Zero-Emission Geothermal Power Generation: Experimental Study on Carbonate Mineralization through CO₂-Andesitic Pyroclastic Rock Interaction at Oku-Aizu Geothermal Plant

Ryosuke Oochi ¹, Kengo Aoyama ², Akira Ueda ^{3, *}

1. Life Cycle Valuation Headquarters, Shimizu Corporation, 2-16-1 Kyobashi, Chuoku, Tokyo 104-8370, Japan

2. Oku-Aizu Geothermal Co. Ltd., 1034-1, Sunagohara, Yanaizu, Kawanuma-gun, Fukushima 969-7321, Japan

3. Graduate School of Science and Engineering for Research, University of Toyama, 3490 Gofuku, Toyama 980-8555, Japan

* Correspondence: Akira Ueda

Academic Editor: Andres Navarro Flores

Special Issue: Geothermal Energy Exploration and Production

Received: November 27, 2019 | Accepted: February 03, 2020 | Published: February 12, 2020

Journal of Energy and Power Technology 2020, Volume 2, Issue 1, doi:10.21926/jept.2001002

Recommended citation: Oochi R, Aoyama K, Ueda A. Zero-Emission Geothermal Power Generation: Experimental Study on Carbonate Mineralization through CO 2 -Andesitic Pyroclastic Rock Interaction at Oku-Aizu Geothermal Plant. Journal of Energy and Power Technology 2020; 2(1): 002; doi:10.21926/jept.2001002.

© 2020 by the authors. This is an open access article distributed under the conditions of the Creative Commons by Attribution License, which permits unrestricted use, distribution, and reproduction in any medium or format, provided the original work is correctly cited.

1. Introduction

According to a report provided by the Japanese government, by 2050, Japan’s current geothermal power generation of 530 MW from 17 localities will increase tenfold. However, the current CO₂ emissions from the geothermal power plants in Japan amount to ca. 500,000 tons/y [1]. The global, CO₂ emissions from geothermal plants have been discussed worldwide by considering the Life Cycle Assessment [2,3,4,5,6,7]. The current installed capacity worldwide is approximately 11 GW as of 2010, and the CO₂ emissions from all the geothermal fields amount to 120 kg/MWh (11.5 Mtons/y) [4] and to 48 kg/MWh (4.6 Mtons/y) [6] as reported by LCA for the average scenario. In general, CO₂ emissions from the geothermal plants in Japan are low; however, a few Japanese geothermal power plants emit levels of CO₂ equivalent to the thermal power plants [1]. In such geothermal power plants, CO₂ emissions could be reduced to zero by reinjecting all the generated steam including CO₂ and H₂S as well as the brine used for power generation into the reservoir (the process is referred to as zero-emission geothermal power generation). Binary power generation involves generating electricity using a heat medium such as ammonia or pentane. However, no previous study has reported reinjecting large amounts of steam containing CO₂ and H₂S along with brine (~1,000 tons/h) into the reservoir. Therefore, it is necessary to assess what type of reaction would occur when the geothermal fluids containing large amounts of CO₂ and H₂S are reduced underground at the temperatures in the range of 150–250 °C. In geothermal power plants, brine is injected naturally into the reinjection wells because these wells are generally drilled at a lower elevation compared to the production wells. This enables brine injection without the requirement of a pump. In the zero-emission system, the pressure of the brine and steam is maintained at high levels, so that it may be returned directly to the injection wells.

Xu and Pruess (2008) [8] examined the behavior of CO₂ fluid in an enhanced geothermal system using theoretical simulation with TOUGHREACT [9] and demonstrated carbonate precipitation at high temperature. In Iceland, sub-surface mineral storage of carbon and sulfur at the CarbFix site at 40–80 °C was reported to have been achieved. The CarbFix approach involves accelerating the mineralization of the acid gases (CO₂ and H₂S) injected into the basaltic layers [10,11,12,13,14,15]. Basaltic rocks prove to be advantageous for the mineralization of CO₂ and other acid gases as they are relatively reactive compared to most of the other rock types. Their dissolution releases the divalent metal cations, such as Ca²⁺, Mg²⁺, and Fe²⁺, as well as assists in the neutralization of the acidic waters as the released cations provide something for the dissolved carbonate to react with [14,15]. Natural investigations on basalt–CO₂ interaction have been reported in the oceanic flood basalts [16] and in Idaho Springs in Idaho, USA [17]. Laboratory-based experimental approaches have also been performed in other parts of the world [18,19,20,21].

In Japan, the injection of CO₂ into high-temperature geothermal areas composed of granitic rocks for the immobilization of minerals has been studied [22,23]. This technology, referred to as “Georeactor”, injects the CO₂ dissolved in the groundwater to allow it to chemically react with the granitic rocks in order to fix the carbonate minerals below the ground. As with all the chemical reactions, this reaction also occurs faster at higher temperatures, and the solubility of the resultant carbonate minerals is sufficiently low for their formation at higher temperatures. In the Sumikawa geothermal field, Akita Prefecture, Japan, the Ca²⁺, Mg²⁺, and CO₂ dissolved in the water react with andesitic to dacitic lava and the tuff reservoir rocks, resulting in the formation of a natural carbonate layer, suggesting the possibility of the immobilization of the carbonate minerals in the geothermal area [24]. Yoo et al. (2013) [25] evaluated the change in permeability of the reservoir layer during the carbonate deposition through stepwise numerical calculations and provided predictions regarding when and where the carbonate deposits might be found in the reservoirs. The authors also demonstrated that the permeability around the injection point tends to significantly reduce upon carbonate accumulation, which might result in the insufficient injection of CO₂.

The Yanaizu–Nishiyama Geothermal Power Plant in the Oku–Aizu area of Japan had an electricity generation capacity of 65 MW when it began operating in the year 1995, and the capacity was increased by 30 MW from August 2017 onward. The steam and brine from the production well are separated using a separator, and the power is generated by the steam. The brine is reinjected to the underground, and at this point, most of the CO₂ gas in the steam is released into the atmosphere. In zero-emission geothermal power generation, a two-phase fluid of steam and brine is subjected to heat-exchange with a heat medium such as ammonia, and the heat medium is then sent to a turbine for power generation. The heat-exchanged two-phase fluid is reinjected directly to the underground without any component being released into the atmosphere. The geothermal reservoir contains open fractures associated with steeply dipping faults. Initially, the reservoir used to be water dominated, although, after the commencement of operations, it was changed to two-phase fluid (steam and brine) with a few steam dominated areas. The maximum temperature inside the wells is 340 °C [26]. The estimated concentration of the total dissolved components in the reservoir fluid is ~2% (wt.%), while the CO₂ concentration is 1% (wt.%), which is a high salt and gas concentration compared to the other Japanese geothermal fields; however, the H₂S concentration is considerably low [26,27,28]. The amount of CO₂ (450–550 g/kWh) discharged from this power plant is higher than that discharged by the other geothermal power plants (15–150 g/kWh), and corresponds to the amount of CO₂ discharged by the LNG (liquefied natural gas) plants (~600 g/kWh) [1]. Oochi et al. (2014) [29] analyzed the carbon and oxygen isotope composition of calcite in the wells and suggested that the source of these high levels of CO₂ and salt in the geothermal fluids was of volcanogenic origin.

The present study carried out the experimental simulation of the behavior of CO₂ during the reinjection of brine containing a large amount of CO₂ into the geothermal reservoir of the Yanaizu–Nishiyama Power Plant for zero-emission operation. The study focused on understanding the interaction of CO₂ with the andesitic pyroclastic rock (tuff) at 150 °C (the temperature corresponding to the reservoir temperature around the reinjection wells) and predicting the involved chemical reaction using geochemical simulations.

2. Outline of the Geology

The Oku–Aizu geothermal area is located in the mountainous region of the south-western margin of Aizu Basin in the central part of the Fukushima Prefecture, Japan (Figure 1). The basement of this area consists of early Triassic sedimentary and granitic rocks. Mid-Miocene marine (and partly non-marine) pyroclastic and volcanic rocks overlay this basement [30,31,32]. Three major north-wests to south-east trending faults occur in this geothermal field: the Chinoikezawa, Sarukurazawa, and the Oizawa faults (Figure 1); the production wells were drilled into the former two faults. In the vicinity of the Sunagohara area, which occurs north of the production area, a Na⁺–Cl⁻-type hot spring (Nishiyama Hot Spring) originates from the intersection of two faults, the Takiyagawa and the Oizawa faults [23].

Figure 1 Geological map and cross-section of the Oku–Aizu geothermal field [28].

Two stages of hydrothermal alteration have been recognized in the host rocks of this region. The first one is the mid-Miocene submarine volcanic related and the other is currently ongoing alteration from the hydrothermal activity. Large quantities of chlorite and sericite are commonly observed in the reservoir rocks (rhyolitic pyroclastic rocks present in the lower Miocene Takiya River Formation), which are thought to have been formed as a result of the submarine volcanic activity [29], along with shallow clay mineralization (smectite, sericite, kaolinite, and zeolite). Anhydrite is present in the high-temperature zone (100-200 °C) of the reservoir, and small amounts of calcite are present at both shallow and deep zones [26,27,28,32,33].

3. Experimental Procedure

A 100-mL Teflon beaker, which was used as the reaction vessel, was placed in a stainless-steel vessel (SUS; HUT–100; San-Ai Science Co., Ltd.) (Figure 2a). These vessels were connected to a CO₂ and/or vacuum line via a high-pressure valve. An andesitic pyroclastic rock sample was obtained from a 1,691 m deep core sample collected from well 84N–3t located in the Oku–Aizu geothermal area. The microscopic observations and the chemical composition of the rock sample are presented in Figure 2b and Table 1, respectively. The microscopic observation of the rock sample was performed at the Mitsubishi Materials Techno Co., Ltd, which revealed a volcanic clastic texture and a composition consisting of quartz, plagioclase, potassium feldspar, iron hydroxide, zircon, apatite, volcanic glass, pumice, andesite, and sandstone, along with alteration minerals (chlorite, illite, calcite, and iron hydroxide). The CaO content in the rock sample was 0.9% (wt.%) (Table 1), which is lower than that of the other Japanese andesitic rocks such as a standard rock JA–1 (5.7% wt.%) [34].

Table 1 Components of the volcanic tuff used in the experiments.

Figure 2 (a) Experimental device used for the laboratory experiments and (b) microscopic observations of the rock sample; (upper) open and (lower) crossed Nicolls.

According to the analysis of the carbon and oxygen isotopic composition of the calcites in the core samples, the geothermal fluids in the Oku–Aizu geothermal plant are believed to have originated from meteoric water with oxygen shift and/or fossil seawater [29]. This CO₂–rock interaction experiment was the first test to examine the behavior of CO₂ in the zero-emission geothermal system at Oku-Aizu. Therefore, tap water that originated from river water at the University of Toyama was used as a representative fluid for meteoric water. The composition of the tap water used has been presented in Table 2. It was observed that the tap water contained small amounts of Ca²⁺ and other minor cations.

Table 2 Chemical composition of the tap water collected at the University of Toyama.

The rock sample (7 g) and water (70 mL) were placed in the Teflon beaker in the SUS reaction vessel. A valve was attached to the lid of the vessel in order to either degas it or inject the gas in the vessel. After degassing the vessel for 1 min using a vacuum pump at <1 KPa, CO₂ gas was injected into the vessel using a syringe pump. At 150 °C, the CO₂ pressure was controlled and maintained at 10 MPa, at which 3% (wt.%) CO₂ was dissolved in the water [35]. Ar gas (without CO₂) was also injected into the vessel for comparative experiments. The reaction vessel was subsequently placed in a high-temperature reaction furnace and the reaction (1 rpm) was allowed to occur at 150 °C for a predetermined period. Subsequent to the rapid cooling of the vessel, when the temperature reached below 100 °C, the water and rock samples were collected for analysis. An aliquot of the water sample was filtered through a 0.22-µm filter for chemical analyses. The rock sample was dried in an oven at 110 °C for two days, following which it was subjected to SEM (scanning electron microscope) and XRD (X-ray diffraction) analyses.

4. Analytical Procedures

The water samples were analyzed for Na⁺, K⁺, Ca²⁺, and Mg²⁺, using an atomic absorption spectrometer (AA–6650; Shimadzu Corporation), and also for Cl and SO₄ using ion chromatography (761 Compact IC; Metrohm). The contents of Mg²⁺, Mn²⁺, Al³⁺, and total-Fe were measured using ICP-MS (7500 series) at the Okayama University. SiO₂ content was measured using the molybdenum yellow method and an absorptiometer (UV-VIS Spectrophotometer; Shimadzu Corporation). The chemical compositions of the rock samples prior to and after conducting the experiments were observed under a fluorescence X-ray analyzer (Supermini 200; Rigaku). Rock surface observations were performed using a scanning electron microscope (SEM; MiniscopeTM 3030; Hitachi).

The reaction temperature for this experiment was maintained at 150 °C, assuming the depth for the CO₂-rich fluid-rock interaction to be 1,000 m. The internal pressure of the vessel at this temperature was maintained at 10 MPa. The amount of CO₂ that could be dissolved in the water samples at 150 °C was calculated using the geochemical calculation software PhreeqC [36]. These calculations were then used to calculate the weight of CO₂ to be placed inside the reaction vessel in the laboratory environment, and consequently, the initial filling pressure of CO₂ was set at 4 MPa. Similarly, the initial filling pressure of Ar gas was set at 7.3 MPa. The reaction periods used were 1, 2, 5, 10, and 15 days.

5. Results and Discussion

5.1 Chemical Compositions of Solutions after CO₂–Rock Reaction

Table 3 and Figure 3 present the concentrations of chemical components in the solutions after the CO₂–rock reaction experiments. The concentrations of Na⁺, Ca²⁺, and SiO₂ in the solutions reacted with CO₂ gas were observed to increase with the increasing reaction time, whereas, the K⁺ and Mg²⁺ concentrations remained constant (Figure 3a). In contrast, the chemical components in the solutions reacted without CO₂ (i.e., solutions reacted with Ar gas) exhibited an increase within the first two days and remained constant thereafter (Figure 3b). Figure 4 presents the concentrations of the chemical components in the solutions with and without CO₂. It is clearly observable that the solutions reacted with CO₂ contained higher amounts of Ca²⁺ compared to the solutions reacted with Ar (Figure 4a). Mg²⁺ content was observed to have increased in the solutions reacted with CO₂. In addition, in one case among the solutions reacted without CO₂, Mg²⁺ content was observed to increase within the first two days and decrease thereafter (Figure 4b). In one case, Na⁺ and K⁺ in the solutions reacted both with and without CO₂ were observed to increase within the first two days and remain almost constant thereafter.

Table 3 Chemical compositions of the solutions after the reactions.

Figure 3 Chemical compositions of the solutions during the reaction with (a) CO₂ and (b) Ar.

The silica concentration in the solutions reacted with CO₂ increased from 96 to 272 mg/L when the duration increased from 1 to 15 days and was higher than that in the solutions reacted without CO₂ (57–134 mg/L) for the same period (Figure 4e). The solubility of quartz and amorphous silica at 150 °C was determined to be 125 mg/L and 619 mg/L, respectively [37], which implied that silica concentration in the solution reacted without CO₂ was saturated with respect to quartz. In contrast, silica concentration in the solution reacted with CO₂ continued to increase the solubility level of amorphous silica. Kuroda et al. (2009) [38] determined the chemical compositions of solutions after reactions with plagioclase and granodiorite at 150 °C for 15 days and reported that the silica concentrations in the solutions reacted with CO₂ and N₂ were 263 mg/L and 132 mg/L, respectively, which are similar to the values obtained in the present study.

Figure 4 Chemical compositions of the solutions during the reaction showing the evolution of the concentrations of (a) Ca²⁺, (b) Na⁺, (c) Mg²⁺, (d) K⁺, (e) SiO₂, (f) SO₄^2-, and (g) Cl^-.

The rock sample used in the present study contained anhydrite and silicate minerals (i.e., volcanic glass and plagioclase), which implied that Ca²⁺ could have been released from both anhydrite and silicate minerals. The Ca²⁺ concentrations in the solutions were observed to remain almost constant after two days of reaction (Figure 3a). The Ca²⁺/SO₄²⁻ mole ratios in the solutions after reaction with CO₂ were determined to be in the range of 2.7–3.3, higher than those in the solutions reacted without CO₂ gas (0.7–1.0 ratios; Table 3). These results indicated that Ca²⁺ in the solution after reaction with CO₂ gas could be released because of the dissolution of anhydrite as well as because of the reaction with silicate minerals. Ueda et al. (2005) [22] reported similar results, i.e., the Ca²⁺/SO₄²⁻ mole ratios in the solutions reacted with CO₂ gas after reaction with granitic rocks and distilled water at 150 °C were higher compared to those in the solutions reacted with Ar gas.

5.2 Comparison of Ca²⁺ Concentration in Solutions after Andesitic Tuff–CO₂ Interaction with other Rock Types

Several studies have been conducted on rock–CO₂ interaction for carbon sequestration, although the majority of these studies have investigated the temperature conditions below 100 °C. Figure 5 compares the results of the reaction test with granodiorite–CO₂–water at the temperatures of 150 °C and 200 °C [22,38]. In the reaction with granodiorite at 150 °C, the behavior of Ca²⁺ concentration exhibited the same trend as that observed for the andesitic tuff in the present study. The Ca²⁺ concentration increased to 200 mg/L within five days and decreased thereafter. On the other hand, Ca²⁺ concentration maintained a low value (100 mg/L; Figure 5) in the test conducted at 200 °C. At both the reaction temperatures, the solutions reacted with CO₂ exhibited higher Ca²⁺ concentrations compared to those reacted in the absence of CO₂.

Luhmann et al. (2017) [21] conducted flow-through experiments with basalt and CO₂-rich brine at 150 °C and observed that Ca²⁺ concentration increased at the beginning of the reaction up to ca. 80 mg/L and then decreased rapidly to <1 mg/L with time (within 30 days). The authors noted that after the reaction, the solutions were enriched in Fe (up to 86 mg/L), undersaturated with respect to calcite and dolomite, and oversaturated with respect to siderite. In the experiments conducted on the granodiorite and CO₂ reaction at 200 °C, total Fe concentration in the solutions after the reaction was observed to reach a maximum value of 0.7 mg/L, while the concentration was less than 0.1 mg/L in most of the cases [18]. In these experiments, calcite was observed to be oversaturated in the solutions. These results suggested that carbonate precipitation during the interaction between CO₂-rich brine and the rocks is controlled by siderite for the basaltic rocks and by calcite for the granitic rocks.

Griffioen and Appelo (1993) [39] demonstrated that for the solutions containing Fe²⁺, and for those containing Ca²⁺ with a high concentration of CO₂, siderite is precipitated first, and is followed by calcite. Using the oxygen isotope compositions of the carbonate minerals and the XRD analytical data of the rock samples, Ueda et al. (2001) [22] demonstrated that during the infiltration of CO₂-rich groundwater, siderite was precipitated at a shallower depth in the Sumikawa geothermal area, while calcite and dolomite were observed at deeper zones. The authors also demonstrated that the carbonate minerals precipitated with smectite during the interaction and acted as a cap rock for the reservoir. These results indicated carbonate formation during the interaction with CO₂-rich water.

Figure 5 Comparison of Ca concentrations in solutions reacted with and without CO₂ for Oku–Aizu tuff (this study) and granodiorite [22,38] at 150 °C.

5.3 Saturation of Minerals in Solution after CO₂–Rock Reaction

In order to investigate the possibility of precipitation of minerals including the carbonate minerals, the saturation index (SI) for several minerals was calculated with the geochemical software PhreeqC [32], using the obtained chemical composition values for the solution after the reaction. PhreeqC is a geochemical code that utilizes solution equilibrium models to calculate aqueous speciation and the saturation state of mineral phases and has been applied previously to CO₂–rock reactions [40,41,42,43,44].

The SI has been represented by log (Q/K), where Q represents the activity product (Ion Activity Product), i.e., the activity product of the actual ions in the mineral precipitation/dissolution reaction equation, and K represents the solubility product. If the SI value is positive, the mineral tends to be oversaturated, i.e., it tends to precipitate, and if the SI value is negative, it indicates that the mineral is in a dissolved state.

The results of the calculations using the analyses of the solutions produced by the reactions between the Oku–Aizu tuff and the CO₂-rich water are presented in Table 4 and Figure 6. The carbonate minerals, such as calcite, aragonite, and dolomite, were observed to be over-saturated and undersaturated in the solutions reacted with and without CO₂, respectively. Anhydrite was always undersaturated in the solutions reacted with and without CO₂. Ueda et al. (2005) [22] indicated that when no CO₂ is present, anhydrite dissolves initially, adding Ca²⁺ to the solution, thereby causing dolomite, calcite, and anorthite to become oversaturated. However, when CO₂ is added, anorthite becomes substantially undersaturated and dissolves, adding more Ca²⁺ to the solution and resulting in an even more positive saturation index (SI) for calcite and dolomite.

Figure 6 depicts that after the reaction occurred for 1 to 14 days with and without CO₂, the solutions were always saturated or oversaturated with calcite and dolomite. In particular, the SI for dolomite in the solutions reacted with CO₂ was observed to increase from approximately 0 to 2 with the increasing reaction time. In contrast, the SI of dolomite in the solutions reacted without CO₂ remained less than zero. These results demonstrated that plagioclase and volcanic glass were dissolved by the reaction with the CO₂-saturated water along with the formation of calcite and dolomite.

Table 4 Saturation indices of mineral phases calculated from chemical compositions of solutions after the reaction.

Figure 6 Saturation indexes for the minerals in the solution after reaction with (a) CO₂ and (b) Ar. The saturation index (SI) was calculated with the geochemical software PhreeqC [36].

5.4 SEM Analyses of Rock Surface after the Reaction

In the present study, after undergoing the rock–CO₂/Ar reaction test, the rock surface was observed under an SEM (Figure 7a and 7b). Rhomboid crystals were observed on the surface of the rock samples reacted for 15 days (Figure 7a), indicating a high possibility of the presence of aragonite, compared to the rock sample reacted with Ar (Figure 7b). No clear carbonate mineral precipitates were observed in the samples on the other reaction days. Kuroda et al. (2009) [38] observed calcite crystals on the surface after reacting plagioclase and CO₂ at 150 °C, although such crystals were not observed in the granodiorite samples subjected to the same experimental condition. The plagioclase in that study exhibited 21.6% (wt.%) Ca²⁺ content, which was higher than that in the granodiorite (1.2% wt.%) and the andesitic tuff (0.9% wt.%) examined in the present study (Table 1). When the saturation index for the solutions after the rock–CO₂ reaction was considered, it was inferred that the solution was oversaturated with respect to calcite. These results suggested that in the present study, aragonite and calcite must have been precipitated after the reaction and observed only for the sample reacted with CO₂ for 15 days. In the case of the other samples, the precipitated amount might have been too small to be observed in the SEM analyses.

Figure 7 SEM images of the surface of rock samples after reaction with (a) CO₂ and (b) Ar.

5.5 Simulation of the Behaviors of Chemical Components in Solution during the Rock-CO₂ Interaction

In order to theoretically predict the chemical reactions occurring among the rock, water, and CO₂, PhreeqC code [36], using the experimental conditions (such as temperature, pressure, and the composition of the solution) and a thermodynamic database provided by the Lawrence Livermore National Laboratory, National Institute of America were used. Using this database, calculations could be applied to high temperatures of up to 300 °C. Besides the reaction temperature used in the experiments of the present study (150 °C), the reaction temperatures of 50 °C, 100 °C, 200 °C, and 250 °C were also used in the simulation. Calculated changes in the concentration of each element under the conditions identical to the experimental conditions were analyzed. The simulation model was created with an emphasis on Na⁺, Ca²⁺, and SiO₂. In the model, when the rock sample was assumed to be a sphere of size 1 mm, the reactive surface area was calculated to be 0.0023 m²/g. However, using this value, the eluted amounts determined for the main components (Na⁺, Ca²⁺, and SiO₂) were extremely small. Therefore, in the model, the surface areas of the initial minerals (quartz, albite, anorthite, etc.) were set to values 100–1,000 times greater than the prescribed values. Figure 8 illustrates that the experimental observation data were the most approximated among the simulation results, although the initial increase in Ca²⁺ estimated by the simulation was not observed. In the present study, the Ca²⁺ concentration in the solutions after the reaction was observed to gradually increase, similar to the previously reported data for labradorite at 120 °C [22]. The present simulation did not agree sufficiently with the experimental results, although it provided the approximate change in chemical compositions (Figure 8).

Figure 8 Simulation results for the solutions after reaction with (a) CO₂ and (b) Ar.

5.6 Investigation of CO₂ Mineralization

In the present study, the experimental temperature used for the Oku–Aizu tuff and water interaction was 150 °C. When conducting the zero-emission geothermal power generation, the temperature for reinjection of brine with high CO₂ concentration into the underground is not necessarily 150 °C. Therefore, the temperature was changed using the parameters obtained from the simulation results in order to predict the Ca²⁺ concentration change in the injected brine during the interaction with the andesitic tuff. In order to predict the optimum temperature for CO₂ mineralization, several other reaction temperatures (50, 100, 150, 200, and 250 °C) were analyzed (Figure 9), and it was observed that the lower the reaction temperature, the more Ca was eluted. This may have occurred because the solubility of calcium carbonate increased as the temperature decreased, implying that the elution is simply a dissolution of the precipitated carbonates.

Figure 9 Simulation results for the Ca²⁺ concentrations in the solutions of Oku–Aizu tuff-water interaction with CO₂ at different temperatures.

Using both field and batch experimental studies, Mito et al. (2008) [45] analyzed the carbonate mineral fixation of CO₂ at ~50 °C through geochemical reactions in a sandstone reservoir at Nagaoka, Japan (the place where the first Japanese pilot project of geological CO₂ storage was conducted). Seawater was reacted with powdered rock samples (rock/water wt. ratio was 100) with and without CO₂ at 10 MPa. The Ca²⁺ concentration was observed to increase from 406 mg/L to 529 mg/L within two days and remained constant thereafter. The SI values for calcite in the solutions after the reactions with and without CO₂ were determined to be −2.7 and −0.4, respectively. Mito et al. (2008) [45] demonstrated using field observation and theoretical calculations that the dissolution of plagioclase and chlorite has great potential for enhancing the neutralization of the acidified formation water after CO₂ injection and that the bivalent cations Ca²⁺, Mg, and Fe provided by these minerals are expected to increase the net carbonate mineral precipitation. The authors estimated that ~5% and half of the injected CO₂ could be fixed as carbonate minerals and aqueous carbonate ions, respectively, within ten years. The authors also demonstrated that mineral trapping could occur from the early stage of CO₂ storage. In the present study, Ca²⁺ concentration in the solution after rock–water interaction at 50 °C was estimated to be ca. 500 mg/L (Figure 9), similar to the results observed in the previously stated study, although the solutions used in the experiments were different for both the studies (tap water in the present study and seawater in the other study) [45].

5.7 New Scaling Problem of Carbonate during Zero-Emission Geothermal Operation

Ground facilities for zero-emission geothermal power generation are expected to be identical to the existing binary power generation facilities. The existing binaries generate electricity through heat exchange with a part of brine or between a two-phase fluid of steam and brine and an ammonia medium, and so on. In the zero-emission geothermal power generation facilities, all the fluid is reinjected into the underground through heat exchange.

In the existing geothermal plants, silica scaling is one of the major problems encountered during the geothermal operation worldwide. Silica in the brines becomes oversaturated with respect to amorphous silica after the vapor–brine separation at the surface, in which scenario, the silica in the brine precipitates in the brine pipes, reinjection wells, and the layer surrounding the reinjection wells.

As for the carbonate scales, it is well known that calcite and/or aragonite precipitate close to the flashing (vapor–brine separation) point in the production wells due to the oversaturation of the Ca²⁺–HCO₃^- concentration. Several scale inhibitors such as polyacrylic acid soda may be used to prevent the precipitation. In contrast, the carbonate scales do not precipitate in the reinjection wells due to the emission of most of the CO₂ into the atmosphere along with the vapor at the geothermal plants. When injecting a large amount of brine with high CO₂ concentration through the reinjection wells for zero-emission power generation at a geothermal power plant, the carbonate minerals such as calcite and aragonite must be precipitated in the reservoir layer close to the reduction wells within a few days. At lower temperatures, the reinjected brine might be more enriched in Ca²⁺ because of interaction with the reservoir rocks and may be precipitated as carbonate minerals during neutralization. In such a scenario, the permeability of the formation is lowered through the carbonate precipitation [25,46]. Therefore, it is necessary to control the temperature conditions and the CO₂ concentration in the fluid for precipitation in the reservoir without precipitating the carbonate minerals in the formation around the reinjection well. In the present study, it was observed that Ca²⁺ could have been dissolved from the carbonates in the rocks and also from the glass at 150 °C. When the CO₂-rich acidic fluid flows into the geothermal reservoir, it is expected that Ca²⁺ would be leached from glass and plagioclase as a result of the reaction with the surrounding rocks. Since the solubility of calcium carbonate decreases with increasing temperature, carbonate minerals are assumed to undergo precipitation.

6. Conclusions

Despite the low Ca²⁺ content (0.9% wt.%) in the sample rock used in the experiments, the Ca²⁺ concentration in the solutions at 150 °C increased sharply over the several days of observation, and the solutions became oversaturated with respect to carbonate minerals. This suggested that CO₂ may be stored as carbonate minerals such as aragonite and calcite. The simulation results for the rock–brine interaction revealed that sufficient Ca²⁺ elution is expected even in a low-temperature region at 50 °C. The results demonstrated that CO₂ emissions may be reduced to zero by reinjecting all the steam including CO₂ and brine used for power generation into the reservoir and that the zero-emission geothermal power generation would serve as one of the most useful techniques for environmental assessments.

Acknowledgments

The experiments on rock-CO₂ fluid interaction in this study were performed at the University of Toyama, with part of the simulation work carried out at the Shimizu Corporation. The authors would like to thank members of Oku–Aizu Geothermal Co., Ltd. and Mitsui Mining and Smelting Co. Ltd. for various kinds of convenience and cooperation during the rock sampling and data analyses. K. Yamaguchi of Mitsubishi Materials Corporation is acknowledged for his advice and help on the simulation of the rock-water reaction. We would like to express our gratitude to A. Ozawa of Mitsubishi Materials Techno Co., Ltd. for the microscopic observation of rock samples used in this study. We express our deep appreciation to J. Zhang, K. Marumo, M. Kusakabe, and K. Horikawa from the University of Toyama for their useful advice and guidance. Also, we would like to thank members of the Geochemical Laboratory of the University of Toyama, especially Y. Yamada, R. Isaji, E. Hanajima, and R. Ikeda.

We would like to thank J. Jin, an academic editor and two anonymous reviewers for their critical, kind and constructive comments and polishing English which improved the manuscript.

The authors would like to thank Enago (www.enago.jp) for the English language review for our first draft.

Author Contributions

These authors contributed equally to this work.

Competing Interests

The authors have declared that no competing interests exist.