Available via license: CC BY 4.0
Content may be subject to copyright.
minerals
Article
Phase Relations in MAFSH System up to 21 GPa:
Implications for Water Cycles in Martian Interior
Chaowen Xu 1, * and Toru Inoue 1,2,3
1Geodynamics Research Center, Ehime University, 2-5 Bunkyo-cho, Matsuyama 790-8577, Japan;
toinoue@hiroshima-u.ac.jp
2Department of Earth and Planetary Systems Science, Hiroshima University, 1-3-1 Kagamiyama,
Higashi-Hiroshima 739-8526, Japan
3Hiroshima Institute of Plate Convergence Region Research (HiPeR), Hiroshima University,
Higashi-Hiroshima, Hiroshima 739-8526, Japan
*Correspondence: dkchaowen@126.com; Tel.: +81-050-3699-0952
Received: 3 August 2019; Accepted: 14 September 2019; Published: 16 September 2019
Abstract:
To elucidate the water cycles in iron-rich Mars, we investigated the phase relation of
a water-undersaturated (2 wt.%) analog of Martian mantle in simplified MgO-Al
2
O
3
-FeO-SiO
2
-H
2
O
(MAFSH) system between 15 and 21 GPa at 900–1500
◦
C using a multi-anvil apparatus. Results
showed that phase E coexisting with wadsleyite or ringwoodite was at least stable at 15–16.5 GPa
and below 1050
◦
C. Phase D coexisted with ringwoodite at pressures higher than 16.5 GPa and
temperatures below 1100
◦
C. The transition pressure of the loop at the wadsleyite-ringwoodite
boundary shifted towards lower pressure in an iron-rich system compared with a hydrous pyrolite
model of the Earth. Some evidence indicates that water once existed on the Martian surface on ancient
Mars. The water present in the hydrous crust might have been brought into the deep interior by the
convecting mantle. Therefore, water might have been transported to the deep Martian interior by
hydrous minerals, such as phase E and phase D, in cold subduction plates. Moreover, it might have
been stored in wadsleyite or ringwoodite after those hydrous materials decomposed when the plates
equilibrated thermally with the surrounding Martian mantle.
Keywords: high pressure; high temperature; Martian interior; water storage; water transport
1. Introduction
Water is an important volatile material that affects the physical and chemical properties of planetary
interiors, such as those of Earth and Mars. Water transportation and storage are crucially important
components of the water cycle, strongly affecting geodynamic processes. On Earth, several studies
have indicated that some hydrous minerals can hold and transport water to the deep Earth by cold
subducting slabs [
1
–
3
]. The so-called dense hydrous magnesium silicates (DHMSs) in MgO-SiO
2
-H
2
O
(MSH) system, such as phase A (Mg
7
Si
2
O
14
H
6
), phase E (Mg
2.3
Si
1.25
O
6
H
2.4
), superhydrous phase B
(Mg
10
Si
3
O
18
H
4
), phase D (MgSi
2
O
6
H
2
), and phase H (MgSiO
4
H
2
), are considered to be important
carriers of subducted water from mantle transition zone down to the middle part of the lower
mantle [
4
–
9
]. By contrast, the major minerals—wadsleyite and ringwoodite—in the Earth’s mantle
transition zone (MTZ), might act as a large water reservoir because they might hold several oceans’
masses of H2O [10].
The existence of water on Mars has long been controversial. Recent studies of topographic
features, for example, the northern plains, sedimentary deposits, and valley networks [
11
–
13
] and the
detection of subsurface ice, as well as various hydrous minerals in Lyot crater, suggest the existence of
an ancient Martian ocean on the surface [
14
–
16
]. The convecting Martian mantle may be hydrated
Minerals 2019,9, 559; doi:10.3390/min9090559 www.mdpi.com/journal/minerals
Minerals 2019,9, 559 2 of 9
when reacting with overlying hydrous crust, bringing water into the deep interior [
17
]. Therefore,
similarly to an Earth-like planet, some hydrous minerals might exist in cold region of iron-rich Mars,
and wadsleyite and ringwoodite might also hold a huge amount of water in the Martian interior, as it
was argued for the present-day Earth.
Several studies have identified phase relations in the MSH and MgO-Al
2
O
3
-SiO
2
-H
2
O (MASH)
systems, with the observation of various hydrous minerals at P–Tconditions related to the cold
subduction slabs [
3
,
4
,
6
,
7
,
18
,
19
]. Nevertheless, few data are available for hydrous iron-bearing systems:
data for iron-rich systems, such as Mars, are rarely reported [
5
]. A better understanding of the phase
relations in MgO-Al
2
O
3
-FeO-SiO
2
-H
2
O (MAFSH) system might help to elucidate the geodynamic
processes associated with the deepwater cycles of Mars. Therefore, we determined the phase relations in
iron-rich MAFSH system between 15 GPa and 21 GPa to systematically ascertain the stability of DHMSs,
and further estimate the possible water transportation into the Martian interior by subducting processes.
2. Materials and Methods
High-pressure and high-temperature experiments were conducted at Geodynamics Research Center
(GRC), Ehime University, using a Kawai-type 1000 ton multi-anvil apparatus. Tungsten carbide cubes
with truncation edge length (TEL) of 4 mm were used in combination with Co-doped MgO-octahedra of
10 mm edge length (10/4 assemblage). Preformed pyrophyllite gaskets were used between the cubes, and
LaCrO
3
was used as the heater. A gold sample capsule was used in the cell assemblage. Pressures were
calibrated at room temperature (25
◦
C) by diagnostic changes in the electrical resistances of ZnTe (9.6 and
12 GPa), ZnS (15.5 GPa), GaAs (18.3 GPa), and GaP (23 GPa) induced by the semiconductor-metal phase
transitions at high pressures. The temperature was monitored using a W
97
Re
3
-W
75
Re
25
thermocouple.
The electromotive force (EMF) was not corrected for the effects of pressure. The sample was compressed
to the desired pressure. Then, the oil pressure was held constant. Subsequently, alternating current (AC)
power was supplied to the heater in the furnace assemblage. After heating for 40–240 min, the power
was stopped by shutting off the electric power supply. Samples were recovered after releasing pressure
slowly during 12 hrs. The recovered run products were mounted in epoxy resin and were polished for
phase identification and chemical composition analysis.
The simplified Martian composition by Dreibus and Wänke [
20
], which is an analog of the primitive
Martian mantle composition corresponding to mantle +crust, was adopted. The oxide mixture of MgO,
Al
2
O
3
, SiO
2
, and FeO was prepared in appropriate proportions. FeO was put in the reduced furnace at
1000
◦
C for 24 h before mixing to ensure that the ferrous ion was used. We added 2 wt.% H
2
O in the
form of Mg(OH)
2
. The chemical compositions are presented in Table 1. To create a reduced environment,
we inserted some Mo foil in a gold capsule before the starting material was encapsulated into the capsule.
The phase assemblages were identified using a micro-focus X-ray diffractometer (MicroMax-007HF;
Rigaku Corp., Tokyo, Japan) with Cu K
α
radiation. The obtained data were processed using 2PD
software, which can display and process two-dimensional data, including smoothing, background
correction, and 2D to 1D conversion. The micro-textures and composition were obtained using a field
emission scanning electron microscope (FESEM, JSM7000F, JEOL, Akishima-shi, Japan) combined with
an energy-dispersive X-ray spectrometer (EDS, X-MaxN, Oxford Instruments, Plc., Abingdon, UK).
The Fe
2
SiO
4
, Mg
2
SiO
4
, and Al
2
O
3
were used as standards in the EDS analyses. Working parameters of
15 kV, 1 nA, and collection times of 30–50 s were used. The chemical composition was analyzed using
EDS. We used software (Aztec ver. 2.4, Oxford Instruments Nanotechnology Tools Ltd., Abingdon, UK)
to process EDS data. The Raman spectrum was obtained using laser Raman spectrometer (NRS-5100gr)
to identify some of the recovered phases, with 532 nm laser excitation. Laser power applied to the
sample was 10 mW. The Raman spectra were obtained from a linear baseline, and peak characteristics
were carried out using the commercial software package. The EDS measurements suggested that we
obtained the homogeneous composition of Mo where it appeared. On the other hand, we calculated the
chemical formula of garnet obtained in this study. The result showed that only the Fe
2+
in the formula
Minerals 2019,9, 559 3 of 9
could make the charge balance. Therefore, we created a reduced environment in the sample chamber by
using Mo foil.
Table 1. Chemical composition (wt.%) of starting materials.
Composition MgO Al2O3FeO SiO2H2O Total
MAFSH 30.2 3.5 19.9 44.4 2 100
MAFSH, in simplified MgO-Al2O3-FeO-SiO2-H2O composition.
3. Results and Discussion
3.1. Phase Relations
Experimental conditions and the results are presented in Table 2and Figure 1. Nominally anhydrous
phases, such as clinopyroxene (Cpx), garnet (Gt), wadsleyite (Wd), and ringwoodite (Rw), were presented
in quenched samples (Figure 2). Based on experimental conditions, different hydrous phases were
observed with increased pressure. Phase E (PhE) was observed in the low-temperature region between
15–16.5 GPa to coexist with Gt and Wd or Rw, as presented in Figure 2a. It became unstable at 16.5 GPa
with a temperature higher than 1100
◦
C. However, the thermal stability limit of PhE in this study was
about 100
◦
C higher than that reported in iron-free hydrous peridotite system [
21
]. Phase D (PhD)
remained as a major hydrous phase at the present pressure range from 18 GPa to 21 GPa below 1100
◦
C
(Figure 2b). With increased temperature, PhD decomposed to Rw, stishovite (St), and Gt. It seems that
PhD in MAFSH system has the same stability region, as reported for a MASH system [
21
]. Therefore,
it was expected that the PhD had a positive pressure-temperature stability slope, as presented in Figure 1.
The high iron content in MAFSH system might inhibit the superhydrous phase B (SuB) formation, leading
to the disappearance of SuB in the whole pressure range.
An earlier report described that the loop in the Wd-Rw boundary shifted towards higher pressure
by the effect of water [
22
] or towards lower pressure with increased iron content [
23
]. In this study,
Wd was stable up to 16 GPa at 1200
◦
C; then it transformed to Rw, as identified by Raman spectrum.
The Rw formation was also found shifting to low pressure, which was observed at 16 GPa and 1200
◦
C.
Therefore, this loop was influenced more strongly by the higher iron content than by the effects of H
2
O
in the hydrous Martian mantle.
Table 2. Experimental conditions and results.
Pressure (GPa) Temperature (◦C) Time (min) Phase
15 900 240 Gt, Wd, PhE
15 1100 240 Gt, Wd, Cpx
15 1250 120 Gt, Wd, Cpx
15 1450 90 Gt, Wd, Cpx
16 1200 120 Gt, Wd, Rw, Cpx
16.5 1100 120 Gt, Rw, PhE, St
16.5 1300 90 Gt, Rw, Cpx
18 1000 240 Rw, PhD
18 1200 120 Gt, Rw, St
18 1550 40 Gt, Rw, St, Melt
19.5 1400 40 Gt, Rw, St
21 900 240 Rw, PhD, St
21 1050 240 Rw, PhD, h-Fe
21 1250 120 Gt, Rw, St
21 1500 40 Gt, Rw, St
Gt, garnet; Cpx, clinopyroxene; Wd, wadsleyite; Rw, ringwoodite; St, stishovite; PhE, phase E; PhD, phase D; h-Fe,
iron-rich hydrous phase.
Minerals 2019,9, 559 4 of 9
Minerals 2019, 8, FOR PEER REVIEW 4 of 9
Cpx was observed at 16.5 GPa and 1300 °C, which further transformed to Gt at higher pressures.
Stishovite (St) was found to be coexisting with Rw at temperatures higher than 1100 °C. Some amount
of ferrous oxide was detected at 21 GPa and 1050 °C. At 21 GPa and 1000 °C, we found iron-rich
hydrous phase, based on analyzing the deficit of weight total. Both, the total weight and diffraction
pattern, were similar to εFeOOH; however, the diffraction peaks were slightly shifted compared with
εFeOOH, probably because of the incorporation of Mg due to the similar ionic radius of VIMg2+ (0.72
Å) and VIFe2+ (0.78 Å). It was also a little strange that this phase did not contain Si and Al. Perhaps,
the solubility of Al and Si decreases when Mg is included. More work is needed to clarify this issue.
Figure 1. Phase relations in the system MgO-Al2O3-FeO-SiO2 with 2% H2O. Solidus lines are obtained
according to the quenched samples combined with reported phase relation in CaO-MgO-Al2O3-SiO2-
pyrolite with 2% H2O [24]. Dashed black lines are the proposed phase boundaries based on phase
assemblages in recovered samples due to the limited data point at these regions. The dashed red line
is proposed areotherm [17,25]. Gt, garnet; Cpx, clinopyroxene; Wd, wadsleyite; Rw, ringwoodite; St,
stishovite; PhE, phase E; PhD, phase D; h-Fe, iron-rich hydrous phase.
Figure 2. Backscattered electron images of representative run products under various pressure and
temperature conditions: (a) 15 GPa and 900 °C; (b) 21 GPa and 1050 °C; (c) 15 GPa and 1250 °C; (d) 21
GPa and 1500 °C. Gt, garnet; Cpx, clinopyroxene; Wd, wadsleyite; Rw, ringwoodite; St, stishovite;
PhE, phase E; PhD, phase D; h-Fe, iron-rich hydrous phase.
Figure 1.
Phase relations in the system MgO-Al
2
O
3
-FeO-SiO
2
with 2% H
2
O. Solidus lines are obtained
accordingtothequenchedsamples combinedwith reported phaserelationinCaO-MgO-Al
2
O
3
-SiO
2
-pyrolite
with 2% H
2
O [
24
]. Dashed black lines are the proposed phase boundaries based on phase assemblages
in recovered samples due to the limited data point at these regions. The dashed red line is proposed
areotherm [
17
,
25
]. Gt, garnet; Cpx, clinopyroxene; Wd, wadsleyite; Rw, ringwoodite; St, stishovite; PhE,
phase E; PhD, phase D; h-Fe, iron-rich hydrous phase.
Minerals 2019, 8, FOR PEER REVIEW 4 of 9
Cpx was observed at 16.5 GPa and 1300 °C, which further transformed to Gt at higher pressures.
Stishovite (St) was found to be coexisting with Rw at temperatures higher than 1100 °C. Some amount
of ferrous oxide was detected at 21 GPa and 1050 °C. At 21 GPa and 1000 °C, we found iron-rich
hydrous phase, based on analyzing the deficit of weight total. Both, the total weight and diffraction
pattern, were similar to εFeOOH; however, the diffraction peaks were slightly shifted compared with
εFeOOH, probably because of the incorporation of Mg due to the similar ionic radius of VIMg2+ (0.72
Å) and VIFe2+ (0.78 Å). It was also a little strange that this phase did not contain Si and Al. Perhaps,
the solubility of Al and Si decreases when Mg is included. More work is needed to clarify this issue.
Figure 1. Phase relations in the system MgO-Al2O3-FeO-SiO2 with 2% H2O. Solidus lines are obtained
according to the quenched samples combined with reported phase relation in CaO-MgO-Al2O3-SiO2-
pyrolite with 2% H2O [24]. Dashed black lines are the proposed phase boundaries based on phase
assemblages in recovered samples due to the limited data point at these regions. The dashed red line
is proposed areotherm [17,25]. Gt, garnet; Cpx, clinopyroxene; Wd, wadsleyite; Rw, ringwoodite; St,
stishovite; PhE, phase E; PhD, phase D; h-Fe, iron-rich hydrous phase.
Figure 2. Backscattered electron images of representative run products under various pressure and
temperature conditions: (a) 15 GPa and 900 °C; (b) 21 GPa and 1050 °C; (c) 15 GPa and 1250 °C; (d) 21
GPa and 1500 °C. Gt, garnet; Cpx, clinopyroxene; Wd, wadsleyite; Rw, ringwoodite; St, stishovite;
PhE, phase E; PhD, phase D; h-Fe, iron-rich hydrous phase.
Figure 2.
Backscattered electron images of representative run products under various pressure and
temperature conditions: (
a
) 15 GPa and 900
◦
C; (
b
) 21 GPa and 1050
◦
C; (
c
) 15 GPa and 1250
◦
C; (
d
)
21 GPa and 1500
◦
C. Gt, garnet; Cpx, clinopyroxene; Wd, wadsleyite; Rw, ringwoodite; St, stishovite;
PhE, phase E; PhD, phase D; h-Fe, iron-rich hydrous phase.
Cpx was observed at 16.5 GPa and 1300
◦
C, which further transformed to Gt at higher pressures.
Stishovite (St) was found to be coexisting with Rw at temperatures higher than 1100
◦
C. Some amount
of ferrous oxide was detected at 21 GPa and 1050
◦
C. At 21 GPa and 1000
◦
C, we found iron-rich
hydrous phase, based on analyzing the deficit of weight total. Both, the total weight and diffraction
pattern, were similar to
ε
FeOOH; however, the diffraction peaks were slightly shifted compared with
ε
FeOOH, probably because of the incorporation of Mg due to the similar ionic radius of
VI
Mg
2+
(0.72 Å)
Minerals 2019,9, 559 5 of 9
and
VI
Fe
2+
(0.78 Å). It was also a little strange that this phase did not contain Si and Al. Perhaps,
the solubility of Al and Si decreases when Mg is included. More work is needed to clarify this issue.
3.2. Mineral Chemistry in DHMSs, Wadsleyite, and Ringwoodite
The measured chemical compositions of the phases in the experiments are presented in Table 3.
Several phases exhibited broad compositional variations. PhE had a composition of 4.5–3.1 wt.% for
Al
2
O
3
and 12.8–8.5 wt.% for FeO, at elevated pressures and temperatures from 15 GPa and 900
◦
C to
16.5 GPa and 1100
◦
C. With increasing pressure from 18 GPa to 21 GPa, the Al
2
O
3
content decreased from
7.9 to 4.7 wt.%
. However, FeO remained almost stable at around 4.3 wt.%. The water contents in PhE and
PhD were both estimated for 13 wt.% on average, based on the deficit from the EDS weight total (each
phase was calculated by summing all the deficit of weight total in Table 3, and then divided by counted
number). PhD in the MAFSH system generally had low amounts of
FeO <4.4 wt.%
. The Al
2
O
3
contents
were 3.5–7.9 wt.%, whereas the FeO contents exhibited a small variation of
3.9–4.8 wt.%
throughout the
samples quenched under various pressure and temperature conditions (Table 3).
Table 3. Representative mineral compositions.
P(GPa) T(◦C) Phase MgO Al2O3SiO2FeO Total
21
1500
Gt 26.48 (54) 13.27 (142) 48.63 (84) 12.33 (39) 100.71 (23)
Rw 36.20 (29) 0 38.11 (32) 25.42 (45) 99.73 (66)
St 0 1.45 (46) 100.68 (81) 0.77 (6) 102.90 (52)
1250
Gt 25.01 (60) 15.21 (142) 46.27 (85) 12.34 (37) 98.84 (33)
Rw 35.72 (33) 0 37.49 (23) 24.95 (43) 98.15 (58)
St 38.20 0.55 (38) 99 (63) 0.78 (33) 100.33 (83)
1050
Rw 41.60 (98) 0 39.27 (62) 15.99 (27) 96.86 (88)
PhD 20.24 (42) 4.71 (71) 56.11 (85) 4.79 (34) 85.86 (145)
h-Fe 7.25 (33) 0 0.87 (60) 81.74 (106) 89.87 (120)
900
Rw 43.57 (86) 1.33 (7) 40.54 (73) 12.11 (87) 96.22 (143)
PhD 29.36 (78) 3.50 (48) 52.95 (90) 3.93 (48) 89.74 (81)
St 1.53 (71) 0 99.08 (78) 1.15 (27) 101.72 (58)
19.5 1400
Gt 26.02 (39) 12.76 (65) 47.78 (60) 12.21 (39) 98.78 (92)
Rw 35.26 (13) 0 36.96 (19) 25.55 (34) 97.77 (44)
St 0 0.96 (47) 98.42 (58) 0.72 (32) 100.10 (81)
18 1550
Gt 27.78 (29) 11.04 (37) 49.70 (30) 12.45 (80) 100.45 (37)
Rw 34.88 (37) 2.05 (18) 37.77 (30) 26.70 (24) 99.35 (58)
St 0 1.54 (58) 99.90 (97) 0.62 (36) 102.05 (63)
Melt 20.65 1.62 16.45 16.57 55.31
18 1200
Gt 25.88 (90) 12.61 (99) 47.96 (67) 12.41 (64) 98.86 (76)
Rw 35.55 (35) 0 37.56 (34) 25.61 (47) 98.71 (79)
St 0 0.58 (23) 99.49 (71) 0.70 (40) 100.77 (82)
18 1000 Rw 32.89 (61) 0 36.98 (46) 27.86 (81) 97.73 (57)
PhD 20.48 (75) 7.86 (59) 54.68 (73) 4.57 (46) 87.59 (53)
16.5 1300
Gt 25.42 (32) 11.23 (48) 47.18 (29) 12.76 (57) 96.59 (50)
Rw 28.89 (32) 0 34.88 (27) 32.21 (25) 95.98 (64)
Cpx 33.79 (24) 19.19 (60) 55.66 (53) 6.98 (47) 96.44 (95)
16.5 1100
Gt 25.76 (80) 12.81 (36) 47.64 (77) 14.74 (57) 100.94 (63)
Rw 34.59 (57) 0 37.72 (17) 26.67 (63) 98.97 (37)
PhE 38.42 (75) 3.12 (20) 37.82 (18) 8.47 (28) 87.82 (57)
St 1.38 (67) 0.76 (64) 96.75 (75) 1.14 (20) 100.03 (27)
Minerals 2019,9, 559 6 of 9
Table 3. Cont.
P(GPa) T(◦C) Phase MgO Al2O3SiO2FeO Total
16 1200
Gt 24.28 (47) 13.42 (37) 46.67 (33) 14.26 (83) 98.63 (44)
Rw 28.36 (42) 0 35.76 (39) 33.43 (59) 97.55 (69)
Wd * - - - - -
Cpx 34.99 (51) 20.70 (54) 56.73 (58) 6.52 (37) 98.25 (37)
15 1450
Gt 26.61 (71) 11.35 (25) 49.48 (61) 13.10 (75) 100.54 (82)
Wd 34.43 (35) 0 37.35 (33) 27.61 (46) 99.40 (81)
Cpx 34.68 (74) 0 58.09 (41) 7.56 (70) 100.34 (82)
15 1250
Gt 27.39 (13) 11 (54) 49.79 (37) 13.06 (34) 101.24 (40)
Wd 30.91 (34) 0 36.54 (18) 32.11 (47) 99.55 (46)
Cpx 35.80 (39) 0 58.17 (57) 6.65 (35) 100.61 (67)
15 1100
Gt 25.19 (57) 12.32 (28) 47.52 (64) 13.08 (61) 98.11 (81)
Wd 31.03 (96) 0 36.49 (80) 30.23 (94) 97.76 (46)
Cpx 35.32 (70) 30.91 (34) 56.57 (72) 6.27 (81) 98.16 (13)
15 900
Gt 27.61 (31) 12.45 (79) 44.08 (83) 17.22 (96) 101.35 (74)
Wd 27.91 (46) 0 36.15 (50) 35.08 (27) 99.14 (84)
PhE 34.01 (84) 4.54 (26) 35.92 (47) 12.79 (54) 87.27 (79)
* Measuring the chemical composition by EDS was difficult because of the small crystal size. This phase was
identified by Raman spectrum. Gt, garnet; Cpx, clinopyroxene; Wd, wadsleyite; Rw, ringwoodite; St, stishovite;
PhE, phase E; PhD, phase D; h-Fe, iron-rich hydrous phase.
The chemical composition changes in Rw and Gt that occur with increasing temperature at some
different pressures are shown in Figure 3. The MgO and SiO
2
contents decreased concomitantly with
increasing temperature and then increased slightly in Rw at 16 and 21 GPa (Figure 3a). However,
the opposite trend was observed for FeO content. At 18 GPa, opposite trends were observed in MgO,
SiO
2
, and FeO contents below 1250
◦
C compared to pressures at 16 and 21 GPa. Both Wd and Rw have
near-stoichiometric bulk composition. However, the (Mg +Fe)/Si ratio of Wd and Rw was lower than 2,
indicating incorporation of H
+
. The H
2
O contents in Rw were greater than those in Wd, based on
deficit total weight estimation. Generally, the Al
3
O
2
and SiO
2
contents in Gt exhibited an opposite
tendency because of Tschermak substitution (Mg
2+
+Si
4+
=2Al
3+
). The Al
3
O
2
content decreased,
and the SiO
2
content increased concomitantly with increasing temperature. However, with increasing
pressure, the Al
3
O
2
content increased, and the SiO
2
content decreased (Figure 3b). The FeO content
remained fundamentally unchanged in all quenched samples, except under the condition of 15 GPa
and 900 ◦C (Table 3).
Minerals 2019, 8, FOR PEER REVIEW 6 of 9
St 1.38 (67) 0.76 (64) 96.75 (75) 1.14 (20) 100.03 (27)
16 1200
Gt 24.28 (47) 13.42 (37) 46.67 (33) 14.26 (83) 98.63 (44)
Rw 28.36 (42) 0 35.76 (39) 33.43 (59) 97.55 (69)
Wd * - - - - -
Cpx 34.99 (51) 20.70 (54) 56.73 (58) 6.52 (37) 98.25 (37)
15 1450
Gt 26.61 (71) 11.35 (25) 49.48 (61) 13.10 (75) 100.54 (82)
Wd 34.43 (35) 0 37.35 (33) 27.61 (46) 99.40 (81)
Cpx 34.68 (74) 0 58.09 (41) 7.56 (70) 100.34 (82)
15 1250
Gt 27.39 (13) 11 (54) 49.79 (37) 13.06 (34) 101.24 (40)
Wd 30.91 (34) 0 36.54 (18) 32.11 (47) 99.55 (46)
Cpx 35.80 (39) 0 58.17 (57) 6.65 (35) 100.61 (67)
15 1100
Gt 25.19 (57) 12.32 (28) 47.52 (64) 13.08 (61) 98.11 (81)
Wd 31.03 (96) 0 36.49 (80) 30.23 (94) 97.76 (46)
Cpx 35.32 (70) 30.91 (34) 56.57 (72) 6.27 (81) 98.16 (13)
15 900
Gt 27.61 (31) 12.45 (79) 44.08 (83) 17.22 (96) 101.35 (74)
Wd 27.91 (46) 0 36.15 (50) 35.08 (27) 99.14 (84)
PhE 34.01 (84) 4.54 (26) 35.92 (47) 12.79 (54) 87.27 (79)
* Measuring the chemical composition by EDS was difficult because of the small crystal size. This
phase was identified by Raman spectrum. Gt, garnet; Cpx, clinopyroxene; Wd, wadsleyite; Rw,
ringwoodite; St, stishovite; PhE, phase E; PhD, phase D; h-Fe, iron-rich hydrous phase.
The chemical composition changes in Rw and Gt that occur with increasing temperature at some
different pressures are shown in Figure 3. The MgO and SiO2 contents decreased concomitantly with
increasing temperature and then increased slightly in Rw at 16 and 21 GPa (Figure 3a). However, the
opposite trend was observed for FeO content. At 18 GPa, opposite trends were observed in MgO,
SiO2, and FeO contents below 1250 °C compared to pressures at 16 and 21 GPa. Both Wd and Rw
have near-stoichiometric bulk composition. However, the (Mg + Fe)/Si ratio of Wd and Rw was lower
than 2, indicating incorporation of H+. The H2O contents in Rw were greater than those in Wd, based
on deficit total weight estimation. Generally, the Al3O2 and SiO2 contents in Gt exhibited an opposite
tendency because of Tschermak substitution (Mg2+ + Si4+ = 2Al3+). The Al3O2 content decreased, and
the SiO2 content increased concomitantly with increasing temperature. However, with increasing
pressure, the Al3O2 content increased, and the SiO2 content decreased (Figure 3b). The FeO content
remained fundamentally unchanged in all quenched samples, except under the condition of 15 GPa
and 900 °C (Table 3).
(a) (b)
Figure 3.
Chemical composition changes in ringwoodite (
a
) and garnet (
b
) with increasing temperature
at different pressures.
Minerals 2019,9, 559 7 of 9
3.3. Stability and Water Contents of Hydrous Phases in Iron-Rich Martian Mantle
Several reports of petrological studies have described that dense hydrous magnesium silicates
(DHMSs) remain stable in the hydrous pyrolite mantle compositions along a cold subducting slab [
1
,
3
,
21
].
In the present study, we observed that the stability regions of DHMSs in the hydrous iron-rich Martian
mantle (2 wt.% H
2
O) were generally consistent with those obtained in CaO-MgO-Al
2
O
3
-SiO
2
(CMAS)
pyrolite with 2 wt.% of H
2
O and a water-saturated MSH system. However, some differences
were apparent.
The stability region of PhE partially overlapped with Wd or Rw below 17 GPa in the low-temperature
field. Although Wd could hold up to 3 wt.% of H
2
O in its crystal structure [
24
], it only accommodated
approx. 0.8 wt.% of H2O in a water-undersaturated condition at 15 GPa and 900 ◦C. The water content
increased drastically to approx. 2.2 wt.% after PhE decomposed at elevated temperature 1100
◦
C and
15 GPa. It subsequently decreased to approx. 0.6 wt.% with the temperature increased to 1450
◦
C.
The water content in Wd was consistent with those proposed in CMAS pyrolite with 2 wt.% of H
2
O at
15 GPa and 1450
◦
C [
24
], indicating that Wd also has large water storage capacity within the stability
field of the Martian mantle.
Reportedly, SuB appears in CMAS or MHS systems at pressures higher than 17 GPa [
24
], but
we observed PhD instead of SuB at 18–21 GPa, which means that much water might be held in the
low-temperature region because of the higher water solubility in PhD. It is expected that PhD might
transport water to the deepest part of the Martian interior because PhD was reported to be stable up to
44 GPa, which corresponds to a depth of 1250 km [6].
Hydrous Rw appeared at pressures higher than 16.5 GPa, which exhibited a wider stability
region than that in hydrous pyrolite. It seemed readily apparent that water contents in Rw decreased
concomitantly with increasing temperature, as shown in Table 2. We observed trace amounts of melt
at 18 GPa and 1550
◦
C, suggesting that it was very close to the wet solidus in MAFS system under
a water-unsaturated condition. We also observed some amount of iron-rich hydrous phase at 21 GPa
and 1050
◦
C, implicating it is a potential water carrier in low-temperature regions of the hydrous
iron-rich Martian mantle. Because of its greater density than its surrounding materials, it might bring
water to the Martian core.
4. Implications
The existence of water on Mars has long been controversial. Some evidence has shown that water
disappeared from the Mars surface after its formation [
17
,
26
,
27
]. Although the whereabouts of the
water have long been debated, some parts of the hydrated crust of Mars may have been brought into
the deep interior by the convecting mantle. We inferred that DHMSs might act as an important water
carrier in the deep region of Fe-rich Mars, although the temperature profile of the Mars interior remains
unclear. Several models have been proposed to constrain the structure of the Martian interior based on
geophysical observations and high-pressure petrological studies. Then, they suggested the pressure of
the core-mantle boundary as 19–25 GPa [
25
,
28
,
29
]. Our results demonstrated that DHMSs could be
expected to transport water even to the iron-rich Martian core in the cold region in this pressure range.
PhE and PhD were potentially relevant DHMSs in the Martian transition zone in a simplified
MAFSH model. The DHMSs would dehydrate completely if the temperature of the surrounding
Martian mantle was higher than their stability limits (Figure 2). The released water would be stored
in Wd or Rw. These phases might subsequently act as a large water reservoir in the Martian mantle,
as in the Earth’s mantle. The results of the present phase relation in iron-rich Martian mantle were
fundamentally consistent with data for hydrous pyrolite, except for the SuB absence in the present
study. In addition, it is noteworthy that much more water could be held in the Martian mantle than in
Earth’s mantle because the phase transition pressure of the loop in both olivine (Ol)–Wd and Wd–Rw
boundary shifted towards lower pressure [22,24].
Furthermore, it seems that the thermal structure of Mars is key to understand water cycles
in a deeper region; however, the areotherm of Mar’s is still under debate [
17
,
25
,
28
]. In this study,
Minerals 2019,9, 559 8 of 9
we adopted a model areotherm, as shown in Figure 1, which was determined according to the present
day core-mantle boundary temperature of ~1400
◦
C suggested by Hauck and Phillips [
17
]. In this
model, Mars was slowly cooling down from 4.5 Ga to the present whose areotherm was still at least
higher than the dehydration temperature of PhE and PhD, as shown in Figure 1. During the evolution
of early Mars, much of the water was possibly lost because of low melting temperatures of DHMSs,
hydrous Wd, and Rw compared with areotherm [
17
]. In addition, water is an incompatible and volatile
component in a solid-melt system, which easily causes partial melting of Martian mantle. Water also
affects the thermal evolution of the planet. The generated magma may bring the released water to
the shallow region, and help conduct and transfer heat more efficiently. This process may help early
Martian mantle quickly cool down than a dry system. By contrast, this may become a barrier for
water transportation at least to some critical depth during some part of Martian history. However,
the accumulated water in the shallow region may hydrate some part of the crust. Unlike the Earth
system in which water is recycled by subducting slabs, the mantle of Mars might be convected, but the
crust is stagnant and not subducted [
17
]. The mantle might be hydrated by reaction with the overlying
hydrous crust. Eventually, the water should be restored in Wd and Rw during mantle convection due
to their high thermal stability region than that of DHMSs.
Our present result indicated that both DHMSs and nominally anhydrous minerals, Wd and Rw,
have the potential to accommodate a certain amount of water, elucidating model geodynamic processes
associated with the deepwater cycles of Mars.
Author Contributions:
Conceptualization, C.X.; data curation, C.X.; writing (original draft preparation), C.X.;
writing (review and editing), T.I.; supervision, T.I.; funding acquisition, C.X. and T.I.
Funding:
This research was funded by JSPS KAKENHI Grant Numbers 18J12511 for C.X. and 26247073, 15H05828,
and 18H03740 for T.I. In addition, C.X. was supported by Research Fellowships of the Japan Society for the
Promotion of Science (JSPS) for Young Scientists (DC2).
Acknowledgments: The authors are grateful to Takeshi Arimoto for the preparation of starting materials.
Conflicts of Interest: The authors declare no conflict of interest.
References
1. Gasparik, T. The role of volatile in the transition zone. J. Geophys. Res. 1993,98, 4287–4299. [CrossRef]
2.
Ohtani, E.; Touma, M.; Litasov, K.; Kubo, T.; Suzuki, A. Stability of hydrous phases and water storage capacity
in the transitional zone and lower mantle. Phys. Earth Planet. Inter. 2001,124, 105–117. [CrossRef]
3.
Komabayashi, T.; Omori, S. Internally consistent thermodynamic dataset for dense hydrous magnesium
silicates up to 35 GPa, 1600
◦
C: Implications for water circulation in the Earth’s deep mantle. Phys. Earth
Planet. Int. 2006,156, 89–107. [CrossRef]
4.
Litasov, K.; Ohtani, E.; Sano, A.; Suzuki, A.; Funakoshi, K. Wet subduction versus cold subduction. Geophys.
Res. Lett. 2005,32, 1–5. [CrossRef]
5.
Ghosh, S.; Schmidt, M. Melting of phase D in the lower mantle and implications for recycling and storage of
H2O in the deep mantle. Geochim. Cosmochim. Acta 2014,145, 72–88. [CrossRef]
6.
Nishi, M.; Irifune, T.; Tsuchiya, J.; Tange, Y.; Nishihara, Y.; Fujino, K.; Higo, Y. Stability of hydrous silicate at
high pressures and water transport to the deep lower mantle. Nat. Geosci. 2014,7, 224–227. [CrossRef]
7.
Pamato, M.; Myhill, R.; Ballaran, T.; Frost, D.; Heidelbach, F.; Miyajima, N. Lower-mantle water reservoir
implied by the extreme stability of a hydrous aluminosilicate. Nat. Geosci. 2015,8, 75–79. [CrossRef]
8.
Walter, M.J.; Thomson, A.R.; Wang, W.; Lord, O.T.; Ross, J.; McMahon, S.C.; Baron, M.A.; Melekhova, E.;
Kleppe, A.K.; Kohn, S.C. The stability of hydrous silicates in Earth’s lower mantle: Experimental constraints
from the systems MgO–SiO2–H2O and MgO–Al2O3–SiO2–H2O. Chem. Geol. 2015,418, 16–29. [CrossRef]
9.
Liu, X.; Matsukage, K.; Nishihara, Y.; Suzuki, T.; Takahashi, E. Stability of the hydrous phases of Al-rich
phase D and Al-rich phase H in deep subducted oceanic crust. Am. Mineral. 2019,104, 64–72. [CrossRef]
10.
Inoue, T. Effect of water on melting phase relations and melt composition in the system Mg
2
SiO
4
-MgSiO
3
-H
2
O
up to 15 GPa. Phys. Earth Planet. Inter. 1994,85, 237–263. [CrossRef]
Minerals 2019,9, 559 9 of 9
11.
Cardenas, B.; Mohrig, D.; Goudge, T. Fluvial stratigraphy of valley fills at Aeolis Dorsa, Mars: Evidence for
base-level fluctuations controlled by a downstream water body. GSA Bull. 2018,130, 484–498. [CrossRef]
12.
Chan, N.; Perron, J.; Mitrovica, J.; Gomez, N. New Evidence of an Ancient Martian Ocean from the Global
Distribution of Valley Networks. J. Geophys. Res. 2018,123, 2138–2150. [CrossRef]
13.
Ivanov, M.; Erkeling, G.; Hiesinger, H.; Bernhardt, H.; Reiss, D. Topography of the deuteronilus contact on
mars: Evidence for an ancient water/mud ocean and long-wavelength topographic readjustments. Planet.
Space Sci. 2017,144, 49–70. [CrossRef]
14.
Byrne, S.; Dundas, C.; Kennedy, M. Distribution of mid-latitude ground ice on Mars from new impact craters.
Science 2009,325, 1674–1676. [CrossRef] [PubMed]
15.
Conway, S.J.; Balme, M.R.; Kreslavsky, M.A.; Murray, J.B.; Towner, M.C. The comparison of topographic
long profiles of gullies on Earth to gullies on Mars: A signal of water on Mars. Icarus
2016
,253, 189–204.
[CrossRef]
16.
Pan, L.; Ehlmann, B. Aqueous Processes from Diverse Hydrous Minerals in the Vicinity of Amazonian-Aged
Lyot Crater. J. Geophys. Res. 2018,123, 1618–1648. [CrossRef]
17.
Hauck, S.A.; Phillips, R.J. Thermal and crustal evolution of Mars. J. Geophys. Res. Planets
2002
,107. [CrossRef]
18.
Frost, D.J.; Fei, Y.W. Stability of phase D at high pressure and high temperature. J. Geophys. Res.
1998
,103,
7463–7474. [CrossRef]
19.
Ohira, I.; Ohtani, E.; Sakai, T.; Miyahara, M.; Hirao, N.; Ohishi, Y.; Nishijima, M. Stability of a hydrous
δ
-phase, AlOOH-MgSiO
2
(OH)
2
, and a mechanism for water transport into the base of lower mantle. Earth
Planet. Sci. Lett. 2014,401, 12–17. [CrossRef]
20. Dreibus, G.; Wänke, H. Mars, a volatile-rich planet. Icarus 1987,71, 225–240. [CrossRef]
21.
Ohtani, E.; Litasov, K.; Hosoya, T.; Kubo, T.; Kondo, T. Water transport into the deep mantle and formation of
a hydrous transition zone. Phys. Earth Planet. Inter. 2004,143–144, 255–269. [CrossRef]
22.
Litasov, K.D.; Ohtani, E.; Sano, A. Influence of water on major phase transitions in the Earth’s mantle. In Earth
Deep Water Cycle; Jacobsen, S.D., van der Lee, S., Eds.; Geophysical Monograph Series; Wiley Blackwell:
Hoboken, NJ, USA, 2006; Volume 168, pp. 95–111.
23.
Akaogi, M.; Ito, E.; Navrotsky, A. Olivine-modified spinel-spinel transitions in the system Mg
2
SiO
4
–Fe
2
SiO
4
:
Calorimetric measurements, thermochemical calculation, and geophysical application. J. Geophys. Res.
1989
,
94, 15671–15685. [CrossRef]
24.
Litasov, K.; Ohtani, E. Stability of various hydrous phases in CMAS pyrolite-H
2
O system up to 25 GPa. Phys.
Chem. Miner. 2003,30, 147–156. [CrossRef]
25.
Bertka, C.M.; Fei, Y.W. Mineralogy of Martian interior up to core-mantle boundary pressures. J. Geophys. Res.
1997,102, 5251–5264. [CrossRef]
26. Baker, V.R. Water and the Martian landscape. Nature 2001,412, 228–236. [CrossRef] [PubMed]
27.
Masson, P.; Carr, M.H.; Costard, F.; Greeley, R.; Hauber, E.; Jaumann, R. Geomorphologic evidence for liquid
water. Space Sci. Rev. 2001,96, 333–364. [CrossRef]
28.
Zharkov, V.N.; Gudkova, T.V. Seismic model of Mars: Effects of hydration. Planet. Space Sci.
2014
,104,
270–278. [CrossRef]
29.
Raevskiy, S.N.; Gudkova, T.V.; Zharkov, V.N. Diagnostic possibilities of body waves for studying the interior
structure of Mars. Izv-Phys. Solid Eart. 2015,51, 143–155. [CrossRef]
©
2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).
