Transcript sublimation of water ice in low pressure environments

sublimation of water ice in low pressure environments:
isotopic effects and implications for the martian paleoclimate record
J.E. Moores , R.H. Brown, D.S. Lauretta and P.H. Smith
Lunar and Planetary Laboratory, 1629 E University Blvd, Tucson, Arizona, 85721 United States
conclusions
introduction
Experiments carried out [1] have shown that vacuum sublimation of water ice possesses two distinct complexities. One of
these is a bulk isotopic effect which consists of a significant fractionation enhanced by high dust entrainment [2] in
prepared samples. Since the polar regions of Mars are made up of dusty ices [3] and are thought to contain an isotopic
record of past climate change locked in the ice [4], it is important to ask about the significance of this experimental work in
the context of the putative Martian paleoclimate record. It can be shown that even climate changes consisting only of
redistribution of ices by sublimation can leave a long lasting isotopic imprint if the amount of material moved is sufficiently
great (meters of ice).
simulating martian ice
Samples have been prepared containing various mixtures of insoluble TiO2 dust grains (density = 4.23g/cc) with pure water having
concentrations covering 0-50wt% of dust, a range consistent with the northern polar caps of Mars [3]. These dust grains were of
similar size to martian dust particles (radius of 1.6 microns), as determined by the Imager for Mars Pathfinder [5] and consistent with
atmospheric dust observed by Viking and MER [6]. Next, these ice-dust mixtures were been salted with D2O and flash frozen by
pipetting into liquid nitrogen. This prevented the dust from being excluded from the ice matrix.
An easily detectable 1xVSMOW (~20%) average deviation in the observed D/H ratio will be preserved for 38 billion years
and the deuterium peak found in the stratigraphy of the polar cap ice will have a mean width of ~10m. This preservation
figure should be viewed as a lower bound for near surface deposits since during low obliquity, the temperature of the
polar cap is on average lower [9] than during high obliquity. However, deeper deposits, being somewhat warmer should
age more rapidly.This preservation age when combined with the depth of the PLD suggests that it may be possible to
preserve a very long record of climate change in the polar caps on Mars provided that vapor diffusion may be
suppressed by compaction in the ice sheet. Here the D/H ratio may function analogously to the O18/O16 ratio in ice cores
on the earth. This signal would be best detected by in-situ coring, however, since the PLDs are exposed by the polar
troughs, it may be possible to detect the climate signal by performing a
transect with a rover on the surface.
The actively sublimating surface of
the ice cap pictured here is
composed of a mixture of ice and
fine-grained airfall dust. This
material was simulated.
The resulting samples were then allowed to sublimate under a quartz tungsten halogen lamp set at a low current level to produce
heating from above while the base of the sample was kept at 160K using a liquid helium cold finger. Analysis of the initial sublimate
gas indicated that the temperature at the top of the sample was 202.88 K and temperature sensors imbedded in the chamber did not
detect a significant change over the course of the experiment. During sublimation, the gas pressure was monitored by a capacitance
manometer and the composition was recorded using a mass spectrometer.
experimental results for dusty ices
Two of our mass spectra are presented here. The first of these [below, top] illustrates the
isotopic evolution of a crystalline sample containing no dust at all. In this case, the D/H
ratio in the sublimate gas begins below the ratio in the bulk solid near the level predicted
by the difference in vapor pressure between deuterated (HDO) and non-deuterated (H2O)
water molecules. This preferential sublimation of H2O over HDO causes the surface
concentration of HDO to increase resulting in a corresponding increase in the number of
HDO molecules escaping the surface.
implications for the NPLD
The classical next argument would be that once the sublimated gas achieves the same
D/H ratio as the bulk solid we reach a state of equilibrium where HDO molecules are
escaping from the actively sublimating surface as quickly as they are becoming available
from the bulk by the advance of the sublimation front down through the sample. However,
as can be seen from the figure we never reach a state of equilibrium, in fact by the end of
the experiment the concentration of HDO in the sublimate gas has increased past the
concentration in the bulk solid. The reason for this lies in the mechanics of solid state
diffusion which at the temperature of interest takes place at a rate equivalent to or greater
than the rate of advance of the sublimation front [1]. This allows the bulk solid to become
enriched in HDO, a fact which eventually shows up in the sublimate gas emitted.
A second sample [Below, bottom] was prepared using 25wt% TiO2 dust. Unlike a clean
sample, this type of sample will leave behind a refractory lag deposit as the sublimation
front travels through the sample. This creates a barrier to the escape of molecules from
the surface of the sample. This has two effects. First, the escape rate of all species is
significantly decreased: from 2.0x10-3 torr total pressure down to 1.5x10-4 torr after 35
days and 160m of sublimation was observed. Secondly since the lighter H2O molecules
can diffuse faster though this layer then the heavier HDO molecules, the D/H ratio of the
gas escaping the sample decreases as the lag layer increases in thickness. An additional
effect comes from our method of creating the dusty sample by flash freezing. This
produced a porous material analogous to an ice laid down by precipitation. In such a unit,
heavier molecules, once activated by sublimation can diffuse rapidly and preferentially to
the colder areas of a sample by simple vapor diffusion, thus depleting the amount of HDO
available for the sublimate gas even further.
The dusty ice case is most applicable to the northern polar regions of
Mars, in particular the Northern Polar Layered Deposits. Like our second
sample, these deposits are thought to contain large quantities of airfall
dust and also, like our sample, the icy component periodically receives
sufficient insolation to cause it to sublimate. Thus, our results indicate that
these deposits will also form dust lags which will both decrease their
sublimate gas output and concentrate the heavier isotopic species of
water in the remaining solid.
A rover transect of layered
deposits such as those pictured
here could allow the story of
past climate on Mars to be read
Volcanic Caldera like this one
are the birthplace of the very
fine material which makes
heavy fractionation possible
This relatively high D/H layer would be laid down at the interface between
any dust lag and the residual ice. This enhancement over the course of 35
days (the length of our experimental run) totals 1.47x1021 molecules/m2
and represents 160m of sublimation. If we assume that the of order 10m
spacing of the dark units in the PLD represents the typical amount of
sublimated material over an obliquity cycle [7], we have a total
enhancement of 150mol/m2 each obliquity cycle. This begs the question of
how long this isotopic signal can be preserved once this enriched material
has been buried and its vapor pathways closed. To answer this question
we must consider again solid state diffusion of HDO in H2O. As a parallel
we make use of an equation from semiconductor engineering governing
diffusion of a fixed amount of doping agent into a semi-infinite substrate:

1m3
 
x2 

[12]


D / H ( x, t ) 

 exp 

50944m ol  2 Dt 
4
Dt


2
Where β is the amount of the D pulse (in mol/m ) diffusing. If we assume
that the ice – now buried below the annual temperature wave with most
vapor pathways filled in – remains initially close to the surface, it will have
a typical temperature (at high obliquity) of 170K [10] and we may use a
diffusion coefficient of HDO16 into H2O16 of 6.38x10-18m2/s [9]. Using these
figures, the changing distribution of D with time is shown above.
background image credit:
ESA/DLR/FU Berlin (G. Neukum) [11]
References: [1] Brown, R.H. et al (2006) Submitted, ApJ. [2] Moores, J.E. et al (2005) LPSC
XXXVI abstract n°1973. [3] Clifford et al (2000) Icarus v144 pp210-242. [4] Saunders and Hecht
(2002) AGUFM abstract n°P12A-0369 [5] Tomasko et al (1999) JGR vol 104-E4 p8987-9008 [6]
Lemmon, M.T. (2004) Science 306 n°5702 pp.1753-1756 [7] Milkovich and Head (2005) JGR
110 E01005 [8] Glicksman, M.E. (2000) Diffusion in Solids. [9] Livingston et al (1998) J. Chem.
Phys. 108 2197-2207. [10] Larsen and Dahl-Jensen (2000) Icarus 144 456-462. [11] Ice and
Dust at Martian North Pole (2006) ESA Image ID n°SEMKO5D3M5E [12] Schaeffer et al (1999)
The Science and Design of Engineering Materials 2nd Ed. P136.
This oblique image illustrates several important aspects of this poster.
Shown prominently is the layered terrain which may contain a record of climate
change locked in D/H variations. Also visible is a sublimating water-ice surface rich in dust
above calderas which may contain fine volcanic ash [11], the source of future dust slags within
the polar layered deposits. Navigate this poster by reading boxes next to the relevant feature.