- Email: [email protected]
nitrite molten salt in concentrated solar technology
Renewable Energy 80 (2015) 177e183
Contents lists available at ScienceDirect
Renewable Energy journal homepage: www.elsevier.com/locate/renene
Corrosion properties of a ternary nitrate/nitrite molten salt in concentrated solar technology ndez a, b, c, *, M. Cortes b, c, E. Fuentealba b, c, F.J. Pe rez a A.G. Ferna a
Surface Engineering and Nanostructured Materials Research Group, Complutense University of Madrid, Av. Complutense s/n, Madrid, Spain Energy Development Center, University of Antofagasta, Avenue Universidad de Antofagasta 02800, Antofagasta, Chile c Solar Energy Research Center (SERC-Chile), Av Tupper 2007, Piso 4, Santiago, Chile b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 12 November 2014 Accepted 30 January 2015 Available online
The enhancements in the storage systems developed by solar thermal power plants have provided to renewable energy a considerable increase in efficiency. Thermal Energy Storage (TES) using HITEC mixture could be used as Heat Transfer Fluid (HTF) in concentrated solar linear technology. In this research, the corrosive effects of HITEC mixture composed by 40 wt% NaNO2 þ 7 wt % NaNO3 þ 53 wt% KNO3 were assessed at 390 C on a carbon steel (A516) and on low-Cr alloy steels (T11 and T22). The corrosion rates were determined by gravimetric tests, measuring the weight gain during 2000 h, identifying the corrosion products via scanning electron microscopy and X-ray diffraction. T22 steel shows a corrosion layer of 6.05 microns, with a protective layer formed in the inner zone to the material, identified through DRX as the K2CrO4 protective spinel. Fe2O3 and MgO were the others important products found on the tests performed at 390 C, being observed also the formation of some stable compounds with the impurities of the salt, as carbonates. The use of the HITEC mixture in solar technology would provide a less aggressive behaviour for materials in contact with it, providing an increase in operational life cycles in current solar technology. © 2015 Elsevier Ltd. All rights reserved.
Keywords: Thermal energy storage (TES) Molten salt Corrosion Carbon steel
1. Introduction Molten salt are widely applied as heat transfer media in Concentrating Solar Power (CSP) plants. An important component of Thermal Energy Storage (TES) systems is the choice of heat transfer fluids used in a solar plant. The main challenges to making electrical power generation from solar thermal plants more profitable are the following:
Reducing the investment, operational and maintenance costs; Extending the hours of operation and energy supply; Increasing the temperatures in the thermal cycle; Increasing the lifetime of power plants.
Research and development in the solar energy sector has focused on reducing the high cost associated with the operation
* Corresponding author. Energy Development Center, University of Antofagasta, Avenue Universidad de Antofagasta 02800, Antofagasta, Chile. E-mail address: [email protected] (A.G. Fern andez). http://dx.doi.org/10.1016/j.renene.2015.01.072 0960-1481/© 2015 Elsevier Ltd. All rights reserved.
and maintenance of these plants. One of the most important lines of research in this regard is the study, design and characterisation of the salts that are used as energy storage fluids. The thermal properties of the HITEC mixture, a molten salt composed of 40 wt% NaNO2 þ 7 wt% NaNO3 þ 53 wt% KNO3, are highly suited to its use as storage fluid in CSP plants [1]. This ternary mixture was first proposed for use as a storage fluid in the THEMIS experimental solar plant in France in the 1980s [2,3]. It is worth noting that one of the disadvantages of this mixture is the need to use a protective layer of inert gas at temperatures above 350 C to avoid the oxidation of the nitrites upon contact with the oxygen in the atmosphere. This is the main reason which prevents at present the widespread use of this salt on thermoelectric solar plants. However the use of the HITEC mixture in solar technology is once again under consideration, albeit for specific applications only. A number of research projects are currently seeking to improve the physical properties of the mixture to a point in which the mixture enhanced properties could compensate its higher costs. To this end, M. Xi Ho and R. Olivares [1,4] tried to improve the mixture's thermal properties as well as its thermal stability under
ndez et al. / Renewable Energy 80 (2015) 177e183 A.G. Ferna
178
different atmospheres, to allow for its use as Heat Transfer Fluid in concentrated linear solar technologies (Parabolic Trough Collectors and Linear Fresnel). The present research introduces a new parameter to be taken into account when using the mixture as solar energy storage fluid, the corrosion potential. This is parameter seldom reported in the bibliography but of vital importance when developing a new generation of solar plants that are both cheaper and profitable. The present study shows the results of an isothermal corrosion study at 390 C (the storage temperature in solar plants using parabolic trough collectors) in a number of carbon steels and with low Cr content. The objective of this study is to establish the influence of the Cr additions in commercial steels, for which a comparative study was carried out in the commercial steel A516, which is currently used as a building material in CSP technology.
Table 1 Chemical composition of HITEC salt. Parameter
Initial values
Cl (%) SO4 2 (mg/g) Ca (mg/g) Cr (mg/g) Fe (mg/g) Mg (mg/g) Moisture (%) NO2 (mg/g)
0,0795 431,57 35,76 <1 <1 70,14 0,14737 268001,4
The corrosive effect of these salts is based on the following reduction reaction:
process to reduce as much as possible their content in the melt. When the temperature increases, the perchlorate ion starts to dissociate into oxygen and chloride, which exponentially increases the corrosion of materials in contact with this environment. The amount of perchlorate ðClO4 Þ can be reduced by increasing the amount of chlorides in the melt. It is important to note that the perchlorate ion is an impurity present in the commercial salts used in this research.
NO3 þ 2e 4NO2 þ O2
ClO4 /2O2 þ Cl
1.1. Corrosion processes in molten nitrate salts
(1)
which results in the oxidation of iron atoms that diffuse from the material [5,6]: Fe þ O2 4 FeO þ 2e
(2)
3FeO þ O2 4 Fe3O4 þ 2e
(3)
To understand this process in alkaline nitrates, it is important to note the formation of several different oxidized ions during the corrosion tests. Experiments performed by I.B. Singh et al. [7] indicate the existence of O2 (oxide), O2 2 (peroxide) and O2 (superoxide), which arise from unstable oxide ions in the nitrate melt, as described by the following equations:
O2 þ NO3 4NO2 þ O2 2
(4)
O2 2 þ 2NO3 42NO2 þ 2O2
(5)
Several authors have studied the formation of these oxides in these salts during the corrosion process [8,9], which results in the formation of Na2O and KO2; the Kþ and Naþ ions have different affinities for the ions formed in equations (5) and (6). 2Naþ þ O2 / Na2O
(6)
2Kþ þ 2 O 2 / KO2
(7)
The formation of these oxides hinders the electronic movement required to generate the cathodic reaction, which decreases the corrosiveness of the salts. The salt mixtures studied in this paper are commercial grade, so impurities will be a factor to consider in the analysis and discussion of the thermal properties and corrosion potential. The common oxidant in molten salts, which accelerate corrosion, are O2, H2O, Hþ and OH [10] Furthermore, water may dissociate into Hþ and OH, and Cl2 can be present in the molten salt after reacting with oxygen:
2
2Cl þ ½ O2 4 Cl2 þ O
The presence of both, oxygen and chlorine, increases the corrosion ability of these salts. Kleppa et al. [11] determined the influence of steric hindrance on the formation of compounds of ClO4 2 with Na or K ions. The formation of KClO4 was favoured because K ion is smaller, which a melting point of 510 C. The objective of this study is to establish the corrosion ability in HITEC mixture in order to evaluate the most appropriate materials as well to determine the plant's life cycle. 2. Material and methods The saline nitrates that were used in the research were NaNO2, NaNO3 and KNO3 (SigmaeAldrich 98%). The impurity content of tested Hitec salt is shown in Table 1 and was determined using the Volhard procedure (Cl), Gravimetry (% moisture, SO4 2 ), UVeVis spectroscopy (NO2 ) and ICP-OES (Mg). The importance of the studied parameters is based on the following factors: Cl: Appears as a corrosion enhancing agent. A slight increase from the initial value is due to the decomposition reaction of the perchlorate in the mixtures. SO4 2 : An important parameter to follow because sulphates are able to form insoluble compounds that can clog pumps and pipes, which affects the circulation of the salt in the solar power plant. Mg is an important impurity in the salt that forms oxides and is able to form magnesiumferrite, a very stable compound formed by the interaction between Mg and hematite (Fe2O3). The corrosion tests were performed on commercial steels with the compositions shown in Table 2 at storage temperature used in parabolic trough solar power plants (390 C).
Table 2 Chemical composition of carbon steels studied. Steel
(8)
Perchlorates are present in commercial molten nitrates as impurities; these impurities need to be controlled during the melting
(9)
A516 T11 T-22
Weight % Si
Mn
Cr
P
Mo
C
S
0.1 0.79 0.3
0.93 0.44 0.4
1.2 2.25
0.035 0.008 0.3
0.5 1
0.27 0.1 0.12
0.035 0.002 0.3
ndez et al. / Renewable Energy 80 (2015) 177e183 A.G. Ferna
179
The specimens used in the corrosion test have dimensions of 20 10 2 mm. They were polished with #600 SiC abrasive papers and washed with distilled water and acetone. The isothermal immersion tests were analysed via gravimetric analysis at 24, 48, 150, 350, 500, 750, 1000, 1250, 1500, 1750 and 2000 h; two specimens of each alloy were removed from the nitrate mixtures for examination and analysis. Once the samples were removed from the oven they were cooled slowly in warm distilled water in order to eliminate the salt in which they had been immersed. They were then dried and weighed, average taken from five values of their weights. The formula (equation I) used to calculate the mass gain over time is:
mi mf Dm ¼ S0 S0 Fig. 1. Gravimetric behaviour of steels in 40% NaNO2 þ 7% NaNO3 þ 53% KNO3 at 390 C for 2000 h of test.
[Ec. I]
where mi is the initial mass of the specimen, mf is the mass of the sample at time t and S0 is the initial area of the specimen. 3. Results and discussion
Fig. 2. Surficial image of A516 after 2000 h of test.
The steels under analysis at temperatures of 390 C are carbon steels with low chrome content (see Table 2 for their composition) which were immersed in molten salt for 2000 h. The increase in gravimetric mass experienced by these steels is shown on Fig. 1. All steels analysed on the saline medium showed similar behaviour against corrosion, with increases in mass of around 0.37 mg/cm2. Individual studies of each steel where made, in order to expound the corrosion processes affecting carbon and low Cr steels in the experiment and the associated increase in mass. The surface of the A516 steel at the end of the experiment (2000 h) shows a thin homogeneous layer of corrosion with oxygen and iron content (Fig. 2). Together with the formation of small needles (with a greater content in calcium, analysis Q5-2) and other less defined particles, in which magnesium and calcium (Q5-1) can be detected. The analysis of a cross section of the sample at the end of the experiment is shown on Fig. 3. The steel shows a good behaviour against 2 corrosion, with a layer.66 microns deep. EDX (Energy Dispersive X-ray) analysis of this layer shows its main elements to be iron, oxygen and magnesium. X-ray analysis (XRD, X-ray diffraction) showed the presence of the following corrosion by-products (Fig. 4). The corrosion layer was composed of hematite (Fe2O3) and magnetite (Fe3O4). MgO and CaCO3 were also detected, originating from the impurities of the salt; their formation in the surface of the steel explains the levels found in the analysis shown in Fig. 2.
Fig. 3. Cross section of A516 steel after 2000 h of test.
180
ndez et al. / Renewable Energy 80 (2015) 177e183 A.G. Ferna
The corroded layer which formed at the end of the experiment (2000 h) was 2.66 microns deep, with Fe2O3 and MgO present. Fe2O3 and MgO had not yet reacted to form MgFe2O4, a compound which has been detected in previous studies with layers over 5 micron thickness [12]. The formation of Fe2O3 in preference to other compounds in this reactive medium has been studied by Picard et al. [13], in which formation of this oxide is related to the values of PO2 in the salt. The thermodynamic calculations based on potential PO2 relationships (vertical axis shows in Fig. 5) suggested that the basicity of the melt increases with temperature increase, which favours incorporation of impurities in the scale. These values affect the passivation quality of the oxidised layers as well as its associated kinetic abilities. The authors showed the balance between the oxides and nitrates in a diagram (Fig. 5). These diagrams have been created based on the thermodynamic constants of the reactions:
Fig. 4. XRD of A516 steel after 2000 h of test.
Fig. 5. Equilibrium diagram of nitrides and iron oxides.
NO3 4NO2 þ 1=2O2
(10)
3NO2 42NO þ NO3 þ O2
(11)
The study of the surface of the steel T11 (Fig. 6) shows an oxidised layer composed of magnesium, iron and oxygen (analysis R53) which two different types of particles are adhered: small particles in the shape of needles with calcium content (R5-2) and larger spherical particles, with higher contents of oxygen and magnesium. The analysis of a cross section of the sample (Fig. 7) shows a metallographic layer of 7.05 microns, almost wholly composed of iron and oxygen. The x-ray analysis (Fig. 8) confirms the presence of corrosion byproducts first glimpsed in the EDX analysis as part of the metallographic study using scanning electron microscopy. Hematite and Magnetite were identified as the two components of the oxidised layer, together with calcium carbonate on the surface of the material. MgO did not seem to be present, owing to the particles forming only occasionally on the surface of the sample. Other stable compounds in this medium that appear at elevated temperatures are carbonate species. C.M. Kramer et al. [14] studied the formation of carbonates via the interactions of nitrates with atmospheric CO2 in this medium, which resulted in the following reactions: 2NaNO3 þ CO2 / Na2CO3 þ N2 þ 5/2O2
Fig. 6. Surficial image of T11 after 2000 h of test.
(12)
ndez et al. / Renewable Energy 80 (2015) 177e183 A.G. Ferna
181
Fig. 7. Cross section of a) T11 steel and b) EDX study during 2000 h of test.
Na2O þ CO2 / Na2CO3
(13)
Similarly, CO3 2 ions formed in the salt can interact with the calcium or magnesium present in the NaNO3 and KNO3 salts as well as the other ionic components in these salts. The surface of the T22 steel after the end of the 2000 h experiments is shown in Fig. 9. Both the image obtained and the EDX analysis show a greater number of acicular structures with a calcium content (S5-3), together with a layering of the same structures with a whitish appearance, corresponding with the S5-5 analysis. These growth structures appear on top of an oxidised layer and are mainly composed of iron, oxygen and magnesium. The cross section of the sample after 2000 h immersed in the ternary mixture of nitrite is shown in Fig. 10. The oxidised layer detected through the EDX analysis of the steel T22 is composed of magnesium, iron and oxygen and a thickness of 6.05 microns. This increase in the oxidised layer is due to its protective character as the chrome present in the base material diffuses from the substrate to increase weight gains, as registered in the gravimetric curve in Fig. 1. However this protective character gives the material better performance against corrosion and isothermal experiment could be
Fig. 8. XRD of T11 steel after 2000 h of test.
extended during the time, maintaining the oxidising parabolic kinetic presently shown. The x-ray study, detected the presence of the corrosion products shown in Fig. 11. In this instance, it appears that Magnesium ferrite (MgFe2O4) was formed, and the analysis shows it is the main component of the oxides generated on the surface of the steel. In this regards, MgO and K2CrO4 were also present, the latter giving a certain level of protection to the sample. The salt used in the corrosion experiments was analysed with x rays to detect any components that could had been detached from the base material as well as any insoluble components of the salts. Since the behaviour of the steels has been shown to be very similar, the by-products of the corrosion were exemplified in the diffractogram (Fig. 12a) and the diffractograms obtained of the three steels have been represented (Fig. 12b) as the only differences shown were with regards to the values obtained of the referenced by products. Hematite (Fe2O3) and NaFeO2 were shown to be the main products of the corrosion, appearing at levels similar to the ones
Fig. 9. Surficial image of T22 after 2000 h of test.
182
ndez et al. / Renewable Energy 80 (2015) 177e183 A.G. Ferna
Fig. 10. Cross section of T22 steel after 2000 h of test.
obtained for this compounds in the salts used to submerge the steels. The excellent performance show by the ternary mixture of nitrite was evaluated by S. Al Omer et al. [15], showing the importance of Na2O2 peroxide to that effect. In this regard, the mixture starts losing weight as the reaction proceeds: Na2O2 4 Na2O þ ½ O2
Fig. 11. XRD of T22 steel after 2000 h of test.
(14)
This would justify the presence of Na2O in the ternary nitrite salt analysed at the end of the corrosion experiment. The results of the chemical analysis of the major parameters of the salt properties are shown in Table 3. Chloride levels in the salt are the lowest detected in the new molten salt formulation with low melting point present in the literature [16e20]. Equally, the chemical analysis on the remaining parameters show similar values to those obtained initially, and there are no traces of chrome or iron content in the salt at the end of the experiment, as was to be expected, in view of the salt's optimum performance against corrosion.
Fig. 12. XRD of HITEC salt after corrosion test a) Salt in contact with A516 b) difractogram comparison.
ndez et al. / Renewable Energy 80 (2015) 177e183 A.G. Ferna Table 3 Chemical composition of HITEC salt after corrosion test compared with the initial values. Parameter
Initial values
Salt A516
Salt T11
Salt T22
Cl (%) SO4 2 (mg/g) Ca (mg/g) Cr (mg/g) Fe (mg/g) Mg (mg/g) Moisture (%) NO2 (mg/g)
0,0795 431,57 35,76 <1 <1 70,14 0,14737 268001,4
0,094 1029 65 e <1 96 0,15 244091
0,091 915 48 <1 <1 59 0,14 260809
0,09 903 53 <1 <1 64 0,11 261303
Moisture level detected in the salt remains, was also close to the initial content, and in particular the level of nitrites shows a slight reduction from the initial level analysed. This last fact is of importance and worth noting since it will imply that the thermal properties of the HITEC salt do not decrease after the change in the final composition of the mixture. The thermal study of this mixture after the corrosion experiment will be duly analysed in future research. 4. Conclusions The corrosive properties of the salt were evaluated at 390 C, simulating the normal working conditions of a solar plant using parabolic trough collectors in contact with carbon steel and low chrome steels. The three steels showed excellent behaviour against corrosion, with mass gains of around 0.35 mg/cm2. A516 steel showed a metallography layer 2.66 microns deep, composed mainly of magnetite and hematite. T11 steel, as deduced by the gravimetric gain, shows a thickness layer of 7.05 microns, due to the occasional presence on its surface of adhered particles of MgO. T22 steel shows a corrosion layer of 6.05 microns, with a protective layer formed in the inner zone to the material, identified through XRD as the K2CrO4 protective spinel. This compound will give to the material a protective character over 2000 h of test, so the corrosion properties of the HITEC salt will not be an impediment to its use in the current concentrated solar power technology. Acknowledgement The authors would like to acknowledge the financial support provided by CONICYT/FONDAP 15110019 “Solar Energy Research
183
Center” SERC-Chile, Fondecyt Postdoctoral grant n 3140014, FIC-R 30137092 funded by Atacama Government and the Education Ministry of Chile Grant PMI ANT 1201.
References [1] Ho MX, Pan C. Optimal concentration of alumina nanoparticles in molten hitec salt to maximize its specific heat capacity. Int J Heat Mass Transf 2014;70(0): 174e84. [2] Mochinaga J. Molar volumes of the molten NaNO3-NaNO2-KNO3 system. J Chem Eng Data 1985;30:274e6. [3] Janz GJ. Melting and premelting properties of the KNO3-NaNO2-NaNO3 eutectic system. J Chem Eng 1983;28:201e2. [4] Olivares RI. The thermal stability of molten nitrite/nitrates salt for solar thermal energy storage in different atmospheres. Sol Energy 2012;86(9): 2576e83. [5] Baraka A, El Hosary A. Corrosion on mild steel in molten sodium nitratepotassium nitrate eutectic. Br Corros J 1976;11(1):44e6. [6] De Jong JM, Broers GHJ. A reversible oxygen electrode in an equimolar KNO3dNaNO3 melt saturated with sodium peroxide II. A voltammetric study. Electrochimica Acta 1976;21(11):893e900. [7] Singh IB. The effect of NaCl addition on the corrosion of mild steel in NaNO3 melt. Corros Sci 1993;34(10):1733e42. [8] Palmisano F. Catalytic currents in fused salts: discharge mechanism of nitrite in molten alkali nitrates(II). J Electroanal Chem Interfacial Electrochem 1978;89(2):311e20. [9] Desimoni E, Palmisano F, Zambonin PG. Catalytic currents in fused salts: discharge mechanism of nitrite in molten alkali nitrates. J Electroanal Chem Interfacial Electrochem 1977;84(2):315e22. [10] Fernandez AG, Galleguillos H, Perez FJ. Thermal influence in corrosion properties of Chilean solar nitrates. Sol Energy 2014;109(0):125e34. [11] Kleppa OJ. Thermochemistry of anion mixtures in simple fused salt systems. II 2 solutions of some salts of MO 4 and MO4 anions in the corresponding alkali nitrates. J Phys Chem 1963;67(12):2750e3. [12] Fernandez AG, Galleguillos H, Perez FJ. Corrosion ability of a novel heat transfer fluid for energy storage in CSP plants. Oxid Metals 2014;82:331e45. http://dx.doi.org/10.1007/s11085-014-9494-3. [13] Picard G, L H. Thermodynamic study of corrosion of iron in NaNO3-NaNO2 mixtures. J Electrochem Soc 1987;134(1). [14] Kramer CM, Munir ZA, Volponi JV. Screening tests of sodium nitrate and potassium nitrate decomposition. Sol Energy 1982;29(5):437e9. [15] Al Omer S. Molten sodium nitrite-potassium nitrite eutectic: the reactions of sodium peroxide. J Inorg Nucl Chem 1978;40:975e8. [16] Fernandez AG, Lasanta MI, Perez FJ. Molten salt corrosion of stainless steels and low-Cr steel in CSP plants. Oxid Metals 2012;78(5e6):329e48. [17] Bradshaw RW, Meeker DE. High-temperature stability of ternary nitrate molten salts for solar thermal energy systems. Sol Energy Mater 1990;21(1): 51e60. [18] Goods SH, Bradshaw RW. Corrosion of stainless steels and carbon steel by molten mixtures of commercial nitrate salts. J Mater Eng Perform 2003;13. [19] Robert W, Bradshaw WMC. Effect of chloride content of molten nitrate salt on corrosion of A516 carbon steel. Sandia report. 2010. [20] Bradshaw R., Low melting point inorganic nitrate salt heat transfer fluid, United States Patent, 2009, US7588694B1.
Contents lists available at ScienceDirect
Renewable Energy journal homepage: www.elsevier.com/locate/renene
Corrosion properties of a ternary nitrate/nitrite molten salt in concentrated solar technology ndez a, b, c, *, M. Cortes b, c, E. Fuentealba b, c, F.J. Pe rez a A.G. Ferna a
Surface Engineering and Nanostructured Materials Research Group, Complutense University of Madrid, Av. Complutense s/n, Madrid, Spain Energy Development Center, University of Antofagasta, Avenue Universidad de Antofagasta 02800, Antofagasta, Chile c Solar Energy Research Center (SERC-Chile), Av Tupper 2007, Piso 4, Santiago, Chile b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 12 November 2014 Accepted 30 January 2015 Available online
The enhancements in the storage systems developed by solar thermal power plants have provided to renewable energy a considerable increase in efficiency. Thermal Energy Storage (TES) using HITEC mixture could be used as Heat Transfer Fluid (HTF) in concentrated solar linear technology. In this research, the corrosive effects of HITEC mixture composed by 40 wt% NaNO2 þ 7 wt % NaNO3 þ 53 wt% KNO3 were assessed at 390 C on a carbon steel (A516) and on low-Cr alloy steels (T11 and T22). The corrosion rates were determined by gravimetric tests, measuring the weight gain during 2000 h, identifying the corrosion products via scanning electron microscopy and X-ray diffraction. T22 steel shows a corrosion layer of 6.05 microns, with a protective layer formed in the inner zone to the material, identified through DRX as the K2CrO4 protective spinel. Fe2O3 and MgO were the others important products found on the tests performed at 390 C, being observed also the formation of some stable compounds with the impurities of the salt, as carbonates. The use of the HITEC mixture in solar technology would provide a less aggressive behaviour for materials in contact with it, providing an increase in operational life cycles in current solar technology. © 2015 Elsevier Ltd. All rights reserved.
Keywords: Thermal energy storage (TES) Molten salt Corrosion Carbon steel
1. Introduction Molten salt are widely applied as heat transfer media in Concentrating Solar Power (CSP) plants. An important component of Thermal Energy Storage (TES) systems is the choice of heat transfer fluids used in a solar plant. The main challenges to making electrical power generation from solar thermal plants more profitable are the following:
Reducing the investment, operational and maintenance costs; Extending the hours of operation and energy supply; Increasing the temperatures in the thermal cycle; Increasing the lifetime of power plants.
Research and development in the solar energy sector has focused on reducing the high cost associated with the operation
* Corresponding author. Energy Development Center, University of Antofagasta, Avenue Universidad de Antofagasta 02800, Antofagasta, Chile. E-mail address: [email protected] (A.G. Fern andez). http://dx.doi.org/10.1016/j.renene.2015.01.072 0960-1481/© 2015 Elsevier Ltd. All rights reserved.
and maintenance of these plants. One of the most important lines of research in this regard is the study, design and characterisation of the salts that are used as energy storage fluids. The thermal properties of the HITEC mixture, a molten salt composed of 40 wt% NaNO2 þ 7 wt% NaNO3 þ 53 wt% KNO3, are highly suited to its use as storage fluid in CSP plants [1]. This ternary mixture was first proposed for use as a storage fluid in the THEMIS experimental solar plant in France in the 1980s [2,3]. It is worth noting that one of the disadvantages of this mixture is the need to use a protective layer of inert gas at temperatures above 350 C to avoid the oxidation of the nitrites upon contact with the oxygen in the atmosphere. This is the main reason which prevents at present the widespread use of this salt on thermoelectric solar plants. However the use of the HITEC mixture in solar technology is once again under consideration, albeit for specific applications only. A number of research projects are currently seeking to improve the physical properties of the mixture to a point in which the mixture enhanced properties could compensate its higher costs. To this end, M. Xi Ho and R. Olivares [1,4] tried to improve the mixture's thermal properties as well as its thermal stability under
ndez et al. / Renewable Energy 80 (2015) 177e183 A.G. Ferna
178
different atmospheres, to allow for its use as Heat Transfer Fluid in concentrated linear solar technologies (Parabolic Trough Collectors and Linear Fresnel). The present research introduces a new parameter to be taken into account when using the mixture as solar energy storage fluid, the corrosion potential. This is parameter seldom reported in the bibliography but of vital importance when developing a new generation of solar plants that are both cheaper and profitable. The present study shows the results of an isothermal corrosion study at 390 C (the storage temperature in solar plants using parabolic trough collectors) in a number of carbon steels and with low Cr content. The objective of this study is to establish the influence of the Cr additions in commercial steels, for which a comparative study was carried out in the commercial steel A516, which is currently used as a building material in CSP technology.
Table 1 Chemical composition of HITEC salt. Parameter
Initial values
Cl (%) SO4 2 (mg/g) Ca (mg/g) Cr (mg/g) Fe (mg/g) Mg (mg/g) Moisture (%) NO2 (mg/g)
0,0795 431,57 35,76 <1 <1 70,14 0,14737 268001,4
The corrosive effect of these salts is based on the following reduction reaction:
process to reduce as much as possible their content in the melt. When the temperature increases, the perchlorate ion starts to dissociate into oxygen and chloride, which exponentially increases the corrosion of materials in contact with this environment. The amount of perchlorate ðClO4 Þ can be reduced by increasing the amount of chlorides in the melt. It is important to note that the perchlorate ion is an impurity present in the commercial salts used in this research.
NO3 þ 2e 4NO2 þ O2
ClO4 /2O2 þ Cl
1.1. Corrosion processes in molten nitrate salts
(1)
which results in the oxidation of iron atoms that diffuse from the material [5,6]: Fe þ O2 4 FeO þ 2e
(2)
3FeO þ O2 4 Fe3O4 þ 2e
(3)
To understand this process in alkaline nitrates, it is important to note the formation of several different oxidized ions during the corrosion tests. Experiments performed by I.B. Singh et al. [7] indicate the existence of O2 (oxide), O2 2 (peroxide) and O2 (superoxide), which arise from unstable oxide ions in the nitrate melt, as described by the following equations:
O2 þ NO3 4NO2 þ O2 2
(4)
O2 2 þ 2NO3 42NO2 þ 2O2
(5)
Several authors have studied the formation of these oxides in these salts during the corrosion process [8,9], which results in the formation of Na2O and KO2; the Kþ and Naþ ions have different affinities for the ions formed in equations (5) and (6). 2Naþ þ O2 / Na2O
(6)
2Kþ þ 2 O 2 / KO2
(7)
The formation of these oxides hinders the electronic movement required to generate the cathodic reaction, which decreases the corrosiveness of the salts. The salt mixtures studied in this paper are commercial grade, so impurities will be a factor to consider in the analysis and discussion of the thermal properties and corrosion potential. The common oxidant in molten salts, which accelerate corrosion, are O2, H2O, Hþ and OH [10] Furthermore, water may dissociate into Hþ and OH, and Cl2 can be present in the molten salt after reacting with oxygen:
2
2Cl þ ½ O2 4 Cl2 þ O
The presence of both, oxygen and chlorine, increases the corrosion ability of these salts. Kleppa et al. [11] determined the influence of steric hindrance on the formation of compounds of ClO4 2 with Na or K ions. The formation of KClO4 was favoured because K ion is smaller, which a melting point of 510 C. The objective of this study is to establish the corrosion ability in HITEC mixture in order to evaluate the most appropriate materials as well to determine the plant's life cycle. 2. Material and methods The saline nitrates that were used in the research were NaNO2, NaNO3 and KNO3 (SigmaeAldrich 98%). The impurity content of tested Hitec salt is shown in Table 1 and was determined using the Volhard procedure (Cl), Gravimetry (% moisture, SO4 2 ), UVeVis spectroscopy (NO2 ) and ICP-OES (Mg). The importance of the studied parameters is based on the following factors: Cl: Appears as a corrosion enhancing agent. A slight increase from the initial value is due to the decomposition reaction of the perchlorate in the mixtures. SO4 2 : An important parameter to follow because sulphates are able to form insoluble compounds that can clog pumps and pipes, which affects the circulation of the salt in the solar power plant. Mg is an important impurity in the salt that forms oxides and is able to form magnesiumferrite, a very stable compound formed by the interaction between Mg and hematite (Fe2O3). The corrosion tests were performed on commercial steels with the compositions shown in Table 2 at storage temperature used in parabolic trough solar power plants (390 C).
Table 2 Chemical composition of carbon steels studied. Steel
(8)
Perchlorates are present in commercial molten nitrates as impurities; these impurities need to be controlled during the melting
(9)
A516 T11 T-22
Weight % Si
Mn
Cr
P
Mo
C
S
0.1 0.79 0.3
0.93 0.44 0.4
1.2 2.25
0.035 0.008 0.3
0.5 1
0.27 0.1 0.12
0.035 0.002 0.3
ndez et al. / Renewable Energy 80 (2015) 177e183 A.G. Ferna
179
The specimens used in the corrosion test have dimensions of 20 10 2 mm. They were polished with #600 SiC abrasive papers and washed with distilled water and acetone. The isothermal immersion tests were analysed via gravimetric analysis at 24, 48, 150, 350, 500, 750, 1000, 1250, 1500, 1750 and 2000 h; two specimens of each alloy were removed from the nitrate mixtures for examination and analysis. Once the samples were removed from the oven they were cooled slowly in warm distilled water in order to eliminate the salt in which they had been immersed. They were then dried and weighed, average taken from five values of their weights. The formula (equation I) used to calculate the mass gain over time is:
mi mf Dm ¼ S0 S0 Fig. 1. Gravimetric behaviour of steels in 40% NaNO2 þ 7% NaNO3 þ 53% KNO3 at 390 C for 2000 h of test.
[Ec. I]
where mi is the initial mass of the specimen, mf is the mass of the sample at time t and S0 is the initial area of the specimen. 3. Results and discussion
Fig. 2. Surficial image of A516 after 2000 h of test.
The steels under analysis at temperatures of 390 C are carbon steels with low chrome content (see Table 2 for their composition) which were immersed in molten salt for 2000 h. The increase in gravimetric mass experienced by these steels is shown on Fig. 1. All steels analysed on the saline medium showed similar behaviour against corrosion, with increases in mass of around 0.37 mg/cm2. Individual studies of each steel where made, in order to expound the corrosion processes affecting carbon and low Cr steels in the experiment and the associated increase in mass. The surface of the A516 steel at the end of the experiment (2000 h) shows a thin homogeneous layer of corrosion with oxygen and iron content (Fig. 2). Together with the formation of small needles (with a greater content in calcium, analysis Q5-2) and other less defined particles, in which magnesium and calcium (Q5-1) can be detected. The analysis of a cross section of the sample at the end of the experiment is shown on Fig. 3. The steel shows a good behaviour against 2 corrosion, with a layer.66 microns deep. EDX (Energy Dispersive X-ray) analysis of this layer shows its main elements to be iron, oxygen and magnesium. X-ray analysis (XRD, X-ray diffraction) showed the presence of the following corrosion by-products (Fig. 4). The corrosion layer was composed of hematite (Fe2O3) and magnetite (Fe3O4). MgO and CaCO3 were also detected, originating from the impurities of the salt; their formation in the surface of the steel explains the levels found in the analysis shown in Fig. 2.
Fig. 3. Cross section of A516 steel after 2000 h of test.
180
ndez et al. / Renewable Energy 80 (2015) 177e183 A.G. Ferna
The corroded layer which formed at the end of the experiment (2000 h) was 2.66 microns deep, with Fe2O3 and MgO present. Fe2O3 and MgO had not yet reacted to form MgFe2O4, a compound which has been detected in previous studies with layers over 5 micron thickness [12]. The formation of Fe2O3 in preference to other compounds in this reactive medium has been studied by Picard et al. [13], in which formation of this oxide is related to the values of PO2 in the salt. The thermodynamic calculations based on potential PO2 relationships (vertical axis shows in Fig. 5) suggested that the basicity of the melt increases with temperature increase, which favours incorporation of impurities in the scale. These values affect the passivation quality of the oxidised layers as well as its associated kinetic abilities. The authors showed the balance between the oxides and nitrates in a diagram (Fig. 5). These diagrams have been created based on the thermodynamic constants of the reactions:
Fig. 4. XRD of A516 steel after 2000 h of test.
Fig. 5. Equilibrium diagram of nitrides and iron oxides.
NO3 4NO2 þ 1=2O2
(10)
3NO2 42NO þ NO3 þ O2
(11)
The study of the surface of the steel T11 (Fig. 6) shows an oxidised layer composed of magnesium, iron and oxygen (analysis R53) which two different types of particles are adhered: small particles in the shape of needles with calcium content (R5-2) and larger spherical particles, with higher contents of oxygen and magnesium. The analysis of a cross section of the sample (Fig. 7) shows a metallographic layer of 7.05 microns, almost wholly composed of iron and oxygen. The x-ray analysis (Fig. 8) confirms the presence of corrosion byproducts first glimpsed in the EDX analysis as part of the metallographic study using scanning electron microscopy. Hematite and Magnetite were identified as the two components of the oxidised layer, together with calcium carbonate on the surface of the material. MgO did not seem to be present, owing to the particles forming only occasionally on the surface of the sample. Other stable compounds in this medium that appear at elevated temperatures are carbonate species. C.M. Kramer et al. [14] studied the formation of carbonates via the interactions of nitrates with atmospheric CO2 in this medium, which resulted in the following reactions: 2NaNO3 þ CO2 / Na2CO3 þ N2 þ 5/2O2
Fig. 6. Surficial image of T11 after 2000 h of test.
(12)
ndez et al. / Renewable Energy 80 (2015) 177e183 A.G. Ferna
181
Fig. 7. Cross section of a) T11 steel and b) EDX study during 2000 h of test.
Na2O þ CO2 / Na2CO3
(13)
Similarly, CO3 2 ions formed in the salt can interact with the calcium or magnesium present in the NaNO3 and KNO3 salts as well as the other ionic components in these salts. The surface of the T22 steel after the end of the 2000 h experiments is shown in Fig. 9. Both the image obtained and the EDX analysis show a greater number of acicular structures with a calcium content (S5-3), together with a layering of the same structures with a whitish appearance, corresponding with the S5-5 analysis. These growth structures appear on top of an oxidised layer and are mainly composed of iron, oxygen and magnesium. The cross section of the sample after 2000 h immersed in the ternary mixture of nitrite is shown in Fig. 10. The oxidised layer detected through the EDX analysis of the steel T22 is composed of magnesium, iron and oxygen and a thickness of 6.05 microns. This increase in the oxidised layer is due to its protective character as the chrome present in the base material diffuses from the substrate to increase weight gains, as registered in the gravimetric curve in Fig. 1. However this protective character gives the material better performance against corrosion and isothermal experiment could be
Fig. 8. XRD of T11 steel after 2000 h of test.
extended during the time, maintaining the oxidising parabolic kinetic presently shown. The x-ray study, detected the presence of the corrosion products shown in Fig. 11. In this instance, it appears that Magnesium ferrite (MgFe2O4) was formed, and the analysis shows it is the main component of the oxides generated on the surface of the steel. In this regards, MgO and K2CrO4 were also present, the latter giving a certain level of protection to the sample. The salt used in the corrosion experiments was analysed with x rays to detect any components that could had been detached from the base material as well as any insoluble components of the salts. Since the behaviour of the steels has been shown to be very similar, the by-products of the corrosion were exemplified in the diffractogram (Fig. 12a) and the diffractograms obtained of the three steels have been represented (Fig. 12b) as the only differences shown were with regards to the values obtained of the referenced by products. Hematite (Fe2O3) and NaFeO2 were shown to be the main products of the corrosion, appearing at levels similar to the ones
Fig. 9. Surficial image of T22 after 2000 h of test.
182
ndez et al. / Renewable Energy 80 (2015) 177e183 A.G. Ferna
Fig. 10. Cross section of T22 steel after 2000 h of test.
obtained for this compounds in the salts used to submerge the steels. The excellent performance show by the ternary mixture of nitrite was evaluated by S. Al Omer et al. [15], showing the importance of Na2O2 peroxide to that effect. In this regard, the mixture starts losing weight as the reaction proceeds: Na2O2 4 Na2O þ ½ O2
Fig. 11. XRD of T22 steel after 2000 h of test.
(14)
This would justify the presence of Na2O in the ternary nitrite salt analysed at the end of the corrosion experiment. The results of the chemical analysis of the major parameters of the salt properties are shown in Table 3. Chloride levels in the salt are the lowest detected in the new molten salt formulation with low melting point present in the literature [16e20]. Equally, the chemical analysis on the remaining parameters show similar values to those obtained initially, and there are no traces of chrome or iron content in the salt at the end of the experiment, as was to be expected, in view of the salt's optimum performance against corrosion.
Fig. 12. XRD of HITEC salt after corrosion test a) Salt in contact with A516 b) difractogram comparison.
ndez et al. / Renewable Energy 80 (2015) 177e183 A.G. Ferna Table 3 Chemical composition of HITEC salt after corrosion test compared with the initial values. Parameter
Initial values
Salt A516
Salt T11
Salt T22
Cl (%) SO4 2 (mg/g) Ca (mg/g) Cr (mg/g) Fe (mg/g) Mg (mg/g) Moisture (%) NO2 (mg/g)
0,0795 431,57 35,76 <1 <1 70,14 0,14737 268001,4
0,094 1029 65 e <1 96 0,15 244091
0,091 915 48 <1 <1 59 0,14 260809
0,09 903 53 <1 <1 64 0,11 261303
Moisture level detected in the salt remains, was also close to the initial content, and in particular the level of nitrites shows a slight reduction from the initial level analysed. This last fact is of importance and worth noting since it will imply that the thermal properties of the HITEC salt do not decrease after the change in the final composition of the mixture. The thermal study of this mixture after the corrosion experiment will be duly analysed in future research. 4. Conclusions The corrosive properties of the salt were evaluated at 390 C, simulating the normal working conditions of a solar plant using parabolic trough collectors in contact with carbon steel and low chrome steels. The three steels showed excellent behaviour against corrosion, with mass gains of around 0.35 mg/cm2. A516 steel showed a metallography layer 2.66 microns deep, composed mainly of magnetite and hematite. T11 steel, as deduced by the gravimetric gain, shows a thickness layer of 7.05 microns, due to the occasional presence on its surface of adhered particles of MgO. T22 steel shows a corrosion layer of 6.05 microns, with a protective layer formed in the inner zone to the material, identified through XRD as the K2CrO4 protective spinel. This compound will give to the material a protective character over 2000 h of test, so the corrosion properties of the HITEC salt will not be an impediment to its use in the current concentrated solar power technology. Acknowledgement The authors would like to acknowledge the financial support provided by CONICYT/FONDAP 15110019 “Solar Energy Research
183
Center” SERC-Chile, Fondecyt Postdoctoral grant n 3140014, FIC-R 30137092 funded by Atacama Government and the Education Ministry of Chile Grant PMI ANT 1201.
References [1] Ho MX, Pan C. Optimal concentration of alumina nanoparticles in molten hitec salt to maximize its specific heat capacity. Int J Heat Mass Transf 2014;70(0): 174e84. [2] Mochinaga J. Molar volumes of the molten NaNO3-NaNO2-KNO3 system. J Chem Eng Data 1985;30:274e6. [3] Janz GJ. Melting and premelting properties of the KNO3-NaNO2-NaNO3 eutectic system. J Chem Eng 1983;28:201e2. [4] Olivares RI. The thermal stability of molten nitrite/nitrates salt for solar thermal energy storage in different atmospheres. Sol Energy 2012;86(9): 2576e83. [5] Baraka A, El Hosary A. Corrosion on mild steel in molten sodium nitratepotassium nitrate eutectic. Br Corros J 1976;11(1):44e6. [6] De Jong JM, Broers GHJ. A reversible oxygen electrode in an equimolar KNO3dNaNO3 melt saturated with sodium peroxide II. A voltammetric study. Electrochimica Acta 1976;21(11):893e900. [7] Singh IB. The effect of NaCl addition on the corrosion of mild steel in NaNO3 melt. Corros Sci 1993;34(10):1733e42. [8] Palmisano F. Catalytic currents in fused salts: discharge mechanism of nitrite in molten alkali nitrates(II). J Electroanal Chem Interfacial Electrochem 1978;89(2):311e20. [9] Desimoni E, Palmisano F, Zambonin PG. Catalytic currents in fused salts: discharge mechanism of nitrite in molten alkali nitrates. J Electroanal Chem Interfacial Electrochem 1977;84(2):315e22. [10] Fernandez AG, Galleguillos H, Perez FJ. Thermal influence in corrosion properties of Chilean solar nitrates. Sol Energy 2014;109(0):125e34. [11] Kleppa OJ. Thermochemistry of anion mixtures in simple fused salt systems. II 2 solutions of some salts of MO 4 and MO4 anions in the corresponding alkali nitrates. J Phys Chem 1963;67(12):2750e3. [12] Fernandez AG, Galleguillos H, Perez FJ. Corrosion ability of a novel heat transfer fluid for energy storage in CSP plants. Oxid Metals 2014;82:331e45. http://dx.doi.org/10.1007/s11085-014-9494-3. [13] Picard G, L H. Thermodynamic study of corrosion of iron in NaNO3-NaNO2 mixtures. J Electrochem Soc 1987;134(1). [14] Kramer CM, Munir ZA, Volponi JV. Screening tests of sodium nitrate and potassium nitrate decomposition. Sol Energy 1982;29(5):437e9. [15] Al Omer S. Molten sodium nitrite-potassium nitrite eutectic: the reactions of sodium peroxide. J Inorg Nucl Chem 1978;40:975e8. [16] Fernandez AG, Lasanta MI, Perez FJ. Molten salt corrosion of stainless steels and low-Cr steel in CSP plants. Oxid Metals 2012;78(5e6):329e48. [17] Bradshaw RW, Meeker DE. High-temperature stability of ternary nitrate molten salts for solar thermal energy systems. Sol Energy Mater 1990;21(1): 51e60. [18] Goods SH, Bradshaw RW. Corrosion of stainless steels and carbon steel by molten mixtures of commercial nitrate salts. J Mater Eng Perform 2003;13. [19] Robert W, Bradshaw WMC. Effect of chloride content of molten nitrate salt on corrosion of A516 carbon steel. Sandia report. 2010. [20] Bradshaw R., Low melting point inorganic nitrate salt heat transfer fluid, United States Patent, 2009, US7588694B1.