Earth Sciences sciences de la Terre

Canadian Journal of

Volume 50

An NRC Research Press Journal

2013

Une revue de NRC Research Press

www.nrcresearchpress.com

Earth Sciences Revue canadienne des

sciences de la Terre

In cooperation with the Geological Association of Canada

Avec le concours de l’Association géologique du Canada

64

Investigation of three Fe–Ti oxide deposits associated with Grenvillian anorthosite massifs as potential source for lunar analogue ilmenite1,2 Caroline-Emmanuelle Morisset, Marie-Claude Williamson, and Victoria Hipkin

Can. J. Earth Sci. Downloaded from www.nrcresearchpress.com by CSP Staff on 09/11/13 For personal use only.

Abstract: Terrestrial ilmenite has been assessed for its suitability in the preparation of a simulant of lunar high-Ti mare regolith, which is needed to test oxygen production on the lunar surface during future missions. Petrographic observations using optical and electron microscopes revealed that ilmenite grains in the Mirepoix, Sanford Lake, and Degrosbois Fe–Ti oxide deposits from Grenvillian anorthosite massifs contain hercynite in their rims but that hematite exsolution is absent in three of eight studied samples. Electron microprobe analyses showed that the ilmenite grains vary in TiO2 from 38.82 to 52.34 wt.%; in FeO from 33.15 to 45.88 wt.%; in Fe2O3 (calculated from stoichiometry) from 2.18 to 27.03 wt.%; in MgO from 0.05 to 2.49 wt.%; and in MnO from 0.26 to 1.51 wt.% and are thus comparable in composition to lunar ilmenite. Ilmenite, free of hematite, in two samples was homogeneous in composition (e.g., sample SL1: 50.4 –51.4 wt.% TiO2). Two sub-solidus reactions occurred amongst the oxide minerals in the studied deposits: (i) exsolution of hercynite in magnetite; and (ii) reaction between the hematite component in the ilmenite and the ulvöspinel component of the magnetite, forming new ilmenite, which includes the hercynite. The reaction between the hematite component in the ilmenite and the ulvöspinel component of the magnetite, when conditions are favorable for diffusion, produces ilmenite grains that are homogenous in composition and free of hematite exsolution. Ilmenite with little or no Fe3⫹ may occur in massive, fine-grained, and metamorphosed Fe–Ti oxide deposits and provide a terrestrial analogue source of ilmenite useful for the production of lunar mare simulant. Résumé : L’emploi de l’ilménite terrestre pour développer un simulant de régolithe de mare lunaire riche en Ti, ce qui est essentiel pour tester les méthodes de production d’oxygène lors de missions future, a été évalué. L’observation pétrographique au microscope optique et au microscope électronique a` balayage a révélé que les échantillons provenant des gisements de d’oxydes de Fe–Ti de Mirepoix, de Sanford Lake et de Degrosbois, compris dans les massifs anorthositiques du Grenville, contiennent des grains d’ilménite avec une bordure de spinelle alumineux (hercynite) mais que trois des huit échantillons étudiés ne contenaient pas d’exsolutions d’hématite. Les analyses a` la microsonde électronique ont permis de déterminer que la composition des grains d’ilménite varie de 38,82 a` 52,34 % poids de TiO2; de 33,15 a` 45,88 % poids de FeO; de 2,18 a` 27,03 % poids de Fe2O3 (le Fe2O3 étant calculé a` partir de la stœchiométrie de l’ilménite); de 0,05 a` 2,49 % poids de MgO et de 0,26 a` 1,51 % poids de MnO. Ces concentrations se comparent donc aux concentrations des mêmes éléments dans l’ilménite lunaire. Dans deux échantillons étudiés, l’ilménite ne contenait pas d’exsolution d’hématite et était de composition homogène (p. ex. échantillon SL1 : 50,4 –51,4 % poids de TiO2). Deux réactions sub-solidus ont eu lieu entre les oxydes dans les échantillons étudiés : (i) l’exsolution de l’hercynite dans la magnétite; et (ii) la réaction entre l’hématite, contenue dans un grain d’ilménite, et l’ulvöspinel, contenu dans un grain de magnétite, ce qui a produit de l’ilménite en continuité optique avec le grain d’ilménite primaire, incluant l’hercynite. Lorsque les conditions sont favorables a` la diffusion, cette réaction produit des grains d’ilménite ne contenant pas d’exsolution d’hématite et qui sont de composition homogène. L’ilménite contenant peu ou pas de Fe3⫹ est analogue en composition a` l’ilménite lunaire et pourrait se trouver dans des gisements massifs d’oxyde de Fe–Ti a` grains fin et métamorphosés. Cette ilménite pourra servir a` la production de simulant de régolithe de mare riche en Ti.

Received 15 February 2012. Accepted 18 August 2012. Published at www.nrcresearchpress.com/cjes on 4 January 2013. Paper handled by Associate Editor Paul Sylvester. C.-E. Morisset* and V. Hipkin. Canadian Space Agency, 6767, route de l’Aéroport, Saint-Hubert, QC J3Y 8Y9, Canada. M.-C. Williamson. Geological Survey of Canada, 601 Booth Street, Ottawa, ON K1A 0E8, Canada. Corresponding author: Caroline-Emmanuelle Morisset (e-mail: [email protected] and [email protected]). *Present address: Golder Associates, 1001 boul de Maisonneuve, 7e étage, Montréal, QC H3A 3C8, Canada. 1 2

This article is one of a series of papers published in this CJES Special Issue on the theme of Canadian contributions to planetary geoscience. Geological Survey of Canada Contribution 20120122.

Can. J. Earth Sci. 50: 64 –77 (2013)

doi:10.1139/e2012-059

Published by NRC Research Press

Morisset et al.


Introduction In the list of recommendations agreed upon by the Canadian Planetary Geology and Geophysics (PG&G) community and submitted to the Canadian Space Agency (CSA), there is clear interest in scientific investigations that target the composition and resource potential of the lunar regolith (CSEW6 2009). The outcomes of these investigations at terrestrial analogue sites will eventually be applicable to mineral prospecting and In situ Resource Utilization (ISRU) activities carried out on the Moon. Previous studies have emphasized the potential of reduced ilmenite (e.g., Gibson and Knudsen 1988; Briggs and Sacco 1991; Vaniman et al. 1991), reduced high-Ti basalt (e.g., Gibson et al. 1994), or reduced lunar mare soil (e.g., Allen et al. 1994) containing ilmenite for the in situ production of oxygen on the Moon. This is because ilmenite is the most reactive mineral in the lunar regolith for the production of oxygen by reduction (Allen et al. 1994). The ability to produce oxygen on the Moon has been discussed for many years (e.g., Taylor and Carrier 1993), as oxygen is valuable both as a fuel and for human life support and thus can enable extended human presence in space. Although presence of water in lunar volcanic glass (46 ppm H2O; Saal et al. 2008) and in apatite (7000 ppm OH; McCubbin et al. 2010) has been recognized recently, it is present in only small quantities. Even if total recovery of this water were achieved, the quantity of oxygen produced would be considerably less than the amount of oxygen that can be produced by reducing a high-Ti basalt containing ilmenite (2.9%– 4.6% O2 yield, Gibson et al. 1994). The presence of water ice on the lunar surface in permanently shadowed craters has been inferred from various remote sensing methods: (i) neutron spectroscopy (Feldman et al. 1998; Feldman et al. 2001); (ii) infrared (IR) spectroscopy (Colaprete et al. 2010); and (iii) radar (Spudis et al. 2010). At present, the quality and accessibility of these reserves are still being investigated and their quantities are poorly estimated. Operational extraction from permanently shadowed craters with their extremely low temperatures presents an additional engineering challenge. While the presence of water ice or OH bound in the regolith is very promising for oxygen production on the Moon (Sanders and Larson 2011), until the quantities available are known and their extraction methods are shown to be relatively efficient, ilmenite reduction remains a strong candidate for lunar oxygen production and is worthy of further research. National Aeronautics and Space Administration (NASA), in a design reference mission for ISRU called RESOLVE, has included a reactor that can heat samples of regolith up to 900 °C to liberate oxygen by hydrogen reduction (George et al. 2012). This has been proven to be an efficient technique (e.g., Heiken and Vaniman 1990; Taylor and Carrier 1993; Allen et al. 1994; Gibson et al. 1994), but further testing is required to optimize the extraction of oxygen (e.g., size of particles, Gibson et al. 1994) and to develop exploration methods to prospect for and process regolith rich in ilmenite. One priority from a NASA–USGS (United States Geological Survey) partnership in simulant development is to produce a high-Ti basalt regolith simulant for ISRU (Edmunson et al. 2010), which should contain appreciable amounts of ilmenite. However, textural and compositional differences between lu-

65

nar and terrestrial ilmenite make it difficult to find ilmenite on Earth with the appropriate textures and compositions for ISRU engineering tests. On Earth, ilmenite typically contains ferric iron in variable quantities because most terrestrial magmas have an oxygen fugacity that allows some of the iron to be ferric and the existence, at temperatures greater than ⬃800 °C, of a complete solid solution between ilmenite (FeTiO3) and hematite (Fe2O3) (Lindsley 1973). As a result, grains of ilmenite on Earth are typically composed of ilmenite (host) and hematite lenses, and end up richer in Fe (and so in Fe–O bonds) and poorer in Ti than the Moon’s ilmenite. These exsolutions textures affect the physical properties of the ilmenite. For example, the spectral reflectance properties of ilmenite differ when hematite is present (Cloutis et al. 2008), and the natural remnant magnetization increases when exsolved hematite occurs as opposed to pure ilmenite (Robinson et al. 2002). To develop a high-Ti basalt regolith simulant that can be of value for engineering, it is, therefore, very important to identify ilmenite on Earth with chemical and physical properties that are comparable with those of lunar ilmenite. The simulant can be used to test ISRU instrumentation, drilling, and processing systems in preparation for future missions to the Moon. To find suitable quantities of ilmenite to use as simulant, we investigate in this study three primary magmatic Fe–Ti ore deposits from the Grenville Province: (i) the Mirepoix deposit in the Mattawa anorthosite massif (Morisset 2002, 2003); (ii) the Sanford Lake deposits in the Adirondacks anorthosite massif (Kelly 1979); and (iii) the Degrosbois deposit in the Morin anorthosite massif (Rose 1969; Fig. 1). Chemical exchange reactions between magnetite and ilmenite that occur during cooling (Buddington and Lindsley 1964; Duchesne 1972) are considered to explain the recognition of relatively hematite-poor to hematite-free ilmenite.

Description of sample localities and samples selected Mirepoix deposit in the Mattawa anorthosite The Mirepoix deposit is a layered intrusion within the 1016 ⫾ 3 Ma Mattawa anorthosite (Hébert et al. 2005) that intrudes the Lac-Saint-Jean anorthosite along the eastern border of the massif (Hébert 2003; Hébert and Cadieux 2003; Owens and Dymek 2005; Fig. 1). The Mattawa anorthosite covers an area of 250 km2 and is found in the Allochthonous Polycyclic Belt (Rivers 1997) of the Grenville Province. The mineralogy of the oxide minerals of the deposit consist of layers of massive ilmenite (Xhem ⫽ 23, see Table 1 footnote for a definition of Xhem) in the lower part of the stratigraphy, overlain by layers of magnetite (Xusp ⫽ 5– 0; see Table 2 footnote for a definition of Xusp) and ilmenite (Xhem ⫽ 10 –5) and of nelsonite (magnetite, ilmenite, and apatite) (Morisset 2002). Minor amounts (1%–3%) of hercynite [(Fe,Mg)Al2O4] are found throughout the layered intrusion. Plagioclase, orthopyroxene, and clinopyroxene form silicate layers interbedded with the oxide-rich layers. The Mattawa intrusion was emplaced prior to the Rigolet orogeny (1010 – 990 Ma, Rivers et al. 2002), which marked the last metamorphic event associated with the Grenville Province. The metamorphism associated with the Rigolet orogeny was concentrated along the Grenville Front (Rivers 2008) and did not Published by NRC Research Press

66

Can. J. Earth Sci., Vol. 50, 2013


Fig. 1. Simplified geological map of the Grenville Province adapted from Davidson (1998) showing the studied localities: (a) Mirepoix deposit; (b) Sanford Lake deposit; (c) Degrosbois deposit. Other Fe–Ti ⫾ P oxide deposits associated with anorthosite massif from Corriveau et al. 2007 are also shown: (d) Saint-Hypolyte; (e) Saint-Urbain; (f) Mine Canada Iron; (g) Saint-Charles; (h) La Hache-Est; (i) Lac Brulé; (j) Lac Dissimieu; (k) Lac La Blache; (l) Rivière Pentecôte; (m) Canton Arnaud; (n) Lac Raudot; (o) Magpie; (p) Big Island; (q) Tio Mine; (r) Everett.

er Riv

affect the Mattawa anorthosite. One sample (TR1-6) was selected for this study. It consists of anorthosite with a fine layer of oxide minerals (⬃0.5 cm) where ilmenite and magnetite are present in equal proportions (based on visual estimates as are all the subsequent given proportions in the following text) and where the grain size varies from 0.5 mm ⫻ 1 mm to 2.5 mm ⫻ 2.5 mm. Sanford Lake deposits in the Adirondack anorthosite massif The Sanford Lake deposits occur as four discrete bodies of massive to semi-massive ores of ilmenite (Xhem ⫽ 16 – 8) and magnetite (Xups ⬍ 1) in the Marcy anorthosite massif (Kelly 1979). The 1054 ⫾ 6 Ma Marcy anorthosite (McLelland et al. 2004) covers an area of 3000 km2 (Ashwal 1993) (Fig. 1). The massive sections of the ore (90%–100% ore minerals) are characterized by magnetite and ilmenite in similar proportions (Kelly 1979) while ilmenite is more abundant than magnetite in semi-massive ores (50%–90% ore minerals) (Gross 1968). The silicate minerals identified in the Sanford Lake deposits consist of plagioclase, clinopyroxene, orthopyroxene, and garnet. The Marcy anorthositic massif was affected by high-grade metamorphism during the 1080 –1020 Ma (Rivers 2008) Grenvillian Ottawan orogeny (McLelland et al. 2001, 2004). Five samples, labelled SL1 to SL5, were selected for this study from the massive oxide ore deposits. Grain sizes and proportions of the oxide minerals vary from one sample to an other.

In sample SL1, ilmenite and magnetite grains are of comparable size (0.2 mm ⫻ 0.2 mm – 1.5 mm ⫻ 1.5 mm) and are present in the same proportions. In samples SL2 and SL5, the ilmenite grains are somewhat larger overall (1 mm ⫻ 1 mm – 3 mm ⫻ 3 mm) than the magnetite grains (0.5 mm to 0.75 mm – 1 ⫻ 2 mm) and more abundant (55% ilmenite versus 45% magnetite). In sample SL3, the ilmenite grains are smaller (0.25 mm ⫻ 0.75 mm – 2 mm ⫻ 3 mm) than the magnetite grains (0.75 mm ⫻ 0.75 mm – 2 mm ⫻ 3.5 mm) and, as observed in SL1, are present in the same proportions. Finally, the ilmenite and magnetite grains in sample SL4 are of comparable size (0.25 mm ⫻ 0.33 mm – 2 mm ⫻ 2 mm) but ilmenite is less abundant than magnetite (30% versus 70%, respectively). Degrosbois deposit in the Morin anorthosite The Degrosbois deposit is found in the 1155 ⫾ 3 Ma Morin anorthosite (Doig 1991) in the Central Metasedimentary Belt of the Grenville Province (Fig. 1). The Morin anorthosite massif covers an area of 1500 km2. The Degrosbois deposit consists of massive oxide ore that contains ilmenite and magnetite as well as silicate minerals (plagioclase and pyroxene) and apatite (Rose 1969). The anorthosite massif was metamorphosed to granulite facies, most likely during the 1080 –1020 Ma (Rivers 2008) Ottawan orogeny (Martignole and Schrijver 1970; Emslie 1975). Two samples, DEG6 and DEG16, were Published by NRC Research Press

Morisset et al.

Samples: TR1-6-117B TR1-6-31 SL1-54 Deposits: Mirepoix SiO2 TiO2 Al2O3 V2O3 Cr2O3 FeO MgO MnO ZnO Total FeO Fe2O3 Total Si Ti Al V Cr Fe3⫹ Fe2⫹ Mg Mn Total Xilm Xhem RO R2O3 TO2

0.02 38.82 0.08 0.07 0.05 57.47 0.44 0.96 b.d. 97.91 33.1 27.0 100.5 0.001 0.739 0.002 0.001 0.001 0.515 0.702 0.017 0.021 1.999 73 27 43 15 42

SL2-39

SL2-44

SL3-136B

SL3-67

SL4-31

SL5-53

DEG6-13

DEG16-22 12065

Mirepoix Sandford Lake Sandford Lake Sandford Lake Sandford Lake Sandford Lake Sandford Lake Sandford Lake Degrosbois Degrosbois Lunar sample 0.01 50.81 0.03 0.00 0.00 46.70 0.71 1.51 0.00 99.77 42.9 4.2 100.2 0.000 0.960 0.001 0.000 0.000 0.079 0.901 0.026 0.032 1.999 96 4 49 2 49

0.03 51.40 0.02 b.d. 0.03 48.50 0.05 0.30 b.d. 100.33 45.9 2.9 100.6 0.001 0.972 0.000 0.000 0.001 0.055 0.964 0.002 0.006 2.001 97 3 49 1 49

0.03 49.34 2.83 0.01 0.11 46.60 1.28 0.33 0.00 100.53 41.8 5.4 101.1 0.001 0.908 0.082 0.000 0.002 0.099 0.855 0.047 0.007 2.001 95 5 48 5 48

0.04 50.70 0.02 0.08 0.06 47.70 0.97 0.26 b.d. 99.83 43.6 4.6 100.3 0.001 0.954 0.001 0.002 0.001 0.087 0.912 0.036 0.006 2.000 95 5 49 2 49

b.d. 52.34 b.d. b.d. b.d. 44.59 2.24 0.54 b.d. 99.71 42.5 2.3 99.9 0.000 0.979 0.000 0.000 0.000 0.043 0.884 0.083 0.011 2.000 98 2 49 1 49

0.01 51.80 0.07 0.00 0.00 44.20 2.49 0.51 b.d. 99.08 41.7 2.8 99.4 0.000 0.972 0.002 0.000 0.000 0.053 0.869 0.092 0.011 1.999 97 3 49 1 49

b.d. 51.40 b.d. b.d. b.d. 44.50 2.32 0.58 b.d. 98.80 41.5 3.3 99.1 0.000 0.969 0.000 0.000 0.000 0.062 0.870 0.087 0.012 2.000 97 3 49 2 49

0.02 51.00 0.02 b.d. 0.05 46.40 2.00 0.37 b.d. 99.86 41.9 5.0 100.4 0.000 0.952 0.001 0.000 0.001 0.094 0.870 0.074 0.008 2.000 95 5 49 2 49

0.02 45.77 0.03 0.21 0.03 50.60 0.85 0.35 0.00 97.86 39.3 12.6 99.2 0.001 0.876 0.001 0.004 0.001 0.242 0.836 0.032 0.008 2.001 87 13 47 7 47

b.d. 51.90 0.02 b.d. 0.05 45.90 1.11 0.78 b.d. 99.76 43.9 2.2 100.0 0.000 0.979 0.001 0.000 0.001 0.041 0.921 0.041 0.017 2.001 98 2 49 1 49

— 53.9 0.42 0.04 0.39 45.4 0.1 0.36 — 100.61 45.4 0 100.61 — 1.007 0.012 0.001 0.007 — 0.943 0.003 0.007 1.98 100 0 50 0 50

Note: Electron-microprobe data. Cations are calculated on the basis of two cations and Fe3⫹ as Fe2⫹ ⫹ Mg ⫹ Mn – Ti followed by a calculation based on three oxygens. Xilm ⫽ [ilm/(ilm ⫹ hem)]100; Xhem ⫽ [hem/(ilm ⫹ hem)]100; ilm ⫽ Ti – Mg – Mn ⫹ Al/2; hem ⫽ 0.5(Fe2⫹ ⫹ Fe3⫹ ⫹ Mg ⫹ Mn – Ti) (Lindsley & Frost 1992). RO ⫽ MgO ⫹ CaO ⫹ MnO ⫹ ZnO ⫹ FeO (all in molar %); R2O3 ⫽ Al2O3 ⫹ V2O3 ⫹ Cr2O3 ⫹ Fe2O3 (all in molar %); TO2 ⫽ TiO2 (in molar %). b.d., below detection limits. Lunar ilmenite composition from sample 12065 is from Cameron (1971) and is shown for comparison.

67



Table 1. Representative compositions of ilmenite.

Samples:

TR1-6-34

TR1-6-112B

TR1-6-123B

SL1-158B

SL2-38

SL3-138B

SL4-21

SL5-24

DEG6-3

DEG16 -163B

12065

Deposits:

Mirepoix

Mirepoix

Mirepoix

Sanford Lake

Sanford Lake

Sanford Lake

Sanford Lake

Sanford Lake

Degrosbois

Degrosbois

Lunar sample

SiO2 TiO2 Al2O3 V2O3 Cr2O3 FeOT MgO MnO ZnO Total FeO Fe2O3 Total Si Ti Al V Cr Fe3⫹ Fe2⫹ Mg Mn Zn Total Xmgt xusp

b.d. 0.45 0.23 0.24 0.07 92.68 0.05 0.00 — 93.71 31.6 67.9 100.5 0.000 0.013 0.010 0.007 0.002 1.955 1.010 0.003 0.000 0.000 3.000 98 2

b.d. 0.09 0.37 0.28 0.04 91.83 0.07 b.d. 0.04 92.71 30.9 67.7 99.5 0.000 0.003 0.017 0.009 0.001 1.969 0.997 0.004 0.000 0.001 3.000 100 0

b.d. 0.02 0.30 0.32 0.06 92.50 0.04 0.01 0.00 93.24 31.1 68.3 100.1 0.000 0.001 0.013 0.010 0.002 1.974 0.998 0.002 0.000 0.000 3.000 100 0

b.d. 0.52 0.76 0.88 0.34 89.70 0.29 b.d. b.d. 92.53 31.0 65.3 99.1 0.000 0.015 0.035 0.027 0.010 1.895 0.999 0.017 0.000 0.000 3.000 98 2

0.03 0.46 0.23 0.74 0.36 91.60 0.08 0.02 b.d. 93.52 31.5 66.8 100.2 0.001 0.013 0.010 0.023 0.011 1.927 1.009 0.005 0.001 0.000 3.000 99 1

0.04 9.31 3.91 0.61 0.23 79.45 0.82 0.17 0.04 94.60 38.7 45.3 99.1 0.001 0.262 0.173 0.018 0.007 1.276 1.211 0.046 0.005 0.001 3.000 71 29

0.04 8.57 2.61 0.70 0.30 82.83 0.54 0.29 0.04 95.92 38.6 49.1 100.8 0.002 0.239 0.114 0.021 0.009 1.374 1.201 0.030 0.009 0.001 3.000 75 25

0.02 0.05 0.27 1.09 0.45 92.65 0.06 b.d. 0.05 94.64 31.5 68.0 101.4 0.001 0.001 0.012 0.033 0.013 1.937 0.998 0.003 0.000 0.001 3.000 100 0

0.01 0.01 0.29 0.74 0.17 92.09 0.01 b.d. b.d. 93.32 31.1 67.7 100.1 0.000 0.000 0.013 0.023 0.005 1.958 1.000 0.001 0.000 0.000 3.000 100 0

0.03 0.92 0.61 0.38 0.15 90.27 0.04 0.04 b.d. 92.44 31.6 65.2 99.0 0.001 0.027 0.028 0.012 0.005 1.900 1.024 0.002 0.001 0.000 3.000 97 3

— 33.6 2.3 0.3 0.8 63.1 0.1 0.3 — 100.5 63.1 0 100.5 — 0.926 0.1 0.009 0.024 — 1.936 0.007 0.009 — 3.011 2 98

Note: Electron-microprobe data. Cations are calculated on the basis of three cations and Fe3⫹ as (2/3)(FeT ⫹ Mg ⫹ Mn – 2Ti – Al/2) followed by a calculation based on four oxygens. Xmgt ⫽ [mgt/ (mgt ⫹ usp)]100; Xusp ⫽ [usp/(usp ⫹ mgt)]100; mgt ⫽ (1/3)(FeT ⫹ Mg ⫹ Mn – 2Ti – Al/2); usp ⫽ {Ti – [Mg ⫹ Mg/(Mg ⫹ Fe2⫹)(Al/2)]/2 –Mn/2}/(0.5Fe3⫹ ⫹ Ti – Mn/2) (Lindsley and Frost 1992). b.d., below detection limits. Lunar ulvöspinel composition from sample 12065 is from Cameron (1971) and is shown for comparison.




68

Table 2. Representative compositions of magnetite.

Morisset et al.

selected for this study. Sample DEG6 consists of disseminated (10% oxide minerals) ore in which ilmenite and magnetite are comparable in size (0.15 mm ⫻ 0.15 mm – 0.75 mm ⫻ 1 mm) but where ilmenite is far more abundant (80% ilmenite versus 20% magnetite). Sample DEG16 belongs to the semi-massive portion of the ore where ilmenite and magnetite occur in the same proportions but ilmenite grains are notably larger in size (0.3 mm ⫻ 0.5 mm – 1 mm ⫻ 1 mm versus 0.3 mm ⫻ 0.3 mm – 0.4 mm ⫻ 1 mm for magnetite).


Mineral textures The oxide minerals from the studied samples selected from the Mirepoix, Sanford Lake, and Degrobois deposits exhibit a range of textures. In some of the samples (e.g., SL1, SL3, DEG16), no hematite was observed in the ilmenite grains using optical or electron microscopes similar to ilmenite found in lunar samples (Fig. 2a). In cases where hematite is found in the ilmenite, abundances decrease from the core to the rim (near the edge of a grain) of the ilmenite grain where it is at the contact with magnetite (Fig. 2b) to the point where it is not observed (Figs. 2c, 2d). At the contact between ilmenite and magnetite grains, hercynite is observed in the rim of the ilmenite grains (Figs. 2b–2e). Small magnetite grains (7– 60 ␮m) can be associated with the hercynite in the rim of ilmenite grains (Fig. 2d). Hercynite also occurs in the rim of magnetite grains that are in contact with each other (Fig. 2f). The magnetite grains may contain exsolution of hercynite or display a trellis texture, the latter being formed by oxidation– exsolution of ilmenite (Buddington and Lindsley 1964). Unlike hematite exsolution in ilmenite that decrease in abundance at the contact of magnetite, trellis texture in magnetite, when present, is found throughout the grain.

Microprobe analyses Analytical technique Ilmenite (n ⫽ 129), magnetite (n ⫽ 81), and hercynite (n ⫽ 6) compositions were determined using a JEOL 8900 electron microprobe with five wavelength-dispersive spectrometry (WDS) detectors located at the Department of Earth and Planetary Sciences at the McGill University, Montréal. The backscattered electron image in Fig. 2a was taken using the scanning electron microprobe of the Geological Survey of Canada in Ottawa while the backscattered electron images in Figs. 2b–2e were taken using the imaging system of the McGill University electron microprobe. Where possible, hematite exsolution lamellae in ilmenite and exsolution in magnetite were avoided during analysis. Data reduction was carried out using the “PRZ” method (Armstrong 1995). The accelerating voltage was set at 15 kV and the beam current was set at 10 nA with a beam diameter of 2–3 ␮m. Counting times were 20 s on peak and 10 s on background, except for Zn that was 100 s on peak and 50 s on background. Vanadium was corrected for Ti interference analyzing the intensity of V measured on the pure rutile Ti standard, which has no V. This correction was performed with JEOL software. Mineral compositions The ilmenite compositions of samples from the three locations are shown in Table 1. We found significant variations in TiO2 (38.82–52.34 wt.%) and FeOT (44.20 –57.47 wt.%) that are reflected in different proportions of the hematite end mem-

69

ber (Xhem ⫽ 2–27) (Table 1; Fig. 3a). Calculated FeO vary from 33.2 to 45.9 wt.% and Fe2O3 from 2.2 to 27.0 wt.%. For all the studied samples with the exception of SL1, SL3, and TR1-6, the centers of the grains are richer in TiO2 and poorer in Fe2O3 than the border (Fig. 3a). Ilmenite grains from samples SL1 and SL3 are almost homogeneous in compositions (cf. SL1: 50.4 –51.4 wt.% TiO2; SL3: 50.6 –52.3 wt.% TiO2) and ilmenite grains from TR1-6 have a core that is much richer in hematite lenses than the rim (Fig. 3a). MgO (0.05–2.49 wt.%) (Table 1) is not well correlated with any of the other measured elements, and no systematic variation from core to border is observed (Fig. 3b). MnO (0.26 –1.51 wt.%) is positively correlated with TiO2 and negatively correlated with total Fe. As a result, the mineral cores are slightly richer in Fe than along the periphery of grains (Fig. 3b). V2O3 and Cr2O3 vary from below detection limits to 0.21 wt.% and 0.11 wt.%, respectively. The magnetite compositions of samples from the Mirepoix, Sanford Lake, and Degrosbois deposits are presented in Table 2. TiO2 values range from below the detection limit to 8.57 wt.%, calculated FeO values range from 30.9 to 38.7 wt.%, and calculated Fe2O3 values range from 45.3 to 68.3 wt.%. These variations are a consequence of the different proportions of the magnetite– ulvöspinel end members in individual samples. For samples DEG6 and SL2, the center of the magnetite grains is richer in Fe2O3 (67.5– 68.1 wt.%) and poorer in TiO2 (below detection, 0.7 wt.%) than the border of the grains (Fe2O3: 66.8 – 67.6 wt.%; TiO2: 0.3– 0.8 wt.%). For sample SL3, the inverse relationship is observed: the center of the grains is poorer in Fe2O3 (45.8 –53.9 wt.%) and richer in TiO2 (6.2–9.3 wt.%) than the border of the grains (Fe2O3: 55.6 –57.7 wt.%; TiO2: 4.6 – 6.2 wt.%). Other samples show overlap in compositions between the core and the border of the grains for these elements. Al2O3 (0.23–3.91 wt.%) is positively correlated with TiO2, Fe2O3 (Fig. 4a), and MgO (below detection, 0.82 wt.%) (Fig. 4b). Some magnetite grains have a core that is richer in Al2O3 and MgO than the borders (e.g., SL3 Al2O3: 3.9 wt.% versus 1.4 wt.%; MgO: 0.82 wt.% versus 0.33 wt.%) (Fig. 4b). V2O3 (0.24 –1.09 wt.%) and Cr2O3 (below detection, 0.45wt.%) are positively correlated and form a trend where values for each samples almost do not overlap but no core to border composition differences are observed. The only sample that does not plot along that trend is DEG6 because it has higher values of V2O3 for its concentration in Cr2O3 compared with other samples (Fig. 4c). Finally, MnO (below detection, 0.29 wt.%) is not well correlated with other elements analyzed for the study. The hercynite compositions found in samples from Mirepoix, Sandford Lake, and Desgrosbois are presented in Table 3. Al2O3 values range from 59.56 to 62.38 wt.%. FeOT values range from 20.56 to 27.44 wt.%, and MgO values range from 11.35 to 13.58 wt.%, which are illustrated in XMg (100[Mg/(Mg ⫹ Fe2⫹)] ⫽ 47–56). Small quantities of TiO2 are present from below detection limits to 0.33 wt.%. Finally, Cr2O3 varies from 0.37 to 0.60 wt.%; MnO varies from 0.05 to 0.17 wt.%; and ZnO varies from 1.52 to 3.20 wt.%.

Discussion Description of the sub-solidus reactions between magnetite and ilmenite Ilmenite (FeTiO3) forms a complete solid solution with hematite (Fe2O3) at high temperature (Lindsley 1973). During Published by NRC Research Press

70



Fig. 2. Backscattered (a–e) and photomicrograph (f) images showing ilmenite grains in samples from the Moon, Mirepoix deposit, and Sanford Lake deposits. (a) Lunar thin section 70017, 557: a grain of ilmenite in contact with a grain of plagioclase and a grain of clinopyroxene in a lunar basalt. Note that the ilmenite grain is free of hematite exsolution. (b) TR1-6 (Mirepoix): a grain of ilmenite in contact with a magnetite grain is showing a strong zoning in its hematite content from the core rich in hematite to the rim poor in hematite. Note the rim of hercynite at the contact between the ilmenite and the magnetite grains; (c) TR1-6 (Mirepoix): Al-spinel rim showing that some spinel grains (7–30 ␮m) are completely included in the ilmenite and others are in contact with magnetite. Note the hematite exsolution present in the right side of the picture diminishing and disappearing toward the contact with magnetite in the closeup image of the view shown in (b; outline with black box); (d) TR1-6 (Mirepoix): magnetite is observed in the hercynite rim found in ilmenite at the contact of magnetite in the closeup image of the view shown in (b; outline with black box); (e) SL5 (Sanford Lake): grain of ilmenite surrounded by magnetite and showing a large (⬃100 ␮m) Al-spine rim. Note that the dashed line represents the limit of the primary grain of ilmenite; (f) SL3 (Sanford Lake): magnetite grains containing hercynite exsolution. Note the rim of hercynite found between grains of magnetite or between grains of magnetite and ilmenite. Symbols: ilm, ilmenite; pl, plagioclase; cpx, clinopyroxene; mgt, magnetite; hc, hercynite; hem, hematite.

c

d


Morisset et al.

71

(wt.%)

(wt.%)

Fig. 4. Al2O3 (wt.%) versus (a) Fe2O3 (wt.%) and (b) MgO (wt.%) and (c) V2O3 (wt.%) versus Cr2O3 (wt.%) in magnetite from Mirepoix (TR1-6), Sanford lake (SL), and Degrosbois (DEG) deposits. The symbols in the legends are given in (a). Analyses from the grain centers are shown in white and from the grain borders are shown in black to see compositional zoning in mineral grains when present in some samples.

(wt.%)

(wt.%)

(wt.%)

(wt.%)

(wt.%)

(wt.%)

(wt.%)

(wt.%)


Fig. 3. (a) TiO2 (wt.%), (b) MgO (wt.%), and (c) MnO (wt.%) versus FeO ⫹ Fe2O3 (wt.%) in ilmenite from Mirepoix (TR1-6), Sanford Lake (SL), and Degrosbois (DEG) deposits. FeO ⫹ Fe2O3 is used because electron microprobe analyses and the Lindsley and Frost (1992) calculation does not provide the exact amount of Fe2O3 present in the ilmenite. The symbols legend is given in (a), and symbols for analyses from the grain borders are filled.

(wt.%)

cooling, the proportion of hematite in the mineral structure will determine the abundance and size of the exsolved hematite lenses present in the ilmenite (Duchesne 1970). Magnetite (Fe3O4) forms a complete solid solution with ulvöspinel

(wt.%)

(Fe2TiO4) at high temperature that will produce a variety of exsolution lamellae in the magnetite during cooling (e.g., cloth and trellis microtextures; Ramdohr 1953; Duchesne 1972). The temperature and the oxygen fugacity that prevail during Published by NRC Research Press

72


Table 3. Representative compositions of hercynite.


Samples:

SL4-3

SL5-7

Deposits:

Sanford Lake

Sanford Lake

SiO2 TiO2 Al2O3 V2O3 Cr2O3 FeOT MgO MnO ZnO Total FeO Fe2O3 Total Si Ti Al V Cr Fe3⫹ Fe2⫹ Mg Mn Zn Total XMg

b.d. b.d. 59.56 b.d. 0.37 27.44 11.35 0.17 1.52 100.56 22.85 5.10 101.05 0.000 0.001 1.885 0.001 0.008 0.103 0.513 0.454 0.004 0.030 3.000 47

b.d. 0.33 62.38 b.d. 0.60 20.56 13.58 0.05 3.20 100.77 18.70 2.07 100.97 0.000 0.007 1.932 0.001 0.012 0.041 0.411 0.532 0.001 0.062 3.000 56

Note: Electron-microprobe data. Cations are calculated on the basis of charge balance followed by a calculation based on four oxygens. XMg ⫽ 100[Mg/(Mg⫹Fe2⫹)]. b.d., below detection limits.

crystallization of a magmatic system dictate the primary composition of the oxide minerals (Buddington and Lindsley 1964). Most magmas on Earth crystallize with an oxygen fugacity equivalent to the fayalite–magnetite– quartz buffer. Under these conditions, the ilmenite should have a hematite component and the magnetite should have an ulvöspinel component (Buddington and Lindsley 1964). During the last stages of crystallization of magmas, under oxidizing conditions due to the presence of interstitial melt rich in O and H, the ulvöspinel in the magnetite can be oxidized to ilmenite by the reaction 6Fe2TiO4 ⫹ O2 ⫽ 2Fe3O4 ⫹ 6FeTiO3 (Buddington and Lindsley 1964). Duchesne (1970, 1972) proposed that hematite contained in the ilmenite could act as an oxidant and react with ulvöspinel contained in the magnetite during last stage of crystallization and cooling of an intrusion. In this reaction, the hematite in ilmenite would react with the ulvöspinel in magnetite to produce ilmenite and magnetite [Fe2O3 (in ilm) ⫹ Fe2TiO4 (in mt) ¡ FeTiO3 ⫹ Fe3O4], resulting in a lower hematite content of the primary ilmenite and lower Ti concentration in the primary magnetite. The newly formed ilmenite is then in optical continuity with the primary grain of ilmenite and the new rim of the ilmenite grain may contain hercynite (Duchesne 1972). Evidence for the sub-solidus reactions between ilmenite and magnetite in the studied samples The reaction between ilmenite and magnetite, as described by Duchesne (1972), can be observed in slowly cooled mag-

matic rocks of the Mirepoix deposit where compositional variations of the ilmenite found in contact with magnetite include a notable decrease in the occurrence of hematite exsolution in the ilmenite from core to rim (Fig. 2a). However, the reaction between the ilmenite and the magnetite is not the first one to occur within the oxide minerals in these rocks. Hercynite is observed in the rim of ilmenite grains in contact with magnetite grains (Figs. 2a–2e), and small magnetite grains (7– 60 ␮m) are noted with hercynite in the rim of some ilmenite grains (Figs. 2c, 2d). A hercynite rim can also be observed between two grains of magnetite (Fig. 2e). The observed enrichment in Al2O3 and MgO of some magnetite grains from core to rim, contrary to V2O3 and Cr2O3, suggests that diffusion of these elements to the border of the grain have led to the exsolution of the hercynite in the magnetite grains. That is also supported by the fact that the hercynite rim is noted in the rim of magnetite. Thus, based on observations, the hercynite exsolution in magnetite is indicated as the first sub-solidus reaction that has occurred in the oxide minerals during cooling. The reaction between ilmenite and magnetite described by Duchesne (1972), as evidenced by the presence in most grains of TiO2-rich cores and Fe-rich rims, occurs second also during cooling. The observed variations can be explained by diffusion of the hematite component to the border of the grains to react with magnetite. Moreover, the composition of the magnetite grains in sample SL3, where the rim is richer in TiO2, could be explained by the diffusion of the ulvöspinel component to the border of the grain to react with ilmenite. The newly produced ilmenite includes the hercynite that was found previously in the magnetite and is now found in the rim of the ilmenite grains (Figs. 2a–2d). As observed by Duchesne (1972), the newly formed ilmenite is in optical continuity with the primary grain of ilmenite. If this is the reaction process, the original boundaries of the primary grains of ilmenite and magnetite should correspond to the location of the hercynite rim (Fig. 2d). Conditions that influence the sub-solidus reactions Three of the samples (SL1, SL3, and TR1-6) do not contain ilmenite grains with TiO2 rich cores and Fe-rich rims. However, these samples provide insight on the conditions that promote the sub-solidus reaction between ilmenite and magnetite. The ilmenite grains in samples SL1 and SL4 show exceptionally homogeneous compositions (Fig. 3). This suggests that a mechanism of diffusion controlled the concentrations of elements in these samples more efficiently than in others. Diffusion is a chemical process that only takes place when sufficient energy is available in the system (e.g., Dodson 1986; Hodges 2003), and is, therefore, temperature dependent. These two samples have been metamorphosed. The diffusion that occurred during cooling from the magmatic temperature to the crystallization point might have continued (or restarted) when high temperatures associated with metamorphism provided sufficient energy to support diffusion again, leaving more time for homogenization. However, samples SL2, SL3, and SL5 are from the same sampling locality and are not as homogeneous (Fig. 3). Remarkably, these samples are coarsergrained (up to 3 mm), and this may have influenced the diffusion process. Even though the high temperature reached during metamorphism would have favoured the diffusion, the larger grain size could have made the homogenization of the Published by NRC Research Press

Comparison of the studied samples with lunar ilmenite compositions When calculating the Fe3⫹ content of the ilmenite using the electron microprobe analyses and the calculation from Lindsley and Frost (1992), none of the analyzed ilmenite in this study is free of Fe3⫹ (Table 1). However, the calculation of Lindsley and Frost (1992) to determine the Fe3⫹content is not a method that can absolutely demonstrate that there is Fe3⫹ in the analyzed mineral. For example, using this calculation for a lunar ilmenite that contains 51.70 wt.% TiO2 and 37.70 wt.% FeO (ilmenite from Apollo 14 soil 14258, Powell and Weiblen 1972), it results in an ilmenite with 6.69 wt.% Fe2O3, which is incorrect. To be certain of the proportion of iron present as Fe2O3 in the studied ilmenite, micro-XANES analyses that enable spot size (7 ␮m) analyses need to be pursued. Spot size analyses will be necessary as the cores of the grains are rich in hematite for some of the samples (e.g., Fig. 2b). We are thus investigating this method as the next step for this study. We compare the compositions of the ilmenite in this study with lunar ilmenite using FeOT (Fig. 5a). The analyses of ilmenite from Mirepoix, Sanford Lake, and Degrosbois deposits that are higher than 50 wt.% TiO2 overlap in compositions with lunar ilmenite in FeOT and in TiO2 (FeO: 32.3– 48.14 wt.%; TiO2: 50.7–56.3 wt.%, Papike et al. 1991 and references therein) (Fig. 5a). If some of the Fe is present as Fe2O3 in the samples from this study, the data would plot more to the left

(wt.%)

Fig. 5. (a) TiO2 (wt.%) versus FeOT (wt.%) and (b) MgO (wt.%) and (c) MnO (wt.%) versus TiO2 wt.% for ilmenite from the Mirepoix, Sanford Lake, and Degrosbois deposits. Ilmenite from the six Apollo missions and from two Luna sample returned missions compiled by Papike et al. (1991; see references therein for individual analysis) are shown for comparison. Ilmenite from the Mirepoix, Sanford Lake, and Degrosbois might contain some Fe2O3, but their analyses are plotted as FeOT to compare with lunar ilmenite compositions.

(wt.%)

(wt.%)

grains more difficult, as diffusion is faster in small grain size than in larger (e.g., Dodson 1986; Hodges 2003). The most inhomogeneous sample is TR1-6 where cores of ilmenite grains are extremely rich in hematite exsolution (Fig. 2a). This sample from the Mirepoix deposit has not been metamorphosed, and the diffusion of the hematite to the border of the ilmenite may have only occurred during the cooling of the intrusion, thus limiting the homogenization process. Other factors that could have influenced the efficiency of the reaction are the proportions and abundances of the phases. For example, sample DEG6 that shows a wide variation in TiO2 and total Fe (Fig. 3a) contains four times more ilmenite than magnetite grains and is from a disseminated part of the deposit. It is suggesting that not all grains were able to reequilibrate because the reactants were not in contact with each other during cooling or metamorphism. The optimal conditions for the crystallization of ilmenite with no trace of Fe3⫹ would, therefore, occur in massive ilmenite and magnetite deposits that were very fine grained and metamorphosed. The massive structure of the deposit would ensure that the oxide minerals (the reactants) are in contact with each other. Metamorphism would help to reach the high temperatures that favour diffusion of the hematite in the ilmenite to react with magnetite and lower the hematite content of the ilmenite. Finer grain sizes would also favour diffusion. Finally, although not specifically discussed here, we could argue that ores containing ilmenite in smaller proportions than magnetite would ensure that the reaction between the two oxide minerals is complete. The massive ilmenite– magnetite deposits associated with anorthosite massifs of the Grenville Province that are older than the pervasive Ottawan Orogeny (1080 –1020 Ma, Rivers 2008) are thus good prospection targets for finding lunar-like ilmenite composition.

73

(wt.%)

(wt.%)


Morisset et al.

(wt.%)



74

of the diagram and would still overlap with the lunar ilmenite compositions in total Fe. The ilmenite compositions from this study also overlap in MgO and MnO concentrations with the lunar compositions (MgO: 0.15–9.63 wt.%; MnO: 0.15– 0.52 wt.%, Papike et al. 1991 and references therein). However, most lunar ilmenite is richer in TiO2 for a given MgO or MnO concentration compared with the ones from this study, forming a distinct field in a diagram MgO or MnO versus TiO2 (Figs. 5b, 5c). There is no overlap between the Cr2O3 concentrations of the ilmenite from this study and the lunar ilmenite analyses that have higher Cr2O3 concentrations (0.16 –2.12 wt.%, Papike et al. 1991 and references therein; Table 1). Ilmenite from magmatic deposits where it is the major oxide mineral phase contains higher Cr2O3 concentrations (e.g., Saint-Urbain deposit 0.08 – 0.47 wt.%, Morisset et al. 2010; Tellnes deposit: 0.01– 0.21 wt.%, Charlier et al. 2007) but the lack of, or low concentrations of, magnetite prevent the oxide minerals from reacting during cooling and thus the ilmenite from losing some of its Fe2O3. Proposed terrestrial ilmenite candidate for lunar simulant production: Sanford Lake Overall, ilmenite from the studied deposits that overlap in compositions (except for Cr2O3) with the lunar ones are those from Sanford Lake: samples SL1, SL3, and SL4. The ilmenite compositions in these samples are the most homogeneous (Fig. 3a) and poorest in calculated Fe2O3 of all the samples and thus the ones where conditions were optimal for the ilmenite to lose its Fe2O3 initial composition (i.e., temperature and cooling history, proportion and composition of the ilmenite and the magnetite, size of the grains). This result has important implications for the testing of exploration methods for ISRU using terrestrial materials with comparable ilmenite compositions to those in lunar rocks and regolith because it provides a geological context in which to look for ilmenite simulant (i.e., massive fine-grained ilmenite and magnetite deposits that haves been metamorphosed). Ilmenite from the Sanford Lake deposits could be mixed with other components to produce high-Ti mare soil as needed for ISRU testing (Edmunson et al. 2010). As stated in the preceding text, ilmenite from the Sanford Lake samples are the ones where the ilmenite is the closest in composition to the lunar ilmenite but it does not overlap in Cr2O3 and it overlaps with the highest concentrations of FeOT and the lowest concentrations of MgO with lunar ilmenite compositions (Figs. 5a, 5b). The production of oxygen from the reduction of ilmenite and lunar mare basalt or soil using hydrogen is based mostly on the reduction of FeO to Fe metal (e.g., Allen et al. 1994; Gibson et al. 1994) and on partial reduction of TiO2 to suboxides (Gibson et al. 1994) while other major oxides have not been reported to reduce. We can thus deduce that what is important in an ilmenite simulant is similarity in the FeO concentration and to a lesser extent TiO2 concentration. We do not think that the difference in Cr2O3 and the fact that the Sandford Lake ilmenite do not overlap with the whole field of MgO concentrations would jeopardize the validity of oxygen extraction testing with this simulant. Moreover, since MgO substitutes for FeO in ilmenite, we could stipulate that ilmenite poor in MgO will be richer in FeO, and thus soils containing ilmenite poor in MgO should be targeted for oxygen extraction.


The Sanford Lake deposits were open pits that have been abandoned since 1983 (USGS Mineral resource data system: deposit identification number 60001558). Sampling material for simulant is easily accessible, as large piles of crushed material are found surrounding the abandoned pits. It would, however, require some sorting and processing to liberate ilmenite grains from the ore. The crushed material is a mixture of massive ore and waste material. For separation of ilmenite from other mineral phases present in the ore deposits (magnetite and minor silicates: plagioclase, pyroxene, and garnet), crushing to the size of grain liberation would be required. The size of grain liberation is based on the grain sizes observed under the optical microscope. For Sanford Lake, the smallest ilmenite grains are 0.2 mm ⫻ 0.2 mm, and thus the grains could be crushed until they reach a size of 180 –160 ␮m or less. This would be sufficient to free the ilmenite from the hercynite-bearing rim. After that, the separation technique of Duchesne (1966) could be used. First, a magnet would remove most of the magnetite, then the Frantz Magnetic separator would remove the remaining magnetite. Ilmenite, if hematitefree, would be nonmagnetic (Ishikawa and Akimoto 1957). For the small amount of silicates that could be present in the nonmagnetic fraction with ilmenite, heavy liquids can be used, as is common in ore processing. Consideration of other possible sources for ilmenite simulant Ilmenite found in the extremely reduced andesitic to dacitic of Disko Island (Greenland) are most probably free of Fe3⫹, as native Fe is present in the mineral assemblage (Pedersen 1981). These lavas have been strongly contaminated by reducing sediments, and ilmenite phenocrysts are now found as Fe–Ti oxide-metal-sulphide aggregates that contain ilmenite, armacolite, and rutile in various proportions (Pedersen 1981). Even though the composition of the ilmenite (TiO2: 55.3–54.6 wt.%; FeO: 34.5–35.1 wt.%; MgO: 7.6 – 8.0 wt.%, Pedersen 1981) is lunar-like, it is present in low proportions in the rock (⬍2%, Pedersen 1981) and it is found as aggregates, which makes it a problematic target for simulant production, as it would be difficult to separate the ilmenite from the intertgrown phases in the aggregates. Ilmenite from placer deposits would be comparatively easy to recover, as they are already crushed from the host rock and concentrated in the heavy sand. Unfortunately, ilmenite from sand deposits does not have a composition that is comparable with the lunar ilmenite, as it is commonly altered and often contains leucoxene and rutile (e.g., Lynd et al. 1954; Bailey et al. 1956; Temple 1966; Pownceby 2010). In some cases, ilmenite has been reported as being hydrated (mixture of ilmenite and pseudorutile that contains structural water: Frost et al. 1983; Pownceby 2010). Where the ilmenite is less altered, hematite is still present (Lynd et al. 1954). The alteration process diminishes the Fe content of the ilmenite and favours the formation of ilmenite much richer in TiO2 than magmatic ilmenite or lunar ilmenite. The TiO2 concentrations in ilmenite from sand deposits can range from 40 wt.% to over 80 wt.% (e.g., Pownceby 2010), implying leaching of Fe during the alteration process. Oxidation and Fe-leaching creates porosity in the grains (Pownceby 2010) that might facilitate the reduction process compared with lunar ilmenite and would thus not be a good analogue for ISRU testing. Even Published by NRC Research Press

Morisset et al.

though it would be easy to extract, its severe alteration makes ilmenite from placer deposits unsuitable for use as a lunar simulant.


Conclusions The Mirepoix (Mattawa anorthosite massif), the Sanford Lake (Adirondack anorthosite massif), and the Degrosbois (Morin anorthosite massif) Fe–Ti deposits were investigated using petrographic techniques and electron microprobe analyses. Sub-solidus reactions occurring between ilmenite and magnetite were considered to determine if they could be an efficient mechanism to produce ilmenite free of Fe3⫹, and thus provide reaction conditions favorable to the production of terrestrial ilmenite that is lunar-like in composition. The major findings of the study are as follows: 1.

Hercynite found in the rim of ilmenite grains in contact with magnetite grains is first formed as an exsolution in the magnetite grains. The reaction between the hematite contained within the ilmenite and the ulvöspinel contained within magnetite [Fe2O3 (in ilm) ⫹ Fe2TiO4 (in mt) ¡ FeTiO3 ⫹ Fe3O4] as described by Duchesne (1972) produces new ilmenite and magnetite. The newly formed ilmenite is in optical continuity with the primary ilmenite grain, and includes hercynite. 2. Ilmenite grains are free of hematite exsolutions in three of the eight samples analyzed. 3. In most of the studied samples, the core of the ilmenite grains is richer in TiO2 and poorer in total Fe, suggesting diffusion of the hematite component to the border of the grains to react with magnetite. In other samples, ilmenite grains can either be homogeneous in composition, suggesting that diffusion was more effective, or are much richer in hematite in their core, suggesting that diffusion was not as effective. 4. The optimal conditions that favour diffusion and reaction of ilmenite with magnetite to produce ilmenite free of Fe3⫹ would occur in a massive fine-grained ilmenite and magnetite deposit that has been metamorphosed. The studied ilmenite compositions are comparable with lunar ilmenite compositions in TiO2, total Fe, MgO, and MnO, but samples from the Sanford Lake deposits are closest in composition and are proposed as the best candidate for use in the production of simulant for ISRU testing for future missions to the Moon. Micro-XANES analyses are required to confirm whether these studied terrestrial ilmenites are indeed completely free of Fe3⫹.

Acknowledgements C.-E. Morisset thanks Marian Lupulescu, New York State Museum, for providing samples SL1, SL2, and SL3 for this study; James McLelland for organizing a field trip to the Adirondack mountains to collect more samples for this study; and Trijet Mining Corporation, for access to the Degrosbois field site. Mr. Lang Shi is thanked for his assistance with the electron microprobe analyses at McGill University. Special thanks to J.-C. Duchesne for having shared his contagious passion for Fe–Ti oxide minerals. Review of the manuscript by Richard Herd was extremely helpful. We are grateful to John Spray, an anonymous reviewer, and Doug Rickman for their

75

constructive and useful reviews and editorial assistance from Paul Sylvester and John Greenough. During the course of the study, C.-E. Morisset was supported by a Natural Sciences and Engineering Research Council of Canada (NSERC) Grant for Visiting Fellowships in Canadian Government Laboratories held at the Canadian Space Agency (2009 –2012).

References Allen, C.C., Morris, R.V., and Mckay, D.S. 1994. Experimental reduction of lunar mare soil and volcanic glass. Journal of Geophysical Research, 99(E11): 23173–23185. doi:10.1029/ 94JE02321. Armstrong, J.T. 1995. CITZAF: a package of correction programs for the quantitative electron microbeam X-ray analysis of thick polished materials, thin films, and particles. Microbeam Analysis, 4: 177–200. Ashwal, L.D. 1993. Anorthosite. Minerals and rocks. Edited by P.J. Wyllie, A. El Gorey, W. von Engelhardt, and T. Hahn. SpingerVerlag, Berlin. pp 422. Bailey, S.W., Cameron, E.N., Spedden, H.R., and Weege, R.J. 1956. The alteration of ilmenite in beach sands. Economic Geology and the Bulletin of the Society of Economic Geologists, 51(3): 263– 279. doi:10.2113/gsecongeo.51.3.263. Briggs, R.A., and Sacco, A., Jr 1991. Hydrogen reduction mechanisms of ilmenite between 823 and 1353 K. Journal of Materials Research, 6(03): 574 –584. doi:10.1557/JMR.1991.0574. Buddington, A.F., and Lindsley, D.H. 1964. Iron-titanium oxide minerals and synthetic equivalents. Journal of Petrology, 5: 310 – 357. Cameron, E.N. 1971. Opaque minerals in certain lunar rocks from Apollo 12. Proceedings of the 2nd Lunar Science Conference, pp. 193–206. Charlier, B., Skår, Ø., Korneliussen, A., Duchesne, J.-C., and Vander Auwera, J. 2007. Ilmenite composition in the Tellness Fe–Ti deposit, SW Norway: fractional crystallization, postcumulus evolution and ilmenite-zircon relation. Contributions to Mineralogy and Petrology, 154(2): 119 –134. doi:10.1007/s00410-007-0186-8. Cloutis, E.A., McCormack, K.A., Bell, J.F., III, Hendrix, A.R., Bailey, D.T., Craig, M.A., et al. 2008. Ultraviolet spectral reflectance properties of common planetary minerals. Icarus, 197(1): 321–347. doi:10.1016/j.icarus.2008.04.018. Colaprete, A., Schultz, P., Heldmann, J., Wooden, D., Shirley, M., Ennico, K., et al. 2010. Detection of water in LCROSS ejecta plume. Science, 330(6003): 463– 468. doi:10.1126/science. 1186986. PMID:20966242. Corriveau, L., Perrault, S., and Davidson, A. 2007. Prospective metallogenic settings of the Grenville Province. In Mineral deposits of Canada: a synthesis of major deposit-types, district metallogeny, the evolution of geological provinces, and exploration methods. Edited by W.D. Goodfellow. Geological Survey of Canada, Mineral Deposits Division, Special Publication 5, pp. 819 – 847. CSEW6 Steering Committee. 2009. Canadian scientific priorities for the Global Exploration Strategy. Proceedings of the 6th Canadian Space Exploration Workshop, Dec 1–3 2008, Quebec, Canada. Davidson, A. 1998. An overview of the Grenville Province geology, Canadian Shield. In Geology of the Precambrian Superior and Grenville provinces and Precambrian fossils in North America. Edited by S.B. Lucas and M.R. St-Onge. Geological Survey of Canada, Geology of Canada, 7, pp. 205–270. Dodson, M.H. 1986. Closure profiles in cooling systems. Materials Science Forum, 7: 145–154. doi:10.4028/www.scientific.net/MSF. Published by NRC Research Press


76 7.145. Doig, R. 1991. U-Pb zircon ages of Morin anorthosite suite rocks, Grenville Province, Québec. The Journal of Geology, 99(5): 729 – 738. doi:10.1086/629535. Duchesne, J.-C. 1966. Séparation rapide des minéraux des roches. Annales de la Société Géologique de Belgique, 87: 347–356. Duchesne, J.-C. 1970. Microtextures of the Fe–Ti oxide minerals in the south-Rogaland anorthositic complex (Norway). Annales de la Société Géologique de Belgique, 93: 527–544. Duchesne, J.-C. 1972. Iron-titanium oxide minerals in the BjerkremSogndal Massif, South-western Norway. Journal of Petrology, 13: 57– 81. Edmunson, J., Betts, W., Rickman, D., McLemore, C., Fikes, J., Stoeser, D., Wilson, S., and Schrader, C. 2010. NASA lunar regolith stimulant program. 41st Lunar and Planetary Science Conference, Houston, Texas. Abstract # 1786. Emslie, R.F. 1975. Major rock units of the Morin Complex, southwestern Quebec. Geological Survey of Canada Paper 74-48, 37 pp. Feldman, W.C., Maurice, S., Binder, A.B., Barraclough, B.L., Elphic, R.C., and Lawrence, D.J. 1998. Fluxes of fast and epithermal neutrons from Lunar Prospector: evidence of water ice at the lunar poles. Science, 281(5382): 1496 –1500. doi:10.1126/science.281. 5382.1496. PMID:9727973. Feldman, W.C., Maurice, S., Lawrence, D.J., Little, R.C., Lawson, S.L., Gasnault, O., et al. 2001. Evidence for water ice near the lunar poles. Journal of Geophysical Research, 106(E10): 23231– 23251. doi:10.1029/2000JE001444. Frost, M.T., Grey, I.E., Harrowfield, I.R., and Mason, K. 1983. The dependence of alumina and silica contents on the extent of alteration of weathered ilmenites from Western Australia. Mineralogical Magazine, 47(343): 201–208. doi:10.1180/minmag.1983.047. 343.10. George, J.A., Mattes, G.W., Rogers, K.N., Magruder, D.F., Paz, A.J., Vaccaro, H.M., et al. 2012. Resolve mission architecture for lunar resource prospecting and utilization. 43rd Lunar and Planetary Science Conference, Houston, Texas. Abstract # 2583. Gibson, M.A., and Knudsen, C.W. 1988. Lunar oxygen production from ilmenite. Proceedings from the Second Conference on Lunar Bases, NASA-JSC, Houston, Texas. Gibson, M.A., Knudsen, C.W., Brueneman, D.J., Allen, C.C., Kanamori, H., and McKay, D.S. 1994. Reduction of lunar basalt 70035: Oxygen yield and reaction product analysis. Journal of Geophysical Research, 99(E5): 10887–10897. doi:10.1029/94JE00787. Gross, S.O. 1968. Titaniferous ores of the Sanford Lake district, New York. In Ore deposits of the United States, 1933–1967 (Graton Sales Volume), Vol. 1. Edited by J.D. Ridge. American Institute of Mining, Metalurgical and Petroleum Engineers, New York. pp. 140 –153. Hébert, C. 2003. Géologie de la région du lac Moncouche (22D/15). Ministère des Ressources naturelles de la Faune et des Parcs, Québec, SI-C3G-22D15-03X. Hébert, C., and Cadieux, A.-M. 2003. Géologie de la région des lacs Portneuf et Maria-Chapdelaine (22E/02 et 22E/01). Ministère des Ressources naturelles, Québec, RE 2002-13, 44 pp. Hébert, C., Cadieux, A.-M., and van Breemen, O. 2005. Temporal evolution and nature of Ti–Fe–P mineralization in the anorthosite– mangerite– charnockite– granite (AMCG) suites of the southcentral Grenville Province, Saguenay – Lac St. Jean area, Quebec, Canada. Canadian Journal of Earth Sciences, 42(10): 1865–1880. doi:10.1139/e05-050.

Can. J. Earth Sci., Vol. 50, 2013 Heiken, G.H., and Vaniman, D.T. 1990. Characterization of lunar ilmenite resources. 20th Lunar and Planetary Science Conference, Houston, Texas. pp. 239 –247. Hodges, K.V. 2003. Geochronology and thermochronology in orogenic systems. In Treatise on Geochemistry, 3, pp. 263–292. Ishikawa, Y., and Akimoto, S.I. 1957. Magnetic properties of the FeTiO3-Fe2O3 solid solution series. Journal of the Physical Society of Japan, 12(10): 1083–1098. doi:10.1143/JPSJ.12.1083. Kelly, W.M. 1979. Chemistry and genesis of titaniferous magnetite and related ferromagnesian silicates, Sanford Lake deposits, Tahawus, New York. Ph.D. thesis, University of Massachusetts, Amherst, US. Lindsley, D.H. 1973. Delimitation of the hematite-ilmenite miscibility gap. Geological Society of America Bulletin, 84(2): 657– 662. doi:10.1130/0016-7606(1973)84⬍657:DOTHMG⬎2.0.CO;2. Lindsley, D.H., and Frost, B.R. 1992. Equilibria among Fe–Ti oxides, pyroxenes, olivine, and quartz. I. Theory. The American Mineralogist, 77: 987–1003. Lynd, L.E., Sigurdson, H., North, C.H., and Anderson, W.W. 1954. Characteristics of titaniferous concentrates. Mining Engineering, 6: 817– 824. Martignole, J., and Schrijver, K. 1970. Tectonic setting and evolution of the Morin anorthosite, Grenville Province, Quebec. Geological Society of Finland Bulletin, 42: 165–209. McCubbin, F.M., Steele, A., Hauri, E.H., Nekvasil, H., Yamashita, S., and Hemley, R. 2010. Nominally hydrous magmatism on the Moon. Proc. Natl. Acad. Sci. USA, 107(25): 11223–11228. doi: 10.1073/pnas.1006677107. PMID:20547878. McLelland, J., Hamilton, M., Selleck, B., McLelland, J., Walker, D., and Orrell, S. 2001. Zircon U–Pb geochronology of Ottawan Orogeny, Adirondack Highlands, New York: regional and tectonic implications. Precambrian Research, 109(1–2): 39 –72. doi: 10.1016/S0301-9268(01)00141-3. McLelland, J.M., Bickford, M.E., Hill, B.M., Clechenko, C.C., Valley, J.W., and Hamilton, M.A. 2004. Direct dating of Adirondack massif anorthosite by U–Pb SHRIMP analysis of igneous zircon: implications for AMCG complexes. Geological Society of America Bulletin, 116(11): 1299 –1317. doi:10.1130/B25482.1. Morisset, C.-E. 2002. Étude pétrologique et géochimique de la minéralisation en oxydes de fer et titane et en apatite du lac Mirepoix (Anorthosite du lac Saint-Jean, Québec). Master’s thesis, Département des sciences de la Terre et de l’atmosphère, Université du Québec a` Montréal, Montréal, Que. Morisset, C.-E. 2003. A study of mineral compositions of the Lac Mirepoix layered complex, Lac St-Jean anorthosite complex (Québec, Canada). In Ilmenite deposits and their geological environment. Edited by J.-C. Duchesne and A. Korneliussen. Norges Geologiske Undersøkelse, Special Publication No. 9, Trondheim. pp. 84 – 85. Morisset, C.-E., Scoates, J.S., Weis, D., Sauvé, M., and Stanaway, K.J. 2010. Rutile-bearing ilmenite deposits associated with the Proterozoic Saint-Urbain and Lac Allard anorthosite massifs, Grenville Province, Quebec. Canadian Mineralogist, 48(4): 821– 849. doi:10.3749/canmin.48.4.821. Owens, B.E., and Dymek, R.F. 2005. Rediscovery of the Mattawa Anorthosite Massif, Grenville Province, Quebec. Canadian Journal of Earth Sciences, 42(10): 1699 –1718. doi:10.1139/e05-073. Papike, J., Taylor, L., and Simon, S. 1991. Lunar minerals. In Lunar sourcebook: a user’s guide to the moon. Edited by G.H. Heinken, D.T. Vaniman, and B.M. French. Cambridge University Press, New York. pp. 121–181. Published by NRC Research Press


Morisset et al. Pedersen, A.K. 1981. Armacolite-bearing Fe–Ti oxide assemblages in graphite-equilibrated salic volcanic rocks with native iron from Disko, Central West Greeland. Contributions to Mineralogy and Petrology, 77(4): 307–324. doi:10.1007/BF00371560. Powell, B.N., and Weiblen, P.W. 1972. Petrology and origin of lithic fragments in the Apollo14 regolith. In Proceedings of the 3rd Lunar and Planetary Science Conference, Houston, Texas, pp. 837– 852. Pownceby, M.I. 2010. Alteration and associated impurity element enrichment in detrital ilmenites from the Murray Basin, southeast Australia: a product of multistage alteration. Australian Journal of Earth Sciences, 57(2): 243–258. doi:10.1080/08120090903521705. Ramdohr, P. 1953. Ulvöspinel and its significance in titaniferous iron ores. Economic Geology and the Bulletin of the Society of Economic Geologists, 48(8): 677– 688. doi:10.2113/gsecongeo.48.8. 677. Rivers, T. 1997. Lithotectonic elements of the Grenville Province: review and tectonic implications. Precambrian Research, 86(3– 4): 117–154. doi:10.1016/S0301-9268(97)00038-7. Rivers, T. 2008. Assembly and preservation of lower, mid, and upper orogenic crust in the Grenville Province – Implications of the evolution of large hot long-duration orogens. Precambrian Research, 167(3-4): 237–259. doi:10.1016/j.precamres.2008.08.005. Rivers, T., Ketchum, J., Indares, A., and Hynes, A. 2002. The High Pressure belt in the Grenville Province: architecture, timing and exhumation. Canadian Journal of Earth Sciences, 39(5): 867– 893. doi:10.1139/e02-025. Robinson, P., Harrison, R.J., McEnroe, S.A., and Hargraves, R.B.

77 2002. Lamellar magnetism in the hematite-ilmenite series as an explanation for strong remanent magnetization. Nature, 418(6897): 517–520. doi:10.1038/nature00942. PMID:12152075. Rose, E.R. 1969. Geology of titanium and titaniferous deposits of Canada. Economic Geology Report 25, Geological Survey of Canada, Ottawa, Ont. Saal, A.E., Hauri, E.H., Lo Cascio, M., Van Orman, J.A., Rutherford, M.C., and Cooper, R.F. 2008. Volatile content of lunar volcanic glasses and the presence of water in the Moon’s interior. Nature, 454(7201): 192–195. doi:10.1038/nature07047. PMID:18615079. Sanders, G.B., and Larson, W.E. 2011. The impact of lunar polar volatiles on in-situ resource utilization. Wet vs. Dry Moon Conference, Houston, Texas. Abstract # 6060. Spudis, P.D., Bussey, D.B.J., Baloga, S.M., Butler, B.J., Carl, D., Carter, L.M., et al. 2010. Initial results for the north pole of the Moon from Mini-SAR Chandrayaan-1 mission. Geophysical Research Letters, 37, 6 pp. Taylor, L.A., and Carrier, W.D., III. 1993. Oxygen production on the moon: an overview and evaluation. Edited by J.S. Lewis, M.S. Mathews, and M.L. Guerrieri. In Resources of Near-Earth Space. University of Arizona Series, Tucson. pp. 69 –108. Temple, A.K. 1966. Alteration of ilmenite. Economic Geology and the Bulletin of the Society of Economic Geologists, 61(4): 695– 714. doi:10.2113/gsecongeo.61.4.695. Vaniman, D., French, B., and Heiken, G. 1991. Afterword. In Lunar sourcebook. Edited by G.H. Heiken, D.T. Vaniman, and B.M. French. Cambridge University Press, New York. pp. 633– 654.
