Article Mineralogy, Mineral Chemistry and SWIR Spectral Reflectance of Chlorite and White Mica
Jonathan Cloutier 1,*, Stephen J. Piercey 2 and Jonathan Huntington 3
1 Centre for Ore Deposits and Earth Science, University of Tasmania, Private Bag 79, Hobart, TAS 7001, Australia 2 Department of Earth Sciences, Memorial University of Newfoundland, 300 Prince Philip Drive, St. John’s, NL A1B 3X5, Canada; [email protected] 3 CSIRO Mineral Resources, PO Box 218, Lindfield, NSW 2070, Australia; [email protected] * Correspondence: [email protected]
Abstract: Hyperspectral reflectance has the potential to provide rapid and low-cost mineralogical and chemical information that can be used to vector in mineral systems. However, the spectral signature of white mica and chlorite, despite numerous studies, is not fully understood. In this study, we review the mineralogy and chemistry of different white mica and chlorite types and investigate what mineralogical and chemical changes are responsible for the apparent shifts in the shortwave infrared (SWIR) spectroscopic absorption features. We demonstrate that the spectral signature of white mica is more complex than previously documented and is influenced by the Tschermak substitution, as well as the sum of interlayer cations. We show that an increase in the interlayer deficiencies towards illite is associated with a change from steep to shallow slopes between the wavelength position of
the 2200 nm feature (2200 W) and Mg, Al(VI) and Si. These changes in slope imply that white micas with different elemental chemistry may be associated with the same 2200 W values and vice versa, contrary to traditional interpretation. We recommend that traditional interpretations should only be used in true white mica with sum interlayer cations (I) > 0.95. The spectral signature of trioctahedral chlorite (clinochlore, sheridanite, chamosite and ripidolite) record similar spectral relationships to Citation: Cloutier, J.; Piercey, S.J.; Huntington, J. Mineralogy, Mineral those observed in previous studies. However, dioctahedral Al-rich chlorite (sudoite, cookeite and Chemistry and SWIR Spectral donbassite) has a different spectral response with Mg increasing with 2250 W, which is the opposite of Reflectance of Chlorite and White traditional trioctahedral chlorite spectral interpretation. In addition, it was shown that dioctahedral Mica. Minerals 2021, 11, 471. https:// chlorite has a 2200 W absorption feature that may introduce erroneous spectral interpretations of doi.org/10.3390/min11050471 white mica and chlorite mixtures. Therefore, care should be used when interpreting the spectral signature of chlorite. We recommend that spectral studies should be complemented with electron Academic Editor: Bernard Hubbard microprobe analyses on a subset of at least 30 samples to identify the type of muscovite and chlorite. This will allow the sum I of white mica to be obtained, as well as estimate the slope of 2200 W Received: 17 March 2021 absorption trends with Mg, Al(vi), and Si. Preliminary probe data will allow more accurate spectral Accepted: 27 April 2021 interpretations and allow the user to understand the limitations in their hyperspectral datasets. Published: 30 April 2021
Keywords: mineralogy; mineral chemistry; SWIR; white mica; chlorite; hyperspectral reflectance Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affil- iations. 1. Introduction Hydrothermal ore deposits are commonly associated with phyllosilicate alteration haloes that contain variable mineralogical, geochemical, geophysical and spectral charac- teristics depending on the deposit type [1,2]. Of these, hyperspectral reflectance has the Copyright: © 2021 by the authors. Licensee MDPI, Basel, Switzerland. potential to provide rapid and low cost mineralogical and chemical information that can This article is an open access article be used to vector towards and within mineralized zones [3]. The spectral characteristics distributed under the terms and of white mica and chlorite are widely used in exploration, as they have been the focus of conditions of the Creative Commons several studies (e.g., [4–11]) and are thought to be well understood. The spectral signature Attribution (CC BY) license (https:// of white mica is strongly influenced by the Tschermak substitution between the octahedral creativecommons.org/licenses/by/ (Al(VI), Mg, Fe) and tetrahedral (Si, Al(VI)) sites, whereas the spectral signature of chlorite is 4.0/). influence by Mg and Fe substitutions in the octahedral site [5–8]. The correlation between
Minerals 2021, 11, 471. https://doi.org/10.3390/min11050471 https://www.mdpi.com/journal/minerals Minerals 2021, 11, x FOR PEER REVIEW 2 of 16
of white mica is strongly influenced by the Tschermak substitution between the octahe- dral (Al(VI), Mg, Fe) and tetrahedral (Si, Al(VI)) sites, whereas the spectral signature of chlo- Minerals 2021, 11, 471 2 of 16 rite is influence by Mg and Fe substitutions in the octahedral site [5–8]. The correlation between white mica and chlorite chemistry and spectral response is generally high for each specific area studied (e.g., [7,8]), but greatly decreases when data from different areas whiteare combined mica and and chlorite compared, chemistry suggesting and spectral that unknown response mineralogical is generally high or chemical for each specificparam- areaeters studied also control (e.g., [the7,8 ]),spectral but greatly signature decreases of white when mica data and from chlorite different. Understanding areas are combined these andparameters compared, is critical, suggesting as there that are unknown often specif mineralogicalic mica and or chlorite chemical species parameters associated also with con- trolmineralization the spectral (e.g., signature [10,11]), of such white that mica improved and chlorite. interpretation Understanding of spectral these data parameters may lead isto critical,the identification as there are of oftennew exploration specific mica target ands chlorite that may species have associatedbeen missed with in historical mineral- izationdatasets. (e.g., [10,11]), such that improved interpretation of spectral data may lead to the identificationThis study of aims new explorationto review the targets mineralogy that may and have chemistry been missed of different in historical white mica datasets. and chloriteThis types study to aims investigate to review what the mineralogical mineralogy and and chemistry chemical ofchanges different are white responsible mica and for chloriteshifts in typesthe shortwave to investigate infrared what (SWIR) mineralogical spectroscopic and chemical absorption changes features. are responsibleIt uses original for shiftsand publicly in the shortwave available databases infrared (SWIR) to demonstrate spectroscopic that absorptionthe spectral features. signature It of uses white original mica andand publiclychlorite is available more complex databases than to previously demonstrate documented. that the spectral signature of white mica and chlorite is more complex than previously documented. 2. Mineralogy Background 2. Mineralogy Background 2.1.2.1. White Mica Group WhiteWhite micasmicas areare phyllosilicatesphyllosilicates inin whichwhich thethe unitunit structurestructure consistsconsists ofof oneone octahedraloctahedral sheetsheet betweenbetween twotwo opposingopposing tetrahedraltetrahedral sheets.sheets. TheseThese areare separatedseparated fromfrom adjacentadjacent layerslayers byby planes of of non-hydrated non-hydrated interlayer interlayer cations cations with with no no less less than than 0.6 0.6cations cations per performula formula unit unit[12]. [If12 the]. If interlayer the interlayer cations cations are arebetween between 0.6 0.6and and 0.85, 0.85, the the white white mica mica is characterized is characterized as asan aninterlayer-deficient interlayer-deficient mica. mica. If Ifthe the interlayer interlayer cations cations are are greater greater than than 0.95, they are ofof thethe muscovite-celadonitemuscovite-celadonite series,series, andand ifif thethe interlayerinterlayer cationscations areare nearnear 0.650.65 theythey areare ofof thethe illiteillite series.series. TheThe simplifiedsimplified formulaformula isis writtenwritten asas II0.6–1.00.6–1.0 MM2–32–3 ❏ 1–0 T41–0O T104 OA102, A where:2, where: I: interlayered cation that is K, Na and Ca, and less commonly Cs, NH4, Rb and Ba I: interlayered cation that is K, Na and Ca, and less commonly Cs, NH4, Rb and Ba M: octahedral cation that is Fe (di- or trivalent), Mg, Al, Ti and Li, and less commonly M: octahedral cation that is Fe (di- or trivalent), Mg, Al, Ti and Li, and less commonly Mn (di- or trivalent), Zn, Cr and V Mn (di- or trivalent), Zn, Cr and V : vacancy in the octahedral sheet ❏: vacancy in the octahedral sheet T: tetrahedral cation that is Si, Al and Fe (trivalent), and less commonly Be and B T: tetrahedral cation that is Si, Al and Fe (trivalent), and less commonly Be and B A: anion that is OH, F and Cl, and less commonly O (oxy-micas) and S. A: anion that is OH, F and Cl, and less commonly O (oxy-micas) and S. White micas can be divided into true micas if more than 50% of I cations are mono- valent,White and micas brittle can micas be divided if more into than true 50% micas of I if cations more than are divalent. 50% of I cations White micasare mono- are dioctahedralvalent, and brittle if the Mmicas site containsif more than less than50% of 2.5 I cationscations andare trioctahedraldivalent. White if it micas contains are moredioc- thantahedral 2.5 cations if the M [12 site]. contains less than 2.5 cations and trioctahedral if it contains more than Muscovite-celadonite2.5 cations [12]. series contain true dioctahedral mica with I greater than 0.95.Muscovite-celadonite They vary between series a K-rich contain muscovite true dioctahedral end-member mica with with a generalI greater formula than 0.95. of KAlThey2 vary(Si3Al)O between10(OH) 2aand K-rich a Fe-Mg-rich muscovite celadonite end-member end-member with witha general a general formula formula of 3+ VI ofKAl KFe2❏(Si(Mg,3Al)O Fe10(OH)2+) Si2 and4O10 a(OH) Fe-Mg-rich2. The mica celadonite is classified end-member as muscovite with a when general Al# formula ( Al/ VI (of AlKFe +3+ Fe(Mg,tot +Fe Mg)2+)❏ isSi4 greaterO10(OH) than2. The 0.5 mica and is celadonite classified as when muscovite it is less when than Al# 0.5 ( [VI12Al/(,13].VIAl Al- + thoughFetot + Mg) not is a greater recognized than mineral,0.5 and celadonite in this study when we it identifyis less than the 0.5 mica [12,13]. as phengite Although when not thea recognized Si:Al ratio mineral, in the tetrahedral in this study layer we is identi greaterfy the than mica 3.25, as following phengite when the classification the Si:Al ratio of Deerin the et tetrahedral al. [14]. The layer main is substitutions greater than in 3.25, the muscovite-celadonitefollowing the classification series areof Deer the Tschermak et al. [14]. 3+ 3+ ↔ 4+ 2+ 2+ substitutionThe main substitutions (Al tet + Al in theoct muscovite-celadoSi tet + Mg octniteor Feseriesoct are) in the the Tschermak octahedral andsubstitution tetrahe- dral sites, and the Na+ ↔ K+ substitution in the interlayer plane. Tischendorf et al. [12] 3+tet 3+oct ↔ 4+tetint 2+octint 2+oct (Al + Al Si + + Mg + or Fe ) in the octahedral and tetrahedral sites, and the showed that the Na int ↔ K int substitution does not constitute a solid-solution with a Na+int ↔ K+int substitution in the interlayer plane. Tischendorf et al. [12] showed that the paragonite end-member (NaAl2 (Si3Al)O10(OH)2). They identified an immiscibility gap Na+int ↔ K+int substitution does not constitute a solid-solution with a paragonite end-mem- between approximately 0.4 to 0.6 Na atoms per formula unit and argue that this immis- ber (NaAl2❏(Si3Al)O10(OH)2). They identified an immiscibility gap between approxi- cibility gap results from the large difference in ionic radius between Na+ (0.95 Å) and mately 0.4 to 0.6 Na atoms per formula unit and argue that this immiscibility gap results K+ (1.33 Å). from the large difference in ionic radius between Na+ (0.95 Å) and K+ (1.33 Å). Illite series contain dioctahedral interlayer-deficient micas between 0.6 and 0.85 with an ideal general formula K0.65Al2 (Si3.35Al0.65)O10(OH)2 [12]. The deficiency in illite +4 originates from excess Si tet that is compensated by a deficiency in K in the interlayer +4 3+ + plane (Si tet ↔ Al tet + K int; Deer et al., 1992). Other substitutions that can occur + + 2+ 2+ 2+ are Na int ↔ K int, similar to the muscovite series, and Ca int ↔ Mg oct or Fe oct 3+ substituting for Al oct. Minerals 2021, 11, x FOR PEER REVIEW 2 of 16
of white mica is strongly influenced by the Tschermak substitution between the octahe- dral (Al(VI), Mg, Fe) and tetrahedral (Si, Al(VI)) sites, whereas the spectral signature of chlo- rite is influence by Mg and Fe substitutions in the octahedral site [5–8]. The correlation between white mica and chlorite chemistry and spectral response is generally high for each specific area studied (e.g., [7,8]), but greatly decreases when data from different areas are combined and compared, suggesting that unknown mineralogical or chemical param- eters also control the spectral signature of white mica and chlorite. Understanding these parameters is critical, as there are often specific mica and chlorite species associated with mineralization (e.g., [10,11]), such that improved interpretation of spectral data may lead to the identification of new exploration targets that may have been missed in historical datasets. This study aims to review the mineralogy and chemistry of different white mica and chlorite types to investigate what mineralogical and chemical changes are responsible for shifts in the shortwave infrared (SWIR) spectroscopic absorption features. It uses original and publicly available databases to demonstrate that the spectral signature of white mica and chlorite is more complex than previously documented.
2. Mineralogy MineralsBackground2021, 11 , 471 3 of 16 2.1. White Mica Group White micas are phyllosilicates in which the unit structure consists of one octahedral sheet between two opposing tetrahedral sheets.2.2. These Chlorite are Group separated from adjacent layers by planes of non-hydrated interlayer cations with Chloriteno less than is a 0.6 hydrous cations phyllosilicateper formula unit in which the unit structure consists of alter- [12]. If the interlayer cations are between 0.6nating and 0.85, a negatively the white chargedmica is characterized tetrahedral–octahedral–tetrahedral as talc-like 2:1 layer (TOT; an interlayer-deficient mica. If the interlayerM cations6T8O20 are(OH) greater4) with than a positively0.95, they chargedare of the octahedral brucite-like layer (O; M6(OH)12), muscovite-celadonite series, and if the interlayerresulting cations in are a 2:1:1near 0.65 stacking they are pattern of the [illite14–16 ]. The simplified formula is written as 2+ 3+ 3+ series. The simplified formula is written as I0.6–1.0((R MxR2–3 y ❏ 6-x-y )61–0(Si Tz4R O104-z A)24, Owhere:10(OH) 8, where: R2+ : octahedral divalent cation that is Fe, Mg and Mn and less commonly Ni, Co, Zn, I: interlayered cation that is K, Na and Ca, and lessx commonly Cs, NH4, Rb and Ba and Cu M: octahedral cation that is Fe (di- or trivalent),3+ Mg, Al, Ti and Li, and less commonly Mn (di- or trivalent), Zn, Cr and V R y: octahedral trivalent cation that is Al and Fe, and less commonly Cr and V : vacancy in the octahedral sheet ❏: vacancy in the octahedral sheet R3+ : tetrahedral trivalent cation that is Al and less commonly B and Fe. T: tetrahedral cation that is Si, Al and Fe (trivalent),y and less commonly Be and B Chlorites are divided into type I (dioctahedral) and type II (di-trioctahedral and tri- A: anion that is OH, F and Cl, and less commonly O (oxy-micas) and S. octahedral) based on their Mg, Fe, Al and content. They are classified as dioctahedral White micas can be divided into true micaswhen if the more sum than of Mg 50% and of FeI cations is greater are thanmono- the sum of Al and vacancy, and as trioctahedral valent, and brittle micas if more than 50% ofwhen I cations the are sum divalent. of Al and White vacancy micas is are greater dioc- than the sum Mg and Fe [15] Chlorites are tahedral if the M site contains less than 2.5 furthercations classifiedand trioctahedral based on if their it contains Si and vacancymore (or R2+ or R3+) content and, lastly, based than 2.5 cations [12]. on their Mg number [17]. Type I chlorite with low Si (2.5 apfu) is classified as sheridanite Muscovite-celadonite series contain true (Mg#dioctahedral > 0.5) andmica ripidolitewith I greater (Mg# than < 0.5), 0.95. with moderate Si (3.0 apfu) as clinochlore They vary between a K-rich muscovite (Mg#end-member > 0.5) andwith chamosite a general (Mg# formula < 0.5), of and with high Si (3.5 apfu) as pennantite 2+ KAl2❏(Si3Al)O10(OH)2 and a Fe-Mg-rich celadonite(Mg# > 0.5)end-member and diabantite with a (Mg# general < 0.5). formula Type II chlorite with vacancy of 1 apfu (R = 2 or 3+ 2+ of KFe3+(Mg, Fe2+)❏Si4O10(OH)2. The mica is classifiedR = 3 apfu) as muscovite is classified when as sudoite Al# (VIAl/( andVI asAl donbassite+ with vacancy of 1.5 apfu (R = 0.5 3+ Fetot + Mg) is greater than 0.5 and celadonite orwhen R it= is 4 apfu)less than [17]. 0.5 [12,13]. Although not a recognized mineral, in this study we identify the mica as phengite when the Si:Al ratio 3. Spectral Reflectance Background in the tetrahedral layer is greater than 3.25, following the classification of Deer et al. [14]. The main substitutions in the muscovite-celadoniteSpectral series reflectanceare the Tschermak is a non-destructive substitution technique that measures the relative absorp- (Al3+tet + Al3+oct ↔ Si4+tet + Mg2+oct or Fe2+oct) in tionthe octahedral of photons and interacting tetrahedral with sites, minerals and the over a specific wavelength interval relative to Na+int ↔ K+int substitution in the interlayer plane.a reference Tischendorf standard et al. [4 ].[12] Absorption showed that features the in the visible–near infrared–shortwave in- + + frared (380–2500 nm, VNIR/SWIR) occur due to electronic processes from wavelengths Na int ↔ K int substitution does not constitute a solid-solution with a paragonite end-mem- of 10 to about 1300 nm and vibrational processes for wavelengths above 900 nm [18–21]. ber (NaAl2❏(Si3Al)O10(OH)2). They identified an immiscibility gap between approxi- − mately 0.4 to 0.6 Na atoms per formula unit Forand aargue molecule that this with immiscibility N atoms, there gap are results 3N 6 normal modes of vibrations, which are from the large difference in ionic radius betweencalled Na fundamentals.+ (0.95 Å) and K Additional+ (1.33 Å). vibrations, termed overtones, occur near multiples of original fundamental frequency using wave numbers (cm−1). Libration, which is a type reciprocating motion, also occurs at low frequencies. Additional weaker vibrations occur at multiples of a single fundamental mode and/or combinations thereof [4].
3.1. OH Stretching Vibrations In hydroxyl-bearing silicates (e.g., amphibole, chlorite, mica), the stretching bond with OH is responsible for the main absorption features in the near infrared (NIR) and shortwave infrared (SWIR) regions. The fundamental OH stretching vibrations (ν(OH)) occur near 3600 cm−1 (2775 nm) and have been the focus of several studies (Table1, e.g., [18,22–27]). The frequency of vibration depends on the strength of the bond in a molecule, which is affected by the combined mass of the pair forming the bond, the valence of the cation (e.g., divalent versus trivalent), and the stacking order angle (ρ) (or crystallinity). Bonds with the heavier Fe2+–Fe2+ pair are associated with the lowest frequency ν(OH) at 3505 cm−1 (2853 nm), whereas bonds with the lighter Al–Al pair have the highest frequency at 3621 cm−1 (2762 nm) (Table1). In addition, cations with the highest charge have absorption features with the lowest frequency (e.g., ν(OH)Fe2+–Fe2+ −1 −1 at 3505 cm (2853 nm) versus ν(OH)Fe3+–Fe3+ at 3535 cm (2829 nm); [22]. Besson and Drits [25,26] observed that the ν(OH)Al–Al is associated with multiple frequencies and documented a change in ν(OH) frequency based on the ρ angle between sheets in micas, wherein a small ρ angle is associated with stronger OH–O interactions and lower ν(OH) frequency. They estimated that on average, a change in the ρ angle by 1◦ may change the band frequency by up to 2.5 cm−1. Besson and Drits [25,26] also observed that in mica with very low Fe and Mg, the dominant absorption feature tends to occur at higher frequency Minerals 2021, 11, 471 4 of 16
near 3640 cm−1 (2747 nm) instead of 3621 cm−1 (2762 nm). They attributed this shift to the increased amount of Al(VI) cations and the different Al(VI) and Al(VI) configuration in the tetrahedral and octahedral layers, which may result in a greater ρ angle [28] and an increase in the OH stretching frequency. It should be remembered that, until recently, these fundamental vibrations, so important for observing OH-bearing silicates, were outside the range of most hand-held field instruments.
Table 1. OH stretching vibration, overtones and the octahedral cations bonded to OH group.
ν(OH) δ(OH) Bond Pair Wave Number (cm−1) Wave Number (cm−1) Al–Al 3658 (1) 915 (2) Al–Al 3641 (1) 915 (2) Al–Al 3621 (1) 915 (2) Al–Mg 3604 (1) 840 (2) Al–Fe3+ 3573 (1) 875 (3) Al–Fe2+ 3559 (1) N/A Al–Fe2+ 3600 (4) N/A Mg–Mg 3583 (1) 600–670 (2) Mg–Fe3+ 3558 (1) 800 (2) Mg–Fe2+ 3559 (1) N/A Fe3+–Fe3+ 3535 (1) 818 (2) Fe2+–Fe3+ 3521 (1) 800 (2) Fe2+–Fe2+ 3505 (1) N/A (1) Besson and Dritz [25], (2) Farmer [18], (3) Martínez-Alonzo et al. [27], and (4) Redhammer et al. [29]. N/A: Not available.
3.2. White Mica Group White micas have diagnostic overtone absorption features at 2200 nm and 2340 nm (Figures1 and2;[ 4,6,7]). The wavelength position of the 2200 nm feature (herein referred to as 2200 W) varies between 2190 and 2225 nm and is the result of overtones between the fundamental OH stretching vibrations of the OHAl–Al, OHAl–Mg and OHAl–Fe near −1 −1 3600 cm (2775 nm) and OHAl–Al librational vibration (δ(OH)) near 915 cm (10,929 nm) (Table2;[ 22]). The position of the absorption feature was documented to negatively correlate with the abundance of Al(VI) in the octahedral layer, which also coincides with an increase in Si and Mg–Fe due to the Tschermak substitution [5,7,30,31]. White micas have been empirically classified as paragonite when the 2200 nm feature occurs between 2180 and 2195 nm, muscovite between 2195 and 2215 nm, and phengite between 2215 and 2225 nm [6], whereas illite has been distinguished by its deeper 1900 nm absorption feature, Minerals 2021, 11, x FOR PEER REVIEW 5 of 16 but it cannot be reliably distinguished from muscovite in the SWIR region [32].
FigureFigure 1.1. StackedStacked NIR/SWIR NIR/SWIR reflectance reflectance spectra of muscovite, spectra chlorite of muscovite, and sudoite.chlorite and sudoite.
Figure 2. Variation in spectral absorption feature for (A) white mica and (B) trioctahedral chlorite (after Clark et al. [4]; Pontual et al. [6], and Kokaly et al. [33]; modified from Cloutier et al. [3]). Minerals 2021, 11, x FOR PEER REVIEW 5 of 16
Minerals 2021, 11, 471 5 of 16
Figure 1. Stacked NIR/SWIR reflectance spectra of muscovite, chlorite and sudoite.
FigureFigure 2. 2.Variation Variation inin spectralspectral absorption feature for for ( (AA)) white white mica mica and and (B (B) trioctahedral) trioctahedral chlorite chlorite (after(after Clark Clark et et al. al. [ 4[4]];; Pontual Pontual et etal. al. [[6],6], andand KokalyKokaly etet al.al. [[3333];]; modifiedmodified from Cloutier et al. [3] [3]).).
Table 2. OH stretching vibration, overtones and the octahedral cations bonded to OH group for white mica.
ν(OH) δ(OH) Overtone Wave Number Wave Number Wave Number Bond Pair Bond Pair nm (cm−1) (cm−1) (cm−1) Al–Al 3658 (1) Al–Al 915 (2) 4573 2187 Al–Al 3641 (1) Al–Al 915 (2) 4556 2195 Al–Al 3621 (1) Al–Al 915 (2) 4536 2205 Al–Mg 3604 (1) Al–Al 915 (2) 4519 2213 Al–Fe3+ 3573 (1) Al–Al 915 (2) 4488 2228 Al–Fe2+ 3559 (1) Al–Al 915 (2) 4474 2235 (1) Besson and Dritz [25], (2) Farmer [18].
3.3. Chlorite Group Trioctahedral chlorites have diagnostic overtone absorption features at 2250 and 2340 nm (Figures1 and2;[ 6,8,34]. These absorption features arise from bonds in the octa- hedral sheets and are due to combinations of fundamental Al–, Fe– and Mg–OH stretching near 3600 cm−1 (2775 nm) with OH librational vibrations near 800 cm−1 (12,500 nm) for the 2250 nm feature and near 650 cm−1 (15,385 nm) for the 2340 nm feature [18,25,26,34]. The position of the 2250 nm (herein referred to as 2250 W) and 2340 nm absorption features Minerals 2021, 11, 471 6 of 16
correlate with their Mg number and their Mg and Fe content, wherein an increase in Mg number shifts the position of the absorption feature to shorter wavelengths [6,8,34]. Chlo- rites have been empirically classified as Mg-rich chlorite when their absorption features vary between 2240 and 2249 nm and 2320 and 2329 nm, intermediate (or Mg–Fe) when they vary between 2250 and 2255 nm and 2330 and 2349 nm, and Fe-rich when they vary between 2256 and 2265 nm and 2350 and 2360 nm [6,8,34]. Al-rich dioctahedral chlorites, such as sudoite (Al), have a well-developed 2200 nm feature in addition to the 2250 nm (shallow) and 2340 nm (deep) features, due to their high Al content (Figure1;[8,35,36]).
4. Methodology Fourteen white mica samples and 16 chlorite samples were analysed for their mineral chemistry by electron probe microanalysis (EPMA) on polished thin sections and by approximately 1.5 cm wide spot VNIR/SWIR spectral reflectance on thin section off cuts. The samples were selected on the basis of their 2200 (white mica) and 2250 nm (chlorite) absorption feature to cover the variability of the datasets. The mineral chemistry of samples from the Lemarchant volcanogenic massive sulfide (VMS) deposit in Newfoundland and Labrador, Canada were analysed at the Central Science Laboratory of the University of Tasmania, Australia, using a JEOL JXA-8530F Plus electron microprobe equipped with 5 tunable wavelength dispersive spectrometers, whereas the Athabasca Basin samples from Saskatchewan, Canada, were analysed on an automated 4-spectrometer Cameca Camebax MBX electron probe by the wavelength dispersive X-ray method (WDS) at Carleton University in Ottawa, Canada. At the University of Tasmania, operating conditions were: 15 kV accelerating voltage, 10 nano-amperes (nA) beam current. Specimens were analysed using a defocused electron beam of 5 µm in diameter. Counting times for analysis and background varied from 10 to 30 s. Background positions were carefully selected to avoid instances of peak overlap. Raw X-ray data were converted to elemental wt percent by the Armstrong/Love-Scott matrix correction algorithm. A suite of well-characterized natural and synthetic minerals and compounds were used as calibration standards. Precision is <1% relative for majors and 1–3% relative for minors. At the University of Ottawa, operating conditions were: 15 kV accelerating voltage, 20 nano-amperes (nA) beam current. Specimens were analysed using a rastered electron beam 5 × 5 to 10 × 10 µm in size. Counting times were 15 to 40 s or 40,000 accumulated counts. Background measurements were made at 50 percent peak counting time on each side of the analysed peak. Background positions were carefully selected to avoid instances of peak overlap. Raw X-ray data were converted to elemental wt percent by the Cameca PAP matrix correction program. A suite of well-characterized natural and synthetic minerals and compounds were used as calibration standards. Analyses are accurate to 1 to 2 percent relative for major elements (>10 wt%), 3 to 5 percent relative for minor elements (>0.5 to <5.0 wt%). Spectral reflectance was measured using an Analytical Spectral Devices (ASD) Ter- raSpec 4 for the Lemarchant samples at the Centre for Ore Deposits and Earth Sciences (CODES), University of Tasmania, Australia, and a Portable Infrared Mineral Analyzer II (PIMA) for the Athabasca Basin samples at Queen’s University, Canada. To enable comparison with spectral data where the spot size is approximately 1.5 cm in diameter compared to 5–10 µm for EPMA, an average of at least 6 EPMA analyses from across the thin section was used in order to account for mineral chemistry heterogeneity in each sample. The ASD TerraSpec 4 standard spectrometer has a spectral range of 350 to 2500 nm across VNIR/SWIR light with a spectral resolution of 10 nm, whereas the PIMA II has a spectral range of 1200 to 2500 nm across NIR/SWIR light with a spectral resolution of 4 nm. Published datasets from Scott and Yang [7] (102 white mica), Scott et al. [8] (78 chlorite), and Buschette and Piercey [10] (10 white mica and 10 chlorite) were used to complement the datasets used in this study. Minerals 2021, 11, x FOR PEER REVIEW 8 of 16 Minerals 2021, 11, 471 7 of 16
respectively. They plot near the muscovite pole in the Si versus Al(tot) plot (Figure 3C) and Mineral formulas for white mica were calculated on the basis of 11 oxygen (22 charge) plot closer to the muscovite–illite trend compare to Type 1 and 2 white micas. We classify following the method described in Rieder et al. [12] and 14 oxygen (28 charge) for chlorite them as Al-rich illite due to their highly deficient nature. following the method described in Zane and Weiss [15]. A mineral name was obtained based on atoms per formula units (apfu) as per the descriptions in the mineralogy back- ground section above.
5. Results Minerals 2021, 11, x FOR PEER REVIEW 8 of 16 5.1. White Mica Chemistry Three types of white mica were identified based on interlayer deficiency (I), Al, Si, 2+ 3+ R (oct),R (oct) andrespectively. Ca (Figure They3). Allplot whitenear the micas muscovite in this pole studyin the Si are versus true Al micas(tot) plot with(Figure more 3C) and than 50% monovalentplot Icloser cations. to the muscovite–illite trend compare to Type 1 and 2 white micas. We classify them as Al-rich illite due to their highly deficient nature.
2+ 3+ FFigureigure 3. White 3. White mica micascatter scatter plots for plots (A) Al for(tot) vs. (A sum) Al I (interlayervs. sum cations), I (interlayer (B) R (VI) cations),vs. R (VI), (C (B) Si)R vs.2+ Al(tot),vs. (D)R Al3+(tot) , vs. Na, (E) Ca vs. K, and (F) Si vs. Ca showing the three types(tot) of whiteFigure mica used 3. White in this study. mica Symbols scatter are:plots(VI) • (this for study), (A) Al(VI) (tot) vs. sum I (interlayer cations), (B) R2+(VI) vs. R3+(VI), (C) Si vs. Al(tot), (D) Al(tot) +( C(Scott) Si and vs. Yang, Al(tot) [7,(]) Dand) Al(tot) (Buschettevs. Na, and (E Piercey,) Ca vs. [10] K,). and (Fvs.) Si Na, vs. (E Ca) Ca showing vs. K, and the ( threeF) Si vs. types Ca showing of white the three types of white mica used in this study. Symbols are: • (this study), mica used in this study. Symbols are: • (this study), + (Scott (Scott and and Yang, Yang, [7 [7]])) and and (Buschette(Buschette andand Piercey, [10]). Piercey, [10]). Minerals 2021, 11, x FOR PEER REVIEW 8 of 16
respectively. They plot near the muscovite pole in the Si versus Al(tot) plot (Figure 3C) and plot closer to the muscovite–illite trend compare to Type 1 and 2 white micas. We classify them as Al-rich illite due to their highly deficient nature.
Minerals 2021, 11, 471 8 of 16
Type 1 are not interlayer deficient (sum I > 0.95) and contain high K and variable Si, Al, 2+ 3+ 3+ R (oct), and R (oct). Si records inverse abundance relationships with Al and R (oct), and a 2+ positive abundance relationship with R (oct). In contrast, Al records an inverse abundance 2+ 3+ relationship with R (oct) and a positive relationship with R (oct) (Figures3A–C and4). Type 1 white mica is part of the muscovite series and records a trend between muscovite (Si Minerals 2021, 11, x FOR PEER REVIEW=3.05) and muscovitic phengite (Si = 3.3), with the majority of analyses being muscovite9 of 16
(Si < 3.25) (Figure3C). Na shows a positive correlation with Al (Figure3D).
Figure 4. SumSum of interlayer cations (sum I) in white mica versus (A) K, (B) Na, (C) Ca, ((D)) Fe,Fe, ((E)) Mg,Mg, ((F)) AlAl(VI),,( (GG)) Ti, Ti, ( (HH)) Figure 3. White mica scatter plots for (A) Al(tot) vs. sum I (interlayer cations), (B) R2+(VI) vs. R3+(VI), (C) Si vs. Al(VI)(tot), (D) Al(tot) Si, and (I) Al(IV) showing the elemental variability in apfu of the white micas used in this study. Symbols are: • (this study), vs.Si, Na, and (E (I)) Ca Al (IV)vs. showingK, and (F the) Si elementalvs. Ca showing variability the three in apfu types of theof white white mica micas used used in inthis this study. study. Symbols Symbols are: are: • •(this(this study), study), ++ (Scott (Scott and and Yang, Yang, [7 [[7]7]])) )and andand (Buschette (Buschette(Buschette and and and Piercey, Piercey, Piercey, [ 10[10] [10]]).).).
Type 2 are interlayer deficient with sum I between 0.68 and 0.95, with the majority 2+ 3+ above 0.75 (Figures3A and4). They contain variable Si, Al, R (oct), and R (oct), with similar abundance relationships to Type 1 white mica, but are enriched in Si and Mg relative 3+ 3+ to Type 1 at similar Al(VI) or R (oct) values. They are also enriched in Si and R (oct) relative 2+ to R (oct) (Figure3B) and depleted in K and Na (Figure3D,E) and record a different trend relative to Type 1 in an Al(tot) vs. Na scatter plot (Figure3D). Analyses with the highest Ca record the lowest K and highest Si (Figure3E,F), suggesting that Ca substitutes for K in the +4 3+ + interlayer position and complements the Si tet ↔ Al tet + K int substitution. Type 2 are technically part of the illite series due to their I site deficiency of less than 0.95. However, they form a continuum between muscovite and phengite, similar to Type 1 white mica but with greater Si, and do not show a progression towards the ideal illite composition (Figure3C). Overall, they are enriched in Ca, Mg, Al (VI), sum M, and Si, and depleted in K, Na, sum I, Fe, Ti, Al(IV) and Al(tot) relative to Type 1 white mica (Figure5).
Figure 5. White mica split probability plots for (A) K, (B)Na, (C) Ca, (D) Fe, (E) Mg, (F) Al(VI), (G) Ti, (H) Si, and (I) Al(tot) showing the elemental variability in apfu in the white mica used in this study. Symbols are: • (this study), + (Scott and Yang, [7]) and (Buschette and Piercey, [10]). Minerals 2021, 11, x FOR PEER REVIEW 9 of 16
Minerals 2021, 11, x FOR PEER REVIEW 8 of 16
respectively. They plot near the muscovite pole in the Si versus Al(tot) plot (Figure 3C) and plot closer to the muscovite–illite trend compare to Type 1 and 2 white micas. We classify them as Al-rich illite due to their highly deficient nature.
Minerals 2021, 11, 471 9 of 16 Figure 4. Sum of interlayer cations (sum I) in white mica versus (A) K, (B) Na, (C) Ca, (D) Fe, (E) Mg, (F) Al(VI), (G) Ti, (H) Si, and (I) Al(IV) showing the elemental variability in apfu of the white micas used in this study. Symbols are: • (this study), + (Scott and Yang, [7]) and (Buschette and Piercey, [10]).
FigureFigure 5. 5. WhiteWhite mica mica split split probability probability plots plots for for ( (AA)) K, K, ( (BB)Na,)Na, ( (CC)) Ca, Ca, ( (DD)) Fe, Fe, ( (EE)) Mg Mg,, ( (FF)) Al(VI), Al(VI), ( (GG)) Ti, Ti, ( (HH)) Si Si,, and and ( (II)) Al(tot) Al(tot) Figure 3. White mica scatter plots for (A) Al(tot) vs. sum I (interlayer cations), (B) R2+(VI) vs. R3+(VI), (C) Si vs. Al(tot), (D) Al(tot) showing the elemental variability in apfu in the white mica used in this study. Symbols are: • (this study), + (Scott and vs. Na, (E) Cashowing vs. K, and the (F) elemental Si vs. Ca showing variability the in three apfu types in the of white white mica mica used used inin thisthis study.study. Symbols are: •• (this study), study), + (Scott and Yang, [7]) and (Buschette and Piercey, [10]). + (Scott and Yang,Yang, [7 [7]])) and and (Buschette(Buschette and and Piercey, Piercey, [ 10[10]]).).
Type 3 are highly deficient white micas with sum I between 0.52 and 0.67 (Figures3A and4). They have the highest Al content with total Al varying between 2.82 and 2.99 apfu and lowest K and Si content ranging between 0.45 and 0.53 and 2.99 and 3.31 apfu, respectively. They plot near the muscovite pole in the Si versus Al(tot) plot (Figure3C) and plot closer to the muscovite–illite trend compare to Type 1 and 2 white micas. We classify them as Al-rich illite due to their highly deficient nature.
5.2. White Mica SWIR Reflectance Type 1 and 2 white micas show similar but distinct relationships with the wavelength position of the 2200 nm absorption feature, except for the Athabasca Basin samples (open black circles; Figure6), which do not show any correlation with 2200 W. Overall, Mg, Al (VI), Si and Al(tot) show different slopes in their correlation with 2200 W. However, differences in the trends of each white mica type are observed. In general, Type 1 is associated with steeper slopes and higher correlation coefficients with 2200 W than Type 2 white mica. Interlayer cations do not correlate with 2200 W (Figure6G–I), except possibly the Type 2 Athabasca Basin samples (Figure6I) and Na in Type 1 mica, where a negative correlation exists with 2200 W up to ~2200 nm. However, when Na is greater than 0.15 apfu, this correlation ceases and the 2200 W value remains constant near 2200 nm (Figure6G). This correlation does not appear to be as consistent in Type 2 and 3 white micas. No correlation is observed with K and Ca. The 2200 W values of Type 3 Al-rich illite show a weak correlation with Mg and Na, with the other elements not correlating with 2200 W. No correlation is observed with Fe in the three types of white mica. This may be due to variable oxidation states of Fe or due to erroneous spectral analysis arising from the different spot size between ASD TerraSpec (~1.5 cm) and EPMA (~15 µm), which may introduce mineral mixes. Minerals 2021, 11, x FOR PEER REVIEW 8 of 16
respectively. They plot near the muscovite pole in the Si versus Al(tot) plot (Figure 3C) and plot closer to the muscovite–illite trend compare to Type 1 and 2 white micas. We classify them as Al-rich illite due to their highly deficient nature.
Minerals 2021, 11, x FOR PEER REVIEW 10 of 16
5.2. White Mica SWIR Reflectance Type 1 and 2 white micas show similar but distinct relationships with the wavelength position of the 2200 nm absorption feature, except for the Athabasca Basin samples (open black circles; Figure 6), which do not show any correlation with 2200W. Overall, Mg, Al(VI), Si and Al(tot) show different slopes in their correlation with 2200W. However, differences in the trends of each white mica type are observed. In general, Type 1 is associated with steeper slopes and higher correlation coefficients with 2200W than Type 2 white mica. Interlayer cations do not correlate with 2200W (Figure 6G–I), except possibly the Type 2 Athabasca Basin samples (Figure 6I) and Na in Type 1 mica, where a negative correlation exists with 2200W up to ~2200 nm. However, when Na is greater than 0.15 Minerals 2021, 11, 471 apfu, this correlation ceases and the 2200W value remains constant near 2200 nm (Fig10ure of 16 6G). This correlation does not appear to be as consistent in Type 2 and 3 white micas. No correlation is observed with K and Ca.
Minerals 2021, 11, x FOR PEER REVIEW 2 of 16
of white mica is strongly influenced by the Tschermak substitution between the octahe- dral (Al(VI), Mg, Fe) and tetrahedral (Si, Al(VI)) sites, whereas the spectral signature of chlo- rite is influence by Mg and Fe substitutions in the octahedral site [5–8]. The correlation between white mica and chlorite chemistry and spectral response is generally high for each specific area studied (e.g., [7,8]), but greatly decreases when data from different areas are combined and compared, suggesting that unknown mineralogical or chemical param- eters also control the spectral signature of white mica and chlorite. Understanding these parameters is critical, as there are often specific mica and chlorite species associated with
mineralization (e.g., [10,11]), such that improved interpretation of spectral data may lead FigureFigure 6. 6.Relationship Relationship between between wavelengthwavelength of the 2200 nm absorption absorption feature feature (2200 (2200 W) W) and and content content of ofmajor major cations cations in in to the identification of newFigure exploration 3. White targetmica scatters that plotsmay for have (A) beenAl(tot) vs.missed sum Iin (interlayer historical cations), (B) R2+(VI) vs. R3+(VI), (C) Si vs. Al(tot), (D) Al(tot) Type 1 (red), Type 2 (blue) and Type 3 (green) white micas. (A) Al(VI), (B) Mg, (C) Fe, (D) Ti, (E) Si, (F) total Al, (G) Na, (H) Type 1 (red), Type 2 (blue) and Type 3 (green) whitevs. Na, micas. (E) Ca (A vs.) Al K(VI), and,( B(F)) Mg, Si vs. (C Ca) Fe, showing (D) Ti, the (E) three Si, (F )types total of Al, white (G) Na,mica used in this study. Symbols are: • (this study), K, and (I) sum I. Symbolsdatasets. are: • (this study), + (Scott and Yang, [7]) and (Buschette and Piercey, [10]). Additionally, (H) K, and (I) sum I. Symbols are: • (this study),+ + (Scott (Scott and and Yang, Yang, [7 [7]])) and and (Buschette(Buschette and and Piercey,Piercey, [ 10[10]]).). Additionally, shown are trends of TypeThis 1, Typestudy 2 andaims Type to review3 (excluding the mineralogyAthabasca Basin and samples; chemistry open of black different circles) white and their mica correla- and showntion coefficient are trendss. of Achlorite Type separate 1, Typetypes trend 2 andtois showninvestigate Type 3for (excluding the what Athabasca mineralogical Athabasca Basin samples. Basin and samples; chemical open changes black circles) are andresponsible their correlation for coefficients. A separateshifts trend in the is shown shortwave for the infrared Athabasca (SWIR) Basin samples.spectroscopic absorption features. It uses original and publicly Theavailable 2200W databases values of to Type demonstrate 3 Al-rich illitethat showthe spectral a weak signature correlation of with white Mg mica and Na, 5.3. Chlorite Chemistry and chloritewith is the more other complex elements than not previously correlating documented.with 2200W. TwoNo correlation types of chlorite is observed were with identified Fe in the based three on types Si, Mg,of white Fe and mica. Al This (Figure may7 ).be Type due 1 2. Mineralogychloritesto variable Background correspond oxidation to states trioctahedral of Fe or chloritedue to erroneous of Zane and spectral Weiss analysis [15], and arising define afrom trend the that variesdifferent from spot Si-deficient size between sheridanite ASD TerraSpec (Mg# > (~1.5 0.5) andcm) ripidoliteand EPMA (Mg# (~15 < µm), 0.5) towhich clinochlore may 2.1. White Mica Group (Mg#introduce > 0.5) mineral and chamosite mixes. (Mg# < 0.5) (Figure7). Chlorites near the sheridanite/ripodilite White micas are phyllosilicates in which2+ the unit structure consists of one octahedral end-member contain low Si and R cations and high Al(VI) (Figure7). As the chemical com- sheet betweenposition two approaches opposing thetetrahedral clinochlore/chamosite sheets. These are end-member, separated from they adjacent become progressivelylayers 2+ by planesenriched of non-hydrated in Si and interlayer R cations cations and depleted with no less in Al than(VI), 0.6 reflecting cations per Tschermak formula substitutionunit 3+ 3+ 4+ [12]. If thebetween interlayer the octahedralcations are and between tetrahedral 0.6 and layers 0.85, inthe the white chlorite mica (Al is characterizedtet + Al oct ↔ asSi tet + 2+ 2+ an interlayer-deficientMg oct or Fe mica.oct). If the interlayer cations are greater than 0.95, they are of the muscovite-celadoniteType 2 chlorites series, and correspond if the interlayer to dioctahedral cations are Type near II chlorite0.65 they of are Zane of the and illite Weiss [15] 3+ series. Theand simplified show a progressive formula is increasewritten as in I Si,0.6–1.0 R Mand2–3 ❏ from 1–0 a T Type4 O10 1A chlorite2, where: pole with an Mg number of approximately 0.8 towards the Si-rich sudoite composition with an end-member I: interlayered cation that is K, Na and Ca, and less commonly Cs, NH4, Rb and Ba at 3.5 apfu Si, 3.2 apfu R3+ and 1.5 apfu (Figure7). This progressive increase in Si, R 3+ M: octahedral cation that is Fe (di- or trivalent), Mg, Al, Ti and Li, and less commonly and is associated with a progressive decrease in R2+, and reflects the substitution of Mn (di- or trivalent), Zn, Cr and V 3R2+ ↔ + 2Al and/or 3R2+ ↔ 2 Si4+ + 2 . ❏: vacancy in the octahedral(VI) sheet T: tetrahedral5.4. Chlorite cation SWIR that Reflectance is Si, Al and Fe (trivalent), and less commonly Be and B A: anion that is OH, F and Cl, and less commonly O (oxy-micas) and S. The two types of chlorite show similar relationships with the wavelength position of Whitethe 2250micas nm can absorption be divided feature, into true except micas for if Mg,more Al than(oct), 50% and octahedralof I cations vacancy are mono- (Figure8). valent, andIn Type brittle 1 chlorite, micas if Fe,more Al (VI)thanand 50% Si of correlate I cations positively are divalent. with White 2250 W, micas whereas are dioc- Mg and Mg tahedralnumber if the M correlate site contains negatively. less than In Type2.5 cations 2 chlorite, and Mgtrioctahedral and Fe show if it acontains positive more relationship than 2.5with cations 2250 [12]. W, whereas Al(VI) shows an inverse relationship. In both types, Fe and Mg Muscovite-celadonite series contain true dioctahedral mica with I greater than 0.95. They vary between a K-rich muscovite end-member with a general formula of KAl2❏(Si3Al)O10(OH)2 and a Fe-Mg-rich celadonite end-member with a general formula of KFe3+(Mg, Fe2+)❏Si4O10(OH)2. The mica is classified as muscovite when Al# (VIAl/(VIAl + Fetot + Mg) is greater than 0.5 and celadonite when it is less than 0.5 [12,13]. Although not a recognized mineral, in this study we identify the mica as phengite when the Si:Al ratio in the tetrahedral layer is greater than 3.25, following the classification of Deer et al. [14]. The main substitutions in the muscovite-celadonite series are the Tschermak substitution (Al3+tet + Al3+oct ↔ Si4+tet + Mg2+oct or Fe2+oct) in the octahedral and tetrahedral sites, and the Na+int ↔ K+int substitution in the interlayer plane. Tischendorf et al. [12] showed that the Na+int ↔ K+int substitution does not constitute a solid-solution with a paragonite end-mem- ber (NaAl2❏(Si3Al)O10(OH)2). They identified an immiscibility gap between approxi- mately 0.4 to 0.6 Na atoms per formula unit and argue that this immiscibility gap results from the large difference in ionic radius between Na+ (0.95 Å) and K+ (1.33 Å). Minerals 2021, 11, x FOR PEER REVIEW 8 of 16
respectively. They plot near the muscovite pole in the Si versus Al(tot) plot (Figure 3C) and plot closer to the muscovite–illite trend compare to Type 1 and 2 white micas. We classify them as Al-rich illite due to their highly deficient nature.
Minerals 2021, 11, x FOR PEER REVIEW 11 of 16
5.3. Chlorite Chemistry Two types of chlorite were identified based on Si, Mg, Fe and Al (Figure 7). Type 1 chlorites correspond to trioctahedral chlorite of Zane and Weiss [15], and define a trend Minerals 2021, 11, x FOR PEER REVIEW 2 of 16 that varies from Si-deficient sheridanite (Mg# > 0.5) and ripidolite (Mg# < 0.5) to clino- Minerals 2021, 11, 471 chlore (Mg# > 0.5) and chamosite (Mg# < 0.5) (Figure 7). Chlorites near the sheridanite/rip-11 of 16 odilite end-member contain low Si and R2+ cations and high Al(VI) (Figure 7). As the chem- of white mica is strongly influenced by the icalTschermak composition substitution approaches between the clinochlore/chamositethe octahe- end-member, they become pro- dral (Al(VI), Mg, Fe) and tetrahedral (Si, Al(VI))gressively sites, whereas enriched the spectralin Si and signature R2+ cations of chlo-and depleted in Al(VI), reflecting Tschermak sub- rite is influence by Mg and Fe substitutionsstitutionnumber in the octahedral recordbetween similar the site octahedral relationships,[5–8]. The and correlation wherein tetrahedral an increase layers in in the Fe correlateschlorite (Al with3+tet longer+ Al3+oct 2250 ↔ between white mica and chlorite chemistrySiW an4+ absorptiontetd + spectral Mg2+oct or features.response Fe2+oct). is generally high for each specific area studied (e.g., [7,8]), but greatly decreases when data from different areas are combined and compared, suggesting that unknown mineralogical or chemical param- eters also control the spectral signature of white mica and chlorite. Understanding these parameters is critical, as there are often specific mica and chlorite species associated with mineralization (e.g., [10,11]), such that improved interpretation of spectral data may lead to the identification of new exploration targets that may have been missed in historical datasets. This study aims to review the mineralogy and chemistry of different white mica and chlorite types to investigate what mineralogical and chemical changes are responsible for shifts in the shortwave infrared (SWIR) spectroscopic absorption features. It uses original and publicly available databases to demonstrate that the spectral signature of white mica and chlorite is more complex than previously documented.
2. Mineralogy Background 2.1. White Mica Group White micas are phyllosilicates in which the unit structure consists of one octahedral sheet between two opposing tetrahedral sheets. These are separated from adjacent layers by planes of non-hydrated interlayer cations with no less than 0.6 cations per formula unit [12]. If the interlayer cations are between 0.6 and 0.85, the white mica is characterized as an interlayer-deficient mica. If the interlayer cations are greater than 0.95, they are of the muscovite-celadonite series, and if the interlayer cations are near 0.65 they are of the illite Figure 3. White mica scatter plots for (A) Al(tot) vs. sum I (interlayer cations), (B) R2+(VI) vs. R3+(VI), (C) Si vs. Al(tot), (D) Al(tot) 2+ 2+ series. The simplifiedMineralsFigure formula 2021 7. 7. ,( A11(A), )isTernaryx TernaryFOR written PEER diagram diagram asREVIEW I0.6–1.0 of of AlM Al 2–3+ ❏ + (octahedral❏ (octahedral 1–0vs. T vacancy) 4Na, O vacancy)-Mg–Fe,10 ( EA) 2Ca,- Mgwhere: vs.–Fe, K, and and(BF)) Si (B vs.) Si RCa vs. (VI)showing Rshowing(VI) theshowing the three two types thetypes two of of white typeschlorite12 micaof of 16 used in this study. Symbols are: • (this study),
usedchlorite in this used study. in this Symbols study. Symbols are: • (this are: study),• (this + study), (Scott+ (Scott et + (Scottandal. [ 8Yang,]) et and al. [7 [8]])) and(Buschette and (Buschette(Buschette and Piercey, and and Piercey, Piercey, [10]). [ 10[10]]).). I: interlayered cation that is K, Na and Ca, and less commonly Cs, NH4, Rb and Ba M: octahedral cation that is Fe (di- or trivalent),Type Mg, 2 chloritesAl, Ti and correspond Li, and less to commonly dioctahedral Type II chlorite of Zane and Weiss [15] Mn (di- or trivalent), Zn, Cr and V and show a progressive increase in Si, R3+ and ❏ from a Type 1 chlorite pole with an Mg ❏: vacancy in the octahedral sheet number of approximately 0.8 towards the Si-rich sudoite composition with an end-mem- T: tetrahedral cation that is Si, Al and Fe (trivalent), and less commonly Be and B ber at 3.5 apfu Si, 3.2 apfu R3+ and 1.5 apfu ❏ (Figure 7). This progressive increase in Si, A: anion that is OH, F and Cl, and less commonly O (oxy-micas) and S. R3+ and ❏ is associated with a progressive decrease in R2+, and reflects the substitution of White micas can be divided into true micas3R2+ ↔ if ❏more + 2Al than(VI) and/or50% of 3RI cations2+ ↔ 2❏ are Si 4+mono- + 2❏. valent, and brittle micas if more than 50% of I cations are divalent. White micas are dioc- tahedral if the M site contains less than 2.55.4. cations Chlorite and SWIR trioctahedral Reflectance if it contains more than 2.5 cations [12]. The two types of chlorite show similar relationships with the wavelength position of Muscovite-celadonite series contain true thedioctahedral 2250 nm absorption mica with feature, I greater except than for 0.95. Mg, Al(oct), and octahedral vacancy (Figure 8). They vary between a K-rich muscovite Inend-member Type 1 chlorite, with Fe, aAl general(VI) and Siformula correlate of positively with 2250W, whereas Mg and Mg KAl2❏(Si3Al)O10(OH)2 and a Fe-Mg-rich celadonitenumber end-membercorrelate negatively. with a generalIn Type formula 2 chlorite, Mg and Fe show a positive relationship 3+ VI VI of KFe (Mg, Fe2+)❏Si4O10(OH)2. The mica is withclassified 2250W as, muscovitewhereas Al when(VI) show Al#s (anAl/( inverseAl + relationship. In both types, Fe and Mg num- Fetot + Mg) is greater than 0.5 and celadoniteber when record it is similarless than relationships, 0.5 [12,13]. Although wherein annot increase in Fe correlates with longer 2250W a recognized mineral, in this study we identiabsorptionfy the mica features. as phengite when the Si:Al ratio in the tetrahedral layer is greater than 3.25, following the classification of Deer et al. [14]. The main substitutionsFigureFigure in 8.the (8A. (muscovite-celado–AF–)F Relationship) Relationship between betweennite series wavelength wavelength are the ofof Tschermak thethe 22502250 nm substitution absorption feature feature and and content content of of major major cations cations in inType Type 1 (Al3+tet + Al3+oct ↔ Si(light4+tet1 (light+ green:Mg green:2+oct clinochlore; or clinochlore; Fe2+oct) in dark darkthe green: octahedral green: chamosite; chamosite; and light tetrahedrallight blue: blue: sheridanite;sheridanite; sites, and dark dark the blue: blue: ripidolite) ripidolite) and and Type Type 2 (pink: 2 (pink: sudoite) sudoite) chlorite. Symbols are: • (this study), + (Scott et al. [8]) and (Buschette and Piercey, [10]). Na+int ↔ K+int substitutionchlorite. in Symbols the interlayer are: • (this plane. study), Tischendorf + (Scott et al. et [ 8al.]) and[12] showed(Buschette that and the Piercey, [10]). Na+int ↔ K+int substitution does not constitute a solid-solution with a paragonite end-mem- 6. Discussion ber (NaAl2❏(Si3Al)O10(OH)2). They identified6. Discussion an immiscibility gap between approxi- mately 0.4 to 0.6 Na atoms per formula unit 6.1.and6.1. Spectral Spectralargue that SignaturesSignatures this immiscibility of White Mica Mica gap results from the large difference in ionic radius between WhiteNaWhite+ (0.95 micasmicas Å) and in this K+ study(1.33 Ådisplay display). similar similar spectral spectral relationships relationships to tothose those observed observed in in previous studies (e.g., [5–7]). The 2200W correlates negatively with Al(VI), and positively previous studies (e.g., [5–7]). The 2200 W correlates negatively with Al(VI), and positively withwith Mg Mg andand SiSi (Figure(Figure6 6),), highlightinghighlighting the the importance importance of of the the Tschermak Tschermak substitution substitution in in influencinginfluencing thethe spectralspectral signature of of white white micas. micas. However, However, a new a new spectral spectral correlation correlation betweenbetween 22002200W W and the the sum sum of of interlayer interlayer cations cations is isidentified, identified, wherein wherein muscovite muscovite series series white micas (Type 1) are associated with steeper slopes with 2200W compared to inter- layer depleted white micas of the muscovitic illite series (Type 2) (Figures 6I and 9A–D). The slope of these trends is controlled by the amount of Al, Mg and Fe in the octahedral site. In white mica, the position of the 2200 nm absorption feature is strongly controlled by the interaction of Al with OH bonds in the mineral structure and has been referred to as the Al–OH feature (e.g., [5–7]). The exact position of the absorption feature is influenced by minor variations in Mg and Fe concentration. When Mg and Fe concentrations are very low, fundamental OH stretching vibrations are associated with several higher frequency ν(OHAl–Al) absorption features near 3621, 3641 and 3658 cm−1, which when coupled with librational 훿(OH)Al–Al near 915 cm−1 produce overtones with short wavelengths at 2205, 2195 and 2186 nm, respectively (Table 1; [25–27]). The 3621 cm−1 fundamental frequency appears to be the dominant ν(OHAl–Al) but it is not clear what chemical or structural min- eralogical parameters activate each centre [25–27]. As Mg and Fe concentrations increase, the position of the absorption feature is associated with an apparent shift to longer wave- lengths, due to lower frequencies of fundamental OH stretching vibrations near 3573 cm−1 for Fe (ν(OHAl–Fe)) compared to 3604 cm−1 for Mg (ν(OHAl–Mg)). When coupled with a libra- tional Al–Al frequency (훿(OH)Al–Al) near 915 cm−1, this results in overtones near 2228 nm for Fe-enriched white mica and near 2213 nm for Mg-enriched white mica (Table 1; [25– 27]). Together, this suggests that samples with higher concentrations of Fe relative to Mg are expected to be associated with longer 2200 nm absorption features and will produce a steeper correlation with 2200W over the expected range of 2200W, as the maximum pre- dicted variability of 2200W in a white mica suite with Al-rich and Fe-rich end members is 42 nm (2186 nm for Al-rich vs. 2228 nm for Fe-rich) compared to 27 nm (2186 nm for Al- rich vs. 2213 nm for Mg-rich) for a mica suite with Al-rich and Mg-rich end members. Minerals 2021, 11, 471 12 of 16
white micas (Type 1) are associated with steeper slopes with 2200 W compared to interlayer depleted white micas of the muscovitic illite series (Type 2) (Figures6I and9A–D). The slope of these trends is controlled by the amount of Al, Mg and Fe in the octahedral site. In white mica, the position of the 2200 nm absorption feature is strongly controlled by the interaction of Al with OH bonds in the mineral structure and has been referred to as the Al–OH feature (e.g., [5–7]). The exact position of the absorption feature is influenced by minor variations in Mg and Fe concentration. When Mg and Fe concentrations are very low, fundamental OH stretching vibrations are associated with several higher frequency −1 ν(OHAl–Al) absorption features near 3621, 3641 and 3658 cm , which when coupled with −1 librational δ(OH)Al–Al near 915 cm produce overtones with short wavelengths at 2205, 2195 and 2186 nm, respectively (Table1;[ 25–27]). The 3621 cm−1 fundamental frequency appears to be the dominant ν(OHAl–Al) but it is not clear what chemical or structural mineralogical parameters activate each centre [25–27]. As Mg and Fe concentrations increase, the position of the absorption feature is associated with an apparent shift to longer wavelengths, due to lower frequencies of fundamental OH stretching vibrations near −1 −1 3573 cm for Fe (ν(OHAl–Fe)) compared to 3604 cm for Mg (ν(OHAl–Mg)). When coupled −1 with a librational Al–Al frequency (δ(OH)Al–Al) near 915 cm , this results in overtones near 2228 nm for Fe-enriched white mica and near 2213 nm for Mg-enriched white mica (Table1;[ 25–27]). Together, this suggests that samples with higher concentrations of Fe relative to Mg are expected to be associated with longer 2200 nm absorption features and will produce a steeper correlation with 2200 W over the expected range of 2200 W, as the maximum predicted variability of 2200 W in a white mica suite with Al-rich and Fe-rich end members is 42 nm (2186 nm for Al-rich vs. 2228 nm for Fe-rich) compared to 27 nm
Minerals(2186 2021, 11 nm, x FOR for PEER Al-rich REVIEW vs. 2213 nm for Mg-rich) for a mica suite with Al-rich and Mg-rich13 of 16
end members.
FigureFigure 9. Summary 9. Summary of features of features that influence that influencethe spectral theresponse spectral in white response mica. (A in–C) white Si (apfu) mica. vs. Al ((tot)A– (apfu)C) Si showing (apfu) vs. chemical variations due to Tschermak and interlayer deficiency substitutions as well as the poles for muscovite, phengite andAl (tot)illite,(apfu) with d showingata points colo chemicalured by variations (A) sum I, ( dueB) 2200 to TschermakW, and (C) Mg and (apfu). interlayer (D) 2200W deficiency vs. Mg (apfu) substitutions showing a as decreasewell as in the slope poles with for decreasing muscovite, sum of phengite interlayer andcations illite, (Sum with I). Symbols data points are: • (this coloured study) and by (+A (Scott) sum and I, Yang, (B) 2200 [7]). W, and (C) Mg (apfu). (D) 2200 W vs. Mg (apfu) showing a decrease in slope with decreasing sum of Type 1 white micas have on average greater abundances of Fe compared to Type 2 interlayer cations (Sumwhite I). Symbolsmicas (Fig are:ure 5•D(this). As study)interlayer and deficiency + (Scott andincreases Yang, towards [7]). Type 2 white mica, Mg, Al(VI) and Si slopes progressively decrease with decreasing sum I (Figures 6A,B,E and 9B). However, the data show some variability between Mg and 2200W that may be asso- ciated with chemical variability due to difference in spot size of the analyses (~1.5 cm for ASD TerraSpec versus ~15 µm for EPMA), or may be due to differences in oxidation state of Fe (Fe2+ is associated with longer 2200W than Fe3+; Table 1). Nevertheless, the observed trends suggest that sum I has an influence on the position of the 2200 nm absorption fea- ture (Figures 6I and 9). The decrease in sum I correlates with an increase in Si, Ca and Mg, a decrease in K and Na and does not correlate with Al(VI) (Figure 4A–C,E,F,H). This agrees with the substitutions observed in illites, where (1) Si+4(IV) ↔ Al3+(IV) + K+ int is associated with an increase in interlayer deficiency and (2) Ca2+int ↔ Mg2+(VI) or Fe2+(VI) corresponds to a progressive enrichment in Ca in the interlayer position [12]. In addition, the Tschermak substitution ((3) Al3+(IV) + Al3+(VI) ↔ Si4+(IV) + Mg2+(VI)) also appears to have an important role by increasing Si and Mg concentrations in the white mica, which results in a narrower 2200W range and shallower slopes with 2200W in interlayer deficient white mica (Figure 9A–D). Exceptions to this are the samples with short 2200W that are enriched in Ca from the Athabasca Basin, which do not plot along the Mg vs. 2200W trend and show a positive trend between sum I and 2200W, contrary to both types of white mica (e.g., black open circles in Figure 6I). They suggest that the presence of Ca in the interlayer position of the white mica structure may disrupt the crystal structure and activate different fundamental OHAl–Al stretching vibration centres (3641 cm−1 compared to the dominant 3621cm−1 Minerals 2021, 11, 471 13 of 16
Type 1 white micas have on average greater abundances of Fe compared to Type 2 white micas (Figure5D). As interlayer deficiency increases towards Type 2 white mica, Mg, Al(VI) and Si slopes progressively decrease with decreasing sum I (Figures6A,B,E and9B). However, the data show some variability between Mg and 2200 W that may be associated with chemical variability due to difference in spot size of the analyses (~1.5 cm for ASD TerraSpec versus ~15 µm for EPMA), or may be due to differences in oxidation state of Fe (Fe2+ is associated with longer 2200 W than Fe3+; Table1). Nevertheless, the observed trends suggest that sum I has an influence on the position of the 2200 nm absorption feature (Figures6I and9). The decrease in sum I correlates with an increase in Si, Ca and Mg, a decrease in K and Na and does not correlate with Al(VI) (Figure4A–C,E,F,H). This agrees +4 3+ + with the substitutions observed in illites, where (1) Si (IV) ↔ Al (IV) + K int is associated 2+ 2+ 2+ with an increase in interlayer deficiency and (2) Ca int ↔ Mg (VI) or Fe (VI) corresponds to a progressive enrichment in Ca in the interlayer position [12]. In addition, the Tschermak 3+ 3+ 4+ 2+ substitution ((3) Al (IV) + Al (VI) ↔ Si (IV) + Mg (VI)) also appears to have an important role by increasing Si and Mg concentrations in the white mica, which results in a nar- rower 2200 W range and shallower slopes with 2200 W in interlayer deficient white mica (Figure9A–D). Exceptions to this are the samples with short 2200 W that are enriched in Ca from the Athabasca Basin, which do not plot along the Mg vs. 2200 W trend and show a positive trend between sum I and 2200 W, contrary to both types of white mica (e.g., black open circles in Figure6I). They suggest that the presence of Ca in the interlayer position of the white mica structure may disrupt the crystal structure and activate different −1 fundamental OHAl–Al stretching vibration centres (3641 cm compared to the dominant −1 3621 cm ν(OHAl–Al)) that result in shorter wavelengths. More research is needed to elucidate the role of Ca in the spectral response of white mica. In summary, our data show a transition from steep to shallow slopes between 2200 W and Al(VI), Mg and Si with increasing interlayer deficiencies in the micas as they approach illite compositions. This implies that micas with different elemental chemistry may be associated with the same 2200 W values, and that shifts in 2200 W may be associated with micas with similar elemental chemistry (Figure9D). The net result of this effect is that the spectral signature of white mica creates data clouds instead of clear linear relationships between 2200 W and Al(VI), Mg and Si, and care is needed when interpreting spectral signatures using traditional interpretations (e.g., paragonite between 2180 and 2195 nm, muscovite between 2195 and 2215 nm, and phengite between 2215 and 2225 nm). Traditional interpretations should only be used in true muscovite series white mica with sum I > 0.95. It is recommended that a statistically representative subset of white mica from each area studied should be analysed using EPMA to obtain sum I and to facilitate more robust interpretations of real time spectral data.
6.2. Spectral Signature of Chlorite The chlorites analysed in this study record similar spectral relationships to those observed in previous studies (e.g., [6,7,34]) with trioctahedral chlorite recording negative correlations between 2250 W and Mg and Mg number, and positive correlations between 2250 W and Fe, Al(VI), and Si (Figure7). This study shows that dioctahedral chlorites (sudoite, cookeite and donbassite) have a different spectral response, recording negative correlations between 2250 W and Al(VI) and Mg number, and positive correlations between 2250 W and Mg and Fe. This shows that traditional spectral interpretation can be used on trioctahedral chlorite, regardless of the subtype, but cannot be used for dioctahedral chlorite. In addition, dioctahedral chlorite has a 2200 nm absorption feature that may affect correct mineral identification in samples containing more than one infrared respon- sive mineral. Fe and Mg# are the only shared feature between both types of chlorite and this suggests that the current interpretation nomenclature of Mg-rich, intermediate and Fe-rich chlorite should only be used for trioctahedral chlorite (clinochlore, sheridan- ite, chamosite and ripidolite). It is recommended that chlorites with spectral signatures Minerals 2021, 11, 471 14 of 16
between 2240–2249 nm should be identified as Fe-poor, as high Mg# or Mg-chlorite is erroneous when trioctahedral chlorite is present.
7. Conclusions This study demonstrates that the spectral signature of white mica is more complex than previously interpreted and shows that in addition to the Tschermak substitution, the sum of interlayer cations is also important in determining the wavelength position of the 2200 nm absorption feature. With an increase in interlayer deficiency, the slope of the trend between 2200 W and Mg, Al(VI) and Si decreases due to the progressive Mg and Si enrichment during illitization. This shift in the trend implies that white micas with similar chemistry but different interlayer deficiency can have variable 2200 W, and that white mica with variable chemistry may have similar 2200 W. It is recommended that spectral reflectance studies on white mica are accompanied by a subset of EPMA result on various mineral species. It is recommended that a minimum of 30 EPMA analyses of each mineral type be obtained for a statistically robust verification of spectral results and to provide guidance on the expected trends between 2200 W and the interlayer cation distributions. It is possible that EPMA analysis will also assist in separating background mineralogical variations from alteration processes, thus potentially aiding with vectoring within a given mineral deposit. Although not the scope of this study, no paragonite was identified even though 2200 W values shorter than 2195 nm were observed. We recommend that the spectral nomenclature of white mica be updated from paragonite to Na-bearing muscovite when the 2200 W absorption feature is between 2180 and 2195 nm. More studies are needed to fully characterize the spectral response of paragonite and white mica with 2200 W absorption features below 2195 nm. Lastly, we demonstrate that the current nomenclature of white mica should only be used for white mica with sum I > 0.95 and cannot be used for white mica with sum I < 0.95. The spectral signature of trioctahedral chlorite (clinochlore, sheridanite, chamosite and ripidolite) appears to record similar spectral relationships to those observed in previous studies (e.g., [6,7,34]). However, dioctahedral chlorite (sudoite, cookeite and donbassite) has a different spectral response with Mg increasing with 2250 W, which is the opposite of trioctahedral chlorite. This demonstrates that the traditional spectral interpretation should only be used for trioctahedral chlorite, regardless of the subtype. It is recommended that chlorites with spectral signatures between 2240 and 2249 nm should be identified as Fe-poor, as high Mg# or Mg-chlorite can be misleading when trioctahedral chlorite is present.
Author Contributions: Conceptualization, J.C.; methodology, J.C., S.J.P. and J.H.; formal analysis, J.C.; writing—original draft preparation, J.C.; writing—review and editing, J.C., S.J.P. and J.H. All authors have read and agreed to the published version of the manuscript. Funding: Financial support for this project was provided by the Centre for Ore Deposits and Earth Science (CODES). Stephen Piercey and his research program is also funded by an NSERC Discovery Grant. Acknowledgments: The manuscript benefited from constructive reviews by David Green from Mineral Resource Tasmania. Conflicts of Interest: The authors declare no conflict of interest.
References 1. Ridley, J. Ore Deposit Geology; Cambridge University Press: Cambridge, UK, 2013; 398p. 2. Robb, L. Introduction to Ore-Forming Processes; Blackwell Science, Ltd.: Oxford, UK, 2013; 373p. 3. Cloutier, J.; Piercey, S.J. Tracing Mineralogy and Alteration Intensity Using the Spectral Alteration Index and Depth Ratios at the Northwest Zone of the Lemarchant Volcanogenic Massive Sulfide Deposit, Newfoundland, Canada. Econ. Geol. 2020, 115, 1055–1078. [CrossRef] 4. Clark, R.N.; King, T.V.V.; Klejwa, M.; Swayze, G.A.; Vergo, N. High spectral resolution reflectance spectroscopy of minerals. J. Geophys. Res. 1990, 95, 12653–12680. [CrossRef] Minerals 2021, 11, 471 15 of 16
5. Post, J.L.; Noble, P.L. The near-infrared combination band frequencies of dioctahedral smectites, micas and illites. Clays Clay Miner. 1993, 41, 639–644. [CrossRef] 6. Pontual, S.; Merry, N.; Gamson, P. G-Mex Volume l: Spectral Interpretation Field Manual; Ausspec International, Pty. Ltd.: Arrowtown, New Zealand, 1997. 7. Scott, K.M.; Yang, K. Spectral Reflectance Studies of White Micas; CSIRO/AMIRA Project P435, Report 439R; Australian Mineral Industries Research Association, Ltd.: Parkville, Australia, 1997; 35p. 8. Scott, K.M.; Yang, K.; Huntington, J.F. The Application of Spectral Reflectance Studies of Chlorites in Exploration; CSIRO/AMIRA Project P435A, CSIRO Exploration and Mining Report P545R; North Ryde NSW: Sydney, Australia, 1998; 88p. 9. Herrmann, W.; Blake, M.; Doyle, M. Short Wavelength Infrared (SWIR) Spectral Analysis of Hydrothermal Alteration Zones Associated with Base Metal Sulfide Deposits at Rosebery and Western Tharsis, Tasmania, and Highway-Reward, Queensland. Econ. Geol. 2001, 96, 939–955. [CrossRef] 10. Buschette, M.J.; Piercey, S.J. Hydrothermal alteration and lithogeochemistry of the Boundary volcanogenic massive sulphide deposit, central Newfoundland, Canada. Can. J. Earth Sci. 2016, 53, 506–527. [CrossRef] 11. Ross, P.-S.; Bourke, A.; Schnitzler, N.; Conly, A. Exploration Vectors from Near Infrared Spectrometry near the McLeod Vol- canogenic Massive Sulfide Deposit, Matagami District, Québec. Econ. Geol. 2019, 114, 613–638. [CrossRef] 12. Rieder, M.; Cavazzini, G.; D’Yakonov, Y.S.; Frank-Kamenetskii, V.A.; Gottardi, G.; Guggenheim, S.; Koval, P.V.; Müller, G.; Neiva, A.M.R.; Radoslovich, E.W.; et al. Nomenclature of the micas. Miner. Mag. 1999, 63, 267–279. [CrossRef] 13. Tischendorf, G.; Förster, H.-J.; Gottesmann, B.; Rieder, M. True and brittle micas: Composition and solid-solution series. Miner. Mag. 2007, 71, 285–320. [CrossRef] 14. Deer, W.A.; Howie, R.A.; Zussman, J. An Introduction to the Rock-Forming Minerals; Longman Group Limited: London, UK, 1992; 696p. 15. Zane, A.; Weiss, Z. A procedure for classifying rock-forming chlorites based on microprobe data. Rend. Lincei Sci. Fis. Nat. 1998, 9, 51–56. [CrossRef] 16. Meunier, A. Argiles; Contemporary Publishing International—GB Science Publisher: Paris, France, 2003; 433p. 17. Wiewióra, A.; Weiss, Z. Crystallochemical classifications of phyllosilicates based on the unified system of projection of chemical composition: II. The chlorite group. Clay Miner. 1990, 25, 83–92. [CrossRef] 18. Farmer, V.C. The Infrared Spectra of Minerals; Mineralogical Society: London, UK, 1974; 539p. 19. Hunt, G.R. Spectral signatures of particulate minerals in the visible and near infrared. Geophysics 1977, 42, 501–513. [CrossRef] 20. Clark, R.N. Spectroscopy of Rocks and Minerals, and Principles of Spectroscopy. In Remote Sensing for the Earth Sciences: Manual of Remote Sensing, 3rd ed.; Rencz, A.M., Ed.; Wiley: New York, NY, USA, 1999; Chapter 1; Volume 3, 728p. 21. Swayze, G.A.; Clark, R.N.; Goetz, A.F.; Livo, K.E.; Breit, G.N.; Kruse, F.A.; Sutley, S.J.; Snee, L.W.; Lowers, H.A.; Post, J.L.; et al. Mapping Advanced Argillic Alteration at Cuprite, Nevada, Using Imaging Spectroscopy. Econ. Geol. 2014, 109, 1179–1221. [CrossRef] 22. Vedder, W.; McDonald, R.S. Vibrations of the OH Ions in Muscovite. J. Chem. Phys. 1963, 38, 1583–1590. [CrossRef] 23. Slonimskaya, M.V.; Besson, G.; Dainyak, L.G.; Tchoubar, C.; Drits, V.A. Interpretation of the IR spectra of celadonites and glauconites in the region of OH-stretching frequencies. Clay Miner. 1986, 21, 377–388. [CrossRef] 24. Robert, J.-L.; Kodama, H. Generalization of the correlations between hydroxyl-stretching wavenumbers and composition of micas in the system K2O–MgO–Al2O3–SiO2–H2O: A single model for trioctahedral and dioctahedral mica. Am. J. Sci. 1988, 288, 196–212. 25. Besson, G.; Drits, V.A. Refined relationships between chemical composition of dioctahedral fine-dispersed mica minerals and their infrared spectra in the OH stretching region. Part I. Identification of the stretching bands. Clays Clay Miner. 1997, 45, 158–169. [CrossRef] 26. Besson, G.; Drits, V.A. Refined relationship between chemical composition of dioctahedral fine-dispersed mica minerals and their infrared spectra in the OH stretching region. Part II. The main factors affecting OH vibration and quantitative analysis. Clays Clay Miner. 1997, 45, 170–183. [CrossRef] 27. Martínez-Alonso, S.; Rustad, J.R.; Goetz, A.F. Ab initio quantum mechanical modeling of infrared vibrational frequencies of the OH group in dioctahedral phyllosilicates. Part II: Main physical factors governing the OH vibrations. Am. Miner. 2002, 87, 1224–1234. [CrossRef] 28. Bookin, A.S.; Drits, V.A. Factors affecting orientation of OH vector in mica. Clays Clay Miner. 1982, 30, 415–421. [CrossRef] 29. Redhammer, G.J.; Beran, A.; Schneider, J.; Amthauer, G.; Lottermoser, W. Spectroscopic and structural properties of synthetic micas on the annite- siderophyllite binary: Synthesis, crystal structure refinement, Mössbauer, and infrared spectroscopy. Am. Mineral. 2000, 85, 449–465. [CrossRef] 30. Duke, E.F. Near infrared spectra of muscovite, Tschermak substitution, and metamorphic reaction progress: Implications for remote sensing. Geology 1994, 22, 621–624. [CrossRef] 31. Duke, E.F.; Lewis, R.S. Near infrared spectra of white mica in the Belt Supergroup and implications for metamorphism. Am. Miner. 2010, 95, 908–920. [CrossRef] 32. van Ruitenbeek, F.J.A.; Cudahy, T.; Hale, M.; van der Meer, F.D. Tracing fluid pathways in fossil hydrothermal systems with near-infrared spectroscopy. Geology 2005, 33, 597–600. [CrossRef] Minerals 2021, 11, 471 16 of 16
33. Kokaly, R.F.; Clark, R.N.; Swayze, G.A.; Livo, K.E.; Hoefen, T.M.; Pearson, N.C.; Wise, R.A.; Benzel, W.M.; Lowers, H.A.; Driscoll, R.L.; et al. USGS Spectral Library Version 7: U.S. Geological Survey Data Series 1035; U.S. Geological Survey: Reston, VA, USA, 2017; 61p. [CrossRef] 34. McLeod, R.L.; Gabell, A.R.; Green, A.A.; Gardavsky, V. Chlorite infrared spectral data as proximity indicators of volcanogenic massive sulphide mineralization. In Proceedings of the Pacific Rim Congress ’87, Gold Coast, QLD, Australia, 26–29 August 1987; pp. 321–324. 35. Zhang, G.; Wasyliuk, K.; Pan, Y. The characterization and quantitative analysis of clay minerals in the athabasca basin, saskatchewan: Application of shortwave infrared reflectance spectroscopy. Can. Miner. 2001, 39, 1347–1363. [CrossRef] 36. Percival, J.B.; Bosman, S.A.; Potter, E.G.; Peter, J.M.; Laudadio, A.B.; Abraham, A.C.; Shiley, D.A.; Sherry, C. Customized Spectral Libraries for Effective Mineral Exploration: Mining National Mineral Collections. Clays Clay Miner. 2018, 66, 297–314. [CrossRef]