Engineering molten MgCl2–KCl–NaCl salt for high-temperature thermal energy storage: Review on salt properties and corrosion control strategies

https://doi.org/10.1016/j.solmat.2021.111344Get rights and content

Highlights

  • Most relevant salt properties of MgCl2–KCl–NaCl to engineer this salt are reviewed.

  • A full set of their recommended temperature-dependent values are given.

  • Working temperature range of MgCl2–KCl–NaCl for TES/HTF is suggested.

  • Corrosion control strategies including corrosion monitoring and mitigation techniques are briefly reviewed.

  • Concept of corrosion control system integrated into TES/HTF system is introduced.

Abstract

Conventional thermal energy storage (TES) media and heat transfer fluids (HTFs) currently used in commercial concentrated solar power (CSP) plants are nitrate-based molten salts with working temperature up to about 565 °C. Current interest in raising the working temperature in next-generation CSP for higher energy conversion efficiency leads to chloride-based molten salt technology (>700 °C). MgCl2–KCl–NaCl salt is selected to be such a TES/HTF material due to its excellent salt properties, abundance and low costs. In this review, survey and evaluation of its exact eutectic composition by evaluation of the phase diagram as well as most relevant salt properties for design, construction and operation of a test loop and projection of a real TES/HTF system are presented. They include minimum melting temperature, vapor pressure, heat capacity, density, thermal conductivity and dynamic viscosity. The working temperature range 420–800 °C is suggested for the TES/HTF system using this salt. Moreover, the recommended values of the heat capacity, density, thermal conductivity and dynamic viscosity in the suggested working temperature range are given. In addition, corrosion control strategies for this salt including corrosion monitoring and mitigation techniques are briefly reviewed, since controlling the salt corrosivity is one main concern of the engineering aspects using the molten chlorides at high temperatures. Overall, this review is expected to supply reliable values of the most relevant salt properties and recommend the most promising corrosion control strategies for further material studies, as well as modeling, simulation, detailed design and construction of molten MgCl2–KCl–NaCl salt TES/HTF systems.

Introduction

Global warming exacerbates the effects of climate change due to the increasing trend of the global energy demand. Efforts of the scientific community and industry have been made in the diversification of sustainable clean energy sources [1]. For instance, the developed concentrated solar power (CSP) plant uses the solar energy to produce electricity and has been an attractive option due to its facile coupling to thermal energy storage (TES) to produce dispatchable clean electricity [2]. It is estimated by International Energy Agency (IEA) that the global capacity of CSP and its TES will grow by 4.3 GW and 21 GWhel from 2018 to 2023, respectively [3].

The TES technology has the advantage that it is possible to store the thermal energy for later use, thereby reducing the gap between solar energy supply and energy demand due to the intermittency during cloudy conditions and at night [4]. Among various TES systems, the most widely used is sensible thermal energy storage (STES). The materials for STES generally are liquids or solids and utilize the heat capacity due to temperature changes during charging and discharging [5,6]. Conventional STES liquids include pressurized water, thermal oils, liquid metals and molten salts. Among them, molten salts have been successfully applied in commercial CSP plants. The commercial TES systems based on molten salts have two different configurations. One is the indirect storage system with molten salt as storage medium often used in a parabolic trough CSP plant with a lower maximum temperature level of about 400 °C. Another configuration is direct storage system often used in a tower CSP plant with molten salt as both, heat transfer fluid (HTF) and TES up to about 565 °C [7,8].

At the time of writing, virtually almost all the commercial CSP plants use nitrate-based molten salt mixtures as TES media because of their properties such as favorable melting temperatures and thermal stabilities for the given thermal application, low viscosity, high heat capacity, low vapor pressure and relatively low costs. However, the maximum working temperature of the nitrate-based molten salt mixtures is limited at about 565 °C due to thermal decomposition [9]. Next-generation (next-Gen) CSP technologies with higher TES/HTF working temperature for higher energy conversion efficiency and lower levelized cost of electricity (LCOE) are being developed, e.g., in the project of U.S. Department of Energy (DOE) - Gen3 CSP [9]. One of the main challenges for the next-Gen CSP technologies is the development of alternative HTF and TES materials with lower costs that could work at temperatures higher than 565 °C of the current nitrate-based molten salt mixtures.

CSP technology has higher solar-to-electrical energy conversion efficiency than photovoltaics (PV) and represents an attractive energy conversion option, since the thermal energy collected from the sun is stored and later transformed into electrical energy through a conventional power block [5]. A conventional commercial CSP plant generally consists of heliostats, a solar receiver, a HTF/TES system and a power generation system [10,11]. As illustrated in Fig. 1, CSP technologies generally are classified into parabolic trough (PT), linear Fresnel (LF), solar tower (ST) and parabolic dish (PD).

Parabolic trough (PT) technology belongs to linear-focusing CSP technologies. It is the preferred choice in working plants due to its high technology maturity and low construction costs. In 2018, ~90% of the commercial CSP plants in operation used this technology [12,13]. As shown in Fig. 1a, its standard configuration includes a set of parabolic mirrors which reflect the sunlight on an absorber tube [14]. PT-type CSP plants generally have TES/HTF working temperatures below 400 °C due to low solar concentration ratio (generally below 100) and thus mostly employ synthetic oils or molten nitrate salts as TES/HTF materials and water steam in the power block [7,8,15]. The current state-of-the-art PT-type CSP has a two-tank TES/HTF configuration using Solar Salt (a mixture of NaNO3–KNO3 60–40 wt%) with a working temperature up to about 385 °C [1,16], which is also considered to be the most reliable and safe TES/HTF system for this type of CSP plants [17]. Some representative operational PT plants include the Solar Electric Generating Station (SEGS I-IX) located in California's Mojave Desert using Therminol as HTF without storage [18], and the Andasol plant, which is the first commercial PT plant built in Europe (southern Spain) with a two-tank indirect configuration using synthetic oil as HTF and molten salt as TES material [19,20].

Linear Fresnel (LF) technology is another linear-focusing CSP technology. As illustrated in Fig. 1b, it uses a series of long flat mirrors placed at different angles to concentrate the sunlight into a line on the receiver [21]. Molten salts have been utilized as TES/HTF in LF-type CSP plants [14,22]. However, the LF-type CSP plants currently are not the mainstreaming CSP plants as the PT-type plants, although their structural simplicity offers low structural support and reflector costs, as well as long focal lengths [23]. SolarPACES (short for Solar Power and Chemical Energy Systems), which is an international cooperative network of International Energy Agency (IEA) for the development and marketing of CSP systems, presents the status of all CSP projects around the world, either operational, under construction, or under development on its official website (solarpaces.org/csp-technologies/csp-projects-around-the-world/). At the time of writing, among the total CSP capacity of about 10 GW, there is about 400 MW LF-type CSP (i.e., about 4% of the total capacity), while PT-type has more than 5 GW capacity, i.e., >50% of the total capacity.

Solar tower (ST) technology is a point-focusing CSP technology, which has much higher solar concentration ratio (in the range of 300–1000) and working temperatures (565 °C or higher) compared to the linear-focusing CSP technology [7]. It is considered as a very promising alternative since higher efficiencies and lower LCOEs are achieved by converting heat into electricity at higher working temperatures [19]. The official website of SolarPACES shows that most CSP plants built recently use this technology, and more than 3 GW ST-type CSPs are either operational, under construction, or under development. This technology comprises a receiver on a tower surrounded by thousands of mirrors that can efficiently achieve high temperatures [24] (see Fig. 1c). Some representative commercial large-scale ST plants are Solar Two and Crescent Dunes in United States, Atacama-1 in Chile, Gemasolar in Spain, and NOOR 3 in Morocco. The state-of-the-art ST CSP is working up to about 565 °C with nitrate-based molten salts as both heat transfer and storage medium (i.e., direct TES system), and the power cycle is a superheated steam Rankine cycle [25].

Parabolic dish (PD) technology is another common point-focusing CSP technology. A PD plant generally consists of a parabolic dish-shaped concentrator, a receiver set on its focal point and a power generation system (e.g., a Stirling engine). The heat from the sun radiation is transferred from the receiver to a Stirling engine for power generation with a working temperature between 700 and 1200 °C [26]. However, although PD has higher solar-to-electrical energy conversion efficiencies owing to higher working temperatures, those systems are still at the demonstration stage due to high installation costs.

Nowadays, molten salt TES technology is the most dominant commercial solution for CSP. The commercial TES capacity with molten salts (almost 17 GWhel) represents about 77% of the globally deployed TES used for electricity applications by the end of 2018 [25,27]. The NaNO3–KNO3 (60–40 wt %, non-eutectic) mixture, called Solar Salt, is used in most commercial CSP plants due to their low costs (~0.8 $/kg [9]), high heat capacity of about 1.52 J/g °C (at 390 °C) and low liquidus temperature of about 250 °C with practical lower operation limit of about 290 °C [28,29]. Next-Gen CSP plants are expected to operate at higher temperatures (>700 °C) than the nitrate mixtures currently in use, in order to achieve higher power cycle efficiencies by using e.g., sCO2 Brayton power cycle, and reduce the LCOE [2] (see Fig. 2). The selection and definition of suitable new salt mixtures is a key aspect to meet the high-efficiency target considered in the next-Gen CSP [9].

In the1950s at Oak Ridge National Laboratory (ORNL) molten chloride salts have been examined as coolants and storage media in advanced nuclear reactors [30]. Despite the first commercial application of molten nitrate salts in CSP technology occurred in 2009 in the Andasol plant [16], molten chloride salts have not been used in commercial CSP plants due to technical challenges such as extremely strong salt corrosivity at high temperatures. According to the SunShot Initiative of DOE [9], the target TES/HTF temperature in the next-Gen CSP technologies is higher than 720 °C. In principle, there are different salt classes to address this temperature. They include sulfates, carbonates, fluorides and chlorides [9]. Considering selection criteria such as melting temperature, salt price, salt density, heat capacity, vapor pressure, corrosivity, thermal stability and toxicity, chloride mixtures were selected as the most promising salt class [9,[31], [32], [33]].

Several authors identified a number of binary and ternary eutectic chloride salt mixtures for high-temperature TES applications. In 2010 Kenisarin [34] summarized several eutectic salts such as KCl–MgCl2, KCl–ZnCl2, LiCl–KCl, NaCl–MgCl2, CaCl2–KCl–MgCl2, CaCl2–KCl–MgCl2, LiCl–BaCl2–KCl, MgCl2–KCl–NaCl with a melting temperature range of 320–487 °C. Wang et al. [35] proposed three different eutectic mixtures of NaCl–KCl–ZnCl2 with a minimum temperature of about 204 °C. One of them is also known as Saltstream 700e®, developed by Halotechnics [19]. Quaternary mixtures have been also investigated by Wei et al. [36] with a melting temperature of 385 °C for the KCl–NaCl–CaCl2–MgCl2 system. Myers et al. [37] investigated a total of 133 chloride salt systems including LiCl, NaCl, KCl, MgCl2, CaCl2, BaCl2 pure salts and their mixtures using a thermodynamic database software - FactSage™. The lowest melting temperatures are achieved with mixtures containing LiCl [38]. However, their price is a disadvantage. AlCl3 and ZnCl2-containing molten salt mixtures have also low melting temperatures but the disadvantage of a relatively high vapor pressure at high temperatures (i.e., a limited maximum upper working temperature) [38,39].

With intensive research in the last few years [[39], [40], [41], [42]], the eutectic MgCl2–KCl–NaCl salt mixture has been identified by different research groups as one of the most promising TES/HTF candidates for the next-Gen CSP due to its low cost (<0.35 $/kg [9,39]), abundance and favorable properties (e.g., low vapor pressure at high temperatures, low melting temperature and high thermal stability). In 1978 Nemecek et al. [43] worked on a latent TES boiler tank using the MgCl2–KCl–NaCl mixture. Fundamental knowledge as sensible TES on this chloride mixture arises from nuclear research of Williams [44] in 2006 as well as Ambrosek [45] in 2011, who explored the use of binary KCl–MgCl2 as HTF in nuclear systems. Since about 2015, there has been a strong renewed interest in the ternary MgCl2–KCl–NaCl mixture as one of the most promising HTF/TES materials in next-Gen CSP due to its abundance and low price [46,47]. Our research group in the German Aerospace Center (DLR) in Germany started in 2015 with the MgCl2–KCl–NaCl ternary mixture for sensible TES [31] and National Renewable Energy Laboratory (NREL) in US considered in 2017 the binary mixture MgCl2–KCl for next-gen CSP [9]. In 2019, Australian National University (ANU) and NREL [40] explored also the MgCl2–KCl–NaCl mixture for sensible heat storage applications in Australia and US. In 2020, Shanghai Institute of Applied Physics, Chinese Academy of Sciences (SINAP-CAS) [41] started intensive study on the MgCl2–KCl–NaCl ternary mixture as HTF/TES in China.

However, MgCl2 is a strong hygroscopic chloride salt and one of the main issues for MgCl2–KCl–NaCl is its strong salt corrosivity due to reactions of impurities (e.g., hydrated water, O2) with the chloride anion, which produce corrosive gases like HCl, Cl2 and corrosive impurities dissolved in the melt like MgOHCl [47]. For this reason, a corrosion control system including corrosion monitoring and mitigation is essential in next-Gen CSP plants to use affordable structural materials (e.g., alloys used in salt tanks) and should be integrated in the molten salt TES/HTF system, as shown in Fig. 2. Several research groups in e.g., NREL, SRNL and DLR, have developed some effective methods for corrosion control and monitoring of metallic structural materials in contact with MgCl2–KCl–NaCl mixture in the lab for potential use on an industrial scale [9,48,49]. Corrosion control strategies of MgCl2-containing molten chlorides for high-temperature HTF/TES will be discussed in Section 3.

There have been some test loops/facilities for next-Gen molten salt TES/HTF technologies (>700 °C) under design, under construction or in operation to alleviate the molten salt technology risks, e.g., FASTR facility of ORNL in US for MgCl2–KCl–NaCl [50], Avanza2 loop of Abengoa in Spain for Li2CO3–Na2CO3–K2CO3 [51]. Compared to Li2CO3–Na2CO3–K2CO3 (>1.3 $/kg, melting temperature of ~397 °C [9,39]), MgCl2–KCl–NaCl (<0.35 $/kg, melting temperature of ~385 °C) has lower material cost and melting temperature, but generally higher corrosivity to metallic materials [9,48,49]. As shown in Fig. 3, the FASTR facility for development of the MgCl2–KCl–NaCl TES/HTF technology is under construction, which includes two parts: (1) salt purification to reduce salt corrosivity and preparation of purified salt for the test loop (Fig. 3 left) and (2) the test loop mainly for testing corrosion control technologies (Fig. 3 right). With the molten salt test loop, the corrosion monitoring and mitigation methods, as well as performance demonstrations of major components such as flanges, heat exchangers and pumps, can be investigated and tested close to real conditions.

For design of a test loop or real TES/HTF system based on molten MgCl2–KCl–NaCl, reliable values of the most relevant thermophysical and physicochemical properties are vital and essential. Moreover, controlling the salt corrosivity is one main concern of the engineering aspects using the molten chlorides at high temperatures. To support engineering of the molten MgCl2–KCl–NaCl salt for high-temperature TES and HTF application, this review presents survey and evaluation of its most relevant thermophysical and physicochemical properties, as well as of the corrosion control strategies for the molten MgCl2–KCl–NaCl salt HTF/TES system. In Section 2, different eutectic compositions of the MgCl2–KCl–NaCl salt system are discussed according to phase diagrams from the literature and simulation using FactSage, in order to determine the exact eutectic composition and melting temperature (minimum working temperature) [52]. Moreover, the maximum working temperature is defined based on the vapor pressure data available in literature and simulated with FactSage. Additionally, this section will review most relevant thermophysical and physicochemical properties of MgCl2–KCl–NaCl for engineering this salt, include heat capacity, salt density, thermal conductivity and dynamic viscosity. In Section 3, the corrosion control strategies on this salt including corrosion monitoring and mitigation are briefly reviewed, since controlling the salt corrosivity is one main concern of the engineering aspects using molten chlorides. Moreover, based on the promising corrosion monitoring and mitigation techniques, a concept of a corrosion control system integrated into the chloride-based TES/HTF system is introduced. In the last section, the recommendations based on Sections 2 Thermophysical and physicochemical properties of molten MgCl, 3 Corrosion control of molten MgCl are summarized for further material studies, as well as modeling, simulation, detailed design and construction of molten MgCl2–KCl–NaCl salt HTF/TES systems.

Section snippets

Thermophysical and physicochemical properties of molten MgCl2–KCl–NaCl salt

As given in Eq. (1), the equation for sensible heat storage Q (J) considers the density of storage materialρ (g/cm3), the volume of storage material V (cm3), the specific heat capacity cp (J/g °C) and the temperature change ΔT (°C) [11,53].Q=ρVcpΔT

Using molten salts with a larger ΔT (i.e., allowed working temperature range) and cp the required material mass (i.e., ρV) and consequently the capital costs of the TES system can be reduced. Besides ΔT and cp, the reliable value of ρ is required for

Corrosion control of molten MgCl2–KCl–NaCl salt

Molten salt corrosion is one of the main challenges to be faced by using chloride mixtures in CSP plants since compared to nitrate mixtures they are strongly corrosive to structural materials (e.g., metallic alloys) at high working temperatures [9,47]. In order to ensure long lifetime and economy of the molten chloride TES system, an effective and affordable corrosion control system (CCS) including corrosion monitoring and mitigation is essential and should be integrated in the TES system [55].

Summary and recommendations

The MgCl2–KCl–NaCl salt mixture has shown to be a promising TES/HTF material in next-Gen CSP plants with large working temperature range of 420–800 °C, excellent salt properties and low material cost of <0.35 USD/kg. By reviewing literature and comparison with the thermodynamic simulations, this work gives recommendations on the exact composition of the eutectic MgCl2–KCl–NaCl salt mixture (47.1–22.7–30.2 mol. %) and its salt properties relevant for engineering this salt. The properties include

CRediT authorship contribution statement

Carolina Villada: Conceptualization, Investigation, Writing – original draft, preparation, Visualization. Wenjin Ding: Conceptualization, Investigation, Writing – original draft, preparation, Visualization, review & editing, revision. Alexander Bonk: Writing – review & editing. Thomas Bauer: Writing – review & editing, Funding acquisition.

Declaration of competing 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.

Acknowledgements

This research has been performed within the DLR-DAAD fellowship programme (Grant number 91716495), which is funded by German Academic Exchange Service (DAAD) and German Aerospace Center (DLR).

References (112)

  • F. De Luca et al.

    On the performance of CSP oil-cooled plants, with and without heat storage in tanks of molten salts

    Energy

    (2015)
  • M. Liu et al.

    Review on storage materials and thermal performance enhancement techniques for high temperature phase change thermal storage systems

    Renew. Sustain. Energy Rev.

    (2012)
  • K. Vignarooban et al.

    Heat transfer fluids for concentrated solar power systems – a review

    Appl. Energy

    (2015)
  • U. Desideri et al.

    Analysis and comparison between a concentrated solar and a photovoltaic power plant

    Appl. Energy

    (2014)
  • N. Kincaid et al.

    An optical performance comparison of three Concentrated solar power collector designs in linear Fresnel, parabolic trough, and central receiver

    Appl. Energy

    (2018)
  • G. Morin et al.

    Molten salt as a heat transfer fluid in a linear Fresnel collector – commercial application backed by demonstration

    Energy Procedia

    (2015)
  • M.R. Rodríguez-Sánchez et al.

    Comparison of simplified heat transfer models and CFD simulations for molten salt external receiver

    Appl. Therm. Eng.

    (2014)
  • M. Aneke et al.

    Energy storage technologies and real life applications – a state of the art review

    Appl. Energy

    (2016)
  • T. Wang et al.

    Novel high thermal stability LiF–Na2CO3–K2CO3 eutectic ternary system for thermal energy storage applications

    Sol. Energy Mater. Sol. Cells

    (2015)
  • A.G. Fernández et al.

    Mainstreaming commercial CSP systems: a technology review

    Renew. Energy

    (2019)
  • A. Bonk et al.

    Advanced heat transfer fluids for direct molten salt line-focusing CSP plants

    Prog. Energ. Combust.

    (2018)
  • D. LeBlanc

    Molten salt reactors: a new beginning for an old idea

    Nucl. Eng. Des.

    (2010)
  • J. Wang et al.

    Temperature and composition dependences of shear viscosities for molten alkali metal chloride binary systems by molecular dynamics simulation

    J. Mol. Liq.

    (2019)
  • M.M. Kenisarin

    High-temperature phase change materials for thermal energy storage

    Renew. Sustain. Energy Rev.

    (2010)
  • X. Wei et al.

    Quaternary chloride eutectic mixture for thermal energy storage at high temperature

    Energy Procedia

    (2015)
  • P.D. Myers et al.

    Thermal energy storage using chloride salts and their eutectics

    Appl. Therm. Eng.

    (2016)
  • G. Mohan et al.

    Sensible energy storage options for Concentrated solar power plants working above 600 °C

    Renew. Sustain. Energy Rev.

    (2019)
  • X. Li et al.

    Unrevealing the thermophysical properties and microstructural evolution of MgCl2–NaCl–KCl eutectic: FPMD simulations and experimental measurements

    Sol. Energy Mater. Sol. Cells

    (2020)
  • J.J. Nemecek et al.

    Demand sensitive energy storage in molten salts

    Sol. Energy

    (1978)
  • G. Mohan et al.

    Assessment of a novel ternary eutectic chloride salt for next generation high-temperature sensible heat storage

    Energy Convers. Manag.

    (2018)
  • W. Ding et al.

    Molten chloride salts for next generation CSP plants: electrolytical salt purification for reducing corrosive impurity level

    Sol. Energy Mater. Sol. Cells

    (2019)
  • C. Prieto et al.

    Carbonate molten salt solar thermal pilot facility: plant design, commissioning and operation up to 700 °C

    Renew. Energy

    (2020)
  • C.W. Bale et al.

    FactSage thermochemical software and databases, 2010–2016

    Calphad

    (2016)
  • C. Villada et al.

    Design and development of nitrate-nitrite based molten salts for Concentrated solar power applications

    Sol. Energy

    (2019)
  • W. Ding et al.

    Progress in research and development of molten chloride salt technology for next generation concentrated solar power plants

    Engineering

    (2021)
  • Y. Li et al.

    Survey and evaluation of equations for thermophysical properties of binary/ternary eutectic salts from NaCl, KCl, MgCl2, CaCl2, ZnCl2 for heat transfer and thermal storage fluids in CSP

    Sol. Energy

    (2017)
  • X. Xu et al.

    Basic properties of eutectic chloride salts NaCl-KCl-ZnCl2 and NaCl-KCl-MgCl2 as HTFs and thermal storage media measured using simultaneous DSC-TGA

    Sol. Energy

    (2018)
  • S. Polimeni et al.

    Comparison of sodium and KCl-MgCl2 as heat transfer fluids in CSP solar tower with sCO2 power cycles

    Sol. Energy

    (2018)
  • L. Du et al.

    Thermal properties and thermal stability of the ternary eutectic salt NaCl-CaCl2-MgCl2 used in high-temperature thermal energy storage process

    Appl. Energy

    (2017)
  • K. Vignarooban et al.

    Vapor pressure and corrosivity of ternary metal-chloride molten-salt based heat transfer fluids for use in concentrated solar power systems

    Appl. Energy

    (2015)
  • J.C. Gomez-Vidal et al.

    Corrosion resistance of alumina-forming alloys against molten chlorides for energy production. I: pre-oxidation treatment and isothermal corrosion tests

    Sol. Energy Mater. Sol. Cells

    (2017)
  • G.J. Janz

    I.B - Density

  • R. Serrano-López et al.

    Molten salts database for energy applications

    Chem. Eng. Process. - Process Intensification

    (2013)
  • Y. Marcus

    Volumetric behavior of molten salts

    Thermochim. Acta

    (2013)
  • A. Palacios et al.

    Thermal conductivity measurement techniques for characterizing thermal energy storage materials – a review

    Renew. Sustain. Energy Rev.

    (2019)
  • A.E. Gheribi et al.

    Recommended values for the thermal conductivity of molten salts between the melting and boiling points

    Sol. Energy Mater. Sol. Cells

    (2014)
  • J.J. Healy et al.

    The theory of the transient hot-wire method for measuring thermal conductivity

    Phys. B+C

    (1976)
  • H. Tian et al.

    Thermal conductivities and characteristics of ternary eutectic chloride/expanded graphite thermal energy storage composites

    Appl. Energy

    (2015)
  • N. Ren et al.

    Preparation and thermal properties of quaternary mixed nitrate with low melting point

    Sol. Energy Mater. Sol. Cells

    (2014)
  • K. Sridharan et al.

    Corrosion in molten salts (Chapter 12)

  • Cited by (52)

    View all citing articles on Scopus
    View full text