Prototype Tests for the Treatment of Geothermal Waters for Raw Material Extraction and Freshwater Production

Geothermal fluids are a proven resource for sustainable baseload energy worldwide. Recently the fluids circulated in large volume streams in geothermal power plants have also come into focus for the potential of raw material extraction, such as Lithium. Geothermal fluids are the product of high-temperature and high-pressure water-rock interaction. This results in varying degrees of enrichment of different elements. Among them are elements of economic or strategic interest as lithium (Li), magnesium (Mg), cesium (Cs), but also elements typically causing unwanted mineral precipitations (scaling) within geothermal power plants such as silicon (Si). The extraction of specific target elements is still challenging in terms of integrating the process technology in a geothermal cycle. Especially in volcanic geothermal systems target elements are lower concentrated than in conventional brine resources and further the fluids tend to silica scaling. To enable the efficient extraction of the target elements, a treatment strategy was developed, consisting of an effective precipitation unit for silica reduction and two combined concentration steps for the enrichment of the target elements. The method uses the controlled precipitation of Si as Calcium-Silica-Hydrate phases and the concentration is conducted by reverse osmosis (RO) and membrane distillation (MD) using geothermal heat. The treatment approach was incorporated in a field demonstrator and tested in a geothermal system in the Southern Volcanic Zone in Chile. The results showed an effective silica reduction, enabling concentration rates up to a concentration factor of 16 under continuous flow-through conditions. The concentration of the dissolved solids on one side of the membranes enabled further the production of freshwater from geothermal fluids on the other side. Moreover, the MD process shows high energy efficiency in comparison to conventional evaporation processes and by geothermal sourcing also has direct potential for fossil fuel saving. Since volcanic resources are the most used in the global geothermal sector in terms of in­stalled capacity, the effective handling of the fluids has the potential for unleashing a global geothermal raw material potential.

1  Introduction

The energy transition and associated electrification shifts the demand for raw materials from hydrocarbons to various metal resources. Several of these, partly rare, metals have poorly diversified value chains, e. g., Lithium (Li) or cobalt (Co), as well as high associated environmental impacts during their production. Geothermal fluids as circulated in geothermal power plants can globally have considerable enrichments of valuable and even critical raw materials such as Li, antimony (Sb), magnesium (Mg) or cesium (Cs) (1, 2). The economic potential unfolds by combining the element concentration with the high flow rates (> 60 l/s) prevailing in geothermal energy production (3). Chile, with more than 200 active volcanoes, shows a large abundance of hydrothermal systems. This abundance leads to one of the largest geothermal potentials worldwide (4). The large potential for the Chilean energy sector is proofed by the first geothermal power plant Cerro Pabellón with an installed capacity of 81 MWe (5). Its raw material potential can be outlined by combining the flowrates of the eight production wells (up to 560 l/s) (6) with the measured Li concentrations (60 mg/l) (7) resulting in 1,060 t of pure Li circulated every year. This amount equals roughly 3 % of the whole Chilean Li production in 2022 (8) and would be sufficient for about 100,000 electric vehicle batteries. A multi-use of geothermal resources could be of several benefits. The additional revenue would strongly improve the economics of geothermal projects. The approach of direct Li extraction (DLE) from geothermal fluids would be more­over less consuming in terms of water and land use compared to conventional mining and additionally be self-supplying in terms of energy. Moreover, the accumulated amounts are in an order of magnitude that would enlarge Chiles global market share in Li production significantly.

A key challenge for geothermal energy production and even more for DLE application on geothermal fluids is the uncontrolled precipitation of minerals supersaturated in the solution. Especially silica scaling poses high demands during processing (9, 10). For this purpose, a controlled precipitation and treatment strategy was developed in the laboratory (11) and transferred into the design and construction of a demonstrator, which was already tested in an operating geothermal power plant in Germany (12). The herewith presented study shows the application of an updated version of the tested demonstrator customized for fluids from volcanic geothermal systems. Since volcanic fluids are the most frequently used source to provide geothermal power globally, successful raw material extraction would unleash an enormous potential. Therefore, the demonstrator was tested in continuous operation processing fluids from a hot spring in the Chilean Southern Volcanic Zone.

The research work presented was conducted in the framework of the BrineMine project funded by the German Federal Ministry of Education and Research (BMBF). The project was realized between 1st March 2019 and 31st March 2023 as a bi-national research project between German and Chilean research and industrial partners. The project structure features two focal points:

1. Determination of the economic potential of thermal waters as a raw material resource; and
2. pre-treatment of thermal waters prior to raw material extraction.

Project partners on the German site are the Fraunhofer Institute ISE (Institute for Solar Energy Systems), Freiburg, the Karlsruhe Institute of Technology (KIT), Karlsruhe, SolarSpring Membrane Solutions GmbH, Freiburg, and Geothermie Neubrandenburg (GTN), Neubrandenburg. Associated partners in Chile are the Andean Geothermal Centre of Excellence (CEGA) at the Universidad de Chile, Fraunhofer CSET, GTN Latin America and Transmark Renewables.

2  Fundamentals

2.1  Geothermal setting

Geothermal resources in Chile are strongly related to the active Andean volcanism. This volcanism is the result of the ongoing subduction of the Nazca Plate under the South American Plate that has taken place since Jurassic times (5) and hosts over 300 geothermal areas (4). The developed treatment system was tested in the field using geothermal volcanic fluids from the Termas de Puyehue, which are located in the region Los Lagos, about 100 km northeast of Puerto Montt. They were chosen because of the good accessibility via an installed well, sufficiently high temperatures (50 to 60 °C), and the availability of infrastructure for transporting and installing the demonstrator. The fluid has a low mineralization (~ 500 mg/l) yet is still enriched concerning various valuable elements such as Li, vanadium (V), molybdenum (Mo) and Cs (Table 1).

Thus, the effect of the treatment process on different elements can be determined. The chemical signature shows an intermediate water type with volcanic and meteoric influences as indicated by the ratios of the main anions and cations (13). Elevated SO₄ concentrations indicate volcanic origin respectively heating by volcanic sulfuric gases or sulfate weathering. High silicon (Si) concentrations (50 mg/l) are the result of the elevated temperatures and enhanced water-rock interactions, indicating subsurface temperatures of about 80 °C using conventional geothermometers (14). The geothermal fluids are currently used for different thermal baths, sanitary hot water and water supply of the various kitchens and laundries.

2.2  Treatment strategy

For extracting raw materials from geothermal fluids, various approaches are currently being investigated (15, 16). In terms of extraction efficiency, higher raw material concentrations are favorable (17). Thus, concentrating the dissolved element content before the extraction process is advisable. For enabling an efficient treatment in geothermal power plants, membrane processes open up great possibilities. Especially reverse osmosis (RO) powered by high pressures and membrane distillation (MD) powered by thermal energy could efficiently be incorporated into geothermal power plants. Moreover, these processes offer the additional possibility of fresh-water production. However, membrane technologies are susceptible to silicate precipitation in volcanic areas (18, 19).

For removing silica and thus, implementing membrane processes in geothermal settings, lime precipitation was evaluated (11). The approach aims at transferring the aqueous silica into the negatively charged H₃SiO₄-species by increasing the pH value > 10 (20, 21). In the presence of positively charged, divalent cations such as Ca²⁺, Calcium (Ca)-Silicate-Hydrate (CSH) phases can form (22, 23) and precipitate, thus reducing the Si and Ca content in solution. The described approach reached a silica removal of 98 % in previous studies (11, 12). In the demonstrator field tests in the geothermal power plant in Germany, a membrane distillation module downstream enabled concentrating of the already saline brine (~105 g/L TDS) by a concentration factor (CF) of three (24). The successfully tested demonstrator was then optimized for less saline, volcanic fluids.

2.3  Technical set-up

The demonstration plant (Figure 2a) for continuous operation was developed and built by Fraunhofer ISE in collaboration with SolarSpring GmbH. The hydraulic scheme (Figure 1) displays the major process steps: Process heat extraction, initial softening, RO pre-concentration, Si-reduction, liquid/solid separation, and MD post-concentration. The demonstrator is connected to the thermal water directly at the thermal spring with temperatures of about 60 °C and a defined flow rate of 110 l/h.

After the MD heat extraction, the fluid enters the chemical pretreatment unit (Figure 2b) at about 55 °C. To avoid spontaneous reaction with Si in the following steps the water is first passing through an initial softening step for removing bivalent ions such as Ca and Mg. Afterwards, the fluid enters the 40 l recirculation buffer tank. Here the fluid is introduced by an open inflow and the pressure is relived to atmospheric level. Furthermore, in the buffer tank, the pH is adjusted to pH 10.5 by a controlled NaOH dosing system, to increase the solubility of Si for the RO step and at the same time transferring Si into the negative aquatic species for later precipitation. This allows a significant pre-concentration of the Si-rich solution by a RO system operating at 25 bar in a feed-and-bleed operation mode and concentrates the brine by a factor of six to ten. While the main part of the water (96 l/h) is collected as high-quality permeate, the small fraction (14 l/h) of high pH concentrate leaves the recirculation tank through an overflow and gets introduced into a continuously steered tank reactor (CSTR) with a working volume of 14 l corresponding to a residence time of about 60 min. The reactor includes a flow-controlled dosing system that doses Ca ions in the form of an aqueous CaCl₂ solution to generate a super saturation with regard to amorphous CSH phases and thus initiates their precipitation. Along with the CSH phases, Ca carbonates form. The suspension is then passed through a custom-made miniature vacuum drum cake filter system (Figure 2c), which separates and dewaters the CSH precipitates from the suspension (Figure 2d). The pH of the filtrate is then re-adjusted using hydrochloric acid before being led into the next module. The last step in the process chain is a MD post-concentration that is thermally driven by the heat extracted from the inlet section of the demonstrator. Here, a spiral-wound membrane module with an active membrane area of 9 m² is applied. The module is built in a feed-gap configuration (FGMD) (25). The small feed stream is passed through a narrow gap and receives heat through a thin polymeric film from a heating solution. Pure water is extracted in vapor form that passes through a hydrophobic, microporous membrane and condenses in a cold permeate stream. The permeate is recirculated through a cooling tower that acts as a heat sink and allows the control of its temperature. The integrity of the membrane is monitored by conductivity measurement of the produced water. The quantification of the water extraction is done utilizing an electronic scale. The demonstrator includes nine sampling points, indicated with S1 to S9, that allow a chemical characterization of the water and solids in all relevant process steps.

2.3  Sampling and analysis

During the field operation, fluid temperature and pH values were measured using a compact precision handheld meter (WTW Multi 340i). To remove particles, all fluid samples were filtered with a syringe filter using a cellulose acetate filter (0.45 μm). To avoid precipitation or further reactions after the sampling process, samples for Si measurements were directly diluted by a factor of 1:10 using distilled water. For major and trace element analysis, separate samples were acidified using suprapure (37 %) hydrochloric acid. Major cations were measured with inductively coupled plasma optical emission spectrometry (ICP-OES, ICap 7000, ThermoFisher), anions were measured via ion chromato­graphy (IC, Compact 930, Methrom), and trace elements with inductively coupled plasma mass spectrometry (ICP-MS, ICap RQ, ThermoFisher). Precipitates were dried in the oven overnight at 105 °C. The mineralogy was analyzed with X-ray diffraction (XRD, D8 Discover, Bruker) and the major element chemistry using wavelength-dispersive X-ray spectroscopy (WDX, S4 Explorer, Bruker AXS) on fused beads. The analyses were conducted at the Laboratory for Environmental and Raw Materials Analysis (LERA), Institute of Applied Geosciences, KIT.

3  Results

The results of the multi-element analysis are displayed in table 1 and the relative changes between the individual process steps, induced by the treatment can be seen in figure 3. Si, as the major scaling forming element, passes through the first chemical treatment step.

It becomes enriched by a CF of five to six during the RO and is reduced in its concentration through precipitation by about 50 % due to the addition of CaCl₂. During the MD step, it becomes concentrated concurrently with other elements. Natrium (Na) concentration increases strongly at S3 because of the NaOH addition to reach a pH value >10 and increases further along with the water extraction of the membrane processes. Ca is initially totally removed by the softener (between S1 and S2) but its concentration increases again between S3 and S5 due to the CaCl₂ addition. Chlorine (Cl) concentration enriches conservatively during the membrane processes and shows a strong increase at S5 due to the CaCl₂ addition.

The comparison with the solid chemistry (Figure 4) shows, that the formation of CSH-phases and associated co-precipitation of Ca carbonates was successful. In contrast to the previous studies (11, 12), the silica reduction was only 50 % instead of 98 %.

The alkali metals Li, Cs, rubidium (Rb) and potassium (K) are negatively influenced by the primary water-softening step (between S1 and S2) and become strongly reduced by 80 to 90 %. After this step, the alkalis do not show an influence by the chemical treatment as in the previous studies (11, 12) and become concentrated in the order of magnitude expected from the membrane flowrates (CF 5-7 for RO and ~2 MD). Elements such as fluorine (F), aluminum (Al), V, and Sb pass through the softening step but partly co-precipitate during the CSH-precipitation step (between S3 and S5). SO₄, boron (B), and Mo do not interact with the chemical processes and just become concentrated linearly with the water extraction. The analysis of the final permeate (S9) shows, that the membrane integrity could be maintained during the production and the production of pure water (25 mg/l TDS) was successful during continuous operation. In further experiments, the MD was operated with the treated fluid in a batch feed-and-bleed mode to reach higher enrichment rates. The experiments enabled further concentration to a CF of about five reaching a final CF of up to 100 and a final TDS of about 40.000 mg/l.

The MD manufactured on average a flowrate of 14 l/h creating 7 l/h of permeate as well as 7 l/h of concentrate. The heat demand of the process was derived by cooling 300 l/h (~0,08 l/s) of spring water from ~60 to ~55 °C before entering the chemical treatment unit. The thermal springs show at the central well a joint seasonal flow rate between 4.5 and 9.0 l/s. With a density of 983.55 kg/m³ and a specific heat capacity of 4.18 kJ/kgK, different energy production scenarios were determined depending on the flow rate and potential heat extraction (Table 2).

The thermal heating demand of the MD, derived by the geothermal spring was in the range of about 1.6 kW leading to an average specific heat consumption of 225 kWh for one ton of permeate respectively concentrate. In comparison to an ideal evaporation of one ton of water consuming about 655 kWh, the MD process needs even under prototype conditions, just around one-third for evaporating the same amount of water. The used energy demand equals in the tested configuration 0.9 to 1.7 % of the energy of the whole accessible volume stream. Extrapolating these values on one production week of the demonstrator, leads to a production of 1,176 l permeate using 260 kWh. This energy amount is comparable to the energy of approximately 24 l oil (assuming a calorific value of 11 kWh/l).

In an upscaling scenario using one-third of the total volume stream (1.5 to 3 l/s) for process heat, 5,180 to 10,360 kWh thermal energy could be produced per week for processing 23 to 46 t of fresh water and an equal amount of concentrated brine using geothermal energy. The geothermal energy could substitute on this scale 471 to 942 l of oil per week. Without using the heat recovery of the MD process, the energy demand and consequently the fossil fuel demand would even be higher by a factor of three. Also, it can be highlighted, that this just represents the extraction of 5 K heat from the brine. Depending on the overall cooling of the fluid this displays just a fractional part of the overall producible energy at the given site (Table 2).

4  Concluding remarks

The field study investigated the application of the field demonstrator for controlled Si precipitation and concentration using RO and MD in low-mineralized volcanic fluids. The treatment enabled successfully the production of fresh water. Elements were concentrated by a factor of 20 in continuous operation and up to 100 in batch operation using sustainable geothermal heat. The direct heating use of the MD technology was shown to be energy efficient for water treatment in geothermal systems demonstrating enormous fossil fuel saving potential even when using just a fraction of the overall water flow.

It was demonstrated that the SI reduction under controlled formation of CSH phases was proofed in this environment, but less efficient than in the high-salinity waters in earlier studies. This could be due to the lack of a “salting out” effect in the low mineralization fluids. A further reduction of the Si concentration could still be achieved by longer reaction times or a higher pre-concentration. The results showed that potentially valuable elements such as Li, Cs, Rb and K are not negatively influenced by the chemical Si treatment and can be enriched using the membrane processes. However, the fluid analysis revealed that for the alkali metals, the water softener was a sink, already reducing their concentrations before entering the demonstrator. Depending on the target elements for a potential extraction, another configuration should be applied. Elements like B and Mo were not influenced by the chemical treatment and concentrated along with the water extraction.

Overall, the field study has demonstrated that even low-mineralized volcanic fluids can be successfully processed using membrane technologies in combination with the developed chemical treatment approach. The processing of these water types for raw material extraction increases the applicability of combinatorial geothermal use enormously, since these high-enthalpy reservoirs supply most of the world’s installed geothermal power plants and thus represent a tremendous source of raw materials.

5  Acknowledgments

The authors would like to thank the Helmholtz Association for research funding within the Geoenergy subtopic in the MTET (Materials and Technologies for the Energy Transition) program of the Energy research field. Further, the BMBF (Federal Ministry of Education and Research) is thanked for funding the BrineMine project (Grant Number 033R190B) in the Client II framework. In addition, they appreciate the support of the Chilean ANID Fondap program (projects 15090013, 15200001 and ACE210005).

They thank Chantal Kotschenreuther, Maya Denker and Claudia Mößner of the Laboratory of Environmental and Raw Materials Analysis (LERA), Chair for Geochemistry & Economic Geology (Karlsruhe Institute of Technology, Division of Applied Geosciences) for the conduction of the analysis and the access to laboratories and equipment, as well as for fruitful discussions. Finally, they want to thank the company Tánica for facilitating the implementation of the demonstrator at their Termas de Puyehue and especially Pablo Sánchez and Ronny Oyarzo for their great effort and support before, during and after the testing phase.