Spectral reflectance properties of magnetites: Implications for remote sensing

Spectral reflectance properties of magnetites: Implications for remote sensing

Accepted Manuscript Spectral reflectance properties of magnetites: Implications for remote sensing Matthew R.M. Izawa , Edward A. Cloutis , Tesia Rhi...

NAN Sizes 0 Downloads 6 Views

Accepted Manuscript

Spectral reflectance properties of magnetites: Implications for remote sensing Matthew R.M. Izawa , Edward A. Cloutis , Tesia Rhind , Stanley A. Mertzman , Daniel M. Applin , Jessica M. Stromberg , David M. Sherman PII: DOI: Reference:

S0019-1035(18)30225-2 https://doi.org/10.1016/j.icarus.2018.10.002 YICAR 13046

To appear in:

Icarus

Received date: Revised date: Accepted date:

17 May 2018 19 September 2018 1 October 2018

Please cite this article as: Matthew R.M. Izawa , Edward A. Cloutis , Tesia Rhind , Stanley A. Mertzman , Daniel M. Applin , Jessica M. Stromberg , David M. Sherman , Spectral reflectance properties of magnetites: Implications for remote sensing, Icarus (2018), doi: https://doi.org/10.1016/j.icarus.2018.10.002

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT 1

AC

CE

PT

ED

M

AN US

CR IP T

Highlights:  Reflectance spectra of magnetite were measured over the 0.2-2.5 μm wavelength range  Electron delocalization leads to high extinction and surface-dominant scattering  Strong scattering leads to local maxima near transition energies (Fresnel peaks)  Magnetite is spectrally distinct from most Fe2+-Fe3+-Ti4+ oxides

ACCEPTED MANUSCRIPT 2

Spectral reflectance properties of magnetites: Implications for remote sensing Matthew R. M. Izawa1,2, Edward A. Cloutis1, Tesia Rhind1, Stanley A. Mertzman3, Daniel M. Applin1 Jessica M. Stromberg4 and David M. Sherman5 Corresponding author: Matthew R. M. Izawa - [email protected]

CR IP T

1. Department of Geography, Univeristy of Winnipeg, Winnipeg MB R3B 2E9 Canada; [email protected]; [email protected]; [email protected]

2. Institute for Planetary Materials, Okayama University – Misasa, 827 Yamada, Misasa, Tottori 682-0193, Japan [email protected] 3. Department of Earth and Environment, Franklin and Marshall College, Lancaster, Pennsylvania, USA 17604-2615; [email protected]

AN US

4. CSIRO Mineral Resources Flagship, 26 Dick Perry Avenue, WA 6151, Australia, [email protected]

Dr. Matthew R. M. Izawa Institute for Planetary Materials, Okayama University – Misasa, 827 Yamada, Misasa, Tottori, Japan

ED

Corresponding author:

M

5. School of Earth Sciences, University of Bristol, Bristol BS8 1RJ United Kingdom; [email protected]

AC

CE

PT

Highlights:  Reflectance spectra of magnetite were measured over the 0.2-2.5 μm wavelength range  Electron delocalization leads to high extinction and surface-dominant scattering  Strong scattering leads to local maxima near transition energies (Fresnel peaks)  Magnetite is spectrally distinct from most Fe2+-Fe3+-Ti4+ oxides

ACCEPTED MANUSCRIPT 3

Abstract Magnetite (Fe3+(Fe2+Fe3+)2O4) is ubiquitous in Earth and planetary materials, forming in igneous, metamorphic, and sedimentary settings, sometimes influenced by microbiology. Magnetite can be used to study many and varied planetary processes, such as the oxidation state of magmas, paleomagnetism, water-rock interactions such as serpentinization, alteration and metamorphism

CR IP T

occurring on meteorite parent bodies, and for astrobiology. The spectral reflectance signature of magnetite in the ultraviolet, visible, and near-infrared is somewhat unusual compared to common planetary materials, suggesting that remote detection and characterization of magnetite should be possible. Here we present a systematic investigation of the reflectance spectral properties of magnetite using natural and synthetic samples. We investigate the effects of chemical substitutions, grain size variations, and mixtures with other phases in order to better constrain

AN US

remote spectral searches for, and interpretation of, magnetite-bearing lithologies. Magnetite is characterized by high extinction over the entire wavelength range considered here, and therefore surface scattering dominates over volume scattering. Magnetite reflectance spectra are strongly influenced by the presence of delocalized electrons above the Verwey transition temperature (~120 K), leading to metal-like scattering behavior, that is, high extinction, surface scattering

M

dominant, and a general increase in reflectance with increasing wavelength, ―red-sloped and featureless‖. Superimposed upon the metal-like reflectance are local reflectance maxima which

ED

we ascribe to Fresnel reflectance peaks corresponding to Fe-O oxygen-metal charge transfer processes (~0.27 and ~0.39 μm) and Fe-related field-d orbital transitions (~0.65 μm). We also

PT

find a systematic shift in the wavelength position of the 0.65 μm Fresnel peak with increasing chemical impurity in magnetite. Magnetite reflectance spectra are most similar to those of

CE

titanomagnetite and wüstite, and unlike those of other Fe-(Ti) oxides, such as ilmenite, hæmatite, ulvospinel, maghemite, pseudobrookite, and armalcolite.

AC

Keywords: Magnetite; Reflectance spectroscopy; B-type asteroids; Carbonaceous Chondrites

1. Introduction

Magnetite (ideal formula: Fe3+(Fe2+Fe3+)2O4) is found in a wide range of extraterrestrial materials, as well as in terrestrial igneous, metamorphic, and sedimentary environments (e.g.,

ACCEPTED MANUSCRIPT 4

Waychunas, 1991). Its ubiquity, and the relationship between its structure and composition with its petrogenetic conditions make it a particularly valuable ―tracer‖ of a wide variety of processes. For instance, the compositions and intergrowth relations of magnetite and associated Fe-Ti oxide phases have been used to constrain the oxidation state of magmas (e.g., Powell & Powell, 1977). Magnetite also plays an important role in preserving magnetic signatures in rocks (e.g., Rochette

CR IP T

et al., 1992). As a product of serpentinization processes, magnetite records past water-rock interactions (e.g., Mayhew et al., 2013; McCollom 2016), and can also act as a catalyst for the formation of organic molecules in these settings (Mao et al., 1994). Magnetite-bearing lithologies are common in aqueously altered carbonaceous chondrites (e.g., Hua and Buseck, 1998; Izawa et al., 2010a; b; Krot et al., 2013), and spectral studies of some asteroids are

AN US

consistent with significant quantities of magnetite (e.g., Yang and Jewitt, 2010).1.1 Magnetite in carbonaceous chondrites and dark asteroids: Magnetite-bearing lithologies are common in aqueously-altered carbonaceous chondrites (e.g., Brearley and Jones, 1998), where the magnetite occurs as a reaction product of the aqueous alteration of olivine and to a lesser extent pyroxene. Aqueously-altered carbonaceous chondrites (CM, CI, CR, and some ungrouped C chondrites)

M

contain abundant magnetite, at abundances up to 10 wt% (Brearley and Jones, 1998), with some Tagish Lake (C2 ungrouped) material containing as much as 17 wt. % magnetite (Izawa et al,

ED

2010a), and in a variety of morphologies (e.g., Hua and Buseck, 1998). Because magnetite is one of the most abundant alteration products formed in carbonaceous chondrites during aqueous alteration, it can be useful in providing a timeline and record of the metamorphic conditions

PT

experienced by aqueously altered meteorites (Herzog et al., 1973; Lewis and Anders, 1975; Swindle, 1998; Barth et al., 2013; Daulton et al., 2013; Krot et al., 2013). Petrographic evidence

CE

suggests that magnetite can form by multiple processes in carbonaceous chondrites, so that detailed studies of magnetites as a function of morphology and composition may be able to

AC

provide insights into carbonaceous chondrite parent body processes (Hua and Buseck, 1998; Barth et al., 2013; Daulton et al., 2013). Due to the ubiquity of magnetite in aqueously altered carbonaceous chondrites, magnetite is a plausible major rock-forming mineral on the surfaces of many dark asteroids. Yang and Jewitt (2010), in an investigation of B-type asteroid spectra, observed a broad local minimum in reflectance near 1 µm consistent with magnetite. Bennu, the target of the OSIRIS-REX sample return mission, is a B-type asteroid and may contain a significant amount of magnetite (Lauretta et al., 2017). Magnetite may also form a significant

ACCEPTED MANUSCRIPT 5

component of the Cerean surface, which contains a substantial fraction of material spectrally similar to aqueously altered carbonaceous chondrite meteorites containing abundant magnetite (e.g., Schäfer et al., 2018). Ryugu, the target of the JAXA Hayabusa-2 mission (Watanabe et al., 2017), may also contain magnetite based on its spectral similarity to aqueously-altered carbonaceous chondrites (e.g., Le Corre et al., 2018, and references therein). Magnetite is also

CR IP T

found in some in chondritic interplanetary dust particles (e.g., Christoffersen and Buseck, 1986), in comet nuclei based on analysis of terminal grains in type-B Stardust aerogel tracks (Changela et al., 2012; Hicks et al., 2017), and occurs as one form of presolar grains found in the LAP 031117 CO3 chondrite (Zega et al., 2015).

1.2 Magnetite in non-carbonaceous extraterrestrial materials: Titanomagnetite is a common

AN US

late-stage crystallization product in Martian shergottite magmas (e.g., Herd et al., 2001). Magnetite in the Martian orthopyroxenite ALHA 84001 has been the source of considerable controversy due to its resemblance to magnetite produced by terrestrial magnetotactic bacteria. Nanodimensional magnetites have been found in the fracture zones of this meteorite that are similar in size and shape to those found in terrestrial magnetotactic bacteria (McKay et al., 1996;

M

Gibson et al., 2001), though further study has revealed evidence for an abiotic formation mechanism (e.g., Scott et al.,1997; Scott 1999; Golden et al., 2001; 2004). Regardless of the

ED

origin of the ALHA 84001 magnetite, it is certainly true that prokaryotic microorganisms can produce morphologically distinct magnetite that can be preserved in the rock record (e.g.,

PT

Kirschkvink et al., 1985; Bazylinski et al., 2007). As Mars surface conditions may have destroyed any organic remains of past Martian biota, biogenic magnetite may constitute one of

CE

the only remaining physical signatures of extinct Martian microorganisms (McKay et al., 2004). Magnetite and the closely-related mineral maghemite are also a significant component of the Martian regolith breccia meteorites (NWA 7034 and pairing group) and may be major carriers of

AC

Martian crustal magnetization signatures (Gattacceca et al., 2014). 1.3 Role of magnetite in geobiological processes and records: Magnetite is one of the reaction products of serpentinization, the reaction of ferromagnesian silicates with water to produce hydrogen, serpentine-group minerals, and magnetite. Olivine and Fe-Mg pyroxene are the most common mafic silicates affected by serpentinization, though calcic pyroxenes, amphibole, chlorite, talc, and ferromagnesian micas can also be affected (e.g., Wicks and Whittaker, 1977).

ACCEPTED MANUSCRIPT 6

Hydrogen generated by serpentinization reactions is an important source of energy for microbial life in Earth‘s subsurface (e.g., Mayhew et al., 2013), and serpentinization in the presence of carbon-bearing species (e.g., CO2, CO, HCO3) can result in the formation of methane and other organic compounds, especially aliphatic hydrocarbons (e.g., Berndt et al., 1996; McCollom 2016). Magnetite can also play a catalytic role in organic synthesis in such serpentinizing

CR IP T

systems (Berndt et al., 1996). Recent detections of methane in Martian meteorites (Blamey et al., 2015) and possible methane detections in the Martian atmosphere (e.g., Mumma et al., 2009; Formisano et al., 2004) could be related to past or ongoing serpentinization-type reactions involving magnetite.

Given the wide distribution of magnetite and its potential as an indicator of many important

AN US

geological processes, a systematic investigation of the spectral signatures of magnetite is warranted. Here we present a systematic investigation of the reflectance spectral properties of magnetite using natural and synthetic samples. Comparison of magnetite reflectance spectra with related oxide phases shows that magnetite is spectrally distinct. We investigate the effects of chemical substitutions, grain size variations, and mixtures with other phases in order to better

ED

M

constrain remote spectral searches for magnetite-bearing assemblages.

1.4 Crystal chemistry of magnetite: To understand what effect elemental substitutions and

PT

vacancies may have on magnetite reflectance spectra and how they relate to formation conditions, we first review the structure of magnetite. Magnetite is a spinel-group mineral, crystallizing in space group Fd-3m. Oxide spinels have the general formula of A8B16O32 (often

CE

expressed as AB2O4), where A and B refer to cations of either 2+ and 3+ valence (A2+B3+2O4, socalled 2-3 spinels) or of 4+ and 2+ valence (A4+B2+2O4, so-called 4-2 spinels). Spinels are further

AC

classified based on the distribution of cations on the A and B sites, into ―normal‖ and ―inverse‖ structures. The A and B cations are distributed over one-eighth of all tetrahedral and half of all octahedral sites in the spinel unit cell, leading to two different cation ordering schemes, termed ―normal‖ and ―inverse‖. In a normal spinel, the A cation occupies the tetrahedral site and the two B cations occupy the octahedral sites; in an inverse spinel, one of the B cations occupies the tetrahedral site and the remaining A and B cations occupy the octahedral sites. Chemical substitution causes variations in the degree of inversion in spinels so that many natural spinel

ACCEPTED MANUSCRIPT 7

compositions are intermediate between normal and inverse spinel structures (e.g., Wechsler et al., 1984; Waychunas, 1991).

The spinel structure allows for numerous cation substitutions, both simple and complex, leading to many compositional end-members. In the case of magnetite, trivalent cations including Cr3+,

CR IP T

Mn3+, and Al3+ can substitute for Fe3+, while numerous divalent cations including Mg2+, Mn2+, Ni2+, Co2+, and Zn2+ can substitute for Fe2+ (Waychunas, 1991; Deer et al., 2013). Magnetite forms solid solutions with numerous other spinel phases. A particularly important crystal chemical variation in magnetite is the coupled substitution, (Ti4+, Fe2+)↔ 2Fe3+ leading to solid solution with ulvöspinel (Ti4+Fe2+2O4). Ti-bearing magnetite intermediate between magnetite and

AN US

ulvöspinel is commonly referred to as titanomagnetite. Under oxidizing conditions, magnetite commonly contains some amount of excess Fe3+. The fully oxidized endmember is maghemite (Fe3+)[Fe3+5/3⧠1/3]O4, in which charge balance is maintained by the introduction of vacancies (denoted ⧠) on octahedral sites (e.g., Cornell and Schwertmann, 2006). Both ulvöspinel and excess Fe3+-bearing magnetite can undergo exsolution, and the resulting exsolution textures are

M

an indicator of the P-T-fO2 evolution of the host rock (e.g., Powell & Powell, 1977). The chemical substitutions involved in solid solutions cause perturbations in the electronic structure

PT

ED

of magnetite which can, in principle, lead to systematic differences in reflectance spectra.

1.5: Review of magnetite reflectance spectra: Magnetite exhibits low reflectance over the

CE

ultraviolet, visible, and near-infrared (here 0.2-2.5 μm) due to high absorption coefficients over this spectral range (Schlegel et al., 1979). Reflectance spectra of magnetite in the visible and near-infrared (~0.4-2.5 μm) have been reported in numerous studies, but have usually been

AC

restricted to one or a few samples (e.g., Hunt et al., 1971; Buchenau and Müller 1972; Hapke et al., 1978; Strens and Wood, 1979; Boppart et al., 1980; Morris et al., 1985; Wagner et al., 1987; Cloutis et al., 2008). Magnetite reflectance spectra in the 0.4-2.5 μm range are further characterized by two regions of locally lower reflectance, one centered near 0.55 μm and another, broader region centered near 1 μm. In general, previous investigators have interpreted magnetite reflectance spectra in terms of a mixture of intravalence electronic transitions between d orbital levels, and intervalence electronic transitions including metal-metal charge transfers

ACCEPTED MANUSCRIPT 8

(Fe3+-Fe2+) and oxygen-metal charge transfers (tetrahedral Fe3+-O and octahedral Fe2+-O and Fe3+-O) (e.g., Loeffler et al., 1974; Tossell et al., 1974; Vaughan and Tossell, 1978; Strens and Wood, 1979; Vaughan, 1985). These studies and their key findings are summarized below.

In the 0.2-0.4 µm range, a number of features of varying intensities and widths have been

CR IP T

identified in previous studies. These include a local minimum near 0.23 µm (near a broad local maximum in absorption coefficient; Pang et al., 1982; Strens and Wood, 1979; Hapke et al., 1978; Wagner et al., 1987), a broad, weak band near 0.29 µm and a broad weak feature near 0.35 µm (Cloutis et al., 2008).

AN US

Hunt et al. (1971) measured reflectance spectra of two magnetites, one titaniferous, for four different grain sizes. Their spectra all exhibited low overall reflectance (<15%) and were described as spectrally featureless, although a local minimum interpreted as a broad absorption band near 1 µm was noted in one of the sample‘s spectra and was assigned to ferrous iron. As expected for opaque minerals, reflectance increased with increasing grain size. Previous spectral

M

reflectance studies of opaque materials such as metal powders and iron meteorites have shown that reflectance increases with increasing particle size (Cloutis et al., 1990; Hoffmann et al.,

ED

1991; Cloutis et al., 2010; Cloutis et al., 2015). In a fine-grained powder, less incident light reaches the detector because of multiple scattering which causes so-called structural absorption. Structural absorption refers to the absorption of light due to multiple scattering caused by the

PT

microscale structure of a material, common examples include bird feathers (McCoy et al., 2018), butterfly scales (Vukusic et al., 2004), and carbon nanotube composites such as Vantablack®

CE

(e.g., Theocharous et al., 2006). In a coarse-grained powder (or compacted surface), there is far less multiple scattering and therefore less structural absorption. This results in more reflected

AC

light reaching the detector and overall higher reflectance. This differs from materials with lower extinction coefficients (e.g., most silicates), where smaller particle sizes lead to a greater influence of volume scattering and, consequently, trends of increased reflectance and band contrast with decreasing particle size (e.g., Cooper & Mustard 1999).

Buchenau and Müller (1972) measured reflectivity of single-crystal magnetite and found that the local minimum in reflectance near 0.5-0.6 µm corresponds to a local minimum in the real part of

ACCEPTED MANUSCRIPT 9

the dielectric constant and optical dielectric constant, while the 1 µm local reflectance minimum is at or close to a minimum in the imaginary part of the refractive index and in optical conductivity.

Strens and Wood (1979) interpreted a 1 µm local minimum as an absorption band in a

CR IP T

reflectance spectrum of magnetite and ascribed it to a d-d orbital transition in octahedrally coordinated Fe2+. Their spectrum also shows a weak local minimum near 0.35 µm that was not discussed. Strens and Wood (1979) mentioned an expected d-d orbital transition in tetrahedrally coordinated Fe2+ near 1.6 µm that was not present in their spectrum.

AN US

Electronic structure calculations by Vaughan and Tossell (1978) and Vaughan (1985) indicated that at least two discrete metal-oxygen charge transfers occur in the 0.3-0.4 µm region, and that crystal field transitions for octahedral Fe2+ are expected at longer wavelengths.

Schlegel et al. (1979) measured single crystal reflectance spectra of magnetite and reported a

M

local minimum near 0.3 µm, a broader, more complex, but still shallow local minimum near 0.50.6 µm, and a deep, broad local minimum band near 1 µm. The 1 µm feature was found to

ED

correlate with a local minimum in the real part of the dielectric function, and was interpreted as a ligand-field transition of Fe2+ in the A site. The 0.3 and 0.5-0.6 µm features were assigned to

PT

transitions in B site Fe3+.

Boppart et al. (1980) measured reflectivity spectra of single crystal magnetite at temperatures

CE

from 4.2 to 300 K and identified a feature at 2 µm due to crystal field transitions in Fe2+ in the B site, another feature at 1.2 µm ascribed to spin-forbidden transitions in Fe3+ in the B site, and

AC

near 0.5 µm due to spin-forbidden transitions in Fe3+ in the A site. They also identified a thin, ―metallic‖ Fe 3d band above the Verwey transition. However, their brief report provided few details on methodology and only a cursory interpretation of the data. Morris et al. (1985) studied the reflectance spectra of several Fe2+-deficient magnetite with different Fe3+/Fe2+ ratios. They found that the reflectance spectra are uniformly dark, with a broad local minimum interpreted as an absorption feature in the 0.5 µm region that becomes

ACCEPTED MANUSCRIPT 10

narrower with increasing Fe2+ content. At longer wavelengths, reflectance peaks near 0.73-0.81 µm (increasing wavelength with increasing Fe2+ content) were identified. Beyond this peak, reflectance declines and a well-resolved absorption feature in the 1 µm region becomes more evident with increasing Fe2+ content. They also found no spectral changes over the temperature

CR IP T

range -110 °C to +20 °C (163.15-293.15 K). 1.3 Electron delocalization and the “metal-like” reflectance of magnetite: The electrons associated with the two octahedral Fe cations in magnetite are delocalized at temperatures above the Verwey transition (e.g., Perversi et al., 2016; Burns, 1981; Verwey, 1939), which occurs at 120 K for stoichiometric bulk magnetite. The delocalized electrons associated with the

AN US

octahedral Fe cations lead to metal-like reflectance, in particular, high extinction (high values of the imaginary part of the refractive index, k) over a wide range of wavelengths (Burns, 1981). Superimposed on this metal-like ―red-sloped and featureless‖ reflectance are spectral features due to other processes, most notably O 2p to Fe 3d ligand-to-metal charge transfers, and strong ligand-field transitions due to Fe3+. Structural and spectroscopic studies suggest that a number of

M

factors could lead to changes in magnetite reflectance spectra in terms of the presence or absence of absorption bands, and their positions and intensities. These include: structural defects, grain

ED

size, cation substitutions, and cation deficiencies (e.g., Schlegel et al., 1979; Morris et al., 1985; Bosi et al., 2009).

PT

We undertook a systematic study of the spectral reflectance properties of a suite of magnetite samples, both synthetic and natural, to better elucidate spectral-compositional-structural

CE

relationships. This was motivated by the general lack of quantitative studies of the spectral reflectance properties of magnetites, and lingering uncertainties concerning the causes of some

AC

of their spectral features. We also examined other factors that could affect magnetite reflectance spectra, such as grain size.

2. Methods and samples:

ACCEPTED MANUSCRIPT 11

Samples and reflectance spectra incorporated into this study include samples acquired specifically for this work and characterized at the University of Winnipeg‘s Planetary Spectrophotometer Facility (PSF; http://psf.uwinnipeg.ca), as well as reflectance spectra from on-line

digital

archives:

Brown

University‘s

RELAB

facility

(http://

www.planetary.brown.edu/relab/) and the United States Geological Survey Spectroscopy Lab

CR IP T

(http://speclab.cr.usgs.gov/). The PSF samples were also characterized by X-ray fluorescence (XRF) and wet chemistry and powder X-ray diffraction (XRD). We also included related Feoxides to better understand the causes of spectral features in magnetite reflectance spectra. The suite of samples included in this study are described in Table 1.

AN US

2.1. Sample preparation

All magnetite samples were hand crushed and dry sieved to <45 μm particle size. For the grain size series, dry sieving was also used to produce 45-90 μm, 90-250 μm, 250-500 μm, and 5001000 μm fractions. Physically separable contaminants were removed through a combination of

M

hand picking under a binocular microscope and repeated passes with a hand magnet. Dry sieving consisted of vigorous brushing with a soft artist‘s brush and shaking of the powders in stainless

ED

steel sieves. Porosity effects were investigated by preparing aliquots of one <45 μm magnetite sample in four different ways, similar to the approach of Cloutis et al. (2018). Normal samples were prepared by pouring the powder into a sample cup and leveling the surface with a clean

PT

glass slide. ―Fluffy‖ samples were prepared by stirring and prodding the surface with just the tip of a clean needle to create a porous, low-density surface. Dense packed samples were firmly

CE

compressed by hand with the face of a clean glass slide. Pressed pellets were created by compressing 1 g of fine powder (<45 um) for 5 minutes at 30 tonnes reading (59205 t/m 2, or

AC

5900 bars) with polished steel anvils, creating a slab-like surface.

2.2. Reflectance spectroscopy

Reflectance spectra were acquired both at the RELAB spectrometer facility at Brown University (Pieters, 1983; http://www.brown.edu/relab/) and at the University of Winnipeg HOSERLab (http://psf.uwinnipeg.ca). RELAB spectra were acquired from 0.3 to 2.5 µm with 5 nm

ACCEPTED MANUSCRIPT 12

resolution at i=30° and e=0° relative to a calibrated halon disk. HOSERLab spectra were acquired with an ASD Field Spec Pro HR spectrometer (0.35-2.5 µm) with a 50 watt QTH light source and a viewing geometry of i=30º and e= 0º. Spectra were measured relative to a calibrated Spectralon® standard. Both RELAB and HOSERLab data were corrected for minor irregularities in the reflectance of halon and Spectralon in the 2-2.5 μm region. The HOSERLab data were also

CR IP T

corrected for minor offsets in the spectra at 1.0 and 1.83 μm, where detector changeovers occur. Between 500 and 1000 spectra of the dark current, standard, and sample were acquired to improve signal to noise.

Ultraviolet reflectance (0.2-0.4 µm) spectra were measured with an Ocean Optics (Dunedin, Fl)

AN US

Maya2000 PRO – a symmetric crossed Czerny-Turner miniaturized spectrometer equipped with an HC-1 grating and 50 μm entrance slit, giving an effective sampling interval between 0.48 nm at 0.20 µm and 0.46 nm at 0.40 µm, and a spectral resolution of ~1.85 nm throughout. This grating is variably blazed with varied groove densities of 300-600 lines/mm. The sensor is a 2D back-thinned linear CCD-array (Hamamatsu S10420) with 2048x64 active pixels, each 14 µm2.

M

Some of the remaining pixels are optically masked to provide continuous dark current estimation and subtraction. Sample illumination was provided by an Analytical Instrument Systems Inc.

ED

Mini-DTA light source with a 30 W deuterium lamp fed through a bifurcated fiber optic bundle consisting of six illumination fibers surrounding a central pick-up fiber feeding into the detector array. This assembly consisted of 400-μm diameter solarization-resistant fibers, with

PT

transmission efficiencies between 23 and 40% across the 0.20-0.40 µm range. A collimating lens is attached to the light house in order to focus more light into the fibers. This lens was manually

CE

focused to maximize signal. The fiber optic bundle was used in normal incidence and we used an integration time of 500 ms and average 700 individual spectra (or higher depending on required

AC

S/N). On Spectralon®, this produces a signal of about ~90% of the pixel-saturation level on that of the highest quantum efficiency – at about ~0.24 µm in wavelength - which represents ~95% of the capacitor full well depth (2×105 electrons). Measurements for each sample were made by first acquiring a dark current spectrum, a reference spectrum, followed by measurement of the sample. The average value of the optically masked pixels is subtracted from all collected spectra to provide a more consistent dark current correction. All three measurements were made using an identical viewing geometry, integration time, and number of averaged spectra. The reference and

ACCEPTED MANUSCRIPT 13

target were both placed at the same distance from the end of the fibers bundle (~3 mm). The fibers have a 25.4º field of view, and a ~3 mm working distance provides optimal signal. The most appropriate definition of the instrument configuration is a biconical arrangement with i and e centered on ~0º. The spectra were corrected to absolute reflectance by ratioing the sample spectra to a calibrated Spectralon® 99% diffuse reflectance standard (AS-01163-060) corrected

CR IP T

to a deep ultraviolet (DUV) mirror. A theoretical spectrum of the mirror was provided by Edmond Optics, which was adjusted only slightly based on standardless measurements. The Spectralon® standard was baked and treated to remove any impurities which could undergo photochemical reactions when exposed to the UV light source. The DUV mirror correction removes irregularities (weak absorption below 0.25 µm) in the Spectralon® standard (Cloutis et

AN US

al., 2008). The reflectance spectra included in this study are provided in the on-line supplement to this paper.

Positions of local maxima and minima reported in this study were measured by first subtracting a linear continuum, then fitting a 3rd-order polynomial to the continuum-removed feature, as

ED

2.3 X-ray diffraction.

M

described by McCraig et al. (2017).

X-ray diffraction analysis of magnetite powders was conducted in continuous scan mode from 5° to 80° 2θ using a Bruker D8 Advance powder X-Ray Diffractometer with a DaVinci

PT

automated powder diffractometer. A Bragg-Brentano goniometer with a theta-theta setup was equipped with a 2.5º incident Soller slit, 1.0 mm divergence slit, a 2.0 mm scatter slit, a 0.2 mm

CE

receiving slit, a curved secondary graphite monochromator, and a scintillation counter collecting at an increment of 0.02º and integration time of 10 seconds per step. The line focus Co X-ray

AC

tube was operated at 40 kV and 40 mA, using a take-off angle of 6º. Minerals were identified using Bruker EVA pattern analysis software and the Crystallography Open Database (COD; Grazulis et al., 2009). Phases identified by XRD are provided in Table 1.

2.3. Bulk chemical analyses

ACCEPTED MANUSCRIPT 14

Sample compositions for HOSERLab magnetite samples were determined using X-ray fluorescence with ferrous/ferric iron abundances measured by wet chemical methods as described by Mertzman (2000), and are provided in Table 1. Reflectance spectra for all the samples in this study are provided in the on-line supplement. Further sample information, analytical data, and complete description of the University of Winnipeg analytical facilities are

(HOSERLab) website at http://psf.uwinnipeg.ca/.

3. Results

CR IP T

available through the University of Winnipeg Planetary Spectrophotometer Facility

Magnetite reflectance spectra in the 0.35-2.5 μm region are characterized by low overall

AN US

reflectance and an overall increase in reflectance with increasing wavelength. Upon this are superimposed local regions of lower reflectance, one centered near ~0.55 μm and another near ~1 μm, and broad local maxima near 0.39 and 0.65 μm. Over the 0.2-0.4 μm region, reflectance decreases with increasing wavelength, and there are poorly-resolved features near ~0.23 μm and ~0.39 μm (Figure 1). Magnetite samples with variable chemistry show subtle differences in their

M

reflectance spectra. Magnetite is, however, spectrally distinct from most related oxide minerals investigated here with the possible exception of wüstite and titanomagnetite (as discussed

PT

4. Discussion

ED

below).

4.1 Reinterpretation of magnetite reflectance spectra – effects of electron delocalization and

CE

high absorption: Our interpretation of the spectral features observed in reflectance spectra of magnetite differs from most previous studies, but are consistent with interpretations that derive

AC

from optical constant measurements (Figure 1), see also Izawa et al. (2018) We identify local maxima at ~0.23 μm, ~0.39 μm, and ~0.65 μm as Fresnel peaks associated with the centers of strong absorptions (or groups of closely spaced absorptions). The electrons associated with the two octahedral Fe cations in magnetite are delocalized at temperatures above the Verwey transition (~120 K) (e.g., Perversi et al., 2016; Burns, 1981; Verwey, 1939). The delocalized electrons associated with the octahedral Fe cations lead to metal-like behavior (i.e., increasing reflectance with increasing wavelength), in particular, high absorption (high values of the

ACCEPTED MANUSCRIPT 15

imaginary part of the refractive index (k) over a wide range of wavelengths. Superimposed on this metal-like reflectance are localized regions of higher or lower reflectance due to other processes, most notably oxygen-metal charge transfers and ligand-field transitions. Ascertaining the nature of the transitions underlying the observed features reflectance spectra is complex, in large part due to lingering uncertainties regarding the electronic structure of magnetite (e.g.,

CR IP T

Antonov et al., 2001; Perversi et al., 2016). Several different band assignments have been made for diffuse reflectance spectra in the 0.2-2.5 μm region, but in general all have identified local minima in reflectance as absorption features (Vaughan and Tossell, 1978; Strens and Wood 1979; Burns 1981; Morris et al., 1985; Vaughan, 1985), though Morris et al., (1985) briefly

AN US

alluded to the possibility that some reflectivity maxima could be due to strong absorptions.

Consideration of the wavelength dependence of the complex refractive index over this range (Querry, 1985; Müller and Buchenau, 1975; Buchenau and Müller, 1972) supports our interpretation of the reflectance spectrum of magnetite. Figure 1 shows the spectrum of near endmember magnetite, MAG102, which contains very low levels (~1%) of hæmatite (as exsolution

M

lamellae), and best illustrates the spectral characteristics of pure magnetite. Also shown are the optical constants n and k for magnetite calculated from near-normal incidence specular

ED

reflectance by Querry (1985), and Fresnel reflectance calculated following Heavens (1991). Local maxima in reflectance near 0.23 μm, 0.39 μm, and 0.65 μm correspond with local maxima in n and k and are, therefore, best interpreted as Fresnel peaks associated with strong absorptions.

PT

Fresnel peaks occur near the centers of absorption bands where the extinction coefficient k is high (typically greater than ~0.1), causing ‗mirror-like‘ or coherent reflectance from the grain

CE

surfaces (e.g., Salisbury et al., 1987; Hapke, 2012). In regions of high extinction, the entry of light into the material is inhibited, photons that do enter are absorbed, and volume scattering

AC

becomes insignificant (e.g., Salisbury et al., 1987; Hapke, 2012). Considering the optical constants of magnetite, it is clear that a ―strong surface scattering‖ interpretation of the reflectance spectra is appropriate, and that the local reflectance maxima are best interpreted as locations of Fresnel peaks at wavelengths corresponding to electronic transitions in magnetite. In the 0.2-2.5 μm wavelength range, three major possible mechanisms of absorption have been identified by previous workers. The first mechanism is ligand-to-metal charge transfers involving the 2p levels of oxygen and the 3d orbitals of Fe. The second is ligand-field d-d orbital

ACCEPTED MANUSCRIPT 16

transitions. The d-d or crystal field transitions of Fe3+ are nominally spin-forbidden; however, these transitions can be greatly intensified when Fe3+ cations are magnetically coupled to adjacent Fe3+ and Fe2+ cations (e.g., Sherman and Waite, 1985). We would expect this to be the case in magnetite. Even above the Verwey transition, we should be able to interpret aspects of the electronic spectra of magnetite in terms of the multiplet (crystal field) states of discrete

CR IP T

Fe2+ and Fe3+ cations since electron delocalisation occurs via small polaron hopping (thermally activated Fe2+-Fe3+ charge transfer).. Included in ligand-field d-d transitions is the phenomenon of ―double excitation‖, wherein a single incident photon excites two strongly correlated electrons. The double excitation mechanism produces a spectral feature at approximately twice the energy of the corresponding ―single‖ excitation (Sherman and Waite; 1985). The third is

AN US

metal-metal charge transfer processes in Fe2+ and Fe3+ in neighboring edge-sharing octahedral sites.

We have assigned the Fresnel peaks in reflectance spectra of magnetite to groups of transitions, bearing in mind that multiple closely-spaced transitions are expected to be part of each feature envelope. The 0.23 μm feature is due to ligand-to-metal charge transfers between

M

the O 2p levels and Fe 3d levels. The 0.39 Fresnel peak is ascribed to ligand-field transitions of octahedral Fe (most likely 6A1 → 4E(4D) and 6A1 → 4T2(4D), possibly including double

ED

excitation contributions such as 6A1 + 6A1 → 4T2(4D) + 4T2(4D). The 0.65 μm Fresnel peak is ascribed to 6A1(6S) → 4T2(4G) ligand-field transitions. Notably absent are features ascribable to metal-metal (Fe2+-Fe3+) intervalence charge transfer processes, which are unresolved due to the

PT

effects of electron delocalization above the Verwey transition. In contrast to most previous investigations, we interpret the local minimum near 1 μm

CE

not as a d-d orbital ―crystal field‖ transition, but instead as a local minimum due to the Fresnel peak caused by Fe-O d-d orbital transitions at 0.65 μm superimposed on the broad, metal-like

AC

reflectance produced by delocalized electrons. The high extinction coefficient (imaginary part of refractive index) of magnetite means that very few photons are transmitted through the material, therefore surface scattering will dominate. Because crystal field absorptions such as the proposed d-d orbital transitions of octahedral iron are typically very weak, it is unlikely that such processes would produce detectable features when superimposed on the metal-like reflectance caused by delocalized electrons. However, a partial contribution from this mechanism cannot be ruled out. Similarly, we interpret the local minimum near 0.55 μm as a minimum between the Fresnel

ACCEPTED MANUSCRIPT 17

peaks at ~0.39 μm and ~0.65 μm, rather than an absorption feature. Regardless, the location and magnitude of these local transition minima may provide useful spectral parameters for magnetite identification and characterization along with the positions of Fresnel reflectance peaks. Speculatively, similar Fresnel peaks might occur in the spectra of other oxide minerals in spectral regions where strong ligand-to-metal or metal-metal charge transfers coexist with large

CR IP T

extinction coefficients, as in magnetite.

4.2 Grain size effects: Reflectance spectra of magnetite at different grain sizes are compared in Figure 2 The primary grain size effect is a small general increase in reflectance and increase in spectral contrast (band depth) with increasing grain size, similar to the results of Hunt et al.

AN US

(1971). This grain size trend is opposite that observed in materials with lower extinction coefficients (e.g., most silicates) and is similar to that observed for metal powders (Hoffmann et al., 1991; Cloutis et al., 1990; Cloutis et al., 2010; Cloutis et al., 2015). In a normalized reflectance plot, spectra of the different magnetite grain sizes are very similar, with no discernable differences in the location of local maxima (Fresnel peaks) or minima. This

M

illustrates the dominance of surface scattering over volume scattering at all particle sizes investigated here. Grain size effects on magnetite spectra are negligible in the ultraviolet

ED

reflectance spectra measured here, and are very minor in the VNIR. Dry sieving may leave some clinging fine particles. There is no reason to suspect, however, that clinging fines would have

PT

preferentially affected any of the particle size fractions measured here.

4.3. Spectrum-altering effects of magnetite in mixtures with other minerals: Magnetite is a

CE

plausible ―bluing agent‖ in the visible and near-infrared. A number of mineral mixtures

AC

involving magnetite have been spectrally characterized to assess its spectrum-altering effects.

4.3.1. Magnetite + CM2 carbonaceous chondrite

Mixtures of magnetite and phyllosilicates form a significant volume fraction of carbonaceous chondrites and are expected to be common on the surfaces of dark asteroids.Increased concentrations of magnetite may account for some of the spectral properties (e.g., blue spectral slopes) of B-type asteroids including Bennu (Clark et al., 2011). In general, powdered mixtures

ACCEPTED MANUSCRIPT 18

of magnetite with the Murchison CM2 carbonaceous chondrite (Figure 3) show lower overall reflectance, which decreases with increasing magnetite content. The spectral contrast of some of the bands is also reduced with the addition of magnetite.

A detailed analysis of the dataset shows that the addition of magnetite to the Murchison CM2

CR IP T

chondrite produces a progressively deeper local minimum near ~1 μm corresponding to the interband minimum in magnetite. In a normalized reflectance plot, a systematic increase in reflectance at wavelengths from ~1.9 to 2.5 μm and a systematic decrease in spectral slope below ~0.65 μm are observed (Figure 3). However, reflectance spectra of various mixtures of Murchison CM2 chondrite and magnetite demonstrate that even at a 50:50 mixing ratio, the

AN US

overall ―red‖ spectral slope due to CM2 chondrite persists (Figure 3). The red spectral slope of Murchison in the 0.35-0.65 μm range is mostly due to the long-wavelength wings of strong ultraviolet absorptions due to mixed-valence sp2- and- sp3-bonded carbon in the organic fraction of Murchison (e.g., Applin et al., 2018). The spectra of magnetite-CM2 chondrite mixtures demonstrate that magnetite is a plausible candidate for creating ―blue‖ spectral slopes (i.e.

M

decreasing reflectance with increasing wavelength) in the UV and visible (at wavelengths below ~0.65 μm wavelength) only in the absence of significant concentrations of macromolecular

ED

carbon. This result may have significant implications for the interpretation of blue spectral slopes for magnetite-rich surface regions (e.g. B-type asteroids). These slopes may suggest the presence of magnetite but the absence of carbonaceous material. In contrast, progressively greater

PT

concentrations of magnetite also create significant spectral bluing in the ~0.6-1.0 μm wavelength range, which carbonaceous material does not. However, the introduction of a broad local

CE

minimum by magnetite in the 1 µm region may complicate interpretations in this spectral region, particularly in analyses that use only a few channels of a hyperspectral measurement, or a few

AC

multispectral bands, to calculate a ―1 μm band depth‖ parameter. Variations in such a parameter that do not account for the effects of magnetite could significantly change interpretations of concentration and/or composition of olivine.

4.3.2. Mixtures of olivine and phyllosilicates with magnetite

ACCEPTED MANUSCRIPT 19

Olivine, phyllosilicates and magnetite can co-exist in CI, CM, CR, CO, and CV chondrites (Brearley and Jones, 1998). Reflectance spectra of mixtures involving olivine and various phyllosilicates, intimately mixed with magnetite were the subject of earlier studies (Cloutis et al., 2011a, b). For powdered olivine, the addition of magnetite reduced overall reflectance and olivine band depth, and flattened overall spectral slope (i.e., causes more spectral bluing) for 10

CR IP T

wt.% magnetite over the 0.4 to 2.5 µm interval. As expected, these effects were more pronounced for finer-grained magnetite (<45 vs. 45-90 µm) and with increasing magnetite abundance.

For mixtures of serpentine and magnetite, increased abundance of <45 µm magnetite led to lower overall reflectance, shallower serpentine absorption bands near 0.9 and 1.1 µm, and a bluer

AN US

overall spectral slope (Cloutis et al., 2011a). Bluer spectral slopes in magnetite-bearing mixtures were also confirmed by comparison of reflectance spectra of a magnetite-free and magnetitebearing serpentine (Cloutis et al., 2011b). Here, we have examined additional mixtures to better constrain the effects of magnetite. For mixtures of serpentine and magnetite (both <45 µm grain size), reflectance decreases, spectral slope becomes bluer, and serpentine absorption bands

M

become shallower, with increasing magnetite (Figure 4A, B).

ED

4.3.3 “Nanophase magnetite” or maghemite? A cautionary tale: We have examined a commercially available nanophase magnetite (HOSERLab samples MAG 200), nominal particle size 20 nm. The spectrum of this material is strikingly different from magnetite and instead is

PT

virtually identical to maghemite (cf. Fig 4C-J and 5B). While both magnetite and maghemite are dark, magnetite has a local reflectance minimum near 1 µm and a flattish spectral slope from

CE

~0.35 to ~0.7 µm. Maghemite is strongly red sloped below ~0.75 µm, has a local reflectance

AC

maximum near 0.95 µm, and a local reflectance minimum near 1.5 µm.

We have conducted a number of mixing experiments with this nanophase powder and various phyllosilicates. These mixture spectra may still be of use in interpreting some remotely-sensed mineral assemblages involving maghemite, or in situation where nanophase magnetite may have been oxidized. Recent results reported by Sklute et al. (2018) have emphasized the need for precautions to prevent the rapid oxidation of nanophase magnetite in air. When the nanophase maghemite powder is mixed with serpentine, the decreases in reflectance and reduction in

ACCEPTED MANUSCRIPT 20

serpentine band depths are more pronounced than with the <45μm magnetite powder, and the spectra assume an overall shape more akin to the maghemite end member (Figures 4C, D). Comparison of two mixtures with identical abundances of magnetite versus nanophase maghemite (Figures 4E, F) shows how the nanophase maghemite more effectively lowers absolute reflectance, suppresses serpentine absorption bands (e.g., band near 0.75 µm), and

CR IP T

causes a more concave spectral shape.

The spectrum-altering effects of nanophase maghemite on a saponite (the most common phyllosilicate in CI1 chondrites; Brearley and Jones, 1998) are similar to those seen for serpentine + maghemite. These include reduction in overall reflectance, flattening of spectral

AN US

slope, reduction in saponite absorption band depths (near 1.9 and 2.3 µm), and a more maghemite-like reflectance spectrum, including a shift of the local maximum near 0.85 µm toward a shorter (pure maghemite) wavelength (Figures 4G, H).

Finally, cronstedtite (Fe2+2Fe3+(Si,Fe3+)O5(OH)4), a common phyllosilicate in CM2 chondrites

M

(Brearley and Jones, 1998) was used as the end member mixed with nanophase (<20 nm) maghemite. Cronstedite is darker than the other phyllosilicates that were examined, consequently

ED

the spectral changes with increasing nanophase maghemite are less dramatic (Figures 4I, J). Reflectance does decrease and the spectra become flatter. Therefore, recognizing the presence of

PT

maghemite in spectra of mixtures with cronstedtite is not straightforward.

4.3.4 Effects of grinding on magnetite-serpentine mixtures, and nanophase maghemite-

CE

serpentine mixtures: Powdered laboratory samples in reflectance spectroscopy are commonly prepared by grinding. Grinding may physically mimic some regolith physical processing. The

AC

spectral effects of repeated grinding of the same sample are rarely documented. We have investigated the effects of grinding on mixtures of serpentine (HOSERLab sample SRP111) with both magnetite (MAG102) and nanophase maghemite (MAG200), the results of which are shown in Figure 5. Mixtures of magnetite and serpentine show an overall decrease in reflectance and band contrast with progressive grinding, though the shape (both of the overall spectrum and of the individual spectral features) remains similar (Figures 5A, B). In contrast, nanophase maghemite shows a marked change with any grinding, including a much more pronounced UV

ACCEPTED MANUSCRIPT 21

drop-off and an enhanced local minimum centered near 1.5 μm (Figures 5C-F). The local maximum, however, does not change wavelength position. In all cases, repeated grinding of the sample increases the influence of the darker material. The greater influence of the nanophase material with grinding may be due to physical changes caused by the grinding process such as enhanced coating of the serpentine grains by more effectively dispersed and increasingly finer-

CR IP T

grained dispersed nanoparticles. An important point here is that the physical state of the mixture can be sufficiently changed by a physical process to an extent that band parameters such as spectral slope, absolute and normalized reflectance values, and band depths can all be affected. This emphasizes the need for caution in the interpretation of band depths and related parameters in terms of relative or absolute abundances, at least in the case of phyllosilicates and strongly-

AN US

scattering opaques. This concern is especially true in VNIR data, where the relationships between composition and spectral features are not generally linear, and where the empirical relationships used to derive composition commonly come from only a few studies with limited sample compositions.

M

4.4 Comparison of 0.35-2.5 μm magnetite reflectance spectra with selected FeO-Fe2O3-TiO2 minerals: It is illustrative to contrast the reflectance of magnetite with several other related oxide

ED

minerals. Magnetite commonly incorporates Ti , , and is also commonly associated with the FeTi oxide minerals ulvöspinel and ilmenite. Magnetite reflectance spectra in the 0.35-2.5 μm range are very similar to those of available titanomagnetite (Figure 6A). In contrast, ilmenite and

PT

ulvöspinel are spectrally distinct from magnetite and are characterized by local maxima near 1 μm and 1.5 μm, respectively, (Figure 6). Magnetite is also commonly associated with maghemite

CE

and hæmatite. Both hæmatite and maghemite have red spectral slopes in the visible spectral region compared to magnetite, due to very strong UV absorptions (Rossmann 1996; Sherman

AC

and Waite 1985). Hæmatite is further characterized by very high overall reflectance beyond ~1 μm, and distinct features due to crystal-field transitions of Fe3+ near 0.85 and 0.44 μm (Sherman and Waite, 1985). The reflectance spectrum of wüstite (Fe1-xO, 0.04
ACCEPTED MANUSCRIPT 22

ligand-to-metal charge transfers and Fe-related d-d orbital transitions. In detail, the structure of wüstite is

[Fe2+1−3x Fe3+2x−t ⧠x+t]IVFe3+ tO, where 0.04 < x < 0.12 and 2.0 < (x + t)/t < 4.5

VI

(Hazen and Jeanloz, 1984). The defect structures and vacancies in wüstite lead to a magnetitelike mixed valence system with significant electron delocalization. Magnetite reflectance spectra are distinct from available RELAB spectra of pseudobrookite ((Fe3+,Fe2+)2(Ti,Fe2+)O5) and

CR IP T

armalcolite ((Mg,Fe2+)Ti2O5), as shown in Figure 6C. A lack of precise knowledge of the mineral and chemical impurities present in the samples measured to produce the available reflectance spectra of these two rare minerals precludes a more detailed discussion of their spectral properties. The comparison of magnetite reflectance spectra with those of related minerals further demonstrates the dominance of the delocalized electrons on the reflectance

AN US

properties of magnetite. Overall, magnetite is spectrally distinct from most related oxide minerals with the exception of those that are similarly affected by electron delocalization (wüstite, titanomagnetite).

4.5 Effects of chemical substitutions: Magnetite undergoes many possible chemical substitutions

M

as detailed in section 1.1. Figure 7 compares reflectance spectra of chemically-pure magnetite (MAG102) with those of magnetite containing significant concentrations of TiO2, Al2O3, MgO

ED

and MnO as shown in Table 1. These samples were chosen because they contain very low levels of other phases (MAG105 and MAG109 contain traces of hæmatite, the others have no mineral impurities detectable by XRD; Table 1). The trace hæmatite may contribute to the spectral slope

PT

near 200 nm due to an Fe3+-O transition (e.g., Cloutis et al., 2008).Despite containing a significant solid solution component of ulvöspinel, the titaniferous magnetites investigated here

CE

do not show spectral features similar to those of ulvöspinel, including the distinctive inter-band reflectance maximum near 1.47 µm. None of the Ti-bearing magnetite spectra here show an

AC

identifiable Fe2+-Ti4+ charge-transfer absorption feature (Figures 7A, B). This may be ascribable to the fact that above the Verwey transition, absorption by delocalized electrons dominates, unless cation substitution is sufficient to greatly lower the number of nearest-neighbor iron pairs (Bahgat et al., 1980). There is, however, a subtle shift in the position of the local Fresnel peak near 0.65 μm in chemically impure magnetite samples (Figure 7C). Our spectral data set, while limited, suggests that one of the biggest spectral variations seen, the shifting of the 0.65 μm Fresnel peak to longer wavelengths (which induces a corresponding shift in the 1 µm region

ACCEPTED MANUSCRIPT 23

local minimum to longer wavelengths) is correlated with the most Fe-deficient samples (MAG109, MAG111). 4.7 Effects of packing: Figure 8 compares spectra of <45 μm magnetite powders with various particle packing states. Many asteroid surfaces are likely covered by loosely-packed ―fluffy‖

CR IP T

regolith. At present, there remains a great deal of uncertainty regarding how well ―asteroid-like‖ regolith particle packing can be reproduced on Earth, where microgravity, electrostatic forces, and high-vacuum effects are difficult to replicate. In general, increasing surface porosity leads to decreased overall reflectance, and compaction increases both overall reflectance and the contrast of spectral features and increased blue slope over visible wavelengths (Figure 8A). In normalized

AN US

reflectance, the spectral slope effect of compaction is even more pronounced, with the pressed pellet (effectively a slab surface) being extremely blue-sloped in the visible (Figure 8B).

5. Summary and Conclusions

We have conducted a survey of the spectral reflectance properties of magnetite for remote

Reflectance spectra of magnetite illustrate an important point, familiar in studies at longer

ED



M

sensing, covering the 0.2-2.5 µm in wavelength. Our results lead to the following conclusions:

mid-infrared wavelengths: Where surface scattering dominates, bands tend to be expressed as local maxima in reflectance. Conversely, where volume scattering

PT

dominates, features are expressed as local minima in reflectance. In magnetite, surface scattering dominates at ultraviolet, visible, and near-infrared wavelengths primarily due 

CE

to the delocalized nature of the electrons associated with octahedral Fe. Magnetite reflectance spectra are characterized by Fresnel peaks near 0.27 µm due to

AC

ligand-to-metal charge transfer from O 2p to Fe 3d bands, and at 0.39 and 0.65 µm due to Fe field-d orbital transitions. These features are superimposed on a ―metal-like‖ (i.e., red sloped and featureless) reflectance spectrum due to scattering from the delocalized electrons associated with the magnetite octahedral Fe cations. The local minimum in reflectance near 0.55 µm is due to lower reflectance between the 0.39 and 0.65 μm Fresnel peaks. The local minimum near 1 µm is due the combined effects of the 0.65 μm

ACCEPTED MANUSCRIPT 24

Fresnel peak and the general rise in reflectance with increasing wavelength into the infrared. 

In a limited set of mineralogically-pure, variably chemically substituted magnetites, there is a systematic shift in the position of the ~0.65 μm Fresnel peak to longer wavelengths with increasing chemical impurity. This also shifts the position of the ~1 μm local

CR IP T

reflectance minimum to longer wavelengths, and either feature could in principle be used to spectrally detect and characterize magnetite composition, at least in terms of ―pure‖ versus ―chemically-substituted‖ magnetite samples. 

Magnetite-rich assemblages that also contain complex organic matter such as those found in many carbonaceous chondrite meteorites will show a progressively deeper local

AN US

minimum near 1 µm due to magnetite, but will show a visible and UV spectral slope dominated by the organic carbon UV absorptions. 

Magnetite is spectrally distinct from many commonly associated oxide minerals including ilmenite, maghemite, hæmatite, and ulvöspinel. Magnetite is also spectrally distinct from the rarer minerals pseudobrookite and armalcolite. Magnetite is spectrally

ED

M

similar to wüstite, due to similar electron delocalization effects.

Acknowledgements: The University of Winnipeg's Planetary Spectrophotometer Facility was established with funding from the Canada Foundation for Innovation, the Manitoba Research

PT

Innovations Fund, the Natural Sciences and Engineering Research Council of Canada (NSERC), the Canadian Space Agency (CSA), and the University of Winnipeg (UW), whose support is

CE

gratefully acknowledged. This study was supported by research grants from NSERC, CSA, and UW. We also wish to thank Dr. Takairo Hiroi and Dr. Carlé Pieters of the NASA-supported

AC

RELAB spectrometer facility at Brown University for acquiring reflectance spectra of a number of the samples used in this study. This manuscript was greatly improved by the suggestions and comments of two anonymous reviewers.

References Antonov V.N., Harmon B.N., Antropov, V.P. (2001), Electronic structure and magneto-optical Kerr effect of Fe3O4 and Mg2+-or Al3+-substituted Fe3O4, Physical Review B, Vol. 64, 134410

ACCEPTED MANUSCRIPT 25

Applin D. M., Izawa M. R. M., Cloutis E. A., Gillis-Davis J., Pitman K. M., Roush T. L., Hendrix A. R., and Lucey P. G. (2018) Ultraviolet spectral reflectance of carbonaceous materials, Icarus 307, 40-82 Bahgat A. A., Eissa N. A., Fayek M. K. and Hamalaway A. A. (1980) The influence of substitution of iron ions on the electron hopping in magnetite Journal of Physics C: Solid State Physics, Volume 13, Number 13

CR IP T

Barth, M., C. Vollmer, A. Bischoff, and M. Horstmann (2013) TEM observations of aqueously altered phases in Maribo (CM2) and Murchison (CM2) Matereoritical Society Meeting, 76th, abstract #5090. Bazylinski, D.A., R.B. Frankel, and K.O. Konhauser (2007) Modes of biomineralization of magnetite by microbes. Geomicrobio. J., 24, 465-475.

AN US

Berndt M.E., Allen D.E., Seyfried W.E. (1996) Reduction of CO2 during serpentinization of olivine at 300°C and 500 bar. Geology 24:351–354. Blamey, N.J.F., Parnell, J., McMahon, S., Mark, D.F., Tomkinson, T., Lee, M., Shivak, J., Izawa, M.R.M., Banerjee, N.R., and Flemming, R.L. (2015) Evidence for methane in Martian meteorites. Nature Communications, 6, doi:10.1038/ncomms8399.

M

Boppart, H., A. Schlegel, and P. Wachter (1980) Optical properties of magnetite Fe3O4. Phil. Mag. B, 42, 431-432. Bosi, F., U. Hålenius, and H. Skogby (2009) Crystal chemistry of the magnetite-ulvöspinel series. Amer. Mineral., 94, 181-189.

ED

Brearley, A.J., and R.H. Jones (1998) Chondritic meteorites. In: Planetary Materials (Papike J.,J., Ed.). Reviews in Mineralogy, 36, Mineralogical Society of America, Washington, D.C., Ch. 3.

PT

Buchenau, U., and I. Müller (1972) Optical properties of magnetite. Solid State Comm., 11, 1291-1293.

CE

Burns, R.G. (1981) Intervalence transitions in mixed-valence minerals of iron and titanium. Ann. Rev. Earth Planet. Sci., 9, 345-383.

AC

Changela, H.G., J.C. Bridges, and S.J. Gurman (2012) Extended X-ray absorption fine structure (EXAFS) in Stardust tracks: Constraining the origin of ferric iron-bearing minerals. Geochim. Cosmochim. Acta, 98, 282-294. Christoffersen, R, and P.R. Buseck (1986) Mineralogy of interplanetary dust particles from the ―olivine‖ infrared class. Earth Planet. Sci. Lett., 78, 53-66. Clark B. E., Binzel R. P., Howell E. S., Cloutis E. A., Ockert-Bell M., Christensen P., Barucci M. A., DeMeo F., Lauretta D. S., Connolly H. Jr., Soderberg A., Hergenrother C., Lim L., Emery J., and Mueller M. (2011) Asteroid (101955) 1999 RQ36: Spectroscopy from 0.4 to 2.4 mu m and meteorite analogs. Icarus 216:462–475.

ACCEPTED MANUSCRIPT 26

Cloutis, E. A., Pietrasz V. B., Kiddell C., Izawa M. R. M., Vernazza P., Burbine T. H., DeMeo F., Tait K. T., Bell III J. F., Mann P., Applin D. M., and Reddy V.. 2018. Spectral reflectance ―deconstruction‖ of the Murchison CM2 carbonaceous chondrite and implications for spectroscopic investigations of dark asteroids. Icarus 305: 203–224. Cloutis, E.A., and J.F. Bell III (2004) Mafic silicate mapping on Mars: effects of palagonitic material, multiple mafic silicates, and spectral resolution. Icarus, 172, 233-254.

CR IP T

Cloutis, E.A., and M.J. Gaffey (1991) Pyroxene spectroscopy revisited: Spectral-compositional correlations and relationship to geothermometry. J. Geophys. Res., 96, 22,809-22,826. Cloutis, E.A., M.J. Gaffey, and T.F. Moslow (1995) Characterization of minerals in oil sands by reflectance spectroscopy. Fueld, 74, 874-879.

AN US

Cloutis, E.A., K.A. McCormack, J.F. Bell III, A.R. Hendrix, D.T. Bailey, M.A. Craig, S.A. Mertzman, M.S. Robinson, and M.A. Riner (2008) Ultraviolet spectral reflectance properties of common planetary minerals. Icarus, 197, 321-347. Cloutis E.A., Hardersen P.S., Bish D., Bailey D.T., Gaffey M.A., Craig M.A. (2010) Reflectance spectra of iron meteorites: Implications for spectral identification of their parent bodies. Meteoritics and Planetary Science 25 (2) 304-332

M

Cloutis E.A., Sanchez J.A., Reddy V., Gaffey M.J., Binzel R.P., Burbine T.H., Hardersen P.S., Hiroi T., Lucey P.G., Sunshine J.M., Tait K.T. (2015) Olivine–metal mixtures: Spectral reflectance properties and application to asteroid reflectance spectra, Icarus 252 39-82

ED

Cloutis E. A., Gaffey M. J., Smith D. G. W., and Lambert R. St. J. 1990. Reflectance spectra of ―featureless‖ materials and the surface mineralogies of M‐ and E‐ class asteroids. Journal of Geophysical Research 95:281–293.

PT

Cloutis, E.A., Hiroi, T., Gaffey, M.J., Alexander, C.M.O‘D., and Mann, P. (2011a) Spectral reflectance properties of carbonaceous chondrites: 1. CI chondrites. Icarus, 212, 18-209. Cloutis, E.A., Hudon, P., Hiroi, T., Gaffey, M.J., and Mann, P. (2011b) Spectral reflectance properties of carbonaceous chondrites: 2. CM chondrites. Icarus, 216, 309-346.

CE

Cooper C.D., Mustard J.F., (1999) Effects of Very Fine Particle Size on Reflectance Spectra of Smectite and Palagonitic Soil. Icarus 142 (2) 557-570

AC

Cornell, R. M. and Schwertmann, Udo (2006) The Iron Oxides: Structure, Properties, Reactions, Occurrences and Uses, 2nd Edition. Wiley-VCH. pp. 703. Daulton, T.L., O. Pravdivtseva, A. Meshik, C.M. Hohenberg, and A.N. Krot (2013) Microstructure and formation of Orgueil magnetites studied by transmission electron microscopy. Meteoritical Society Meeting, 76th, abstract #5275. Deer, W.A., R.A. Howie, and J. Zussman (2013) An Introduction to the Rock-Forming Minerals, 3rd Edition. Mineralogical Society, London, UK.

ACCEPTED MANUSCRIPT 27

Formisano V., S. Atreya, T. Encrenaz, N. Ignatiev N., M. Giuranna (2004) Detection of Methane in the Atmosphere of Mars Science Vol. 306, Issue 5702, pp. 1758-1761. Gattacceca, J., Rochette, P., Scorzelli, R. B., Munayco, P., Agee, C., Quesnel, Y., Cournède C., Geissman, J. (2014). Martian meteorites and Martian magnetic anomalies: A new perspective from NWA 7034. Geophysical Research Letters, 41(14), 4859-4864.

CR IP T

Gibson Jr., E.K., D.S. McKay, K.L. Thomas-Keprta, S.J. Wentworth, F. E. Westall, A. Steele, C.S. Romanek, M.S. Bell, and J. Toporski (2001) Life on Mars: evaluation of the evidence within Martian meteorites ALH84001, Nakhla, and Shergotty. Precambrian Res., 106, 15-34. Golden, D.C., D.W. Ming, C.S. Schwandt, H.V. Lauer Jr., R.A. Socki, R.V. Morris, G.E. Lofgren, and G.A. Mckay (2001) A simple inorganic process of formation of carbonates, magnetite, and sulfides in martian meteorite ALH84001. Amer. Mineral., 86, 370.

AN US

Golden, D.C., D.W. Ming, R.V. Morris, A.J. Brearley, H.V. Lauer Jr., A.H. Treiman, M.E. Zolensky, C.S. Schwandt, G.E. Lofgren, and G.A. Mckay (2004) Evidence for exclusively inorganic formation of magnetite in Martian meteorite ALH84001. American Mineralogist, 89, 681-695. Grazulis S., Chateigner D., Downs R. T., Yokochi A. F. T., Quirós M., Lutterotti L., Manakova E., Butkus J., Moeck P. and Le Bail A. 2009. Crystallography Open Database - an open-access collection of crystal structures. Journal of Applied Crystallography 42: 726-729.

M

Hapke, B., 2012. Theory of Reflectance and Emittance Spectroscopy. Cambridge Univ. Press, Cambridge, UK.

ED

Hapke, B.W., W.D. Partlow, J.K. Wagner, and A.J. Cohen (1978) Reflectance measurements of lunar materials in the vacuum ultraviolet. Proc. Lunar Planet. Sci. Conf., 9th, 2935-2947.

PT

Hoffman H., Berger M., and Neukum G. (1991). Ni and Fe‐ powders: Grain size effects and compositional influence on the spectral reflectance (abstract). 22nd Lunar and Planetary Science Conference. pp. 279–280. Abstract #1288

CE

Hazen R. M. and Jeanloz R (1984) Wüstite (Fe1‐xO): A review of its defect structure and physical properties. Reviews of Geophysics Vol. 22(1) 37-46 Heavens, O. S. (1991), Optical Properties of Thin Solid Films, 261 pp., Dover, Mineola, N. Y.

AC

Herd C.D.K, Papike J.J., Brearley A.J., (2001), Oxygen fugacity of martian basalts from electron microprobe oxygen and TEM-EELS analyses of Fe-Ti oxides Am. Mineral., 86, pp. 1015–1024 Herzog, G.F., E. Anders, E.C. Alexander Jr., P.K. Davis, and R.S. Lewis (1973) Iodine129/xenon-129 age of magnetite from the Orgueil meteorite. Science (180), 489-491. Hicks, L.J., J.L. MacArthur, J.C. Bridges, M.C. Price, J.E. Wikcham-Eade, M.J. Burchell, G.M. Hansford, A.L. Butterworth, S.J. Gurman, and S.H. Baker (2017) Magnetite in comet Wild 2: Evidence for parent body aqueous alteration. Meteorit. Planet. Sci., 52(10), 2075-2096.

ACCEPTED MANUSCRIPT 28

Hua, X., and P.R. Buseck (1998) Unusual forms of magnetite in the Orgueil carbonaceous chondrite. Meteorit. Planet. Sci., 33, A215-A220. Hunt, G.R., J.W. Salisbury, and C.J. Lenhoff (1971) Visible and near-infrared spectra of minerals and rocks: III. Oxides and hydroxides. Mod. Geol., 2, 195-205.

CR IP T

Izawa, M. R. M., Flemming, R. L., King, P. L., Peterson, R. C., McCausland, P. J. A., 2010b. Mineralogical and spectroscopic investigation of the Tagish Lake carbonaceous chondrite by Xray diffraction and infrared reflectance spectroscopy. Meteoritics & Planetary Science. 45, 675698. Izawa, M. R. M., Flemming, R. L., McCausland, P. J. A., Southam, G., Moser, D. E., Barker, I. R. (2010a) Multi-technique investigation reveals new mineral, chemical, and textural heterogeneity in the Tagish Lake C2 chondrite. Planetary and Space Science, 58(10), 1347-1364

AN US

Izawa, M. R. M., Applin, D. M., Cloutis, E. A., (2018) The ―Infrared-Ultraviolet Connection‖ in reflectance spectroscopy 49th Lunar and Planetary Science Conference, Houston, TX, abstract #2930 Kirschvink, J.L., D.S. Jones, and B.J. MacFadden (Eds.) (1985) Magnetite Biomineralization and Magnetoreception in Organisms: A new Biomagnetism. Plenum, NY, 682 pp.

M

Krot, A.N., P. Doyle, and K. Nagashima (2013) Secondary fayalite, hedenbergite, and magnetite in the CO3.0-3.1 carbonaceous chondrites Y-81020, EET 90043, and MAC 88107. Lunar Planet. Sci. Conf., 44, abstract #1754.

ED

Lauretta, D.S. + 47 co-authors (2017) OSIRIS-REx: Sample return from asteroid (101955) Bennu. Space Science Reviews, 212, 925-984.

PT

Le Corre, L., Sanchez, J.A., Reddy, V., Takir, D., Cloutis, E.A., Thirouin, A., Becker, K.J., Li, J.-Y., Segita, S., and Tatsumi, E. (2018) Ground-based characterization of Hayabusa2 mission target asteroid 162173 Ryugu: constraining mineralogical composition in preparation for spacecraft operations. Monthly Notices of the Royal Astronomical Society, 475, 614-623.

CE

Lewis, R.S., and E. Anders (1975) Condensation time of the solar nebula from extinct 129I in primitive meteorites. Proc. Natl. Acad. Sci. USA, 72, 268-273.

AC

Loeffler, B.M., Burns, R.G., Tossell, J.A., Vaughan, D.J., and Johnson, K.H. (1974) Charge transfer in lunar materials: Interpretation of ultraviolet-visible spectral properties of the Moon. Proceedings of the Fifth Lunar Conference. Supplement 5, Geochimica et Cosmochimica Acta, 3, 3007-3016. Mao, K. R. P. M., Huggins, F. E., Mahajan, V., Huffman, G. P., Rao, V. U. S. (1994) The role of magnetite in Fischer-Tropsch synthesis Hyperfine Interactions, 93(1), 1745–1749 Mayhew, L.E., E.T. Ellison, T.M. McCollom, T.P. Trainor, and A.S. Templeton (2013) Hydrogen generation from low-temperature water-rock reactions. Nature Geosci., 6, 478-484. McCollom, T.M. (2016) Abiotic methane formation during experimental serpentinization of olivine. PNAS (49) 13965-13970.

ACCEPTED MANUSCRIPT 29

McCoy D.E., Feo T., Harvey T.A., Prum R.O. (2018) Structural absorption by barbule microstructures of super black bird of paradise feathers, Nature Communications 9(1) McCraig M.A., Osinski G.R., Flemming R.L., Cloutis E.A., Izawa M.R.M., Reddy V., FieberBeyer S., Pompilio L., van der Meer F., Berger J.A., Bramble M., Applin D.M. (2017) Fitting the curve in Excel: Systematic curve fitting of laboratory and remotely sensed planetary spectra, Computers and Geosciences Volume 100, March 2017, Pages 103–114

CR IP T

McKay, G., D. Lindstrom, S.-R. Yang, and J. Wagstaff (1988) Petrology of unique achondrite Lewis Cliff 86010. Lunar Planet. Sci. Conf. 20, 762-763 (abstract). McKay, D.S., E.K. Gibson Jr., K.L. Thomas-Keprta, H. Vali, C.S. Romanek, S.J. Clemett, X.D. F. Chiller, C.R. Maechling, and R.N. Zare (1996) Search for past life on Mars: Possible relic biogenic activity on Martian meteorite ALH84001. Science, 273, 924-930.

AN US

McKay, C.P., E.I. Friedmann, R.B. Frankel, and D.A. Bazylinski (2004) Magnetotactic bacteria on Earth and on Mars. Astrobiology, 3, 2630270. Mertzman, S.A., 2000. K-Ar results for from the southern Oregon-northern California cascade range. Oregon Geol. 62, 99-122.

M

Morris, R.V., Lauer Jr., H.V., Lawson, C.A., Gibson Jr., E.K., Nace, G.A., Stewart, C., 1985. Spectral and other physicochemical properties of submicron powders of hematite (α-Fe2O3), maghemite (γ-Fe2O3), magnetite (Fe3O4), goethite (α-FeOOH), and lepidocrocite (γ-FeOOH). J. Geophys. Res. 90, 3126-3144.

ED

Müller, I., and Buchenau, U. (1975) Measurement of the magneto-optical properties of magnetite in the near infrared. Physica 80B, 69-74.

PT

Mumma M. J., G. L. Villanueva, R. E. Novak, T. Hewagama, B. P. Bonev, M. A. DiSanti, A. M. Mandell, M. D. Smith (2009) Strong Release of Methane on Mars in Northern Summer 2003 Science Vol. 323, Issue 5917, pp. 1041-1045 Pang, K.D. S.F.S. Chun, J.M. Ajello, Z. Nansheng, and L. Minji (1982) Organic and inorganic interpretations of the martian UV-IR reflectance spectrum. Nature, 295, 43-46.

CE

Perversi, G., J. Cumby, E. Pachoud, J. P. Wright and J. P. Attfield (2016) The Verwey structure of a natural magnetite Chemical Communications 52, 4864-4867

AC

Pieters, C.M. (1983) Strength of mineral absorption features in the transmitted components of near-infrared light: First results from RELAB. J. Geophys. Res., 88, 7534-9544. Powell, R., Powell, M., (1977). Geothermometry and oxygen barometry using coexisting irontitanium oxides: a reappraisal. Mineralogical Magazine 41 (318), 257-263. Querry, M. R., (1985) Optical constants, Contractor Report CRDC-CR-85034; Defense Technical Information Center Accession Number: ADA158623 http://www.dtic.mil/docs/citations/ADA158623

ACCEPTED MANUSCRIPT 30

Rochette, P., Jackson, M., Aubourg, C. (1992) Rock magnetism and the interpretation of anisotropy of magnetic susceptibility, Reviews of Geophysics 30(3) 209-226 Rochette, P., J. Gattacceca, M. Bourot-Denise, G. Consolmagno, L. Folco, T. Kohout, L. Pesonen, and L. Sagnotti (2009) Magnetic classification of stony meteories: 3. Achondrites. Meteorit. Planet. Sci., 44, 405-427.

CR IP T

Rossman G. R. (1996) Why hematite is red: Correlation of optical absorption intensities and magnetic moments of Fe3+ minerals. Geochemical Society Special Publication No. 5, Editors M. D. Dyar, C. McCammon and M. W. Schaefer pages 23-27. Salisbury, J. W., Walter, L. S., Vergo, N. (1987) Mid-infrared (2.1-25 μm) spectra of minerals: first edition. US Geological Survey Open-File report 87-263

AN US

Schäfer, M., Schäfer, T. , Izawa, M. R.M., Cloutis, E. A., Schröder, S. E., Roatsch, T. , Preusker, F. , Stephan, K. , Matz, K. , Raymond, C. A. and Russell, C. T. (2018), Ceres‘ spectral link to carbonaceous chondrites—Analysis of the dark background materials. Meteoritics & Planetary Science . doi:10.1111/maps.13079 (Early View) Schlegel, A., Alvarado, S.F., and Wachter, P. (1979) Optical properties of magnetite (Fe3O4). J. Phys. C Solid State Phys., 12, 1157-1164. Schlegel, A., S.F. Alvarado, and P. Wachter (1979) Optical properties of magnetite (Fe3O4). J. Phys. C: Solid State Phys., 12, 1157-1164.

M

Scott, E.R.D. (1999) Origin of carbonate-magnetite-sulfide assemblages in Martian meteorite ALH84001. J. Geophys. Res., 104, 3803-3813.

ED

Scott E.R.D., Yamaguchi A., Krot S.A. (1997) Petrological evidence for shock melting of carbonates in the martian meteorite ALH84001, Nature 387, pages 377–379

PT

Sherman D.M. (1985) The electronic structures of Fe3+ coordination sites in iron oxides; applications to spectra, bonding and magnetism. Phys Chem Mineral 12 : 161-175

CE

Sherman D.M., Waite T.D. (1985) Electronic spectra of Fe3+ oxides in the near UV to near IR. Am Mineral 70:1262-1269

AC

Sklute, E.C., Kashyap, S., Dyar, M.D., Holden J.F., Tague T., Wang P., Jaret S.J. (2018) Spectral and morphological characteristics of synthetic nanophase iron (oxyhydr)oxides. Phyics & Chemistry of Minerals 45: 1-26. Strens, R.G.J., and Wood B.J. (1979) Diffuse reflectance spectra and optical properties of some iron and titanium oxides and oxyhydroxides. Mineral. Mag., 43, 347-354. Swindle, T.D. (1998) Implications of iodine-xenon studies for the timing and location of secondary alteration. Meteorit. Planet. Sci. 33, 1147-1155. Theocharous, E., Deshpande, R., Dillon, A. C., Lehman, J. () "Evaluation of a pyroelectric detector with a carbon multiwalled nanotube black coating in the infrared". Applied Optics. 45 (6): 1093.

ACCEPTED MANUSCRIPT 31

Tossell, J.A., Vaughan, D.J., and Johnson, K.H. (1974) The electronic structure of rutile, wustite, and hematite from molecular orbital calculations. American Mineralogist, 59, 319-334. Vaughan, D.J. (1985) Spectroscopy and chemical bonding in the opaque minerals. In: Chemical Bonding and Spectroscopy in Mineral Chemistry (Berry, F.J., and D.J. Vaugahn, Eds.). Chapman and Hall, London, pp. 251-292.

CR IP T

Vaughan, D.J., and J.A. Tossell (1978) Major transition-metal oxide minerals: Their electronic structures and the interpretation of mineralogical properties. Canadian Mineralogist, 16, 159-168. Verwey, E. J. W. (1939) Electronic Conduction of Magnetite (Fe3O4) and its Transition Point at Low Temperatures Nature, Vol. 144, No 3642, pp. 327-328. Vukusic, P., Sambles, J. R. & Lawrence, C. R. (2004) Structurally assisted blackness in butterfly scales. Proc. R. Soc. Lond. Ser. B Biol. Sci. 271, S237–S239.

AN US

Wagner, J.K., B.W. Hapke, and E.N. Wells (1987) Atlas of reflectance spectra of terrestrial, lunar, and meteoritic powders and frosts from 92 to 1800 nm. Icarus, 69, 14-28. Watanabe, S., Tsuda, Y., Yoshikawa, M., Tanaka, S., Saiki, T., and Nakazawa, S. (2017) Hayabusa2 mission overview. Space Science Reviews, 208, 3-16.

M

Waychunas, G.A. (1991) Crystal chemistry of oxides and oxyhydroxides. In: Oxide Minerals: Petrologic and Magnetic Significance. Reviews in Mineralogy, 25 (Lindsley, D.H., Ed.), Mineralogical Society of America, Washington, D.C., 11-68. Wechsler, B.A., Lindsley, D.H., and Prewitt, C.T. (1984) Crystal structure and cation distribution in titanomagnetites (Fe3–xTixO4). American Mineralogist, 69, 754–770.

ED

Wicks, F. J., and Whittaker, E.J.W. (1977) Serpentine textures and serpentinization. Canadian Mineralogist Vol. 15, 459-488

PT

Yang, B., and D. Jewitt (2010) Identification of magnetite in B-type asteroids. Astron. J., 140, 692-698.

CE

Yamamoto, A. (1982) Modulated structure of wüstite (Fe1-xO) (three-dimensional modulation). Acta Crystallographica: B38: 1451-1456.

AC

Zega, T.J., P. Haenecour, C. Floss, and R.M. Stroud (2015) Circumstellar magnetite from the LAP 031117 CO3.0 chondrite. Astrophys. J., 808, 55 (9pp); doi:10.1088/0004-637X/808/1/55.

Figures and Captions

ACCEPTED MANUSCRIPT

AN US

CR IP T

32

AC

CE

PT

ED

M

Figure 1: Optical constants n and k for magnetite calculated from near-normal incidence specular reflectance measured by Querry (1985) (http://www.dtic.mil/docs/citations/ADA158623) compared with Fresnel reflectance calculated following Heavens (1991), multiplied by a factor of 10 for ease of comparison; and reflectance spectra (this study) from 0.2-0.4 μm (biconical) and 0.35-2.5 μm (i=0º, e=30º). The 0.2-0.4 μm spectra were scaled to equal the 0.4 μm reflectance in the 0.35-2.5 μm spectra, and both were multiplied by a factor of 25 for ease of comparison. Magnetite spectra are of sample MAG102, which has minimal chemical impurities, and no mineralogical contaminants detectable by X-ray diffraction. This figure demonstrates that the local maxima in reflectance near 0.27, 0.39, and 0.65 μm correspond with local maxima in n and k and are interpreted as Fresnel peaks associated with strong absorptions due to ligand-tometal charge transfer (0.27 and 0.39 μm) and ligand-field transition (0.65 μm) processes. Positions of local maxima were determined by subtracting a linear continuum in the region of each peak, followed by fitting 3rd-order polynomials to the continuum-removed feature; the wavelength of maximum of the fit polynomial was then taken as the Fresnel peak location.

ACCEPTED MANUSCRIPT

AN US

CR IP T

33

AC

CE

PT

ED

M

Figure 2: Grain size effects on reflectance spectra of magnetite. A) Ultraviolet spectral region, biconical geometry, absolute reflectance, sample MAG102. Absolute reflectance does not show a systematic variation with grain size. B) Ultraviolet reflectance normalized at 0.25 μm. C) 0.352.5 μm spectral region, i=0º, e=30º, sample MAG107. D) reflectance normalized at 0.58 μm. Grain size effects on positions and intensities of Fresnel peaks and apparent band minima are minimal, with no discernable changes in the position of Fresnel peaks or local minima. The smallest grain size fraction shows slightly lower normalized reflectance at wavelengths longward of ~1.5 μm. The smallest grain size, <45 μm shows higher reflectance overall. The other grain size fractions show a small increase in reflectance with grain size over all wavelengths.

ACCEPTED MANUSCRIPT

PT

ED

M

AN US

CR IP T

34

AC

CE

Figure 3: Spectra of mixtures of magnetite (MAG102) and the CM2 chondrite Murchison, both <45 μm grain size powders, and spectral slope. A) Spectral effects of mixing magnetite with CM2 chondrite include decreased spectral slope at wavelengths below ~0.5 μm, a progressively deeper local minimum near 1 μm, and increased reflectance in the infrared from ~1.9 up to 2.5 μm. Even at a very high 50:50 ratio of magnetite to CM2 material, however, the spectral effects of the macromolecular organic carbon in the CM2 chondrite dominate spectral slope behaviour in the visible and near-infrared wavelengths up to ~1 μm. B) The spectral slope between 0.65 and 1.0 μm shows a systematic decrease (―bluer‖ spectral slope) with increasing magnetite concentration. Also shown is the spectral slope of a mixture of 80 wt. % magnetite and 20 wt. % serpentine (pink triangle) to illustrate that the general trend in this spectral slope parameter with increasing magnetite continues to hold at higher magnetite concentrations (HOSERLab samples MAG107 and SRP110).

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

M

AN US

CR IP T

35

ACCEPTED MANUSCRIPT 36

AC

CE

PT

ED

M

AN US

CR IP T

Figure 4: Mixtures of phyllosilicates with magnetite (MAG102) and nanophase maghemite (MAG200). A) For mixtures of serpentine and magnetite (both <45 µm grain size), reflectance decreases, spectral slope becomes bluer, and serpentine absorption bands become shallower, with increasing magnetite. B) Same as A, but normalized at 0.653 µm. C) Mixtures of nanophase maghemite (MAG200) with serpentine (SRP117), showing that the decreases in reflectance and reduction in serpentine band depths are more pronounced than with magnetite. D) Same as A, but normalized at 0.65 µm. E) Nanophase maghemite more effectively lowers absolute reflectance, suppresses serpentine absorption bands (e.g., band near 0.75 µm), and causes a more concave spectral shape than magnetite. F) Same as E, but normalized at 0.653 µm. G) Mixtures of nanophase maghemite (MAG200) and saponite (SAP103) showing decreased overall reflectance and shallower absorption bands with increasing nanophase maghemite concentration. H) Same as G, but normalized at 0.653 µm. I) The spectrum-altering effects of nanophase maghemite (MAG200) on cronstedtite (CRO101) spectra are less dramatic, and are largely confined to a reduction in overall reflectance. J) Same as I, but normalized at 0.653 µm.

ACCEPTED MANUSCRIPT

PT

ED

M

AN US

CR IP T

37

AC

CE

Figure 5: Effects of grinding on serpentine-magnetite and serpenite-nanophase maghemite mixtures. Spectra are shown both in absolute reflectance and normalized to one at 0.65 μm. A) Spectra of a mixture of 90% <45 μm serpentine (HOSERLab sample SRP111) and 10 wt. % <45 μm magnetite (MAG102) showing a progressive decrease in spectral contrast with grinding. B) Same as A, normalized to reflectance at 0.65 μm. C) Variation in reflectance spectra of a mixture of 5 wt. % nanophase maghemite with 95 wt. % <45 μm serpentine (SRP117) with grinding. There is a very significant change between the unground and ground samples, with an increased UV drop-off and enhanced local minimum centered near 1.5 μm, indicating that the spectral influence of the nanophase material is enhanced by grinding. The increased UV drop-off may also result from the formation of hæmatite or other chemical changes during grinding. The greater influence of the nanophase material with grinding may be due to physical changes caused by the grinding process such as enhanced coating of the serpentine grains by progressively finergrained and better dispersed nanoparticles. The locations of the serpentine OH-overtone/combination bands near 1.4 and 2.3 microns are unchanged, but their spectral

ACCEPTED MANUSCRIPT 38

AC

CE

PT

ED

M

AN US

CR IP T

contrast is generally reduced. D) Same as C but normalized at 0.65 μm. E) Reflectance spectra of a mixture of 10 wt.% nanophase maghemite and 90 wt.% <45 μm serpentine (SRP117), showing similar effects as in C including increased UV dropoff and decrease in spectral contrast. F) Same as E, but normalized at at 0.65 μm.

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

M

AN US

CR IP T

39

Figure 6: Magnetite reflectance spectra compared with various related Fe-Ti oxides. A) Comparison of magnetite reflectance spectra (0.35-2.5 μm) with those of ulvöspinel,

ACCEPTED MANUSCRIPT 40

AC

CE

PT

ED

M

AN US

CR IP T

titanomagnetite, and ilmenite. Ulvöspinel (RELAB data) sample grain size is <125 μm, other samples are <45 μm. Ilmenite and ulvöspinel have distinctive local maxima near 1 μm and 1.5 μm, respectively. Titanomagnetite is spectrally similar to magnetite. The offset in absolute reflectance between magnetite and titanomagnetite is likely due mainly to grain size. B) Comparison of magnetite reflectance spectra (0.35-2.5 μm) with those of hæmatite and maghemite. All samples <45 μm grain size. Both hæmatite and maghemite have red visible spectral slopes due to very strong UV absorptions compared to magnetite, and hæmatite is further characterized by very high overall reflectance and distinct features due to crystal-field transitions of Fe3+ near 0.85 and 0.44 μm. C) Comparison of magnetite reflectance (0.35-2.5 μm) with those of wüstite (Fe1-xO, 0.04
ACCEPTED MANUSCRIPT

AC

CE

PT

ED

M

AN US

CR IP T

41

Figure 7: Variation in reflectance spectra of magnetite in chemically pure versus highly substituted magnetite samples of <45 μm grain size. There is a general trend of decreasing

ACCEPTED MANUSCRIPT 42

CE

PT

ED

M

AN US

CR IP T

reflectance with chemical substitution. A) 0.2-0.4 μm wavelength, biconical geometry. B) 0.352.5 μm wavelength, i=0º, e=30º. C) There is a small systematic shift in the position of the ~0.65 μm Fresnel peak with increasing chemical impurity in magnetite, at least for this limited sample set. Relevant minor element concentrations are listed in the inset table, full compositions are given in Table 1.

AC

Figure 8: Surface porosity and aggregation state effects on the reflectance spectra of magnetite, sample MAG107, <45 μm grain size. A) Porous or ―fluffy‖ packing and normal (poured into sample cup and levelled) spectra are very similar. Spectra of compacted samples (packed dense or pelletized) show higher reflectance, bluer spectral slopes shortward of ~1 μm, and increased spectral contrast due to reduced multiple scattering in compacted samples. B) Spectral differences due to surface porosity are small in normalized reflectance, but the same trend of increasing blue spectral slope shortward of ~1 μm is observable.

Tables

CR IP T

ACCEPTED MANUSCRIPT

43

Table 1: Chemical compositions for magnetite samples from XRF and wet chemistry, and summary of mineralogy from XRD Sample

TiO2

FeO

Fe2O3

Cr2O3

Al2O3

MgO

MnO

NiO

ZnO

CaO

V2O5

SiO2

MAG101

3.04

26.52

63.71

0.03

0.33

3.63

0.60

n.d.

n.d.

n.d.

0.36

n.d.

MAG102

0.04

28.37

66.88

0.01

0.38

0.61

0.06

0.03

n.d.

n.d.

0.17

n.d.

MAG103

7.28

28.80

60.38

<0.01

0.43

<0.01

2.00

<0.01

n.d.

n.d.

0.05

MAG104

0.01

29.44

56.22

0.02

0.02

0.13

0.08

0.00

MAG105

0.01

29.82

59.22

<0.01

0.03

0.01

0.06

0.00

MAG106

4.55

86.82

n.d.

0.27

<0.01

0.35

0.53

0.06

Total

XRD

n.d.

98.22

magnetite, hematite

n.d.

96.55

magnetite, ilmenite

n.d.

98.94

magnetite, hematite

MAG108

0.06

21.51

66.44

<0.01

0.13

1.48

0.24

0.00

MAG109

11.29

20.76

50.28

0.00

3.87

6.74

1.77

0.00

MAG110

0.01

19.08

70.66

<0.01

0.08

0.00

0.24

0.01

MAG111

4.42

17.13

52.25

0.01

4.28

6.81

4.41

0.03

MAG112

0.04

26.81

61.56

0.01

0.18

0.02

0.04

0.09

AN US

n.d.

Na2O

MAG113

0.38

19.39

67.08

<0.01

1.39

0.43

0.01

MAG117

1.24

11.69

73.74

0.01

0.57

1.04

0.02

MAG118

0.13

23.53

65.23

0.01

0.19

0.07

0.02

MAG119

0.10

22.66

64.83

0.01

0.21

0.23

0.07

MAG120

0.04

19.51

67.49

0.00

0.17

0.18

0.01

MAG121

0.25

22.43

64.82

0.01

0.54

0.32

0.04

28.25

n.d.

0.03

n.d.

n.d.

85.96

n.d.

n.d.

0.03

n.d.

n.d.

89.19

n.d.

n.d.

0.26

n.d.

92.84

magnetite

28.25

magnetite, hematite magnetite, hematite

n.d.

magnetite, hematite

0.00

0.00

0.00

0.00

89.87

0.01

0.00

0.35

0.00

0.00

95.08

0.00

0.00

0.03

0.00

0.00

90.12

magnetite

0.11

0.09

0.00

0.39

0.36

90.29

magnetite

0.02

0.00

0.01

0.00

0.00

88.79

magnetite, hematite, ilmenite

0.02

0.00

0.00

0.03

0.00

0.00

88.74

magnetite, hematite, goethite

0.05

0.01

0.00

0.57

0.00

0.00

88.95

magnetite, hematite, quartz

0.04

0.00

0.00

0.31

0.00

0.00

89.53

0.03

<0.01

0.01

0.21

0.01

0.00

88.38

0.07

0.07

0.02

0.22

0.03

0.00

87.82

0.05

0.03

0.01

0.18

0.07

0.00

88.76

M

0.00

ED

MAG107

n.d.

magnetite, hematite, goethite

magnetite

AC

CE

from XRD in bold.

PT

FeO determined by wet chemistry; Fe2O3 taken as difference between total and ferrous iron, n.d. = not determined. Major minerals