A COMPREHENSIVE SIMS STUDY OF HYDROGEN, FLUORINE, AND CHLORINE IN NOMINALLY
ANHYDROUS MINERALS FROM 15 LUNAR SAMPLES. J. L. Mosenfelder, J. R. Caseres, and M. M.
Hirschmann, Department of Earth Sciences, University of Minnesota, 310 Pillsbury Dr. SE, Minneapolis, MN,
55455 (jmosenfe@umn.edu).
Introduction: The role of volatile elements in the
origin and differentiation of the Moon remains contro-
versial [e.g., 1,2,3], despite a burgeoning number of
recent studies on lunar materials using modern, low
detection limit analytical techniques. One of the most
surprising findings is the FTIR measurement by Hui et
al. [4] of trace amounts of H in plagioclase (Pl) from
three Apollo samples: one troctolite (76535) and two
ferroan anorthosites (15415 and 60015). This H was
inferred to have been partitioned from the lunar mag-
ma ocean (LMO) and then preserved throughout an
~4.4 billion year history including cooling of the LMO
and surface processes such as cataclasis and shock
from impact events. However, preliminary data we
presented at LPS XXXXVII [5] and new results herein
challenge details of this hypothesis.
Using ultra-low blank SIMS, we obtained ~250
new analyses of H and F in 15 Apollo samples, includ-
ing the three studied by [4,6]. Cl was also analyzed in
selected samples. Although we focused on Pl from the
ferroan anorthosite suite (FAN; 13 samples total), we
also investigated a norite (77215) and troctolite
(76535) from the Mg-suite.
Methods: SIMS was performed on the Cameca 7f-
GEO at Caltech using previously outlined methodolo-
gy [7] enabling us to achieve limits of detection (LOD)
for H
2
O and F of ~ 1 ppmw and 0.1 ppmw, respective-
ly. The effective LOD for H
2
O in anorthitic Pl varied
from 1.5 to 2 ppmw, owing to differences in calibra-
tion between anorthite and more Na-rich feldspars [7]
and slightly worse vacuum conditions compared to
analyses reported in [5]. Cl was under the LOD with
the exception of some analyses of 60015 (see below).
Particular attention was paid to screening analyses for
organic surface contamination resulting from sample
preparation, owing to extremely low inherent H con-
centrations. We also examined SIMS craters after the
session using BSE imaging and EDS mapping in order
to assess possible contamination from cracks and con-
firm phase identifications where necessary (Fig. 1).
Samples: The FAN samples for which we obtained
uncontaminated analyses of anorthitic Pl are: 15415,
60015, 60618, 60619, 60639, 62237, 65315, 65325,
65326, 67635, 67746, 68515, and the "big white" anor-
thosite clast from 73255, 251 [8]. Some of these FAN
exhibit high degrees of shock (e.g., 60015), were con-
tained within impact melt (e.g., 60618), or have grano-
blastic textures (60619, 67746) suggesting protracted
cooling histories after crystallization, allowing us to
test the effects of these secondary processes on H and
F retention.
From the Mg-suite, we measured newly polished Pl
and Ol grains from troctolite 76535 [cf. 5] as well as Pl
and Opx from noritic breccia 77215 (Fig. 1).
Figure 1. a. BSE image of chip from cataclastic norite 77215
studied by SIMS. The breccia is composed primarily of anor-
thitic Pl (dark grey) and Opx (light grey), with larger clasts
embedded in a fine-grained matrix, as previously described
[9]. Red rectangle outlines area shown at higher magnifica-
tion in BSE (
b) and selected EDS maps (
c-e) showing loca-
tion of a SIMS analysis in an Opx clast that yielded 5.4
ppmw F. Smaller clasts of opx, Pl, SiO
2
, spinel, and troilite
(?) are also visible in this image, as labeled. We also meas-
ured 3.0-5.6 ppmw F in Opx separates from this sample.
Results: FAN studied by Hui et al. Hui et al. [4]
reported 0.5-5 ppmw H
2
O in four Pl single crystals
from 15415. We acquired 24 analyses from five Pl
crystals from this sample; 23 analyses are below the
LOD for H
2
O with only one being above it (at 2
ppmw). In contrast to this potentially anomalous result,
2473.pdf
Lunar and Planetary Science XLVIII (2017)
F contents in 15415 are consistently elevated above the
LOD, ranging up to 0.6 ppmw; a graphical example
demonstrating our low LOD is depicted in Fig 2.
Figure 2. Sequential SIMS analyses of the blank and two
anorthite single crystals from two different splits of FAN
65315 demonstrating low limit of detection (LOD) for fluo-
rine.
16
O
1
H is below LOD in the analyses of 65315 but
19
F is
well above LOD, with concentrations calculated based on F-
bearing basalt glass standards. 17 analyses of 6 single crys-
tals from 65315 yielded 0.5 to 1 ppmw F.
Following up on our initial results presented on two
Pl crystals from 60015 [5], we acquired 28 new anal-
yses on 4 additional Pl crystals. In contrast to our pre-
vious estimate of 2±1 ppmw H
2
O in this sample, our
new measurements are characterized by highly varying
amounts of H
2
O (from below LOD up to 28 ppmw)
and F (2.6 to 45 ppmw, compared to 6±4 ppmw F re-
ported previously), with no obvious systematic spatial
variation. Furthermore, some measurements yield sig-
nificant Cl (up to 20 ppmw); we have not detected Cl
in any other terrestrial or lunar feldspar and would not
expect to see it, because of the large ionic radius of Cl.
The high variability in H
2
O, F and Cl supports a pre-
viously unconsidered hypothesis that the volatiles re-
side primarily or entirely in melt inclusions (MI),
which are known to be present in 60015 and inferred to
be shock induced [10]. Further work is needed to as-
sess the distribution of MI in the Pl we measured in
both of our SIMS sessions.
Addtional FAN samples. We failed to find H above
the LOD in Pl from the other 11 FAN samples we
measured. On the other hand, F in Pl in 9 of these FAN
is consistently above the LOD, ranging from 0.3 to 1.1
ppmw (Fig. 2). F was below LOD in 60619 and ranged
from below LOD to just above it (0.2 ppmw) in 67746.
Notably, these two samples differ from the others in
having granoblastic textures indicative of subsolidus
recrystallization; protracted cooling and/or reheating
events may have resulted in volatile loss.
Mg-suite. We analyzed 4 additional Pl and 3 Ol
single crystals from troctolite 76535 and once again
failed to reproduce the results of [4,6] on this sample,
finding no H or F in Pl above the LOD. Our previously
reported value of 2 ppmw H
2
O in one Ol analysis from
this sample appears to be a singular anomaly.
Investigation of noritic breccia 77215 proved to be
much more enlightening. Two Opx single crystal sepa-
rates yielded 3 to 5.6 ppmw F, with H
2
O below the
LOD; this result was confirmed by analysis of a poly-
crystalline slab (Fig. 1) containing Opx and Pl clasts.
Discussion: Our results call into question the inter-
pretation of [4] that original H has been preserved in
FAN and troctolite 76535. Our hypothesis that the vol-
atiles in 60015 are partly or wholly contained in MI
bears further study and raises further questions about
the impactor and impact process. Nevertheless, even if
the original estimate of 6.4 ppmw by [4] is considered
to be robust, the calculated H
2
O concentration of the
melt in equilibrium with the Pl is reduced by an order
of magnitude, from 1600 ppm to 160 ppm. This results
from two factors: a revision of the FTIR calibration for
H in plagioclase [7] that reduces the 6.4 ppmw esti-
mate to 3.3 ppmw, and our preliminary, experimentally
determined anorthite-melt partition coefficient of
~0.02, presented by Caseres et al. at this meeting [11].
Our measurements of up to 1.1 ppmw F in a large
number of FAN also allow us to place constraints on F
in the FAN parent melt; using a preliminary F partition
coefficient of 0.02 [11] we calculate 50 ppmw in the
melt, and application of a simple fractional crystalliza-
tion model with the bulk partition coefficient from [12]
yields 15 ppmw F in the LMO, which is in the range of
previous estimates for F in the bulk silicate moon [1,2].
The higher F concentrations measured in 77215 are
consistent with crystallization of this Mg-suite rock
from an incompatible element enriched magma [13].
References: [1] McCubbin F.M. et al. (2015)
Am.
Mineral.,
100, 1668–1707. [2] Hauri E.H. et al. (2015)
EPSL, 409, 252-264. [3] Lin Y. et al. (2017)
Nat. Ge-
osci., 10, 14-19. [4] Hui H. et al. (2013)
Nat. Geosci.,
6, 177–180. [5] Mosenfelder J.L. and Hirschmann
M.M. (2016) LPS XXXXVII, Abstract #1716. [6] Hui
H. et al. (2015) LPS XXXXVI, Abstract #1927. [7]
Mosenfelder J.L. et al. (2015) Am. Mineral., 100,
1209-1221. [8] Blanchard D.P. and Budahn J.R. (1979)
Proc. 10th Lunar Sci. Conf., 803-816. [9] Chao, E.C.T.
et al. (1976) Proc. 7th Lunar Sci. Conf., 2287-2308.
[10] Sclar, C.B. and Bauer J.F. (1974) Proc. 5th Lunar
Sci. Conf., 319-336. [11] Caseres et al. (2017) LPS
XXXXVIII. Abstract #2303. [12] Rosenthal A. et al.
(2015) EPSL, 412, 77-87. [13] James O.B. and Flohr
M.K. (1983)
JGR, 88, A603-A614.
2473.pdf
Lunar and Planetary Science XLVIII (2017)