JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 113, E09011, doi:10.1029/2008JE003119, 2008
Aqueous depletion of Mg from olivine surfaces enhanced by ion
irradiation
E. D. Cantando,1 C. A. Dukes,1 M. J. Loeffler,1 and R. A. Baragiola1
Received 15 February 2008; accepted 15 July 2008; published 27 September 2008.
[1] Laboratory experiments demonstrate that olivine irradiated with keV ions and
exposed to water exhibits up to a 60% reduction from the original near-surface
concentration of magnesium, as measured by X-ray photoelectron spectroscopy. The
depth of this depletion layer is measured to be 15 nm. Irradiations were performed with
4 keV ions at fluences from 1014 –1019 ions cm – 2 and water immersion times ranging
from 3 s to more than 100 h in neutral (pH = 6.8), high-purity water. The depletion of Mg
depends strongly on ion fluence but weakly on immersion time after 3 min, when it
saturates. Remarkably, ion irradiation enhances the rate of surface depletion of Mg from
olivine by a factor of 26,000. This effect must be considered when assessing the surface
composition of samples exposed to simulated space weathering in the laboratory and
during the handling and analysis of irradiated extraterrestrial specimens acquired via
sample return missions.
Citation: Cantando, E. D., C. A. Dukes, M. J. Loeffler, and R. A. Baragiola (2008), Aqueous depletion of Mg from olivine surfaces
enhanced by ion irradiation, J. Geophys. Res., 113, E09011, doi:10.1029/2008JE003119.
1. Introduction
[2] Surfaces of airless bodies in space are bombarded by
energetic photons, ions, and by meteorites. These impacts
alter the chemical composition of the surfaces and modify
other properties, such as optical appearance. This phenomenon, known as ‘‘space weathering,’’ can be studied by
examining extraterrestrial materials from sample return
missions or by laboratory simulations on planetary surface
analogs. In such studies it is critical to minimize alterations
of the surfaces by environmental factors during sample
handling.
[3] Space weathering effects on the structure and composition of minerals have been the focus of a number of
recent laboratory studies. Bradley [1994] measured the
compositions of interplanetary dust particles (IDPs) and
chondritic meteorites using energy dispersive X-ray analysis
(EDX) on a transmission electron microscope (TEM). Cross
sections of IDPs revealed a depletion of magnesium and
calcium relative to oxygen from the weathered silicate
surfaces with respect to the bulk. A similar Mg depletion
of olivine from the Allende meteorite was measured after
20 keV proton irradiation and attributed to nonstoichiometric sputtering of the mineral surface [Bradley, 1994]. Demyk
et al. [2001] measured with EDX a 30% reduction in the
Mg:Si ratio of olivine due to bombardment with 4 and
10 keV He ions. Similar results were reported by Toppani et
al. [2006] in silicate samples analyzed ex situ by EDX.
1
Laboratory for Atomic and Surface Physics, University of Virginia,
Charlottesville, Virginia, USA.
Copyright 2008 by the American Geophysical Union.
0148-0227/08/2008JE003119$09.00
[4] In contrast, Dukes et al. [1999] showed that keV ion
bombardment of olivine does not significantly change the
relative concentrations of Mg, Fe, O, and Si as measured by
X-ray photoelectron spectroscopy (XPS) in situ; i.e., without removing the sample from vacuum. We also note that
Jäger et al. [2003] found no change in chemical composition of enstatite (MgSiO3) due to 50 keV He+ bombardment
in contrast with the Mg depletion reported by Bradley
[1994].
[5] We posit that environmental changes on irradiated
surfaces are the cause of conflicting results between experiments where surfaces irradiated in vacuum were analyzed in
situ or analyzed after handling outside vacuum. The latter
occurs typically in specimen preparation for electron microscopy. In section 2 we present experimental results that
suggest that the reason for the contradictory results on
olivine may be the preferential dissolution of magnesium
due to exposure of the ion-irradiated surfaces to water in ex
situ analyses.
2. Experimental Methods
[6] Olivine, (Mg,Fe)2SiO4, consists of isolated silicate
tetrahedra linked by metal cations. The relative concentrations of Mg and Fe in the olivine mineral group vary
between the end-members Mg2SiO4 (forsterite) and Fe2SiO4
(fayalite). Our experiments used natural forsterite samples
from the Twin Sisters Range in Washington (Ward’s Scientific, Rochester, New York,) cut with a diamond blade into
2 mm thick samples that were then rinsed in successive
ultrasonic baths of acetone, methanol, and high-performance liquid chromatography (HPLC) water. We measured
surface composition with XPS using a Physical Electronics
560 spectrometer maintained at ultra high vacuum (base
E09011
1 of 6
E09011
CANTANDO ET AL.: Mg DEPLETION FROM IRRADIATED OLIVINE
E09011
Figure 1. X-ray photoelectron spectra of olivine upon admission to vacuum is shown. The inset shows
the high-resolution spectra of olivine upon admission to vacuum (open circles) after 1017 Ar+ cm2
irradiation (line) and after 3 min exposure to water (solid circles). No changes in stoichiometry were
observed in situ as a result of the irradiation, except for the removal of adventitious carbon. Note that the
dramatic decrease in Mg content occurs only after ion irradiation and water immersion.
pressure 109 torr). XPS spectra were excited by Al-K Xrays and measured with a cylindrical mirror analyzer in
fixed analyzer transmission mode. The energy resolution
was 3.2 eV for survey spectra and 0.6 eV for highresolution spectra. A representative XPS spectrum of an
unirradiated olivine is shown in Figure 1, where the electron
binding energy is the X-ray energy (1486.6 eV) minus the
measured electron kinetic energy. The surface composition
of our natural forsterite is presented as atomic ratios Mg:Si
= 1.72 ± 0.06, Fe:Si = 0.28 ± 0.08 and was calculated using
sensitivity factors measured for our instrument using MgO,
SiO2, and Fe calibration samples. Use of measured sensitivity factors, rather than those provided by the manufacturer, provides greater accuracy in quantification and is a
recent improvement in our analysis technique. Ion irradiations were done primarily with 4 keV Ar+ at normal
incidence; similar results were obtained with 4 keV H+2
and 4 keV D+2 . The ion beam was rastered over a 6 6 mm2
region of the sample to ensure uniform irradiation of the
entire analyzed area. Ion fluences in the range 1014 – 1019
ions cm2 were obtained using current densities below
10 mA cm2, as measured with a Faraday cup located next
to the sample. Electrostatic charging was avoided by flooding
the sample with low-energy (<1 eV) electrons. Additional
experimental details are given by Dukes et al. [1999].
[7] Water soaks were performed by removing the irradiated sample from vacuum and submerging the specimen in a
vial of 20 mL HPLC water (pH = 6.8) for times ranging
from 3 s to 166.5 h. After removal from the water bath the
sample was laid briefly on a tissue to wick away moisture
without contacting the processed surface and then reintroduced to the vacuum chamber.
[8] Our procedure for all experiments was to acquire
initial XPS spectra of the sample, irradiate it to a prescribed
fluence (ions cm2), collect a second set of XPS spectra, soak
the specimen, and then obtain a final set of XPS spectra. We
also took data with unirradiated samples and varied soak time
to quantitatively assess the effect of ion irradiation and/or
water immersion time on surface Mg depletion.
[9] The Mg:Si atomic ratios were calculated from the
areas of the Mg 2s (binding energy (BE) = 90 eV) and Si 2p
(BE = 103 eV) photoelectron peaks, shown in Figure 1
(inset). The peak areas were calculated after Shirley background subtraction and removal of satellite peaks and then
corrected with the atomic sensitivity factors. We note that
for the Si and Mg peaks in olivine, the depth of information
given by XPS is 2.6 nm [Hochella and Carim, 1988].
3. Results
[10] High-resolution XPS spectra for olivine upon admission to vacuum after 1017 Ar+ cm2 irradiation and after a
3-min soak in water are shown in Figure 1 (inset). One can
see, clearly, the decrease in the intensity of the Mg 2s (BE =
90 eV) peak relative to the Si 2p (BE = 103 eV) peak. This
drop in the Mg signal occurs only after ion irradiation and
exposure to water; no changes in the Mg:Si ratio were
observed as a result of the irradiation alone, within 5%.
Additionally, we observed that the Fe:Si ratio remained
constant within error after irradiation as well as after
irradiation plus soaking. That is, there is no detectable Fe
leaching from the olivine. Similar to the studies of dissolution of forsterite in acidic solutions by Seyama et al. [1996],
we found that the Mg and Si photoelectron lines did not
change position or width as a result of processing. Interestingly, we found that implanted Ar projectiles are removed
during soaking, as evidence by the disappearance of the Ar
2p photoelectron line (BE = 240 eV).
2 of 6
E09011
CANTANDO ET AL.: Mg DEPLETION FROM IRRADIATED OLIVINE
E09011
Figure 2. The magnesium to silicon ratio of olivine after irradiation with 4 keV Ar+ and subsequent
soaking in water for 3 min. The horizontal band indicates the range of the Mg:Si atomic ratio measured
for olivine before and after ion irradiation.
[11] We conducted control experiments to test whether
magnesium would leach preferentially from an unirradiated
olivine sample within our sensitivity and time scales. We
found no significant change in composition due to water
immersion except for an increase in carbon due to exposure
to the laboratory air.
[12] To quantify the effect of irradiation on Mg depletion,
we measured the Mg:Si atomic ratio versus irradiation
fluence up to a fluence of 1018 Ar+ cm2. These results
(Figure 2) show strong fluence dependence with a depletion
of more than 40% after 1016 Ar+ cm2 and a 3-min soak in
water. We also measured the surface depletion of Mg versus
soak time for an irradiation fluence of 1– 2 1017 Ar+
cm2. Figure 3 shows that that preferential dissolution of
magnesium occurs quickly, reaching a steady state after
3 min of water immersion. Prolonged exposure to water up
to a few hours did not decrease the Mg:Si ratio any further.
Soak times longer than 10 h produced inconsistent results,
possibly due to the spallation of the Mg-depleted layer, as
observed in leached alkali-silicate surfaces by Casey and
Bunker [1990].
[13] Since transporting and storing samples in air is
common practice in many laboratory studies, it is also of
interest to determine the effect of exposing irradiated
minerals to the humidity in the atmosphere. We analyzed
several irradiated samples that were exposed only to air
(30% relative humidity (RH)) and not directly immersed
in water. We found that irradiation to 1017 Ar+ cm2
followed by exposure to air for 2 days resulted in up to
15% Mg depletion. However, this did not always occur,
possibly due to variations in relative humidity or other
environmental variables. We hypothesize that the irradiated
area reacted with water and other molecules in the atmosphere, producing volatiles that evaporated upon admission
to the ultra high vacuum. This phenomenon will be explored further in future studies.
[14] To determine the thickness of the depleted layer, we
obtained a depth profile by sputtering. A sample was
irradiated with 4 keV Ar+ to a fluence of 1017 ions cm2
and then immersed in water for 3 min. After reintroducing
the sample to vacuum, 4 keV Ar+ ions were again used to
remove the depleted surface layer by sputtering, while
monitoring the composition with XPS. Figure 4 shows that
the Mg:Si ratio returned to its initial value, within error,
Figure 3. Surface depletion of Mg versus soak time for an
irradiation fluence of 1 – 2 1017 Ar+ cm2. The
preferential dissolution of magnesium occurs rapidly,
reaching a steady state after 3 min of water immersion.
Prolonged exposure to water up to 168 h did not decrease
the Mg:Si atomic ratio any further.
3 of 6
E09011
CANTANDO ET AL.: Mg DEPLETION FROM IRRADIATED OLIVINE
Figure 4. The Mg:Si atomic ratio versus depth for an
olivine sample irradiated with 1017 4 keV Ar+ cm2 and
soaked for 3 min in water. The bracket on the y axis indicates
Mg:Si after irradiation with 1017 4 keV Ar+ cm2 (Irr) but
without exposure to water. The depth scale was calculated
assuming a sputtering yield of 0.4 (Mg1.72Fe0.28SiO4) ion1.
after a fluence of 3 1017 Ar+ cm2. To obtain a depth
scale, we used a density of 3.27 g cm3 and a sputtering
yield of 0.4 (Mg1.72Fe0.28SiO4) ion1, derived from the
sputtering yield for silica, measured to be 1 (SiO2)/ion with
4 keV Ar+ [Betz and Wehner, 1983; Nenadović et al., 1990] in
the absence of data for olivine. Thus, we estimate the
thickness of the magnesium-depleted layer to be 15 nm,
comparable to the maximum range of the Ar ions, 15 nm,
calculated with the Monte Carlo simulation code Transport of
Ions in Matter (TRIM) (available at www.SRIM.org). In
addition, the concurrent disappearance of implanted Ar upon
soaking, mentioned above, implies that dissolution occurs
over the entire irradiated area, as in the acidic dissolution of
forsterite [Seyama et al., 1996], rather than preferentially in
etch pits. To test this conclusion, we examined the surface of a
polished olivine sample covered with a 100 line/inch Au mesh
(20 mm wires spaced 230 mm) ion irradiated and immersed in
water for 3 min. Using a scanning electron microscope, we
took images of the sample coated with 20 nm carbon to
reduce charging. We saw no discernible differences between
irradiated and unirradiated regions (Figure 5).
4. Discussion
[15] Our finding that ion irradiation enhances the depletion rate of Mg from olivine in neutral solutions in very
Figure 5. SEM photomicrographs of the (top) surface
(magnification 200X) of olivine covered with a 100 line/
inch Au mesh then ion irradiated to a fluence of 3 1017
Ar+ cm2 followed by immersion in water for 3 min and
removal of the grid; (middle) 6000X magnified image of
irradiated and water soaked area 1; and (bottom) 6000X
magnified image of unirradiated and water soaked area 2.
We see no discernible differences between irradiated and
unirradiated regions that can be attributed to etch pits.
4 of 6
E09011
E09011
CANTANDO ET AL.: Mg DEPLETION FROM IRRADIATED OLIVINE
short times is quite dramatic. It is generally known that even
in the absence of irradiation, olivine dissolves very slowly
and incongruently in acidic and, to a lesser extent, neutral
aqueous solutions [Banfield et al., 1995; Casey and Ludwig,
1995; White and Brantley, 1995; Edwards and Russell,
1996; Chardon et al., 2006]. Casey and Bunker [1990]
have shown that the dissolution rate depends strongly on pH
as well as on the Mg++ linkage to the native silicate bonds at
the surface. Pokrovsky and Schott [2000a] found that a Mgdepleted surface layer, less than 1 – 2 nm thick, formed on
forsteritic olivine in aqueous solutions with pH < 9. As
olivine is exposed to water, initially there is preferential Mg
loss but, at longer times, a steady state is reached where
dissolution becomes congruent (stoichiometric) [Pokrovsky
and Schott, 2000b]. The strong increase in the dissolution
rate as the solution becomes more acidic supports the idea
that Mg depletion occurs via the cation exchange reaction
2H+ ! Mg2+ [Blum and Lasaga, 1988]. Pokrovsky and
Schott [2000a] have shown that the dissolution leads to the
formation of silicic acid through
Mg2 SiO4 ðsÞ þ 4Hþ ðaqÞ ! 2Mg2þ ðaqÞ þ H4 SiO4 ðaqÞ;
leaving a thin silica layer on the surface. Pokrovsky and
Schott [2000a] used XPS to analyze forsterite leached in
acid solutions. They observed a broadened O 1s peak and
interpreted it as evidence of hydrogen penetration and/or the
polymerization of SiO4 units. In contrast, our measurements
on aqueous alteration of irradiated olivine do not show
either an increased O 1s peak width or a change in the O 1s
or the Si 2p binding energies, which would indicate a
significant change in chemistry.
[16] It is interesting to compare the dissolution rate for
virgin and ion-irradiated olivine samples. For unirradiated
specimens, the steady state Mg dissolution rate is 1.1 1014 mol cm2 s1 or 6.6 109 atoms cm2 s1 in
neutral solutions at room temperature [Blum and Lasaga,
1988; Pokrovsky and Schott, 2000b]. In contrast, the Mg
dissolution of ion-irradiated olivine occurs at an enormously
higher rate. The 15 nm thick altered layer loses on average
2.0 1016 Mg cm2 (half the initial value) in 2 min.
This gives a specific dissolution rate of 1.7 1014 atoms
cm2 s1, or 26,000 times larger than the reported values
for unirradiated olivine.
[17] Petit et al. [1989] have studied dissolution of minerals after irradiation with 207 keV Pb ions. They saw
enhanced dissolution in some silicate minerals [Dran et al.,
1984] but not in olivine. This contrasts with our work using
inert gas ion irradiation and may be due to either the
chemical alteration of their surfaces by the accumulation
of lead or the insufficient depth resolution of their Rutherford backscattering technique.
[18] We suggest that the radiation damage produced by
energetic ion irradiation enhances the penetration of protons
into the mineral by breaking the bonds between the Mg++
and the silicate tetrahedra. Previous measurements have
shown the formation of a hydrated layer on the surface of
olivine soaked in water [Petit et al., 1989; Fujimoto et al.,
1993]. Radiation damage amorphizes the surface region
enhancing hydrogen diffusion into the solid, where they
exchange with the Mg cations. That broken bonds enhance
chemical reactions at minerals surfaces is known and
E09011
applied in the mineral processing technique of mechanical
activation [Tkáčová, 1989]. In the particular case of olivine,
mechanical activation by milling increases the specific
dissolution rate ninefold between long and short milling
times [Kleiv and Thornhill, 2006].
5. Conclusion
[19] Magnesium is easily and substantially depleted from
the surface of ion-irradiated olivine subject to water immersion and even ambient atmospheric conditions. The depth of
alteration is equivalent to the range of the ions. The rate of Mg
depletion of olivine in a neutral solution is enhanced dramatically when the mineral is damaged by ion irradiation. Our
results suggest that the conflicting reports of preferential
magnesium removal from irradiated olivine surfaces can be
attributed to ex situ measurement techniques, where the
surface of an ion-irradiated crystalline material is susceptible
to alteration by the atmosphere or specimen preparation
(rinsing). In addition, chemically active elements in a damaged layer may be preferentially removed by water in the
Earth’s atmosphere. Therefore, surface analysis of irradiated
materials should always be conducted in situ to prevent
alteration of the surface by the atmosphere. When in situ
examination is impossible, as in sample return missions from
space, care should be taken to avoid alteration by the
atmosphere or by water exposure during sample handling.
[20] Acknowledgments. We would like to thank T. Herlihy at the
Nanoscale Materials Characterization Facility for his assistance in acquiring
the SEM images. This research was supported by the NASA Cosmochemistry program. E.D.C. thanks the Virginia Space Grant Consortium for a
fellowship.
References
Banfield, J. F., G. G. Ferruzzi, W. H. Casey, and H. R. Westrich (1995),
HRTEM study comparing naturally and experimentally weathered pyroxenoids, Geochim. Cosmochim. Acta, 59, 19 – 31, doi:10.1016/00167037(94)00372-S.
Betz, G., and G. K. Wehner (1983), Sputtering of multicomponent materials, in Sputtering by Particle Bombardment II, edited by R. Behrisch,
pp. 11 – 90, Springer, Berlin.
Blum, A., and A. Lasaga (1988), Role of surface speciation in the lowtemperature dissolution of minerals, Nature, 331, 431 – 433, doi:10.1038/
331431a0.
Bradley, J. P. (1994), Chemically anomalous, preaccretionally irradiated
grains in interplanetary dust from comets, Science, 265, 925 – 929,
doi:10.1126/science.265.5174.925.
Casey, W. H., and B. Bunker (1990), Leaching of mineral and glass surfaces during dissolution, in Mineral and Water Interface Geochemistry,
edited by M. F. Hochella and A. F. White, pp. 397 – 426, Mineral. Soc.
Am., Chelsea, Mich.
Casey, W. H., and C. Ludwig (1995), Silicate mineral dissolution as a
ligand-exchange reaction, Rev. Mineral. Geochem., 31, 87 – 117.
Chardon, E. S., F. R. Livens, and D. J. Vaughan (2006), Reactions of
feldspar surfaces with aqueous solutions, Earth Sci. Rev., 78, 1 – 26,
doi:10.1016/j.earscirev.2006.03.002.
Demyk, K., P. Carrez, H. Leroux, P. Cordier, A. P. Jones, J. Borg, E.
Quirico, P. I. Raynal, and L. d’Hendecourt (2001), Structural and chemical alteration of crystalline olivine under low energy He+ irradiation,
Astron. Astrophys., 368, L38 – L41, doi:10.1051/0004-6361:20010208.
Dran, J.-C., Y. Langevin, and J.-C. Petit (1984), Effects of ion implantation
on the dissolution properties of materials and its suitability for simulating
internal irradiation due to a-decay, Nucl. Instrum. Methods Phys. Res.,
B1, 557 – 568.
Dukes, C. A., R. A. Baragiola, and L. A. McFadden (1999), Surface modification of olivine by H+ and He+ bombardment, J. Geophys. Res., 104,
1865 – 1871, doi:10.1029/98JE02820.
Edwards, B. R., and J. K. Russell (1996), A review and analysis of silicate
mineral dissolution experiments in natural silicate melts, Chem. Geol.,
130, 233 – 245, doi:10.1016/0009-2541(96)00025-3.
5 of 6
E09011
CANTANDO ET AL.: Mg DEPLETION FROM IRRADIATED OLIVINE
Fujimoto, K., K. Fukutani, M. Tsunoda, H. Yamashita, and K. Kobayashi
(1993), Hydrogen depth profiling using 1H (15N,ag)12C resonant nuclear
reaction on water-treated olivine surfaces and characterization of hydrogen species, Geochem. J., 27, 155 – 162.
Hochella, M. F., and A. H. Carim (1988), A reassessment of electron escape
depths in silicon and thermally grown silicon dioxide thin films, Surf.
Sci., 197, L260 – L268, doi:10.1016/0039-6028(88)90625-5.
Jäger, C., D. Fabian, F. Schrempel, J. Dorchner, T. Henning, and W. Wesch
(2003), Structural processing of enstatite by ion bombardment, Astron.
Astrophys., 401, 57 – 65, doi:10.1051/0004-6361:20030002.
Kleiv, R. A., and M. Thornhill (2006), Mechanical activation of olivine,
Miner. Eng., 19, 340 – 347, doi:10.1016/j.mineng.2005.08.008.
Nenadović, T., B. Perrallon, Ž Bogdanov, Z. Djordjević, and M. Milić
(1990), Sputtering and surface topography of oxides, Nucl. Instrum.
Methods Phys. Res., Sect. B., 48, 538 – 543.
Petit, J.-C., J.-C. Dran, A. Paccagnella, and G. Delle Mea (1989), Structural
dependence of crystalline silicate hydration during aqueous dissolution,
Earth Planet. Sci. Lett. , 93, 292 – 298, doi:10.1016/0012821X(89)90077-0.
Pokrovsky, O. S., and J. Schott (2000a), Forsterite surface composition in
aqueous solutions: A combined potentiometric, electrokinetic, and spec-
E09011
troscopic approach, Geochim. Cosmochim. Acta, 64, 3299 – 3312,
doi:10.1016/S0016-7037(00)00435-X.
Pokrovsky, O. S., and J. Schott (2000b), Kinetics and mechanism of forsterite dissolution at 25°C and pH from 1 to 12, Geochim. Cosmochim.
Acta, 64, 3313 – 3325, doi:10.1016/S0016-7037(00)00434-8.
Seyama, H., M. Soma, and A. Tanaka (1996), Surface characterization of
acid-leached olivines by X-ray photoelectron spectroscopy, Chem. Geol.,
129, 209 – 216, doi:10.1016/0009-2541(95)00142-5.
Tkáčová, K. (1989), Mechanical Activation of Minerals, Elsevier,
Amsterdam.
Toppani, A., C. Dukes, R. Baragiola, and J. Bradley (2006), Segregation of
Mg, Ca, Al, and Ti in silicates during ion irradiation. Lunar Planet. Sci.,
XXXVII, abstract 2056.
White, A. F., and S. L. Brantley (1995), Chemical Weathering Rates of
Silicate Materials, Rev. Mineral. Geochem. Ser., vol. 31, edited by A. F.
White and S. L. Brantley, pp. 1 – 22, Mineral. Soc. Am., Chantilly, Va.
R. A. Baragiola, E. D. Cantando, C. A. Dukes, and M. J. Loeffler,
Laboratory for Atomic and Surface Physics, University of Virginia,
Charlottesville, VA 22904, USA. ([email protected])
6 of 6