1
Gas trapping in ice and its release upon warming
Akiva Bar-Nun and Diana Laufer
Department of Geophysics and Planetary Sciences, Tel-Aviv University, Israel
Oscar Rebolledo, Serguei Malyk, Hanna Reisler, and Curt Wittig
Department of Chemistry, University of Southern California, Los Angeles, CA 90089,
USA
NASA's Deep Impact was a turning point in our measurements of comet properties. For
the first time we obtained direct measurements of the density, thermal inertia of the
surface, and, most importantly, the tensile strength of the upper layers. The very small
tensile strength of only 1-10 kPa (like that of Talcum powder) tells us that comet Tempel
1 is a loose agglomerate of fluffy ice particles (Bar-Nun et al., 2007). In what follows, we
describe how gases are trapped in fluffy ice particles, how they are released from them
when the temperature is increased, either by overall heating or by pulsed infrared laser
irradiation and finally, what happens when a "time bomb" releases its trapped gases when
the heat wave penetrates inward. In addition, it will be shown that laboratory
measurements can now be carried out that address fundamental transport issues such as
the release of trapped gases in such ice environments and their transport through thin and
thick ice layers.
Gas Trapping
When water vapor freezes at low temperatures, the ice forms an "open" amorphous
structure (often referred to as amorphous solid water, ASW) with many open pores. A
CO content of a few percent in comets requires the temperature of formation of the ice
grains, which agglomerate to form comet nuclei, to be about 25 K (Notesco and Bar-Nun,
2005). This is in agreement with the nuclear spin temperature of ~ 25 K derived from the
2
ortho/para ratio of H2O in comet Halley (Mumma et al., 1988) and Comet Hale-Bopp
(Crovisier, 1999), and also the temperature of ~ 28 K derived from the ortho/para ratio of
NH3 in Comet Linear C/1999 S4 (Kawakita et al., 2001).
At such low temperatures, water molecules stick wherever they hit the ice surface, as
they do not have enough energy to migrate and form a more stable structure with more
hydrogen bonds between the water molecules. This metastable structure rearranges
during the annealing of the amorphous ice between 35 and 120 K, and more vigorously
during transformation to the cubic structure at 120-135 K.
Figure 1. Fluxes of gas and water vapor versus temperature. Gas-rich amorphous ice was condensed at 50
K from the mixture H2O:CH4:CO:Ar = 1:0.33:0.33:0.33. Note: changes in gas evolution occur over 6
orders of magnitude. Ranges b, d and d' – gas release due to annealing of the amorphous ice; c – desorption
of a monolayer of adsorbed gas; e – gas release during the transformation of amorphous ice to cubic ice and
restrained amorphous ice; f – gas release during the partial transformation to hexagonal ice; g – release of
gases that were trapped even more tightly during the transformations and can be released only when the ice
matrix itself sublimes (Bar-Nun et al., 1988).
When amorphous ice is formed in the presence of gas, the gas molecules enter the
open pores and reside there for a time that depends on the interactions with the dangling
3
OH's on the ice walls, as well as the masses of the gas molecules and the ambient
temperature. If, before escaping the pore, an overlayer of ice forms, the gas molecules
become trapped, and they can escape only when the water molecules move during
annealing and/or the transformation to cubic ice. Molecules such as HCN, NH3, CO2,
CH4, etc. are trapped more efficiently than CO, Ar, Kr, and Xe (Bar-Nun, et al., 1988).
Due to the energy of the molecules in the open pores, the trapping efficiency of each gas
drops exponentially with increasing temperature; see Figures 1 and 2.
Figure 2a. Total amount of trapped gas vs. deposition temperature for water vapor – gas mixtures. ◊ CH4,
∆ CO, ○ N2, □ Ar. Open symbols: results from deposition of a 1:1 ratio of a single gas species to water
vapor. Filled symbols: results from the deposition of gas mixtures, H2O:CH4:Ar:CO(or N2) = 1:0.33:0.33
0.33.
4
Figure 2b. Logarithmic decrease of trapping efficiency versus temperature: Ne, Ar, and CO2.
The extreme case, which demonstrated the dependence on the energy of the gas in a
pore, was a comparison of the trapping of Ar, Kr, and Xe. The ratios of residence times in
the open pore were (m1/m2)1/2, and the ratios of their trapping efficiencies were observed
to scale according to this factor for
36
Ar / 40Ar and
82
Kr / 84Kr (Notesco et al., 1999). The
differences among the heavier Xe isotopes were too small to be significant, especially in
view of their higher polarizability and consequent adherence to the ice walls.
Another mechanism of gas trapping is through the formation of clathrate-hydrate,
where water molecules form a cage around a guest molecule (Petrenko and Whitworth,
2002). In the laboratory, ice is crushed by steel balls in the presence of a high gas
pressure (Schmitt, 1991). Since steel balls were not available in the region of comet
formation, we resorted to gentler procedures. In this region the gas pressures were very
low, and consequently clathrate-hydrates of CH4, CO, CO2, N2, HCN, Ar, Kr, Xe could
not have been formed (Davidson et al., 1987, Schmitt, 1991). The only gases capable of
forming clathrate-hydrates at low gas pressures are methanol, tetrahydrofurane, and
oxirane (ethylene oxide) (Notesco and Bar-Nun, 2000, Richardson et al., 1985), but not
even the fairly polar HCN.
5
Another trapping mechanism involves ammonia. It is trapped in the ice like the other
gases, but upon warming, H2O and NH3 form a hemihydrate NH3·2H2O at ~ 130 K, and a
hydrate NH3·H2O at 158 K (Moore et al., 2007, Notesco and Bar-Nun, unpublished).
Both clathrate-hydrates and ammonia hydrates decompose at elevated temperatures and
release their gases.
Gas Release
The entire description of gas release is shown in Figure 1. The first step of gas release
from its cages in the amorphous ice occurs between 35 and 120 K, where it anneals
slowly, in temperature dependent steps (Bar-Nun et al., 1987). During these slow
structural changes, some trapped gas is released. Gas release stops altogether when the
temperature is kept constant for several days and is resumed when the temperature
increases by a mere few degrees, showing that the annealing is strongly temperature
dependant. This gas release can be responsible for the observed activity of comets at
large heliocentric distances preperihelion (Meech et al., 2008).
Beginning at about 120 K, water molecules have enough energy to move and form the
more stable cubic structure, with about 70% remaining as "restrained amorphous" ice
(Jenniskens and Blake, 1994). The rate of transformation increases exponentially with
temperature and becomes large at ~ 135 K. During this process, the gases that were
trapped in the amorphous ice are released from it by dynamic percolation (Laufer et al.,
1987) following the increasing fraction of transformed ice.
6
Figure 3. A plot of the fluxes of evolved Ar (black curve) and water (gray curve) during warming up of
0.1 µm (a) and 5 µm (b) ice layers. The gas-laden ice was deposited at 27 K from a H2O:Ar = 1:1 mixture,
at a rate of 10-3 µm min-1. The reason for the large amount of frozen Ar (30-40K) compared with the
amount of water was explained by Notesco et al., (2003). This frozen gas would not accumulate during ice
formation in micron and sub-micron grains during 105 years.
7
The dependence of gas release on the thickness of the ice sample is shown in Figure 3.
In Figure 3a, the 0.01 µm ice layer releases all its trapped gas during the transformation
to cubic ice and restrained amorphous ice. However as seen in Figure 3b, with a 5 µm
thick ice layer, some of the gas is left in the cubic and restrained amorphous ice to be
released only during the transformation of both ices to the hexagonal structure at 160 K.
This is because in a thicker ice layer, movement of water molecule to form cubic and
restrained amorphous structures also closes pores and traps some of the gases, until
further movement, during the transformation to hexagonal ice releases them. In thicker
ice samples, 50 µm as in Figure 1, even the transformation to hexagonal ice can not
release all the gas, which is trapped at a greater depth. Only the sublimation of the entire
ice layer releases the trapped gases (region "g" in Figure 1).
Laser Induced Gas Release
A complementary strategy for studying the release of trapped gas in ASW films enlists
pulsed IR radiation. This is easily obtained by stimulated Raman scattering (in D2) of the
1064 nm output of a Nd-YAG laser: the 3425 cm–1 (second Stokes) line lies within the
broad water absorption band. Specific vibrational degrees of freedom targeted by this
photoexcitation are converted to heat on a picosecond time scale. For a few hundred
layers, energy deposition is quite uniform because the attenuation length is an order of
magnitude larger than the film thickness. Were the film on a perfectly insulating substrate, photoexcitation could create a liquid that would rid itself of CO2. Alternatively,
facile heat transfer to the substrate can result in large temperature gradients, and the
release of trapped gas can vary markedly with distance from the substrate.
This is illustrated with an example. Referring to Figure 4, pulsed IR radiation is
incident on a CO2/ASW film, and a time-of-flight (TOF) mass spectrometer probes
molecules that leave the surface. An important feature is that the electron beam can be
operated at repetition rates up to 200 kHz. Thus, temporal behavior can be recorded for
different species leaving the surface.
8
microchannel plates
10
drift tube
ion extraction
region
MgO(100)
substrate
−10
Torr base pressure
3425 cm −1 radiation
Nd:YAG laser,
Raman shifter
focused e-beam
1-2 mm diameter
ASW/CO2 film
500 µ m diameter
3425 cm −1 radiation
Figure 4. Arrangement for studying gas release from ASW/CO2 films: The UHV chamber contains
additional diagnostics: FTIR, TPD, and Auger spectroscopies (Hawkins et al., 2005, Kumi et al., 2006,
Malyk et al., 2007).
Figure 5(a) shows TOF spectra for a single IR pulse. Each 10 µs interval corresponds
to a mass spectrum. Good S/N enables removal efficiencies to be determined layer-bylayer. The IR beam diameter exceeds greatly the film thickness, so heat transfer takes
place into the substrate.
9
CO 2
H 2O
0
50
150
100
200
250
time / µ s
0. 6
300
CO 2+
H 2O+
0. 5
0. 4
0. 3
0. 2
OH + H O+
3
CO+
O+
0. 1
0. 0
28
28
30
30
32
32
34
34
time / µ s
Figure 5. Upper: a burst of molecules produced by a single IR pulse is analyzed by closely spaced TOF
mass spectra. Blue and red lines indicate H2O and CO2 contributions, respectively. Lower: the region near
30 µs is expanded. H3O+ indicates the presence of higher-than-monomer species. There is ~ 3 times more
CO2 than H2O in the released gas even though the deposited sample contains 3 times more H2O than CO2.
Note that CO2 leaves the surface with lower speed than H2O.
Figure 6 shows data from two pulses incident at the same location. The first addresses
a fresh sample (300 layers, CO2 : H2O = 1:3). Released CO2 exceeds released H2O by a
factor of ~ 3. Thus, the CO2 content is ~ 9 times higher than in the initially condensed
sample. The CO2:H2O ratio for released gas decreases with continued irradiation: (b)
shows that for the sixth pulse the amounts of CO2 and H2O are comparable. The system
can be calibrated by recording signal vs. film thickness, enabling the amount of material
removed per pulse to be determined.
10
CO 2
(a)
H 2 O:
blue
boxes
0
50
150
100
time / µ s
200
250
300
CO 2
(b)
H 2O:
blue
boxes
0
50
100
150
200
250
300
Figure 6. Entry (a) is the same as Figure 8(a). Entry (b) is from the sixth pulse incident on the same area.
The amount of CO2 exceeds the amount of H2O by 20%. Blue boxes correspond to H2O peaks.
The laser-induced temperature increase facilitates CO2 removal, whereas rapid cooling
by the substrate acts to lessen CO2 removal. Consequently, material close to the substrate
remains amorphous, trapping CO2 efficiently, whereas the region near the film surface
has much less CO2. The physics is the same as that of the fluffy agglomerates
experiments, but here processes transpire in the presence of a heat sink. On an insulating
substrate, a few mJ is expected to release nearly 100% of the CO2 in a film of several
hundred layers thickness, over an area > 1 mm2.
11
Summing up the results of the experimental studies of gas release from thin layers of
gas laden amorphous ice, we conclude that changes in the ice, due to heating the entire
ice, or locally by laser heating, release the trapped gases. Yet, what happens in much
thicker, comet like ice samples, composed of these gas-laden grains of amorphous ice?
Large Ice Samples
As described above, the trapping of gas, as well as its release from ice, were studied in
our laboratory with ice samples of thickness 0.1–100 µm. The machine used for these
studies is shown in Figure 7.
Figure 7. A picture of the laboratory machine where thin ice samples are studied. Water vapor and gases
flow onto a cold plate. Amorphous ice traps gases that are released during warming of the cold plate.
A further step in comet simulation was achieved by constructing a unique machine
capable of producing large (20 cm diameter, up to 10 cm high) samples of gas-laden
amorphous ice (Figure 4). One cannot produce large samples of amorphous ice by
12
depositing water vapor on a cold plate, as is done routinely for samples up to hundreds of
micrometers thick used in small-ice-sample studies. This is because it is necessary to get
rid of the large heat of condensation of water vapor into ice (2.7 × 1010 erg g−1) through
the growing ice sample, whose heat conductivity is poor (less than ~104 erg cm−1 K−1 s−1)
(Klinger 1981, Prialnik and Bar-Nun 1992, Bar-Nun and Laufer 2003).
Figure 8. Picture of the laboratory machine that is used to study large ice samples.
If one grows an ice layer that is too thick, the heat of deposition cannot be transmitted
to the cold surface, and the newly deposited ice layer reaches ~ 120 K and becomes
crystalline rather than amorphous. Consequently, it does not trap the gas that is flowed
together with the water vapor. The solution was to slowly form thin ~ 200 µm ice layers
on a 80 K (liquid nitrogen) cold plate, through which the heat of condensation of water
could still be transmitted to the cold surface, remain amorphous, and trap the
accompanying gas in it. Once a thin amorphous gas-laden ice layer is formed, it is
scraped from the plate on which it is formed by an 80 K cold knife into the 80 K sample
container, which is covered by an 80 K dome (Bar-Nun and Laufer 2003).
13
Figure 9. Schematic drawing of the machine producing large (200 cm2 × 10 cm) gas-laden amorphous ice
samples, as a "comet" simulation: (1) vacuum chamber; (2) 80 K cold plate; (3) 200 µm amorphous
gas-laden ice; (4) homogeneous flow of water vapor and gas; (5) water vapor and gas pipes; (6) 200 cm2
and 5–10 cm thick ice sample; (7) heating dome; (8) 80 K cold knife; (9) thermocouples; (10) density
measurements; (11) mass spectrometer; (12) ionization gauge; (13) heating tape; (14) LN2 cooling pipes.
Thus a 20 cm diameter and up to 10 cm thick sample of a fluffy agglomerate of 200
µm particles of porous amorphous gas-laden ice is formed. This sample is a good representative of the "snowball" so cleverly proposed by Fred Whipple (1987). Upon warming
by the heating dome (7 in Figure 9), a heat wave penetrates inward, releasing the trapped
gases and sublimating the upper layer. A typical plot, together with the response of the
embedded thermocouples to the penetrating heat wave, is shown in Figure 10.
14
Figure 10. Evolution of trapped Ar and sublimation of water from the surface, together with temperatures
measured by thermocouples in the interior.
A major observation is that trapped Ar (which behaves like CO) emanates from the
entire sample, whereas water vapor sublimes only from the surface, as expected. This is
demonstrated by first blocking the irradiation, followed by resuming it. Whereas the
surface cools quickly and ice sublimation ceases, the warmer interior releases its trapped
gas. As seen in Figures 1-6, the gases thus released escape quiescently. However, when
more vigorous gas release occurs, the gases carry with them huge fluxes of ice grains.
These grains, each having about 1010 water molecules, are propelled by gas jets, each
having 1010-1012 molecules (Laufer et al., 1987). In large samples, the gas emanating
from below produces a bulge that eventually breaks, releasing a huge flux of gas and a
plethora of ice grains. These phenomena are discussed in another chapter in this book.
15
Summary
The processes occurring during the formation of ice grains in the presence of gas are
rather simple: At low enough temperatures around 20-50 K, the water molecules stick to
the ice matrix where they hit, forming an open porous structure. Gas molecules enter
these pores and reside there for a while, attracted by van der Waals forces. If before they
leave the pore it is covered by a subsequent ice layer, they are trapped inside.
The trapped gases can escape the ice when the water molecules which form the cold
ice move due to heating. This process occurs during the annealing of the ice between
50 to 120 K in a stepwise, temperature depended manner, and then during the vigorous
transformation to cubic ice. If buried in deeper layers, further transformation of the cubic
ice to hexagonal ice is needed for release another fraction of the trapped gases. All these
movements of water molecules in the ice both open gas filled pores and close some.
Hence the final release of all the remaining gas awaits the sublimation of the entire layer.
This description encompasses the entire gas release mechanism. However, with submicron ice particles, all the trapped gas will be released during the transformation to
cubic ice.
Another matter altogether is the release of trapped gas from a few cm thick
agglomerate of ~100 µm gas-laden amorphous ice particles. Upon warming from above,
gas is released from the interior while the water sublimates only from the surface. This
means that the gas/water vapor in cometary comae are enriched by gas, and raises the
question of what is the real composition of cometary nuclei.
Further consequences of gas release from underneath are discussed in other chapters
of this book.
Acknowledgements
This work has been supported by the US- Israel Binational Science Foundation, BSF
grant 2006339 and by the U.S. Army Research Office under grant number
W911NF-07-1-0081.
16
References
Bar-Nun A. and Laufer D., 2003. First experimental studies of large samples of gas-laden
amorphous "cometary'' ices. Icarus 161, 157-163.
Bar-Nun, A. and Kleinfeld, I., 1989. On the temperature and gas composition in the
region of comet formation. Icarus 80, 243-253.
Bar-Nun A., Kleinfeld I., Kochavi E., 1988. Trapping of gas mixtures by amorphous
water ice. Phys. Rev. B 38, 7749-7754.
Bar-Nun A., Dror J., Kochavi E., Laufer D. 1987. Amorphous water ice and its ability to
trap gases. Phys. Rev. B 35, 2427-2435.
Crovisier J., 1997. Infrared Observations of Volatile Molecules in Comet Hale-Bopp.
Earth, Moon and Planets, 79, 125-143.
Davidson D.W., Desando M.A., Gough S.R., Handa Y.P., Ratcliffe C.I., Ripmeester J.A.,
Tse J.S., 1987. A clathrate-hydrate of carbon monoxide. Nature 328, 418-419.
Hawkins S., Kumi G., Malyk S., Reisler H., Wittig C. 2005. Temperature programmed
desorption and infrared spectroscopic studies of thin water films on MgO(100). Chem.
Phys. Lett. 404, 19-24.
Jenniskens P. and Blake D.F., 1994. Structural transitions in amorphous water ice and
astrophysical implications. Science 265, 753-756.
Kawakita H., Watanabe J., Ando H., Aoki W., Fuse T., Honda S., Izumiura H., Kajino T.,
Kambe E., Kawanomoto S., Noguchi K., Okita K., Sadakane K., Sato B., Takada-Hidai
M., Takeda Y., Usuda T., Watanabe E., Yoshida M., 2001. The Spin Temperature of NH3
in Comet C/1999S4 (LINEAR). Science 294, 1089-1091.
Klinger J., 1981. Some consequences of a phase transition of water ice on the heat
balance of comet nuclei. Icarus 47, 320–324.
Kumi G., Malyk S., Hawkins S., Reisler H., Wittig C. 2006. Amorphous solid water
films: Transport and guest-host interactions with CO2 and N2O dopants. J. Phys. Chem.
A 110, 2097-2105.
Laufer D., Kochavi E., and Bar-Nun A., 1987. Structure and dynamics of amorphous
water ice. Phys. Rev. B. 36, 9219-9227.
Malyk S., Kumi G., Reisler H., Wittig C., 2007. Trapping and release of CO2 guest
molecules by amorphous ice. J. Phys.Chem. A 111, 13365-13370.
17
Mumma M.J., Blass W.E., Weaver H.A., Larson H.P., 1988. Measurements of the OrthoPara Ratio and Nuclear Spin Temperature of Water Vapor in Comets Halley and
Wilson(1986l) and Implications for Their Origin and Evolution. Bull. Amer. Astron. Soc.
20, 826.
Moore M.H., Ferranteb, R.F., Hudsonc R.L., Stoneb J.N., 2007. Ammonia–water ice
laboratory studies relevant to outer Solar System surfaces. Icarus 190, 260-273 .
Notesco G. and Bar-Nun A., 2005. A ~25 K temperature of formation for the submicron
ice grains which formed comets. Icarus 175, 546-550.
Notesco G. Bar-Nun A., Owen, T., 2003. Gas trapping in water ice at very low deposition
rates and implications for comets. Icarus 162, 183-189.
Notesco G. and Bar-Nun, A., 2000. The Effect of Methanol Clathrate-Hydrate Formation
and Other Gas-Trapping Mechanisms on the Structure and Dynamics of Cometary Ices.
Icarus 148, 456-463.
Notesco G., Laufer D., Bar-Nun A., Owen T., 1999. An Experimental Study of the
Isotopic Enrichment in Ar, Kr and Xe When Trapped in Water Ice. Icarus 142, 298–300.
Petrenko V.F. and Whitworth R.W., 2002. In "Physics of Ice" Oxford University Press.
Inc., NY.
Prialnik D. and Bar-Nun A., 1992. Crystallization of amorphous ice as the cause of
Comet P/Halley's outburst at 14 AU. Astron. Astrophys. 258, L9-L12.
Richardson H.H., Wooldridge P.J., Devlin J.P., 1985. FT-IR spectra of vacuum deposited
clathrate hydrates of oxirane, H2S, THF, and ethane. J. Chem. Phys. 83, 4387-4394.
Schmitt B., 1991. Thermal and physicochemical processes in cometary nuclei .
In "Interrelations between physics and dynamic for minor bodies in the solar system".
Benest, E. and Foershle, C. eds., pp. 265-307.
Whipple F.L., 1987. The Cometary Nucleus. Astron. Astrophys. 187, 852-858.