Trace Element Characteristics of Tourmaline in Porphyry Cu Systems: Development and Application To Discrimination

Tourmaline supergroup minerals (TSM) are common accessory phases found in many geologic environments and deposit types (Slack 1996, Slack & Trumbull 2011), and their major, minor, and trace element chemistries have been widely used as indicators of the primary environments of formation, owing to the wide variation in elements that may be accommodated (e.g., Henry & Guidotti 1985, van Hinsberg & Schumacher 2011, van Hinsberg et al. 2011, Li et al. 2021a, Sciuba et al. 2021). The TSM are characterized by the general formula: XY₃Z₆T₆O₁₈(BO₃)₃V₃W, with X = Na⁺, K⁺, Ca²⁺, □ (vacancy); Y = Al³⁺, Cr³⁺, V³⁺, Fe^2+/3+, Mg²⁺, Mn²⁺, Li⁺, Ti⁴⁺; Z = Al³⁺, Cr³⁺, V³⁺, Fe^2+/3+, Mg²⁺; T = Si⁴⁺, Al³⁺, B³⁺; V = (OH)⁻, O²⁻; W = (OH)⁻, F⁻, O²⁻ as the common major constituents. The characters in the general formula represent groups of cations including ^[9]X, ^[6]Y, ^[6]Z, and ^[4]T. The characters V and W represent groups of anions at the ^[3]O3 and ^[3]O1 sites (Bosi 2018). Chemical and isotopic partitioning of TSM is shown to be influenced by sector growth (van Hinsberg et al. 2006, van Hinsberg & Marschall 2007, van Hinsberg & Schumacher 2007b, Marks et al. 2013). It has been suggested that the a-sector should be analyzed for major, minor, and trace elements, as it is considered neutral and is not affected by intra-crystalline fractionation (van Hinsberg et al. 2006, Marks et al. 2013). Sector zoning varies as a function of temperature and, due to the strongly anisotropic nature of TSM, also varies along the c-axis, resulting in differential surface charge on the associated crystal faces, which leads to sector growth (Henry et al. 1999, van Hinsberg et al. 2006, Marks et al. 2013).

In general, TSM are found as trace to minor mineral constituents in many hydrothermal environments, which has led to numerous studies into the occurrence, chemistry, and stable-isotope features to better understand the origin and development of a variety of mineralized systems, including volcanogenic massive sulfides (VMS) (Palmer & Slack 1989, Slack et al. 1993, Griffin et al. 1996), orogenic gold (Kalliomäki et al. 2017, Manéglia et al. 2018, Sciuba et al. 2021), polymetallic W-Sn-Cu deposits (Codeço et al. 2017, 2019, 2020), porphyry Cu-Au-Mo deposits (Frikken 2003, Frikken et al. 2005, Baksheev et al. 2010, 2011, 2012, 2017), and even emerald mineralization (Galbraith et al. 2009). Whereas several key mineral groups have been evaluated as potential indicators of mineralized porphyry systems (e.g., apatite, epidote, and chlorite group minerals; Wilkinson et al. 2015, Bouzari et al. 2016, Cooke et al. 2020b), no comprehensive studies involving TSM trace elements have similarly been conducted to date.

Studies of TSM trace-element chemistry from porphyry systems can be found in the literature (e.g., Iveson et al. 2016, Qiao et al. 2019, Li et al. 2021a), but such studies are generally restricted to the deposit-scale (Smith 1985, Warnaars et al. 1985, Koval et al. 1991, Lynch & Ortega 1997, Yavuz et al. 1999, Skewes et al. 2003, Frikken et al. 2005, Baksheev et al. 2010, 2011, Dill et al. 2012, Bačík et al. 2016, Baksheev et al. 2017, Testa et al. 2018, Baksheev et al. 2019), with only very few having been conducted at the broader scale (Baksheev et al. 2011, Beckett-Brown et al. 2021, Li et al. 2021a). It is within this context that the need for a comprehensive, broad-scale study of the trace element features present, along with their respective trends and ratios, has been deemed timely and critical considering the number of porphyry systems that contain TSM.

Tourmaline is a relatively common accessory mineral that develops in porphyry systems, not all of which are necessarily mineralized (Sillitoe & Sawkins 1971). The presence of TSM has been noted in some of the most prolific porphyry deposits worldwide, e.g., El Teniente (Chile), Los Bronces – Rio-Blanco district (Chile), Highland Valley Copper (British Columbia, Canada), Cadia district (eastern Australia), and Yerington district (Nevada, USA). A detailed understanding of which trace elements are present, the ranges in the concentrations of those elements, their internal chemical zoning relationships to other trace elements, and how these relate to the environments in which they occur are key factors to developing a holistic understanding of the genesis and evolution of the porphyry environment and, in turn, how that is reflected in TSM trace-element chemistry. The purpose of this study is to investigate and evaluate the applicability and efficacy of TSM trace-element chemistry to identify porphyry Cu (±Au, ±Mo) related TSM. Several key elements of porphyry TSM trace elements are addressed, including: (1) elements present, (2) concentrations, (3) their relationship with texture development (internal and external texture, spatial distribution, etc.), (4) inter/intra-deposit variations, and (5) determination of those important trace-element concentrations or ratios that have the highest efficacy of distinguishing TSM in mineralized porphyries from other geologic environments. This contribution will provide the first detailed documentation and comparison of the trace elements that occur in TSM associated with seven different mineralized porphyry systems and builds on the textural and major element characterization of Beckett-Brown et al. (2023), with an emphasis on using these as tools useful in the discrimination between mineralized porphyry systems and other geologic environments. This applies a similar approach to the longstanding application of TSM as a petrogenetic indicator in sedimentary rocks (Henry & Guidotti 1985). The longer-term goal is to develop a methodology that applies to a wide range of sample media, spanning those found in bedrock to those recovered from surficial sediment samples.

Approximately 70% of the world's Cu production is supplied from porphyry Cu deposits (Arndt et al. 2017). Porphyry Cu deposits are large magmatic-hydrothermal systems of high-volume and low-grade concentrations of primarily Cu, Mo, and Au (Sillitoe 2010). An evaluation of ∼470 porphyry systems worldwide (that report grade) shows that nearly 30% contain TSM (Fig. 1) (using the USGS porphyry Cu database, Singer et al. 2008, and data compiled in Beckett-Brown et al. 2023).

At the Salikvan Cu-Mo porphyry system (Turkey), it was demonstrated that those TSM associated with a TSM-dominant rock had higher Zn/(Zn+Pb) and Sr values than those with TSM occurring in veins (Yavuz et al. 1999). Concentrations of Zn, Pb, and Mn of TSM from the Cadia East Cu-Au porphyry (Australia) were used to highlight variations with alteration types and stratigraphic depth, with higher concentrations of Zn, Pb, Sb, and Mn being linked to the shallow and late-forming TSM (Cooke et al. 2017). An investigation of TSM from porphyry Cu systems in Mongolia and Russia reported that TSM contain the highest As contents among TSM from any mineralized systems (Koval et al. 1991).

To the author's knowledge, there exist three detailed publications on TSM trace elements in porphyry systems: the Margaret Cu-Mo porphyry system in Washington (Iveson et al. 2016), the Hadamiao porphyry Au deposit in China (Qiao et al. 2019), and the Bilihe porphyry Au deposit in China (Li et al. 2021a), of which key trace-element variations are outlined here. Low Li concentrations are observed among the three systems, ranging from 0.1 to 70.6 ppm. Low to intermediate V concentrations (2.87–643 ppm) reflect a dearth of mafic host rock contribution. Intermediate Cu concentrations (0.23–146 ppm) potentially reflect the involvement of Cu-rich hydrothermal fluids. Porphyry TSM generally contain low Zn concentrations (6.01–1029 ppm). The TSM grains contain intermediate concentrations of Sr (23.4–648 ppm) with primary hydrothermal grains containing the highest Sr concentrations. Porphyry TSM in general contain elevated Sn concentrations (<LOD–57.6 ppm), with the paragenetically later-forming grains containing higher Sn concentrations in general. The TSM grains contain high light rare earth element concentrations (e.g., La 0.15–86.8 ppm). Finally, concentrations of Pb in porphyry TSM are low (0.4–18.5 ppm). An investigation of TSM from the Bilihe porphyry Au deposit (Inner Mongolia, China; Li et al. 2021a) found that hydrothermal vein TSM show negative Eu anomalies, whereas TSM replacing magmatic plagioclase feldspar show positive Eu anomalies. Li et al. (2021a) also presented a series of plots potentially applicable in the discrimination between porphyry deposits, orogenic Au deposits, IOCG deposits, and TSM forming in unmineralized granitic rocks using a variety of elements including (Sn + Li) versus (Ni + V + Zn), (ΣREE + Y + Zr) versus (Ni + V + Zn) and V versus Zn. Unfortunately, a clear justification for the selection of elements or why they are beneficial for discriminating environment of formation of TSM was not presented. The challenge with using the summation of elements in a discrimination diagram is that one single element can skew the results and thus mask the contribution of the other components, making it, in many cases, unreliable. Many of the elements listed above can be strongly influenced by host rock composition, and as such, the addition of elements for discrimination would be ineffective. A more systematic approach is needed for the effective discrimination of TSM using trace elements.

Bedrock, hand, and drill core samples of TSM from Casino (Yukon, Canada), Coxheath (Nova Scotia, Canada), Donoso breccia (Chile), Highland Valley Copper (British Columbia, Canada), New Afton (British Columbia, Canada), Schaft Creek (British Columbia, Canada), and Woodjam (British Columbia, Canada) were examined (Table 1, Appendix 1¹). Images of all samples investigated can be found in Appendix 2 and detailed deposit descriptions can be found in Beckett-Brown et al. (2023), with a summary presented here.

The Casino deposit is a calc-alkaline Cu-Au-Mo porphyry deposit hosted in Late Cretaceous quartz-monzonite and associated breccias (Casselman & Brown 2017) from which six drill core samples were investigated. One sample from the core of the deposit within the potassic zone was investigated, along with two additional samples within the phyllic alteration zone that surrounds the potassic zone. Three final samples were collected toward the periphery of the deposit, consisting of a TSM breccia within argillic to phyllic zones of alteration at various depths. The Coxheath deposit is a calc-alkaline Cu-Mo-Au porphyry system hosted in a ca. 620 Ma hornblende-diorite (Lynch & Ortega 1997). Two Geological Survey of Canada (GSC) archive samples have been investigated, consisting of one surface sample from the sodic alteration zone which overlaps the potassic zone collected by A. Soregaroli in 1975 and another sample collected by R. Kirkham in 1976 from shaft #2. The Donoso breccia is the youngest breccia in the Rio-Blanco – Los Bronces copper district, where TSM breccias host a large volume of the mineralization (Warnaars et al. 1985). One grab sample from the Donoso breccia within the phyllic zone was investigated. The sample consists of a TSM cemented breccia with chalcopyrite, pyrite, and quartz. The Highland Valley Copper district is hosted in the Late Triassic calc-alkaline granodiorite, referred to as the Guichon Creek batholith (Byrne 2019). This GSC archive sample was collected by R. Kirkham from the Highmont E pit, 3^rd bench on the south wall in 1982. The sample consists of a mineralized and altered TSM breccia in a porphyry dike. The New Afton Cu-Au porphyry system is a silica-saturated alkalic deposit hosted in the Late Triassic Iron Mask Batholith (Kwong 1981). Two drill core samples have been investigated; one from the phyllic zone (524 m depth) and the other from the outer phyllic zone near the potassic zone (66 m depth). The first sample consists of a pervasively altered sample crosscut by TSM veins. The second sample consists of a TSM breccia with accessory pyrite and quartz. The calc-alkaline Schaft Creek Cu-Mo-Au deposit is hosted by Late Triassic volcanics and associated with porphyritic granodiorite dikes of the Hickman batholith (Scott et al. 2008). Ten samples from the paramount zone from three separate drill holes have been investigated: SCK-13-435 (302 and 306 m depth), SCK-15-443 (87, 104, 438, and 467 m depth), and T80CH139 (415, 427, 440, 446 m depth). The Woodjam Cu-Au deposits are hosted in calc-alkaline high K-monzonite porphyries (del Real et al. 2017). Two samples were investigated: one from a surface sample at the Takom deposit and an additional drill core sample from the Deerhorn deposit (157.6 m depth).

In the companion study to this (Beckett-Brown et al. 2023), the textural and major/minor element characteristics of TSM from the seven porphyry systems discussed above were described and evaluated. Tourmaline should be considered a significant phase in porphyry systems, as it occurs in nearly 30% (137 out of 470) of worldwide porphyry occurrences (Fig. 1). The physical characteristics of TSM from porphyry systems are remarkably consistent in terms of macro color, grain size, morphology, and color in transmitted light. In addition, the paragenetic position of TSM from porphyry systems is remarkably consistent in being one of the earliest hydrothermal phases. These features provide some information for the discrimination of TSM in mineralized porphyry systems compared to that in other geologic environments. Some generalities that exist in porphyry TSM include their black to dark brown color at the macro-scale, weak pleochroism (green, brown, and blue), and rare presence of inclusions, with zircon and rutile being the most common when present. Porphyry TSM form early in the paragenesis of these systems, denoted by their euhedral morphology and repeated observations of alteration and mineralization crosscutting individual grains. Major element chemistry of porphyry TSM is characterized by sodic group TSM, primarily dravite with some schorl compositions. The TSM have remarkably consistent Mg apfu values of ∼2.0 and almost exclusively follow the oxy-dravite–povondraite trend. In short, although there are some trends, the physical and major/minor element data alone cannot definitively distinguish TSM in porphyry systems from that in other geologic environments. The integration of these data with the findings of the trace element data presented here provides an effective method by which TSM in mineralized porphyry Cu systems can be distinguished.

Data for TSM from a variety of mineralized systems as well as unmineralized geologic settings have been compiled and compared for reference. Data for TSM in porphyry Cu ±Au, ±Mo deposits were collected in this study. Data for TSM in porphyry Au deposits have been compiled from Qiao et al. (2019) and Li et al. (2021a). Data for TSM forming in granites have been compiled from Drivenes et al. (2015), Kalliomäki et al. (2017), Zall et al. (2019), Zhao et al. (2019), Xiang et al. (2020), Aysal et al. (2021), Hong et al. (2021), and Zhao et al. (2021a), and unpublished analyses collected by the lead author from the Seagull Batholith, Yukon. Data for TSM in pegmatites have been compiled from Novák et al. (2011), Marks et al. (2013), Čopjaková et al. (2015), Zhao et al. (2019), Chakraborty & Upadhyay (2020), Long et al. (2021), Sciuba et al. (2021), Zhao et al. (2021c), and unpublished analyses collected by the lead author from the Usakos pegmatite (Namibia), as well as the Beryl pit (Ontario, Canada). Data for TSM in metamorphic rocks has been compiled from van Hinsberg (2011), Berryman et al. (2017), Kalliomäki et al. (2017), Wang et al. (2018), Hong et al. (2021), and unpublished analyses collected by the lead author from Jordan Falls (Nova Scotia, Canada). Data for TSM in orogenic Au deposits have been compiled from Jiang et al. (2002), Kalliomäki et al. (2017), Trumbull et al. (2018), Jin & Sui (2020), and Sciuba et al. (2021). Data for TSM in VMS deposits have been compiled from Slack et al. (1999), Klemme et al. (2011), and Sciuba et al. (2021). Data for TSM in polymetallic W ± Sn ± Cu deposits have been compiled from Jiang et al. (2004), Codeço et al. (2017), Launay et al. (2018), Harlaux et al. (2019), Carocci et al. (2021), Carr et al. (2020), Harlaux et al. (2020), Hu & Jiang (2020), Xiang et al. (2020), Ghosh et al. (2021), and Zhao et al. (2021b). Data for TSM from other localities were also compiled, including samples from a Be-U system (Zhu et al. 2021) and a Au-Co system (Tapio et al. 2021). A comparison of trace element generalizations for the main TSM environments and ore systems is presented in Table 2.

Trace element analyses of TSM grains were made in situ on polished thin sections using laser-ablation inductively coupled-plasma mass spectrometry (LA-ICP-MS) at Laurentian University. Samples were ablated using Resonetics Resolution M-50 coupled to a Thermo Electron X-Series II quadrupole ICP-MS. This employs a 193 nm ArF excimer laser which was operated at a rate of 8 Hz with a 30 μm beam and a line scan speed between 5 μm/s and 15 μm/s with a fluence of ∼3 J/cm². The total analyte sweep time was 300 ms. Beam size was optimized to be large enough for better count resolution but also small enough to collect enough data on individual chemical zones. Slow line scans were conducted to allow for an investigation of variations across complex chemically zoned grains (generally oriented perpendicular to the c-axis), allowing for the integration and processing of data from individual chemical zones while also allowing for the ability to remove contamination from inclusions. Tourmaline is known to exhibit significant crystallographic fractionation of specific trace elements between the a- and c-sectors (van Hinsberg et al. 2006, van Hinsberg & Schumacher 2007a, van Hinsberg & Marschall 2007, Iveson et al. 2016, Marks et al. 2013). It has been suggested that the a-sector is the ideal zone for analyses, as it approximates a sector-free tourmaline in grains that exhibit sector zonation (van Hinsberg et al. 2006). Porphyry TSM has been shown to exhibit a range of internal zoning textures (Beckett-Brown et al. 2023); they are almost exclusively dominated by concentric zonation, with minimal sector zonation present across the deposits and grains examined. Specifically, only two samples [Woodjam (Takom: 13CBWJ05) and Highland Valley Copper (KQ-82-52B)] displayed sector zonation. Where grains with sector zonation were analyzed, they are noted in the associated data tables (Appendix 3). It is shown later that the absolute concentrations are not as important as element ratios (which are significantly less affected by sector zonation) in discriminating the TSM environment of formation. External reference materials included: (1) NIST 610 (Jochum et al. 2011), (2) NIST 612 (Jochum et al. 2011), and (3) BHVO2G (Raczek et al. 2001). The NIST 610 standard served as the primary reference material, with Si (obtained from quantitative SEM-EDS analyses) being used as the internal standard. Data reproducibility was assessed using NIST 612, BHVO2G, and an in-house tourmaline standard, which resulted in an accuracy of concentrations for NIST 612 and BHVO2G within 10–15% of the published values. The in-house tourmaline standard was analyzed as a matrix-matched quality control standard, producing <10% relative error from its accepted values. Standards were ablated before and after every 10–15 analyses made on TSM. Although TSM from porphyry systems rarely contain inclusions, they are present and need to be avoided or removed using the data processing software program Iolite (Paton et al. 2011). During data processing, specific attention was paid to P, Ti, and Zr, which account for the most common inclusions observed (apatite, oxides, and zircon), whereby spikes in these elements were, in general, treated as inclusions and not included in the data set. Additionally, channels of B and Si were also referenced to assure that count rates were consistent across zones that were integrated for analyses. A drift correction was applied using the baseline-reduction scheme available in Iolite (Paton et al. 2011). Detection limits were calculated in Iolite following the method of Pettke et al. (2012) and presented as an average value. A summary of the average trace-element data for porphyry-related TSM is presented in Table 3, with full results for the porphyry TSM presented in Appendix 3. Trace-element concentrations are presented in ppm with the notion that they are by weight (i.e., mg/kg).

Trace element analyses (LA-ICP-MS) were completed on 527 TSM grains from 24 samples from the seven porphyry deposits examined in this study (Fig. 2, Table 3). Many elements were found to be at or below the lower detection limit (Li, Be, Mo, Ag, Cs, Ta, Bi, and Hf). The elements detected were divided into three populations based on the concentrations obtained. The lowest concentration trace elements (averaging 0.5–10 ppm) include large ion lithophile elements (LILE): Rb, Ba, and Pb; high field strength elements (HFSE): Y, Zr, Nb, U; and semi-metals: Ge, In, and Sb. Others were present in low concentrations (averaging 10–100 ppm), including transition metals: Sc, Cr, Co, Ni, Cu, and Zn; semi-metals: Ga, Ge, As, and Sn; and HFSE: Th. The remainder, alkali and alkali-earths: K and Sr; and transition metals: Ti, V, and Mn, were found to occur in intermediate concentrations (averaging 100–1000s of ppm).

Two key trace elements were particularly examined. The first is Li, which has been used as a discriminator based on its abundance in TSM among deposit types in other studies (Iveson et al. 2016, Kalliomäki et al. 2017, Codeço et al. 2020, Li et al. 2021a). Results from the current study found Li to be present in concentrations at or below the limit of detection (∼10 ppm), in contrast to the higher concentrations (e.g., orogenic Au TSM contain tens of ppm Li, granitic TSM containing >80 ppm Li in the Spirit Lake pluton, and TSM from the Panasqueira W-Sn-Cu deposit contains hundreds of ppm Li) noted in other studies (Iveson et al. 2016, Kalliomäki et al. 2017, Codeço et al. 2020). The second group of elements that warranted careful examination was the rare earth elements (REEs). In many cases, several of the middle rare earth elements (MREE) occurred below the limit of detection. In cases where REEs concentrations were above detection limits, the resulting distribution patterns are concave, with variable Eu anomalies (from <LOD to 59.43, averaging 2.1) (calculated using the values from Taylor & McLennan 2008). Depletion of MREEs is consistent with amphibole fractionation, typical for hydrous porphyry magmas (Chiaradia et al. 2012, Loucks 2014). A dichotomy in Eu anomalies was observed with HVC, Los Bronces, Schaft Creek, and Woodjam samples exhibiting negative Eu anomalies on average while the Casino, Coxheath, and New Afton samples exhibited positive Eu anomalies. These contrasting anomalies for the two groups of samples reflect differing fluid sources or if plagioclase alteration was involved in the formation of TSM (Čopjaková et al. 2013, Marks et al. 2013).

The results from this study show that porphyry TSM contain variable concentrations of LILE expected to reside at the X-site of TSM, including K (range: 72–6900 ppm, averaging 378 ppm), Rb (range: <LOD–61.6 ppm, averaging 2.4 ppm), Sr (range: 35–977 ppm, averaging 352 ppm), Ba (range: <LOD–300 ppm, averaging 9.8 ppm), and Pb (range: <LOD–45 ppm, averaging 3.3 ppm). High field strength elements include Ti (range: 111–9239 ppm, averaging 2620 ppm), Zr (range: 0.59–116 ppm, averaging 9 ppm), Nb (range: <LOD–32.6, averaging 0.9 ppm), and Th (range: <LOD–13.8 ppm, averaging 1.3 ppm). Transition metals include V (range: 51–3076 ppm, averaging 626 ppm), Cr (range: <LOD–1241 ppm, averaging 70 ppm), Mn (range: 16–1300 ppm, averaging 185 ppm), Co (range: <LOD–260 ppm, averaging 14.3 ppm), Ni (range: <LOD–120 ppm, averaging 18.5 ppm), Cu (range: <LOD–510 ppm, averaging 24.7 ppm), and Zn (range: 5.7–151 ppm, averaging 36 ppm). Finally, semi-metals include Ga (range: 21–114 ppm, averaging 58 ppm), As (range: <LOD–1499 ppm, averaging 85 ppm), Sn (range: 1–62 ppm, averaging 11.8 ppm), and Sb (range: <LOD–41.5 ppm, averaging 1.7 ppm). The latter elements (As, Sn, and Sb) potentially reflect a lower temperature, distal portion of these systems.

Among the deposits examined in this study, important inter-deposit TSM trace-element variations exist, notably in terms of the following elements for which the average values are presented below: LILE – K, Sr, and Pb; transition metals – Ti, V, Mn, and Cu; semi-metals – As and Sn; and REEs – La. Inter-deposit variations are important in that they reflect physiochemical differences among the deposits examined. Among the TSM grains from the deposits examined in this study, Coxheath contains the highest average K (1256 ppm), while Woodjam contains the lowest (143 ppm). Strontium is consistent between the deposits; 352 ppm for porphyry TSM, with the Casino deposit being somewhat lower (149 ppm). The Donoso breccia additionally contains the highest Pb concentration (8 ppm), while Woodjam contains the lowest (1.2 ppm). With regards to the transition metals, the Donoso breccia contains the highest Ti concentration (6551 ppm), which is much higher than any other breccia samples examined, whereas the Casino samples contain the lowest (1517 ppm). New Afton contains the highest V contents (1724 ppm), reflecting the association with mafic rocks, while HVC contains the lowest (141 ppm). Highland Valley contains the highest Mn (364 ppm), while Casino contains the lowest (80 ppm). The Donoso breccia contains the highest Cu (44 ppm), while HVC contains the lowest (6.3 ppm), which reflects the timing relationship between TSM formation and Cu-sulfide formation with the samples from HVC forming much later than all other deposits examined. Finally, for the semi-metals, the Donoso breccia contains the highest As (755 ppm), while New Afton contains the lowest (3 ppm), likely reflecting the incorporation of sediments or surrounding host rocks. The Casino deposit contains the highest Sn (21 ppm), while Coxheath contains the lowest (3.9 ppm), with higher concentrations linked to more distal, lower temperature regions of a specific deposit. The Donoso breccia contains the highest La (17.4 ppm), while Coxheath contains the lowest (0.7 ppm).

Intra-grain variations exist in all the deposits examined and reflect changes in the local environment during crystallization. A LA-ICP-MS trace-element map was collected for a TSM grain from Woodjam that showed a particularly complicated BSEI (backscattered electron image), i.e., a combination of concentric and “patchy” zoning, which some might call sector zonation (Fig. 3). This grain is described, as it highlights the two common internal textures observed in TSM from porphyry systems and represents the most complicated example of internal zonation. Similar, but less complicated, intra-grain features and chemistries were observed in other porphyry deposits, specifically, the trends reflected in the concentric zones (concentric being by far the most common). The variations in grayscale primarily reflect changes in major element chemistry, specifically Al and Fe, with the lighter areas having a higher concentration of Fe (13.16 wt.% avg) and being depleted in Al (11.74 wt.% avg) relative to the darker zones. In general Mg (∼5.5 wt.%) remains relatively constant from core to the rim (Fig. 3). Several trace-element associations are evident: LILE, HFSE, metals, semi-metals, and REEs. Regarding the LILE: K, Sr, and Pb are the principal elements detected, capable of substituting into the ^[IX]X-site along with Na and Ca. Potassium concentrations are highest in the core and weakly correlated with the observed concentric zonation pattern, with some correlation with the patchy zonation as well, although that is less clear (Fig. 3). Strontium is distinct in that it is not correlated with the concentric zonation patterns but closely follows the patchy zonation (Fig. 3). Higher Sr concentrations are found predominantly in the brighter (BSEI) areas that coincide with areas of Fe enrichment and Al depletion. Lead is homogeneous in its distribution across the grain (not shown). The HFSE, including Ti and Zr, are enriched in the core regions but have some inverse relationships (Fig. 3). The small <10 μm bright spots in the Zr map (Fig. 3) likely reflect zircon inclusions. The Ti distribution pattern is like that of Sr described above, but the inverse is present for Zr (Fig. 3). Many metals, including Cr, Ni, and Cu, are homogeneously distributed (i.e, show no visible zonation in trace element grain maps), while a few key elements, including V, Mn, and Zn, are not. For example, V (not shown in Fig. 3) displays oscillatory zonation, a pattern similar to that observed for Zn (Fig. 3). Manganese also shows a similar pattern to Zn (Fig. 3). Vanadium is concentrated in the rim and less abundant in the core, which is also like Zn (Fig. 3). Manganese exhibits a similar enrichment in the rim of this grain but is concentrated closer to the contact with the core region rather than closer to the edges of the grain (Fig. 3). The V and Mn distributions exhibit subtle variations matching the patchy zonation but dissimilar to that for Sr (Fig. 3). Variations in semi-metal concentrations (e.g., Ga and As) are less pronounced and appear to coincide with the observed oscillatory zonation (Fig. 3). Gallium is slightly enriched in the rim relative to the core, whereas As (not shown) and Sn are enriched in the core compared to the rim (Fig. 3). Overall, the REEs follow both oscillatory and patchy zonation patterns (Fig. 3). There does appear to be a difference between LREE (light rare earth element) and HREE (heavy rare earth element) concentrations: LREE correlates with bright (BSEI) patchy regions in the core while the HREE appears inverse with higher concentrations in the dark (BSEI) patchy regions of the core (Fig. 3).

Intra-deposit variations in TSM exist, as noted by the min/max variations for individual deposits with multiple samples (e.g., Casino, Coxheath, New Afton, and Schaft Creek). At the Casino deposit, where good sample distribution exists, some spatial trace-element trends become evident, primarily with Zn and Pb. Lower concentrations of Zn and Pb (averaging ∼10 and 1.5 ppm respectively) are observed closer to the porphyry center and concentrations increase as you get more distal from the porphyry center (averaging ∼40 and 5 ppm respectively). Additional element trends are less clear within the current data set, but will be the focus of a future publication.

The trace element variations present in a mineral (and TSM in particular) are dependent on several factors, including (but not limited to) fluid/melt composition (Slack & Trumbull 2011), the nature of co-crystallizing minerals (Taylor & Slack 1984, Slack et al. 1999), host rock compositions (Galbraith et al. 2009), metamorphic grade (Sciuba et al. 2021), P-T (pressure-temperature) conditions (van Hinsberg et al. 2011), oxygen fugacity (van Hinsberg et al. 2011), and crystal structure constraints (Bosi 2018). To understand the variation in the overall trace-element composition of TSM is to holistically understand the interplay among the inter- and intra-variables during crystallization. Doing so requires an appreciation of the fact that, while a complex set of parameters may be in operation collectively or independently, the sum of their effects has been encapsulated within the TSM. Further, the potential use of such variations to discriminate between TSM from different environments of formation becomes an important, feasible outcome, and it is this latter aspect, specifically concerning those characteristics of TSM from mineralized porphyry systems, that serves as the locus of this present study.

Interpretation of the trace element data and patterns in TSM requires knowledge of both the intrinsic and extrinsic factors that collectively control the chemical composition, especially when absolute concentrations are unable to effectively discriminate TSM from different environments (Fig. 4). Intrinsic factors are those related to the crystal chemistry and structure of a TSM, including the ideal size of a given crystallographic site, ionic radii, valence, and differences in substituting ions (i.e., size, charge, and compatibility with the crystallographic site) for a simple ionic substitution or composite ionic radii for more complex substitutions. The systematics of trace element cation incorporation into the TSM structure has been relatively unexplored in terms of crystal chemistry, but an understanding of the intrinsic constraints on the trace element incorporation is necessarily crucial. The most basic manner by which any element can be considered as potentially substituting for another is by considering the relative differences between the ions in terms of radii (primarily) and charge (secondarily) (Goldschmidt 1937). The ideal radii for various crystallographic sites is known: r_X = 1.28 Å, r_Y = 0.68 Å, r_Z = 0.56 Å, and r_T = 0.24 Å, determined from site dimensions from a TSM of intermediate schorl-dravite composition (van Hinsberg 2011). This is presented as a simplistic overview and should not be taken as rigid parameters for element incorporation. The TSM crystal structure is incredibly flexible and can incorporate a wide range of elements that theoretically should not easily fit (e.g., potassium: see Berryman et al. 2016). Figure 5 shows each cation site and the potential substituents. The ideal site size is denoted on the respective diagrams with 5% and 15% (following Goldschmidt's rules for substitution; Goldschmidt 1937) fields also denoted which encapsulate the most likely substituents into their respective sites. Although the ideal crystallographic site size will differ depending on the TSM endmember, the vast majority of the TSM examined in this study lie within the schorl-dravite solid solution (cation site assignments gathered from Bosi 2018, Bačík & Fridrichová 2021, Vereshchagin et al. 2021). Given the size of the X-site (for Na-dominant endmembers), the similar size of ^[9]REE (1.032–1.216 Å) and the incorporation of elements (e.g., alkaline earths such as Ca) here suggest the incorporation of REE into the X-site. Using the diagram, it is predicted that the REE patterns for TSM should show evidence for LREE enrichment, given that the HREE is >15% smaller than the ideal X-site size (Fig. 5a). Other X-site trace elements unlikely to be found in significant concentrations include U and Th, due to their large size and significant charge difference, as well as Rb and Cs, due to their significantly larger (>15% of the ideal size) ionic radius. The Z-site, most commonly occupied by Al, can be expected to accommodate a range of trace elements, including As, Cr, Co, Ga, and V, whereas the Y-site, most commonly occupied by Fe and Mg, can also incorporate Ni, Li, Cu, Zn, Co, Mn, and Sc (Fig. 5b). The T-site, commonly occupied by Si, is unlikely to contain large concentrations of trace elements due to the small site size and the abundance of Si in hydrothermal fluids (Fig. 5c). Boron can be incorporated into the T-site in rare cases when there is excess B and or a lack of sufficient Si, which has been noted in Al-rich TSM (Hughes et al. 2000, Ertl et al. 2018). Due to the flexibility of the TSM structure, it is important to consider the size of prospective crystallographic sites in the TSM structure. It must also be considered that differences in ionic radii among substituting ions may also be, in part, accommodated through relativistic effects associated with substitution in local or adjacent crystallographic sites. The latter may lead to the development of ordering patterns (short- or long-range) (Bosi 2008, Bosi et al. 2018, Andreozzi et al. 2020).

Alternatively, extrinsic controls are those that are imparted by the environment of formation on the formation of a crystal which may include temperature, pressure, melt/fluid composition, fO₂, co-crystalizing phases etc. that will directly affect the potential trace element substitution in TSM. One study indicated that the partitioning of trace elements between TSM and a silicate melt (800 °C and 7.5 kbar) is close to one (van Hinsberg 2011), suggesting that the trace element data of a TSM can closely indicate the composition of the melt from which it crystallized. In terms of overall porphyry chemical signatures, porphyry magmas have been shown to exhibit high contents of Sr relative to Y as well as V relative to Sc (Richards 2011, Chiaradia et al. 2012, Loucks 2014). Other geochemical pathfinders for porphyry deposits include Cu, Mo, and Au (proximal pathfinders) as well as Zn, Ag, Pb, As, Sb, and Mn (distal pathfinders) (Cooke et al. 2020a), that are also many of the key elements for distinguishing TSM in porphyry systems.

Discrimination diagrams are numerous in the mineral chemistry literature (e.g., Dare et al. 2012, Li et al. 2021a), but many authors fail to provide a scientifically valid justification for the use of specific elements, ratios, or addition of element concentrations that were used in developing the discrimination plots presented. These diagrams (in general) rely simply on the visual separation of data and very few consider suitable explanations for the visual interpretations that have been made. Average trace-element concentrations can partially distinguish TSM in porphyry systems from that in other geologic environments, but ratios of key elements are most effective because much of the noise in the data is removed. Ratios may correct for the wide range of trace element concentrations (concentric or sector zonation) and in part normalize the data, removing some of the intra-grain variations observed between distinct chemical zones (Fig. 3). Ratios can also be useful for evaluating the competition of elements for specific crystallographic sites. A number of ratios were examined including V/Sn, Sr/Sn (previously presented by Sciuba et al. 2021), Ti/V, Cr/V, Co/Ni, Sr/Pb, and Zn/Cu, many of which showed little to no variation among the environments examined and their associated fluid/melt sources (i.e., granitic, metamorphic, orogenic, pegmatitic, porphyry Cu, polymetallic W ± Sn ± Cu, and VMS). Meaningful variations in Sr/Sn exist for the TSM groups examined, but the overlap between orogenic and porphyry-related samples limits its utility as a discriminator. For Ti/V, little variation exists in Ti concentrations among the environments examined, while V is much more variable, with pegmatitic and W-Sn samples having significantly lower concentrations with respect to other environments investigated. Unfortunately, this ratio is only effective at separating pegmatitic samples. Finally, V/Sn is effective for separating porphyry and orogenic samples, but the high concentrations of V in the metamorphic samples overlap both porphyry and orogenic TSM, precluding differentiation. The ratios Sr/Pb and Zn/Cu are the most useful in discriminating between TSM in mineralized porphyry systems and those from other potential environments (Fig. 6). Development of this discrimination plot is based on several key features, including intrinsic (crystal-chemical) and extrinsic (geologic environment) factors. In terms of the crystal structure of TSM, both Sr and Pb are known to occupy the X-site, while both Zn and Cu are known to compete for the Y-site (Beurlen et al. 2011, Ertl & Bačík 2020, Bačík & Fridrichová 2021).

A ternary plot of Sr/Pb–Zn/Cu–Ga reveals that TSM associated with mineralized porphyry systems have trace element compositions that predominantly occur along the Sr/Pb–Ga join (Fig. 6). Porphyry TSM contain high Sr/Pb (averaging 297), variable Ga (averaging 58 ppm), and low Zn/Cu (averaging 4.8) values (Fig. 6). Conversely, the TSM associated with orogenic Au deposits cluster along the Sr/Pb–Zn/Cu join, and those having a granitic or granitic pegmatite affiliation plot along the Zn/Cu–Ga join (Fig. 6). In effect, the relatively simple Sr/Pb–Zn/Cu–Ga ternary plot is a powerful tool for discriminating among different types of TSM evaluated in this study. As a note, some of the TSM that have been classified as orogenic in origin (e.g., those forming from hydrothermal fluids in orogenic Au systems) plot outside of what is considered the orogenic field. Such TSM are likely metamorphic in origin or may have experienced some overprinting effects related to metamorphism. These TSM warrant further investigation into their textural relationship with the mineralizing fluids. In general, metamorphic TSM plot in the center of the diagram trending from the Zn/Cu apex along a vector toward the center of the Sr/Pb and Ga side of the diagram (Fig. 6). Even with the high degree of chemical variability that exists at the grain-scale for individual crystals of porphyry-related TSM, their distinct porphyry signature (i.e., high Sr/Pb, low Zn/Cu, and variable Ga) is still retained (Fig. 6). However, some porphyry-related TSM plot to the left of the Sr/Pb–Ga join due to relative enrichments in Zn, which include some of the TSM from HVC. These TSM samples are from a late-stage dike rather than within the porphyry and may explain the Zn enrichment, similar to the zonation of Zn seen in porphyry systems (Cooke et al. 2020a). Based on this assumption, samples of TSM with high Zn could potentially be considered less prospective (i.e., typically have concomitant decreases in Cu) because they may be indicative of occurring distal from the porphyry center or entirely unrelated to the mineralizing process. Orogenic TSM, much like porphyry TSM, exhibit a wide variation along the Sr/Pb–Zn/Cu join. The bulk of this variation reflects changes in Sr and Pb which largely reflect the wide range of host rocks for these deposits. For TSM associated with VMS environments, the most recent data are from the La Ronde deposit (Quebec, Canada) (Sciuba et al. 2021). These data show that two populations exist: one that plots in the metamorphic field and the other that plots close to the top of the diagram in the orogenic field (due to high concentrations of Sr and low concentrations of Zn). Unfortunately, due to a paucity of data for TSM from VMS settings, conclusions cannot be made about their trace element compositions in relation to the other environments examined in this study. However, based on TSM data from the select VMS systems (e.g., Griffin et al. 1996, Slack et al. 1999), elements that should be useful for discrimination include Cu, Pb, and Zn.

The absolute concentrations of individual trace elements within single TSM grains can be orders of magnitude different (e.g., Ti: Fig. 3) and, as such, using absolute concentrations to discriminate TSM becomes challenging, not to mention the difficulty of determining where the best place for trace element analyses might be on any grain with complex chemical zonation (e.g., Fig. 3). Utilizing ratios as presented above removes much of the noise from absolute concentrations. For example, >90% of LA-ICP-MS time slice data (a method presented first presented by Petrus et al. 2017) for the TSM grain map presented in Figure 3 plots within the porphyry TSM field (Fig. 7). This shows that, no matter the internal variation of a grain, the TSM environment of formation can still be determined using Figure 6.

The consistently high Sr content, compared to TSM in unmineralized environments, present in porphyry-related TSM [averaging 352 ± 180 (2S.D.) ppm] is consistent with the enrichment observed in porphyry type magmas (≥400 ppm) and their associated fluids, a feature ascribed to the suppression of plagioclase crystallization in tandem with fractionation of amphibole ± garnet in the lower crust (Richards & Kerrich 2007, Chiaradia et al. 2012, Loucks 2014). The enrichment in Sr in porphyry TSM is also consistent with observations made of Sr in plagioclase from intrusions related to porphyry deposits that contains thousands of ppm Sr (Cao et al. 2019). Mineralizing porphyry fluids are also known to contain elevated concentrations of Sr, hundreds of ppm (Ulrich et al. 1999). The alteration of these Sr-rich feldspars during hydrothermal processes could provide some or all of the Sr budget for the porphyry-related TSM. It is interesting to note that the Sr content of TSM associated with granitic pegmatites is lower by an order of magnitude (averaging 35 ppm in this study), which may be related to the formation of paragenetically earlier feldspars. The Sr content of TSM from TSM granites (magmatic TSM: Andreozzi et al. 2020) is even lower (averaging 5 ppm), reflecting the competition with other phases, such as feldspars. In TSM granites specifically, TSM formation is shown to be one of the last phases in the paragenesis (Samson & Sinclair 1992, Sinclair & Richardson 1992, Hong et al. 2019).

The Pb concentrations (averaging 3.3 ppm) in porphyry-related TSM are an order of magnitude lower than those of Sr, resulting in consistently high values of Sr/Pb. The incorporation of Pb into TSM is linked to the Pb availability and competition with other crystallizing minerals, most importantly alkali feldspars and sulfides, which both have high affinities for Pb, with D_{feldspar/melt} = 0.989–2.72 and D_{sulfide melt/silicate melt} = 140–3300 (Aigner-Torres et al. 2007 and Li et al. 2021b, respectively). Concentrations of Pb and Zn increase with increasing distance from the porphyry center (Jones 1992). Thus, the variations of Pb-Zn in TSM should reveal this spatial relationship, such that the Pb and Zn concentrations should decrease toward the core of the system. This appears to be the case based on the data set presented here, but contrasting observations were made by Cooke et al. (2017). An additional detailed investigation of TSM trace-element chemistry is needed to better constrain this observation. In terms of TSM from unmineralized settings, Pb concentrations are significantly higher, with TSM from granitic pegmatites, for example, averaging 22 ppm. It should be noted that Pb-rich TSM have been observed as natural samples and experimentally, with up to 17.5 wt.% PbO reported in natural TSM from Vietnam (Sokolov & Martin 2009) and up to 14.7 wt.% PbO in TSM synthesized at 700 °C and 200 MPa (Vereshchagin et al. 2020), which could explain the occurrence of spuriously high Pb values in the TSM data set presented in this study.

The concentration of Zn is strongly correlated with the nature of the environment of formation. For example, TSM from mineralized porphyries contain relatively low concentrations of Zn (averaging 36 ppm), whereas granitic pegmatites show more pronounced enrichments (averaging 450 ppm, in data compiled in this study). In porphyry systems, higher concentrations of Zn are observed to be distal to the porphyry center. To date, the highest recorded concentration in a TSM is 7.37 wt.% ZnO from a transitional NYF-LCT granitic pegmatite (Pieczka et al. 2018). In medium to high-grade metapelitic rocks, it has been argued that the principal hosts of Zn (that are not sphalerite), ranked in terms of decreasing concentrations, are gahnite > staurolite > tourmaline (Henry & Dutrow 2018). In this regard, along with their generally high modal abundances, TSM must be considered an important sink for Zn not only in metamorphic rocks, but also in other environments that contain elevated concentrations of Zn, such as pegmatites, especially those that lack gahnite. In systems where gahnite is present (which contains up to 43.75 wt.% ZnO), TSM can still contain significant concentrations of ZnO, up to 0.75 wt.% (Henry & Dutrow 2001). The fact that Zn is considered highly mobile in the presence of a fluid phase may, in part, explain its enrichment during fluid exsolution leading to pegmatite formation (Telus et al. 2012).

The concentration of Cu in porphyry-related TSM is generally higher (averaging 25 ppm) than in other environments. The relative enrichment of Cu in porphyry TSM may be in part related to the high Cu concentrations found in porphyry mineralizing fluids (Hedenquist & Richards 1998, Ulrich et al. 1999). The concentration of Cu also appears to strongly correlate with its environment of formation. The exception for high Cu concentrations are Paraiba and Paraiba-like material (pegmatitic TSM with a striking blue color) where TSM-elbaite contains up to 2.50 wt.% CuO (Abduriyim et al. 2006). The enrichment of Cu in these TSM has been attributed to pegmatitic melts mixing with ultramafic rocks rich in Cu (Beurlen et al. 2011). Additionally, these Cu-bearing to Cu-rich TSM can be distinguished by their low concentrations of Sr, which are generally <10 ppm (Abduriyim et al. 2006), making them easily distinguised from porphyry TSM. Typical pegmatitic TSM, in contrast, has much lower concentrations of Cu (avg: 1.2 ppm). Variation along the Zn/Cu–Ga side of the diagram (Fig. 6) of pegmatitic TSM grains could in part be explained by variability in the parental granitic melts. These melts can arise from single sources, but also from the anatexis of multiple rock types, which could lead to considerable chemical variability in the resulting TSM (Černý et al. 2012).

Gallium has an average concentration in the Earth's upper continental crust of 17.5 ppm (Rudnick & Gao 2003). The presence of Ga³⁺ (^VI0.62 Å), owing to its similarity in charge and size, substitutes for Al³⁺ (^VI0.535 Å) or Fe³⁺ (^VI0.645 Å) (Shannon 1976). Gallium is compatible with TSM, fitting into either the Y or Z sites. The fact that Ga exhibits a wide range in concentrations (Fig. 4), from one to hundreds of ppm in TSM, makes it a potential discriminator. Studies of Ga in natural melts and fluids show that the highest concentrations are associated with fluids of magmatic origin (Prokof'ev et al. 2016) even though Ga relative to Al concentrations have been shown to increase during fractionation (Breiter et al. 2013). The concentration of Ga in TSM has been shown to be useful for separating magmatic and hydrothermal samples, with hydrothermal grains having lower concentrations (Sciuba et al. 2021, their Fig. 11h). From the data presented here it is unclear if this observation is widespread or whether pressure or temperature contribute to the observed Ga contents of TSM; likely both play a role in its abundance in TSM. Concentrations of Ga in other minerals, such as igneous magnetite, are higher than those in hydrothermal magnetite (Nadoll et al. 2014), potentially further linking Ga incorporation as a function of temperature. Similar results were found in the current study when TSM from porphyry versus orogenic settings are considered. Here porphyry and orogenic TSM (hydrothermal) contain lower concentrations of Ga than granitic TSM (magmatic).

It is also important to consider the mineral–melt partition coefficient values (D_tur/melt) in relation to the elements used for discrimination, as well as mineral–mineral coefficients, when available. As noted earlier, the only study that provides any such data for TSM is that of van Hinsberg (2011), which reported K values of ∼1 for most elements. Although this study cannot in any way encapsulate and reflect all the possible conditions under which TSM may develop, it does provide a baseline from which trace element data may be discussed. For example, the D_tur/melt Sr = 1.05 is close to unity and allows for Sr contents to be directly used as an indicator of the bulk melt Sr composition and, by proxy, a discriminator of different melt/fluid sources. In the discrimination diagram presented here (Fig. 6), a Sr/Pb ratio is considered, owing to the hypothesis that both elements are co-competing for the X-site. However, the D_tur/melt Pb = 0.57 (i.e., slightly < 1), reflecting the fact that Pb favors the melt (under the conditions of the experiment), thus Pb is not as favored in the TSM structure. For the Zn/Cu apex, the D_tur/melt Zn = 0.66, indicating a slight preference for the melt; empirically, Zn concentrations in TSM have been a strong indicator for differentiating fractionation and magmatic versus hydrothermal processes. Copper (D_tur/melt = 1.86) in TSM is another strong indicator for mafic systems which have not yet reached conditions of sulfide saturation or hydrothermal systems with elevated Cu concentrations (i.e., porphyry deposits). In systems where sulfide saturation has occurred, D_{sulfide liquid/silicate melt} Cu = 50–220 (Li et al. 2021b), any subsequent TSM would be strongly depleted in Cu. In systems which are not actively crystalizing sulfides or have not reached sulfide saturation, Cu can readily partition into TSM with a reported D_tur/melt Cu = 1.86. Finally, Ga has a reported D_tur/melt Ga = 0.76, suggesting a slight preference for the melt. It has also been shown to be an indicator of magmatic and hydrothermal processes as well as metamorphic grade (Prokof'ev et al. 2016, Sciuba et al. 2021). Considering the partition coefficients (values from van Hinsberg 2011) and crystal-chemical considerations described above for trace elements in TSM, we propose a new ternary plot (Fig. 6) to discriminate TSM from distinct melt/fluid sources. The lack of major constraints on the incorporation of these elements in TSM further reinforces the effectiveness of this diagram as a useful discriminator for its geological environment of formation.

In order to determine how individual elements affect the distribution on the ternary diagram, the absolute abundances (color-coded concentrations separated into ten equal ranges) of each of the elements are overlain on the ternary diagram (Fig. 8). Strontium concentrations are highest in porphyry and orogenic TSM samples and lowest in pegmatitic and granitic samples. Lead concentrations distinctly show lower concentrations toward the Sr/Pb apex and higher concentrations toward the Zn/Cu–Ga line. Orogenic samples exclusively have low concentrations of Pb while the porphyry TSM increase in Pb concentrations toward the Ga apex, likely reflecting TSM forming either on the peripheries or in a higher-level setting where Pb is more concentrated. Gallium concentrations in TSM have been plotted to distinguish magmatic and hydrothermal origins, and they also possibly reflect the temperature of formation. Granitic and pegmatitic TSM contain the highest concentrations of Ga, followed by porphyry and orogenic related TSM. Zinc concentrations reflect competition with other phases and fractionation with porphyry TSM containing the lowest concentrations followed by orogenic and granitic/pegmatitic, which have the highest concentrations. Finally, Cu concentrations are highest in porphyry-related TSM followed by pegmatitic and orogenic related TSM. While Cu and Zn abundances in TSM can be used to broadly define prospective porphyry versus non-porphyry sources, the wide range in Sr, Pb, and Ga abundances highlights the limitation of using individual element concentrations in TSM as a source discriminator and the strength of using element ratios to better define TSM sources.

There are a variety of factors to consider when applying the discrimination diagram (Fig. 6). To start, a thorough investigation of macro and micro textures should be undertaken (e.g., grain size, morphology, color, inclusion populations, and internal zonation; discussed in Beckett-Brown et al. 2023). Grains that do not fit the textural characteristics of porphyry TSM cannot be considered porphyry-derived. For example, grains forming as isolated individual grains with abundant inclusions are not derived from a porphyry system. Once the textures have been investigated, then the major and minor elements should be investigated. Grains which contain >0.5 apfu Ca+K+Vac should not be considered further. Additionally, any Li-bearing (>0.1 wt.%) grains should also be excluded from further consideration. Once all these parameters have been investigated, then and only then should trace element data be collected and plotted on the discrimination diagram (Fig. 6). Grains which plot with in the 99^th percentile field (the larger porphyry Cu TSM region with the line drawn at 10% line on the ternary) are prospective, and those which plot in the 95^th percentile field (the smaller region, drawn at the 3% line on the ternary) are highly prospective. This discrimination is for porphyry systems that contain Cu, so those systems which are Au-only will not fall within this region, as there is no available Cu to substitute into the TSM. Additional work will be required to help separate those Au-only porphyries.

Tourmaline in porphyry Cu (± Au, ±Mo) systems can be distinguished from that in other environments or deposit styles using trace element data if intrinsic (crystal-chemical i.e., site-specific partitioning) and extrinsic (geochemical environment, co-crystallizing minerals, and fluid chemistry) factors that affect the TSM chemistry are considered. Trace elements should not be investigated alone, and an integration of textural (Beckett-Brown et al. 2023) and chemical data provides the most effective method for identification of porphyry-related TSM. This has direct implications for the application of TSM in mineral exploration as an effective indicator mineral.

Large ions, such as Sr and Pb, are controlled by the presence or absence of feldspars. Porphyry deposit intrusions are known for their anomalous Sr concentrations, wherein Sr is primarily hosted in feldspars. The breakdown of primary feldspars during hydrothermal activity in addition to Sr-rich porphyry fluids contributes to the high concentrations in the resulting TSM. Transition metals, including V, Cr, Co, Ni, Cu, and Zn, are strongly affected by the presence of mafic rocks, with porphyry TSM in general containing low concentrations of these elements. Copper is an exception, as it can also be influenced by Cu-rich fluids found in porphyry Cu systems. The Cu contents in TSM reflect its availability in the fluid, with porphyry TSM containing the highest Cu concentrations of the ore systems that were compared. Finally, Zn, in addition to being sensitive to host rock compositions, also reflects fractionation and a lack of Zn-competing phases. Concentrations of Zn in porphyry TSM are the lowest among the environments examined. Gallium is used as a proxy for temperature as well as an indication of magmatic, magmatic–hydrothermal, and hydrothermal samples. Rare earth concentrations in TSM from porphyry systems are generally low with LREE-enriched patterns and variable Eu anomalies.

From an investigation of these intrinsic trace-element effects as well as the role of extrinsic geochemical trends and factors, a ternary plot of Sr/Pb–Zn/Cu–Ga is proposed to effectively discriminate porphyry TSM from that forming in other unique geologic environments (Fig. 6). Strontium and Pb readily substitute into the X-site, Zn and Cu readily substitute into the Y-site, and Ga substitutes into the Z-site. The partition coefficients close to one between TSM and a melt make these elements ideal indicators of distinct geochemical processes. Porphyry tourmaline is characterized by high Sr/Pb (averaging 297), low Zn/Cu (averaging 4.8), and variable Ga (20.7–114.3, averaging 58 ppm). Orogenic tourmaline is characterized by low Sr/Pb (averaging 193), high Zn/Cu (averaging 223), and variable Ga (averaging 43 ppm). Metamorphic tourmaline is characterized by low Sr/Pb (averaging 12), high Zn/Cu (averaging 223), and variable Ga (averaging 72 ppm). Pegmatitic and granitic tourmaline is characterized by low Sr/Pb (averaging 2.91), high Zn/Cu (averaging 472), and higher Ga (averaging 112 ppm).

The following companies provided sample material: Western Copper and Gold, New Gold, and Gold Fields Ltd. (now owned by Consolidated Woodjam Copper Corporation). Scott Casselman (Yukon Geological Survey) is thanked for providing geological information, advice, and access to samples from deposits in the Yukon. Martin McCurdy and Dave Sinclair (Geological Survey of Canada), as well as John Chapman, are thanked for their contributions. Western Copper and Gold are thanked for providing logistical support to access the Casino deposit and for sharing geological information. This study was funded by the Geological Survey of Canada Targeted Geoscience Initiative - 5. The lead author was additionally funded by the NSERC-CGS-Doctoral Scholarship (2018–2021) and additional support from a Society of Economic Geology Research Grant (2018–2019) and Student Fellowship (2017). This paper has benefited from thoughtful reviews by D. Petts, E. Berryman, and one anonymous reviewer, as well as the editorial handling of Dr. S.A. Prevec. This is Natural Resources Canada contribution number 20220493.