A
« Monitoring of gas and fluid transport from the hydrothermal reservoir to the Cerro Pabellón geothermal plant; international validation of analytical protocols »Dr.
PhilippeRobidoux1✉,2Emailrobidouxphilippe@gmail.comDiegoMorata1,2
MaximilianoSeguelCárdenas1,2
GermainRivera1,2
VerónicaRodríguez1,2
SantiagoMaza1,2
MarcoCecioni3
GianniVolpi3
MariannoTantillo4
ManfrediLongo4
FaustoGrassa4
PierangeloRomano4
1Department of Geology, Facultad de Ciencias Físicas y MatemáticasUniversidad de ChilePlaza Ercilla 8038370450SantiagoChile
2
A
A
Centro de Excelencia en Geotermia de Los AndesChile3Enel Green PowerChile
4INGV PalermoItaly
Author names and affiliations: Philippe Robidouxa,b, Diego Morataa,b, Maximiliano Seguel Cárdenasa,b, Germain Riveraa,b, Verónica Rodrígueza,b, Santiago Mazaa,b, Marco Cecionic, Gianni Volpic, Marianno Tantillod, Manfredi Longod, Fausto Grassad, Pierangelo Romanod
aDepartment of Geology, Facultad de Ciencias Físicas y Matemáticas, Universidad de Chile, 8370450 Santiago, Chile
bCentro de Excelencia en Geotermia de Los Andes, Chile
cEnel Green Power, Chile
dINGV Palermo, Italy
Dr. Philippe Robidoux
Department of Geology, Facultad de Ciencias Físicas y Matemáticas, Universidad de Chile, Plaza Ercilla 803, 8370450, Santiago, Chile
E-mail address: robidouxphilippe@gmail.com
Corresponding author.
A
Abstract
A
The geothermal heat extraction of binary type plant and its lifetime can be optimized through fluid monitoring of its gas and the natural system gas discharges. In order to present a novel gas monitoring tool for geothermal exploration and decision-making production, this research at Cerro Pabellón, Chile, since its internationally celebrated opening in 2017, demonstrates that the compositional characterization of gas and CO2-rich phases is beneficial for deepen understanding of the gas compositional array and associated physico-chemical processes produced from wells drilled into high-enthalpy geothermal systems elsewhere. The gas phase, studied through 5 years of direct gas condensate sampling and first-time application of a Principal Component Analysis (PCA) factor as Uniform Manifold Approximation and Projection (UMAP), showed two clear trends in gas mixture fluctuations: cooling as an Apacheta-like gas mixture hydrothermal reservoir with fluctuations between reducing-weakly oxidizing conditions. The equilibrium gas geothermometers obtained by studying the gas chromatography results of CO2-rich (> 96.0%) gases show a liquid-dominated system circuit-reservoir/in equilibrium with steam at a temperature of 225–350 ºC (Md = 252 ± 78°C) and oxidation variations (Md= -3.43 ± 0.64 RH), resulting from a complex cooling process that is distinct at the segments of the vapor separation lines (C, F, J, M) until reaching 140–160°C at the plant steam/brine separators. The monitoring data show that cooling of the vapour phase causes an increase in oxidation state (RH) and vapour fraction (%). When the condensate fraction increases during cooling (> 80%) and subsequent HCO32− dissolve in the condensate as liquids, this thermal effect can be detected by the enrichment of [Ar] content (ppm) and lower N2/Ar ratios in gas samples; such compounds in equilibrium with CO2-rich fluids should be monitored and controlled in geothermal plants to improve long term maintenance and binary type power plants sustainability by combining condensate liquid survey.A
1. Introduction
Cerro Pabellón, located at 21°50′56″S and 68°09′24″W, hosts the first geothermal power plants developed in South America, representing a binary type plant (Mine, 2016)(Fig. 1a-b). Cerro Pabellón is a blind geothermal field (Maza et al., 2018; Taussi et al., 2021), which means that the ground surface does not explicitly receive hydrothermal discharges. In order to maintain the production frequency since its establishment in 2017, the fluids from the geothermal power plants have been carefully extracted and reinjected, so as to maintain a balance between brine production, gas-steam and reinjection, while providing the required energy. To date, several sensors and measurements are required to accurately model the stability of the gases and fluids of the entire circuits (Battistelli et al., 1997; Cecioni et al., 2020). It remains a challenge at any geothermal plant worldwide to maintain a complete understanding of the gas influence on the performance of production flow and energy. At Cerro Pabellón, as other plants, the other challenge is to prevent further mineral precipitation or/and pipe corrosion (Seguel, 2023), as the latest expansion years of the plant (2017–2024) to review any changing thermodynamic (pressure/temperature parameters) (Cecioni et al., 2020) as physico-chemical conditions of the transported fluids (Arnórsson et al., 2007).
Fig. 1
– Location of study area.
Geographical location, b) Physiological map location, c) Geothermal well location
A
The crucial importance of the characterization of the gases in any geothermal plant is based on the fraction of CO2 that may extract heat from fractured rock with superior efficiency in comparison to water rate extraction, implying using gas-vapor CO2-rich phases becomes an advantage (Apps and Pruess, 2011). Geothermal heat extraction of binary type plant thus requires a deep understanding of the gas compositional array and related physico-chemical processes produced from wells drilled into these systems. The gas phase thus provides important information about various characteristics of the whole volcanic-geothermal context (Gudmundsson and Arnórsson, 2002; Stimac, 2015). This understanding is essential in order to define the best conditions for preserving fluxes and, in the long term, potentially extending the lifetime of any high enthalpy geothermal plant systems (Carroll and Stillman, 2014; Budisulistyo et al., 2017; Yuan et al., 2021).The main question here is therefore to determine how we can monitor fluids from geothermal power plants and its natural environment such as the binary cycle plant Cerro Pabellón to provide an international standard of validation of the stability of the gases? Since the binary plant conceives several production wells and steam as brine transport lines, and, considering that each production well is fed by specific site depth geological as hydrothermal context (Fig. 1c) (Vidal et al., 2021 and references therein), one should integer engineering characteristics of the site but understand the background conditions that favor hot fluids transport to the plant itself. This research is therefore aimed at understanding natural as industrial performance characteristics already on set of the best segments of the production lines and takes into account the natural hydrothermal manifestations that persist locally (Cerro Apacheta volcano; Urzua et al., 2002) as regionally (Tassi et al., 2010), as to understand the evolution of gas-fluid mixtures between geothermal reservoir and vapor transport lines (e.g. El Tatio field; Fig. 1b)).
2. Favorable geologic context for Cerro Pabellón heat-fluid source
The Cerro Pabellón field lies within a NW-SE graben structure (Pabelloncito Graben, Baker and Francis, 1978) that controls the NW-SE trending Pliocene stratovolcanoes. It is worth noting the tremendous magmatic control on the region, as below is the inferred presence of the Altiplano Puna Magma Body (APMB), the largest magma body on Earth (~ 500,000 km 3 volume). Several extrusive volcanic centers are found among the Pliocene stratovolcanoes of the Altiplano Puna Volcanic Complex (APVC; De Silva et al., 2006 ; Silva and Kay 2018)(Fig. 1b), which is one of the youngest (< 10 Ma) and most voluminous felsic volcanic provinces of the Central Volcanic Zone of the Andes (CVZA; Stern 2004). For example, stratovolcanoes of basaltic andesite to dacitic composition (Apacheta-Aguilucho Volcanic Complex; Ahumada & Mercado, 2009) and rhyolitic domes (Chac-Inca and Pabellón domes; Tierney et al., 2016; Prat et al., 2023).
Among the main sources of heat, tectonic and structural controls may have favored many influences (Tassi et al., 2010; Maza et al., 2018; Vidal et al., 2023). Regionally, the area experienced a local Pliocene extensional phase within a regional compressional regime related to the subduction of the Nazca Plate under the South American Plate (e.g., González et al., 2023). This phase generated a NW striking normal fault system that extends from the Azufre volcano in the NW to the Inacaliri volcano in the SE (Tibaldi et al., 2016; Godoy et al., 2019; Hübner, 2024).
3. Methodology
3.1 Classic Gas Survey Strategy
The method proposed here is to investigate gas transport conditions. The gas-condensate chemistry becomes key here to understand how the dominant gas species circulate in the geothermal circuits, from fluid extraction to reinjection into the reservoir, behave and change their composition since 2017 can reflect temperature, pressures the solubility/precipitation factors on the solid phases (oxiding-reducing control effect). Those parameters control mineral precipitation (Galarce, 2022; Seguel, 2023), considering the physico-chemical characteristics of each transport conduct (well, production lines; Fig. 2) that are possible to monitor with sensors.
Fig. 2
– Geothermal plant line parameters.
Geothermal plant segments, b) Physical parameters/per segment
To study the "hydrothermal gases” at saturated conditions, this study compare the fumarole sampling at a given pressure and temperature on the nearest open conduit discharge for hydrothermal fluids such as the Apacheta system (Fig. 1b) (Urzua et al., 2002; Tassi et al., 2010; Taussi et al., 2021). The hydrothermal fluids extracted far below the surface in the Pabelloncito grabbed may reflect distinct hydrothermal reservoir conditions, but such fluids are assumed to be the source of typical production as exploration well gases (H2O, CO2, H2S, HCl, CO, CH4; e.g. Giggenbach et al., 1988) and for this reason the sampling tools are inspired by classic instrumental as methodology to treat gas chemistry (Arnórsson et al., 2010). The gas from geothermal plant wells and enthalpy tap were sampled in double valve, pre-evacuated alkaline glass containers called Giggenbach bottles which are used to store gas samples. Well sampling is resumed by connecting the heat-resistant hose to the enthalpy tap of the well enthalpy tap (Fig. 2a; Sampling bifase fluid well). Vapor lines enthalpy tap is sampled after separator (Fig. 2a; Sampling vapor).
The liquid (brine) is also sampled after the Separator but using its own connection line (Fig. 2a; Sampling line). Gas samples were collected in Giggenbach bottles with a 4N NaOH solution, while liquid samples (brine and condensate) were collected in HDPE bottles, following the protocols established by Geotermica del Norte S.A., a group of Enel. All samples were analyzed at the CEGA Fluid Geochemistry Labs and Mass Spectrometry Lab in the Geology Department, University of Chile. Condensable gases (CO2 and H2S) were dissolved in the 4N NaOH solution and later analyzed using Neutralization Potentiometric Titration and Iodometric Titration, respectively (Giggenbach and Goguel, 1989). Non-condensable gases including He, H2, Ar, O2, N2, CH4, and CO were analyzed by Gas Chromatography using an Agilent Technologies 7890A Gas Chromatograph with a customized Wasson-ECE Instrumentation sample inlet system, auxiliary oven, two Thermal Conductivity Detectors (TCD) and one Flame Ionization Detector (FID) (Tassi et al., 2010).
This setup ensures precise identification and quantification of each gas component. Liquid samples (brine or condensate) were analyzed for both cations and anions. Major cations (Na, K, Ca, Mg, Si) were measured by Flame Atomic Absorption Spectrometry using a Perkin Elmer PinAAcle 900F, with an individual hollow cathode lamp for each analyte. Other cations, including Li, B, Al, Fe, As, Rb, Sr, Sb, and Cs, were quantified by Inductively Coupled Plasma Mass Spectrometry (ICP-MS) using a Thermo Scientific iCAP Q. Anions (F−, Cl−, SO4=, Br−, I−) were determined by Ion Chromatography with a Thermo Scientific Dionex ICS-2100 system equipped with a conductivity detector. Alkalinity was determined via Neutralization Potentiometric Titration using an automatic Hanna Titrator HI902C and a combined glass electrode. Ammonium content was analyzed using the Nessler reaction, a method based on molecular absorption in the visible range.
3.2 New Adapted-Separation Line Gas Approach
The gas-mixture is sampled and gas contents used to model thermodynamic conditions the closest to gas mixture flow extracted as gas from the heating source Apacheta. As a first step here to identify geochemical relationship with source, classification diagrams are proposed based on literature to discriminate gas mixture groups and thermal conditions (Giggenbach et al., 1996). The results are discussed to evaluate potential transport conditions that may affect chemical element contents in gas as liquid fases which could be involved in mineral precipitation (pre-heater segment; Fig. 2a).
The physico-chemical parameters are measured by control room during sampling and merged into databases to perform multi-component statistical approach by section line (C, F, J, M) and wells (CP5 and CP6). The objective of associating physico-chemical conditions to the identified gas-liquid-solid phases chemistry (Supplementary material Table 1) is possible through reviewing a control room database of sensors installed in the plant, indicating vapour/liquid fraction (steam fraction), gas/steam ratio (weight %), pressure (kb), temperature (°C), all obtained at the time of sampling (Fig. 2a-b), which is carried out for gas and liquid analysis. Access control room survey parameters collected during field surveys and dataset are stored by the ENEL company (Cecioni et al., 2020).
3.3 Statistic approach
Cluster analysis is a powerful tool in geothermal exploration for discovering subgroups of gas species (Petrillo et al., 2023) or solute water major element chemistry/dissolved gases (Golla, 2018; Daniele et al., 2020). At the stage of reviewing chemistry of gas-condensate and physico-parameters, a statistical approach is thus proposed here to decipher groups of variables (clusters) that behave the same between the systems of the geothermal plant and to determine the chemical characteristics of natural (Apacheta, modelled hydrothermal reservoir) and industrial system (Cerro Pabellón). In order to reduce the large number of variables accumulated in recent years in monitoring database, the PCA-based method here simplifies understanding of data correlation. In the first case, variables yet unknown and assumed correlations will be presented as the “data-driven” statistic approach (Barnett, 2017).
The classified database DB in this study contains analyses from gases (n = 23 variables/n = 90 data), as liquid condensates (n = 16 variables/n = 22 data). As rule of thumbs for PCA and multivariable statistics, databases are built as to maintain 1:1 ratios data/variables. The Kaiser’s criterion was used to determine the factors using eigenvalues higher than 1 (Barnett, 2017; Golla, 2018). KMO tests are then calculated to quantify the significance of the factor analysis (FA). The components are rotated using the varimax method, to maximize the variance of the squared loading for each factor (Barnett, 2017). Finally, the weight of the variables of each factor is called “loading”, and each factor (Principal Components, PC) is associated with at least one variable (e.g. Yellowstone natural geothermal system; Golla, 2018). Loading values > 0.5 define the representative variables for the factors using the IBM SPSS Statistics Software V28.0.1.
Since several types of relationships are ubiquitous in geochemical DBs because several processes are in “competition”, we propose the Uniform Manifold Approximation and Projection (UMAP) (Bonham-Carter and Grunsky, 2018; McInnes et al., 2018) approach to discuss the observed statistical PCA tendencies in vapor separation line systems, such method attribute statistical non-linear correlations between confounding variables (non-linear relationships such as logarithmic, hyperbolic, quadratic, exponential, polynomial, etc.; McInnes et al., 2018). The UMAP cluster treatment used the umap-learn library for Python (The python codes including parameters for UMAP). The aim of this approach is to reduce the data (Bonham-Carter and Grunsky, 2018), which here are typically linked in gas geochemistry by logarithmic geothermometers for assessing equilibrium (Giggenbach, 1980, 1987), and thus identify the most strongly related variables. Within this category of mathematical relationships, for example, there are known geochemical processes that are controlled by redox factors (RH = Log [H2/H2O]; Log[CH4/CO2]), among other gas tracers (H2/Ar; N2/Ar, CO2/CO, etc.) (Chiodini and Marini, 1998) which are normally non-linear in binary diagrams (Cumming and Powell, 2000).
We therefore propose PCA (Ventura et al., 2011; Tong et al., 2013; Barnett, 2017) to observe group of linear statistical tendencies, then UMAP for non-linear relationships (McInnes et al., 2018), to evaluate the grouping process and demonstrate the best geothermo-chemical gas tracers. Since gas composition systems behave directly as a function of P-T-V, our method evaluates the optimized transport conditions on potential heat transfer which depends on temperature (°C or °K), flux (t/h) and vapor fraction (%), thus assuming those are the variables that alter geothermal compositional gas chemistry. The compositional gas-mixture variation can be explained with such leading variables to distinguish between the effects of cooling rates and oxido-reduction (RH) of each separation lines (here C, F, J, M), if a clear pattern is observed between well extraction to vapor separation lines of the geothermal plant.
4. Results
4.1 Gas chemistry
A
The chemical composition of hydrothermal field gases at discharge sites is reported in Supplementary Material Table 1 (Gas Database). For the Cerro Pabellón wells, contaminated samples (< 71 mol% CO2) were quickly discarded from graphical/statistical analysis such as the three following sample code ID (#52, 53 from well CP6 on day 07-09-2024; #79, 84 from Duplicate F). Samples ID #31 and 57 are considered for analysis but labelled as outliers despite being lower in CO2 contents (86–88 mol%). Samples in series CP reach the interval of 5.67–0.81 mol% N2 (12.5% outlier in Línea F/CP-5 + 5A on day 22-03-2022), while the well samples of CODELCO exploration campaigns have only 0.2–2.2% (Urzua et al., 2002) (Fig. 3a). For comparison with the well as vapor separation line sampling, the fumaroles observed from literature in the Apacheta sector are dominated by CO2, which comprises > 96.0% of the total gas in each sample using acidimetric titration (Fig. 3b; Table 1) (96.2–99.8 mol% in literature; Urzua et al., 2002; Tassi et al., 2010; Taussi et al., 2019). Overall, the CO2 content at Cerro Pabellón transport lines vary between 92 and 99,1 mol% (SD = 1.41%). In contrast, 0.5–113 ppm carbon monoxide (CO) but relatively low methane is detected (CH4 = 9.5–36.6 ppm), except outliers’ sample ID #61–62 (227–295 ppm). Methane concentration reaches 4–3970 ppm at regional values (4–9.6 ppm at Apacheta) in measurements of Tassi et al. (2010), but do not exceed 10 ppm, limited by reported precision in Urzua et al. (2002)(well or neither fumarole dataset). The H2S was measured by iodometric volumetric titration, (dissolved in soda) giving values of 0.003–2.02 mol% for this study, 0.01–0.61 mol% for Urzua et al. (2002) wells and regional values reach ~ 500 ppm, and extreme contents at Apacheta (~ 8000 ppm) according to Tassi et al. (2010).Fig. 3
– Major gas compositions.
Ar–N2–He, b), He–CO2–H2
The oxygen (O2) measure between not detected (n.d.) and 1.96 mol% while argon (Ar) 0,001–0,078 mol% at the geothermal plant. The majority of the Apacheta fumarole samples have oxygen O2 contents lower than 0.1% (< 0.92% in wells). All fumaroles-wells sampled have much higher N2/O2 values than that of the air ratio of 3.73 (minimum of 6.76), suggesting only weak atmospheric contamination (except samples ID # 31, 51, 52, 53, 57, 84). Published helium isotopic data by Urzua et al. (2002) (Rc/Ra = 1.85) as Tassi et al. (2010) (Rc/Ra = 1.81) showed that the gases were not affected by atmospheric/cortical helium in nearest hydrothermal system. Our samples from Cerro Pabellón in this study showed traces of He (> 9–23 ppm), being 6.2–8.4 ppm at Apacheta. The ratio N2/Ar, used to discriminate magmatic, air and meteoric-dominant mixture shows clear tendency in Fig. 3 for well literature data CP6 (ID#54) and Apacheta (> 363) being farther from air contamination typical value (< 84).
4.2 Condensate chemistry
A
The chemical compositions of the liquid condensate samples produced by the steam collected at Cerro Pabellón are listed in Supplementary Material, Table 2 (Condensate Database). Cation-anion reported by thesis work (CEGA/U. Chile) between years 2021–2022, classified plant condensates and brines as chlorinated sodium waters, Total dissolved solids (TDS) = 14–17 g/L, with steam fraction reaching 24%, steam ratio: 1% CNG − 99% water vapor, could vary by factor 100% (Giudetti and Tempesti, 2021; Galarce, 2022). The physical parameters of the monitored sensor systems, recorded since the gas monitoring sessions (10-02-2019 to 02-08-2024), show a variation of the separation pressures (before the separator) of 4.2–6.5 bars, resulting in a variable vapor fraction when the vapor is separated in the separation vapor lines (44.5–176.2 t/h), where the fluxes are higher in the brines (65–690 t/h). Each circuit is discussed in detail in sections 5.2–5.3.The company dataset report fluids that are dilute (< 37 mg/L Cl), except two contaminated, CP-5A, Production well (27-06-2016) (> 37 mg/L Cl), recording extreme conducing conditions (> 170 µS/cm)/high TDS > 900 mg/L. The variations in the rest of samples (0.026–6.28; SD 1.72 mg/L) is above typical natural systems recharged by meteoric water in Apacheta sector (Tassi et al., 2010). Of the 3 wells sampled, HCO32−, Na, Ca, Cl and SO4 are the dominant components in the liquid phase with concentrations in the range of 15.9–45.1 mg/L, 0.19–15.9 mg/L, 0.1–1.87 mg/L, 0.25–6.72 mg/L, and 0,55–4.54 mg/L, respectively. In general, all the samples have relatively low TDS (< 250 mg/L), but dominated by waters of the bicarbonate and sulfate type.
The relative proportions of Cl and B, both considered conservative elements on the C-SVZ (Benavente et al., 2016; Tassi et al., 2010), are useful to trace the source of the thermal fluids, but boron could not be measured. Due to its relative immunity to secondary processes, Li was chosen here as a reference constituent (ibid. Tassi et al., 2010) to highlight the variations in B/Cl ratios and/or semivolatile behavior; its values are between 0.001–0.19 mg/L. The content of anions is sensitive to pH 4.597–5.762, measured at 18.95–25.4°C temperature on sites and quickly condensates in the sampling circuit (from ~ 150–160°C).
4.3 Gas and condensate chemistry: Multivariate statistical analysis
Multivariate statistical analyses within gas and condensate geochemical data were used to identify internal patterns between the separation lines and the parameters controlling their composition (e.g., d'Amore et al., 1987; Golla 2018; Petrillo et al., 2023). To verify the validity of the method, the direct relationship between variables was tested with correlation matrix, using Pearson coefficient (El Chichón Crater; Cuoco et al., 2013; Campi Flegrei, Ebrahimi et al., 2023), taking into account the correlation index (Supplementary material Tables 3 to 6), show relatively high positive (> 0.5; yellow cells) or negative (<-0.5; orange cells) correlations.
The number of intercorrelated variables in the gas table (23 variables) processed to SPSS KMO test (Supplementary material Table 3), here where n = 3, which correlate between gas contents (O2, He, CH4) and physico-chemical monitoring parameters (brine flux and vapor fraction). Vapor fraction (%) is highly correlated with RH and anticorrelated with equilibrium temperature as N2/O2 ratio. At least 18 variables are positively correlated between gas and gas ratios, but 18 variables are anticorrelated.
For the condensate table (Supplementary material Table 4) (16 variables), of which n = 13 correlate with in-situ physico-chemical parameters (81% of them), while 80% of the variables are detected in the analysed samples (otherwise n.d.). Sulfate (SO4), iron (Fe) and Aluminium (Al) anticorrelate or show not significant control on other variables, while temperatures (°C) anticorrelate with other physico-chemical parameters. The rest of the anion-cation contents and physico-chemical parameters display positive correlation (n = 66 cases).
The Kaiser rule overlays yielded multiple principal components (PCs), but often result in an unnecessary (excessive) number of retained variables in applied statistics (e.g., n > 8 in Yellowstone geothermal water chemistry; Golla, 2018). Here, the PCA method was also tested with a processed Pearson coefficient matrix, but to avoid complexity in cluster interpretation, we retained components representing > = 5% of the explained original variables. The gas and condensate databases were each evaluated to generate a Pearson matrix table based on mean standardisation with natural log transformation to improve detection of sample group anomalies (Supplementary material Table 5, 6).
The results indicate that physicochemical parameters measured in situ are closely related to chemical variations of anion-cation contents of condensates formed in gas separation lines, while the gas contents themselves have a complex network of correlations. Gas samples represent four main PC components (other PCs do not have correlation index > or < 0.5) (Supplementary material Table 5); PC1 contains RH anticorrelation with CO2 dominant and dominant oxidising gases (H2S, N2, O2), PC2 contains vapour fraction dominant/equilibrium temperature anticorrelation (see calculation in section 5.1), PC3 contains RH dominant (reducing gas H2, CO), PC4 contains equilibrium temperature dominant (as CO content or CO/CO2 used in thermometry model). Condensate samples represent three main PC components (Supplementary Material Table 6); PC1 contains anion-cation contents and physico-chemical positive correlations, but negative on Al-Fe pair; PC2 contains anticorrelation between Fe and SO4; PC3 contains Al-SO4-Fe pair positively correlated.
5. Discussion
5.1 Dominant fluid chemistry species/per separation Lines
The separation vapour line conditions model shows distinct thermal than oxido-reduction pathways that diverge from the natural Apacheta fumarole system, according to review of compositionally plotted fumarole gas classification datasets and their modelled thermal range values. N2-CO2-Ar and N2-He-Ar ternary diagram of well and fumarole gas chemistry (Fig. 3) showed the similar tendency in the C-F-J separation line data cluster to CO2-He enrichment/depletion, controlled by variations in forced cooling processes. Cooling between reservoir fluid extraction > 350°C (Urzua et al., 2002), wells to separators, as to induce controlled secondary processes of secondary boiling and steam condensation (e.g. Chiodini and Marini, 1998) within the transported lines here are identified as the main causes of gas chemical composition variations.
The internal vapour cooling processes (boiling and vapour condensation) can cause large chemical changes when the heated liquid phase is boiled, which can be critical in inducing variations in the major gas species (Chiodini and Marini, 1998). For example, in this work the results show geothermal vapour compositions representing crustal to meteoric sources when compared to the natural system of the Ollague and El Tatio region (Figs. 3, 4). This near-linear correlation (non-linear for binary compounds) is observed at constant relative N2 contents, despite two identified outliers (as Urzua et al., 2002 dataset) and El Tatio series (ET; Tassi et al., 2010), for which N2 contents are below ~ 1.0% (or < ~ 1,000 µmol/mol referring to the authors' units used). Some scatters of high N2 contents in relative N2/Ar ratios (atmospheric values of 84), in n = 10 samples, indicate that 60% N2 prevails over Ar and that > 40% N2 prevails over helium in N2-He-Ar diagrams, for which atmospheric compounds (significant N2 excess) affect samples due to conductive contacts with atmospheric air. Mixing during enthalpy tap connection/sampling processes can be detected when O2 contents increase at low CO2-contents (ID #31, 52, 53, 57, 79, 84; Supplementary material Table 1). If such processes have occurred, the values are still valid as a reference to check trends in vapor lines to detect any significant mixed atmospheric compounds in oxidation-reduction processes.
Fig. 4
– Binary Gas Geothermometer.
See Fig. 3 for legend symbols.
Binary diagram of log(H2 /Ar*) vs. log(CH4/ CO2). Red line tilted box represents zoomed view in figure (b)
Binary diagram log(H2 /Ar*) vs. log(CH4/ CO2) with snapshot from figure (a). Green polygon represents field for natural Apacheta (Urzua et al., 2002; Tassi et al., 2010) and well samples (this study; Urzua et al., 2002). Orange arrows represent two cluster tendencies from separation line F while blue arrows represent two cluster tendencies from separation J. Black tilted arrow represents a single separation line C tendency.
Binary diagram of log(H2 /Ar*) vs. log(CO2/ Ar). The diagram is fixed at RH = -3.5.
Binary diagram of log(CH4/ CO2) vs. log(CO/ CO2))
If the same samples (N2/Ar > 84) preserve thermal/unmixed gases between wells-separators, we could infer that natural processes occurred in the hydrothermal reservoir below Cerro Pabellón, similar to open-field fumaroles; Crustal magmatic-meteoric mixing with magmatic dominance with its tectonic signature (azote-rich sediments overlying the subducted slab; Snyder, 2003), Puchulzida-Tuja and Apacheta have the strongest N2/Ar ratios (> 41 to 442) and also the highest Rc/Ra ratios in the whole region (Fig. 3a), implying that hydrothermal gases are not fully "reset" to atmospheric end-member mixing. The same analogy is used in the N2-CO2-Ar ternary diagram (Fig. 3b), also used to discriminate meteoric mixing in sampled gases and air-saturated water (ASW; Giggenbach, 1993). The result of this comparison still shows high concentrations of Ar and He typical for crustal radiogenic decay in Apacheta and linear-like trends in Ar-to-He or Ar-to-CO2 variations that are similar between lines C, F and J (less visible at Line M that contains less data). Assuming that Ar is mainly derived from ASW, variations of Ar concentrations may depend on the degree of gas-water interaction at shallow depth, which in turn depends on the thickness of the local aquifers (Tassi et al., 2010).
The consistently high detection of CO2 and CO is relatively high compared to natural systems, here inferred because the thermal and closed transport line systems do not involve dissolution in a dominant water reservoir and that CO2 gas compounds were already separated in production wells during flash points. For example, carbon monoxide is below the detection limit (< 1%; <1,000 µmol/mol; e.g. Urzua et al., 2002), possibly because this gas readily dissolves in shallow aquifers to form HCOOH (Shock, 1994), which is the case in the regional bubbling pools and fumaroles sampled in Tassi et al. (2010). In the case of Apacheta and the geothermal plants, the carbon species are sufficiently abundant and preserved in our dataset, but most importantly, they represent a vapor produced by the boiling of a liquid phase, in this case ascending hydrothermal fluids in production wells.
5.2 Geothermal vapor fluid chemistry tracking modified hydrothermal reservoir cooling and oxidation variations
The change in hydrothermal reservoir condition inferred from well fluid chemistry is determined using specific gas geothermometers that account for second boiling of the hydrothermal liquid phase after well extraction (Chiodini and Marini, 1998). The sampled vapor is considered to be a product of the vapor produced by the boiling of a liquid phase. Here is detailed the modeled process used to estimate the Cerro Pabellón reservoir temperature.
As proposed for the Cerro Pabellón (CP) gases to justify the RH values of the CO2-CH4-H2 equilibria and the presence of acidic compounds in the Apacheta/or El Tatio region (Tassi et al., 2010), the highly oxidizing conditions seem to affect the root of the hydrothermal reservoirs at the Apacheta and extraction well CP systems (Fig. 4a). Nevertheless, important shifts of the apparent CO2-CH4-H2 equilibria towards higher temperatures (> 250°C) and more reducing conditions (RH < -3.6) are observed for F as J lines, split towards a single thermal variation (temperature control H2/Ar; Fig. 4b) at isoline C, instead, line F shows two tendencies A) towards lower temperature ranges below 250°C (low H2/Ar) and reducing conditions (high negative RH), but also B) towards lower temperature (low H2/Ar) and constant modelled curves of RH = -3.8 (following the oxidoreduction isoline).
The CO2/Ar - H2/Ar (CAR-HAR) gas geothermometer grid (Fig. 4c) with log (H2 fugacity/H2O fugacity) = -3.5 (the redox state, expressed as RH; Giggenbach, 1987) shows that the vapor lines accumulate gases from a vapor phase in equilibrium with a liquid reservoir at a temperature of 225–300 ºC (~ 325 ºC for Apacheta; Urzua et al., 2002). This high temperature may reflect a significant magmatic component to the geothermal system, but outside the dataset, outliers in the equilibrium pattern are found at 200–225°C, which may be consistent with the propylitic alteration zone characterized by a chlorite-epidote-illite facies ( ~ > 200°C; Maza et al., 2018).
As a first approximation for an industrial system in the context of the Ollague-El Tatio region typical natural conditions for gas chemistry, we present temperature reservoir data measured in CO2-CH4-H2 system, and that aquifer interaction is interpreted previously in the Cerro Pabellón field (Shock, 1993).
The water content was not directly measured by laboratory technique in this study, therefore the systematic measurements of vapour fraction (%), modelling geothermal plant separator, Fig. 2a (vapor flux/ [vapor flux + brine flux]; in tons/hour), as the main indicator of the weighted % fraction gas/liquid. Field separator in this study also helped to approximate the gas/steam weight ratio and derived vapour fraction (%) of gas space/above liquid space (G/S in Supplementary material Table 1). The estimated liquid mol-free basis [H2O] content was considered occupying the rest of the separation vapor line at the time of gas measurements as to recalculate gas-only phase molar contents for the rest of the gas species (H2, among others). Giggenbach's (1987) most dominant parameter was used to describe the redox potentials of the fluids sampled in this study as follows:
(1)
RH = log rH = log (fH2 /fH2O ) = log (XH2 /XH2O )
Assuming the gas volume relative to condensate water volume is constant in each sampling process, the calculated RH gas composition indicates values between − 5.39 and − 1.64 (md= -3.43 ± 0.65) (see Supplementary material Table 1). Therefore, to explore the relationship of oxido-reductores and indicatores of RH with temperature in modelled steam produced by liquid boiling with equilibrium constrains of the CH4-CO2 pair, we estimate Giggenbach, 1987 equation in CO2-CH4-H2 system:
(2)
CH4 + 2H2O = CO2 + 4H2
The estimated temperatures at RH in steam produced by boiling of a liquid phase is now considering Giggenbach, 1987 equation:
(3)
log(XCH4/XCO2)v = 4RH + 5181/T(K)
Assuming that the gases equilibrate as vapor phase "v" in a liquid-dominated reservoir (database filtered to accept only 0% steam/80% of liquid; GDN, 2010), the plotted gas compositions indicate a temperature between 35° and 406°C (md = 226 ± 70°C).
As a second thermometer for method comparison, water fugacity is fixed (Solfatara; Chiodini et al., 2010) and used in the Chiodini et al. (2010) calculation. In this case, water vapour is the main transported phase in the boundary layer and the geothermometer uses instead the CO/CO2-CH4/CO2 system. The RH is modelled but not calculated for each sample as in the previous attempt. The temperature in this system can be expressed by Eq. 6 of Chiodini et al. (2010) as a function of the ratios of the mole fractions in the vapour phase:
(4) log(XH2O /XH2 ) + log(XCO /XCO2 ) = L1 = 2.485–2248/TK
Our estimation of water fugacity, based on fH2O is Md = -1.76 ± 0.32 (-1.76-2.20 range). Then our estimation of temperature, based on variable fH2O (Giggenbach, 1980) is thus Md = 252 ± 78°C (112–440°C range); the temperature dependence of the equilibrium constants of Equations 6 from Chiodini et al. (2010) is thus relevant in this study as it approximate the ranges of 100–374°C for typical hydrothermal systems and fit the previous 250–300 ºC from the CAR-HAR model (Powell and Cumming, 2000). Although several outliers represent gas composition anomalies that overestimate thermal conditions when considering sample values with estimated < 80% liquid water. Modeled CP gas transport context is thus close to a naturally constrained liquid-dominated reservoir (GDN, 2010). Otherwise, with a fixed water fugacity (equation no. 6–7 in Chiodini et al., 2010), the estimated temperatures are outside the accepted literature range (Md = 477 ± 37°C) (Chiodini and Marini, 1998). In addition, some uncertainties on the P-T system factor may interfere, as the vapour transport line represents much lower pressure conditions measured in the geothermal plant system, 4–6 bars, compared to the classic geothermometer application of 1-220 bars, which may represent pressurized systems at higher depths (> 2–3 km).
The inverse path for modeling the hydrothermal reservoir temperature of the extracted fluids below CP may be close to the fluid composition circulating in the plant (Cecioni and Cei, 2020), but strong atmospheric mixing and hydrothermal-meteoric as industrial factors may alter its composition due to oxidizing variations. To test this effect, the RH dependence of the CO/CO2-CH4/CO2 geothermometer grid makes this graph (Fig. 4d) useful for assessing the redox state of the CP and Apacheta systems before using fixed RH data (assuming minimal produced water condensates). From this plot, an RH of -3.5 is obtained to represent the boundary line data sets, suggesting a slightly oxidizing reservoir environment; this is still close to our field method estimate (md= -3.54 ± 0.35 RH). Because this grid pairs compare a fast geothermometer (CO/CO2) with a slow one (CH4/CO2), the predicted temperatures may differ from the actual temperatures, since temperature is mostly determined by CO/CO2 (near horizontal) and RH by CH4/CO2 (near vertical) (Powell and Cumming, 2000). Assuming the gases equilibrate in a liquid-dominated reservoir, the plotted gas compositions indicate a temperature between 300° and 350°C, with outliers (> 10%) found at > 350°C; such values and identified gas species (H2S-rich and non-detected HCl, SO2) agree with scrubbed gas at depths and heat source persistency (Gorini et al., 2018; Taussi et al., 2019; Prat et al., 2023).
The line sensor temperatures measured in the field (150–170°C) and 88°C in the near-surface exploration well are still lower than classical geothermometers (Giggenbach, 1987), which instead represent drilled geothermal reservoir heat (Urzua et al., 2002; Taussi et al., 2019). These measured temperatures at the seeps or wells are not predicted by thermal equilibrium models but reflect the result of a complex cooling process for gas mixtures. For instance, if the temperatures reflect the observed values (> 170°C), the consequence would be to record H2/Ar and CO2/Ar lower (richer in Ar). Within this scenario the ratio of CO/CO2 would be lower, but CH4/CO2 would be higher (richer in methane and lower in CO).
The modeled reservoir temperatures, based on well (after flashing point) and separation (after vapor-liquid separation) lines (Fig. 2), are thus artifacts of the original fluid extracted from the hydrothermal reservoir. The dataset in Figs. 3 and 4 thus represents the natural tendency of a cooling and oxidizing gas system controlled by artificial processes. None of the regional gas mixtures in the literature match the original tendency of the Apacheta compositional cluster (Urzua et al., 2002; tassi et al., 2010), therefore we argue that the data clusters observed in this study are authentic tracers of the Apacheta-CP system, for which a structured review of physico-chemical plant parameters is necessary to explain the gas compositional variation (Section 5.3).
5.3 Dominant variables to improve heat maintenance production: evidence from gas tracers and gas/liquid fractions
The collection of non-condensable gas (NCG) content measured at the Cerro Pabellón geothermal plant provides direct insight into thermal parameters (modelled temperature) and can be used as a symptom of oxido-reduction changes at the plant separator (Fig. 2a).
Visualising our new dataset with a multivariate approach shows what quantitative variations exist in geochemical datsets, using a novel graphical tool (UMAP) for geothermal exploration/production disciplines. This is because the UMAP results are instead presented as a three-component axis with weighted values to simplify visualization of statistical changes (Bonham-Carter and Grunsky, 2018; McInnes et al., 2018). UMAP groups the samples into two large main clusters (Figs. 5, 6), reinforcing the results of the binary plot (Fig. 4). The gas compounds are mainly grouped on the UMAP 1 line, strongly for the J-M-F line, and more strongly for the UMAP 2 line for C and F (Fig. 6a). The C line contains intermediate values between the two clusters but is clearly distinct in the data dispersion. To highlight the origin of the clusters, the quartile color code was applied to both physico-chemical (RH; vapour fraction, modelled temperature) and gas composition parameters in Fig. 6, which were clearly identified by statistical correlation/anticorrelation behaviour in the correlation/Pearson matrix (Supplementary Material Tables 3, 5).
Fig. 5
– Component Factor.
Fig. 6
– Uniform Manifold Approximation and Projection (UMAP).
The most relevant and distinct correlations are seen for the RH (Fig. 6b) as vapor fraction (%) (Fig. 6c) at each vapor line, while the gas signature mainly controlled by [Ar] is best suited to highlight this parameter. The third cluster group (intermediate C) is the vapor fraction that represents the lowest values (%), but points to the coldest equilibrium temperatures (Fig. 6d) as its own thermal gas ratio indicator (lowest N2/Ar; <250°C; Fig. 6e). Line J also shows clusters with lower vapor fractions (%), which may favor condensation in the transport lines (Fig. 7). On the other hand, the gas mixture of line F reflects the highest vapor fraction (%) and a low intermediate cluster of N2/Ar.
Fig. 7
–Chemistry model/Binary Plant Cross section.
This may be explained later by the fact that a larger proportion of the condensate volume may be formed during gas transport and separation. Under these circumstances, the most favorable conditions for these gas mixtures may be for gas volatilization/evaporation down the stack of the plant rather than increased liquid condensate formation (Fig. 7). It is important to estimate the proportion of non-condensable gases (NCG) at the separation stage, as they can alter the production capacity (enthalpy KJ; Li et al., 2022), which instead represents a positive result for the production performance (Hosgor et al., 2016), but only if the boiling of the reservoir is controlled during the drop in the flash point pressure of the water between the wellhead and the separator.
The condensate dataset, being smaller, provides less point density, therefore not suitable for visual PCA as UNAM analyst for samples (Supplementary material Table 2). Nevertheless, interesting clusters are observed by analyzing the variable correlation (Supplementary Tables 2,4,6); it is clear, according to the Pearson matrix (Supplementary Material Table 6), that [Fe] decreases with SO4 represent solid iron precipitates or that aluminum [Al] increases in the liquid favor aluminosilicate dissolution, while the rest of elements dissolve during cooling. Overall, diluted contents of [SO4] (positive PC2 in PCA) and [Fe] (negative PC in PCA), may both explain the formation of sulphate in the condensate by reducing H2S in the vapor and favoring iron sulphate precipitates (Taussi et al., 2010). The carbonate precipitates instead favor HCO32− when the liquid condensate is saturated with this anion (high HCO32−/SO4 ratio reflecting shallow heated aquifer buffer on heated scrubbed gases). Such mineral precipitates have been so far observed in paragenesis of solid scaling mineralogy reported by XRD and SEM results on mineral chemistry (Galarce, 2022; Seguel, 2023).
Overall, the characterization of chemical changes between the gas and liquid phases should be further investigated, and some separators in geothermal plants are more efficient than others in maintaining thermal conditions. For instead at Cerro Pabellón, at neutral pH in line F, carbonic acid HCO32− is less concentrated in condensate solutions (Fig. 6). The condensate chemistry measured by ion chromatography indicates that a neutral pH favors a higher proportion of CO2 dissolved as HCO32− (mg/L), on the other hand, the observed increasing CO2 content in the gas phase occurs when heating conditions are maintained for most oxidized gas phase conditions (RH<-3.4); these conditions should therefore be maintained and monitored by plant operators. This effect is in fact directly related to the cooling of the system (T-temperature variable negatively correlated with the other variables in the Pearson matrix correlation; Supplementary Material Tables 3 to 6). Under these circumstances, it is assumed that the degree of cooling was monitored by recording a greater temperature drop in the C + J lines compared to the F and M lines between 2017–2024 at Cerro Pabellón, so that a greater volume of condensate could be formed under these circumstances. We expect that this gas-liquid phase transition relationship will eventually affect the chemistry of the condensate and the nature of the separated brine, increasing dissolved reactive chemical compounds such as HCO32− when cooling and oxidation interfere with non-condensable gas mixtures.
5. Conclusions
A first gas-fluid novel standard approach is presented based on gas, condensate of a geothermal plant and compared to natural hydrothermal reservoir. The main conclusions are the following:
(1)
The CO2-dominant gas mixture sampled in the wells and steam-separated lines of the Cerro Pabellón geothermal plant has a composition controlled by secondary processes of secondary boiling and steam condensation. Their natural composition, close to the Apacheta fumarole, far from the Ollague-El Tatio fumarole fields, diverge in composition due to the degree of gas-water interaction at shallow depth and industrial processes that force cooling of the gases.
(2)
Geothermal steam chemistry represents a modified Apacheta hydrothermal reservoir (325 ºC for Apacheta; Urzua et al, 2002) cooling of a liquid dominant system circuit-reservoir/in equilibrium with vapor at a temperature of 250–350 ºC (Md = 252 ± 78°C) and oxidation variations (RH md= -3.43 ± 0.65), resulting from a complex cooling process that behaves distinctly at the segments of separation vapor lines (C, F, J, M) until reaching 140–160°C at plant vapor/brine separators.
(3)
Vapor phase cooling in the separation line individually affects the gas vapor fraction (%) and thus the condensate fraction is more important when this occurs, inducing a lowering of the N2/Ar ratio to indicate that such effects occur in geothermal plants. According to UMAP and PCA statistical analysis, condensate chemistry will favor neutral pH and dissolve a higher proportion of CO2 as HCO32- in the condensate when cooling occurs. Consequently, the monitoring of gas ratio N2/Ar among other ratio, is a strong tool to indicate if heat and optimized reducing conditions (RH) are maintained or fluctuate in geothermal plant circuits.
A
Acknowledgement
This work is part of internal CEGA and ENEL fundings. The measurements were also supported by Veronica Rodriguez (CEGA), Erika Rojas (CEGA) and years’ experience Santiago Maza (CEGA), Maximiliano Seguel for sampling and assisting until 2024s field campaign. Laboratory analysis of duplicates were obtained by Marianno Tantillo, Manfredi Longo and Fausto Grassa (INGV-Palermo) through participation at the TransNational Access (TNA) in the framework of the EXCITE (Horizon 2020) research infrastructure. This publication results from work carried out under transnational access / national open access action conducted at Noble Gas Laboratory under the support of WP3 ILGE - MEET project, PNRR - EU Next Generation Europe program, MUR grant number D53C22001400005. The financed project under transnational access is registered with ID # C1_014 and full title "Identification of magmatic gas source and geochemical connection from high-enthalpy geothermal gas system to its magmatic source, Andean Cordillera".
CAPTIONS
Blue tilted arrows represent cooling tendency anticorrelated to high temperature tendency with red tilted arrows. Green tilted arrows represent additional tendencies (see Pearson Matrix table in Supplementary material Tables 5 and 6).
a)
Factor component analysis: Gas variables with 3 principal components (PC1-2-3). Abbreviations of variables are listed as follows: Vapor flux (Qvap), Brine flux (Qsal), Separation pressure (Psep), temperature (°T), Gas vapor fraction (GV). The rest of gas contents are marked with parentheses “[…]”, except when variables are presented as gas ratios. b) Factor component analysis: Condensate variables with 2 principal components (PC1-2). Abbreviation is used for variable of Temperature (°T), the rest of elements are concentrations. See Supplementary material Table 6 for full details on vocabulary and units for each variable.
Analyst plotted with 3 principal components (UMAP1-2-3). Color codes represent the four quartile categories (Q25 − 50−75−>75) (see original dataset in Supplementary Table 1).
a) UMAP: Gas Line-Wells, using single symbols per separation Lines (losanges) and wells (squares) and colored polygons are equivalent illustrated in subsequent figure, b) UMAP: RH (no filter on liquid/gas fraction applied as in section 5.2 for calculations), c) Gas Vapor fraction (%), d) UMAP: Gas [Ar] content (mol%), red polygon illustrate samples that recorded third quartile (Q75) values from vapor fraction (%), e) Gas temperature modelled according to Eq. 3 (section 5.2; Eq. 6 from Chiodini et al., 2010)(see column TK6 in Supplementary material Table 1), UMAP
(N2/Ar), red polygon illustrates samples that recorded second quartile (Q50) values from modelled gas temperatures.
Supplementary Material
Table 1 – Gas Database
Table 2 – Condensate Database
A
Table 3 – Matrix Correlation Gas DatabaseA
Table 4 – Matrix Correlation Condensate DatabaseA
Table 5 – Pearson Matrix Correlation Gas DatabaseA
Table 6 – Pearson Matrix Correlation Condensate DatabaseElectronic Supplementary Material
Below is the link to the electronic supplementary material
A
A
Author Contribution
P.R. wrote the main manuscript text and M.S., S.M. prepared figures 1-2. All authors reviewed the manuscript.
References
Apps, J., & Pruess, K. (2011, January). Modeling geochemical processes in enhanced geothermal systems with CO2 as heat transfer fluid. In Proceedings, Thirty-Sixth Workshop on Geothermal Reservoir Engineering (pp. 583–590).
Arnórsson, S., Stefánsson, A., & Bjarnason, J. O. (2007). Fluid-fluid interactions in geothermal systems. Reviews in Mineralogy and Geochemistry, 65(1), 259–312.
Arnórsson, S., Angcoy, E., Bjarnason, J. Ö., Giroud, N., Gunnarsson, I., Kaasalainen, H., … Stefánsson, A. (2010). Gas chemistry of volcanic geothermal systems. In World Geothermal Congress, Bali, Indonesia (pp. 25–29).
Barnett, R. M. (2017). Principal component analysis. Geostatistics Lessons; Deutsch, JL.
Battistelli, A., Calore, C., & Pruess, K. (1997). The simulator TOUGH2/EWASG for modelling geothermal reservoirs with brines and non-condensible gas. Geothermics, 26(4), 437–464.
Budisulistyo, D., Wong, C. S., & Krumdieck, S. (2017). Lifetime design strategy for binary geothermal plants considering degradation of geothermal resource productivity. Energy Conversion and Management, 132, 1–13.
Cecioni, M., & Cei, M. (2020). Numerical Simulation of Cerro Pabellon Geothermal Field (Chile) with TOUGH-2®. In Proceedings World Geothermal Congress (p. 1).
Chiodini, G., & Marini, L. (1998). Hydrothermal gas equilibria: the H2O-H2-CO2-CO-CH4 system. Geochimica et Cosmochimica Acta, 62(15), 2673–2687.
Chiodini, G., Caliro, S., Cardellini, C., Granieri, D., Avino, R., Baldini, A., … Minopoli, C. (2010). Long-term variations of the Campi Flegrei, Italy, volcanic system as revealed by the monitoring of hydrothermal activity. Journal of Geophysical Research: Solid Earth, 115(B3).
Carroll, S., & Stillman, G. (2014, February). Assessment of key physical and chemical research findings for the use of CO2 as a heat exchanging fluid for geothermal energy production. In Proceedings of the Thirty-Ninth Workshop on Geothermal Reservoir Engineering (pp. 24–26).
Cuoco, E., De Francesco, S., & Tedesco, D. (2013). Hydrogeochemical dynamics affecting steam-heated pools at El Chichón Crater (Chiapas–Mexico). Geofluids, 13(3), 331–343.
d'Amore, F., Fancelli, R., Saracco, L., & Truesdell, A. H. (1987). Gas geothermometry based on CO content–application in Italian geothermal fields (No. SGP-TR-109-35). International Institute for Geothermal Research, CNR, Pisa, Italy; US Geological Survey, Menlo Park, CA.
A
D’Amore F. (1991) Gas geochemistry as a link between geothermal exploration and exploitation. In Application of Geochemistry in Geothermal Reservoir Development. (ed. F. D’Amore), pp. 93–117. UNITARDaniele, L., Taucare, M., Viguier, B., Arancibia, G., Aravena, D., Roquer, T., … Morata, D. (2020). Exploring the shallow geothermal resources in the Chilean Southern Volcanic Zone: Insight from the Liquine thermal springs. Journal of Geochemical Exploration, 218, 106611.
De Silva, S., Zandt, G., Trumbull, R., Viramonte, J. G., Salas, G., & Jiménez, N. (2006). Large ignimbrite eruptions and volcano-tectonic depressions in the Central Andes: a thermomechanical perspective. Geological Society, London, Special Publications, 269(1), 47–63.
De Silva, S. L., & Kay, S. M. (2018). Turning up the heat: high-flux magmatism in the Central Andes. Elements: An International Magazine of Mineralogy, Geochemistry, and Petrology, 14(4), 245–250.Elíasson, E. T. (2001). Power generation from high-enthalpy geothermal resources. GHC Bulletin, 22(2), 26–34.
Ebrahimi, P., Guarino, A., Allocca, V., Caliro, S., Cicchella, D., & Albanese, S. (2023). The relationship between dissolved radon and other geochemical parameters in Campi Flegrei volcanic aquifer (Southern Italy): A follow-up study. Applied Geochemistry, 151, 105607
Galarce Arenas, F. F. (2022). Modelamiento termodinámico de la alteración hidrotermal bajo condiciones de reservorio del sistema geotermal Cerro Pabellón, Cordillera de los Andes, Norte de Chile.
A
GDN-Geotermica Del Norte, 2018. Modelo Conceptual del Sistema Geotermico Cerro Pabellón, internal report. (Unpublished report). p. 68 (in spanish).Giggenbach, W. F. (1980). Geothermal gas equilibria. Geochimica et cosmochimica Acta, 44(12), 2021–2032.
Giggenbach, W. F. (1987). Redox processes governing the chemistry of fumarolic gas discharges from White Island, New Zealand. Applied Geochemistry, 2(2), 143–161.
Giggenbach, W. F. (1988). Geothermal solute equilibria. derivation of Na-K-Mg-Ca geoindicators. Geochimica et cosmochimica acta, 52(12), 2749–2765.
A
Giggenbach W. F. (1991) Chemical techniques in geothermal explo- ration. In Application of Geochemistry in Geothermal Reservoir Development. (ed. F. D’Amore), pp. 119 – 144. UNITARA
Giggenbach WF, Sano Y, Wakita H (1993) Isotopic composition of helium, and CO2 and CH4 contents in gases produced along the New Zealand part of a convergent plate boundary. Geochim Cosmochim Acta 57: 3427–3455Gudmundsson, B. T., & Arnórsson, S. (2002). Geochemical monitoring of the Krafla and Námafjall geothermal areas, N-Iceland. Geothermics, 31(2), 195–243
Golla, J. K. (2018). Using Principal Component Analysis to Aid in Visualization and Interpretation of Geothermal Solute Chemistry: An Application to Yellowstone Thermal Waters. Geothermal Resources Council.
González, R., Oncken, O., Faccenna, C., Le Breton, E., Bezada, M., & Mora, A. (2023). Kinematics and convergent tectonics of the Northwestern South American plate during the Cenozoic. Geochemistry, Geophysics, Geosystems, 24(7), e2022GC010827.
Godoy, B., Taussi, M., González-Maurel, O., Renzulli, A., Hernández-Prat, L., le Roux,P., … Menzies, A. (2019). Linking the mafic volcanism with the magmatic stages during the last 1 Ma in the main volcanic arc of the Altiplano-Puna Volcanic Complex (Central Andes). Journal of South American Earth Sciences, 95, 102295.
Gorini, A., Ridolfi, F., Piscaglia, F., Taussi, M., Renzulli, A., 2018. Application and reliability of calcic amphibole thermobarometry as inferred from calc-alkaline products of active geothermal areas in the Andes. J. Volcanol. Geotherm. Res. 358, 58–7
Hosgor, F. B., Tureyen, O. I., & Satman, A. (2016). Keeping inventory of carbon dioxide in liquid dominated geothermal reservoirs. Geothermics, 64, 55–60.
Li, T., Gao, R., Gao, X., & Liu, Q. (2022). Synergetic Effect of Non-Condensable Gas and Steam Quality on the Production Capacity of Geothermal Wells and Geothermal Power Generation for Hot Dry Rock. Energies, 15(20), 7726.
A
Mercado, J. L., Ahumada, S., Aguilera, F., Medina, E., Renzulli, A., & Piscaglia, F. (2009). Geological and Structural Evolution of Apacheta-Aguilucho Volcanic Complex (AAVC), Northern Chile. Santiago, 22, S7_002.Petrillo, Z., Tripaldi, S., Mangiacapra, A., Scippacercola, S., Caliro, S., & Chiodini, G. (2023). Principal component analysis on twenty years (2000–2020) of geochemical and geophysical observations at Campi Flegrei active caldera. Scientific Reports, 13(1), 18445.
A
Powell, T., & Cumming, W. (2010, February). Spreadsheets for geothermal water and gas geochemistry. In Proceedings (pp. 4–6).Prat, L. H., Cannatelli, C., Godoy, B., Manosalva, D. A., Morata, D., & Buscher, J. T. (2023). Amphibole recycling and polybaric crystallization in rhyolitic lava domes from melt inclusion geochemistry at Cerro La Torta, Altiplano-Puna Volcanic Complex. Journal of South American Earth Sciences, 130, 104569.
A
Seguel Cárdenas, M. A. (2023). Scaling en la central geotérmica Cerro Pabellón, norte de Chile: Química, mineralogía y modelo geoquímico.Shock, E. L. (1994). Application of thermodynamic calculations to geochemical processes involving organic acids. In Organic acids in geological processes (pp. 270–318). Berlin, Heidelberg: Springer Berlin Heidelberg.
A
Snyder, G., Poreda, R., Fehn, U., & Hunt, A. (2003). Sources of nitrogen and methane in Central American geothermal settings: Noble gas and 129I evidence for crustal and magmatic volatile components. Geochemistry, Geophysics, Geosystems, 4(1), 1–28.A
Stimac, J., Goff, F., & Goff, C. J. (2015). Intrusion-related geothermal systems. In The encyclopedia of volcanoes (pp. 799–822). Academic Press.Taussi, M., Nisi, B., Pizarro, M., Morata, D., Veloso, E. A., Volpi, G., … Renzulli,A. (2019). Sealing capacity of clay-cap units above the Cerro Pabellón hidden geothermal system (northern Chile) derived by soil CO2 flux and temperature measurements. Journal of Volcanology and Geothermal Research, 384, 1–14.
Tong, C. D., Yan, X. F., & Ma, Y. X. (2013). Statistical process monitoring based on improved principal component analysis and its application to chemical processes. Journal of Zhejiang University SCIENCE A, 14(7), 520–534.
Ventura, G. T., Hall, G. J., Nelson, R. K., Frysinger, G. S., Raghuraman, B., Pomerantz,A. E., … Reddy, C. M. (2011). Analysis of petroleum compositional similarity using multiway principal components analysis (MPCA) with comprehensive two-dimensional gas chromatographic data. Journal of Chromatography A, 1218(18), 2584–2592.
A
Verma, S. P., Pandarinath, K., & Santoyo, E. (2008). SolGeo: A new computer program for solute geothermometers and its application to Mexican geothermal fields. Geothermics, 37(6), 597–621Yuan, W., Chen, Z., Grasby, S. E., & Little, E. (2021). Closed-loop geothermal energy recovery from deep high enthalpy systems. Renewable Energy, 177, 976–991.