Compositional Variations of Titanite: A Possible New Tool for Cyprus-Type Volcanogenic Massive Sulfide Deposit Prospecting

Titanite in submarine mafic magmatic rocks of Neotethyan origin was studied to reveal its possible use as an indicator mineral of modern Cyprus-type VMS deposits. Four ore deposit bearing (mineralized) and eight barren (unmineralized) Triassic–Jurassic locations from the Apennines, displaced fragments of the Dinarides, as well as the Dinarides and Hellenides were studied in order to gain representative results. Preliminary SEM-EDS and more detailed EPMA analyses were performed to characterize compositional variations of titanite from basalt, dolerite and gabbro. The obtained results show compositional differences according to the mode of formation. Titanite from VMS mineralized zones shows a composition close to stoichiometric values, and thus can be distinguished based on Ti content (Ti (apfu) ≥ 0.85). Due to Fe + Al substitution on Ti site, the Fe + Al vs. Ti binary plot seems to be the most discriminant for distinguishing mineralized and unmineralized locations. However, Fe vs. Al, Al vs. Mn and Si vs. Ca + Mn discrimination diagrams can also be used. Hence, compositional variations of titanite may be a possible new tool for prospection of concealed Cyprus-type deposits in the Neotethyan realm.


Introduction
Titanite (sphene) is a common accessory mineral in igneous and metamorphic rocks as well as in pegmatites and hydrothermal environments [1]. Its compositional variations clearly correlate to the mode of formation (see, e.g., [2][3][4]); thus this mineral is successfully used as a geothermometer (Zr-in-titanite), as an indicator of sources of detrital material or as an indicator of magmatic crystallization processes [3,5,6]. In spite of the obvious possibilities following from these characteristics, titanite is rarely used in exploration geology. Some successful applications are related to Cu-porphyry systems, i.e., the evaluation of porphyry fertile plutons that may be future exploration targets (see, e.g., [7][8][9][10]). There is increasing demand for developing new methods of prospection of concealed deposits, particularly in the field of volcanogenic massive sulfide (VMS) deposits [11]. As titanite occurs in rocks of the oceanic crust either as a late stage accessory mineral or as (metasomatic) alteration products [6], it may have the potential to be an indicator mineral of Cyprus-type VMS deposits (Fanerozoic Cu-bearing VMS deposit type, hosted in mafic rocks of the oceanic crust). In the present paper we discuss this possibility and show the results of a still ongoing, comprehensive study carried out in the European Neotethyan realm.

Geology of the Studied Samples
The study areas in the Northern Apennines (Italy), the displaced fragments of the Dinarides (the NE Hungarian Darnó-and Szarvaskő Units are geologically correlated with the Dinarides; their current position is a result of a~300 km Cenozoic displacement along the Zagorje-Mid-Transdanubian Zone (see, e.g., [15] and the references cited therein)) (NE Hungary), the Dinarides (Croatia, Albania) and the Hellenides (Greece) represent the Triassic rifting-related submarine environment as well as the Jurassic oceanic crust of the Neotethys (see, e.g., [15][16][17][18] and the references cited therein). Each study locality ( Figure 1) comprises basalt, dolerite or gabbro, and can be characterized with similar primary (submarine) alteration features and negligible metamorphic overprinting [18][19][20]. Cu-Zn-Fe-rich Cyprus-type Jurassic VMS deposits were found at the Italian and Albanian ophiolitic locations, comprising massive sulfide lenses or underlying stockwork alteration zones [21,22]. The host rocks are mid-oceanic ridge (MOR)-or suprasubduction zone (SSZ)-type mafic magmatic rocks (see, e.g., [21,23] and the references cited therein). By contrast, barren Triassic and Jurassic submarine within-plate basalt (WPB), MOR-or back-arc-basin (BAB)/island arc (IAB)-type magmatic rocks can be found at the other study locations (see, e.g., [17,18]) (for a brief geological description of each study locality, see Table 1).
Geosciences 2020, 10, x FOR PEER REVIEW 3 of 11 magmatic rocks can be found at the other study locations (see, e.g., [17,18]) (for a brief geological description of each study locality, see Table 1).

Figure 1.
Schematic map showing the Neotethyan ophiolites (black patches) as well as the approximate position of the study locations (red labels) (based on [24]).

Materials and Methods
Samples from four Cyprus-type VMS deposit-bearing (mineralized) and eight barren (unmineralized) locations were studied in details (Table 1). Detailed field work and petrographical analyses were used to choose the representative, titanite-bearing polished thin and block sections for further analyses.

Materials and Methods
Samples from four Cyprus-type VMS deposit-bearing (mineralized) and eight barren (unmineralized) locations were studied in details (Table 1). Detailed field work and petrographical analyses were used to choose the representative, titanite-bearing polished thin and block sections for further analyses.
Preliminary scanning electron microscopy-energy dispersive spectroscopy (SEM-EDS) and more detailed electron probe microanalyses (EPMA) were used to characterize the compositional variation of titanite from the different locations (n = 149). The SEM-EDS study was carried out at the Department of Petrology and Geochemistry, Eötvös Loránd University (Budapest, Hungary), using an Amray 1830I type microscope equipped with an EDS detector. An accelerating voltage of 20 kV and a counting time of 100 s were used. The instrument was calibrated with the help of natural standards and the detection limits were 0.1% for the analyzed major elements. The EPMA study was undertaken at the Eugen Stumpfl Laboratory of the University Center of Applied Geosciences (University of Leoben, Austria), where a Superprobe Jeol JXA 8200 instrument was used in wavelength dispersive spectroscopy (WDS) mode, with 15 kV accelerating voltage and 10 nA beam current. The beam diameter was about 1 µm. The counting time was 20 s on peaks and 10-10 s on backgrounds. The following diffracting crystals were selected: pentaerythritol (PETJ) for Si and Ca; thallium acid phthalate (TAP) for Al; layered dispersive element (LDE1) for F; and lithium fluoride (LIFH) for Fe, Ti, Cu, Mn and Zn. Synthetic and natural corundum, fluorite, quartz, wollastonite, olivine, rutile, chalcopyrite, rhodonite and sphalerite were used as standards. The detection limits were as follows: 100 ppm for Al, 200 ppm for Ti and Fe, 240 ppm for Ca, 300 ppm for Si, 400 ppm for F and 700 ppm for Mn, Cu and Zn.
Analyses data were handled in MS Excel, in which recalculation of mineral formulas, correlation analyses and preparation of several different diagrams as well as representative tables were performed.

Results
Based on field observations as well as petrographical microscopy, the host rocks of mineralized and unmineralized locations were characterized. Basalt often displays a pillow structure and the primary alteration is very similar at each study locations. Spilitisation is a common feature, as well as the occurrence of hydrothermal minerals (e.g., quartz, calcite, chlorite, epidote, smectites, zeolites); in cooling cracks, amygdales and hyaloclastite breccia are typical. Chloritization, Na-enrichment of plagioclase also occur in the studied dolerite and gabbro, as well as hydrothermal mineral filled veinlets (e.g., quartz, calcite, prehnite, epidote, chlorite). Commonly, alteration is stronger in the mineralized zones, though really intense hydrothermal overprinting (hiding the original rock texture) occurs only in the immediate vicinity of the stockwork veins and massive sulfide lenses. Pyrite, chalcopyrite, pyrrhotite and sphalerite are common ore minerals in the studied Cyprus-type VMS occurrences. In addition, arsenopyrite, cobaltite and rare Ni-and Au-Ag-tellurides occur in the studied Albanian sites, while galena and rare Pb-selenide, native Au, freibergite and millerite occur in the studied Italian locations (for details, see [19,20,22]).
In each studied rock type (basalt, gabbro, dolerite), titanite occurs in similar textural positions. It may be found either as disseminated anhedral to euhedral grains in the groundmass, or as alteration product of former magmatic minerals. It is also observed in hydrothermal mineral formed crystal aggregates as well as veins and cavities (Figure 2A-C). The grain sizes are in a large range, though they are similar in the mineralized and in the unmineralized zones and in each rock type (commonly from <10 µm to 0.5 mm). However, in general, a slightly higher amount of titanite occurs in the mineralized zones. Based on SEM observations, compositional inhomogeneity or zonation was not observed in the studied grains ( Figure 2B).
Mineral chemical analyses revealed that titanite had close to stoichiometric Si and Ca values, and substitution in the Ti site was also subordinate. However, besides common Fe and Al contents, some F, Mn, Cu and Zn contents were also measured. Obvious differences among Ti, Al, Fe, F and Mn contents of titanite from mineralized and unmineralized zones could be observed. Slight enrichment in Ti and depletion in Al, Fe, Mn and F contents was observable in titanite from mineralized zones ( Figure 3). By contrast, no significant differences were observable according to different rock types or ages (for details, see Table 2 and Appendices A and B).

Element Substitution on the Ti Site
As expected from [12], the analyzed titanite grains showed a chemical composition close to stoichiometric values. The measured average Ti, Fe and Al amounts were consistent with the observations of [25], i.e., there is no need to assign any Al to the tetrahedral (Si) site. High Al-content (up to 5.365 wt% Al 2 O 3 ) was consistent with the formation conditions, i.e., relatively low temperature and absence of anorthite-rich plagioclase [13]. Similarly to the hydrothermal titanite results of [4], Fe/Al ratios scattered widely and consistent variability was not observed. High FeO content (up to 5 wt%) observed in the mafic rocks of Precambrian Noranda [14] was not found in these young Neotethyan rocks. The highest values were measured from a Hungarian unmineralized occurrence (up to FeO = 3.778 wt%).
The most important Ti site substitutions were Fe + Al both in the mineralized and unmineralized zones, which is proven by the observable strong negative correlation (r = −0.929). There was, however, a distinct difference between the degrees of substitution. Values closer to stoichiometric titanite (i.e., Ti (apfu) = 0.85-1) was characteristic to mineralized zones, while unmineralized zones contained titanite with higher amount of substitution (i.e., Ti (apfu) = 0.7-0.85) (Figure 4). Al content was commonly higher, while Fe content was overlapping or higher in unmineralized zones (Table 2). Interestingly, while there was a rather strong correlation between Fe and Al in mineralized zones (r = 0.697), their relation was independent in unmineralized zones (r = 0.097). Consequently, Ti and Al showed a strong negative correlation in each titanite type (r = −0.898 and r = −0.805, respectively, see Appendices A and B), while the relation of Ti and Fe was strong only in the case of mineralized zones (r = −0.788) and much weaker in unmineralized ones (r = −0.472) (see Appendices A and B). This fact may draw attention to the possibility of excess Fe substitution in Ca sites [12] in unmineralized zones. As a conclusion, on the Ti vs. Fe + Al and Al vs. Fe diagram, different fields can be assigned clearly for titanite compositions from mineralized and unmineralized zones (Figures 4 and 5). To summarize, titanite from mineralized and unmineralized zones could be rather trustworthily distinguished based on Ti site substitutions.

Charge Balancing
Based on the literature [1,4,12], charge balance in hydrothermal titanite due to Fe 2+ / 3+ and Al 3+ substitution into the Ti 4+ site is largely accomplished by the coupled substitution of Ffor O 2− . In the case of unmineralized zone titanites, where the substitution into the Ti site is characteristically higher, a clear correlation (r = 0.972) has been seen between Fe + Al and F as well as between Ti and F (r = −0.811) (see Appendices A and B). These facts support the above findings. Al content was commonly higher, while Fe content was overlapping or higher in unmineralized zones (Table 2). Interestingly, while there was a rather strong correlation between Fe and Al in mineralized zones (r = 0.697), their relation was independent in unmineralized zones (r = 0.097). Consequently, Ti and Al showed a strong negative correlation in each titanite type (r = −0.898 and r = −0.805, respectively, see Appendices A and B), while the relation of Ti and Fe was strong only in the case of mineralized zones (r = −0.788) and much weaker in unmineralized ones (r = −0.472) (see Appendices A and B). This fact may draw attention to the possibility of excess Fe substitution in Ca sites [12] in unmineralized zones. As a conclusion, on the Ti vs. Fe + Al and Al vs. Fe diagram, different fields can be assigned clearly for titanite compositions from mineralized and unmineralized zones (Figures 4 and 5). Al content was commonly higher, while Fe content was overlapping or higher in unmineralized zones (Table 2). Interestingly, while there was a rather strong correlation between Fe and Al in mineralized zones (r = 0.697), their relation was independent in unmineralized zones (r = 0.097). Consequently, Ti and Al showed a strong negative correlation in each titanite type (r = −0.898 and r = −0.805, respectively, see Appendices A and B), while the relation of Ti and Fe was strong only in the case of mineralized zones (r = −0.788) and much weaker in unmineralized ones (r = −0.472) (see Appendices A and B). This fact may draw attention to the possibility of excess Fe substitution in Ca sites [12] in unmineralized zones. As a conclusion, on the Ti vs. Fe + Al and Al vs. Fe diagram, different fields can be assigned clearly for titanite compositions from mineralized and unmineralized zones (Figures 4 and 5). To summarize, titanite from mineralized and unmineralized zones could be rather trustworthily distinguished based on Ti site substitutions.

Charge Balancing
Based on the literature [1,4,12], charge balance in hydrothermal titanite due to Fe 2+ / 3+ and Al 3+ substitution into the Ti 4+ site is largely accomplished by the coupled substitution of Ffor O 2− . In the case of unmineralized zone titanites, where the substitution into the Ti site is characteristically higher, a clear correlation (r = 0.972) has been seen between Fe + Al and F as well as between Ti and F (r = −0.811) (see Appendices A and B). These facts support the above findings. To summarize, titanite from mineralized and unmineralized zones could be rather trustworthily distinguished based on Ti site substitutions.

Charge Balancing
Based on the literature [1,4,12], charge balance in hydrothermal titanite due to Fe 2+ / 3+ and Al 3+ substitution into the Ti 4+ site is largely accomplished by the coupled substitution of F − for O 2− . In the case of unmineralized zone titanites, where the substitution into the Ti site is characteristically higher, a clear correlation (r = 0.972) has been seen between Fe + Al and F as well as between Ti and F (r = −0.811) (see Appendices A and B). These facts support the above findings.

Element Substitutions on the Ca Site
In the studied geological environment, Mn is the most probable Ca site substitution. However, according to our results, it is not particularly enriched in titanite from mineralized zones ( Table 2). Such high amounts reported by [14] from the Noranda VMS deposit (up to 0.2 wt% MnO) were not observed; moreover it is commonly below detection limit in the studied mineralized Neotethyan samples. By contrast, rather high Mn content can be observed in titanite from unmineralized zones (up to 1 wt% MnO). Consequently, weak positive correlation between Ca and Mn content is observable only in the latter case (r = 0.540).
On the Mn vs. Ca diagram, exact fields for titanite from mineralized and unmineralized zones cannot be assigned. However, on the Al vs. Mn and Si vs. Ca+Mn diagrams-despite some overlapping in the latter case-we may be able to distinguish titanite grains of different origin (Figures 6 and 7).

Element Substitutions on the Ca Site
In the studied geological environment, Mn is the most probable Ca site substitution. However, according to our results, it is not particularly enriched in titanite from mineralized zones ( Table 2). Such high amounts reported by [14] from the Noranda VMS deposit (up to 0.2 wt% MnO) were not observed; moreover it is commonly below detection limit in the studied mineralized Neotethyan samples. By contrast, rather high Mn content can be observed in titanite from unmineralized zones (up to 1 wt% MnO). Consequently, weak positive correlation between Ca and Mn content is observable only in the latter case (r = 0.540).
On the Mn vs. Ca diagram, exact fields for titanite from mineralized and unmineralized zones cannot be assigned. However, on the Al vs. Mn and Si vs. Ca+Mn diagrams-despite some overlapping in the latter case-we may be able to distinguish titanite grains of different origin (Figures 6 and 7).  As suggested earlier, we may assume some Fe 2+ substitution in the Ca site in the case of titanite from unmineralized zones. This may explain the strong correlation between Fe and Mn (r = 0.785) in these titanite grains (see Appendix B).

Metal Substitutions
Hannington et al. [14] observed up to 0.14 wt% ZnO substitution in titanite from the Noranda VMS deposit. In most cases, our data scatter in a similar range, though a maximum of 0.236 wt% ZnO content was found. CuO content also scattered in a similar range, with a maximum of 0.368 wt%. However, there was no characteristic difference between Zn and Cu content of titanite from Geosciences 2020, 10, x FOR PEER REVIEW 8 of 11

Element Substitutions on the Ca Site
In the studied geological environment, Mn is the most probable Ca site substitution. However, according to our results, it is not particularly enriched in titanite from mineralized zones ( Table 2). Such high amounts reported by [14] from the Noranda VMS deposit (up to 0.2 wt% MnO) were not observed; moreover it is commonly below detection limit in the studied mineralized Neotethyan samples. By contrast, rather high Mn content can be observed in titanite from unmineralized zones (up to 1 wt% MnO). Consequently, weak positive correlation between Ca and Mn content is observable only in the latter case (r = 0.540).
On the Mn vs. Ca diagram, exact fields for titanite from mineralized and unmineralized zones cannot be assigned. However, on the Al vs. Mn and Si vs. Ca+Mn diagrams-despite some overlapping in the latter case-we may be able to distinguish titanite grains of different origin (Figures 6 and 7).  As suggested earlier, we may assume some Fe 2+ substitution in the Ca site in the case of titanite from unmineralized zones. This may explain the strong correlation between Fe and Mn (r = 0.785) in these titanite grains (see Appendix B).

Metal Substitutions
Hannington et al. [14] observed up to 0.14 wt% ZnO substitution in titanite from the Noranda VMS deposit. In most cases, our data scatter in a similar range, though a maximum of 0.236 wt% ZnO content was found. CuO content also scattered in a similar range, with a maximum of 0.368 wt%. However, there was no characteristic difference between Zn and Cu content of titanite from As suggested earlier, we may assume some Fe 2+ substitution in the Ca site in the case of titanite from unmineralized zones. This may explain the strong correlation between Fe and Mn (r = 0.785) in these titanite grains (see Appendix B).

Metal Substitutions
Hannington et al. [14] observed up to 0.14 wt% ZnO substitution in titanite from the Noranda VMS deposit. In most cases, our data scatter in a similar range, though a maximum of 0.236 wt% ZnO content was found. CuO content also scattered in a similar range, with a maximum of 0.368 wt%. However, there was no characteristic difference between Zn and Cu content of titanite from mineralized and unmineralized locations, though the highest Zn and Cu values were measured from the latter occurrences. As the development of a hydrothermal fluid circulation system leaches the metals (e.g., Cu, Zn, Mn) from the silicates [26], occasionally high metal content of titanite from unmineralized zone may be due to the lack of extensive alteration.

Concluding Remarks
In the present study, Neotethyan mafic igneous rocks from Cyprus-type VMS mineralizations and barren locations were studied. The submarine rocks show similar primary alteration features, and therefore it is hard to localize the concealed, economically potential (i.e., locally mineralized and strongly altered) occurrences. The research presented here, however, revealed some conclusions regarding the possible use of titanite chemistry for modern Cyprus-type VMS deposit prospecting in the Neotethyan realm. This mineral occurs in each possible host rock types and its compositional variations correlate well with the mode of formation. The following distinguishing features can be listed: • The study locations cover a wide geographic area, though all of them represent the same geotectonic situation, i.e., the rifting and spreading of the Neotethys. Therefore the above conclusions have to be handled accordingly and further research should be performed. Future research should involve not only a detailed mineral chemical study of more locations, but also more complimentary geostatistical investigations to better evaluate the use of titanite chemistry in mineral exploration.
Author Contributions: Conceptualization, investigation, methodology, analyses, visualization, writing (original draft preparation) and supervision, G.B.K.; analyses, writing (review), methodology, F.Z. All authors have read and agreed to the published version of the manuscript. Acknowledgments: The authors are thankful to I. Barna, G. Garuti, A. Lovász, A. Lukács, Zs. Molnár, D. Pásztor and P. Skoda for providing some of the analyzed samples. Zs. Bendő is thanked for assistance during SEM-EDS analyses. I. Dódony is thanked for the discussions on the mineralogy of titanite. S. Strmic-Palinkas is thanked for useful comments while finalizing this manuscript. The comments of two anonymous reviewers have contributed greatly to the improvement of the original manuscript. The Eugen F. Stumpfl Electron Microprobe Laboratory (University of Leoben) is thanked for the access to the EPMA laboratory.

Conflicts of Interest:
The authors declare no conflict of interest.