Tourmaline in Geothermal Systems: An Example From Darajat, Indonesia

GRC Transactions, Vol. 39, 2015 Tourmaline in Geothermal Systems: An Example From Darajat, Indonesia Thomas M. Etzel 1,2, Joseph N. Moore 2, John R. Bowman 1, Clay G. Jones 2, Rindu Grahabhakti Intani 3, Glenn Golla 3, and Greg Nash 2 1 Dept. of Geology and Geophysics, University of Utah, Salt Lake City UT 2 Energy & Geoscience Institute, University of Utah, Salt Lake City UT 3 Chevron Geothermal Indonesia, Ltd Keywords Tourmaline, water-rock interaction, vapor-dominated, Indonesia, Darajat, geochemistry Abstract The chemical compositions of two separate tourmaline populations (Stage 1 and Stage 3) from well DRJ-S1 at Darajat have been determined. This data, alongside petrologic observations, is used to improve our understanding of the evolution of the Darajat vapor-dominated geothermal system. Stage 1 tourmalines (replacing feldspar) have distinctly higher Fe/(Fe + Mg) and Na/(Na + Ca) ratios than Stage 3 tourmalines (formed in anhydrite veins). Mineral paragenesis and the high Fe content of the tourmalines suggest that Stage 1 formed in a higher temperature, fluid-dominated environment following the emplacement of subvolcanic intrusives. Ca-abundant Stage 3 tourmalines formed after descending steam condensates causing advanced argillic alteration began to neutralize. Continued anhydrite and calcite deposition (due to the heating of descending steam condensates) at shallow levels reduced porosity and permeability, impeding reservoir recharge, resulting in the current vapor-dominated system. Introduction Tourmaline is a common accessory mineral in a variety of geologic settings (Hawthorne and Henry, 1999; Henry and Dutrow, 2012), but it is rarely reported from geothermal systems. In this paper, we describe the mineral paragenesis and composition of tourmaline from the vapor-dominated geothermal system at Darajat, Indonesia. These compositions are then compared to tourmalines from a variety of geothermal and magmatic systems. Tourmaline is a borosilicate mineral that can incorporate a broad range of major and trace elements in its cation and anion sites. Its general formula is given by the expression XY 3 Z 6 (T 6 O 18 ) (BO 3 ) 3 V 3 W where: X = Na, Ca, K; Y = Li, Mg, Fe 2+, Mn 2+, Al, Ti 4+ ; Z = Mg, Al, Fe 3+, Cr 3+, V 3+ ; T = Si, Al, B; B = B, V = OH, Figure 1. Simplified well and location maps of Darajat. Wells used in the study are highlighted in red circles and red well traces. Darajat and Salak geothermal fields, highlighted in yellow on inset map, are owned and operated by Chevron Geothermal Indonesia. Karaha-Bodas is circled in red. 529

O; and W = OH, O, F (Hawthorne and Henry, 1999; Henry et al., 2011). This compositional spectrum has led to the recognition of seventeen ideal end-members (Henry et al., 2011). Geologic Setting Darajat is located in West Java, approximately 50 km southeast of Bandung and 25 km west of Garut (Figure 1). The field is owned and operated by Chevron Geothermal Indonesia, Ltd and currently has a capacity of 271 MWe. The Darajat reservoir is developed primarily in intrusive rocks that are informally referred to as the Andesite Complex, a feature recognized in many wells by early Amoseas Indonesia Inc. geologists, who first explored the field. The intrusive rocks consist primarily of diorite and biotite hornfels, representing their contact metamorphic equivalents. Lava flows are mainly andesitic in composition, although basalts are also present (Bogie and MacKenzie, 1998; Herdianita, 2006; Rejeki et al., 2010). The intrusive rocks are interpreted to be the remnants of a subvolcanic complex that was altered and eroded prior to the formation of the modern vapor-dominated geothermal system. The Andesite Complex is overlain by a thick sequence of pyroclastic and volcaniclastic rocks that cap the system. Hydrothermal Alteration and Occurrence of Tourmaline Hydrothermal alteration has been characterized in core and well cuttings from four wells DRJ-S1, DRJ-29, DRJ-21, and DRJ- 18ST1 in the Darajat field (Moore unpub., 2007); the locations of these wells are shown in Figure 1. Figures 2-3 show the distribution of hydrothermal alteration minerals (including tourmaline) in wells DRJ-S1 and DRJ-29 based on petrographic and X-ray diffraction analyses. The hydrothermal minerals can be divided into four distinct stages based on textural relationships and cross-cutting veins first described by Moore (unpub., 2007) (Table 1). Stage 1 mineralization represents the development of a liquid-dominated system. This stage includes argillic, phyllic and propylitic alteration that is interpreted to postdate the high-temperature alteration occurring in the Andesite Complex. As the system waned, zeolites (wairakite and laumontite), calcite and anhydrite were deposited. Stage 2 is represented by the formation of chalcedony after quartz. In places, the chalcedony forms overgrowths on earlier formed quartz as well as on calcite, anhydrite and wairakite. The formation of chalcedony is interpreted to reflect the boiling of an evolving liquid-dominated geothermal system and the initial formation of the vapor-dominated regime in a manner similar to that described by Moore et al. Table 1. Stages of alteration in the Darajat system. Modified from Moore (unpub., 2007). Stage Process Alteration Type Mineralogy 1 Upwelling of high temperature NaCl fluids; liquid-dominated 2 Boiling; conversion to vapordominated conditions 3 Descent and neutralization of steam condensates Depth (meters) 300 400 500 600 700 800 900 1000 1100 Shallow: argillic-phyllic alteration Deep: propylitic Silicification Shallow: advanced argillic; sealing of marginal fractures Lithology Alteration Style 1 2 3 4 5 Quartz Smectite I/S C/S WEll: DRJ-S1 Illite Chlorite Pyrite Shallow: smectite, I/S, illite Deep: epidote, chlorite, actinolite, biotite, tourmaline Chalcedony, quartz Shallow: pyrophyllite, diaspore, anhydrite, calcite, tourmaline Deep: Wairakite 4 Surficial oxidation of H 2 S Advanced argillic alteration Alunite, kaolinite Epidote Actinolite Adularia Tourmaline Anhydrite Calcite Pyrophyllite Fluorite Wairakite Prehnite Figure 2. Overview of hydrothermal alteration mineralogy in well DRJ-S1 based on petrographic observations and XRD analyses. Alteration styles are as follows: 1 = argillic, 2 = silicification, 3 = phyllic, 4 = propylitic and 5 = advanced argillic. Black lines indicate the mineral has been observed at that interval; thickness of the line indicates abundance. Depth is in meters from the surface. I/S = interlayered illite-smectite; C/S = interlayered chlorite-smectite. 1200 1300 1400 Lithology Key Tuff Diorite Lahar Biotite Hornfels Abundances >20% 1-5% 5-20% <1% 530

(2008) for the formation of the nearby KarahaTelaga Bodas system. Stage 3 mineralization is represented by the continued deposition of calcite and anhydrite. Stage 4 is represented by continued acid leaching and further silicification; the timing of this event is poorly understood. Tourmaline is widely distributed throughout both wells in trace to minor amounts (Figs 2-3). This study focuses on tourmaline in well DRJ-S1. At least two petrographically distinct tourmaline populations can be identified and are associated with Stages 1 and 3. Stage 1 tourmaline partially replaces plagioclase and clinopyroxene in the diorites. It occurs in shades of blue, green and brown pleochroic crystals, often in weakly altered intrusive rocks (Figure 4). In contrast, Pyrite Zunyite Fluorite Pyrophyllite Diaspore Kaolinite Anhydrite Calcite Wairakite Tourmaline Actinolite Chlorite Epidote Illite I/S C/S Smectite 12345 WEll: DRJ-29 Quartz Alteration Type Lithology Figure 3. Overview of hydrothermal alteration mineralogy in well DRJ-29OH based on petrographic observations and XRD analyses. Alteration styles are as follows: 1 = argillic, 2 = silicification, 3 = phyllic, 4 = propylitic and 5 = advanced argillic. Black lines indicate the mineral has been observed at that interval; thickness of the line indicates abundance. Depth is in meters from the surface. I/S = interlayered illite-smectite; C/S = interlayered chlorite-smectite. Depth (meters) Etzel, et al. 0 500 1000 2000 2500 Lithology Key Figure 4. Tourmaline (Tur), epidote (Ep) and pyrite in the altered diorite. Photomicrographs of core collected from well DRJ-29 at 944.9 m. A) Plane polarized light. B) Crossed nicols. Plane of view is 0.7mm. Abundances Andesite Lava >20% Lahar Tuff 5-20% Diorite 1-5% <1% Figure 5. Open space in tuff filled first by acicular tourmaline (Tur) before wairakite (Wrk) and anhydrite (Anh). Photomicrograph of core collected from well DRJ-S1 at 1807 m. A) Plane polarized light. B) Crossed nicols. Field of view is 1.4 mm. Anh = anhydrite; Tur = tourmaline; Wrk = wairakite. 531

Stage 3 tourmaline forms radiating aggregates of crystals that are commonly distributed on the margins of anhydrite veins above the intrusive rocks (Figure 5). These tourmaline crystals typically lack the strong pleochroism and color characteristic of the deeper crystals. Pleochroism, if present at all, is characteristically weak in shades of light green. Table 2. Calculated tourmaline mineral formulas. Ions are grouped according to site location. Ion values are averages of multiple spot locations within a grain. n = indicates the number of spot analyses contributing to the averaged value. indicates X-site vacancy. Boron is assumed to be stoichiometric (B = 3 apfu). Sample = 975.4-1 975.4-7 975.4-5 821.4-1 821.4-2 670.6-2 670.6-1 670.6-4 670.6-3 807.7-1 807.7-2 n = 3 2 5 4 3 5 5 4 3 3 2 Stage = 1 1 1 3 3 3 3 3 3 1 1 X-Site Na 0.521 0.618 0.490 0.436 0.383 0.318 0.355 0.312 0.384 0.304 0.431 Ca 0.770 0.224 0.170 0.333 0.514 0.306 0.230 0.195 0.622 0.110 0.163 K 4 0 8 0 0.058 0.086 0.169 0.099 0.099 0.255 7 0.194 0.155 0.332 0.229 0.186 0.320 0.280 0.398 0.128 0.332 0.399 Y-Site Fe 1.412 1.303 1.446 0.276 0.554 0.716 0.491 0.657 0.633 1.075 0.868 Mg 1.317 1.497 1.253 1.598 1.736 1.973 1.598 1.674 1.648 1.609 1.528 Mn 0.015 0.014 0.022 4 0.021 0.031 0.019 0.020 0.032 0.040 0.012 Al 0.057 0.123 0.179 7 0.523 0.225 0.384 0.108 0.483 0.372 0.581 Ti 0.434 0.029 0.031 0.038 7 0.025 0.017 0.300 0.011 0.021 0 Z-Site Al 5.697 5.993 6.000 6.000 6.000 5.892 6.000 5.725 6.000 5.674 6.000 T-Site Si 6.060 5.982 6.068 6.577 6.159 6.137 6.492 6.517 6.193 6.210 5.986 Al 9 0.060 0 0 0 0 0 0.106 0 0 0.036 B-Site B 3.000 3.000 3.000 3.000 3.000 3.000 3.000 3.000 3.000 3.000 3.000 W-Site O 0.669 0.670 0.666 0.608 0.634 0.497 0.536 0.558 0.557 0.635 0.707 F 0 5 0 0.022 0.033 0.247 0.177 0.149 0.144 0 0 Cl 0.015 0 0.013 0.067 0.016 8 0.019 0.014 0.021 0.055 0.017 OH 3.335 3.335 3.333 3.304 3.317 3.248 3.268 3.278 3.317 3.354 3.354 Chemical Characteristics of the Tourmaline Electron microprobe analyses were conducted at the University of Utah using a Cameca BX50 automated electron microprobe (EMP) with four wavelength dispersive spectrometers, anaccelerating voltage of 15 kv, 20 na sample current and a 1-micron beam diameter. Standards included albite (Na), sanidine (Al, K), diopside (Mg, Ca, Si), rutile (Ti), rhodonite (Mn), hematite (Fe), fluorite (F), and tugtupite (Cl). Structural formulae were calculated on the basis of the general formula XY 3 Z 6 (T 6 O 18 )(BO 3 ) 3 V 3 W by normalizing the sum of T+Y+Z to 15 (15 cations). Light elements (B, H, O) could not be analytically determined; boron was assumed to be stoichiometric (i.e., B = 3) following the justification provided by Henry and Dutrow (1996); oxygen and hydrogen were determined following the procedure originally described by Grice and Erict (1993). Li, assumed to be a negligible component in these tourmaline samples, has not been accounted for. All iron is assumed to be Fe 2+. Tourmaline was analyzed in core samples from four depth intervals in well DRJ-S1. Tourmalines associated with Stage 1 were analyzed from intervals 975.4 m and 807.7 m while Stage 3 tourmalines were analyzed from intervals 821.4 m and 670.1 m; calculated mineral chemistry is presented in Table 1. X-vacant Group 1.00 X-Site vacancy Interval 975.4 (Stage 1) Interval 821.4 (Stage 3) Interval 807.7 (Stage 1) Interval 670.1 (Stage 3) 1.00 Ca 2+ Calcic Group Alkali Group 1.00 Na 1+ (K 1+ ) Larderello, Italy Karaha - Telaga Bodas, Indon San Jorge porphyry, Argentin Hnilec granite, Slovakia The Geysers, Califronia Figure 6. General chemical composition diagram based on alkali content of tourmaline (Hawthorne and Henry, 1999). Filled circles represent individual analyses of a sample from the Darajat system. Open symbols represent published data from other sites. 532

Na/(Ca+Na) 1.00 0.90 0.80 0.70 0.60 0.40 0.30 0.20 0.10 a) Dravite Uvite Schorl Feruvite 0.10 0.20 0.30 0.40 0.60 0.70 0.80 0.90 1.00 Fe/(Mg+Fe) 0.70 0.60 0.40 0.30 0.20 0.10 NaMg( Al) -1 FeAl -1 TiMgAl -2 NaAl(CaMg) -1 Al(OH)(CaMgO) -1 Al 2 (MgSi) -1 AlOMg -1 (OH) -1 4.00 4.50 5.00 5.50 6.00 6.50 7.00 Al Figure 7. Composition diagrams constructed to further classify tourmalines into accepted categories and depict possible exchange vectors. a) Na/(Ca + Na) ratios plotted as a function of Fe/(Mg + Fe) ratios for each sample; displays the chemical heterogeneity between the two assemblages. b) Al vs. Na plot. All symbols follow the key in Figure 6. Na 0.80 b) 7.50 Tourmaline compositions are plotted in figures 6 and 7. For comparison, compositions of tourmaline from Karaha- Telaga Bodas (Moore et al., 2004), The Geysers (Jones et al., in prep), Larderello (Cavarretta and Puxeddu, 1990) and two intrusive systems (Dill et al., 2012; Jiang et al., 2008) are also shown. The tourmalines from Karaha and The Geysers occur in association with mineral assemblages characteristic of Stage 3 at Darajat, and the tourmalines from Larderello and the two intrusive systems occur in mineral associations characteristic of Stage 1 at Darajat. Chemical variation in Darajat tourmalines is most noticeable in Al, Ca, Na, Mg, and Fe tot while K, Ti, Mn, F and Cl concentrations are minimal and show little variation with depth or between stages. Compositionally, most tourmaline samples plot within the alkali group (Figure 6). Three analyses with no computed X-site vacancies plot within the calcic group on the right arm of the ternary diagram. These analyzed grains are quite small, so this high Ca concentration could, in part, be the result Cabearing minerals such as anhydrite and wairakite inadvertently contributing to those tourmaline analyses. Tourmalines have been further subdivided into specific end-member categories (figure 7) following an accepted computational procedure of Henry et al. (2011). Computed tourmaline compositions range from schorl-dravite (alkali abundant subcategories) to uvite-feruvite compositions (calcic abundant subcategories). Stage 1 tourmalines are intermediate schorl-dravite in composition. Tourmalines from interval 975.4 m are classified as schorl, save one Ca-rich analysis that is classified as uvite. Tourmalines from interval 807.7 are dravitic. Stage 3 tourmalines range from dravite to uvite; tourmalines from interval 670.1 m are predominately dravitic, with a small number of calcic-rich specimens plotting in the uvite field, while tourmalines from interval 821.4 m range from intermediate dravite-uvite to dravite in composition. Stage 1 tourmalines have distinctly higher and more variable Fe/(Mg + Fe) ratios, with interval 975.4 m having the overall highest values (Figure 7a), than Assemblage 3 tourmalines. Henry and Dutrow (1996) pointed out that metagranitoid tourmalines are typically enriched in Fe. For example, main stage metagranitoid tourmalines in Hnilec granites (Jiang et al., 2008) have very Fe-rich compositions (Fig. 7a). Stage 1 tourmalines at Darajat formed during the early stages of hydrothermal alteration shortly after pluton emplacement, and are therefore related to metamorphism in the contact aureole. Tourmalines documented from similar occurrences at Karaha-Telaga Bodas, Indonesia (Moore et al., 2008), Larderello, Italy (Cavarretta and Puxeddu, 1990) and at the Geysers, California have similar Fe/(Mg + Fe) ratios (Figure 7), attesting to the metamorphic nature of Stage 1 tourmalines at Darajat. With one exception, Stage 1 tourmalines have higher Na/(Na + Ca) ratios than Stage 3 tourmalines (Figure 7a), but Stage 3 tourmalines have more variable Na/(Na + Ca) ratios than Stage 1 tourmalines. Both groups of tourmaline exhibit significant variations in Na and Al contents (Fig. 7b). Na is negatively correlated with Al in Stage 1 tourmalines, with two exceptions, and Na is positively correlated with Al in Stage 3 tourmalines, primarily within the tourmalines from interval 670.1 m. The NaAl(CaMg) -1 exchange vector can explain in general the positive correlation of Na and Al, and variable Na/(Ca + Na) (Na-Ca substitution) on the X-site in Stage 3 tourmalines (Figure 7b). However because Na/ (Ca + Na) ratios are relatively constant in Stage 1 tourmalines, the variations in Na content and negative correlation of Na with Al in Stage 1 tourmalines are more likely the result of the NaMg( Al) -1 exchange vector (Fig. 7b). Natural and laboratory-generated data have revealed a correlation between increasing Na content of tourmaline and increasing temperature in a metamorphic environment (von Goerne et al., 2001). The Na-enriched Stage 1 tourmalines likely formed at higher reservoir temperatures than did Stage 3 tourmalines, although fluid inclusion data are needed for validation. However, reservoir rock chemistry will also have an important impact on tourmaline composition. Stage 1 533

tourmalines formed at the expense of magmatic minerals abundant in Na, while Stage 3 tourmalines formed in anhydrite veins sourcing cations from a Ca-enriched hydrothermal fluid. Despite the occurrence of fluorite and zunyite in Stage 3, fluorine is a relatively insignificant component of tourmalines except for interval 670.1 m. Although Henry and Dutrow (2011) found that as X-site charge decreases F content also decreases, our limited data set fails to produce a similar observation. The significance of F- and B-bearing minerals however is their suggestion of a magmatic contribution to hydrothermal fluids during the formation of assemblage 3. Summary and Conclusions A significant feature of the mineral assemblages in the tourmaline-bearing Darajat wells is the presence of minerals characteristic of advanced argillic alteration (Figure 8). These minerals include diaspore, pyrophyllite, zunyite and kaolinite. In places, alunite and fluorite are also present. Calcite and anhydrite are common, as is wide-spread silicification and intense cation leaching. Pyrophyllite and diaspore are diagnostic of temperatures exceeding ~250 0 C and acidic conditions (ph <2-3). Tourmaline and fluorite are temporally associated with the advanced argillic alteration assemblages but persist to greater depths. Stage 3 tourmalines exist as a fracture-filling phase alongside late-stage calcite and anhydrite overprinting earlier propylitic assemblages containing quartz, epidote and actinolite. Due to the retrograde solubility of both calcite and anhydrite, tourmalinebearing anhydrite veins were deposited by downward percolating acid-sulfate waters that developed as steam condensate Figure 8. Correlation of tourmaline occurrences and advanced argillic alteration assemblages at Darajat. percolated downward. Water-rock interactions acted to neutralize the descending acid-condensates by introducing Ca ions, leached from the plagioclase abundant wall-rock, into solution. This resulted in generally higher Ca/(Ca+Na) ratios for stage 3 tourmalines. Moore et al. (2008) described similar relationships at Karaha-Telaga Bodas. They observed tourmaline in well T-2, the well drilled closest to the acid Telaga Bodas Lake. This lake is interpreted to overlie a magmatic vapor chimney whose composition is strongly influenced by magmatic gases, particularly HCl and SO 2. Advanced argillic alteration occurs in the upper part of T-2, and in addition to tourmaline, fluorite and native sulfur are locally present. These minerals were not found in other wells at Karaha-Telaga Bodas. Fluid inclusions trapped in anhydrite indicate that tourmaline deposition occurred at temperatures of about 235 C. Based on these relationships, it was concluded that the B and F were magmatic in origin. The origin of the B and F at Darajat is less certain, although the close association with diaspore, pyrophyllite, and the F-bearing minerals zunyite and fluorite, is also strongly suggestive of a magmatic origin. Darajat is currently a vapor-dominated system. Hydrothermal mineral assemblages identified in this work and previous studies (Moore unpub., 2007; Herdianita, 2006) suggest the system was once liquid dominated. Using paragenetic relationships, the evolution of this geothermal system is summarized in figure 9 and below. Stage 1 alteration minerals document an early liquid dominated system. Shallow argillic-phyllic alteration and deep propylitic alteration indicates temperatures increased with depth. The presence of tourmaline in proximity to the contact near the diorite intrusions suggests: 1) temperatures exceeded 300 o C (Corbett and Leach, 1998) and 2) that magmatic fluids introduced B into the system. Widespread chalcedony deposition, associated with Stage 2, signifies a change in the physio-chemical conditions of the reservoir. Chalcedony in geothermal systems indicates temperatures below 180 C (Fournier, 1985) and is therefore rarely observed in the propylitically altered high temperature portions of a system. Fluid inclusion data from this generation of chalcedony is needed to constrain temperatures during this time. However, vapor rich inclusions trapped in quartz 534

Figure 9. North to south cross-section of the Darajat field. The reservoir consists of diorite intrusions and Stage 1 propylitic alteration (green). The tectonic erosional surface (red line) resulted following a flank collapse. Younger tuffs, lahars and subvolcanic intrusions were then deposited. Ascending acid-rich vapors condensed (white arrows) and began to percolate downward through faults and fractures; these condensates formed advanced argillic alteration halos (dark green) and Stage 3 anhydrite-tourmaline veins. crystals observed by Moore (2007) suggest a rapid reduction in reservoir pressures, allowing fluids to flash, thus creating vapor to be trapped in inclusions during chalcedony deposition. Rapid depressurization resulting in the deposition of chalcedony also occurred in Karaha Telaga Bodas; this was interpreted to have resulted from a flank collapse event (Moore et al., 2008). It is possible this too has occurred at Darajat. Alternatively, movement along the Kendang fault could have led to rapid depressurization. In either case, Stage 2 marks the transition from liquid-dominated to vapor-dominated. During stage 3 alteration ascending magmatic volatiles condensed, forming acidic fluids that migrated laterally and downward through fractures and faults (fig. 9). These migrating acidic fluids enriched in H 2 S, SO 4 and CO 2 interacted with the wall-rock, resulting in advanced argillic alteration at shallow depths. This interaction introduced Ca ions into solution, thus gradually increasing the ph. As these slightly acid-to-near neutral waters continued to descend below the advanced argillic horizon, increasing temperatures resulted in calcite and anhydrite deposition. Physiochemical conditions were such that boric acid in solution also reacted with Ca ions and the surrounding wall-rock to crystallize stage 3 tourmalines. Vapors enriched in SO 2, CO 2, B and F present in the steam reservoir resulted in stage 3 mineral assemblages. Continued fracture filling decreased the porosity and permeability, impeding reservoir recharge, therefore sustaining vapor dominated conditions that still persist. Acknowledgements The authors thank the staff and management of Chevron Geothermal Indonesia, Ltd for providing the samples and well log data for this study. We would also like to thank Wil Mace for his help with the electron microprobe analyses. References Bogie, I., and K.M. MacKenzie, 1998, The application of a volcanic facies model to an andesitic stratovolcano hosted geothermal system at Wayang Windu, Java, Indonesia: Proceedings 20 th New Zealand Geothermal Workshop. Cavarretta, G., and M. Puxeddu, 1990, Schorl-dravite-ferridravite tourmalines deposited by hydrothermal magmatic fluids during early evolution of the Larderello geothermal field, Italy: Economic Geology, v. 85, p. 1236-1251. Corbett, G.J., and T.M. Leach, 1998, Southwest Pacific gold-copper systems: Structure, alteration and mineralization: Special Publication 6, Society of Economic Geologists, 238 p. Dill, H.G., M.M. Garrido, F. Melcher, M.C. Gomez, and L.I. Luna, 2012, Depth-related variation of tourmaline in the breccia pipe of the San Jorge porphyry copper deposit, Mendoza, Argentina: Ore Geology Reviews, v. 48, p. 271-277. Fournier, R.O., 1985, The behavior of silica in hydrothermal solutions, in Geology and Geochemistry of Epithermal Systems, Berger, B. R., and Bethke, P. M. (eds): Reviews in Economic Geology, v. 2, p. 45-62. Grice J.D., and T.S. Erict, 1993, Ordering of Fe and Mg in the tourmaline crystal structure: the correct formula: Neues Jahrbuch fuer Mineralogie Abhandlungen, v. 165, p. 245-266. Hawthorne, F.C., and D.J. Henry, 1999, Classification of the minerals of the tourmaline group: European Journal of Mineralogy, v. 11, p. 201-215. 535

Henry, D.J., and B.L. Dutrow, 1996, Metamorphic tourmaline and its petrologic applications, in Boron Mineralogy, Petrology and Geochemistry, E. S. Grew and L. M. Anovitz (eds): Reviews in Mineralogy, v. 33, 503-557. Henry, D.J., and B.L. Dutrow, 2011, The incorporation of fluorine in tourmaline: Internal crystallographic controls or external environmental influences?, The Canadian Mineralogist, v. 49, p. 41-56. Henry, D.J., and B.L. Dutrow, 2012, Tourmaline at digenetic to low-grade metamorphic conditions: its petrologic applicability: Lithos, v. 154, p. 16-32. Henry, D.J., M. Novak, F.C. Hawthorne, A. Ertl, B.L. Dutrow, P. Uher, and F. Pezzotta, 2011, Nomenclature of the tourmaline-supergroup minerals: American Mineralogist, v. 96, p. 895-913. Herdianita, N.R., 2006, The evolution of the Darajat geothermal field, West Java, Indonesia from water to vapour dominant: unpublished internal report. Jiang, S.Y., M. Radvanec, E. Nakamura, M. Palmer, K. Kobayashi, H.X. Zhao, and K. D. Zhao 2008, Chemical and boron isotopic variations of tourmaline in the Hnilec granite-related hydrothermal system, Slovakia: constraints on magmatic and metamorphic fluid evolution: Lithos, v. 106, p. 1-11. Moore, J.N., R.G. Allis, M. Nemcok, T.S. Powell, C.J. Bruton, P.E. Wannamaker, I.B. Raharjo,, and D.I. Norman, 2008, The evolution of volcanohosted geothermal systems based on deep wells from Karaha Telaga Bodas, Indonesia: American Journal of Science, v. 308, p. 1-48. Moore, J.N., B.W. Christenson, R.G. Allis, P.R.L. Browne, S.J. Lutz, 2004, The mineralogical consequences and behavior of descending acid-sulfate waters: an example from the Karaha Telaga Bodas geothermal system, Indonesia: The Canadian Mineralogist, v. 42, p. 1483-1499. Rejeki, S., D. Rohrs, G. Nordquist, and A. Fitriyanto, 2010, Geologic conceptual model update of the Darajat geothermal field, Indonesia: Proceedings World Geothermal Congress, p. 1-6. von Goerne, G., G. Franz, and W. Heinrich, 2001, Synthesis of tourmaline solid solutions in the system Na 2 O MgO - Al 2 O 3 SiO 2 B 2 O 3 H 2 O HCl and the distribution of Na between tourmaline and fluid at 300 to 700 o C and 200 MPa, Contributions to Mineralogy and Petrology, v. 141, p. 160-173. 536