., hematite,
matched the bright areas of M a r s in color.
Fe20~); the ruddy color of M a r s has long
Of the 12 broad categories of terrestrial
been used as an argument for iron oxides
minerals surveyed, Sharonov found con-
on t h a t planet (see, e.g., Wildt, 1934). The
cordance with the M a r t i a n bright areas
visual reflectivities and color indices of a
only for the yellow ochre variety of the
wide range of common terrestrial minerals
mineral limonite. We emphasize t h a t ap-
have been compared b y Sharonov (1961)
parently only solid mineral samples were
1Alfred P. Sloan Foundation Research Fellow. used in Sharanov's studies.
43

44 e. SAGAN ET AL.

Limonite is the hydrated form of goethite, measurements of the ultraviolet refleetivity
Fe O(OH), also described by the empirical of Mars, these observations permit a fur-
formula Fe20~'H20. A common hydration ther, critical test of the hypothesis that
state of limonite is 2Fe2Oa'3H20, but the ferric oxides are the predominant constitu-
number of water of crystallization bonds ent of the Martian bright areas.
varies from sample to sample; in the pres- In the following discussion, we shall be
ent paper we will describe limonite as concerned only with explaining the proper-
Fe~O~-n H20, where n is a number, not ties of the bright areas. The dark areas
necessarily an integer, greater than unity. show other features altogether; while they
In addition to the water of hydration there have a reddish cast (Kuiper, 1957), they
is, characteristically, adsorbed water in any are not so red as the bright areas (Sharonov,
sample of limonite. On the Earth, limonite 1961); their visible albedo is substantially
rarely appears as a pure mineral. It is most smaller, and their polarization curve cor-
often mixed with goethite, hematite, and responds to some still unidentified material,
bauxite. In mixture with hematite and more strongly absorbing than pulverized
bauxite, it is called laterite; lateritic soils limonite (Dollfus, 1957). Both the reflec-
comprise a significant fraction of the tivity and the polarization are dependent
Earth's surface area (Sivarajasingham et on the local Martian seasons. A study of
al., 1962). the dark areas is most properly undertaken
A third, and most convincing, piece of in a biological context (Sagan, 1965a), and
evidence for ferric oxides in the Martian we will not pursue it further here.
bright areas is provided by polarimetry
II. EXPERIMENTAL PROCEDURE
(Dollfus, 1957; 1961). The polarization-
phase angle curve of the Martian deserts To explain the infrared and ultraviolet
has a negative branch for 0 ° ~ • ~ 23°; photometric results for the Martian bright
positive polarizations prevail between 23 ° areas, we undertook a program of diffuse
and 43 °, the maximum phase angle for reflection spectrophotometry of selected
Earth-based observations of Mars. The terrestrial minerals. Eight mineralogic-
maxima, minima, and zeros of the curve ally distinct samples of limonite, hematite,
point strongly to polarization due to scat- and goethite, as listed in Table I, were
tering by strongly absorbing particles of obtained from Ward's Natural Science Es-
small dimensions--characteristic diameters tablishment, Rochester, New York.
are 100 ~ and smaller. Of the many hun- Mineral fragments were pulverized with
dreds of samples investigated by Dollfus an impact mortar and passed through a
(1957) in the laboratory, only finely pul- 40-mesh screen. The powder from a given
verized limonite matched the polarization sample was then homogenized and a frac-
curve of the Martian bright areas. tion removed for a rough determination of
One additional argument in favor of particle size distribution. The fraction was
limonite on Mars has been provided by suspended in low vapor pressure immersion
Adamcik (1963), who notes that the equi- oil, placed between a glass slide and cover
librium partial pressure of water vapor slip, and examined with a Unitron Model
above limonite at typical Martian daytime MMA monocular metallurgical microscope,
temperatures is just that anticipated fitted with a Bausch and Lomb 10X objec-
earlier, on theoretical grounds (Sagan, tive filar micrometer calibrated in milli-
1961), and subsequently verified spectro- meters. In transmitted light, the particle
scopically (Spinrad, Miinch, and Kaplan, size varied considerably from point to
1963). point on the objective slide, due, e.g., to
In recent years, a series of infrared clumping and nonuniform crushing. In an
observations of Mars has been performed effort to obtain a representative mean par-
from the ground (Sinton, 1959; Moroz, ticle size, 20 randomly selected particles
1964) and from balloon altitudes (Daniel- were measured for each pulverized sample.
son et aI., 1964). Together with older The particles were selected by moving the

TABLE I
PROPERTIES OF FERRIC OXIDE SAMPLES
Particle diameter
Maximum Minimum Mean TGA temperature range TGA fractional
No. Sample Source State (microns) (microns) (microns) ¢'C) mass loss

Limonite Tuscaloosa, Ala. Powder 58O 3.4 127
TGA residue 370 4.2 131 30-350 ° 10.0% H20
Hematite Ironton, Minn. Powder 26 2.4 I0
TGA residue 140 2.4 21 30-1000 ° <0.3%
Geothite Biwabik, Minn. Powder 58O 17.0 130
TGA residue 58O 17.0 180 30-400 ° 9 . 8 % H,O
Fossiliferous Rochester, N.Y. Powder 360 2.5 79 t~
hematite TGA residue 68O 1.7 140 500-850 ° 18.8% COn2-
Oolitic Chilton, N.Y. Powder 310 17.0 74 *q
hematite TGA residue 400 100.0 130 500-850 ° 19.0% CO~I-
Yellow ochre CartersviUe, Powder 290 17.0 71
limonite Ga. TGA residue 1500 17.0 290 200-450 ° 6 . 6 % H~O
Specular Ishpeming, Powder 100 1.8 28
hematite Mich. T G A residue 330 5.1 63 30-1000 ° <o.5%
Limonite Pelican Pt., Powder 390 4.2 140
pseudomorphs Utah TGA residue 650 6.6 220 30-400 ° 8 . 4 % CO2
after pyrite

¢.u

46 C. S A G A N E T AL.

microscope stage at random intervals and ochre hue, while the hematites ranged from
measuring that particle which happened to claret to brownish purple. Samples which
fall on a preselected location on the cross had been heated to high temperatures dur-
hair of the filar micrometer. Maximum, ing TGA took on colors characteristic of the
minimum, and mean particle diameters, unheated hematite samples, but usually
obtained through this procedure, are listed darker. This circumstance may be attrib-
in Table I. The sizes fall within the range uted to the loss of water of crystallization
of those deduced on polarimetric grounds from the limonite and goethite samples
for Mars. during heating (and possibly to a change
For each pulverized mineral of Table I, in oxidation state) ; with the water removed,
a sample was taken through thermogravi- they have been converted to hematite. The
metric analysis (TGA). A Harrop TGA colors of the pulverized samples were al-
apparatus was employed, coupled to a Seko ways more vivid than the eolors of the
recording semimicro analytical balance, unpulverized minerals, either in direct com-
having a sensitivity of 5 X 10-5 gm. Sam- parison, or when streaked on paper. No
ples were suspended i n air beneath the solid sample of yellow ochre limonite,
balance, in a furnace wound with (plati- Variety 6, was available.
num-20% rhodium) wire. The furnace Ultraviolet, visible, and infrared spectra
temperature was increased at a rate of were obtained for monochromatic light
20 C°/min and the fractional mass loss specularly reflected, diffusely reflected at a
measured as a function of temperature fixed angle, or reflected into all 4 ~ ster-
between 30 ° and 1000°C. Bound water is radians from the sample, using a specially
characteristically lost at 350-450°C; carbon constructed integrating sphere on a modio
dioxide evolution from carbonates charac- fled Beckman DK-2 ratio-recording spec-
teristically at 750-850°C. The TGA resi- trophotometer. Sunlight reflected from the
dues were then examined microscopically Martian surface is generally observed at a
by the same technique used for size deter- range of diffuse reflection angles. The re-
mination of the pulverized samples. The flectivity into 4 ,-r sterradians is of interest
results on TGA residue sizes are also dis- if we wish to compare laboratory spectra
played in Table I; we note that the TGA with Russell-Bond albedos. For Mars, the
residue particle sizes tend to be somewhat spherically integrated albedo in the visible
larger than their precursor powders, prob- is approximately equal to a typical diffuse
ably due to high-temperature sintering. reflectivity observed at the same wave-
The original solid samples were, in many lengths; the infrared phase integral is so
cases, mineralogically complex. Limonite, large, however, that tile wavelength-
Variety 1, for example, had faces speckled integrated diffuse reflectivity is about 25%
with small yellow grains; but other faces less than the bolometric Russell-Bond
were smooth and very dark. Thus, observa- albedo (de Vaucouleurs, 1964). To mini-
tions performed on a solid sample of a mize any serious dependence of our results
given variety of mineral were much more on the detailed scattering phase functions,
variable from sample to sample (i.e., face it is the spherically integrated reflectivity
to face) than were observations performed which we have principally measured in the
on various samples of the same mineral laboratory. Above 6000A, these measure-
when pulverized. ments should, therefore, be compared only
Bright colors characterized the pulverized with the corresponding Russell-Bond al-
samples. Limonites ranged in color from bedo of Mars. [Note that the astronomical
dull gray-brown, through bright mustard, and laboratory measurements are still not
to orange ochre, the color of limonite pseu- quite strictly comparable; the integrating
domorph after pyrite, Variety 8, which sphere sums all angles of reflection, but not
seemed, of all the samples, to bear the all angles of incidence. However, the effect
closest resemblance to the Martian bright of this difference will be minor; the polar-
areas. Goethite samples were of yellow ization of sunlight reflected from Mars (and

MARTIAN SURFACE MATERIALS 47

from the Moon) is essentially independent oxide plate with a 5-degree optical wedge
of surface inclination (Dollfus, 1957).] inserted to measure total reflectance (spec-
The spectrometer has a useful wavelength ular and diffuse reflected components). The
range between 2000A and 2.7 ~. The wave- sample path contains the sample to be
lengths are calibrated to an accuracy of measured, and, as in the case of the ref-
approximately 40A in the ultraviolet, 15 A erence sample, a 5-degree wedge to measure
in the visible, and 0.012 ~ in the near infra- total reflectance. Alternation of the light
red. Periodic verification of the wavelength beam provides ratio recording of the sam-
calibration was made with fixed bandpass ple reflectance and the reference magnesium
filters. The apparatus, including the inte- oxide plate. The detector, either a lead sul-
grating sphere, was continuously purged fide cell or photomultiplier (K), measures
with dry nitrogen to minimize residual at- the integrated reflected light within the
mospheric absorption in the light path and magnesium-oxide-coated integrating sphere
to protect the optical components. (L). AneilIary to the integrating sphere is
The optical path of the DK-2 spectropho- a first reflection shield at the detector open-
tometer is illustrated in Fig. 1. An image of ing. The first reflection shield, a tube

B

I~G. I. Schematic diagram of modified Beckman DK-2 spectrophotometer, with integrating sphere.
See text for details.

the light source, either a tungsten or hydro- approximately 2 cm in length, prevents
gen lamp (A), is focussed on a condensing erroneous results when highly specular
mirror (B) and then on a 45-degree mirror components are present.
(C). The light then passes through the Absolute spectral reflectance data, as
entrance slits at (D) into a Littrow mono- presented in this study, were obtained by
chromator. Light falling on the collimating correcting the measured sample reflectivi-
mirror (E) is rendered parallel and re- ties for the reflectivity of the magnesium
fleeted toward the prism (F). The back oxide comparison plate. The absolute re-
surface of the prism is aluminized so that fleetivity of freshly prepared magnesium
light refracted at the first surface is re- oxide surfaces has been measured by
fleeted back through the prism, and under- Edwards et al. (1961). The magnesium
goes further dispersion. The collimating oxide surfaces for the reference sample and
mirror (E) then focuses the spectrum in integrating sphere are prepared by burning
the plane of the slits. Light of the wave- magnesium wire and depositing thicknesses
length for which the prism is driven is of at least 6 mm on the surfaces.
transmitted out of the monochromator Solid mineral samples were placed di-
through the exit slit to the oscillating mir- rectly in the sample port; however, for
ror (G). The oscillating mirror passes the powdered samples, a quartz holder was
light beam alternately to mirrors (H), first used. In these cases, approximately 3 mm
along the reference path (HI), then along thickness of powdered magnesium oxide
the sample path (H J). The reference path and sample were used in the reference and
contains a freshly prepared magnesium sample beams, respectively, thereby corn-

48 c. SAGAN ET AL.

90
1 l t J I I I ' [ I I

80-

70- ~ " ~"
•f '~'/ ". ~-
iSC- /
II1

,., / ~ ",
,J. 4 C - I I ,, -
.. I I " "
-3c- /~,,, //I / SOLID SPECIMEN
2G - / " . " " - ~'~ I ---- POWDERED SPECIMEN
/ ./ " I ..... POWOERED RESIDUE TGA 1200eC

C I I 1 I l ]
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6
WAVELENGTH.microns
Fro. 2. Total reflectance of solid, powdered, and TGA residue forms of limonite, Variety 1, Tusca-
loosa, Alabama.

I 1 I 1 I I r I I F I

8¢
•J
7C

i
ILl
u i
i
h$
j,
W
I 8PEC,.EN I-I
/ I - - - eowoeeEo SPECIMEN II
l:'J I . . . . . ~ o . o e . e o ~SIO.E T6A 1200"C I--I
~." ",.Y 'I
IC--

C I
O.2 0.4 0.6 0.8 1 . 0 I.Z 1•4 1•6 1•8 2•0 2•2 2.4 2.6
WAVELENGTH,microns
Fro. 3. Total reflectance of solid, powdered, and TGA residue forms of hematite, Variety 2, Ironton,
Minnesota•

pensating, in this ratio-recording instru- 2-9. The corresponding TGA plots are
ment, for the effect of the quartz holders given in Figs. 10-17. For the yellow ochre
common to both beams. limonite, Variety 6, no solid samples were
available. For all spectra, we note t h a t the
III. D~scussm~ OF RESULTS total refleetivities of pulverized samples
The infrared spectra of solid, pulverized, and of powdered T G A residues exceeded at
and powdered TGA residue samples of the all wavelengths the corresponding reflectivi-
minerals of Table I are displayed in Figs. ties of the solids. This result is to be ex-


9Q I I t I I i I I I I I

"/~8C- I . . . . . SOLID SPECIMEN I -
_ _ _ POWOERED SPECIMEN
POWOEREO RESIOUE TGA 1200"C _j/
°......,i~.~" - -

,., /..
Z

i'f ........ - ' " , 2
W
E
./
g

o 2C- S"',,-i /

G2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6
WAVELENGTH,microN
Fro. 4. Total reflectance of solid, powdered, and TGA residue forms of goethite, Variety 3, Biwabik,
Minnesota.

I I I I I ! I I I I I

! os, ~°
/ m
bl
¢J
/

,.J~-
E
J3C"
SOLID SPECIMEN

10-
I _ _ _ POWI)EREO SPEGIMEN
. . . . . POWDEREO RESIDUE TGA 1200eC
I
o I I I I I I I I I I I
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 i.8 2.0 2.2 2.4 2.6
WAVELENGTH,microns
FIG. 5. Total reflectance of solid, powdered, and TGA residue forms of fossiliferous hematite, Variety
4, Rochester, New York.

peered; for a given mass of absorber, the tion. Accordingly, the particle phase func-
more interfaces encountered by a random- tions must have a somewhat pronounced
walking photon, the greater the likelihood forward-scattering lobe. Thus, a back-
of back-scattering. The particles of Table I, scattered photon has a significant proba-
like the particles in the Martian bright bility of escape from the particle array.
areas (Dollfus, 1957), have radii a > > )t/ Many of the spectra exhibited a reflec-
2 7r, where X is the wavelength of observa- tivity decline longward of 2.5 or 2,6 ~. Since

50 C. SAGAN ET AL.

9¢
I i J J I I I I I I I |
8£ souo SF~CIMEN l .1-~
___ POWDERED SPECIMEN I /./ |
- 70 . . . . . POWOEREO RES,OUE TGA P,O0"C I f J" -~
. , t "~" -
...,~...,...-~" f J
Z
./"
W
,.-, 4 0 - / -
n..
UJ

J30--
<[
I-- /
~ - /
/
I0-

0 I I I 1 I I I Z t I I
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6
WAVELENGTH,microns
FIG. 6. Total reflectance of solid, powdered, and TGA residue folTns of oolitic hematite, Variety 5,
Chilton, New York.

90 [ I r 1 I i I I I 1 t |

80- J

bJ
¢)

~ 50-

" SY'
.,/.s POWDERED SPECIMEN
2(: J -"----POWDERED RESIDUE TGA 1200"C

IC-

e J I I I I I I t I I f
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6
WAVELENGTH,microns
Fro. 7. Total reflectance of powdered and TGA residue forms of yellow ochre limonite, Variety 6,
Cartersville, Georgia.

Figs. 2-9 terminate at 2.6 #, these declines shows the presence of both folzns of wa~er
are not in all cases evident in the figures; in the samples. For example, in Fig. 10, the
they can be attributed to absorption b y slow mass decline at 30°C ~ T ~ 250°C is
water, either present as water of crystal- a consequence of the loss of adsorbed water.
lization, or adsorbed in the sample. T h e r - The abrupt mass decline at T ~ 300°C is
mogravimetric analysis of limonite and characteristic of the loss of water of hydra-
goethite (Figs. 10, 12, 15, and 17) clearly tion. At higher temperatures, smaller quan-


9O
[ F ]

80-- SOLID SPECIMEN
___ POWDERED SPECIMEN
70-- ..... POWDERED RESIDUE TGA 1200°C

~ -
W
j ~.---- J

~ 5C--
I.-
w
~ 4(]--
W

-J 30--
I,-
o
i- 20--

I0 ~ V
I I I I [ I I J i I I
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6
WAVELENGTH,microns
Fro. 8. Total reflectance of solid, powdered, and T G A residue forms of specular hematite, Variety 7,
Ishpeming, Michigan.

9c
T 1 ~ r ! T ~ 7
8C-
/
z~~.~"~" "~'~'"~
/ '~'~..~ ~. ~ ./ --
-c 7O / ~......~.'~"

W

Z
/ -- -,,,..
(.)
W

W
tE
40
./
/]/I SOL|D SPECIMEN
_ _ _ POWDERED SPECIMEN
-J3C f'~-.J / I ..... POWDERED RESIDUE TGA 1200°C
~-//-,,_~/I [
h-
o

0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6
WAVELENGTH,microns

FIG. 9. Total reflectance of solid, powdered, and T G A residue forms of limonite pseudomorphs after
pyrite, Variety 8, Pelican Point, Utah.

titles of both adsorbed water and water of tinctive carbonate absorption features ap-
crystallization are lost, and the mass of the pear in the spectra of these hematites
sample asymptotically approaches a con- (Figs. 5 and 6) ; they are expected at longer
stant value. I n Figs. 13 and 14, the abrupt infrared wavelengths. The identification of
decline in mass at T ~_ 800°C is due to the the 3 ~ feature of limonite and goethite with
loss of CO,; oolitic and fossiliferous hema- water absorption is borne out by the gen-
tite are both rich in carbonates. No dis- eral absence of such features in the spectea

52 C. SAGAN ET AL.

O~j. I --- T T --T 1
0.02[- ~ ' ~

LL
0.12

0.14

O.l~ I i I J I J
I00 200 300 400 500 600 700 800 900 IOOO
TEMPERATUREIN°C
Fro. 10. Thermogravimetric analysis of pulverized limonite, Variety l, Tuscaloosa, Alabama, per-
formed in air, heating rate 20 C°/min.

¢./) 0
u)
o,
u)
0.02-
=E
.J

Z
o 0.04-
F-
o
Iv

b. 0"06 i I I I I I I I
I00 200 300 400 500 600 700 80O 900 1000
TEMPERATURE IN "C

Fro. 11. Thermogravimetric analysis of pulverized hematite, Variety 2, Ironton, Minnesota, per-
formed in air, heating rate 20 C°/min.

of dehydrated TGA residues of the same - - i n Figs. 2-9, a feature which must be
samples. Water vapor has a well-known ascribed to absorption in the Fe208 moiety
absorption band at 2.7 t~. A better apprecia- of these minerals. The upper boundary of
tion of the depth of the 3 t~ limonite absorp- the shaded area has been smoothed to
tion feature may be gained from the figures remove the 0.85 tL and other features, for
of the accompanying paper by Hovis reasons of clarity. Beyond 2.7 ~, the curve
(1965). is due to Hovis (1965). Spectra of solid
A typical range of the monochromatic samples of limonite, goethite, or hematite
total reflectivity (integrated over all solid tend to fall below the shaded area. A com-
angles) of pulverized limonite and goethite bination of solid and powdered samples will
is displayed by the shaded area of Fig. 18. have lower albedos than the powdered sam-
The lower boundary of this area cor- ples alone. Individual solids will have a
responds to the reflectivity of pulverized wide range of albedos, depending on their
limonite, Variety 1. I t exhibits the charac- surface granularities. For these reasons,
teristic 0.85g absorption feature of all even if an unequivocal demonstration were
samples--both powders and TGA residues forthcoming that limonite is the principal


I I I I l
0.02
~o
t/3
Oj 0.04
¢/3

.J 0 . 0 6
z
9
I..-
U
"~ 0.08
12:

0.10

ol2 I ] i I I I I 1
I00 200 300 400 500 600 700 800 900 I000

TEMPERATURE IN * C

Fro. 12. Thermogravimetric analysis of pulverized goethite, Variety 3, Biwabik, Minnesota, performed
in air, heating rate 20 C°/min.

01 - - 0
I'~,,, I I
0 05 -- 0.02

00~ --
0.04

jO OOE
o 0.06

..J OOE
o i 0.05
~3OIC
OJO
LL
Oi .~ --
0.]2

014 --
L_
0,14 - -

OIE I I I I I
I00 200 300 400 500 600 700 800 800 I000
TEMPERATUREIN=C O.IE I I I I I ~ I
200 300 400 500 600 700 500 900 I000
TEMPERATURE IN °C
:FIG. 13. Thermogravimetric analysis of pul-
verized fossiliferous hematite, Variety 4, Roches- FIG. 14. Thermogravimetrie analysis of pu[-
ter, New York, performed in air, heating rate verized oolitic hematite, Variety 5, Chilton, New
20 C°/min. York, performed in air, heating rate 20 C°/inin.

constituent of the M a r t i a n bright areas, it obtained by Moroz (1964) do not fall
would be difficult to determine the grain within the shaded area of Fig. 18, even
size of the dust from albedo observations when corrected for the probable infrared
alone. phase function of Mars, probably because
L a b o r a t o r y infrared spectra of limonite Moroz' samples were not pulverized. His

54 C. SAGAN ET AL.

I I I 1 I

002 --
o r)
o3
o
.J
~o
¢n 0 , 0 4 -
¢Z
_J

0.06--
p-
o
¢¢
h_
0.08 - -

o,o I I I I I I l I
,oo 200 ~oo 400 ~oo 600 700 800 900 ,ooo
TEMPERATURE IN °C

Fro. 15. Thermogravimetric analysis of pulverized yellow ochre limonite, Variety 6, Cartersville,
Georgia, performed in air, heating rate 20 C°/min.

0
I I I - 7 --t-- I I

,~ 0.02
J

z
0 0.04
I-
o
a,
0.06 1 I 1 I 1 I I
I00 200 500 400 500 600 700 800 900 I000
TEMPERATURE IN *C

Fla. 16. Thermogravimetric analysis of pulverized specular hematite, Variety 7, Ishpeming, Michi-
gan, performed in air, heating rate 20 C°/min.

spectra agree well with those of solid sam- mate continuation from an estimate by
ples in Figs. 2-9. A single spectrum between Woolf (1965), based in part on Stratoscope
0.7t~ and 2.3~ of a mixture of pulverized II infrared spectra, and compatible with
goethite and hematite (at a fixed phase earlier observations by Kuiper (1952; see
angle) by Draper, Adamcik, and Gibson also Harris, 1961). We have converted
(1964) agrees well with our data for pul- Woolf's estimate of p~ to A~ by setting the
verized samples. infrared phase integral at ~ > 1 ~, q~,.
The mineral reflection spectra are com- 1.4 (de Vaucouleurs, 1964). The curve for"
pared in Fig. 18 with the observed mono- wavelengths X > 3t~ is from a rough esti-
chromatic spherical albedos of Mars within mate by Sinton (1965; see also Sinton,
the best current knowledge. In the ultra- 1959), similarly corrected by q~,---- 1.4. We
violet, visible, and near infrared, the Rus- should note that most of the observations
sell-Bond albedos, Ax =-pxqx, are taken leading to the solid and dashed line of
from de Vaucouleurs' (1964) compilation; Fig. 18 were performed on the entire plan-
here px is the geometric albedo (~ = 0°), etary disk. In the present paper, we are
and qx is the phase integral. For 0.8 ~ X concerned exclusively with the Martian
3.0t~, the dashed curve shows an approxi- bright areas; unfortunately, the available


°~,~1 I I I I I I
°°2~ ~
i 0.04--
o.oo-
0.08 --

o,c I 1 I 1 1 I ] "1--"
,oo 200 300 400 500 600 7oo 800 900 I000
TEMPERATURE IN °C

Fro. 17. Thermogravimetric analysis of pulverized limonite pseudomorphs after pyrite, Variety 8,
Pelican Point, Utah, performed in air, heating rate 20 C°/min.

I I I I I i i ~ I I I i I I I
~Typical smoothed laboratory olbedos of
/pulverized limoaite and hematite,
samples
0.60 /'Observed Martian albedo ~ various

x~,~,~ig,~~~ / (schematic -- Wooff or~d Sinton)
0.50

of Moroz are denoted by ~)

0.40

0.50

020

010

t J I I I IIrl I I I I I I
0.1 0.2 0.5 0.5 0.7 1.0 2.0 5 0 4 0
Wavelength in microns -'-'=

FIe. 18. Comparison of measured and estimated Russell-Bond monochromatic albedos of Mars with
laboratory total reflectances of pulverized ferric oxides. See text for details.

data apply to some unspecified linear com- unalloyed Martian bright area will lie
bination of bright and dark area albedos. somewhat above the heavy curve of Fig. 18.
I f we assume that the dark areas have Also plotted in Fig. 18 are four reflectivi-
lower refleetivities in the infrared as well ties measured by Moroz (1964), and cor-
as in the visible (el. Sinton, 1959), it fol- rected to spherical albedos using qtr ~-~ 1.4.
lows that the infrared reflectivities of an These points are in clear disagreement with

56 c. SAGAN ET AL.

the dashed line. We believe that calibration inconsistent with our ultraviolet reflection
errors have crept into Moroz' discussion; spectra of pulverized limonite and goethite
his values of px in the 1-2 ~ region do not (Figs. 2 and 4). Solid samples have even
join smoothly with the observations at lower reflectivities. No reasonable source of
slightly shorter wavelengths (cf. de Vau- Rayleigh or fluorescent scattering has been
couleurs, 1964), corresponding to the thick suggested which provides a high albedo.
solid line of Fig. 18. To derive px in the Despite a recent confirmation of the Jovian
3-4 ~ region, Moroz assumes px independent ultraviolet albedo obtained on the same
of ~, as his laboratory spectra of solid rocket flight (Stecher, 1964), the original
limonite samples suggest. This assumption experimenters give very low weight to the
is equivalent to the statement that a typical high Martian ultraviolet albedo (Boggess,
mean temperature of the observed disk of 1965), and we conclude that no discrepancy
Mars is T = 290°K, a serious overestimate. between the ultraviolet albedo of Mars and
With more reasonable temperatures, Sinton the limonite identification is apparent at
(1959; 1965) derives a reflectivity increas- the present time. Further ground-based and
ing sharply from 3 to 4 ]~; and Hovis (1965) rocket-borne observations of the Martian
finds the same behavior for a sample of ultraviolet albedo--particularly during mo-
pulverized limonite. ments of blue clearing--would be of great
We see from Fig. 18 that between 5500-~ value.
and 2.2~ the albedos of pulverized ferric The steep slope of the Martian albedo
oxides in the 10-100~ size range and the in the visible and near infrared (the thick
albedo of Mars are in good general agree- solid line of Fig. 18) is a significant com-
ment. For ~ > 2.2~, the shapes of the position index. It is characteristic of ferric
Martian and limonite albedos are closely oxides and of very few other materials. In
similar, but there is only marginal agree- Figs. 19 and 20 are displayed the total
ment on the magnitude of the albedo. But reflectances of solid samples of the common
we are here comparing a single laboratory silicates olivine [(Mg, Fe)2Si04] and ser-
measurement (Hovis, 1965) with one ap- pentine [3MgO'2SiO2"2H20]. Pulverized
proximate estimate of the Martian albedo; samples have the same form, but with
more laboratory and observatory measure- amplified refiectivities. The two curves of
ments are clearly needed. Fig. 20 correspond to two different faces of
At X < 5500A, the albedos of ferric ox- the same serpentine sample, and illustrate
ides systematically fall above the albedos the difficulties in spectrum reproducibility
of Mars. This is an indication of the of heterogeneous or conglomerate samples.
existence of the Martian blue haze. Let us We see that silicates do not exhibit the
assume that limonite is the principal con- steep albedo slope of Mars--in accord, of
stituent of the bright areas. The obseura- course, with the common observation that
tion of Martian surface detail in the blue- silicates are not commonly colored red.
violet is then not attributable to a decline In Fig. 21, a serpentine sample suffi-
in contrast between bright and dark areas-- ciently small to be partially transparent
a purely surface effect--but must, instead, was prepared; total reflectance and trans-
be due to extinction by some component of mittance could then be measured, so the
the Martian blue haze (cf. Kellogg and total absorptance could be computed. We
Sagan, 1961; Sagan and Kellogg, 1963; see that the absorption coefficient tends to
Sagan, Hanst, and Young, 1965). A mar- be low in the near infrared, and higher at
ginal rocket observation by Boggess and longer and at shorter wavelengths, cor-
Dunkelman (1959) suggested a geomet- responding, respectively, to vibrational and
rical albedo at 2700A, px ~0.24. The electronic transitions. Indeed, the absence
phase integral of Mars at 3200A is qx of absorption features in the red and near
--~0.96 (de Vaucouleurs, 1964), and is infrared is a general property of most
unlikely to be much smaller at 2700A. A minerals, and leads us to predict a near
spherical albedo Ax ~ 0.23 at 2700A is infrared reflectivity maximum for most


9C
I I I J I I ] I; I~ I: I
• 8C

c 7C

bJ
Z

W
4c

-JSC
o
*--20
i

0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6
WAVELE NGTH,micronl

Fro. 19. T o t a l reflectance of olivine.

I I I I I I 1 I I I [
80 __ ISI" -SURFACE NO.I I
$2 -SURFACE NO.2
70

~60
~J
¢J
z
0
W
40

-30
I-
S2
/
I0 - / $ I "~" . . . .

o I i I I I i I I l l I
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6
WAVELENGTH,micron=
I~G. 20. T o t a l reflectance of t w o faces of the s a m e s a m p l e of a serpentine conglomerate.

minerals, as is indeed observed. It is the reflectivity of felsitic rhyolite and Mars.
steepness of the slope, rather than the The measured spherical albedo of a solid
wavelength of the albedo maximum, which sample is exhibited in Fig. 22. The increased
is; diagnostically useful. reflectivity of pulverized samples will place
k silicate which does display a steep this material within the shaded area of
albedo slope is felsitic rhyolite, generally a Fig. 18, confirming Kuiper's (1952) conclu-
conglomerate of free silica and alkali feld- sions. Yet the similarities between the al-
spar of volcanic origin. Kuiper (1952) first bedo of felsitic rhyolite and, for example,
drew attention to the similarity in infrared yellow ochre limonite, Variety 6 (Fig. 7),

58 C. 8AGAN ET AL.

9C
I" I' I I I I I I I I ~

NCE
7C

6C

5G
I-
z
=
o 4c -

el
3C- t-.. TOTAL
/ / "~-~
TRANSMITTANCE
2¢
- TOTAL /'"~j/ --
_ REFLEC.TANCE/--7"~_" -.t.--...

C' 1./ I f I I I I ] ~'1 -~.21--'~
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6
WAVELENGTH,microns

Fro. 21. Measured t o t s ] reflectance and transmittance and deduced absorpts~ce of a serpent~ae
sample.

90
I ~ I I I I t I I I I I
.8O

-=

Z
~5o

~ 30

~2o

I0

I ! I I I I I I I I I t
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 26
WAVELENGTH, micron==
Fza. 22. Total reflectance of felsitic rhyolite. The steep slope towards the infrared is due to ferric
Oxide impurities.


e.g., the 0.85-0.90 ~ feature, and the decline near 1.4~. Such a feature does exist in
at:. X ~ 2.6~, are not accidental. Ferric limonite pseudomorph after pyrite (Fig. 9),
oxides are often present as a dusty inclusion but Moroz' observations need confirmation.
in the feldspar component of felsitic rhyo-
lite, ancl are responsible for the red color IV. CONCLUSIONS
of the mineral (see, e.g., Chapman, 1960).
It is difficult to avoid the conclusion that Within the range of observational uncer-
ferric oxides are prominent on Mars. tainty, laboratory diffuse reflection spec-
BUt even if ferric oxides are prominent, trophotometry' does tend to support the
need they be a major constituent of the identification of powdered ferric oxides as
Martian surface? Arguing from terrestrial a principal constituent of the Martian
analogy, Van Tassel and Salisbury (1964) bright areas. Pulverized limonite pseudo-
h~.ve argued that silicates must be the morph after pyrite shows the 1.4 ~ Martian
predominant constituent, with ferric oxides feature tentatively identified by Moroz,
present as an impurity. The spectra of and also has a hue and brightness, closely
felsitie rhyolite are not inconsistent with resembling that of the visual image of Mars
this conclusion. However, the polarimetric at the telescope. Mixed with other varieties
evidence (Dollfus, 1957; 1961) argues per- of goethite, limonite, and hematite, a close
suasively for the predominance of ferric fit to the well-observed Martian albedo can
oxides. While large quantities of silicates be generated.
transparent at 2000 A ~ X ~ 2.7 ~ make It is highly unlikely that the Martian
little contribution to the total reflectivity bright areas are covered by a pure mineral.
of Mars, the same is not true for polariza- If limonite is present, we expect goethite
tibn, which occurs, through Fresnel's laws, and hematite in varying proportions. On
at silicate as well as at limonite interfaces. Earth, limonite: appears commonly mixed
The. polarization phase curve of silicates with hematite and bauxite in lateritic soils.
does not resemble that of Mars (Dollfus, Lateritic soils on: Earth are produced in
1957; 1961). In particular, large amounts high-humidity tropical environments and
of rhyolite can be excluded on polarimetric generally require free oxygen for their oxi-
grounds (Dollfus, 1964). While a Martian dation (Sivarajasingham et al., 1962); the
surface of limonite-coated siliceous rocks high oxidation state and water content of
cannot be excluded, it does not provide limonite point to similar conditions for its
a very striking economy o f hypothesis. formation. The present environment of
Explanations have been proposed for the Mars is characterized by low temperatures,
source of a surface layer of ferric Oxides on low water contents, and low oxygen partial
Mars (Sagan, 1965a). We are forced to pressures. We may provisionally conclude
conclude that the Martian bright areas, a t that the presence of limonite on Mars
least down to the depth observed at infrared points to an earlier epoch in which the
frequencies, are largely ferric oxide. environmental conditions more nearly re-
Future astronomical and laboratory spec- sembled the present environment of Earth.
trophotometry could put this identification This conclusion is in accord with the views
on a firmer basis. Almost all ferric oxide of Wildt (1934), who proposed t h a t most of
samples show a reflection minimum in the the oxygen complement of an earlier Mar-
0.85-0.90~ region; there is no published tian atmosphere is now combined in the
photometric investigation of the Martian Martian crust. The source of the extensive
bright areas in the same wavelength inter- quantities of reduced iron required remains
val. Some of the pulverized samples showed to be discussed; the iron may be of mete-
a-pronounced reflection minimum near oritic origin and related to the proximity of
1.9~; Mars has yet to be investigated Mars to the asteroid belt (Sagan, 1965a).
photometrically at 1.9#. Moroz: (1964) The possibility of warmer and more humid
found a relative minimum in the light re- times in early Martian history increases the
flected from the integrated disk of Mars probability of the origin of life on primitive

60 C. SAGAN ET AL.

M a r s (cf. K e l l o g g a n d S a g a n , 1961, p. 33;Evolution of Atmospheres and Oceans" (P. J.
S a g a n , 1965b). Brancazio and A. G. W. Cameron, eds.), p. 298.
T h e possible p r e s e n c e of l i m o n i t e on John Wiley, New York.
DRAPER, A. L., ADAMCIK,J. A., A/qn GIBSON, E. K.
M a r s is also of r e l e v a n c e for p r o b l e m s of
c o n t e m p o r a r y Martian• biology. F o r e v e r y (1964). Comparison of the spectra of Mars and
Fe203 m o i e t y in limonite, t h e r e is m o r e a goethite-hematite mixture in the 1 to 2 micron
region. Icarus 3, 63-65.
t h a n one m o l e c u l e of w a t e r . T h e M a r t i a n
EDWARDS,D. K., GIER,J. T., NELSON,K. E., A/qD
s u r f a c e m a y be c o m p o s e d of as m u c h as
RODDICK, R. D. (1961). Integrating sphere for
5 - 1 0 % w a t e r b y m a s s (cf. T a b l e I ) , w a t e r
imperfectly diffuse samples, d. Opt. 8oc. Am. 51,
b o u n d b y w e a k c r y s t a l l i z a t i o n bonds. I s i t
1279--1288.
n o t possible t h a t M a r t i a n o r g a n i s m s a r e
HARRIS, D. L. (1961). Photometry and colorimetry
c a p a b l e o f t a p p i n g this v a s t s u p p l y of
of planets and satellites. In "Planets and Satel-
w a t e r ? P e r h a p s M a r s seems no m o r e a r i d
lites" (G. P. Kuiper and B. M. Middlehurst,
to the i n d i g e n o u s o r g a n i s m s t h a n do, to t h e
eds.), Chap. 8. Univ. of Chicago Press, Chicago,
o r g a n i s m s c o m m o n in o u r experience, t h e Illinois.
Hovls, W. A. (1965). Infrared reflectivity of Fe~O~
t e r r e s t r i a l oceans. T h e possible r o l e of
•n H..O: Influence on Martian reflection spectra.
l i m o n i t e in h y p o t h e t i c a l M a r t i a n ecologies
is discussed elsewhere (Sagan, 1965a). Icarus 41, (preceding paper).
KELLOGG, W. W., AND SAGA/q, C. (1961). The
ACKNOWLEDGMENTS atmospheres of Mars and Venus, National
Academy of Sciences-National Research Council
w e are indebted to Messrs. Donald Crowley, Publication 944, Washington, D. C.
Leo Fitzpatrick, William J. Ladroga, Jr., and KUIPER, G. P. (1952). Planetary atmospheres and
Charles A. Lermond for assistance with the meas- their origin. In "The Atmospheres of t)he Earth
urements; and to Dr. James B. Pollack for a dis- and Planets" (G. P. Kuiper, ed.). Univ. 0f
cussion of small-particle scattering. Chicago Press, Chicago, Illinois.
KUIPER, G. P. (1957). Visual observations of Mars,
REFERENCES 1956. Astrophys. J. 125, 307-317.
ADAMCIK, J. A. (1963). The water vapor content MOROZ, V. I. (1964). The infrared spectrum of
of the Martian atmosphere as a problem of Mars (X 1.1-4.1 ~). Astronomiche~kii Zhurnal
chemical equilibrium. Planet. Space Sci. l l , 355- (USSR) 41(2), 350-361.
358. SAaA/q, C. (1961). The abundance of water vapor
BOGGESS,A. (1965). In "Astronomical observations on Mars. Astron. J. 66, 52.
from space vehicles, Proc. I.A.U. Symposium SAGA/q,C. (1965a). Primordial ultraviolet synthesis
No. 23, Liege" (J. Steinberg, ed.), in press. of nucleoside phosphates. In "The Origin Of
BOGGESS, A., AND DUNKELMAN, L. (1959). Ultra- Prebiologieal Systems" (S. W. Fox, ed.), Aca-
violet reflectivities of Mars and Jupiter. Astro- demic Press, New York.
phys. J. 129, 236-237. SAGAN, C., ed. (1965b). "Remote Investigations of
CHAPMAN, C. A. (1960). Rhyolite. In "Encyclo- Martian Biology." Conference Proceedings, to
pedia of Science and Technology." McGraw-Hill, be published.
New York. SAGA/q,C., HA/qST,P. L., A/qDYOUNG, A. T. (1965).
DANIELSO/q,R. E., GAUSTAD,J. E., SCHWARZSCHILD, Nitrogen oxides on Mars. Planet. Space Sci., in
M., WEAVER, H. F., AND WOOLF, N. J. (1964). press.
Mars observations from Stratoscope II. In SAGA/q, C., AND KELLGGa, W. W. (1963). The ter-
,'Symposium on Instrumental Astronomy." Univ. restrial planets. Ann. ReD. Astron. Astrophys.
of Arizona Press, Tucson, Arizona. I, 235-266.
DOLLFUS, A. (1957). Etude des planetes par la SHARONOV,V. V. (1961). A lithological interpreta-
polarisation de ]cur lumiere. Ann. Astrophys., tion of the photometric and colorimetric studies
Suppl. 4. of Mars. Soviet Astron.-AJ 5, 199-202.
DOLLFUS,A. (1961). Polarization studies of plan- SINTON, W. M. (1959). Further evidence of vege-
ets. In "Planets and Satellites" (G. P. Kuiper tation on Mars, Science 130, 1234-1237.
and B. M. Middlehurst, eds.), Chap. 9. Univ. of SINTON, W. M. (1965). In "Remote Investigations
Chicago Press, Chicago, Illinois. of Martian Biology" (C. Sagan, ed.), in press~
DOLLFUS,A. (1964). Discussion in "The Origin and SlVARAJASINGttAM, S., ALEXANDER, L. W., CADY,


J. G., AND CLINE, M. G. (1962). Laterite. Advan. VAN TASSEL, R. A., AND SALISBURY, J. W. (1964).
Agronomy 14, 1-60. The composition of the Martian surface. Icarus
SPINRAD, H,, MiiNCH, G., A~D KAPLAN, L. D. 3, 264-269.
(1963). The detection of water vapor on Mars. WraPT, R. (1934). Ozon und Sauerstoff in den
Astrophys. J. 137, 1319-1321. Planeten-Atmosph~ren. Ver6f]entli. Univ.-Stern-
STECHER, T. P. (1964). Private communication. warte G6ttingen, No. 38.
i)z VAUCOULEVRS,G. (1964). Geometric and photo- WOOLF, N. J. (1965). In "Remote Investigations
metric parameters of the terrestrial planets. of Martian Biology" (C. Sagan, ed.), in press.
Icarus 3, 187-235.

Total Reflection Spectrophotometry and "l'hermogravimetric Analysis of Simulated Martian Surface Materials

Recommended

Recommended

More Related Content

What's hot

What's hot (9)

Viewers also liked

Viewers also liked (20)

Similar to Total Reflection Spectrophotometry and "l'hermogravimetric Analysis of Simulated Martian Surface Materials

Similar to Total Reflection Spectrophotometry and "l'hermogravimetric Analysis of Simulated Martian Surface Materials (20)

Total Reflection Spectrophotometry and "l'hermogravimetric Analysis of Simulated Martian Surface Materials