Multistep Reactions of Molten Nitrate Salts and Gas Atmospheres A thesis accepted by the Faculty of Energy-, Process and Bio-Engineering of the University of Stuttgart to fulfil the requirements for the degree of Doctor of Engineering Sciences (Dr.-Ing.) by Julian Steinbrecher born in Garmisch – Partenkirchen Main Referee: Prof. Dr. André Thess Co Referee: PD Dr. Ingo Hartenbach Day of the oral exam: 15.05.2024 Institute for Building Energetics, Thermotechnology and Energy Storage (IGTE) at the University Stuttgart 2024 II Declaration Ich versichere hiermit, dass ich die anliegende Dissertation selbständig verfasst und keine anderen Hilfsmittel und Quellen als die angegebenen benutzt habe. Die Stellen, die anderen Werken (einschließlich des Internets und anderer elektronischer Text- und Datensammlungen) dem Wortlaut oder dem Sinn nach entnommen sind, habe ich in jedem einzelnen Fall durch Angabe der Quelle bzw. der Sekundärliteratur als Entlehnung kenntlich gemacht. __________________________ ________________________ Ort, Datum Unterschrift III Acknowledgements I would like to take the opportunity to thank some of the people who took great part in enabling me to finish this dissertation. I am grateful to Prof. André Thess (University of Stuttgart) for his invaluable guidance, unwavering encouragement in refining my research focus, and his continuous support towards enhancing the quality of my academic work. PD Dr. Ingo Hartenbach (University of Stuttgart), for graciously agreeing to co-referee this thesis. Dr. Alexander Bonk, for his supervision and guidance over the past years, which has been instrumental in shaping this thesis. The countless fruitful discussions, served as a driving force behind the successful completion of this scientific work. To all my colleagues from the group ‘Thermal Systems for Fluids’, especially Dr. Thomas Bauer, Dr. Qing Gong, Dr. Wenjin Ding, Dr. Sebastian Kunkel, Markus Braun Andrea Hanke and Ralf Hoffmann, I express my gratitude for our collaborative efforts and fruitful exchange. Your scientific and technical expertise have been invaluable. Jonina Felbinger, Dr. Marc Linder and many others, who generously dedicated their time to discuss the demanding parts of this thesis. Your patience in finding valuable insights greatly contributed to organize my thoughts. The whole department of Thermal Process Engineering, for fostering a conducive working environment and feedback during numerous doctoral seminars. To my scientific companions, Jochen and Daniel, whose unwavering support throughout my academic journey have been deeply appreciated. My heartfelt thanks to my parents Jutta and Rainer, and my sister Corinna, for their support, open- mindedness, and their invaluable input on seemingly unsolvable issues. Finally, I want to express my deepest gratitude to Lena, for her unwavering emotional support and consistent interest on the progress of this thesis. Her encouragement has been an essential source of motivation throughout this journey. IV Contents Declaration .................................................................................................................................. II Acknowledgements ..................................................................................................................... III List of Abbreviations ..................................................................................................................... V Abstract ...................................................................................................................................... VI Kurzfassung ................................................................................................................................ VII 1 Introduction...................................................................................................................... - 1 - 1.1 Application of Nitrate Salts .............................................................................................. - 2 - 1.2 Scope of the thesis .......................................................................................................... - 6 - 2 Theoretical Background and Experimental Methods .......................................................... - 9 - 2.1 Thermodynamics of a Chemical Reaction ....................................................................... - 9 - 2.2 Reaction kinetics ............................................................................................................ - 10 - 2.3 Experimental Methods .................................................................................................. - 12 - 3 Journal Publication ..........................................................................................................- 15 - 3.1 Paper I ............................................................................................................................ - 15 - 3.2 Paper II ........................................................................................................................... - 26 - 3.3 Paper III .......................................................................................................................... - 37 - 4 Comprehensive Discussion ...............................................................................................- 55 - 4.1 Implications on the Nitrate – Nitrite Equilibrium ......................................................... - 55 - 4.2 Oxide Ion Formation and Equilibrium ........................................................................... - 61 - 5 Future Work and Conclusion ............................................................................................- 64 - 6 Bibliography ....................................................................................................................- 66 - V List of Abbreviations °C degree Celsius g gram h hour min minute s second l litre M molar Mw molar weight wt% weight percentage mol% molar percentage t tonnes rpm rounds per minute MWhth mega Watt hours thermal GWhel giga Watt hours electric n mole c concentration t experiment duration TGA thermogravimetric analysis IC ion chromatography TES thermal energy storage CSP concentrated solar power HTF heat transfer fluid LCOE levelized cost of electricity VI Abstract In the global pursuit of transitioning away from fossil fuels, the need for long-duration energy storage is vital to transform the fluctuating output of renewable energy sources into a reliable and dispatchable electricity supply. Meeting these demands requires technologies with advanced readiness levels, cost-effectiveness, and scalability. Thermal energy storage utilizing molten nitrate salt stands out as a key technology that currently satisfies these requirements. A mixture of NaNO3 and KNO3, commercially known as Solar Salt, is the state-of-the-art sensible heat storage and heat transfer material, currently employed in several concentrated solar power plants over the world. Presently, molten salt is heated to bulk temperatures of 565 °C and utilized to produce steam for electricity generation via traditional Rankine cycles. For superior efficiency and to allow technology transfer to modern Rankine cycles, the salt temperature needs to be increased to above 600 °C. Increasing the temperature of the molten nitrate salt is not readily feasible, because decomposition reactions within the nitrate salt potentially lead to the formation of corrosive oxide ions (O2-, O22-, O2-) and toxic nitrous gases. Stabilization of the nitrate salt at those temperatures is required, which can be accomplished by intelligent gas management techniques. While high oxygen partial pressures can partially decrease salt decomposition, the chemical reactions between molten nitrates, oxygen and/or other reactive gases at temperatures exceeding 600 °C remain not fully understood. This thesis intensively investigates chemical equilibrium reactions of the nitrate salt mixture (Solar Salt) in the temperature range from 500 to 650 °C. Long-term thermodynamic experiments, facilitated by purpose-built autoclave test rigs, were employed to investigate the high temperature induced physicochemical processes of the molten salt solution. This was achieved by taking advantage of the liquid-gas reaction between the molten salt and a suitable atmosphere, enriched with O2 and NOx-gases. The study provided a regeneration pathway for aged, corrosive molten Solar Salt. Shedding light on the impact of congruent reactions, specifically corrosion reactions and oxide ion formation, improved the precision of thermodynamic data for Solar Salt. The kinetics of the oxide ion formation in 100-gram scale experiments were assessed and shown to be slower than previously anticipated. Experimental assessment of different nitrous gas species (N2O, NO, NO2) identified the dominant reaction mechanism. The findings contribute valuable insight into life-time aspects of commercially utilized Solar Salt. Combination of aspects such as gas management, oxide ion impurity control, could allow technically the increase of the operation temperature of Solar Salt beyond the state-of-the-art temperature of 565 °C. This could enlarge the temperature difference of the storage, increase the power block efficiency, and could allow downsizing of power related components due to larger temperature differences. In summary, innovative process strategies could significantly improve the lifetime of Solar Salt, offering substantial potential to reduce capital expenses in Solar Salt heat storage technologies. VII Kurzfassung In dem weltweiten Bestreben, sich unabhängig von fossilen Brennstoffen zu machen, wird die Implementierung langfristiger Energiespeicherung von entscheidender Bedeutung sein, um fluktuierende erneuerbare Energien in eine zuverlässige und regelbare Stromquelle umzuwandeln. Die Erfüllung dieser Anforderungen erfordert Technologien mit einem hohen Reifegrad, Kosteneffizienz und Skalierbarkeit. Die thermische Energiespeicherung auf Basis von geschmolzenem Nitratsalz kann als Schlüsseltechnologie angesehen werden, die all diese Anforderungen heute erfüllt. Eine Mischung aus NaNO3 und KNO3, kommerziell bekannt als Solarsalz, ist der Stand der Technik für sensible Wärmespeicher- und Wärmeübertragungs- Material, das derzeit in mehreren Solarthermie Kraftwerken auf der ganzen Welt eingesetzt wird. Aktuell wird das geschmolzene Salz auf eine Bulk-Temperatur von 565 °C erhitzt und zur Erzeugung von Dampf für die Stromerzeugung über herkömmliche Rankine-Kreisläufe verwendet. Um einen höheren Wirkungsgrad zu erzielen und den Technologietransfer auf moderne Rankine-Zyklen zu ermöglichen, muss die Salztemperatur auf über 600 °C erhöht werden. Eine Erhöhung der Temperatur des geschmolzenen Nitratsalzes ist nicht ohne weiteres möglich, da Zersetzungsreaktionen innerhalb des Nitratsalzes möglicherweise zur Bildung von korrosiven Oxidionen (O2-, O22-, O2-) und toxischen stickoxidhaltigen Gasen führen. Eine Stabilisierung des Nitratsalzes bei diesen Temperaturen ist erforderlich, was durch intelligente Gasmanagementtechniken erreicht werden kann. Während hohe Sauerstoffpartialdrücke den Salzabbau teilweise verringern können, bleiben die chemischen Reaktionen zwischen geschmolzenen Nitraten, Sauerstoff und/oder anderen reaktiven Gasen bei Temperaturen über 600 °C noch nicht vollständig verstanden. In dieser Arbeit werden chemische Gleichgewichtsreaktionen des Nitratsalzgemisches (Solarsalz) im Temperaturbereich von 500 bis 650 °C intensiv untersucht. Um die bei hohen Temperaturen induzierten physikalisch-chemischen Prozesse der Salzschmelze zu untersuchen, wurden eigens entwickelte Autoklaven-Prüfstände für thermodynamische Langzeitexperimente verwendet. Dies wurde erreicht, indem die Flüssig-Gas-Reaktionen zwischen dem geschmolzenen Salz und einer geeigneten Atmosphäre, angereichert mit O2 und NOx-Gasen, ausgenutzt wurden. Dabei wurde eine Regenerationsroute für gealtertes, korrosives geschmolzenes Solarsalz identifiziert. Umfangreiche Erkenntnisse über den Einfluss kongruenter Reaktionen (Korrosionsreaktionen und Oxidionenbildung) auf die Genauigkeit thermodynamischer Daten für Solarsalz wurden vorgestellt und experimentell nachgewiesen. Die Kinetik der Oxidionenbildung in Experimenten im 100-Gramm-Maßstab wurde untersucht und erwies sich als langsamer als bisher angenommen. Der Einfluss verschiedener Nitrose Gase (N2O, NO, NO2) wurde experimentell untersucht und der dominante Reaktionsmechanismus identifiziert. Die Ergebnisse dieser Arbeit geben einen Einblick in die Lebensdaueraspekte von kommerziell genutztem Solarsalz. Die VIII Kombination von Aspekten wie Gasmanagement und Kontrolle der Oxidionenverunreinigung könnte technisch die Erhöhung der Betriebstemperatur von Solarsalz über den Stand der Technik von 565 °C hinaus ermöglichen. Dies könnte den Temperaturunterschied des Speichers vergrößern, den Wirkungsgrad des Leistungsblocks erhöhen und eine Verkleinerung von leistungsbezogenen Komponenten aufgrund größerer Temperaturunterschiede ermöglichen. Zusammenfassend könnten innovative Verfahrensstrategien die Lebensdauer von Solar Salz erheblich verbessern und bieten ein großes Potenzial zur Reduzierung der Kapitalausgaben bei Wärmespeichertechnologien mit Solar Salz. Introduction - 1 - 1 Introduction The green energy transition is pushing the demand for energy storage technologies. Successful implementation of volatile renewable energy sources to reduce world-wide CO2 emissions has become a key challenge and requires the efficient use of storage technologies (e.g. thermal, electrochemical methods). Daily fluctuation of renewable power sources (e.g. solar, wind) makes integration into the market with a demand-based energy request challenging. Upon the different technologies, power generation using integrated molten salt thermal energy storage (TES) can offer flexibility and dispatchability for the electrical grid.[1] At the time of writing, molten salt storage is well developed and commercially used in modern concentrated solar power (CSP) plants. Worldwide, a total of 94 CSP projects achieved commercial online operation till the end of 2018 and most of them include a TES unit with 2.5 to 15 h storage duration.[2] In the year 2019, TES with a capacity of 21 GWhel has been installed worldwide.[1] According to IRENA’s (international renewable energy agency) technology outlook, the construction of at least 74 GWhel additional capacity is planned until the end of 2030. However, their Paris aligned transforming energy scenario (accelerated deployment of CSP and TES, adequate policies) and predicts a higher growth of molten salt TES between 491 to 631 GWhel until the end of 2030.[3] In the state-of-the-art solar tower CSP plant, a salt mixture of 60 wt% NaNO3 and 40 wt% KNO3 (commercial name “Solar Salt”) is utilized as heat transfer fluid (HTF) as well as TES material. The TES system is built up by a two-tank configuration containing hot and cold salt and a heat exchange as shown in Figure 1.1. Figure 1.1 CSP plant scheme with molten salt power tower with direct heat storage. Solar Salt as HTF and storage medium is restricted to operating temperatures from 290 °C to 565 °C. (reprinted and modified with permission from NREL [4]) Introduction - 2 - For power generation, solar radiation is concentrated on the receiver system, where Solar Salt is continuously heated and then pumped into the hot tank. The state-of-the-art temperature limit of this process is approximately 565 °C. Thermal energy storage in the hot salt is subsequently used to generate superheated steam in the heat exchanger unit, which is fed into a Rankine type power block for electricity production (Fig. 1.1). The size of the molten salt storage system can be designed towards the anticipated capacity of the plant but is typically in the range of 1-4 GWhth for modern CSP tower systems. The CSP system can thereby generate dispatchable electricity during day and/or night time (e.g. NOOR3 solar tower plant 134 MWhel, Marocco).[1] To push the efficiency of this technology, it was shown that increasing the upper temperature limit of the molten salt to 620 °C or above, could lead to a significant cost (levelized cost of electricity, LCOE) reduction during the operation of the CSP power block.[5] A different study further presented the benefits of utilizing a supercritical carbon dioxide Brayton power cycle at 600 °C to 650 °C, which would be possible for hot salt temperatures above 620 °C.[6] Yet, the thermal stability of Solar Salt is currently the limiting factor stopping TES technology from increasing the hot tank temperature.[7] However, recent studies[8, 9] revealed that the composition of the gas phase play a crucial role stabilising the molten salt and that proper gas management in fact could extent the upper temperature limit for Solar Salt way beyond 600 °C. The upcoming chapters will give a more detailed summary on the state-of-the-art nitrate salt chemistry and potential research questions, which will be answered in this thesis. 1.1 Application of Nitrate Salts Apart from decomposition studies on the single salts of sodium nitrate and potassium nitrate and the Solar two project (2002, CSP tower plant, US)[10], the investigation of the Solar Salt composition has just recently been intensified.[1, 11, 12] The following subsection, will review the relevant studies on nitrate salts as well as some mixtures. 1.1.1 Thermal degradation Due to the molecular structure of the nitrate anion, the N-O bonds will break eventually, if a certain (thermal) energy level is surpassed. The thermal stability of the nitrate ion is strongly dependent on the present cation in the salt.[13-16] For alkali metal and earth alkali nitrate salts the thermal stability of the salt increases with the size of the cation. Intermediate metal nitrate salts (e.g. Fe(NO3)3, Ni(NO3)2) are in general less stable, due to a weaker bond between the ions, compared to the lower valent alkali metals (e.g. Na, K).[17] Additionally, it has been demonstrated, that Introduction - 3 - impurities like, Mg2+, Ca2+, SiO2, H2O, Fen+, Cl-, SO42- in the molten salt accelerate the degradation of the nitrate ion.[18-20] In this thesis, the Solar Salt mixture (60 wt% NaNO3 and 40 wt% KNO3) and mixtures with addition of nitrite salt and/or oxide salt were exclusively used. To maintain the cation impact on the thermal stability the cation ratio was adjusted each time according to the one of Solar Salt and is not part of the discussion in this thesis. The decomposition behaviour of Solar Salt in terms of anions has been studied to some extent and the adequate thermal stability was proposed to be at or below 600 °C in air.[21-23] With the efforts of Sandia national laboratories and co-workers, state-of-the-art molten salt storage was found to work well at a temperature of 565 °C. [4] One reason for this limitation, is the breakdown of the nitrate ion with increasing temperature to form a nitrite ion accompanied by the release of oxygen (see section 1.1.3, Eq. (N)).[24, 25] Recent studies of the DLR group and other authors showed that the thermal stability increases if the overlaying atmosphere contains increasing partial pressures of oxygen.[8, 26-28] This behaviour, was explained by the oxidation of nitrite ion and the establishment of chemical equilibrium between the nitrate ion and the nitrite ion.[14, 25, 29] At temperatures above 600 °C, the nitrite ion is degrading to form corrosive oxide ions and releases several toxic gases (e.g. NO, NO2, N2O) as well as N2, which was demonstrated in independent studies.[27, 30-32] Despite numerous studies investigating the decomposition mechanisms of the nitrite ion [30, 33-38], there is no agreement on a single reaction mechanism, but it seems likely that a mixture of oxide (O2-), peroxide (O22-) and superoxide (O2-) ions is present in molten Solar Salt. For simplicity, the three different oxide species, will be summarized and expressed as “oxide ion (O2-)”, hereafter. 1.1.2 Corrosion in molten nitrate salt One crucial point stopping nitrate salt-based TES from moving to temperatures above 565 °C is the increased corrosivity of the molten Solar Salt.[39] In this context the corrosivity understood as measure of the ability of molten salt to induces corrosive degradation of metal construction materials.[40] The enhanced level of corrosive ionic species which can form during thermal decomposition, are expected to aggravate corrosion at temperatures above 565 °C. Decades ago the investigation from Slusser et al. already proposed the negative effect of alkali oxides on the durability of Fe-Ni-Cr alloys up to 705 °C.[41] More recently, Kruizenga et. al. found a drastic increase of the corrosion rate with a temperature increase from 400 to 680 °C of several different alloys immersed in Solar Salt.[7, 42] Until now, despite the publication of numerous corrosion studies with a variety of alloys, the corrosion mechanism of the most relevant 300-type steels (e.g. AISI 347, 316 or 321)is not fully understood.[12, 22, 43-46] However, for the mentioned steel types scanning electron micrographs and X-ray diffraction patterns revealed an outer layer of hematite (Fe2O3) and magnetite (Fe3O4) followed by an inner layer of Fe-Cr spinel or presumably Introduction - 4 - pure Cr2O3 can typically be detected after immersion in Solar Salt at 600 °C. It is also known that pure chromium compounds react with the molten nitrate salt and completely dissolve into chromate ions (CrO42-) even after only hours of exposure.[47, 48] Bonk et al. claims, the protectiveness of the formed oxide layer is based on the degree of diffusion of oxygen/oxide ions towards the metal, as well as diffusion and eventually dissolution kinetics of the metal components towards the molten salt.[49] In this context, the oxide ion, which is formed by the degradation of the nitrite ion, is expected to accelerate certain corrosion processes.[30] The detection of chromate ions in the melt is understood as an indicator for corrosion reactions in molten salt.[50] Beside the formation of metal oxide, the evolution of significant amounts of NOx (NO, NO2) gases was detected during several corrosion studies.[51, 52] Peng et al. suggested a closed system configuration and proposed an increased thermal stability of molten salt when reactive gases (O2, NOx) are retained.[53] This approach was experimentally realized by Bonk et. al, who presented a reduced corrosivity of Solar Salt with atmospheres containing O2 compared to a higher corrosivity under pure N2 atmosphere.[47] These findings indicated that stabilisation of the salt, might be the key to solve the corrosion issues for molten salt application. 1.1.3 Stabilisation of Nitrate Salts with Reactive Gases Large scale application of molten Solar Salt requires knowledge transfer of decomposition studies by elaborating data from microbalance experiments. Investigation of the chemical equilibrium, where the salt composition basically remains unchanged and stable under the respective conditions, is a necessary task to be done So far, it can be agreed upon, that the decomposition of the nitrate ion follows a two-step mechanism. In the primary decomposition reaction, the nitrate ion will break down to give a nitrite ion under the evolution of oxygen gas as written in Eq. (N) (“N” for nitrite ion). 𝑁𝑁𝑁𝑁3 − ⇌ 𝑁𝑁𝑁𝑁2 − + 0.5𝑁𝑁2 (N) Equation (N) is reversible if O2 gas is in contact to the molten salt and will establish a stable equilibrium with respect to the partial pressure of oxygen and the temperature.[29, 54, 55] For the successful equilibration according to Eq. (N), it is assumed that the dissolution reaction of O2 via the surface inside the melt is faster than the oxidation of the nitrite ions. This theory was discussed recently by Sötz et al. who studied the kinetics of Eq. (N). In the study of Desimoni et al., the solubility of O2 in eutectic (Na, K)NO3 was measured to be at a ppb-level (at 250 °C), slowly increasing with temperature.[56] Thus, the concentration of O2 dissolved in Solar Salt at 600 °C is expected to be low and the assumption of fast saturation is reasonable.[57] Limitations of this assumption were found in the kinetic investigation of the nitrite oxidation (reverse direction Eq. (N)) by Nissen and Meeker et al., who found an effect of O2 mass transport nitrite oxidation and only vigorous stirring of the molten salt lead to reproduceable data.[29] In general it is accepted, Introduction - 5 - that the Eq. (N) is described with the equilibrium constant KN according to Eq. (1), taking the above-mentioned approximation into account. KN = [𝑁𝑁𝑁𝑁2 −]𝑒𝑒𝑒𝑒 [𝑁𝑁𝑁𝑁3 −]𝑒𝑒𝑒𝑒 ∗ 𝑃𝑃𝑂𝑂2 0.5 (1) With a constant partial pressure of oxygen (𝑃𝑃𝑂𝑂2 0.5), and the equilibrium content of nitrate ([𝑁𝑁𝑁𝑁3 −]𝑒𝑒𝑒𝑒) and nitrite ([𝑁𝑁𝑁𝑁2 −]𝑒𝑒𝑒𝑒) ions, the equilibrium constant 𝐾𝐾𝑁𝑁 can be calculated from the measured data. There are several individual studies available, which comprise thermodynamic data on Eq. (N) and a review can be found in the article of Bauer et. al..[39] However, few researchers have addressed the issue regarding the above mentioned mass transport limitation of the oxidation reaction. Additionally, most of the experiments performed were too short (< 300 h) and mostly below 600 °C, which raises the question if full equilibration of the salt system was achieved. To the knowledge of the author, there is no literature report of a stable Solar Salt up to 650 °C, which will be one part of this thesis. The secondary decomposition reaction of Solar Salt is the degradation of the nitrite ion. The thermal stability of the nitrite ion is principally lower compared to the nitrate ion. In pure form decomposition will start above 330 °C (for NaNO2).[17] In molten Solar Salt or in other words, when the nitrite ion is embedded in a nitrate-rich ionic matrix, several studies found the degradation of the nitrite ion to become significant only above 600 °C.[58] A commonly cited decomposition mechanism for the nitrite ion is shown in Eq. (O) (“O” for oxide ion). 2𝑁𝑁𝑁𝑁2 − ⇌ 𝑁𝑁2− + 𝑁𝑁𝑁𝑁 + 𝑁𝑁𝑁𝑁2 (O) The formation of oxide ion species is accompanied by the evolution of nitrous gases, which in the presence of oxygen will react according to their temperature dependent gas-gas equilibrium (Eq.(2)).[59, 60] 2𝑁𝑁𝑁𝑁 + 𝑁𝑁2 ⇌ 2𝑁𝑁𝑁𝑁2 (2) Based on Eq. (N) and Eq. (O) Bonk et al. presented an increased thermal stability of Solar Salt by sealing the reaction chamber at 600 °C and retaining reactive gases.[9] Sötz et. al. was to the first to report the stabilization of molten Solar Salt at 600 °C and 620 °C with reactive purge gases.[8] In the study, stable anion composition over several hundred hours indicated the full equilibration of the salt with respect to both Eq. (N) and Eq. (O). A theoretical enthalpy-based calculation (listed data from Barin et al. [61]) resulted in a modified equilibrium description for the second decomposition step (Eq. (O)). The modified equation (Eq. (3)) of Sötz contained nitrogen, which can be formed during internal oxidation of nitrite to nitrate.[17, 27, 62] 2 𝑁𝑁𝑁𝑁2 − ⇌ 𝑁𝑁2− + 0.5 𝑁𝑁𝑁𝑁 + 1.25 𝑁𝑁2 + 0.75 𝑁𝑁2 (3) However, Stern et al. stated that nitrogen is too stable to react any further at temperatures below 1000 K. Until know, it remains unclear whether Nitrogen takes part in the stabilisation of Solar Introduction - 6 - Salt and experimental proof was pending. It can be stated that the experimental data situation for the formation of oxide ions in Solar Salt is poor and needs to be extended in order to get a better understanding on the entire high temperature reaction chemistry of the salt. 1.2 Scope of the thesis This thesis presents a variety of experimental results and theoretical considerations, investigating the thermal stability of Solar Salt at temperatures up to 650 °C. The influence of reactive gas compositions on the equilibration process in the molten Solar Salt at different temperatures is of particular interest. From a theoretical point of view, thermodynamic equilibrium would only be possible for a closed and completely isothermal system, which seldomly is realised in a practical application. However, if available thermodynamic data is transferrable to any Solar Salt based study under identical conditions, it is of great value for this research field. Full stabilization of Solar Salt, in terms of composition with respect to nitrate, nitrite and oxide ions, would be close to an equilibrium state. Major tasks of this dissertation focus on providing solutions for establishing a stable molten Solar Salt in temperature regimes above present state. According to recent findings, the corrosivity of the molten salt is closely related to the high temperature salt decomposition reactions, especially the formation of oxide ions in the molten salt. Even small amounts (ppm-level concentration) of oxide ions present could result in sever corrosion attack of construction materials of a TES unit. For this reason, investigation of the oxide ion formation mechanism and the related chemical equilibration is a major aim of all the described experiments. Though, an accurate determination of the ppm-level oxide ion amount is difficult with state of the art detection methods. Hence, it is a challenging task to generate accurate and reproducible results. In the work of Sötz et al. [8] the oxide ion formation (Eq, (O)) was successfully supressed by a constant purge of nitrous gases over the molten salt. However, it remains unclear if the experimental duration was chosen appropriately and if the resulting steady state salt composition corresponds to the equilibrium composition. Based on the kinetic description of the chemical equilibrium (see Section 2.2, Eq. (14)) the reverse direction of reaction (N) and (O) are faster compared to the forward direction (if [𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸]𝐸𝐸𝑒𝑒 > [𝑃𝑃𝑃𝑃𝑃𝑃𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸]𝐸𝐸𝑒𝑒). Thus, approaching the equilibrium from the product side, will lead to a faster equilibration of Eq. (N) and (O). The method of reversing Eq. (N) and Eq. (O) will be referred to as “regeneration” and was applied in Paper I, by intentionally adding an excess amount of nitrite and oxide ions (defined as “synthetically aged”) to the molten salt and a purge gas of reactive gases (O2 and NO). Especially at higher operating temperatures, the possibility of salt decomposition becomes more likely and active control of the oxide forming reaction Eq. (O) has not been addressed sufficiently in literature. The results of this approach are considered the first to present an active regeneration mechanism of a synthetically aged Solar Salt towards a fully equilibrated molten salt. Additionally, Introduction - 7 - experimental data is delivered, which would support the dominant occurrence of equilibrium reaction (O) for the regeneration of Solar Salt, beside many other suggested mechanisms. The second decomposition reaction (O) is inevitably connected with the primary equilibrium of the nitrate ion (N). Consequently, before the investigation of the oxide ion formation is proceeding, a continuous thermodynamic data set of the nitrate – nitrite equilibrium needs to be available. Despite, some studies on this topic are available, they either focused on a narrow temperature regime or were performed with an experiment configuration, which might have affected the resulting thermodynamic data. That is why, Paper II will reassess the thermodynamic data for Solar Salt from 500 to 650 °C and review the difference to literature findings. Especially, the data of the intrinsic kinetic study of the oxide ion formation reaction were of particular interest. It is suggested that the apparent kinetics present in a 100-gram scale system are closer to a TES system than previous data and they are reported in Paper II for temperatures 560 to 650 °C. In addition, Paper II shows the beneficial effect of retaining the decomposition gas inside the molten salt container on the stability of the salt at 650 °C. This so called “closed system” operation mode could be realized in a TES unit at elevated temperatures, which has been shown to result in an increased thermal stability of Solar Salt (up to 600 °C) compared to the open system configuration. This result is explained by pressure built up of reactive decomposition gases, which stabilize the equilibrium reactions (N) and (O). But a pressurized system is costly and not desired at an industrial scale. For that reason, Paper II presents the possibility to control pressure built up in the closed system configuration, by simultaneously maintaining the stability of the molten salt. It is known that reactions between metal components and Solar Salt alter the salt composition (e.g. Paper I). Also, the degradation of nitrite (Eq. (O)) at elevated temperatures will most likely affect the position of nitrate – nitrite equilibrium (suggested in Paper II). Interaction of these processes are poorly understood and conclusive experiments providing data with sufficient accuracy are lacking so far. The study in Paper III was set out to present a data set for the nitrate – nitrite equilibrium data of unique accuracy up to 620 °C. For the first time, the effect on several nitrous gases (NO, NO2, N2O) as well as N2 on the oxide ion producing reaction was investigated to identify the most relevant reactions for the full stabilization of Solar Salt. A comprehensive summery of the specific aim of each publication is given in Figure 1.2, with the intention to visualized differences and the relations between the individual papers. Introduction - 8 - Figure 1.2: Nitrate- nitrite- equilibrium reaction and nitrite- oxide- equilibrium reaction. Investigated temperature range covered by the individual papers, as well as the corresponding aim of the study. Impact of different partial pressures and experimental configuration on the equilibrium reactions is listed on the right. Theoretical Background and Experimental Methods - 9 - 2 Theoretical Background and Experimental Methods 2.1 Thermodynamics of a Chemical Reaction 2.1.1 Fundamentals of Thermodynamics With the help of classic thermodynamics, a system is described in its macroscopic state. State properties are given by so called state functions, which themselves are categorized into intensive and extensive variables.[61] Extensive properties are volume (V), internal energy (U), enthalpy (H), entropy (S), Helmholtz energy (A) and the Gibbs free energy(G), which dependent on the amount of material in the system. Intensive properties on the other hand are independent of the material amount or will not change with the system size, e.g. temperature (T) or pressure (p). Gibbs free energy combines the enthalpy with the temperature dependent entropy term according to Eq. (4). The change of the Gibbs free energy indicates for the system if the reaction will happen spontaneous (Δ𝐺𝐺 < 0) or it requires energy for the reaction to happen (Δ𝐺𝐺 > 0). At thermodynamic equilibrium (Δ𝐺𝐺 = 0) the System is in a stationary state and from a macroscopic view no mass or energy flow is observed Eq. (5). [63] Δ𝐺𝐺 = Δ𝐻𝐻 − 𝑇𝑇Δ𝑆𝑆 (4) Δ𝐺𝐺 = 0, 𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝑃𝑃𝐸𝐸𝐸𝐸𝐸𝐸 (5) From an energetic perspective, the most stable configuration for a chemical system is the equilibrium state. Assuming a system at constant temperature and pressure, at equilibrium the Gibbs free energy (G) is at its minimum value. [17] 2.1.2 Chemical equilibrium Chemical reactions are most of the time written to follow the forward direction only. With this description it is assumed the reverse reaction is not taking place at a considerable rate. In many cases writing reactions only in a single direction is reasonable, if the reverse reaction rate is so small that its influence on the forward direction can be neglected. For the reaction of reactant A to form reactant B an accurate description is given with an equilibrium arrow as well as the stoichiometry coefficients α and 𝛽𝛽 (𝛼𝛼 = 𝛽𝛽 = 1 for Eq. (6)) 𝛼𝛼A ⇌ 𝛽𝛽B (6) The most favourable state for every chemical reaction is the state of lowest Gibbs free energy, which is defined as chemical equilibrium. The progress of the chemical reaction is described with the reaction quotient (Q), which is the ratio between products and reactants (Eq. (7)). At chemical equilibrium the reactant concentration within the system remains unchanged and the reaction quotient (Q) is constant (Eq.(7)). Theoretical Background and Experimental Methods - 10 - 𝑄𝑄 = [B]𝛽𝛽 [𝐴𝐴]𝛼𝛼 = const. (7) The Gibbs free energy change or reaction Gibbs energy (Δ𝐺𝐺𝑅𝑅) is the sum of the standard Gibbs energies (Δ𝐺𝐺𝑅𝑅 ° ) at standard conditions with a pressure of 1 bar and a temperature of 298.15 K. The progress of the reaction (Q) is thereby linked to the free energy change of the system via Eq.(8). Δ𝐺𝐺𝑅𝑅 = Δ𝐺𝐺𝑅𝑅 ° + 𝑅𝑅𝑇𝑇ln𝑄𝑄 (8) At thermodynamic equilibrium (Δ𝐺𝐺𝑅𝑅 = 0), the Gibbs free energy is minimized. With Δ𝐺𝐺𝑅𝑅 = 0, Eq.(8) can be rearranged and the reaction quotient substituted by the equilibrium constant (K). This leads to Eq.(9), where free enthalpy and composition of the system are correlated with the temperature. Δ𝐺𝐺𝑅𝑅 ° = −𝑅𝑅𝑇𝑇ln𝐾𝐾 (9) In practice, the equilibrium constant for reversible reaction can be measured at different temperatures. The van’t Hoff equation (Eq.(10)), is used to relate the thermodynamic entities standard reaction enthalpy (Δ𝐻𝐻𝑅𝑅 ° ) and standard reaction entropy (Δ𝑆𝑆𝑅𝑅 ° ) with the equilibrium Composition of the system. ln𝐾𝐾 = − Δ𝐻𝐻𝑅𝑅 ° 𝑅𝑅𝑇𝑇 + Δ𝑆𝑆𝑅𝑅 ° 𝑅𝑅 (10) Plotting the values of lnK against the inverse temperature, should result in plot with a linear relationship, the so call van’t Hoff plot. According to Eq. (10), the slope of the linear regression between the induvial lnK-values corresponds to the reaction enthalpy (−Δ𝐻𝐻𝑅𝑅 ° /𝑅𝑅) and the y- intercept to the entropy term (Δ𝑆𝑆𝑅𝑅 ° /𝑅𝑅). For many processes, this approach gives sufficient results and predictions about equilibrium of the reactant mixture at different temperatures. But often the thermodynamic equilibrium state is not accessible because of kinetic limitations (e.g. mass transport or diffusion), which will be introduced in the next section. 2.2 Reaction kinetics In contrast to the thermodynamic equilibrium, the chemical kinetics give an answer to the speed of a reaction and the underlying mechanism for the product formation. For the model reaction (Eq.(6)) the rate of the forward reaction is defined as the amount of A reacting to B over time. It is defined as a first order reaction with respect to A and leads to the following rate law Eq.(11): Rate = 𝐸𝐸[A] 𝐸𝐸𝐸𝐸 = 𝑘𝑘[A] (11) Theoretical Background and Experimental Methods - 11 - The temperature dependent kinetic constant k with the unit per second (s-1) is characteristic for this reaction and independent of the reactant concentration. This description is only valid for the forward reaction. For complete description of equilibrium reaction Eq. (6), it can be rewritten into Eq.(12). In this case, forward and revers reaction are assumed to occur with the same rate law Eq.(13) and the kinetic constants k1 (forward) and k-1 (reverse). A ⇌𝑘𝑘−1 𝑘𝑘1 B ; (12) 𝑅𝑅𝑅𝑅𝐸𝐸𝑒𝑒𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 = 𝑘𝑘1[A] ; 𝑅𝑅𝑅𝑅𝐸𝐸𝑒𝑒𝑓𝑓𝑒𝑒𝑟𝑟𝑒𝑒𝑓𝑓𝑟𝑟𝑒𝑒 = 𝑘𝑘−1[B] (13) Because forward and revers reaction rate are both depending on the reactant concentration A (forward) and B (reverse), the reaction rate will decrease over time and eventually reach a steady state. A chemical system in the state, where educt concentration ([A]eq) and product concentration ([B]eq) remain constant over time is defined as equilibrium state. The ratio of forward and revers rate constant thereby is directly connected with the thermodynamic equilibrium constant K (Eq. (14)). K = 𝑘𝑘1 𝑘𝑘−1 = [B]𝑒𝑒𝑒𝑒 [A]𝑒𝑒𝑒𝑒 (14) Many chemical processes tend to be of higher reaction order (multi molecular reactions) and simultaneous reactions make an evaluation complex. However, two approximations are available which allow for an accurate description of first order rate law for many kinetic problems. With the help of the rate determining step or rate limiting step approximation, a set of reactions is reduced to a single reactions step. As an example, the sequential reaction (Eq.(15)) of A to B and B to C is given: A 𝑘𝑘1→ B 𝑘𝑘2→ C (15) This approximation is appropriate, if the rate constant of the first reaction is significantly smaller (20-fold; k1< 99%) and KNO3 (Merck, purity > 99%). If reactions with oxide and nitrite ions are investigated, Na2O2 (Merck, purity > 95%) prepared in a moisture-free atmosphere and/or NaNO2 (Merck, Darmstadt, Hesse, Germany, purity > 99%) are additionally added to the salt mixture. To prevent an impact of the cation content on the thermal stability of the nitrate ion, the concentration is fixed to the composition of the Solar Salt (Na 65.5 mol%; K 35.5 mol%, equivalent to 60 wt% and 40 wt%) for all experiments. To fix the atmospheric composition of the open system configuration experiments, the salt is purged with a constant gas flow (100 ml/min) adjusted by flow meters (Bronkhorst, EL-Flow). To avoid concentration gradients in the molten salt, every experiment is stirred at 60 rpm (or 0, or 120 rpm), if not mentioned otherwise. After melting the salt mixture at 300 °C under a flow of synthetic air (5.0 grade, Linde Gas), the flange for manual sample extraction or the auto sampling unit (Figure 2.1, (b)) is mounted onto the autoclave test rig. Subsequently, the desired purge gas flow is adjusted from mixing O2, N2, and/or a reference gas containing 1000 to 2000 ppm NO (NO2, N2O) in N2 (all 5.0 grade, Linde gas). The concentrations of O2 (vol%), CO2 (ppm), NO2 (ppm) and NO (ppm) in the exhaust gas are continuously measured with an Emerson X-SREAM analyser (Model XEGP). Time zero (t = 0 h) of every experiment is set, when the molten salt reaches the target temperature (500 - 650 °C). Isothermal conditions are kept throughout the experiment and controlled with a maximum deviation of ±4 °C. Over the course of the experiments (approx. 800 – 3000 h), salt samples are extracted and analysed by wet-chemical methods to monitor salt composition. 2.3.2 Salt Analysis Extracted molten salt samples are analyzed using ion chromatography (883 Basic IC plus, Metrohm, Switzerland) and acid base titration (Titrando 905, Metrohm, Switzerland) to yield concentrations of nitrate-, nitrite-, (chromate) and oxide anions. The eluent used in this work for cations (K+, Na+) was 2mM HNO3 with 0.3 mM oxalic acid, for anions it was a solution of 4 mM Na2CO3 with 1 mM NaHCO3 and a suppression liquid of 200 mM phosphoric acid. For every sample about 125 mg salt is dissolved in ultra-pure water (conductivity < 0.055 µScm-1) generated using Theoretical Background and Experimental Methods - 14 - an Astacus² Analytical system (MembraPure, Germany). Subsequently, the solutions are measured automatically by ion chromatography. The oxide ion contents are determined by automated acid base titration under an inert purge gas of N2 (5.0 grade, Linde gas). In general, oxide ions are referred to as O2-, regardless of the actual present oxide ion species (O2-, O2-, O22-) About 500 mg of salt is dissolved in 160 mL CO2 free ultra-pure water and titrated with 0.01M HCl (Titrisol standard) solution. Calibration is done multiple time with a known amount of pre-dried Na2CO3 (Merck, Germany, purity > 99.5%). Oxide ion concentration is calculated from the measured amount of hydroxide (2 OH- convert to 1 O2-or O22- ). The limit of detection (LOD) for oxide ions, represented as O2- in this work, with regard to salt composition was below 0.01 mol% (20 ppm). Equilibrium levels are calculated from anion content after the reaction quotient (Q , ratio between educt and product Eq.(7)) was stable. For NO3- and NO2- (Eq. (N)) this is the case after approximately 300 hours and temperature above 550 °C. Oxide formation rates are often afflicted with a larger error (typically +/- 15 %), especially if corrosion reactions with metal components cannot be excluded [64]. Stable concentration of oxide ions are measured in experiments with atmospheres containing nitrous gases. In the temperature range from 600 °C to 620 °C, the O2- equilibrium (Eq. (O)) takes approximately 500 hours to be attained. After the respective time the median of the remaining data points is calculated and displayed as horizontal lines in the graphs, which is defined as equilibrium composition. Journal Publication - 15 - 3 Journal Publication 3.1 Paper I Investigation of Regeneration Mechanisms of Aged Solar Salt Julian Steinbrecher, Alexander Bonk, Veronika Anna Sötz and Thomas Bauer This article was published in Materials, 2021, 14, 5664. Copyright MDPI (2021). https://doi.org/10.3390/ma14195664 https://doi.org/10.3390/ma14195664 Journal Publication - 16 - Individual contributions in Paper I: • Conceptualization: In the paper the research goal was specified by the lead author, which is the demonstration of regeneration mechanism of an artificially aged molten Solar Salt. This goal was set and achieved with a reactive NOx purge gas atmosphere over the molten salt. • Selection of the parameter variation and execution of the experiments. • Data collection: The lead author was responsible for the experimental supervision as well as the collection of salt samples. Subsequently, post analysis of salt samples and data evaluation were performed. Error calculation for the oxide ion titration method was adapted by the lead author. • Visualization and deductive reasoning: Selection of the relevant data and graphical illustration of the data was categorized and prepared in a clear way. The outline and structuring of the research data were done together. Context between state-of-the-art knowledge was drawn and novelty of the work was developed by the lead author. • Writing and publishing the manuscript in materials by the lead author. materials Article Investigation of Regeneration Mechanisms of Aged Solar Salt Julian Steinbrecher 1,* , Alexander Bonk 1 , Veronika Anna Sötz 2,3 and Thomas Bauer 2 ���������� ������� Citation: Steinbrecher, J.; Bonk, A.; Sötz, V.A.; Bauer, T. Investigation of Regeneration Mechanisms of Aged Solar Salt. Materials 2021, 14, 5664. https://doi.org/10.3390/ma14195664 Academic Editors: Francisco de Paula Montero Chacón and Juan Carlos Serrano-Ruiz Received: 27 August 2021 Accepted: 25 September 2021 Published: 29 September 2021 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affil- iations. Copyright: © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). 1 Institute of Engineering Thermodynamics, German Aerospace Center (DLR), D-70569 Stuttgart, Germany; Alexander.Bonk@dlr.de 2 Institute of Engineering Thermodynamics, German Aerospace Center (DLR), D-51147 Cologne, Germany; veronika.soetz@th-deg.de (V.A.S.); Thomas.Bauer@dlr.de (T.B.) 3 Research Center Modern Mobility, Technology Campus Plattling, D-94447 Plattling, Germany * Correspondence: julian.steinbrecher@dlr.de Abstract: The scope of our study was to examine the potential of regeneration mechanisms of an aged molten Solar Salt (nitrite, oxide impurity) by utilization of reactive gas species (nitrous gases, oxygen). Initially, aging of Solar Salt (60 wt% NaNO3, 40 wt% KNO3) was mimicked by supplementing the decomposition products, sodium nitrite and sodium peroxide, to the nitrate salt mixture. The impact of different reactive purge gas compositions on the regeneration of Solar Salt was elaborated. Purging the molten salt with a synthetic air (p(O2) = 0.2 atm) gas stream containing NO (200 ppm), the oxide ion concentration was effectively reduced. Increasing the oxygen partial pressure (p(O2) = 0.8 atm, 200 ppm NO) resulted in even lower oxide ion equilibrium concentrations. To our knowledge, this investigation is the first to present evidence of the regeneration of an oxide rich molten Solar Salt, and reveals the huge impact of reactive gases on Solar Salt reaction chemistry. Keywords: thermal energy storage; concentrating solar power (CSP); molten nitrate salt; thermal stability; liquid-gas reactions 1. Introduction Thermal energy storage with molten nitrate salts at 565 ◦C is currently employed in several Concentrating Solar Power (CSP) plants to provide dispatchable and renewable electricity in the MW-scale. To increase the solar-to-power conversion efficiency, elevating the maximum operating temperature is one possible strategy [1,2]. For a nitrate based molten salt storage system, chemical stability is challenged at temperatures beyond 565 ◦C and thus effort needs to be made to maintain thermal properties of this sensible heat storage material. The decomposition and equilibrium reactions of a nitrate-based molten salt has extensively been investigated over the past decades and a variety of decomposition mechanisms are reported [3–6]. It is widely accepted that in a first step nitrate ions (NO3 −) compose to form nitrite ions (NO2 −) under the release of oxygen (Equation (1)). NO− 3 NO− 2 + 1/2O2. (1) In the second step, the nitrite ion further decomposes to form oxide ions. A reaction, most commonly referred to as a possible decomposition mechanism for nitrite, is the formation of metal oxide, accompanied by the evolution of NO and NO2 (Equation (2)). 2NO− 2 O2− + NO + NO2. (2) Beside these ions, the evolution of several reactive gas species, namely oxygen, ni- trogen and nitrous oxide (N2O) has been discovered [7]. It has to be emphasized that the exact stoichiometry of Equation (2) has not been determined experimentally and may also depend on atmospheric composition. Additionally, at different temperatures, other reaction paths may become favorable [8,9]. Nevertheless, the influence of nitrous gases on Materials 2021, 14, 5664. https://doi.org/10.3390/ma14195664 https://www.mdpi.com/journal/materials https://www.mdpi.com/journal/materials https://www.mdpi.com https://orcid.org/0000-0003-1692-7487 https://orcid.org/0000-0002-0676-7267 https://orcid.org/0000-0003-4080-7944 https://doi.org/10.3390/ma14195664 https://doi.org/10.3390/ma14195664 https://creativecommons.org/ https://creativecommons.org/licenses/by/4.0/ https://creativecommons.org/licenses/by/4.0/ https://doi.org/10.3390/ma14195664 https://www.mdpi.com/journal/materials https://www.mdpi.com/article/10.3390/ma14195664?type=check_update&version=2 Materials 2021, 14, 5664 2 of 9 the nitrite-oxide equilibrium becomes more tangible through Equation (2). In this regard, Sötz et al. reported the stabilization of molten Solar Salt with a purge gas flow containing nitrous gases at temperatures of 600 ◦C and even 620 ◦C in terms of stable nitrate-nitrite- and low oxide ion concentrations. In this study, the positive effect of the stabilizing gases is tested to its limits. More specifically, the reduction of high oxide ion levels (0.15 wt% or 1 mol%) in Solar Salt-based mixtures (60 wt% NaNO3, 40 wt% KNO3) by reversing Equation (2), is demonstrated. These mixtures are artificially aged by adding unusually high concentrations of nitrite- and oxide ions (10 mol% and 1 mol%, respectively) and the effect of different purge gas compositions (20 % and 80 % O2, with and without 200 ppm NO) on the salt chemistry are investigated. The synthetically aged Solar Salt mixtures represent extreme conditions of decomposed Solar Salt beyond the tolerable level in practical applications. Increasing the operating temperature, the threat of salt decomposition becomes more likely and, so far, active control of reaction Equation (2) has not been sufficiently addressed in the literature. In this paper the reduction of oxide ion concentration is defined as regeneration, which is the process to convert synthetically altered Solar Salt to Solar Salt with a tolerable low level of oxide and nitrite ions. In other words, this work presents a change of perception, where decomposition products are intentionally added to the salt melt, in order to enable the investigation of a recovery/regeneration mechanism, which has not been investigated experimentally yet. 2. Materials and Methods To investigate salt regeneration, synthetically altered Solar Salt was produced by mixing NaNO3 (Merck, Darmstadt, Hesse, Germany, purity > 99%), KNO3 (Merck, purity > 99%), and Na2O2 (Merck, purity > 95%) prepared in a moisture-free atmosphere and NaNO2 (Merck, Darmstadt, Hesse, Germany, purity > 99%), according to the desired composition listed in Table 1. The cation content in all experiments, was fixed to the one of the Solar Salt composition (Na 65.5 mol%; K 35.5 mol%). Each of the experiments is labeled according to their salt and purge gas composition (see Figure 1). Table 1. List of experiments of this work (and literature). Labels are specified with regard to salt composition and purge gas composition. All salts exhibit Na/K ratios equal to that of ideal Solar Salt mixture (Na 65.5 mol %:K 35.5 mol %). Label Anion Content (mol %) Purge Gas (100 mL/min) Nitrate Nitrite Peroxide N2 O2 NO Ref/20_0 [a] 100 0 0 80 vol% 20 vol% 0 Ref/20_200 [b] 100 0 0 80 vol% 20 vol% 200 ppm Ox/20_0 99 0 1 80 vol% 20 vol% 0 Nit-Ox/20_0 89 10 1 80 vol% 20 vol% 0 Nit-Ox/20_200 89 10 1 80 vol% 20 vol% 200 ppm Ref/80_0 100 0 0 20 vol% 80 vol% 0 Ox/80_0 99 0 1 20 vol% 80 vol% 0 Nit-Ox/80_0 89 10 1 20 vol% 80 vol% 0 Ox/80_200 99 0 1 20 vol% 80 vol% 200 ppm Nit-Ox/80_200 89 10 1 20 vol% 80 vol% 200 ppm [a,b] Reference experiments already published by Sötz et al. [7]. Materials 2021, 14, x FOR PEER REVIEW 2 of 9 nitrite-oxide equilibrium becomes more tangible through Equation (2). In this regard, Sötz et al. reported the stabilization of molten Solar Salt with a purge gas flow containing ni- trous gases at temperatures of 600 °C and even 620 °C in terms of stable nitrate-nitrite- and low oxide ion concentrations. In this study, the positive effect of the stabilizing gases is tested to its limits. More specifically, the reduction of high oxide ion levels (0.15 wt% or 1 mol%) in Solar Salt-based mixtures (60 wt% NaNO3, 40 wt% KNO3) by reversing Equation (2), is demonstrated. These mixtures are artificially aged by adding unusually high concentrations of nitrite- and oxide ions (10 mol% and 1 mol%, respectively) and the effect of different purge gas compositions (20 % and 80 % O2, with and without 200 ppm NO) on the salt chemistry are investigated. The synthetically aged Solar Salt mixtures represent extreme conditions of decomposed Solar Salt beyond the tolerable level in practical applications. Increasing the operating temperature, the threat of salt decomposition becomes more likely and, so far, active control of reaction Equation (2) has not been sufficiently addressed in the literature. In this paper the reduction of oxide ion concentration is defined as regeneration, which is the process to convert synthetically altered Solar Salt to Solar Salt with a tolerable low level of oxide and nitrite ions. In other words, this work presents a change of perception, where decomposition products are intentionally added to the salt melt, in order to enable the investigation of a recovery/regeneration mechanism, which has not been investigated experimentally yet. 2. Materials and Methods To investigate salt regeneration, synthetically altered Solar Salt was produced by mixing NaNO3 (Merck, Darmstadt, Hesse, Germany, purity > 99%), KNO3 (Merck, purity > 99%), and Na2O2 (Merck, purity > 95%) prepared in a moisture-free atmosphere and NaNO2 (Merck, Darmstadt, Hesse, Germany, purity > 99%), according to the desired com- position listed in Table 1. The cation content in all experiments, was fixed to the one of the Solar Salt composition (Na 65.5 mol%; K 35.5 mol%). Each of the experiments is labeled according to their salt and purge gas composition (see Figure 1). Figure 1. Experiment labeling according to salt composition and purge gas used. Table 1. List of experiments of this work (and literature). Labels are specified with regard to salt composition and purge gas composition. All salts exhibit Na/K ratios equal to that of ideal Solar Salt mixture (Na 65.5 mol %:K 35.5 mol %). Label Anion Content (mol %) Purge Gas (100 mL/min) Nitrate Nitrite Peroxide N2 O2 NO Ref/20_0 [a] 100 0 0 80 vol% 20 vol% 0 Ref/20_200 [b] 100 0 0 80 vol% 20 vol% 200 ppm Ox/20_0 99 0 1 80 vol% 20 vol% 0 Nit-Ox/20_0 89 10 1 80 vol% 20 vol% 0 Nit-Ox/20_200 89 10 1 80 vol% 20 vol% 200 ppm Ref/80_0 100 0 0 20 vol% 80 vol% 0 Ox/80_0 99 0 1 20 vol% 80 vol% 0 Nit-Ox/80_0 89 10 1 20 vol% 80 vol% 0 Figure 1. Experiment labeling according to salt composition and purge gas used. Materials 2021, 14, 5664 3 of 9 For each experiment an autoclave test rig was loaded with 100 g of salt mix contained in an Al2O3 crucible. The autoclave test rig is temperature controlled and purged with a gas flow, and adjusted with calibrated flow meters. A detailed description of the experimental setup is given by Bonk et al. [10]. In order to remove moisture and carbon dioxide, each of the salt mixtures is heated to 120 ◦C for at least 24 h under a flow of synthetic air (5.0 grade, Linde Gas). Subsequently, Na2O2 is added to the salt and the desired purge gas flow (100 mL/min) is adjusted from mixing O2, N2 and a reference gas containing 1000 ppm NO in N2 (all 5.0 grade, Linde gas). The salt is stirred and heated (2 K/min) to the target temperature (600 ◦C) and the purge gas flow is adjusted according to the desired composition (see Table 1). Upon reaching 600 ◦C, the first salt sample is collected (t = 0 h). Over the course of the isothermal experiments (~1200 h), salt samples are extracted and analyzed to monitor salt composition (see upcoming sub-chapter for details). To reveal the anion (and cation) content of the salt samples, ion chromatography is utilized. For ion chromatography (IC) a Metrohm model 930 Compact IC Flex is used. About 125 mg of the salt sample is dissolved in ultrapure water (500 mL, HiPerSolv, VWR, Darmstadt, Germany) and analyzed. A detailed description of the experimental procedure and calibration is given elsewhere [10]. Chromate was measured in order to monitor undesired side reactions between the steel parts and the molten salt. The limit for quantitative detection of chromate ions accounts to 0.2 mg/L or 0.05 mol% in a salt sample. Standard deviation for IC analysis was accessed via a five-fold measurement of a representative sample composition (c(NO3 −) = 140.13 mg/L, c(NO2 −) = 26.43 mg/L, c(CrO4 2−) = 0.72 mg/L). To identify additional cationic species (other than sodium and potassium) arising from steel corrosion (iron, chromium, nickel), an iCE 3000 series atomic adsorption spec- trometer from Thermo Fisher Scientific (Waltham, Massachusetts, USA) was utilized. In a standardized procedure, about 50 mg of the salt sample are dissolved in 50 mL of ultrapure water including 0.5 mL of 69% HNO3 and 0.5 mL of a 10% CsCl solution. Cali- bration was performed with standard solutions (Roth, certified reference material; (iron (1003.0 ± 2.2 mg/L; chromium (1000.9 ± 3.1 mg/L; nickel (1001.8±3.5 mg/L)) for iron, chromium and nickel and freshly prepared for the desired concentration range (1–5 mg/L). The oxide and carbonate contents are determined by inert gas acid base titration (Metrohm Titrando 800, Herisau, Switzerland). Regardless of the specific oxide ion species actually present in the melt (O2 −, O2 2−, O2−), all are summarized and expressed as O2− values in this work. About 500 mg of each salt sample was dissolved in 160 mL ultrapure water and titrated with 0.01 M HCl (Titrisol standard solution). Calibration of the HCl titrant was performed by multiple titrations of a known amount of pre dried Na2CO3 (Merck, Darmstadt, Hesse, Germany, purity > 99.5%). The oxide ion concentration of the molten salt is calculated back from the measured hydroxide (2 OH− converted to 1 O2−) and carbonate (1 CO3 2- converted to 1 O2−) concentrations, obtained from acid base titration [10–12]. The carbonates are likely to be formed after sample extraction, since the purge gases are of high purity and leak rates of the test rig are low. It is reasonable and already reported that oxide ions react with atmospheric carbon dioxide before or during titration [10,13]. In the presence of nitrite, the found carbonate ion concentration is signifi- cantly lower (10–20%) when compared to the true value. In order to take this systematic error into account, error bars of the oxide ion concentration contain the maximum error for the carbonate detection and for this reason are not symmetric. The detection limit with regard to salt composition was below 0.1 mol% O2−. For mol fractions of each ionic species all uncertainties are combined via quadratic error propagation and represented as error bars in the figures (see Tables S1–S3). Overall, the most relevant post analysis methods were IC and titration to monitor the nitrite and oxide level for the Solar Salt regeneration. Other measurements were supplementary to monitor undesired side reactions with some steel parts. Materials 2021, 14, 5664 4 of 9 3. Results and Discussion 3.1. Behavior of an Oxide Rich Nitrate Melt under Various Conditions In the following, changes in the anion content of the molten salt over the course of the experiment are discussed in order to reveal a potential influence of different purge gas compositions on the molten salt chemistry. In order to avoid an incorrect interpretation of the experimental results, the effect of each reaction parameter needed to be analyzed with caution. For clarity, the different experiments listed in Table 1 are categorized and discussed separately in a twofold manner. To exclude the effect of the nitrate–nitrite equilibrium, for the first set the results of molten salt purged with p(O2) = 0.2 atm are discussed. In the second step, experiments with a purge gas containing p(O2) = 0.8 atm are discussed. The summarized data of experiments performed under p(O2) = 0.2 atm expressed as anion concentrations over time, is illustrated in Figure 2. Reference experiments of pure Solar Salt already published are marked orange and named according to the applied purge gas (Ref/20_0, Ref/20_200) [7]. Within 200 h, all experiments exhibited a steady concentration of nitrates and nitrites. The molar fraction of nitrate stabilized between 89.0, 89.3, 89.6 (±0.1) mol% (Ox/20_0, Nit-Ox/20_200, Nit-Ox/20_0). Similarly, the nitrite concentration equilibrated at 10.4, 9.8, 9.7 (±0.1) mol% respectively. In general, these results agree with previously reported Solar Salt equilibrium data [7,14–16]. Materials 2021, 14, x FOR PEER REVIEW 5 of 9 Figure 2. Content of nitrate, nitrite and oxide ions of investigated nitrate melts in this work with oxide addition. Literature data (orange symbols) shown (Ref/20_0, Ref/20_200) without oxide addition. All melts were stored under synthetic air with/without 200 ppm NO indicated by the label. Error bars not visible are within symbol size. Close examination of the oxide content reveals that the molten salts in each of the three experiments follow a different reaction mechanism. For Nit-Ox/20_200 with nitrite and oxide addition and NO in the purge gas, the molten salt is first regenerated, as ex- pressed by a significant reduction of the oxide ion concentrations. Subsequently, the oxide level is in equilibrium (unchanged), which can be attributed to the presence of NO in the atmosphere. It is reasonable to assume that the oxide ions are regenerated by the back reaction of Equation (2). This effect is visible in the nitrite content of Nit-Ox/20_200 having a maximum value after 116 h (Figure 2), which is not the case for molten salts purged with synthetic air (Ox/20_0, Nit-Ox/20_0). Compared to the literature experiment from Sötz et al. (Ref/20_200), the oxide ion equilibrium concentration at the end of the experiment of Nit-Ox-20_200 is somewhat higher (0.006 mol% vs. 0.11 mol%). In the case of the Ox/20_0 experiment with no NO in the gas stream, the nitrous gases evolving from the melt are constantly purged out of the crucible, resulting in p(NO) being effectively zero and there- fore not allowing for any chemical equilibrium of Equation (2). The high but constant ox- ide ion content in Nit-Ox/20_0, is somewhat confounding, but could be explained as the result of a kinetic effect. Compared to Ref/20_0, the oxide content in our experiment is four times higher. Thus, with respect to Equation (2), nitrite decomposition is less favora- ble, slowed down, and the melt appears stabilized. It is possible that longer experimental durations would show an increasing oxide ion concentration of the Nit-Ox/20_0 experi- ment as a result of ongoing nitrite decomposition. The second set of experiments was performed with oxygen rich purge gas (p(O2) = 0.8 atm) are reduce nitrate decomposition and consequently study its effect on the oxide equilibrium. Similar to the previously discussed set of experiments at p(O2) = 0.8 atm all of the experiments reach nitrate-nitrite-equilibrium with a ratio close to Ref/80_0 (Figure Figure 2. Content of nitrate, nitrite and oxide ions of investigated nitrate melts in this work with oxide addition. Literature data (orange symbols) shown (Ref/20_0, Ref/20_200) without oxide addition. All melts were stored under synthetic air with/without 200 ppm NO indicated by the label. Error bars not visible are within symbol size. The oxide ion concentrations of the molten salt samples (Ox/20_0, Nit-Ox/20_200, Nit-Ox/20_0; Figure 2) decreased steadily during the first 400 h, before they stabilized at 0.3, 0.1 and 0.4 mol%, respectively. It is in contrast to our expectation that the ox- ide ion concentrations decrease if no NO-gas is present in the purge gas. Further salt analysis, however, indicated that the oxide ions were consumed in a concurrent reac- Materials 2021, 14, 5664 5 of 9 tion with the stirrer (see next chapter). At a later stage of the experiments, the oxide ion concentrations began to diverge. For molten salt without NO purge (Ox/20_0), the oxide content started to increase again after 400 h. This process was not observed for Nit-Ox/20_0, where the oxide ion content did not drop as low as in the Ox/20_0 experi- ment and remained constant at 0.44 (+0.01 − 0.02) mol%. Molten salt purged with 200 ppm NO (Nit-Ox/20_200) comprised about a tenth of its initial oxide concentration after 400 h, approximately 0.11 (+0.01 − 0.02) mol% after t = 1125 h. These findings allow for some remarks on the interaction of the gas phase and the oxide reaction equilibrium. The nitrate-nitrite equilibrium (Equation (1)) is most probably not directly affected by the presence of nitrous gas, and thus equilibrium levels for nitrate and nitrite are similar to published equilibrium data without additional oxide ions (Ref/20_0). In contrast to our expectation, for experiments with the addition of nitrite (Nit-Ox/20_0, Nit-Ox/20_200) the initial nitrite ion concentration after heating to the target temperature (at t = 0 h) was below 10 mol%. Presumably this is the case due to the occurrence of catalytic reactions involving the oxide ions and the oxidation of nitrite during the heating process, as it has been reported elsewhere [17]. The results point to the probability that with the presence of oxide ions, the effect of temperature and p(O2) on nitrate-nitrite reactions is occurring more rapidly. As the focus of this study was on equilibrium reactions at 600 ◦C, we are aware that interpretations of reaction kinetics must be done with great caution and qualitatively only. Close examination of the oxide content reveals that the molten salts in each of the three experiments follow a different reaction mechanism. For Nit-Ox/20_200 with nitrite and oxide addition and NO in the purge gas, the molten salt is first regenerated, as expressed by a significant reduction of the oxide ion concentrations. Subsequently, the oxide level is in equilibrium (unchanged), which can be attributed to the presence of NO in the atmosphere. It is reasonable to assume that the oxide ions are regenerated by the back reaction of Equation (2). This effect is visible in the nitrite content of Nit-Ox/20_200 having a maximum value after 116 h (Figure 2), which is not the case for molten salts purged with synthetic air (Ox/20_0, Nit-Ox/20_0). Compared to the literature experiment from Sötz et al. (Ref/20_200), the oxide ion equilibrium concentration at the end of the experiment of Nit-Ox-20_200 is somewhat higher (0.006 mol% vs. 0.11 mol%). In the case of the Ox/20_0 experiment with no NO in the gas stream, the nitrous gases evolving from the melt are constantly purged out of the crucible, resulting in p(NO) being effectively zero and therefore not allowing for any chemical equilibrium of Equation (2). The high but constant oxide ion content in Nit-Ox/20_0, is somewhat confounding, but could be explained as the result of a kinetic effect. Compared to Ref/20_0, the oxide content in our experiment is four times higher. Thus, with respect to Equation (2), nitrite decomposition is less favorable, slowed down, and the melt appears stabilized. It is possible that longer experimental durations would show an increasing oxide ion concentration of the Nit-Ox/20_0 experiment as a result of ongoing nitrite decomposition. The second set of experiments was performed with oxygen rich purge gas (p(O2) = 0.8 atm) are reduce nitrate decomposition and consequently study its effect on the oxide equilibrium. Similar to the previously discussed set of experiments at p(O2) = 0.8 atm all of the experiments reach nitrate-nitrite-equilibrium with a ratio close to Ref/80_0 (Figure 3, orange). Furthermore, nitrite addition did not influence the final equilibrium values. All experiments with artificially added oxides showed a decrease of oxide concen- tration over time. In the reference experiment (Ref/80_0) with pure Solar Salt (Figure 3, orange) the nitrite content increased before it reached a steady state at 6.0 (±0.1) mol%, while the nitrate content stabilized at 93.9 (±0.1) mol%. This result is consistent with thermodynamic expec- tations (93.3(±0.3):6.7(±0.1) mol% NO3 −:NO2 − [7]) and in agreement with Equation (1), where increasing the oxygen partial pressure will push equilibrium composition to the nitrate side, thereby lowering the nitrite content. The detected oxide ion concentration increased continuously, with a final value of 0.14 (+0.01 − 0.02) mol% at 1150 h. This Materials 2021, 14, 5664 6 of 9 behavior is again explained by the fact that all gases evolving from the molten salt are purged out of the test rig, allowing no equilibration of Equation (2). Materials 2021, 14, x FOR PEER REVIEW 6 of 9 3, orange). Furthermore, nitrite addition did not influence the final equilibrium values. All experiments with artificially added oxides showed a decrease of oxide concentration over time. In the reference experiment (Ref/80_0) with pure Solar Salt (Figure 3, orange) the ni- trite content increased before it reached a steady state at 6.0 (±0.1) mol%, while the nitrate content stabilized at 93.9 (±0.1) mol%. This result is consistent with thermodynamic ex- pectations (93.3(±0.3):6.7(±0.1) mol% NO3−:NO2− [7]) and in agreement with Equation (1), where increasing the oxygen partial pressure will push equilibrium composition to the nitrate side, thereby lowering the nitrite content. The detected oxide ion concentration increased continuously, with a final value of 0.14 (+0.01 − 0.02) mol% at 1150 h. This be- havior is again explained by the fact that all gases evolving from the molten salt are purged out of the test rig, allowing no equilibration of Equation (2). Figure 3. Content of nitrate, nitrite and oxide ions of investigated nitrate melts. Reference data shown (Ref/80_0) did not contain oxide dopant. Salts were stored under 0.8 atm O2 with/without 200 ppm NO and with/without added nitrite indi- cated by the label. Error bars not visible are within symbol size. For experiments without NO in the atmosphere and added oxides, the oxide ion con- tent of Ox/80_0, Nit-Ox/80_0 decreased over the course of the experiments, mainly due to consumption of those ions during corrosion reactions. With the appliance of a purge gas containing NO, the oxide ion content was significantly lower compared to experiments without NO. Detected values were as low as 0.05(±0.01) mol% for Nit-Ox/80_200 and be- low the detection limit (<0.01 mol%) for Ox/80_200. Rapid regeneration again resulted in a temporary increase of the nitrite content during the early stages of the experiments. This can be seen for experiment Nit-Ox/80_200 and Ox/80_200 after 21 and 166 h, respectively. Figure 3. Content of nitrate, nitrite and oxide ions of investigated nitrate melts. Reference data shown (Ref/80_0) did not contain oxide dopant. Salts were stored under 0.8 atm O2 with/without 200 ppm NO and with/without added nitrite indicated by the label. Error bars not visible are within symbol size. For experiments without NO in the atmosphere and added oxides, the oxide ion content of Ox/80_0, Nit-Ox/80_0 decreased over the course of the experiments, mainly due to consumption of those ions during corrosion reactions. With the appliance of a purge gas containing NO, the oxide ion content was significantly lower compared to experiments without NO. Detected values were as low as 0.05(±0.01) mol% for Nit-Ox/80_200 and below the detection limit (<0.01 mol%) for Ox/80_200. Rapid regeneration again resulted in a temporary increase of the nitrite content during the early stages of the experiments. This can be seen for experiment Nit-Ox/80_200 and Ox/80_200 after 21 and 166 h, respectively. It is worth mentioning that the oxide ion equilibrium level under 0.8 atm O2 (Figure 3, Nit-Ox/80_200 with a value of 0.05 (±0.01) mol% at the end of the experiment) is about half compared to 0.2 atm O2 purge (Figure 2, Nit-Ox/20_200 with a value of 0.11 (+0.01 − 0.02) mol% at the end of the experiment). This demonstrates that by reducing the equilibrium content of nitrite from about 10 mol% (p(O2) = 0.2 atm) to 6 mol% (p(O2) = 0.8 atm), the equilibrium oxide ion content is also decreased. It is concluded in this work that the oxygen partial pressure indirectly controls nitrite decomposition to oxides. Materials 2021, 14, 5664 7 of 9 3.2. Indicators for Corrosivity in an Oxide Rich Nitrate Melt The addition of sodium peroxide significantly altered the chemical properties of Solar Salt. In our experiments, the shaft of the stainless steel stirrer suffered from corrosion attack through the oxide ion. We utilized the detection rate of chromate as a qualitative indicator for the apparent corrosivity of our molten salts [18]. Molten salt without the addition of oxide ions (Ref/80_0) did not contain detectable chromates and the salt remained colorless over the course of the experiment. In contrast, experiments containing oxide ions showed steeply increasing chromate concentrations to approximately 0.1 mol%, within the first 200 h (Figure 4b), indicating high corrosivity compared to experiments without oxide addition. The salt samples containing even moderate concentrations of chromates exhibited a bright yellow or yellow/green color (Figure 4a). Materials 2021, 14, x FOR PEER REVIEW 7 of 9 It is worth mentioning that the oxide ion equilibrium level under 0.8 atm O2 (Figure 3, Nit-Ox/80_200 with a value of 0.05 (±0.01) mol% at the end of the experiment) is about half compared to 0.2 atm O2 purge (Figure 2, Nit-Ox/20_200 with a value of 0.11 (+0.01 − 0.02) mol% at the end of the experiment). This demonstrates that by reducing the equilib- rium content of nitrite from about 10 mol% (p(O2) = 0.2 atm) to 6 mol% (p(O2) = 0.8 atm), the equilibrium oxide ion content is also decreased. It is concluded in this work that the oxygen partial pressure indirectly controls nitrite decomposition to oxides. 3.2. Indicators for Corrosivity in an Oxide Rich Nitrate Melt The addition of sodium peroxide significantly altered the chemical properties of So- lar Salt. In our experiments, the shaft of the stainless steel stirrer suffered from corrosion attack through the oxide ion. We utilized the detection rate of chromate as a qualitative indicator for the apparent corrosivity of our molten salts [18]. Molten salt without the ad- dition of oxide ions (Ref/80_0) did not contain detectable chromates and the salt remained colorless over the course of the experiment. In contrast, experiments containing oxide ions showed steeply increasing chromate concentrations to approximately 0.1 mol%, within the first 200 h (Figure 4b), indicating high corrosivity compared to experiments without oxide addition. The salt samples containing even moderate concentrations of chromates exhibited a bright yellow or yellow/green color (Figure 4a). Figure 4. (a) Salt sample appearance initially (t = 0 h) and at the end of the experiments (t = 1150 h); (b) Chromate content over time of the molten salts. Limit of quantification (LOQ) indicated as gray bar. Reference data shown (Ref/80_0) did not contain oxide impurity. Error bars lay inside symbol. Figure 4. (a) Salt sample appearance initially (t = 0 h) and at the end of the experiments (t = 1150 h); (b) Chromate content over time of the molten salts. Limit of quantification (LOQ) indicated as gray bar. Reference data shown (Ref/80_0) did not contain oxide impurity. Error bars lay inside symbol. Apart from chromium impurities, iron was also detected (up to 70 ppm) in some of the salt samples which contained oxide ions (see Table S4). By comparing the sample appearance depicted in (Figure 4a), a brown/green color (Ox/20_0, Ox/80_200) is a sign for transition metal impurities (e.g., Fe, Cr) dissolved in the salt [14,19,20]. Because corrosion reactions consume oxide ions and the investigated oxide ion regeneration process is taking place at the same time, great caution was necessary for the interpretation of oxide equilib- rium data (Figures 2 and 3). However, during experiments with NO purge, the chromate Materials 2021, 14, 5664 8 of 9 levels stabilized, which indicates that corrosion reactions stopped at some point (e.g., ex- periments Ox/80_200, Nit-Ox/80_200 and Nit-Ox/20_200) and oxide ion equilibrium data was collected. It can be concluded that the regeneration of aged molten salt suppressed the release of transition metal impurities most likely originating from corrosion reactions. An extensive corrosion study needs to be conducted in order to generate quantitative data on this effect, and that was not within the scope of this investigation. 4. Conclusions In this study, we demonstrated the regeneration of a decomposed oxide ion rich molten Solar Salt towards literature known equilibrium levels with a very low concentration of oxides under synthetic air conditions and 200 ppm NO. Additionally, the regeneration under p(O2) = 0.8 atm lead to even lower oxide equilibrium levels and this showed the indirect influence of pO2 on the oxide formation mechanism. Our data reveals the entire capability of the nitrous gas purge to recover an aged molten Solar Salt towards Solar Salt equilibrium composition at 600 ◦C in terms of anion content. On the basis of literature investigations, thermophysical properties, which are directly dependent on salt composition, were conserved [19,20]. This work gives valuable experimental data on the oxide equilibrium concentration in Solar Salt under different atmospheres. Furthermore, for storage applications at current (565 ◦C) and potentially higher operation temperatures (>565 ◦C), our study is crucial for stable operation of Solar Salt beyond the current operation temperature and innovative salt recovery strategies of decomposed Solar Salt. Additionally, we have shown that at 600 ◦C, molten Solar Salt with high oxide concentration significantly attacks steel components. By purging the molten salt with NO rich gas, the release of transition metal impurities potentially originating from corrosion reactions was retarded. However, given the short duration and varying steel surface quality in our experiments, caution must be applied when interpreting the data. Research into answering the question of a positive effect of NO purge and corrosion process reduction is already in progress. Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/ma14195664/s1, Table S1: Titration error sources. Table S2: Ion chromatography error sources. Table S3: Maximum relative measuring uncertainty with respect to quadratic error propagation for molar ion content. Table S4: Iron content in nitrate salt samples after different exposure times. Images of the respective samples are presented for clarity. Author Contributions: Conceptualization, J.S. and A.B.; Funding acquisition, A.B. and T.B.; Inves- tigation, J.S.; Methodology, J.S. and V.A.S.; Project administration, A.B. and T.B.; Resources, J.S.; Supervision, J.S. and A.B.; Validation, J.S., V.A.S. and A.B.; Visualization, J.S.; Writing—original draft, J.S.; Writing—review & editing, A.B., V.A.S. and T.B. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by the German Federal Ministry for Economic Affairs and Energy (BMWi) in the VeNiTe project (Contract No. 03EE5043). Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Data Availability Statement: The data presented in this study are available on request from the corresponding authors. Acknowledgments: We kindly thank in particular Markus Braun and Andrea Hanke for technical work and great expertise in experimental methods. Conflicts of Interest: The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. https://www.mdpi.com/article/10.3390/ma14195664/s1 https://www.mdpi.com/article/10.3390/ma14195664/s1 Materials 2021, 14, 5664 9 of 9 References 1. Durth, M.; Prieto, C.; Rodriguez-Sanchez, A.; Patino-Rodriguez, D.; Cabeza, L.F. Effects of sodium nitrate concentration on thermophysical properties of solar salts and on the thermal energy storage cost. Sol. Energy 2019, 182, 57–63. [CrossRef] 2. Bauer, T.; Pfleger, N.; Laing, D.; Steinmann, W.-D.; Eck, M.; Kaesche, S. High-temperature molten salts for solar power application. In Molten Salts Chemistry; Elsevier: Amsterdam, The Netherlands, 2013; pp. 415–438. 3. Stern, K.H. High temperature properties and decomposition of inorganic salts part 3, nitrates and nitrites. J. Phys. Chem. Ref. Data 1972, 1, 747–772. [CrossRef] 4. Nissen, D.A.; Meeker, D.E. Nitrate/nitrite chemistry in sodium nitrate-potassium nitrate melts. Inorg. Chem. 1983, 22, 716–721. [CrossRef] 5. Wright, S.; Tran, T.; Chen, C.; Olivares, R.I.; Sun, S. Thermal Stability of Potassium and Sodium Nitrate Molten Salt Mixtures Above 500 ◦C. In Proceedings of the Ninth International Conference on Molten Slags, Fluxes and Salts, Beijing, China, 27–30 May 2012. 6. Mohammad, M.B.; Brooks, G.A.; Rhamdhani, M.A. Premelting, Melting, and Degradation Properties of Molten Alkali Nitrates: LiNO3, NaNO3, KNO3, and Binary NaNO3-KNO3. Metall. Mater. Trans. B 2018, 49, 1482–1498. [CrossRef] 7. Sotz, V.A.; Bonk, A.; Steinbrecher, J.; Bauer, T. Defined purge gas composition stabilizes molten nitrate salt—Experimental prove and thermodynamic calculations. Sol. Energy 2020, 211, 453–462. [CrossRef] 8. Fei, Z.; Zhang, Y.; Ge, M.; Wang, Y.; Li, Y.; Cheng, J.; Wei, B.; Hou, H.; Liu, H. Probing thermal decomposition mechanism of molten nitrite/nitrates salt by time of flight mass spectrometry. Sol. Energy 2019, 183, 823–828. [CrossRef] 9. Freeman, E.S. The kinetics of the thermal decomposition of sodium nitrate and of the reaction between sodium nitrite and oxygen. J. Phys. Chem. 1956, 60, 1487–1493. [CrossRef] 10. Bonk, A.; Martin, C.; Braun, M.; Bauer, T. Material Investigations on the Thermal Stability of Solar Salt and Potential Filler Materials for Molten Salt Storage. AIP Conf. Proc. 2017, 1850, 080008. [CrossRef] 11. Sötz, V.A.; Bonk, A.; Bauer, T. With a view to elevated operating temperatures in thermal energy storage—Reaction chemistry of Solar Salt up to 630 ◦C. Sol. Energy Mater. Sol. Cells 2020, 212, 110577. [CrossRef] 12. Fernandez, A.G.; Cabeza, L.F. Molten salt corrosion mechanisms of nitrate based thermal energy storage materials for concentrated solar power plants: A review. Sol. Energy Mater. Sol. Cells 2019, 194, 160–165. [CrossRef] 13. Nikulshina, V.; Ayesa, N.; Galvez, M.E.; Steinfeld, A. Feasibility of Na-based thermochemical cycles for the capture of CO2 from air—Thermodynamic and thermogravimetric analyses. Chem. Eng. J. 2008, 140, 62–70. [CrossRef] 14. Bonk, A.; Braun, M.; Hanke, A.; Forstner, J.; Rückle, D.; Kaesche, S.; Sötz, V.A.; Bauer, T. Influence of different atmospheres on molten salt chemistry and its effect on steel corrosion. AIP Conf. Proc. 2018, 2033, 090003. [CrossRef] 15. Sötz, V.A.; Bonk, A.; Forstner, J.; Bauer, T. Microkinetics of the reaction NO3 − NO2 −+0.5 O2 in molten sodium nitrate and potassium nitrate salt. Thermochim. Acta 2019, 678, 178301. [CrossRef] 16. Tracey, T.R. Conceptual Design of Advanced Central Receiver Power System; Final Report; Martin Marietta Corp.: Denver, CO, USA, 1978. 17. Paniccia, F.; Zambonin, P.G. Redox mechanism in an ionic matrix. IV. Catalytic effects of peroxide and superoxide on the oxidation of nitrite by molecular oxygen in molten alkali nitrates. J. Phys. Chem. 1974, 78, 1693–1698. [CrossRef] 18. Federsel, K.; Wortmann, J.; Ladenberger, M. High-temperature and corrosion behavior of nitrate nitrite molten salt mixtures regarding their application in concentrating solar power plants. Enrgy Proced 2015, 69, 618–625. [CrossRef] 19. Villada, C.; Bonk, A.; Bauer, T.; Bolivar, F. High-temperature stability of nitrate/nitrite molten salt mixtures under different atmospheres. Appl. Energy 2018, 226, 107–115. [CrossRef] 20. Zhang, C.; Han, Y.; Wu, Y.; Lu, Y. Comparative study on high temperature thermal stability of quaternary nitrate-nitrite mixed salt and Solar salt. Sol. Energy Mater. Sol. Cells 2021, 230, 111197. [CrossRef] http://doi.org/10.1016/j.solener.2019.02.038 http://doi.org/10.1063/1.3253104 http://doi.org/10.1021/ic00147a004 http://doi.org/10.1007/s11663-018-1205-z http://doi.org/10.1016/j.solener.2020.09.041 http://doi.org/10.1016/j.solener.2019.03.067 http://doi.org/10.1021/j150545a005 http://doi.org/10.1063/1.4984429 http://doi.org/10.1016/j.solmat.2020.110577 http://doi.org/10.1016/j.solmat.2019.02.012 http://doi.org/10.1016/j.cej.2007.09.007 http://doi.org/10.1063/1.5067097 http://doi.org/10.1016/j.tca.2019.178301 http://doi.org/10.1021/j100610a003 http://doi.org/10.1016/j.egypro.2015.03.071 http://doi.org/10.1016/j.apenergy.2018.05.101 http://doi.org/10.1016/j.solmat.2021.111197 Journal Publication - 26 - 3.2 Paper II Stabilization of Solar Salt at 650 °C – Thermodynamics and practical implications for thermal energy storage systems Julian Steinbrecher, Andrea Hanke, Markus Braun, Thomas Bauer, Alexander Bonk This article was published in Solar Energy Materials and Solar Cells 2023, 258, 112411. Copyright 2023 Elsevier B.V. https://doi.org/10.1016/j.solmat.2023.112411 https://doi.org/10.1016/j.solmat.2023.112411 Journal Publication - 27 - Individual contributions in Paper II: • Conceptualization: In the paper the research goal was specified by the lead author, which was the generation of thermodynamic data for Solar Salt under various oxygen atmospheres up to a temperature of 650 °C. In this context, the temperature dependent oxide ion formation was defined as one research subject. Further, the impact of pressure control in gas tight reaction vessel with Solar Salt and added sodium nitrite at 650 °C was planned. • Selection of the parameter variation and execution of the experiments. • Data collection: The lead author was responsible for the experimental supervision as well as the collection of salt samples. Subsequently, post analysis of salt samples and data evaluation was performed by the lead author. The appropriate methodology to generate thermodynamic and kinetic data was defined by the lead author. • Visualization and deductive reasoning: Selection of the relevant data and graphical illustration of the data was categorized and prepared in a clear way under the lead of the first author. The outline and structuring of the research data were done together. Context between state-of-the-art knowledge was drawn and novelty of the work was developed. The experimental detail of each previous study was summarized and the impact on the accuracy of the thermodynamic data was evaluated by the lead author. • Writing and publishing the manuscript in Solar Energy Materials and Solar Cells by the lead author. Solar Energy Materials & Solar Cells 258 (2023) 112411 Available online 15 June 2023 0927-0248/© 2023 Elsevier B.V. All rights reserved. Stabilization of Solar Salt at 650 ◦C – Thermodynamics and practical implications for thermal energy storage systems Julian Steinbrecher a,*, Andrea Hanke a, Markus Braun a, Thomas Bauer b, Alexander Bonk a a German Aerospace Center (DLR), Institute of Engineering Thermodynamics, 70569, Stuttgart, Germany b German Aerospace Center (DLR), Institute of Engineering Thermodynamics, 51147, Cologne, Germany A R T I C L E I N F O Keywords: Thermal stability Solar Salt Molten salt High temperature chemistry Concentrating solar power A B S T R A C T Thermal Energy Storage (TES) based on molten salts is thought to play a major role for the transition from fossil fuels to renewable energy carriers in the future. Solar Salt, a mixture of NaNO3–KNO3 is currently the state-of- the-art heat transfer and storage material in Concentrating Solar Power (CSP) plants which produce electricity from a Rankine cycle with steam temperatures up to 550 ◦C. To allow a technology transfer and adapt Solar Salt based TES systems to modern, high temperature Rankine cycles (e.g. Tsteam > 600 ◦C), the thermal stability of Solar Salt needs to be increased well above 615 ◦C. At these temperatures, the formation of nitrites, which depends on the oxygen partial pressure above the melt, needs to be suppressed effectively to prevent further decomposition into corrosive oxide ions. In this work, the thermodynamics of the nitrite-forming reaction at different oxygen partial pressure are explored in a temperature range up to 650 ◦C from isothermal experiments in the 100 g-scale and limitations of the ideal description are revealed. The measured apparent oxide ion for- mation rates at 100 g-scale were below previous findings. The activation energy found was 60 ± 15 kJ/mol and the preexponential factor 1 ∗ 10− 5 ± 0.00005 s− 1. The effect of closing the storage system in terms of gas and salt phase at 645 ◦C are also explored to understand if and how pressure formation and oxygen release correlate. The results of this work finally contribute to an understanding of the decomposition reactions of Solar Salt at pre- viously untouched temperatures. 1. Introduction The efficient use of storage technologies utilizing thermal, chemical or electrochemical methods has become and will remain the key for the successful implementation of volatile renewable energy technologies to reduce world-wide CO2 emissions [1]. Amongst those technologies, thermal energy storage (TES) exhibits the highest potential for direct integration and transformation of former coal fired power plants (CPP) into storage plants [2]. Those plants combined with renewables could efficiently be used for future energy storage with net-zero CO2 emis- sions. Current TES technologies based on molten salts can reach tem- peratures up to 565 ◦C, thereby allowing the generation of steam up to temperatures of ~550 ◦C [3]. Modern steam turbines in CPP operate at steam temperatures at and above 600 ◦C which creates a demand for higher operating temperatures of the TES System [4]. Examples include increasing the efficiency of CSP plants, opening up new industrial pro- cesses for electrification and reducing the costs of power components (e. g. increasing the temperature difference and thus costs in the steam generator) [5]. For all these applications, it is relevant to raise the operating temperatures of Solar Salt beyond the state-of-the-art tem- perature of 565 ◦C. Modern TES systems in Concentrating Solar Power (CSP) plants are operated in non-pressurized two-tank systems and utilize “Solar Salt”, a mixture of 60 wt% NaNO3 and 40 wt% KNO3, as storage and heat transfer medium [6]. These systems are referred to as “open” systems hereafter, since they are taking up atmospheric air during the charging-discharging cycles. In this configuration, the oxygen partial pressure (PO2 ) is constant over long periods of time due to the constant uptake of “fresh” air. In the lab-scale this configuration is mimicked by an external purge gas of synthetic air which is required to stabilize the oxygen partial pressure, such as schematically shown in Fig. 1 (a)). Under such a constant oxygen partial pressure, Solar Salt will form a defined level of nitrite anions and eventually a nitrate-nitrite equilib- rium will establish according to Eq. (1). The concentration of nitrite ions will change with varying temperature and oxygen partial pressure PO2 in the purge gas; e.g. around 4.5 mol% nitrite is present at 565 ◦C, while * Corresponding author E-mail address: julian.steinbrecher@dlr.de (J. Steinbrecher). Contents lists available at ScienceDirect Solar Energy Materials and Solar Cells journal homepage: www.elsevier.com/locate/solmat https://doi.org/10.1016/j.solmat.2023.112411 Received 2 January 2023; Received in revised form 4 May 2023; Accepted 4 June 2023 mailto:julian.steinbrecher@dlr.de www.sciencedirect.com/science/journal/09270248 https://www.elsevier.com/locate/solmat https://doi.org/10.1016/j.solmat.2023.112411 https://doi.org/10.1016/j.solmat.2023.112411 https://doi.org/10.1016/j.solmat.2023.112411 http://crossmark.crossref.org/dialog/?doi=10.1016/j.solmat.2023.112411&domain=pdf Solar Energy Mate