Supplementary
Information
Negligible effects from sample preparation
No OH bands have been observed in the dehydrated grains of An
68
-
70
, An
96
, and GRR
2058 using FTIR. Furthermore,
the unheated An
68
-
70
and An
96
grains contain 10 and
7
ppm H
2
O,
respectively
(Table 1)
, while the dehydrated grains of An
68
-
70
, An
96
, and GRR 2058 all display
uniform OH
-
/
30
Si
-
values, lower than that of the unheated GRR 2058 (~1 ppm) (Fig. 2).
Therefore,
these dehydrated grains have lost their original OH contents, any
hydrogen
detected
in the dehydrated grains is assumed to be background in the SIMS measurements, and sample
preparation has not introduced additional
hydrogen
signal.
Hydrogen isotope
fractionation between mineral and melt
No experimental studies have been carried out on the fractionation of hydrogen isotope
between mineral and melt. The fractionation during mantle melting and magma differentiation,
however, have been discussed using
hydrogen isotope ratios of fresh submarine basalts
(
Kyser
and O’Neil, 1984; Bindeman et al., 2012
)
. It has been suggested differentiation or partial
melting have had little effect on the hydrogen isotope fractionation
(
Kyser and O’Neil, 1984;
Bindeman et
al., 2012
)
. Melt
-
lherzolite fractionation of hydrogen isotope has been suggested to
be less than 10‰ during mantle melting
(
Bindeman et al., 2012
)
. The uncertainties of
δ
D
of
undegassed lunar minerals are rather large, typically >100‰ (
Barnes et al., 201
4
).
Therefore, the
δ
D
of lunar plutonic minerals could represent those of their parent magmas.
The
δ
D
value of GRR1968
The low water content (70 ppm) of GRR 1968 has made the determination of its
hydrogen isotope ratio
using traditional method
(
e.g.,
mass spectrometer)
on bulk sample
difficult. In this study, we use an alternative approach to calculate the
δ
D
value of GRR
1968.
GRR
1968 is an anorthite (An
93.8
Ab
4.0
with 0.37% of FeO) megacryst of island arc basalt
from Miyake Island of Izu
-
Bonin
-
Mariana
island arc
(
Kimata et al.,
1995; Johnson and Rossman,
2003
)
. Because the fractionation of hydrogen isotope between mineral and melt is very small
(see text above), we could use the melt
δ
D
value to represent that of GRR
1968. No hydrogen
isotope ratios have been published for arc basalts from Miyake Island. Therefore, we use the
average
δ
D
value (
-
42
±
29‰) of samples from Izu
-
Bonin
-
Mariana arc system found in the
literature. This dataset includes 87 samples with
δ
D
va
rying from
-
12 to
-
73.7‰ (Table S
4
, and
references therein). Furthermore, it has been suggested that the fractionation of hydrogen
isotope between mineral and melt is less t
han 10‰
(Kyser and O’Neil, 1984; Bindeman et al.,
2012)
.
Therefore, w
e estimate t
hat the
δ
D
value of GRR
1968 is
-
42
±
39‰. This
δ
D
estimation
is reasonable and
sufficiently precise
for the purpose of our study, because most of the
uncertainties in the H isotopic results of lunar plagioclases are from the correction for
cosmogenic produ
ction of hydrogen. Furthermore, the
δ
D
of
An
96
, a plagioclase from Aleutian
arc melt
(
Waythomas et al., 2010
)
, is
-
1
0
±
90‰ using this standard GRR
1968 (
-
42
±
39‰),
overlapping with those (
-
12 to
-
73.7‰) from Izu
-
Bonin
-
Mariana arc system (Table S
4
, and
references therein).
E
stimation of
the
initial LMO water content
The
plagioclase of ferroan anorthosite 60015
contains
5
±
1
ppm water
analyzed
by
SIMS.
Using a plagioclase
-
melt partition coefficient
(Hamada et al., 2013; 0.005
±
0.003 if
the
2
latest infrared
absorption coefficient determined for water in plagioclase by Mosenfelder et al.
(2015) is used)
, t
he
LMO
melt
that
equilibrated
with the ferroan anorthositic
plagioclase
could
have
contained
10
00
ppm water.
The relatively large uncertainty primarily comes from that of
partition coefficient.
Water in the LMO could be lost through degassing into a vacuum. On the
other hand,
water in the
LMO melt
could increase as the LMO solidification continued.
If only
the
se two scenarios were considered, w
e can infer the initial LMO water content
.
The
hydrogen species degassed in the Moon has been suggested to be H
2
(
Sharp et al.,
2013
)
.
The
fraction
of
hydrog
en (equivalent to water)
lost
from
the LMO
in
to
a vacuum
before
the crystallization of
ferroan anorthosite
can be estima
ted using the
Rayleigh
fractionation
equation
!
!
"
=
퐹
%
&
%
'
(
)
, where
R
is the
D
/
H
ratio of the LMO when a fraction
F
of hydrogen
remains in the LMO, and
R
0
is the initial hydrogen isotope ratio of the LMO, and
M
1
and
M
2
are
the masses of the volatile phase isotopologues
(
masses
of
2
.016
for
H
2
, and 3
.022
for
H
D
; Sharp
et al., 2013
)
.
The
δ
D
of
-
28
1
±
49
‰
(represented
by sample
77215
) is used as the initial
hydrogen
isotope ratio of the LMO
, whereas
+31
0
±
110
‰
(represented by sample 60015)
is
used as the final hydrog
en isotope ratio of the LMO
.
Therefore
, the fraction of hydrogen
remained in the LMO
when
60015 crystallized
is
3.8
%
.
Furthermore,
20
±
5
% of the LMO
may
have
remained melt
when ferroan anorthositic plagioclase crystallized
(
Shearer et al., 2006;
Elkins
-
Tanton et al., 2011
)
.
Note almost all of the water
undegassed from
the LMO
(i.e.,
3.8
% of
initial LMO water)
should
have still remained in the LMO melt
residue.
Therefore
, the initial
LMO water content could have been
up to
50
00
±
+,
--
//
--
ppm.
If a new partition coefficient of
0.02
±
0.002 determined under the lunar condition by Caseres et al. (2017), which is also in the
range of 0.006
–
0.04 det
ermined by Lin et al. (2017), i
s used, the
calculated
initial LMO water
content would be 1320
±
/0-
)+1-
ppm
. However, if a
generic
partition coefficient of 0.
001
used by
Elkins
-
Tanton and Grov
e (2011) in their LMO model
is
also
used
here
, the initial LMO water
con
tent would be
~2.6 wt
%. Ther
efore, an accurate hydrogen
partition coefficient between
plagioclase and lu
nar melt is needed to model the degassing process in the LMO.
References:
Barnes, J.J., et al., 2014. The origin of water in the primitive Moon as revealed by the lunar
highlands samples.
Earth Planet. Sci. Lett.
390,
244
-
252.
Bindeman, I.N., Kamenetsky, V.S., Palandri, J., Vennemann, T., 2012. Hydrogen and oxygen
isotope behaviors during variable degrees of upper mantle melting: Example from the
basaltic glasses from Macquarie Island. Chem. Geol. 310
-
311, 126
-
136.
Caseres, J.
R., Mosenfelder, J. L., Hirschmann, M. M., 2017. Partitioning of hydrogen and
fluorine between feldspar and melt under the conditions of lunar crust formation. 48
th
Lunar Planet. Sci. Conf., Abstract #2303.
Elkins
-
Tanton, L.T., Burgess, S., Yin, Q.
-
Z., 201
1. The lunar magma ocean: Reconciling the
solidification process with lunar petrology and geochronology. Earth Planet. Sci. Lett.
304, 326
-
336.
Elkins
-
Tanton, L.T., Grove, T.L., 2011. Water (hydrogen) in the lunar mantle: Results from
petrology and magma
ocean modeling. Earth Planet. Sci. Lett. 307, 173
-
179.
Hamada, M., Ushioda, M., Fujii, T., Takahashi, E., 2013. Hydrogen concentration in plagioclase
as a hygrometer of arc basaltic melts: Approaches from melt inclusion analyses and
hydrous melting experim
ents. Earth Planet. Sci. Lett.
365, 253
-
262.
3
Johnson, E.A., Rossman, G.R., 2003.
The concentration and speciation of hydrogen in feldspars
using FTIR and
1
H MAS NMR spectroscopy.
Am. Mineral. 88, 901
-
911.
Kimata, M., et al., 1995.
Anorthite megacrysts from
island arc basalts.
Mineral. Mag. 59, 1
-
14.
Kyser, T.K., O’Neil, J.R., 1984.
Hydrogen isotope systematics of submarine basalts.
Geochim.
Cosmochim. Acta 48, 2123
-
2133.
Lin, Y., et al., 2017. A lunar hygrometer based on plagioclase
-
melt partitioning of h
ydrogen. 48
th
Lunar Planet. Sci. Conf., Abstract #1286.
Sharp, Z.D., McCubbin, F.M., Shearer, C.K., 2013. A hydrogen
-
based oxidation mechanism
relevant to planetary formation. Earth Planet. Sci. Lett.
380, 88
-
97.
Shearer, C.K., et al., 2006. Thermal and ma
gmatic evolution of the Moon. Rev. Mineral.
Geochem. 60, 365
-
518.
Waythomas, C.F., et al., 2010. The 7
-
8 August 2008 eruption of Kasatochi Volcano, central
Aleutian Islands, Alaska. J. Geophys. Res. 115, B00B06, doi:10.1029/2010JB007437.