Evaluating fluid-rock interactions in geothermal and contact metamorphic systems

The study of fluid-rock interactions provides insight into subsurface geologic processes, such as diagenesis, hydrothermal alteration and metamorphism. Understanding and predicting these interactions also helps us assess the geologic impact of hydrocarbon recovery and geothermal production and injection. Therefore, the study of fluid-rock interactions has both geologic and economic impact. At the Dixie Valley geothermal field, NV, precipitated calcite and aragonite within a production well trapped boiling fluids in fluid inclusions. The trapped gases were analyzed and shown to be compositionally different than those sampled at the well head. The inclusions trapped a greater ratio of light gases CH4 and H2 to CO2 than those sampled at the well head. This result indicates that the fluid inclusions trapped the initial steam fraction during boiling. Declining performance of injection wells at the Coso and Salton Sea geothermal fields, CA, were found to result from mineral deposition in the near-wellbore environment during fluid injection. At Coso, opal-A and minor calcite scale mineral precipitates were found in cuttings from wells drilled near previously operating injection wells. At the Salton Sea, cuttings from a deepened injection well contained banded barite, fluorite, amorphous silica, and minor anhydrite scales. Mineral precipitation was modeled and predicted with the non-isothermal reactive transport modeling code TOUGHREACT. Geochemical simulations were also performed to predict the consequences of injecting H2SO4 modified fluid for mitigating silica precipitation at Coso using TOUGHREACT. The models predict that silica precipitation will be reduced significantly by maintaining pH of 5 or less. This can be accommodated in the models by reducing the kinetic rate constant for silica precipitation. TOUGHREACT simulations also predict that corundum proppants will be chemically stable under geothermal conditions. The formation of talc in the outer aureole of the Alta Stock does not define a regular isogradic surface, unlike isograds in the inner aureole. Examination of mineral and fluid stabilities in the H2O-CO2-NaCl system shows that several fluid evolution scenarios, including fluid immiscibility, may produce the observed talc heterogeneity.

EVALUATING FLUID-ROCK INTERACTIONS IN GEOTHERMAL AND CONTACT METAMORPHIC SYSTEMS by Kristie S. McLin A dissertation submitted to the faculty of The University of Utah in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Geology Department of Geology and Geophysics The University of Utah May 2012 Copyright © Kristie S. McLin 2012 All Rights Reserved The University of Utah Graduate School STATEMENT OF DISSERTATION APPROVAL The dissertation of has been approved by the following supervisory committee members: , Chair Date Approved , Member Date Approved , Member Date Approved , Member Date Approved , Member Date Approved and by , Chair of the Department of and by Charles A. Wight, Dean of The Graduate School. Kristie S. McLin John R. Bowman 3/5/2012 Joseph N. Moore 3/5/2012 John P. Kaszuba 2/6/2012 Erich U. Petersen 3/5/2012 Ronald L. Bruhn 3/5/2012 D. Kip Solomon Geology and Geophysics ABSTRACT The study of fluid-rock interactions provides insight into subsurface geologic processes, such as diagenesis, hydrothermal alteration and metamorphism. Understanding and predicting these interactions also helps us assess the geologic impact of hydrocarbon recovery and geothermal production and injection. Therefore, the study of fluid-rock interactions has both geologic and economic impact. At the Dixie Valley geothermal field, NV, precipitated calcite and aragonite within a production well trapped boiling fluids in fluid inclusions. The trapped gases were analyzed and shown to be compositionally different than those sampled at the well head. The inclusions trapped a greater ratio of light gases CH4 and H2 to CO2 than those sampled at the well head. This result indicates that the fluid inclusions trapped the initial steam fraction during boiling. Declining performance of injection wells at the Coso and Salton Sea geothermal fields, CA, were found to result from mineral deposition in the near-wellbore environment during fluid injection. At Coso, opal-A and minor calcite scale mineral precipitates were found in cuttings from wells drilled near previously operating injection wells. At the Salton Sea, cuttings from a deepened injection well contained banded barite, fluorite, amorphous silica, and minor anhydrite scales. Mineral precipitation was modeled and predicted with the non-isothermal reactive transport modeling code TOUGHREACT. Geochemical simulations were also performed to predict the consequences of injecting H2SO4 modified fluid for mitigating silica precipitation at Coso using TOUGHREACT. The models predict that silica precipitation will be reduced significantly by maintaining pH of 5 or less. This can be accommodated in the models by reducing the kinetic rate constant for silica precipitation. TOUGHREACT simulations also predict that corundum proppants will be chemically stable under geothermal conditions. The formation of talc in the outer aureole of the Alta Stock does not define a regular isogradic surface, unlike isograds in the inner aureole. Examination of mineral and fluid stabilities in the H2O-CO2-NaCl system shows that several fluid evolution scenarios, including fluid immiscibility, may produce the observed talc heterogeneity. iv For Ryan, Liam, Oona, and all of my very supportive family. TABLE OF CONTENTS ABSTRACT.......................................................................................................................iii ACKNOWLEDGEMENTS................................................................................................ix Chapter 1 MINERALOGY AND FLUID INCLUSION GAS CHEMISTRY OF PRODUCTION WELL MINERAL SCALE DEPOSITS AT THE DIXIE VALLEY GEOTHERMAL FIELD, USA……………………………..….......1 Abstract……………………………………………………………………………......1 Introduction……………………………………………………………………………2 Methods………………………………………………………………………………10 Results……………………………………………………………………………..…15 Discussion……………………………………………………………………………29 Conclusion………………………………………………………………………...…34 Acknowledgements……………………………………………………………….….35 References…………………………………………………………………….……...35 2 MODELING THE GEOCHEMICAL EFFECTS OF INJECTION AT COSO GEOTHERMAL FIELD, CA; COMPARISION WITH FIELD OBSERVATIONS…………………………………………………………………...39 Abstract……………………………………………………………………………....39 Introduction..................................................................................................................40 The Coso Geothermal Field.........................................................................................41 Discussion....................................................................................................................50 Mineral Scale Summary...............................................................................................51 One Dimensional Reactive Transport Modeling.........................................................52 Results.........................................................................................................................57 Conclusions and Future Work.....................................................................................63 References....................................................................................................................64 3 MODELING THE GEOCHEMICAL EFFECTS OF INJECTING pH MODIFIED FLUIDS AT COSO GEOTHERMAL FIELD, CA...........................66 Abstract.......................................................................................................................66 Introduction..................................................................................................................67 Study Area...................................................................................................................69 Observations from Well Cuttings................................................................................74 Previous Coso Modeling Studies.................................................................................76 Modeling Approach.....................................................................................................79 Results.........................................................................................................................87 Discussion..................................................................................................................102 Conclusions................................................................................................................109 References..................................................................................................................110 4 MODELING THE GEOCHEMICAL EFFECTS OF INJECTION AT THE SALTON SEA GEOTHERMAL FIELD, CA: COMPARISON WITH FIELD OBSERVATIONS..................................................113 Abstract.....................................................................................................................113 Introduction...............................................................................................................114 Salton Sea Geothermal Field.....................................................................................114 Observations..............................................................................................................116 Modeling Approach...................................................................................................119 Results........................................................................................................................125 Conclusions................................................................................................................126 References..................................................................................................................129 5 GEOCHEMICAL MODELING OF WATER-ROCK-PROPPANT INTERACTIONS......................................................................................................130 Abstract.....................................................................................................................130 Introduction...............................................................................................................131 Previous Studies........................................................................................................132 Batch Model Setup....................................................................................................133 Flow Model Setup.....................................................................................................134 Results.......................................................................................................................139 Discussion.................................................................................................................142 Conclusions...............................................................................................................144 Acknowledgements...................................................................................................145 References.................................................................................................................145 vii 6 EVALUATING FLUID-ROCK REACTIONS IN THE TALC ZONE OF THE ALTA CONTACT AUREOLE.........................................................................147 Abstract......................................................................................................................147 Introduction................................................................................................................148 Geologic Setting and Metamorphism in the Alta Aureole.........................................149 Scenarios for the Development of the Outer Alta Aureole........................................161 Fluid Immiscibility....................................................................................................163 Modeling Approach...................................................................................................163 Results........................................................................................................................165 Discussion: Fluid Evolution.......................................................................................172 Conclusions................................................................................................................184 References..................................................................................................................185 Appendix A METHOD OF CALCULATING LIQUID AND VAPOR CONCENTRATIONS OF GASES WITH BOILING..............................................188 B PHOTOMICROGRAPHS OF AMORPHOUS SILICA SCALE SAMPLES FROM COSO WELLS 68-20RD AND 68B-20RD..............................190 C PHOTOMICROGRAPHS OF MINERAL SCALE FROM SALTON SEA WELL ELMORE IW-3 RD......................................................................................233 D PHASE DIAGRAMS FOR THE H2O-CO2-NaCl SYSTEM...................................255 viii ACKNOWLEDGEMENTS I would like to thank my advisors, Dr. John R. Bowman and Dr. Joseph N. Moore, for making this dissertation possible. Their patience and attention to detail are much appreciated. I would also like to thank my committee members, Dr. John P. Kaszuba, Dr. Erich U. Petersen, and Dr. Ronald L. Bruhn, who put considerable effort into my degree program, as well. My family has been so considerate, patient, and supportive. This is especially true of my husband Ryan, son Liam, and daughter Oona. Nana (Jody) and Dadu (Greg) Diehl, (Aunt) Elisa and (Aunt) Alaina Diehl, and Grandma (Debbie) and Grandpa (Steve) McLin all made considerable efforts to take care of us when we needed it the most. Without their help, this would not have been possible, so I thank them. My grandparents Joseph and Dolores Sobieski and Roland and Edith Diehl also provided much needed support along the way. I would like to thank the US Department of Energy (DE-FG36-04GO14292 & - FG36-08GO18189) and the Petroleum Research Fund for the funding that made my dissertation projects possible. Last, but certainly not least, I want to thank Dr. David I. Norman (1940-2008), who inspired me to pursue the path that led me here. CHAPTER 1 MINERALOGY AND FLUID INCLUSION GAS CHEMISTRY OF PRODUCTION WELL MINERAL SCALE DEPOSITS AT THE DIXIE VALLEY GEOTHERMAL FIELD, USA Abstract At the Dixie Valley geothermal field, Nevada, USA, fluid boiling triggered the precipitation of carbonate scale minerals in concentric bands around tubing inserted into production well 28-33. When the tubing was removed, this mineral scale was sampled at 44 depth intervals between the well head and 1227 m depth. These samples provide a unique opportunity to evaluate the effects of fluid boiling on the scale mineralogy and geochemistry of the vapor and liquid phase. In this study, the mineralogy of the scale deposits and the composition of the fluid inclusion gases trapped in the mineral scales were analyzed. The scale consists mainly of calcite from 670-1112 m depth and aragonite from 1125 to 1227 m depth, with traces of quartz and Mg smectite. Mineral textures, including hopper growth, twinning, and fibrous growth in the aragonite and banded deposits of fine grained calcite crystals, are the result of progressive boiling. The fluid inclusion noncondensable gas was dominated by CO2. However, significant variations in He relative to N2 and Ar provide evidence that the geothermal reservoir consists of mixed source deeply circulating reservoir water and shallow, air saturated meteoric water. Gas analyses for many inclusions also showed higher CH4 and H2 relative to CO2 than measured in gas sampled from this well, other production wells, and fumaroles. These inclusions are interpreted to have trapped CH4 and H2 enriched gas resulting from early stages of boiling. Introduction Geothermal systems provide natural laboratories for understanding the behavior and processes affecting fluids within the crust. The gas composition of geothermal fluids can provide important information on the fluid origin, temperature, and evolution (e.g. Giggenbach, 1980, 1986). Similarly, the gas composition of geothermal fluid inclusions can be used to trace fluid sources and histories. CO2 is the dominant gas in geothermal waters, and where other data is absent, is also considered to be the dominant gaseous species in fluid inclusions from geothermal and epithermal systems (Hedenquist and Henley, 1985). However, investigations by Norman and Musgrave (1994), Graney and Kesler (1995), Norman et al. (1996, 1997, 1999), and Moore et al. (2000, 2001) suggest other gases (CH4, H2) may be present in significant concentrations. Moore et al. (2000, 2001) further suggest that higher concentrations of these light gases in fluid inclusions represent preferential partitioning of these gases into the vapor phase during early boiling. As with CO2, these gases can significantly influence the ice melting temperature and therefore apparent salinities of the inclusion fluid, yielding estimates of salinities that are too high. 2 Boiling occurs in geothermal systems, either naturally or due to pressure changes during production. Phase separation within production wells can result in the deposition of carbonate and silicate minerals. Simmons and Christenson (1994) and Simmons and Browne (2000) described calcite deposition in boiling environments in geothermal wells. These scale minerals can significantly reduce flow in wells, and consequently scale inhibitor is used to mitigate their formation. However, failure of the inhibitor can lead to rapid mineral deposition, as was found in the Dixie Valley geothermal field production well 28-33. Carbonate and silicate minerals formed around tubing that was inserted into the well to deliver scale inhibitor. This scale was sampled at 44 depth intervals from inside the well head down to 1227 m depth when the tubing was removed. Because these mineral deposits spanned the boiling column as estimated by pressure distributions, the samples provide an opportunity to examine the effects of boiling on the scale mineralogy and textures. It also provides an unusual opportunity to compare the composition of gases measured at the well head with the compositions of gases trapped by fluid inclusions in the scale minerals during production of these fluids. There are no previous studies that have related the well gas chemistry to fluid inclusion gas chemistry in geothermal fields. In this study, we directly measure and relate gases trapped in fluid inclusions from these scale minerals to the chemistry of the gases sampled from well 28-33. Geologic Setting Dixie Valley is located 160 km northeast of Fallon, Nevada between the Stillwater and Clan Alpine Ranges (Fig. 1.1A). The valley trends NNE and is 120 km long by 20 km wide. Dixie Valley is located in the southern end of a region of elevated 3 heat flow known as the Battle Mountain High (Bergfeld et al., 2001). The geothermal reservoir is hosted in Triassic to Jurassic marine quartzite, siltstone, shale, and volcaniclastic rocks. The reservoir is overlain by a suite of oceanic crustal rocks that include gabbro, diorite, and basalt (Speed, 1976; Weibel, 1987; Lutz et al., 1997). The Triassic and Jurassic units are imbricated by three thrust faults and intruded by Cretaceous granodiorite (Goff et al., 2002). Oligocene ignimbrites and the Miocene Table Mountain Basalt overlie these older rocks in the valley (Goff et al., 2002). Most geologic units are extensively altered by hydrothermal activity (Goff et al., 2002). Dixie Valley was formed by Basin and Range tectonic events. The geothermal field is classified as an "extensional" geothermal resource (Kennedy and van Soest, 2006). Fluid production is from the Stillwater fault zone (Fig. 1.1B). Thermal gradients indicate that heat is transported advectively through the fault system by upward flow of geothermal fluids (Blackwell et al., 2000, 2002). The Dixie Valley power plant is located southeast of the Stillwater Range (Fig. 1.1B) and has operated continuously since 1988. The plant currently produces around 66 MWe. Production reservoir fluid temperatures are between 225 and 245oC (Benoit, 1992; Bergfeld et al., 2001), though temperatures as high as 285oC are found in wells drilled 5 km to the south (Blackwell, 2000). Production depths are between 2,400 and 2700 m (Kennedy and van Soest, 2006). The power plant and wells are located southeast of the surface expression of the Stillwater fault. Active and fossil hot springs and fumaroles lie along the fault zone from a few km north of the plant (Senator fumarole group) to an area 20 km to the southwest of the power plant (Dixie hot springs) (Bergfeld et al., 2001). In addition, hot fluids discharge into shallow aquifers in the region. Previous studies have 4 Fig. 1.1. Location maps of the Dixie Valley geothermal field. (A) Map showing location of Dixie Valley geothermal field, surrounding mountain ranges, and valleys (modified from Bergfeld et al. 2001). (B) Map showing location of the wells and fumaroles in the vicinity of the Dixie Valley geothermal power plant (modified from Goff et al. 2002). 5 (8) Plant I .. I ~ • o Mountain * Dixie Va lley Geothermal t N * e • Production Well o Observation Wei • Injection Well o Separator Production Line Injection Line Fault measured the gas compositions discharged by wells, fumaroles, springs, and soils in the geothermal field and surrounding areas (e.g. Bergfeld et al., 2001; Goff et al., 2002; Kennedy and van Soest, 2006). The gas composition of the fumaroles indicates a mixture of air or gases from air saturated water and gases from a deeper geothermal source (Bergfeld et al., 2001). Their 13C values suggest that the reservoir CO2 is produced by thermal decarbonation of hydrothermal calcite veins (Bergfeld et al., 2001). Helium isotope data imply mixing between a deeply sourced geothermal reservoir fluid and shallower meteoric waters. Ra values of 0.7-0.76 have been interpreted to indicate the presence of a mantle derived He component (Kennedy and van Soest, 2006). History of Well 28-33 Drilling of well 28-33 was initiated in May, 1990 and completed in July, 1990. The well was drilled to a total measured depth (TD) of 2898 m. The main productive fractures were penetrated between 2784 and 2788 m depth. Temperature surveys for well 28-33 taken while the well was flowing are shown in Fig. 1.2. and indicate that initiation of boiling occurs between 1000 and 1200 m depth. During a 1993 clean out of the well, scale minerals on the well bore walls were found to be approximately 1 cm thick above 950 m depth. Below 950 m, no mineral scale was observed. Tubing was inserted into 28- 33 after this clean out to deliver scale inhibitor into the well. Nalco 1340 HP was used as a scale inhibitor until 1998 when it was replaced by Nalco 9354. In 2001, after an acid cleanout of several production wells, Nalco 1340 HP was again used as the scale inhibitor in the Dixie Valley production wells. The tubing from 28-33 was removed during the 6 Fig. 1.2. Flowing temperature log for well 28-33 from 1993 to 1998. The break in slope between 1000 and 1200 m depth indicates the initiation of boiling. The temperature of the boiling point is depressed compared to that of pure water due to the presence of dissolved gases, such as CO2. 7 Table 1.1 Well 28-33 Fluid Chemistry ____________________________________________________________________________________________________________ Well 28-33 fluid chemistry, values in ppm. ____________________________________________________________________________________________________________ Date Lab pH Na K Ca Mg Fe Al SiO2 B Li HCO3 CO3 Cl F SO4 09/23/93 a 7.58 228 6.13 15.6 2.08 101 2.25 0.35 140 0 70.1 4.28 273 12/12/95 b 433 70.9 9.19 1.55 629 10.5 132 35.0 463 16.4 214 10/07/96 b 420 71.1 8.36 1.48 623 10.5 126 43.0 453 15.0 214 10/30/97 a 9.13 429 70.1 7.40 0.02 642 9.47 2.24 115 52.8 470 15.4 199 11/05/97 b 441 71.7 8.17 0.029 609 10.2 148 30.3 463 16.4 200 04/28/98 a 9.01 447 67.8 7.50 <0.01 550 9.38 2.28 73.2 72.0 446 16.6 199 10/02/98 b 417 63.9 8.54 0.020 534 10.7 1.75 140 31.0 454 15.0 211 10/21/98 a 9.38 412 65.5 7.21 0.03 531 9.73 2.03 75.6 76.8 441 15.6 199 05/05/99 a 9.10 432 66.2 6.68 0.02 561 9.60 2.24 85.0 74.0 483 16.3 213 01/24/00 b 454 74.9 8.80 0.100 542 11.7 2.58 153 28.0 473 15.2 221 02/13/01 b 448 58.7 6.47 587 9.86 2.11 199 23.9 463 15.8 238 07/27/01 b 441 65.0 7.13 645 10.9 1.59 188 31.2 438 16.9 242 01/31/02 b 459 62.4 7.21 0.055 548 10.0 1.98 167 21.5 443 16.0 234 04/27/04 b 415 56.2 5.32 538 8.97 2.07 157 39.2 427 18.4 182 ________________________________________________________________________________________________________________________________ a. Reported by Goff et al. (2002) b. Reported by Terra-Gen 8 Table 1.2 Well 28-33 Gas Chemistry ____________________________________________________________________________________________________________ Well 28-33 gas chemistry as reported by Goff et al. (2002) reported as mol% dry gas (except for H2O). ____________________________________________________________________________________________________________ Date Steam Fraction H2O CO2 H2S H2 CH4 C2H6 N2 Ar He Comment (y) (mol% wet) 10/30/97 0.156 99.5 44.8 0.981 0.012 0.192 43.1 0.558 0.000 Air contamination 10/21/98 99.8 96.0 0.966 0.064 0.406 0.006 1.77 0.030 <0.0002 05/05/99 0.159 99.6 97.8 0.596 0.027 0.253 0.004 0.909 0.023 0.001 _____________________________________________________________________________________________________________________ 9 clean out process in 2001 (Fig. 1.3A). Fluid and gas chemistry analyses for well 28-33 are provided in Tables 1.1 and 1.2, respectively. Methods Scale Mineral Analysis Concentric bands (Fig. 1.3B) and aggregates (Fig. 1.3C) of scale minerals deposited on the tubing pulled from well 28-33 were sampled at 44 depth intervals between 650 and 1200 m depth. The mineralogy and mineral abundances in the scale were estimated for 36 sample intervals using X-ray diffractometry (XRD). X-ray Diffractometry Whole rock and clay XRD were performed on each sample in the XRD laboratory at the Energy and Geoscience Institute at the University of Utah, using a Bruker D8 Advance X-ray diffactometer. Phase quantification using the Rietveld method (Rietveld, 1969) was performed using TOPAS software developed by Bruker Analytical X-ray Systems. The following operating parameters were used during the analyses: Cu-K- radiation at 40 kV and 40 mA, 0.02o2 step size, 0.4 and 0.6 seconds per step for clay and bulk samples respectively. Clay samples (<5 m) were examined from 2 to 45o2 before and after treatment with ethylene glycol and the bulk sample from 4 to 65o2 . The clay mineral abundances were determined from the Rietveld refinement of the bulk scans. 10 Fig. 1.3. Scale minerals deposited on tubing from well 28-33. (A) Tubing pulled from 28- 33 with scale. (B) Fine grained scale from 844 m depth. (C) Coarsely crystalline scale from 1183 m depth. 11 Scanning Electron Microscopy Scanning electron microscope (SEM) analyses were used to characterize the scale mineral textures. The samples were examined on a LEO tungsten filament electron source SEM at the Dixon Laser Institute at the University of Utah. An accelerating voltage of 20 kV was used. Fluid Inclusion Gas Analysis Liquid and vapor rich fluid inclusions are common in the carbonate scale samples (Figs. 1.4A and 1.4B). Although repeated efforts were made to prepare polished chips of the scale for microthermometric measurements, the small size of the fluid inclusions (generally 1-<3 m) and the poor optical quality of the chips precluded measurement of their homogenization and ice melting temperatures. Samples from 24 depth intervals between 700 to 1200 m depth in well 28-33 were selected for fluid inclusion gas analysis at the New Mexico Institute of Mining and Technology. The gases in the fluid inclusions represent a sample of the liquid and vapor phases in the boiling column trapped during a period of five years. Prior to the analysis, the samples are cleansed with NaOH or KOH to remove surface organics. Major and minor gases, including H2O, CO2, CH4, H2S, H2, N2, Ar, He, and C2-7 organic species contained in inclusions were analyzed with a Balzers QMS 420 quadrupole mass spectrometer after being released by crushing under vacuum. Norman and Sawkins (1987) and Norman et al. (1996) present details of this analytical technique. The crush-fast- scan (CFS) method used involves opening inclusions with a swift crush in a vacuum chamber housing the mass spectrometer. The volatiles are removed from the chamber by 12 Fig. 1.4. Photomicrographs of fluid inclusions found in scale. (A) Unusually large (~5 mm) liquid rich fluid inclusions found in crystal from 1095 m depth. Arrow points to a liquid rich inclusion. (B) Fluid inclusions found in large crystal from 877 m depth. Arrow points to a group of small (<1 mm) fluid inclusions. 13 the vacuum pumping system in 1 to 2 s. Measurements are taken every 150 to 225 milliseconds by operating the quadrupole in a fast scan mode. Whole samples within each depth interval, not individual layers, were crushed in the CFS analysis. Each crush yields an analysis that could represent various proportions of liquid and vapor rich inclusions (Norman et al., 1997). Gas concentrations greater than 1 mol% suggest a fluid inclusion population dominated by vapor rich inclusions. Opening a 10 to 20 m inclusion or group of smaller inclusions of equivalent volume provides the ideal amount of volatiles for the analysis. The volatile content of a 40 m inclusion will overload the vacuum system, precluding analysis of the sample. Due to this limitation on the volume of gaseous species that can be collected and analyzed, inclusions that trapped gas rich steam will typically be underrepresented (Moore et al, 2000). Five to twenty crushes can be made on a 0.2 g sample with the expectation that some of the analyses will be failures. The precision of the CFS analyses was estimated from repeatability of gas ratio measurements. It is dependent on the size of the volatile burst, which ranges from 10 to 105 counts. Bursts with counts of less than 103 are rejected due to poor precision. Volatile bursts of about 105 counts yield a precision of 20 percent for gaseous specie/water ratios and 10 percent for gas/gas ratios. Because geothermal waters are low in O2, it is routinely measured to determine if air contamination has occurred. No evidence of air contamination was detected in these analyses. Ammonia is rarely detected due to interferences from secondary H2O peaks at a mass to charge ratio (m/e) = 17 and 16. Helium concentrations below 30 ppm may be masked by the tail on the H2 peak. Organic compounds, principally propane (C3H8) and propene (C3H6), may interfere with the detection of Ar. It is assumed that the peak 14 measured at m/e = 41 or 39 is solely from propene because it interferes more strongly with the Ar peak at m/e = 40 than propane. The amount of Ar is calculated by first subtracting the calculated contribution to the 40 m/e peak from propene. The maximum Ar value is taken as 20 percent of the propene (C3H6) peak height and is used to constrain the minimum N2/Ar ratios in cases where there is such strong interference that data reduction indicates no detectable Ar. Results Mineral Scale Mineral abundances for the 36 depth intervals, as determined by XRD, are plotted in Fig. 1.5. The mineral scales consist mainly of calcite and aragonite with minor Mg rich smectite and quartz. Aluminum bearing amorphous silica was found in the scale near the top of the well (Fig.1.6). Scale found deeper in the well consisted of individual bands of calcite, aragonite, or Mg smectite (Fig. 1.6B). Quartz is found in trace amounts between 844 and 888 m. Between 1112 and 1125 m depth (Fig. 1.5), there is a change from calcite as the dominant carbonate mineral in the scale at shallow depths to aragonite. While trace amounts of aragonite occur in the calcite dominated interval, there is very little calcite in the scale below 1125 m. The aragonite crystals form layers of coarse grained crystals that are loosely bound and easily separated from each other. Aragonite crystals appear to have both irregular and curved surfaces. Hopper growth (Fig. 1.6C), twinning (Fig. 1.6D), and fibrous crystals (Fig. 1.6E) can be observed. In contrast, the calcite layers are denser and 15 Fig. 1.5. Mineral abundance of scale deposits from XRD analysis. Red dashed line indicates the change from aragonite to calcite as the dominant form of CaCO3. 16 650 700 750 800 850 --E 900 {950 ~ 1000 1050 1100 1150 1200 - - - - - - - - - - - -- --------------- - - - I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I o 40 80 0 40 80 0 4 8 Calcite Aragonite Smectite Mineral Abundance (0/0) I I I I I I I I I I I 048 Quartz Fig. 1.6. SEM images and photomicrographs of scale minerals. (A) Aluminum rich amorphous silica scale found near the top of the well. (B) Mg smectite. (C) Aragonite crystals showing hopper growth. (D) Aragonite crystal showing twinning (twinning plane illustrated in red). (E) Both large, coarse and fibrous aragonite crystals. (F) Layers of fine, microcrystalline calcite. 17 more difficult to separate from each other. The calcite is finely crystalline to cryptocrystalline (Fig. 1.6F). Fluid Inclusion Gases Fluid inclusion gas composition data are reported in Table 1.3. The fluid inclusions are H2O rich with less than 4% total gas (Fig. 1.7). Fluid inclusions in aragonite (depths of 1134 to 1183 m) have generally lower mol% gas than those analyzed in calcite (depths of 718 to 1100 m). Many of the analyses from calcite in this depth interval have gas concentrations greater than the solubility of CO2 in water (<1 mol% wet gas) at the measured temperatures in this interval. Data for CO2-CH4-H2 are shown in Fig. 1.8A, and for N2-Ar-He in Fig. 1.9A; also plotted in both Figs. 1.8A and 1.9A are the wellhead compositions from 28-33. Inclusions display an appreciable range of CO2/CH4 and CO2/H2 values that are generally lower than those measured at the well head (Fig. 1.8A). Gas analyses from shallow calcite samples are generally enriched in CO2 relative to CH4 and H2 compared to analyses from the deeper aragonite samples. Most aragonite samples have higher He relative to N2 and Ar compared to the calcite samples. Furthermore, aragonite samples define a linear trend of He enrichment at constant N2/Ar away from the wellhead composition (Goff et al., 2002). Many analyses of the calcite samples are closer in composition to the wellhead gas composition (Fig. 1.9A). However, although many of the calcite samples appear to have lower Ar contents than the present-day well composition, the Ar in these samples may be anomalously low due to peak interference from propane and propene. Ar could not be detected in several samples due to this interference. For comparison, gas analyses of Dixie Valley wells and fumaroles 18 Table 1.3 Fluid inclusion gas chemistry ____________________________________________________________________________________________________________ Fluid inclusion gas chemistry reported as mol % wet gas unless otherwise listed. ________________________________________________________________________________________________________________________________ Depth Burst Size H2O Total Gas H2 CH4 CO2 He N2 Ar H2S C2H6 C1/C2 Geothermometer (m) (counts) (oC) 718 1223 96.9 3.12 0.046 0.155 1.20 4.00E-5 1.65 0.023 0.001 0.007 175 941 10390 99.3 0.72 0.003 0.008 0.526 7.60E-6 0.165 0.003 0.002 0.007 101 941 20760 99.4 0.55 0.003 0.009 0.351 6.87E-6 0.163 0.003 0.163 0.009 99 966 1780 99.2 0.82 0.010 0.012 0.582 0.000 0.191 0.002 0.002 0.004 124 966 1867 99.0 1.03 0.008 0.025 0.695 0.000 0.263 0.004 0.002 0.005 139 1003 3349 99.0 1.00 0.008 0.016 0.789 1.05E-5 0.165 0.002 0.002 0.004 132 1003 9813 98.4 1.57 0.015 0.025 1.21 1.51E-5 0.267 0.002 0.004 0.006 134 1024 7697 96.3 3.69 0.086 0.016 2.97 6.54E-5 0.524 0.007 0.008 0.012 104 1044 1489 99.3 0.74 0.010 0.005 0.628 1.55E-5 0.064 5.00E-4 0.004 0.009 81 1044 2123 99.0 0.98 0.010 0.010 0.856 1.10E-5 0.070 4.00E-4 0.004 0.008 102 1044 7299 98.6 1.41 0.049 0.012 1.17 4.11E-5 0.125 4.00E-4 0.006 0.007 111 1044 2232 99.0 0.95 0.011 0.016 0.779 2.55E-5 0.094 3.00E-4 0.004 0.007 115 1064 1072 98.0 2.01 0.009 0.012 1.76 3.14E-5 0.178 0.002 0.005 0.007 109 1064 2351 97.7 2.30 0.012 0.020 1.94 3.41E-5 0.268 0.004 0.005 0.011 112 1064 6748 98.0 2.02 0.051 0.015 1.65 7.63E-5 0.253 0.003 0.004 0.014 108 1064 2714 98.5 1.55 0.013 0.022 1.28 3.36E-5 0.198 0.003 0.003 0.011 115 1100 13160 97.4 2.63 0.076 0.010 2.11 7.30E-5 0.339 0.003 0.010 0.012 92 1100 23370 97.4 2.57 0.087 0.014 2.07 7.48E-5 0.299 0.002 0.009 0.008 110 1100 4854 98.5 1.52 0.014 0.015 1.24 1.58E-5 0.178 0.001 0.003 0.004 129 1134 5323 99.7 0.33 0.005 0.035 0.233 0.000 0.017 nd 0.003 0.006 142 1134 7498 99.7 0.33 0.012 0.061 0.181 3.90E-6 0.023 nd 0.004 0.008 147 1134 4691 99.7 0.28 0.014 0.038 0.168 6.37E-6 0.018 nd 0.003 0.006 143 1154 82650 99.6 0.35 0.047 0.003 0.196 1.27E-4 0.058 0.001 0.011 0.019 51 19 Table 1.3 continued ____________________________________________________________________________________________________________ Fluid inclusion gas chemistry reported as mol % wet gas unless otherwise listed. ________________________________________________________________________________________________________________________________ Depth Burst Size H2O Total Gas H2 CH4 CO2 He N2 Ar H2S C2H6 C1/C2 Geothermometer (m) (counts) (oC) 1154 91140 99.7 0.26 0.030 0.002 0.146 1.03E-4 0.043 6.00E-4 0.010 0.013 53 1154 107300 99.8 0.19 0.025 0.002 0.102 2.00E-4 0.024 4.00E-4 0.009 0.009 64 1165 3124 99.8 0.22 0.003 0.015 0.147 0.000 0.036 5.00E-4 0.001 0.008 111 1165 48230 99.8 0.19 0.003 0.002 0.136 2.82E-5 0.016 1.00E-4 0.005 0.012 53 1174 93510 99.8 0.15 0.023 0.004 0.092 7.19E-4 0.009 2.00E-4 0.007 0.008 78 1174 48950 99.8 0.22 0.012 0.007 0.162 4.62E-5 0.012 1.00E-4 0.005 0.012 84 1174 81840 99.8 0.18 0.023 0.005 0.111 4.12E-4 0.013 2.00E-4 0.006 0.009 81 1183 85960 99.8 0.19 0.028 0.004 0.111 7.41E-4 0.013 3.00E-4 0.007 0.011 74 1183 6923 99.9 0.13 0.001 0.001 0.095 6.88E-6 0.022 3.00E-4 0.001 0.003 71 1154 36150 99.3 0.73 0.094 0.005 0.421 9.21E-5 0.140 0.002 0.012 0.030 50 1154 22940 99.6 0.45 0.063 0.003 0.252 5.53E-5 0.094 0.001 0.006 0.014 55 1154 55800 99.5 0.47 0.066 0.003 0.272 5.73E-5 0.075 0.001 0.011 0.025 42 ________________________________________________________________________________________________________________________________ nd=not detected 20 Fig. 1.7. Total dry gas from fluid inclusion analyses. 21 Total Gas (mol%) o 1 2 3 4 700 ~~~~---r--~ 750 800 850 -E 900 .-..!...:. • Q.. 950 OJ 0 1000 • • 1050 • • • 1100 • • • 1150 1 200 1....-_.1...-_.1...----'"-------1 Fig. 1.8. CO2-CH4-H2 plots. (A) Relative CO2-CH4-H2 composition of well head gas (n=3) and fluid inclusion gas analyses from calcite (n=19) and aragonite (n=16) scale samples. Aragonite analyses show a broad range in CO2/CH4 and CO2/H2 values, while calcite analyses (with one exception) show a composition overwhelmingly dominated by CO2. Well head gas analyses (Goff et al., 2002) also show high CO2 relative to CH4 and H2. 22 Fig. 1.8. continued. (B) Relative CO2-CH4-H2 composition of Dixie Valley area wells and fumaroles. Gas analyses show dominance of CO2 relative to CH4 and H2. 23 Fig. 1.8. continued. (C) Relative CO2-CH4-H2 compositional changes associated with closed system and open system boiling. The red symbols represent the initial composition of the fluid (43% H2, 15% CH4, 42% CO2 for a closed system and 10% H2, 50% CH4, and 40% CO2 for an open system). The numbers on the plot are the steam fraction (y) associated with each composition plotted. Upon commencement of boiling (y=0.003) in an open or closed system, the vapor (open symbols) becomes enriched in CH4 and H2, while the liquid becomes depleted in these gases, moving the composition of the residual liquid to higher CO2 relative to CH4 and H2 (for example, note compositions of residual liquid at 0.003 and 0.008 steam fractions). With progressive boiling in an open or closed system, both liquid and vapor compositions move to higher CO2 relative to CH4 and H2 compared to their compositions at y=0.003. The measured y value, determined from steam and liquid at the well head of 28-33, was 0.156 in 1997 and 0.159 in 2009. 24 Fig. 1.9. N2-Ar-He plots. (A) Relative N2-Ar-He composition of well head gas (n=3) and fluid inclusion gas analyses from calcite (n=19) and aragonite (n=13) scale samples. Aragonite analyses show a broad range of He contents and relatively consistent N2/Ar ratio. Calcite samples all have low He relative to N2 and Ar. Well head gas analyses plot within this range of He relative to N2 and Ar. Some calcite analyses have high N2/Ar ratios. These anomalous ratios reflect low Ar values due to interference of propane and propene peaks in detecting Ar. The compositions of air and air saturated water (ASW) are plotted for reference. 25 Fig. 1.9. continued. (B) Relative N2-Ar-He composition of Dixie Valley area wells and fumaroles. The geothermal wells follow a trend of varying He content at relatively consistent N2/Ar ratio. Samples from nongeothermal local wells, springs, and fumaroles have lower He relative to N2 and Ar than most of the geothermal well samples in the area. The compositions of air and air saturated water (ASW) are plotted for reference. 26 Fig. 1.9. continued. (C) Relative N2-Ar-He compositional changes associated with closed system and open system boiling. The red symbols represent the initial composition of the fluid. The numbers on the plot are the steam fraction (y) associated with each composition plotted. Upon commencement of boiling (y = 0.003) in an open system, the vapor becomes slightly enriched in N2 relative to Ar, while the liquid becomes depleted in N2, moving the composition of the residual liquid to a lower N2/Ar ratio. With progressive boiling in an open system, N2 becomes depleted relative to Ar in both the liquid and vapor, and Ar also begins to be depleted relative to He in the liquid (see y=0.008 and y=0.014). Upon initiation of boiling in a closed system (y=0.003), the liquid becomes slightly enriched in N2 relative to Ar and He, while the vapor becomes slightly depleted in N2 relative to Ar and He. With progressive boiling in a closed system, the compositions of both liquid and vapor move to higher N2 than their compositions at y=0.003. The measured y value of the steam and fluid discharged at the wellhead was 0.156 in 1997 and 0.159 in 2009. 27 Fig. 1.9. continued. (D) N2-Ar-He compositional fields for magmatic, crustal, deeply circulated meteoric, and shallow meteoric fluid sources (modified from Norman and Musgrave 1994). 28 from Goff et al. (2002) are presented on the CO2-CH4-H2 and N2-Ar-He plots in Figs. 1.8B and 1.9B, respectively. The compositions of 28-33, other Dixie Valley wells, and fumarole gases are all overwhelmingly rich in CO2 and have similar CO2-CH4-H2 compositions to most of the gas analyses from calcite samples. In contrast, most of the aragonite samples are variably enriched in CH4 and/or H2. The gas analyses from the Dixie Valley wells and fumaroles also define a linear trend of He enrichment similar to that defined by the fluid inclusion gas analyses from the aragonite scales (Figs. 1.9B and 1.9A). Discussion Mineralogy The mineral composition, habit, and textural relationships provide a record of the physical and chemical conditions in the well during scale formation. Polyacrylate scale inhibitors, such as Nalco 9354, create chemical conditions that favor the spontaneous nucleation and formation of very small calcite crystals in the fluid (Siega et al., 2005). This could explain the appearance of the calcite crystals found in the mineral scale, but not the presence of large aragonite crystals. The form of the aragonite crystals is interpreted to reflect the degree of saturation of the fluid at the time of their formation. Gonzales et al. (1992) and Rimstidt (1997) suggest that minerals formed under near equilibrium conditions will show smooth faces on well-developed crystals. With higher degrees of supersaturation, these crystal faces become roughened due to surface nucleation (Liu et al., 1992). Hopper growth, penetration twinning, and radial fiber growth result from the growth rate becoming 29 diffusion limited. At the highest degrees of supersaturation, colloidal particles form in solution then precipitate as aggregates (in low flow rate systems) or as banded deposits (in high flow rate systems) (Saunders, 1990). If these principles are applied to this geothermal well environment, the observed textures indicate that aragonite was deposited with a moderate degree of supersaturation in the fluid, whereas the calcite was deposited where progressive boiling significantly increased supersaturation of CaCO3. Differences in the morphology of the calcite (cryptocrystalline calcite or coarser grained calcite aggregates) may also reflect variations in the degree of supersaturation of CaCO3. These variations could, in part, be related to changes in well bore conditions dictated by field operations. Operational changes include putting the well on idle, bringing the well back on line, varying flow rates, and periods of unrestricted flow to atmosphere. Fluid Inclusion Gases Some of the fluid inclusion gases have higher CH4 or H2 (relative to CO2) and/or higher N2 (relative to Ar) than gas samples from the well head, other wells, and fumaroles. The formation of CH4 rich (relative to CO2) inclusions may reflect: 1) the mixing of a higher CO2/CH4 fluid with a lower CO2/CH4 fluid and the subsequent trapping of various mixtures of these fluids over time; 2) Fischer-Tropsch reactions (see below) occurring within the inclusion after formation; 3) bacterial production of CH4 within the inclusion after formation; 4) the consumption of free CO2 by carbonate mineral formation; and/or 5) the trapping of early steam in the fluid inclusions (vapor will have a CO2/CH4 ratio about 1/4 of that in the liquid at temperatures near 250oC (Moore et al., 2000)). The formation of CH4 through Fischer-Tropsch reactions in a natural system 30 is controversial (McCollum and Seewald, 2001). Abiogenic formation of CH4 is postulated at midocean ridges (MOR) and in other high temperature environments where a chemical catalyst may be involved (Fiebig et al., 2004). CO2+4H2=CH4+2H2O (1.1) Comparisons of the C1-6 gases can provide insight into the origins of the CH4. C1 and C2 gases can be used as a geothermometer in geothermal fields. Their relationship to temperature is given by: ToC=57.8log(C1/C2)+96.8 (1.2) This relationship was developed by Darling (1998) based on data from various hydrothermal systems. This relationship assumes that the thermal decomposition of higher order hydrocarbon chains will be temperature dependent. Excess CH4 for a given temperature implies the addition of CH4 from another source, possibly abiogenic production. The calculated temperatures for the Dixie Valley samples using this geothermometer are given in Table 1.3 and range from 42 to 175oC. All but one of the calculated temperatures is lower than the measured production temperatures. This indicates a deficiency of CH4 with respect to C2H6, possibly reflecting the influence of abundant sedimentary hosted organics on these gas compositions (Giggenbach and Corrales, 1992; Darling, 1998). Thus, these geothermometer calculations do not support the formation of CH4 by abiogenic processes within the inclusion after formation. 31 N2-rich (relative to Ar) inclusions can be the result of: 1) decomposition of organic matter providing an additional source of N2 (Norman and Musgrave, 1994); 2) breakdown of ammonia; 2NH3=2N2+3H2 (1.3) or 3) mixing of fluids with different N2/Ar and the subsequent trapping of various mixtures of these fluids. The breakdown of ammonia would increase N2 and H2 in the fluid inclusion gases. However, the reservoir and well temperatures are too low to be conducive to the thermal breakdown of ammonia (Norman and Musgrave, 1994), and there does not seem to be a correlation between increased N2 and increased H2 in gas analyses (Table 1.3). During boiling, CH4, H2, N2, and other light gases preferentially partition into the vapor phase (Henley et al., 1984). Owing to greater solubility, some gases, particularly CO2 and H2S, are partially retained in the residual liquid phase. Calculations of the gas content in either the vapor or liquid phase can be made using gas distribution coefficients following the procedures outlined by Giggenbach, (1980) and Henley et al., (1984) as detailed in Appendix A. Vapor-liquid distribution constants (referred to as Kd or B) reported by Fernandez-Prini et al., (2003) were used to calculate the composition of the liquid and vapor phases at different stages of boiling. Both closed and open system boiling end member processes were considered. Example boiling trends are illustrated in Figs. 1.8C and 1.9C. These calculations indicate that boiling can explain the progression in the fluid inclusion gas compositions from high CH4 and H2 to high CO2. Simmons et 32 al. (2007) observed heterogeneous trapping of coexisting gas and liquid phases as a result of boiling in fluid inclusions from the Broadlands-Ohaaki geothermal system in New Zealand. The measured variations in CH4 and H2 to CO2 in the fluid inclusions analyzed in this study probably reflect heterogeneous trapping of both liquid and vapor phases, as well; therefore, gas compositions of both liquid and vapor phases during progressive boiling were calculated and plotted in Figs. 1.8C and 1.9C. Fluid inclusion analyses with gas contents >1 mol% (Fig. 1.7) also suggest that some of these analyses are sampling predominantly vapor rich fluid inclusions. However, boiling of a single source fluid cannot produce both the high CH4 and high H2 fluid inclusion gas compositions shown in Fig. 1.8A. Boiling also cannot explain the trend of some samples toward higher N2 (relative to Ar) in shallower samples (Fig. 1.9A), since lower N2 would be expected with progressive boiling; nor can boiling explain the trend of the variable He contents at near constant N2/Ar ratio. The pattern of the fluid inclusion gas data indicates that boiling has influenced the variation in the gas compositions, but the data and compositional trends cannot be fully explained by this process. Alternatively, variation in gas compositions suggests heterogeneity in the composition of the production fluids over time. Various studies have concluded that there is mixing between a deeper circulating geothermal reservoir fluid with a shallower meteoric fluid in the Dixie Valley geothermal field (Bergfeld et al., 2001; Kennedy and van Soest, 2006). Ra values of 0.7-0.76 have been interpreted to indicate the presence of a mantle derived He component (Kennedy and van Soest, 2006). The linear trend of increasing He at near constant N2/Ar in both fluid inclusion gas analyses (Fig. 1.9A) and 33 surface gas analyses (Fig. 1.9B) is consistent with variable mixing between deeply circulating meteoric water (reservoir fluid) and shallow meteoric water (Fig. 1.9D) or variation in the input of He from a mantle derived source (Giggenbach, 1986). This trend is also observed in the data from the Dixie Valley well and fumarole gas samples (Fig. 1.9B) and in the well gas sampled from 28-33 (open circles, Fig. 1.9A). The fluid inclusions from depths of 718 and 1100 m (calcite) plot closer to the air or air saturated water end member on the mixing trend. This implies dominance of a shallow source fluid on the composition of gases trapped within this depth interval. Conclusion At the Dixie Valley geothermal field, fluid boiling triggered the precipitation of calcite and aragonite scale in concentric bands around tubing inserted into production well 28-33. The trends and ranges in fluid inclusion gas compositions from these mineral scales appear to result from the combined effects of fluid mixing and boiling. The fluid inclusion gas analyses define a linear trend of increasing He at constant N2/Ar ratios similar to air saturated surface (meteoric) water, a trend consistent with variable mixing of this surface water with deeply circulated meteoric water (deep geothermal reservoir fluid) or variation in the input of He from a mantle derived source. However, the variations in CO2/CH4 and CO2/H2 in these inclusion gases and the entrapment of gas enriched in CH4 and/or H2 relative to the very CO2 rich gas composition discharged at the wellhead demonstrate the impacts of progressive boiling. Because CO2 is assumed to be the dominant gas present in geothermal systems based on well gas sampling (Hedenquist and Henley, 1985), the trapping of CH4 and H2 in fluid inclusions as a result of initial 34 boiling has implications for the calculation and interpretation of salinity and redox conditions in both modern and fossil geothermal systems. Acknowledgments Funding for this project was provided by the US Department of Energy's Geothermal Program under grant DE-FG36-04GO14292. The author would like to thank Terra-Gen Power, LLC for allowing the collection of the samples and providing data on downhole conditions. The author would also like to thank Jungho Park and David Norman for the fluid inclusion gas analyses, David Langton for his work in collecting the mineral scale samples, Louise Spann for processing and analyzing the samples using X-ray diffractometry. The author would also like to thank A.E. Williams-Jones and Stuart Simmons for their thoughtful review and comments. References Benoit, D., 1992. A case history of injection through 1991 at Dixie Valley, Nevada. Geothermal Resources Council Transactions 16, 611-620. Bergfeld, D., Goff, F.E., Janik, C.J., 2001. Elevated carbon dioxide flux at the Dixie Valley geothermal field, Nevada: relations between surface phenomena and the geothermal reservoir. Chemical Geology 177, 43-66. Blackwell, D.D., Golan, B., Benoit, D., 2000. Thermal regime in the Dixie Valley geothermal system. Proceedings World Geothermal Congress 2000, 991-997. Blackwell, D.D., Leidig, M., Smith, R., Johnson, S., Wisian, K.W., 2002 Exploration and development techniques for Basin and Range geothermal systems: examples from Dixie Valley, Nevada. Geothermal Resources Council Transactions 26, 513-518. Darling, W.G., 1998. Hydrothermal hydrocarbon gases: 1, Genesis and geothermometry. Applied Geochemistry 13, 815-824. 35 Fernandez-Prini, R., Alvarez, J., Harvey, A., 2003. Henry's constants and vapor-liquid distribution constants for gaseous solutes in H2O and D2O at high temperatures. Journal of Physical and Chemical Reference Data 32, 903-916. Fiebig, J., Chiodini, G., Caliro, S., Rizzo, A., Spangenberg, J., Hunziker, J.C., 2004. Chemical and isotopic equilibrium between CO2 and CH4 in fumarolic gas discharges: Generation of CH4 in arc magmatic-hydrothermal systems. Geochimica et Cosmochimica Acta 68, 2321-2334. Giggenbach, W.F., 1980. Geothermal gas equilibria. Geochimica et Cosmochimica Acta 44, 2021-2032. Giggenbach, W.F., 1986. The use of gas chemistry in delineating origin of fluids discharges over the Taupo Volcanic Zone: a review. International Volcanological Congress 5, 47-50. Giggenbach, W.F., Corrales, R., 1992. The isotopic and chemical composition of water and steam discharges from volcanic-magmatic-hydrothermal systems of Guanacaste Geothermal Province, Costa Rica. Applied Geochemistry 7, 309-332. Goff, F., Bergfeld, D., Janik, C.J., Counce, D., Murrell, M., 2002. Geochemical data on waters, gases, scales, and rocks from the Dixie Valley region, Nevada (1996-1999). Los Alamos National Laboratory publication LA-13972, 1-80. Gonzales, L., Carpenter, S.J., Lohmann, K.C., 1992. Inorganic calcite morphology: Roles of fluid chemistry and fluid flow: Journal of Sedimentary Petrology 62, 382-399 Graney, J.R., Kesler, S.E., 1995. Gas composition of inclusion fluid in ore deposits: Is there a relation to magmas? In: Magmas, Fluids and Ore Deposits (ed. JFH Thompson) Mineralogical Association of Canada Short Course Series 23, 221-245. Hedenquist, J.W., Henley, R.W., 1985. The importance of CO2 on freezing point measurements of fluid inclusions; evidence from active geothermal systems and implications for epithermal ore deposition. Economic Geology 80, 1379-1406. Henley, R.W., Truesdell, A.H., Barton, P.B., 1984. Fluid-mineral equilibria in hydrothermal systems. Reviews in economic geology, Society of Economic Geologists, El Paso, TX. Kennedy, B.M., van Soest, M.C., 2006. A helium isotope perspective in the Dixie Valley, Nevada hydrothermal system. Geothermics 35, 26-43. Liu, X.Y., Bennema, P., van der Eerden, J.P., 1992. The rough-flat-rough transition at crystal surfaces. Nature 356, 778-780. 36 Lutz, S.J., Moore, J.N., Benoit, D., 1997. Geologic framework of Jurassic reservoir rocks in the Dixie Valley geothermal field, Nevada: implications from hydrothermal alteration and stratigraphy. Proceedings 22nd Workshop on Geothermal Reservoir Engineering 1997, 131-139. McCollum, T.M., Seewald, J.S., 2001. A reassessment of the potential for reduction of dissolved CO2 to hydrocarbons during serpentinization of olivine. Geochimica et Cosmochimica Acta 65, 3769-3778. Moore, J.N., Powell, T.S., Heizler, M.T., Norman, D.I., 2000. Mineralization and hydrothermal history of the Tiwi geothermal system, Philippines. Economic Geology 95, 1001-1023. Moore, J.N., Norman, D.I., Kennedy, B.M., 2001. Fluid inclusion gas compositions from an active magmatic-hydrothermal system: a case study of The Geysers geothermal field, USA. Chemical Geology 173, 3-30. Norman, D.I., Sawkins, F.J., 1987. Analysis of fluid inclusions by mass spectrometer. Chemical Geology 61, 1-10. Norman, D.I., Musgrave, J.A., 1994. N2-Ar-He compositions in fluid inclusions: Indicators of fluid source. Geochimica et Cosmochimica Acta 58, 1119-1131. Norman, D.I., Moore, J.N., Yonaka, B., Musgrave, J., 1996. Gaseous species in fluid inclusions: a tracer of fluids and indicator of fluid processes. Proceedings 21st Workshop on Geothermal Reservoir Engineering 1996, 233-240. Norman, D.I., Moore, J.N., Musgrave, J., 1997. More on the use of fluid-inclusion gaseous species as tracers in geothermal systems. Proceedings 22nd Workshop on Geothermal Reservoir Engineering 1997, 419-426. Norman, D.I., Moore, J.N., 1999. Methane and excess N2 and Ar in geothermal fluid inclusions. Proceedings 24th Workshop on Geothermal Reservoir Engineering 1999, 233- 240. Rietveld, H.M., 1969. A profile refinement method for nuclear and magnetic structures. Journal of Applied Crystallography 2, 65-71. Rimstidt, J.D., 1997. Gange mineral transport and deposition. In: Geochemistry of Hydrothermal Ore Deposits (ed. HL Barnes) New York, John Wiley and Sons, 487-516. Saunders, J.A., 1990. Colloidal transport of gold and silica in epithermal precious-metal systems; Evidence from Sleeper deposit, Nevada. Geology 18, 757-760. 37 Siega, F.L., Herras, E.B., Buning, B.C., 2005. Calcite scale inhibition: the case of Mahanagdong Wells in Leyte geothermal production field, Philippines. Proceedings World Geothermal Congress 2005. Simmons, S.F., Browne, P.L.R., 2000. Hydrothermal minerals and precious metals in the Broadlands-Ohaaki geothermal system: Implications for understanding low-sulfidation epithermal environments. Economic Geology 95, 971-999. Simmons, S.F., Christenson, B.W., 1994. Origins of calcite in a boiling geothermal system. American Journal of Science 294, 361-400. Simmons, S.F., Simpson, M.P., Reynolds, T.J., 2007. The significance of clathrates in fluid inclusions and the evidence for overpressuring in the Broadlands-Ohaaki geothermal system, New Zealand. Economic Geology 102, 127-135. Speed, R.C., 1976 Geologic map of the Humboldt Lopolith. Geological Society of America Map Chart Series MC-14, 1:81050 scale, 1 sheet. Weibel, A.F., 1987. An overview of the geology and secondary mineralogy of the high temperature geothermal system in Dixie Valley, Nevada. Geothermal Resources Council Bulletin Sept/Oct, 5-11. 38 CHAPTER 2 MODELING THE GEOCHEMICAL EFFECTS OF INJECTION AT COSO GEOTHERMAL FIELD, CA; COMPARISON WITH FIELD OBSERVATIONS Abstract Decreased performance of injection wells after 5 to 7 years of injection has been documented in several geothermal fields. In this study, the effects of injecting flashed geothermal fluids into the Coso geothermal field, California are investigated by comparing drill cuttings from the original injection wells with samples from wells drilled on the same pads after injectivities in the original wells had declined. At Coso the fluids injected into well 68-20 had silica contents up to 940 ppm and are significantly supersaturated in silica with respect to quartz, the stable silica phase in the reservoir. X-ray diffraction and scanning electron microscope analyses of the reservoir rock penetrated by redrilled injection well 68-20RD indicate that loss of injectivity in 68-20 was caused by the deposition of silica as opal-A accompanied by trace amounts of calcite near the well bore. As the scale deposits mature, the original 2 m spheres coalesce into larger spheres, up to 10 m in diameter and plate like sheets. Application of TOUGHREACT one dimensional (1D) reactive transport models (Xu et al., 2004) predicts the deposition of amorphous silica with minor calcite, leading to a sharp decrease in porosity 1 to 2 years after initiation of injection. These predictions are consistent with field measurements and the observations made from injection well cuttings from Coso pad 68-20. Furthermore, the modeling predicts silica deposition will occur close to the wellbore, consistent with the observation that amorphous silica is found only in the redrilled wells (68-20RD and 68B-20RD) with trajectories closest to a previously drilled injection well. Introduction The geochemical effects of injecting fluids into geothermal reservoirs are poorly understood and may be significantly underestimated. Decreased performance of injection wells has been observed in several geothermal fields after only a few years of service, although the reason for these changes has not previously been established. The purpose of this investigation is to predict the geochemical effects of the injection fluids on the reservoir rocks and to test the effects of modifying the injection fluid as a strategy for mitigating mineral precipitation. This study consists of petrologic investigations of the scale mineral assemblages in the cuttings from wells drilled on the 68-20 pad at the Coso geothermal field along with application of TOUGHREACT one dimensional reactive transport models to predict the mineral precipitation observed. The effects of changes to the injected fluid chemistry on the reservoir fluids and rocks were also modeled to further understand the consequences of particular injection strategies. 40 The Coso Geothermal Field The Coso geothermal field is developed in Mesozoic granitic rocks of the Sierra Nevada Batholith on the western edge of the Basin and Range (Adams et al., 2000) (Fig. 2.1). The heat driving the geothermal activity is related to shallow intrusions that have given rise to 38 rhyolitic domes during the last million years. The reservoir host rocks range in composition from diorite to granite with varying degrees of alteration and veining (Kovac et al., 2005). Active and fossil fumaroles lie along a northeast to southwest trending belt that extends through Devil's Kitchen and Coso Hot Springs. On the eastern margin of the field, known as the East Flank, fossil sinter and travertine deposits are present (Adams et al., 2000). Installed capacity of the power plant is 240 MWe, and power production has been sustained since 1989. Between 1987 and 1993, six injection wells were drilled on the 68-20 pad and four were drilled on the 67-17 pad in the southern part of the field (Fig. 2.1). Cuttings were examined from wells drilled on both pads, but pad 67-17 is not discussed here because no injection related mineral scale deposits were found in the cuttings of these wells. Well pad 68-20 Between 1987 and 1993, six injection wells were drilled on the 68-20 pad in the southern part of the field. The trajectories of these wells and lost circulation zones obtained from well logs are shown in Fig. 2.2. Injectivity decreased in well 68-20 from a maximum liquid injection rate of over 1000 kph (kilopounds per hour mass flow) in March, 1989 to a minimum rate of 0 kph in November, 1990 after a steady decline (Fig. 2.3). Mechanical and chemical cleanouts increased injectivity to a one time high of 800 41 Fig. 2.1. Simplified geologic map of the Coso geothermal field showing the locations of the major thermal features. The 68-20 injection pad is located in the southern part of the field. 42 • o o .5 63-18 . 53-laRD o57 - laTC•H 1km • Geothermal Well A Active Fumarole Fault (dashed where inferred) • Non-Calcareous Alteration • Calcitic Stockworks and Travertine D Quaternary Surficial Deposits(unpatternedj D Quaternary Rhyolitic Pyroclastics D Quaternary Rhyolite Domes Ter.-Quat. Volcanic / Sedimentary Rocks D Mesozoic Granitesl Metamorphics 15-17 • • 72-19 / / ...... ... • 73-19 '--~ I o / 68-20 • -- I / / / I 1 I. 1 _ -- Coso Hot I I I A~ N ! Site Map Fig. 2.2. Well trajectories for injection wells drilled in pads 67-17 and 68-20. Locations of lost circulation zones are shown as discs, and the amount of fluid lost is represented by the size of the disc. X and Y axes in UTMs, Z axis in feet. 43 Fig. 2.3. Injection history of well 68-20 from October 1988 to March 1992. (kph=kilopounds per hour) 44 LiqlnjRato_kph g .... 8 § [;g c c Oct-88 tj - -=+==-1- Noy-88 Dec-88 Jan-89 l S L L Feb-89 I ~ Mar-89 . Apr-a9 May-89 Jun-89 Jul-89 Aug-89 Sap-89 Oct-89 Noy-a9 Dec-B9 Jan-90 Feb-90 ~~~::~ I I -=-1 May-gO I ;;I; .c. Jun-90 it Jul-90 Aug-90 Sep-9o Oct-gO Nov-90 1r./5u:=.-­Dec- gO Jan-91 Feb-91 Mar-91 -' Apr-91 ~a-;;;;;~~~:J--T~-l May-91 I..: Jun-91 Jul-91 Aug-91 Sep-91 Oct-91 Nov-91 Dec-91 Jan-92 Feb-92 ... § ;., 8 Mar-92 I.' _ __ ....L __ .L ___ L __ --l ___ J. ___ ..J &l N o kph after November, 1990, but never fully recovered, remaining between 0-400 kph through 1992. Redrilled wells also experienced similar injectivity declines. Reservoir temperatures in this part of the field prior to injection ranged from approximately 205-240oC, but cooling of the reservoir was observed around these wells post injection as seen in temperature logs of subsequent redrills. The temperatures of the injected fluids ranged from 110-120oC. Cuttings from the six injection wells were sampled at 3 m intervals. The reservoir host rocks in these wells ranged from diorite to granite with trace to moderate alteration and veining. Fault breccias were observed in the cuttings, indicating major fault or fracture zones intersected by the wells. Table 2.1 shows the reported chemistry of injection fluid for well 68-20. There was a wide range in the composition of the injected fluids, showing variation in produced fluids and practices of adding steam condensate back into the injected fluid. The maximum silica content analyzed was 965 mg/L. Mineral Scale from 68-20RD Cuttings from Coso injection wells 68-20, the original injection well, and 68- 20RD, 68A-20, and 68A-20RD have been examined at 3 m depth intervals. The rock type, the abundance of primary and secondary minerals, and the abundance, mineralogy, and paragenesis of the veins were documented at each interval. Thinly banded opaline silica was observed in the cuttings from 68-20RD and 68B-20RD, but not in the original injection well 68-20 or in wells 68A-20, 68A-20RD, and 68B-20. The banding and textural relationships suggest the silica represents fracture fillings and not alteration of preexisting minerals. The greatest density of silica precipitation was found in cuttings 45 Table 2.1 Well 68-20 Injected Fluid Chemistry _______________________________________________________________________ Chemistry of injected fluid from well 68-20, showing high, low, and average concentrations in mg/kg from 15 analyses. Brines were injected at 110-120oC. _______________________________________________________________________ High Low Average Na+ 4,283 2,897 3,612 K+ 941 362 614 Ca2+ 130 19 45 Mg2+ 8.7 0 1.1 Fe 84.1 0.1 9.5 Al3+ 10.4 0 1.1 SiO2 (aq) 965 97 657 B(OH)3 141.8 83 115.4 Li+ 47 25 34 Sr2+ 8.2 2.6 4.4 Astotal 26.24 2.85 9.54 Ba2+ 116 0 9 HCO3 - 229 77 167 Cl- 6,958 5,015 6,079 F- 5.7 1.6 2.5 SO4 2- 99 27 68 TDS 12300 9233 11103 Lab pH 8.3 6.17 7.44 _______________________________________________________________________ from depths of 869-884 and 1710-1713 m in well 68-20RD. Samples of the precipitate from these two zones were analyzed using a scanning electron microscope and X-ray diffractometer. Additional petrographic images (photomicrographs and SEM images) of silica scale from 68-20RD and 68B-20RD are shown in Appendix B. SEM images and X-ray patterns of scale deposits are separated here by depth for descriptive purposes; however, morphologies and depth are not correlative. 46 869-884 m Depth The silica deposits consist of opal-A spheres and plates. Fig. 2.4 shows the morphological progression associated with maturation of deposits. Textural relationships shown in Fig. 2.4A indicate the silica was deposited initially as spheres 1-2 m in diameter. As the deposit matures, the spheres coalesce to form larger spheres up to 10 m in diameter (Fig. 2.4B). Further maturation is associated with the formation of plates and sheets. Infilling of the spaces between spheres provides a possible explanation for the dense, smooth surfaces seen in Figs. 2.4B, 2.4C, and 2.4D. This maturation sequence is similar to changes observed in young sinter deposits in New Zealand described by Rodgers et al. (2004) and Lynne and Campbell (2004). Traces of calcite locally coat the amorphous silica, suggesting it represents a later stage in the evolution of the deposits. The X-ray diffraction pattern of hand picked silica rich samples from this depth (Fig. 2.4E) indicate that the deposit consists of opal-A with a broad peak centered at 22o 2- theta. In addition, quartz peaks are present in the X-ray diffraction patterns at 21.5o and 26.8o 2-theta, but quartz was not unambiguously documented in the SEM images. It is possible that traces of quartz were deposited by the injected fluid or through interactions with the amorphous silica. Alternatively, the quartz could represent fragments of the host reservoir rock that were incorporated into the deposits of amorphous silica. 1710-1713 m Depth SEM images of silica deposits from a depth of 1710-1713 m (Fig. 2.5) show that they display similar textural and mineralogical relationships as those from depths of 869- 884 m (Fig. 2.4). Both dense and porous layers of silica are present. Textures shown in 47 1000 0 100 200 300 400 500 600 700 800 900 15.0 18.0 20.0 22.0 24.0 26.0 28.0 30.0 32.0 35.0 2-Theta Counts / Sec Fig. 2.4. Mineral scale from 68-20RD at 869-884 m depth. (A)-(D) SEM images of samples taken from 68-20RD at 869-884 m depth. Opal-A spheres 1-2 mm in diameter seen in (A) coalesce to form 10 mm spheres and sheets seen in (B), (C), and (D). (E) X-ray diffraction pattern of scale samples taken from the depth interval 869-884 m showing a broad opal-A peak centered at 22o 2-theta and quartz peaks at 21.5o and 26.8o 2-theta. Calcite 48 1000 0 100 200 300 400 500 600 700 800 900 15.0 18.0 20.0 22.0 24.0 26.0 28.0 30.0 32.0 35.0 2-Theta Counts / Sec Fig. 2.5. Mineral scale from 68-20RD at 1710 to 1713 m depth. (A)-(D) SEM images of samples taken from 68-20RD at 1710-1713 m depth. (A) Alternating silica layers with varying density and visible porosity. (B), (C) Silica spheres aligning to form fibrils (B) and sheets (C). (D) Tube structure covered with silica spheres. (E) X-ray diffraction pattern of scale samples taken from the depth interval 1710-1713 m showing a broad opal-A peak centered at 22o 2-theta and quartz peaks at 21.5o and 26.8o 2-theta. Calcite 49 Figs. 2.5A, 2.5B and 2.5C suggest that the denser layers develop as the silica spheres form strands and sheets. The formation of strands of small spheres suggests a progression to a more stable silica form. The silica plates in Fig. 2.5C appear to be formed from coalesced opal-A spheres. An unusual tube like structure coated with silica spheres is shown in Fig. 2.5D. Similar features, interpreted as silicified bacteria, have also been observed in sinters from geothermal fields in New Zealand (Rodgers et al., 2004). The X-ray pattern (Fig. 2.5E) of a sample from 1710-1713 m indicates the silica consists of opal-A with a broad peak centered at 22o 2-theta and quartz with peaks at 21.5o and 26.8o 2-theta. Discussion Monomeric and polymeric deposition are two mechanisms of silica precipitation (Iler, 1979). Direct deposition of silica molecules onto solid surfaces is referred to as monomeric deposition. The formation of a colloid in solution and its subsequent precipitation is referred to as polymeric deposition. Monomeric deposition tends to form a hard, dense deposit, while polymeric deposition forms a softer, porous silica scale. The textures of silica spheres observed in deposits from Coso well 68-20RD indicate that polymeric deposition is the dominant process of silica deposition. The growth and nucleation stages that accompany the polymerization of amorphous silica are controlled mainly by silica saturation, pH, and salinity (Makrides et al., 1978; Iler, 1979). Precipitation or amorphous silica can be further catalyzed by the presence of iron and aluminum in solution, which reduces the solubility of amorphous silica when these metals are incorporated (Gallup, 1998). At higher degrees of silica supersaturation and 50 near neutral pH, polymeric deposition is favored (Iler, 1979). As the silica precipitates, the degree of supersaturation in the fluid decreases, favoring monomeric deposition. Also, the chemical composition of the injected fluid is variable (see Table 2.1), including varying degrees of silica saturation. These mechanisms can explain the variety in textures and observed porosity in the amorphous silica scale. When compared to the maturation sequence observed and documented in geothermal sinter deposits (Rodgers et al., 2004; Lynne and Campbell, 2004; Lynne et al., 2007), several textures observed in the Coso scale indicate a maturation of silica during and/or after deposition. The observed textures include nano and micro spheres (Fig. 2.4A), botryoidal coalesced microspheres (Fig. 2.4A), bumpy microspheres (Fig. 2.4B), sheets (Fig. 2.5A) and aligned chains of amorphous silica (Fig. 2.5B and 2.5C). Maturation of the silica usually leads to increased porosity and permeability when the opal-A phase progresses to opal-CT. However, infilling of spaces within aggregates of colloidal particles that leads to sheet like textures observed at Coso may provide a barrier to further maturation of the silica by depriving contact with fluid necessary to dissolve and re precipitate silica as a more stable phase. This infill could also lead to difficulty in removing this scale as it becomes a barrier to fluid flow over time. Mineral Scale Summary Examination of cuttings from redrilled injection wells at the Coso geothermal field has yielded direct evidence for relating injectivity losses to mineral precipitation. Deposits of amorphous silica associated with traces of calcite were found in the reservoir rocks adjacent to the original injection well 68-20. This well had experienced a 51 significant loss in injectivity within a period of 1 1/2 years. The silica deposits are layered, with individual layers ranging from tens to hundreds of micrometers. Apparent porosities vary from layer to layer with some displaying little visible pore space. Textural relationships indicate that the silica was originally deposited as 1-2 m spheres of opal- A. The size and uniform diameter of the spheres suggests the silica layers formed as a colloidal precipitate. As the deposits mature, botryoidal, coalesced microspheres and bumpy microspheres up to 10 m in diameter form. With further maturation, infilling of pore spaces between spheres results in the formation of silica plates and sheets, which could form a barrier to fluid flow that could slow or stop the maturation sequence of the silica. One Dimensional Reactive Transport Modeling Modeling Approach Simulations were carried out using the nonisothermal reactive geochemical transport code TOUGHREACT (Xu and Pruess, 2001; Xu et al., 2004). This code was developed by introducing reactive chemistry into the framework of the existing multiphase fluid and heat flow code TOUGH2 V2 (Pruess et al., 1999, see also http://www-esd.lbl.gov/TOUGHREACT/). Interactions between mineral assemblages and fluids can occur under local equilibrium or kinetic rates. Precipitation and dissolution reactions can change formation porosity and permeability. This simulator can be applied to one, two, and three dimensional porous and fractured media with physical and chemical heterogeneity. Simulations can include any number of species present in the liquid, solid, and gaseous phases. Various thermal, physical, and chemical processes are 52 considered under conditions of pressure, temperature, water saturation, ionic strength, pH, and Eh. The current models do not consider processes related to certain types of mineral precipitation and maturation kinetics, including nucleation, formation of metastable phases and their transformation to stable phases, and Ostwald ripening (Xu et al., 2007). Simulation Setup The conceptual model considers a one dimensional flow path between the injection and production wells, which is a small subvolume of the more extensive three-dimensional reservoir. The initial reservoir conditions were 275oC and 30 MPa pressure. An over pressure of 2 MPa was applied to simulate injection. The model is based on conditions during nearly continuous injection over 7 years. The control case model uses measured, observed, and estimated parameters from data gathered through various studies at Coso (Lutz and Moore, 1997; Kovac et al., 2005; McLin et al., 2006). Further cases are based on hypothetical situations where these parameters are adjusted to determine the sensitivity of the modeling, as well as predict alternative reservoir conditions. The simulations were run to a total time of 7 years. Changes in fluid pH, fracture porosity, fracture permeability, fluid temperature, and changes in mineral abundances were monitored out to a distance of 594 m from the injection well. Mineral abundance changes were reported in terms of changes in volume fraction for the following minerals: quartz, potassium feldspar, chlorite, illite, Na smectite, Ca smectite, calcite, dolomite, anorthite, annite, and amorphous silica. Amorphous silica, calcite, and quartz displayed the most significant changes. Changes in porosity were calculated as a function of mineral 53 dissolution and precipitation. A porosity increase indicates that mineral dissolution is dominant, while a porosity decrease occurs when precipitation dominates. Changes in permeability were calculated from changes in porosity using a cubic law to calculate the relationship between porosity and permeability. Fluid and Heat Flow Conditions The geometry and fluid and heat flow conditions are modeled after those described in Xu and Pruess (2004). A one dimensional MINC (multiple interacting continua) model was used to represent the fractured rock. The MINC method can resolve "global" flow and diffusion of chemicals in the fractured rock and interaction with "local" exchange between fractures and matrix. Details on the MINC method for reactive geochemical transport are described by Xu and Pruess (2001). In the simulations, interactions with 1) a zone representing the relatively impermeable, unaltered host rock, and 2) altered, fractured, and veined host rock were considered. In addition, two different reservoir rock types, diorite and granodiorite, were considered. These are the rock types are the dominant lithologies where amorphous silica deposition has been observed. The hydrologic and thermal parameters used in the models are shown in Table 2.2. Density = 2650 kg*m-3, heat capacity = 1000 J*kg-1K-1, and diffusivity = 10-9 m-2s-1 were used for all zones. The cubic law was used to define the porosity-permeability relationship in both zones (Xu et al., 2004). The model generates changes in porosity and permeability based on changes in mineral abundances. 54 Mineralogical Conditions The reservoir rocks at a depth of 878 m in 68-20RD are dominated by hornblende-biotite-quartz diorite. Biotite granodiorite dominates at 1710 m. The mineralogical compositions of these rocks were estimated from petrographic observations of samples from 68-20RD and from X-ray and thin section studies of East Flank wells by Kovac et al. (2005) and Lutz and Moore (1997). Both rock types were found to be only weakly altered in 68-20RD. The veining was found to be especially weak in the diorite at 878 m, so the fracture zone was modeled as nearly empty. The deeper granodiorite zone was found to contain quartz, calcite, and chlorite veins. Mineral parameter inputs are shown in Table 2.3. Table 2.2 Hydrologic and Thermal Parameters _______________________________________________________________________ Hydrologic and thermal parameters of rocks used in the models _______________________________________________________________________ Parameters Fracture Average Weakly Altered Granodiorite Average Weakly Altered Diorite Volume (m3) 0.1 0.9 0.9 Permeability (m2) 2.0E-12 2.0E-18 2.0E-15 Porosity 0.10 0.02 0.05 Thermal Conductivity (W* m-1K-1) 2.9 3.0 3.0 Tortuosity 0.3 0.1 0.1 _______________________________________________________________________ 55 Table 2.3 Mineralogical Parameters _______________________________________________________________________ Simplified initial mineralogical composition of the two rocks used in the preliminary simulations. A temperature of 275oC was used for the initial rock temperature in the simulations. _______________________________________________________________________ Mineral Volume Fraction of Solid Rock 1710 m: Granodiorite hosted 878 m: Diorite hosted Average Weakly Altered Granodiorite Fracture Average Weakly Altered Diorite Fracture Quartz 0.34 0.05 0.135 Potassium Feldspar 0.17 0.045 Chlorite 0.02 0.01 0.010 Illite 0.03 Calcite 0.02 0.04 0.025 0.009 Anorthite 0.33 0.038 Annite 0.06 0.150 ________________________________________________________________________ Mineral Kinetic Rates and Parameters Mineral dissolution and precipitation are considered under kinetic constraints. The general kinetic rate expression is used in TOUGHREACT (Xu et al., 2004) is: rm = ±kmAmaH+ n |1- Qm/Km| (2.1) where m is the mineral index, rm is the dissolution/precipitation rate, (positive for dissolution, negative for precipitation), km is the rate constant (moles per unit mineral surface area and unit time) which is temperature dependent, Am is the specific reactive surface area per kg of H2O, aH+ is the activity of H+ , and n is an empirical reaction order 56 accounting for catalysis by H+ in solution. Km is the equilibrium constant for the mineral water reaction written for the destruction of one mole of mineral m, Qm is the ion activity product. The temperature dependence of the reaction rate constant can be expressed as: k = k25 exp[-Ea/R(1/T-1/298.15)] (2.2) where Ea is the activation energy, k25 is the rate constant at 25oC, R is the universal gas constant, and T is absolute temperature. Table 2.4 shows the parameters used in the kinetic rate expression. Water Chemistry The composition of the reservoir fluid was estimated from the approximate composition of reservoir fluid taken from an East Flank well (Table 2.4). Initial fluid compositions within the fracture and host rock were calculated by equilibrating the reservoir fluid composition with each rock's mineralogical composition at 275oC. An example injection fluid composition that was calculated from 15 well 68-20 injection fluid analyses as representative injection fluid was chosen as the trial injection water (Table 2.5). The injectate composition was not allowed to change over time within the model. Results In the modeled control case for both diorite and granodiorite hosted fracture systems, amorphous silica precipitates soon after injection begins, followed by calcite and a very minor amount of quartz (Figs. 2.6A, 2.6B, 2.6C, and 2.6D), consistent with 57 Table 2.4 Kinetic Rate Parameters _______________________________________________________________________ List of kinetic rate parameters used in Eqns. (1) and (2) for minerals considered in the present paper (Xu and Pruess, 2004; Palandri and Kharaka, 2004). The first line indicates dissolution parameters and the second line precipitation parameters; the same values were used for both where only one line is shown. _______________________________________________________________________ Mineral k25 (moles m-2 s-1) Ea (KJ/mole) n (rxn. order) Surface Area (cm2/g) Quartz 1.2589E-14 87.5 0 9.1 Am. Silica 7.3200E-13 1.0000E-10 60.9 0.00 00 1.0E6 1.0E6 K-feldspar 1.0000E-12 57.78 0 9.1 Anorthite 1.0000E-12 57.78 0 9.1 Na smectite 1.0000E-14 58.62 0 108.7 Ca smectite 1.0000E-14 58.62 0 108.7 Illite 1.0000E-14 58.62 0 108.7 Annite 2.5119E-15 2.5119E-15 66.20 66.20 10 9.1 9.1 Calcite 6.9183E-2 6.4565E-7 18.98 62.76 109.1 9.1 Dolomite 1.0233E-3 4.4668E-10 20.90 62.76 10 9.1 9.1 Chlorite 2.5119E-12 62.76 0 9.1 _______________________________________________________________________ 58 Table 2.5 Fluid Chemistry _______________________________________________________________________ Composition of a reservoir fluid from an East Flank well at 275oC, the temperature used in the simulations, and average injection fluid composition at 120oC temperature based on 15 samples. _______________________________________________________________________ Chemical Component Reservoir (Mol/kg) Injection (Mol/kg) SiO2 1.30E-2 9.96E-03 B(OH)3 8.42E-3 1.01E-02 Na+ 9.50E-2 1.46E-01 K+ 1.20E-2 1.42E-02 Li+ 2.45E-3 4.44E-03 Ca2+ 9.55E-4 1.07E-03 Mg2+ 4.12E-6 2.22E-2 Sr2+ 3.60E-5 5.00E-2 Cl- 1.10E-1 1.60E-01 F- 1.47E-4 1.15E-04 HCO3 - 1.10E-3 2.48E-03 SO4 - 3.12E-4 6.97E-04 HS- 3.02E-5 CH4 6.25E-10 pH 6.84 6.47 As 1.16E-04 _______________________________________________________________________ observed mineral scale deposits in the injection well 68-20RD (Figs. 2.7A and 2.7B). Amorphous silica deposition did not occur in a case where silica concentration was decreased in the injected fluid by an order of magnitude (Figs. 2.6A and 2.6D). In both the diorite and granodiorite hosted cases, the majority of amorphous silica precipitates within a few meters of the well and within the first year of the simulation (Fig. 2.8 A and C). Minor calcite precipitation is also precipitated rapidly within the first few meters (Fig. 2.8 B and D). Modeled porosity and permeability decreases are also predicted within a few meters of the well over the same time period. 59 Fig. 2.6 Results of TOUGHREACT models. (A) and (B) Modeled results for the diorite hosted case showing amorphous silica (A) and calcite (B) as a function of distance in meters at t = 7 years. (C) and (D) Modeled results for the granodiorite hosted case showing amorphous silica (C) and calcite (D) as a function of distance in meters at t = 7 years. 60 Fig. 2.7. SEM images showing calcite in scale. (A) SEM image of opal-A and calcite from the depth interval 869-884 m. (B) SEM image of opal-A with small white calcite crystals from the depth interval 1710-1713 m. 61 Fig. 2.8. Modeled results using the control injection chemistry for the diorite hosted case ((A) and (B)) and the granodiorite hosted case ((C) and (D)) showing amorphous silica and calcite as a function of distance in meters for various time periods over seven years. (A) (B) (C) (D) 62 Well 68-20, the first well drilled on the 68-20 pad, experienced a sharp decline in injectivity within months of initiation of injection. Significant amorphous silica scale deposits have only been observed in two of the subsequently drilled injection wells studied on the 68-20 pad, wells 68-20RD and 68B-20RD. The trajectory of 68-20RD lies only several meters from the original well 68-20. The trajectory of 68B-20RD lies near that of 68A-20RD. In contrast, the trajectories of wells 68A-20 and 68A-20RD place them hundreds of meters from the trajectories of previously drilled injection wells. Although the trajectory of 68B-20 is close to that of both 68-20 and 68-20RD, no scale was observed in the cuttings from this well. Conclusions and Future Work The modeled results closely simulate the observed mineral paragenesis and abundances based on observations made on cuttings from injection well 68-20RD. Future investigations will consider the effects of varying the composition of the injected fluid and changing temperatures. Temperatures profiles were measured for wells 68-20 and 68-20RD. Lower temperatures measured in 68-20RD indicate that the surrounding rock has cooled over time due to the injection of cooler fluids into 68-20 over the period of four and a half years. Attempts are being made at many geothermal fields, including Coso, to increase the solubility of silica in injected fluids by decreasing the pH. It is not yet known what effect acidifying the injected fluid will have on the system. Future modeling will attempt to predict the effect of pH modifications on mineral deposition and dissolution within the reservoir. 63 References Adams, M.C., Moore, J.N., Bjornstad, S., Norman, D.I., 2000. Geologic history of the Coso geothermal system. Geothermal Resources Council Transactions 24, 205-209. Iler, R.K., 1979. The Chemistry of Silica-Solubility, Polymerization, Colloid, and Surface Properties, and Biochemistry. John Wiley & Sons, Inc., New York. Kovac, K.M., Moore, J.N., Lutz, S.J., 2005. Geologic framework of the East Flank, Coso geothermal field: Implications for EGS Development. Proceedings, 30th Workshop on Geothermal Reservoir Engineering. Lutz, S.J., Moore, J.N., 1997. Petrographic and x-ray diffraction study of 130 cuttings samples from six wells in the Coso geothermal area, California. Unpublished CalEnergy Corporation Report. Lynne, B.Y., Campbell, K.A., 2004. Morphologic and mineralogic transitions from opal- A to opal-CT in low-temperature siliceous sinter diagenesis, Taupo Volcanic zone, New Zealand. Journal of Sedimentary Research 74, 561-579. McLin, K.S., Moore, J.N., Hulen, J., Bowman, J.R., Berard, B., 2006. Mineral characterization of scale deposits in injection wells; Coso and Salton Sea geothermal fields, CA. Proceedings 31st Workshop on Geothermal Reservoir Engineering. Palandri, J.L., Kharaka, Y.K., 2004. A compilation of rate parameters of water-mineral interaction kinetics for application to geochemical modeling. U.S. Geological Survey Open File Report 2004-1068. Pruess, K., Oldenburg, C., Moridis, G., 1999. TOUGH2 user's guide, Version 2.0. Lawrence Berkeley Laboratory Report LBL-43134. Rodgers, K.A., Browne, P.R.L., Buddle, T.F., Cook, K.L., Greatrex, R.A., Hampton, W.A., Herdianita, N.R., Holland, G.R., Lynne, B.Y., Martin, R., Newton, Z., Pastars, D., Sannazarro, K.L., Teece, C.I.A., 2004. Silica Phases in sinters and residues from geothermal fields of New Zealand. Earth Science Reviews 66, 1-61. Xu, T., Pruess, K., 2004. Numerical simulation of injectivity effects of mineral scaling and clay swelling in a fractured geothermal reservoir. Geothermal Resources Council Transactions 28, 269-276. Xu, T., Sonnenthal, E., Spycher, N., Pruess, K., 2004. TOUGHREACT user's guide: A simulation program for non-isothermal multiphase reactive geochemical transport in variably saturated geologic media. Lawrence Berkley National Laboratory publication LBNL-55460. 64 Xu, T., Pruess, K., 2001. Modeling multiphase non-isothermal fluid flow and reactive geochemical transport in variably saturated fractured rocks: 1. Methodology. American Journal of Science 301, 16-33. 65 CHAPTER 3 MODELING THE GEOCHEMICAL EFFECTS OF INJECTING pH MODIFIED FLUIDS AT COSO GEOTHERMAL FIELD, CA Abstract Decreased performance of injection wells has been documented in several geothermal fields after very short periods of injection. At the Coso, CA geothermal field, the fluids injected into wells on pad 68-20 were supersaturated with respect to amorphous silica. Examination of drill cuttings from the original and five redrilled injection wells on pad 68-20 indicated that opal-A, accompanied by trace amounts of calcite, was precipitating near the well bore, causing the observed declines in injectivity. Injection fluids at Coso are now modified with H2SO4 to decrease fluid pH as a method of reducing the deposition of amorphous silica scale by reducing the kinetic rate of silica polymerization. One dimensional (1D) models were constructed using the reactive transport code TOUGHREACT (Xu et al., 2004) to assess amorphous silica precipitation and the long term effects of injecting acidic fluid into the geothermal reservoir at Coso. These simulations predict that injection of fluid supersaturated with respect to amorphous silica will cause rapid declines in injectivity due to precipitation of opal-A, consistent with observations of amorphous silica scaling in cuttings from injection wells drilled on pad 68-20 at Coso. The modeling results also predict that an optimal injection fluid pH of 4 would virtually eliminate amorphous silica and calcite deposition within fractures and therefore maintain porosity and permeability. The fracture fluid pH is buffered rapidly upon injection into the reservoir rock due to silicate mineral reactions and SiO2(aq) in solution. Although changes to porosity and permeability of the fracture or reservoir rock are not predicted by these models, mineral dissolution and precipitation are predicted to occur. Porosity and permeability of the fracture and reservoir rock are maintained by equal volume dissolution and precipitation of minerals in these models, but processes that are not considered by the models, such as preferential precipitation in pore throats and pressure solution at grain contacts, may still change porosity and permeability along the flow path from injection to production well. Introduction The geochemical effects of injecting fluids into geothermal reservoirs are poorly understood and may be significantly underestimated. Decreased performance of injection wells has been observed in several geothermal fields after only a few years of service, including the six injection wells on pad 68-20 in the Coso, CA geothermal system. Rock cuttings from original and redrilled injection wells drilled on pad 68-20 at Coso were used to characterize the mineral and geochemical changes that occur as a result of injection. Samples from the original well 68-20 at Coso were used to establish the mineral assemblages and their geochemical characteristics prior to injection. Amorphous silica scale deposits were found in two of the five subsequently drilled injection wells on 67 pad 68-20. Based on these initial studies, McLin et al., (2006a) identified that precipitation of amorphous silica led to declining injectivity in these wells. Because amorphous silica scale is difficult to remove once it has precipitated, modifications to the injected fluid are made to prevent this precipitation. Silica scale control methods used in geothermal operations include injecting at higher temperature, dilution with condensate or cool surface or groundwater, reduction of fluid pH to slow polymerization kinetics, and a variety of treatments with scale inhibitors (Gallup, 1998). Of these options, reduction of fluid pH is often the most practical and economic solution due to the expense of inhibitors and the lack of fresh water for injection at many sites. H2SO4 is commonly used to reduce the pH (Gallup and Kitz, 1998; Gallup, 1997). The target range for injection in a geothermal field is typically between pH=5.25 and pH=4.75 to minimize both silica scale precipitation and corrosion (Darrell Gallup, personal communication). Although pH modification is a commonly employed practice, including at Coso, the effects of these modifications on the reservoir fluids and rocks are poorly understood and should be examined. Reactive transport modeling is an ideal geochemical method for predicting the effects of injecting flashed brine and brine modified with H2SO4 on the reservoir fluids and rocks at the Coso geothermal field. Because it is difficult to measure directly the effects of H2SO4 injection on the reservoir rocks and fluids, models can provide valuable insight into the consequences of acid injection on the productivity of the Coso geothermal field. In this study, the results of previous petrological investigations of amorphous silica scale deposition in Coso reservoir rock (McLin et al., 2006a) and reactive transport modeling of amorphous silica precipitation in fractured Coso reservoir rock (McLin et al., 2006b) were used to 68 constrain numerical modeling of fluid-rock interactions and mineral changes with H2SO4 injection in the near well bore and downstream reservoir environments. The objectives of this study were to evaluate the geochemical and mineralogical impacts of fluid injection and of reducing silica scale precipitation with H2SO4 injection on reservoir rocks and fluids. Study Area The Coso geothermal field is developed in Mesozoic granitic rocks of the Sierra Nevada Batholith on the western edge of the Basin and Range (Adams et al., 2000) (Fig. 3.1). The heat driving the geothermal activity is related to shallow intrusions that have given rise to 38 rhyolitic domes during the last million years. The reservoir host rocks range in composition from diorite to granite with varying degrees of alteration and veining (Kovac et al., 2005). Active and fossil fumaroles lie along a northeast to southwest trending belt that extends through Devil's Kitchen and Coso Hot Springs. On the eastern margin of the field, known as the East Flank, fossil sinter and travertine deposits are present (Adams et al., 2000). The alteration mineral assemblage identified in East Flank wells related to the current geothermal system consists of calcite veins with minor quartz, chlorite, pyrite, and hematite at shallow depths and epidote, chlorite, quartz, adularia, and wairakite in the deepest portions (Kovac et al., 2005). Geothermal power production has been sustained since 1989 with an installed capacity of 240 MWe at the power plant. 69 Fig. 3.1. Simplified geologic map of the Coso geothermal field showing the locations of the major thermal features. The 68-20 injection pad is located in the southern part of the field. 70 63· 18 . 63·18RD o o57 · 18TC•H o .5 lkm Geothermal Well Active Fumarole Fault (dashed where inferred) • Non-Calcareous Alteration • Calcitic 5tockworks and Travertine D Quaternary Surfi cial Deposits(unpanerned) D Quaternary Rhyolitic Pyroclastics D Quaternary Rhyolite Domes D Ter.-Quat. Volcanic/ Sedimentary Rocks D Mesozoic Granites/ Metamorphics 15-17 • • 72-19 • 73-19 '---------../ o 68-20 • - - / / / / / / /. / -- - Coso Hot Wheeler N , Well pad 68-20 Between 1987 and 1993, six injection wells were drilled on the 68-20 pad in the southern part of the field. The trajectories of these wells and lost circulation zones obtained from well logs are shown in Fig. 3.2. Injectivity decreased in well 68-20 from a maximum liquid injection rate of over 1000 kph (kilopounds per hour mass flow rate) in March, 1989 to a minimum rate of 0 kph in November, 1990 after a steady decline. Cleanouts increased injectivity to a one time high of 800 kph after November, 1990, but never fully recovered, remaining between 0-400 kph through 1992. Redrilled wells also experienced similar injectivity declines. Although one well on the East Flank reaches nearly 350oC at depth, reservoir temperatures in the part of the field near pad 68-20 prior to injection ranged from approximately 205-240oC. Cooling of the reservoir was observed around these wells post injection as seen in temperature logs of subsequent redrills. The temperatures of the injected fluids ranged from 110-120oC. Cuttings from the six injection wells were sampled at 3 m intervals. The reservoir host rocks in these wells ranged from diorite to granite with trace to moderate alteration and veining. Fault breccias were observed in the cuttings, indicating major fault or fracture zones intersected by the wells. Table 3.1 shows the reported chemistry of injection for well 68-20. There was a wide range in the composition of the injected fluids, resulting from variation in produced fluids and practices of adding steam condensate back into the injected fluid. The maximum silica content analyzed was 941 mg/L. The minimum silica concentration of 97 mg/L corresponds to mixing produced fluid with steam condensate. 71 Fig. 3.2. Well trajectories for injection wells drilled in pads 67-17 (not discussed in this paper) and 68-20. Locations of lost circulation zones are shown as discs, and the amount of fluid lost is represented by the size of the disc. X and Y axes in UTMs, Z axis in feet. The green and red arrows point north. 72 Table 3.1 Chemistry of Injected Fluid, Well 28-20 ____________________________________________________________________________________________________________ Injected fluid chemistry from well 68-20, showing concentrations in mg/kg from 15 analyses. Brines were injected at 110-120oC. ____________________________________________________________________________________________________________ Date Na K Ca Mg Fe Al SiO2 B Li As HCO3 Cl F SO4 TDS Lab (mm/dd/yy) pH 06/27/88 3340 725 78 8.7 33 10 555 100 37 9.2 229 5570 2.4 73 10900 6.7 10/04/88 3520 941 64 0.27 0.49 >0.61 965 116 47 11 77 6600 2.8 27 12300 7.2 03/30/89 2897 362 19 0.08 0.30 0.37 617 88 25 8.3 156 5015 2.3 80 9233 6.8 06/15/89 2993 465 22 0.09 0.13 0.56 678 83 41 11 195 5465 2.2 90 nr 7.3 09/27/89 3652 501 29 0.14 0.01 0.25 174 119 33 10 178 6440 1.7 95 11237 7.5 11/13/89 3470 580 25 0.12 2.53 0.63 698 114 31 6.9 168 5165 1.6 71 nr 6.8 01/07/90 3540 590 24 0.15 1.88 0.39 826 129 35 8.0 225 6018 2.3 44 nr 6.7 04/20/90 4283 633 36 0.65 84.1 1.6 747 127 35 26 142 5698 5.7 90 nr 6.7 08/19/90 4010 739 32 0.14 2.85 0.50 936 142 40 7.2 151 6958 3.0 27 nr 7.3 10/12/90 3908 672 38 0.15 0.33 0.33 701 121 37 8.8 161 6618 2.1 47 nr 8.2 02/03/91 3774 595 47 0.46 0.12 >0.61 545 109 31 6.7 161 6340 2.5 99 11550 8.2 05/07/91 3480 605 37 >0.16 0.88 >0.02 593 113 28 2.9 174 6070 2.2 87 11000 8.3 08/03/91 4192 590 130 3.94 6.5 1.0 620 130 30 11 185 6855 2.8 53 nr 8.2 11/21/91 3508 601 46 0.14 0.30 0.43 546 125 31 6.1 143 6100 1.9 72 11500 8.3 12/13/96 876 147 18 nd 1.7 nd 97 36 6.8 1.9 7.5 1412 nd 59 2766 6.2 ___________________________________________________________________________________________________________ nd=not detected, nr=not recorded Note: Fluid analysis from 12/13/96 is representative of a mixture of flashed production fluid and condensed steam. 73 Observations from Well Cuttings Thinly banded opaline silica was observed in the cuttings from 68-20RD (McLin et al., 2006a) and 68B-20RD, but not in the original injection well 68-20, 68A-20, 68A- 20RD, or 68B-20. The banding and textural relationships suggest the silica represents fracture fillings and not alteration of preexisting minerals. The greatest density of silica precipitation in well 68-20RD was found in cuttings from depths of 869-884 and 1710- 1713 m. The zone from 1710-1713m depth corresponds to the depth of a zone of lost circulation in well 68-20. Scanning electron microscopy (SEM) was used to examine the textures of the amorphous silica scale. Figs. 3.3A-E are SEM images that show the morphological progression associated with deposition and maturation of the silica scale deposits. The deposits consist of opal-A deposited as colloidal particles of nano and microspheres (Fig. 3.3A), indicating that silica homogeneously nucleated in solution to form polymeric particles (Iler 1979). Several textures related to the evolution of these deposits, including coalesced (Fig. 3.3B) and botryoidal (Fig. 3.3C) microspheres, fibrils (Fig. 3.3D), and sheets (Fig. 3.3E) are observed. Textural relationships indicate the silica was deposited initially as spheres 1-2 m in diameter (Fig. 3.3A). As the deposits mature, the spheres coalesce to form larger spheres up to 10 m in diameter (Fig. 3.3B and C). Further maturation is associated with the formation of fibrils and sheets (Fig. 3.3D and E) through infilling. Traces of calcite are found deposited on the amorphous silica, suggesting it represents a later stage in the evolution of the deposits. Fig. 3.3F shows an interesting feature that has been interpreted as a silicified bacterium. 74 Fig. 3.3. SEM images of amorphous silica scale deposits from well 68-20RD. (A) Micro and nanoparticles of silica. (B) Coalesced particles of silica. (C) Botryoidal microspheres. (D) Fibrils of silica. (E) Sheet of amorphous silica. (F) Possible silicified bacterium. 75 When compared to the maturation sequence observed and documented in geothermal sinter deposits (Rodgers et al., 2004; Lynne and Campbell, 2004; Lynne et al., 2007), several textures observed in the Coso scale indicate a maturation of opal-A during and/or after deposition. Maturation of the silica usually leads to increased porosity and permeability when the opal-A phase progresses to opal-CT. However, infilling of spaces in colloidal particle deposits that leads to sheet like textures observed at Coso may provide a barrier to further maturation of the silica by depriving contact with fluid necessary to dissolve and reprecipitate silica as a more stable phase. This infill could also lead to difficulty in removing this scale as it becomes a barrier to fluid flow over time. The X-ray diffraction patterns of the amorphous silica scale samples are characterized by a broad peak centered at 22o 2-theta representing opal A, and peaks at 21.5o and 26.8o 2-theta representing quartz Fig. 3.4. Although quartz was not documented in the SEM images, it is possible that it represents fragments of the host reservoir rock. Alternatively, it is possible, but less likely, that the quartz represents maturation of the opal-A to a higher degree of crystallinity. Although quartz is common in mature sinter deposits at other geothermal fields (Lynne et al., 2004), there is no evidence from the SEM or X-ray diffraction studies of stages in silica maturity beyond opal-A in the Coso scale deposits. Previous Coso Modeling Studies Prior to examination of the cuttings from the wells on pad 68-20, Adams et al. (2005) examined the geochemical consequences of injecting groundwater at Coso 76 Fig. 3.4. X-ray diffraction pattern of amorphous silica scale from 1710-1713 m depth. geothermal field as a strategy for minimizing precipitation of amorphous silica in the near-wellbore environment. The modeling code REACT (Bethke, 1996) was used to calculate saturation states of minerals to predict fluid-rock interactions with this injection strategy. With the addition of Mg2+ and Ca2+ rich groundwater, the precipitation of anhydrite + dolomite or anhydrite + calcite and a magnesium silicate was predicted. Following the examination of cuttings from wells 68-20 and 68-20 RD, a small set of initial models were run using the nonisothermal reactive transport code TOUGHREACT (Xu et al., 2004) to examine the precipitation of amorphous silica during injection (McLin et al., 2006b). An injection fluid based on the 15 analyses of injection fluid from well 68-20 was used, and fluid was injected into a one dimensional, 594 m long flow path that consisted of a zone of fractures and a zone of altered host rock (either granodiorite or diorite). Based on data from Adams et al. (2005), the reservoir temperature used for these simulations was 275oC. In the modeled cases using injection fluid with 650 ppm SiO2(aq), amorphous silica precipitated and significantly reduced 77 porosity of the fracture within the first meter of the flow path within the first year. Modeled silica precipitation was followed by calcite and a very minor amount of quartz in the near wellbore environment, consistent with observed mineral deposits in the injection well cuttings. These results are consistent with the rapid decline in injectivity experienced by well 68-20. Amorphous silica scale is only found in redrilled injection wells that have a trajectory very close to that of a previously drilled well, and amorphous silica was predicted to deposit only within a few meters of the original well. Thus the modeled results closely simulated the observed mineral paragenesis and abundances based on analyses of the cuttings. The effects of modifying the pH of the injection fluid, from acidic to basic, to mitigate the effects of silica deposition were investigated by Park et al. (2006). Acid injection reduces the kinetic rate of silica precipitation by reducing the rate of silica polymerization (Iler, 1979; Rothbaum et al., 1979; Klein, 1995). Park et al. (2006) constructed a one dimensional reactive transport model to investigate the consequences of injection of acid or alternating acid and base on the reservoir rocks and fluids. Park et al. (2006) explain that injection of a basic solution will lead to SiO2 undersaturation and dissolution by the increased activity of H3SiO4 - and the formation of NaHSiO3(aq). The models predicted that with acid injection SiO2 deposition was mitigated in the immediate vicinity of the injection well, but deposition of SiO2 was predicted at greater distances. Injection of base also mitigated SiO2 precipitation, but the precipitation of calcite was predicted. Thus alternating the pH of the injected fluid to prevent mineral scale, as well as to maintain neutral fluid pH in the reservoir near production wells over time, was predicted to enhance permeabilities at Coso. However, the acid fluids modeled for 78 injection were very low pH (pH=2.26), and thus could cause extensive corrosion of equipment and well casing. Therefore, injection of fluid at this pH is not practical. Modeling Approach The current simulations were carried out using the nonisothermal reactive geochemical transport code TOUGHREACT (Xu and Pruess, 2001; Xu et al., 2004). This code was developed by introducing reactive chemistry into the framework of the existing multiphase fluid and heat flow code TOUGH2 V2 (Pruess et al., 1999, see also http://www-esd.lbl.gov/TOUGHREACT/). Interactions between mineral assemblages and fluids can occur under local equilibrium or kinetic rates. Precipitation and dissolution reactions can change formation porosity and permeability. This simulator can be applied to one, two, and three dimensional porous and fractured media with physical and chemical heterogeneity. Simulations can include any number of species present in the liquid, solid, and gaseous phases. Various thermal, physical, and chemical processes are considered under conditions of pressure, temperature, water saturation, ionic strength, pH, and Eh. However, the current models do not consider processes related to certain types of mineral precipitation and maturation kinetics, including nucleation, formation of metastable phases and their transformation to stable phases, and Ostwald ripening (Xu et al., 2007). Simulation Setup The conceptual model considers a one dimensional flow path between the injection and production wells, which is a small subvolume of the more extensive three 79 dimensional reservoir. The geometry and fluid and heat flow conditions of the model were based on those of Xu and Pruess, (2004). The thermodynamic database provided with the TOUGHREACT program, modified from the EQ3/6 database (Wolery, 1992) as described by Xu et al., (2004), was used. The initial reservoir conditions were 220oC and 30 MPa pressure. The decrease in model reservoir temperature from that used in previous models is based on temperature surveys from well 68-20 and better reflects the initial reservoir temperature prior to injection. An over pressure of 2 MPa was applied at the boundary of the flow path to simulate fluid injection. The model is based on conditions during continuous injection over seven years. The initial models use measured, observed, and estimated parameters from data gathered through various studies at Coso (Lutz and Moore, 1997; Lutz et al., 1999; Kovac et al., 2005; McLin et al., 2006a, 2006b). Further cases are based on hypothetical situations where these parameters are adjusted to determine the sensitivity of the modeling, as well as to predict the effects of alternative reservoir conditions. Finally, injection of H2SO4 modified fluid is modeled for mixtures of injectate with 98% H2SO4 at pH=3, 4, and 5. The simulations were each run to a total time of 7 years. Changes in fluid pH, fracture porosity, fracture permeability, fluid temperature, and changes in mineral abundances were monitored out to a distance of 594 m from the injection well. Mineral abundance changes were reported in terms of changes in volume fraction for the following minerals: quartz, potassium feldspar, chlorite, illite, Na smectite, Ca smectite, calcite, dolomite, albite, oligoclase, anorthite, annite, phlogopite, clinozoisite, anhydrite, and amorphous silica. Changes in porosity were calculated as a function of mineral dissolution and precipitation. A porosity increase indicates that mineral dissolution is 80 dominant, while a porosity decrease occurs when precipitation dominates. Changes in permeability were calculated from changes in porosity using a cubic law to calculate the relationship between porosity and permeability (as discussed in Xu et al., 2004). Fluid and Heat Flow Conditions The geometry and fluid and heat flow conditions are modeled after those described in Xu and Pruess (2004). A one dimensional MINC (multiple interacting continua) model was used to represent the fractured rock. The MINC method can resolve mass transport from "global" flow and diffusion of chemical species within the fractured rock from transport by "local" exchange between fluid within fractures and the minimally permeable rock matrix. Details on the MINC method for reactive geochemical transport are described by Xu and Pruess (2001). In the simulations, interactions with 1) a zone representing the relatively impermeable, unaltered host rock, and 2) a fracture zone within the host rock were considered. The model allows these two zones in the flow column to have individually assigned porosity and permeability. The fluid can flow in and chemically interact with rock both zones. Mass fluxes and reaction progress within each zone will be a function of the assigned parameters. The parameters used in the models are shown in Table 3.2. Density = 2650 kg*m-3, heat capacity = 1000 J*kg-1K-1, and diffusivity = 10-9 m-2s-1 were used for both fracture and reservoir rock zones. The cubic law was used to define the porosity-permeability relationship in both zones (Xu et al., 2004). The model generates changes in porosity and permeability based on changes in mineral abundances. 81 Mineralogical Conditions Mineral modes used for the simulations are shown in Table 3.3. The reservoir rocks at depths of 878 m and 1710 m in 68-20RD are dominated by hornblende biotite quartz diorite and biotite granodiorite, respectively. These are the two depths were the most amorphous silica scale was observed in the cuttings. Because lost circulation was associated with 1710 m depth, the majority of the models are based on injection into granodiorite. The mineralogical compositions of the granodiorite and diorite were estimated from petrographic observations of samples from 68-20RD and from X-ray and thin section studies of East Flank wells by Kovac et al. (2005), Lutz et al. (1999), and Lutz and Moore (1997). Initially, anorthite was used as the composition of the plagioclase feldspar in the reservoir rock. This composition served as a noncalcite source of calcium, and the alteration mineralogy observed by Kovac et al. (2005) is an alteration product of Ca rich feldspar in the presence of CO2. However, the composition of the feldspar in the initial granodiorite is likely more albite rich (e.g., oligoclase), and feldspar near fractures is likely already altered to a Na rich feldspar composition from prior interaction with reservoir fluids. Both rock types were found to be only weakly altered in 68-20RD. The veining was found to be especially weak in the diorite at 878 m. The granodiorite was found to contain quartz, calcite, and chlorite veins. Due to mineralogical variations and uncertainties, sensitivity studies were conducted for a range of mineralogical compositions of the rocks, including diorite as the reservoir host rock, different composition of plagioclase feldspar in the reservoir rock (anorthite (an), albite (al), and oligoclase (olig)), and increased mafic minerals biotite 82 Table 3.2 Hydrologic and Thermal Parameters _______________________________________________________________________ Hydrologic and thermal parameters of rocks used in the models _______________________________________________________________________ Parameters Fracture Weakly Altered Reservoir Rock Volume (m3) 0.1 0.9 Permeability (m2) 2.0E-12 2.0E-18 Porosity 0.10 0.02 Thermal Conductivity (W* m-1K-1) 2.9 3.0 Tortuosity 0.3 0.1 _______________________________________________________________________ (represented by annite and phlogopite) and epidote (represented by clinozoisite) in the reservoir rock. For three models, low albite was allowed to precipitate. These sensitivity studies were constructed to explore the effects of these differing mineralogical compositions on the fluid chemistry and mineral precipitation within the system. Mineral Kinetic Rates and Parameters Mineral dissolution and precipitation are considered under kinetic constraints. A general kinetic rate expression is used in TOUGHREACT (Xu et al., 2004): rm = ±kmAmaH+ n |1- Qm/Km| (3.1) where m is the mineral index, rm is the dissolution/precipitation rate, (positive for dissolution, negative for precipitation), km is the rate constant (moles per unit mineral 83 Table 3.3 Mineralogical Parameters _______________________________________________________________________ Initial mineralogical composition of the rock types used in the preliminary simulation and mineralogical sensitivity studies. A temperature of 220oC was used for the initial rock temperature in the simulations. _______________________________________________________________________ Rock Mineralogical Composition Mineral Granodiorite (an) Diorite (an) Granodiorite (olig) Granodiorite (ab) Diorite (60% mafic) Fracture Albite 0.33 Anorthite 0.33 0.50 0.12 Oligoclase 0.33 Quartz 0.34 0.18 0.34 0.34 0.12 0.05 K Feldspar 0.17 0.07 0.17 0.17 0.06 Illite 0.03 0.03 0.03 0.01 Chlorite 0.02 0.01 0.02 0.02 0.04 0.01 Calcite 0.02 0.02 0.02 0.02 0.02 0.04 Annite 0.06 0.20 0.02 0.02 0.20 Phlogopite 0.02 0.02 0.20 Clinozoisite 0.02 0.02 0.20 _______________________________________________________________________ an=anorthite, olig=oligoclase, ab=albite surface area and unit time) which is temperature dependent, Am is the specific reactive surface area per kg of H2O, aH+ is the activity of H+ , and n is an empirical reaction order accounting for catalysis by H+ in solution. Km is the equilibrium constant for the mineral-water reaction for the dissolution or precipitation of one mole of mineral m, Qm is the ion activity product. The temperature dependence of the reaction rate constant can be expressed as: k = k25 exp[-Ea/R(1/T-1/298.15)] (3.2) 84 where Ea is the activation energy, k25 is the rate constant at 25oC, R is the universal gas constant, and T is absolute temperature. Table 3.4 shows the parameters used in the kinetic rate expression. Because precipitation rate data do not exist for most minerals, parameters for neutral pH dissolution were used to calculate precipitation rates for those minerals without precipitation rate data. The processes, different from dissolution, that are not considered in the calculation of precipitation rates include nucleation, crystal growth and Ostwald ripening processes, as well as the calculation of changes to the reactive surface area. Amorphous silica will homogeneously nucleate, polymerize, and precipitate as a colloidal particle at near neutral pH, as observed in cuttings from the Coso geothermal system, and the code does not provide for a pH dependent impact on the polymerization of silica. A surface area of 106 cm2/g was used for calculation of the rate of amorphous silica precipitation for injection fluid pH above 5 to account for these nucleation and polymerization processes within the framework of the code. This large surface area value takes into account the very small size of amorphous silica particles in solution (Parks, 1990; Xu et al., 2004) and can approximate precipitation rates that match observations of rapid injectivity declines in wells on pad 68-20. To approximate and simulate the reduced rate of polymerization with lower pH (Iler, 1979), the amorphous silica surface area term 85 Table 3.4 Kinetic Rate Parameters ___________________________________________________________________________________________________________ List of kinetic rate parameters used in Eqns. (3.1) and (3.2) for minerals considered in this study(Xu and Pruess, 2004; Palandri and Kharaka, 2004; Xu et al., 2007). The same values were used for both dissolution and precipitation, except for amorphous silica where the first line indicates dissolution parameters and the second indicates precipitation parameters. ___________________________________________________________________________________________________________ Mineral A (cm2/g) Parameters for kinetic rate law Neutral Mechanism Acid Mechanism Base Mechanism k25 (mol/m2/s) E (kJ/mol) k25 (mol/m2/s) E (kJ/mol) n (H+) k25 (mol/m2/s) E (kJ/mol) n (H+) Quartz 9.1 1.023E-14 87.7 Illite 108.7 1.660E-13 35 1.047E-11 23.6 0.34 3.020E-17 58.9 -0.4 Oligoclase 9.1 1.445E-12 69.8 2.138E-10 65 0.457 Albite 9.1 2.745E-13 69.8 6.918E-11 65 0.457 2.512E-16 71 -0.572 Low albite 9.1 2.754E-13 69.8 6.918E-11 65 0.457 2.512E-16 71 -0.572 Anorthite 9.1 7.586E-10 17.8 3.162E-4 16.6 1.411 K-feldspar 9.1 3.890E-13 38 8.710E-11 51.7 0.5 6.310E-12 94.1 -0.823 Chlorite 9.1 3.02E-13 88 7.762E-12 88 0.5 Na smectite 108.7 1.660E-13 35 1.047E-11 23.6 0.34 3.020E-17 58.9 -0.4 Ca smectite 108.7 1.660E-13 35 1.047E-11 23.6 0.34 3.020E-17 58.9 -0.4 Calcite 9.1 1.549E-6 23.5 5.012E-1 14.4 1.000 Dolomite 9.1 2.951E-8 52.2 6.457E-4 36.1 0.5 Annite 9.1 2.818E-13 22 1.445E-10 22 0.525 Phlogopite 9.1 3.981E-13 29 Clinozoisite 9.1 1.023E-12 70.7 2.512E-11 71.1 0.338 4.677E-18 79.1 -0.556 Am. Silica 9.1 or 106 4.900E-13 3.800E-10 76 49.8 Anhydrite 9.1 6.457E-4 14.3 86 was reduced to 9.1 cm2/g for injection fluid pH less than 5. Sensitivity of this amorphous silica surface area term was investigated in a series of models with granodiorite host rock (Granodiorite (an)) and injection fluid pH=6.5. Water Chemistry The composition of the reservoir fluid was estimated from the approximate composition of reservoir fluid produced from an East Flank well and was calculated by equilibrating the East Flank fluid composition with the granodiorite mineralogical composition at 220oC in batch calculations (shown in Table 3.5). An injection fluid composition that was calculated from 15 analyses of injection fluid from well 68-20 was chosen as the trial injection water (Table 3.5). Injection of H2SO4 modified fluid was modeled for mixtures of injectate with 98% H2SO4 at pH=3, 4, and 5 calculated with batch equilibrations. The composition of the injectate was constant over time within each model simulation. Results The initial model considers the injection of fluid supersaturated with respect to amorphous silica into a fracture zone in a granodiorite host rock (Granodiorite (an), Table 3.3). Several minerals dissolve and precipitate within the fracture and the altered reservoir rock. Amorphous silica precipitates in the fracture (Fig. 3.5), with the majority of the precipitation occurring within the first year of the model and within the first few meters of the flow path. Precipitation of amorphous silica fills greater than 80% of the fracture pore volume within the first meter of the flow path after seven years of injection 87 Table 3.5 Fluid Chemistry _______________________________________________________________________ Composition of a reservoir fluid from a produced fluid from Coso East Flank well 38B-9 in equilibrium with Granodiorite (an) at 220oC and injection fluid composition at 110oC based on 15 analyses of fluid injected into well 68-20. _______________________________________________________________________ Chemical Component Reservoir (Mol/kg) Injection (Mol/kg) SiO2 4.44E-3 9.96E-3 B(OH)3 3.88E-2 1.01E-2 Na+ 6.82E-2 1.46E-1 K+ 3.70E-3 1.42E-2 Li+ 1.63E-3 4.44E-3 Ca2+ 9.55E-4 1.07E-3 Al3+ 1.44E-5 8.49E-7 Mg2+ 3.23E-4 1.52E-5 Sr2+ 3.60E-5 5.00E-2 Cl- 8.42E-2 1.60E-1 F- 1.07E-4 1.15E-4 HCO3 - 8.84E-2 2.48E-3 SO4 2- 1.83E-4 6.97E-4 HS- 3.02E-5 CH4 6.25E-10 pH 6.50 6.50 _______________________________________________________________________ (Fig. 3.5). Trace quartz precipitates in the fracture and rock. No dissolution of K-feldspar is observed. However, trace dissolution of anorthite is observed in the rock. Trace illite and smectite precipitate in the rock. There is both trace dissolution and trace precipitation of chlorite within the fracture, with only trace dissolution in the rock. Calcite and trace dolomite precipitate within the fracture. Trace calcite also precipitates in the rock. Trace chlorite dissolves in the fracture and rock, and then it precipitates in the fracture. These results are consistent with the rapid decline in injectivity experienced with well 68-20 and 88 Fig. 3.5. Precipitation of amorphous silica in the fracture zone for simulation with Granodiorite (an). Amorphous silica surface area is 106cm2/g. 89 the observed scale mineralogy. These results are also consistent with the observation that amorphous silica scale was only found in wells 68-20RD and 68B-20RD, as these are