- Email: [email protected]
LiNO3–NaNO3–KNO3 salt for thermal energy storage: Thermal stability evaluation in different atmospheres
Thermochimica Acta 560 (2013) 34–42
Contents lists available at SciVerse ScienceDirect
Thermochimica Acta journal homepage: www.elsevier.com/locate/tca
LiNO3 –NaNO3 –KNO3 salt for thermal energy storage: Thermal stability evaluation in different atmospheres Rene I. Olivares ∗ , William Edwards CSIRO Energy Centre, Newcastle, Australia
a r t i c l e
i n f o
Article history: Received 20 November 2012 Received in revised form 20 February 2013 Accepted 26 February 2013 Available online 16 March 2013 Keywords: Thermal energy storage (TES) Heat transfer fluid (HTF) Concentrating solar power (CSP) Eutectic LiNO3 –NaNO3 –KNO3 Thermal stability
a b s t r a c t The thermal stability of the eutectic LiNO3 –NaNO3 –KNO3 salt was investigated by simultaneous differential scanning calorimetry, thermogravimetry and mass spectrometry (DSC/TG–MS). The work was carried out between room temperature and 1000 ◦ C in blanket gas atmospheres of argon, nitrogen, oxygen and air. The stability of the salt, as measured by the gases evolving from the melt, was influenced by the atmosphere. Evolution of the main gaseous species NO was detected at 325 ◦ C in an atmosphere of argon, at 425 ◦ C in an atmosphere of nitrogen, at 475 ◦ C in an atmosphere of air and at 540 ◦ C in an atmosphere of oxygen. Prior to melting, the eutectic underwent endothermic (␣/) solid–solid type transformation at 87 ◦ C. The melting point was 121 ◦ C, and the solidification point 98 ◦ C. Under-cooling of the salt coincided with the onset of the (␣/) solid–solid transformation upon heating. At a temperature of 500 ◦ C in air, TG analysis showed that the long-term stability of the salt was limited and this was confirmed by DSC. Uncertainty analysis indicated that the measurement of temperature is accurate to ±2.7 ◦ C. © 2013 Elsevier B.V. All rights reserved.
1. Introduction Solar thermal plants which concentrate the Sun’s energy to produce steam and electricity often use molten salt mixtures as a heat transfer fluid (HTF) and/or as thermal energy storage (TES) medium [1–3]. It is desirable that such a salt should have a low melting point in addition to being stable at the highest possible temperature. Nitrates based salts are generally used at temperatures below 500 ◦ C. Operating such salts at temperatures higher than 500 ◦ C may be possible by rigorous control of the atmosphere. In previous contribution [4], using the same methodology described in this work, it was shown that atmosphere had an important effect on the thermal stability of the nitrite/nitrates based ternary salt (NaNO2 –NaNO3 –KNO3 ) allowing for the maximum operating range of the salt to be extended at high pO2 . Oxidising atmosphere however changed the chemistry of the salt by converting nitrite (NO2 − ) to nitrate (NO3 − ) increasing the melting point. In this work, we revise the low melting point nitrates only, eutectic LiNO3 –NaNO3 –KNO3 , under the influence of different gas atmospheres. A low melting point salt that is stable at temperatures higher than 700 ◦ C is needed to increase the heat-to-electricity conversion efficiency for Rankine cycle systems, for use in Brayton cycle
∗ Corresponding author at: 10 Murray Dwyer Circuit, Steel River Estate, Mayfield West, NSW 2304, Australia. Tel.: +61 2 4960 6253. E-mail address: [email protected] (R.I. Olivares). 0040-6031/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.tca.2013.02.029
systems and to supply high grade heat to industrial processes. The high temperature thermal stability of the salt is then an important aspect that requires consideration given that it will determine the long term viability of its use in a high temperature storage system. Numerous reports addressing thermal stability for nitrate based systems are available [5–8], albeit that some ambiguous conclusions have been made as well. It is therefore essential that the thermo-chemical behaviour, thermal stability, and mode of decomposition of the salt are studied in some detail. For this, the low melting point ternary eutectic system LiNO3 –NaNO3 –KNO3 was characterised in the present work between room temperature and 1000 ◦ C under various gas atmospheres. 1.1. Literature The principal mode of thermal decomposition in nitrate salts has been agreed amongst researchers [5–10]; to correspond to the main reaction: NO3 − ↔ NO2 − + 1/2 O2 . Further decomposition also takes place with the evolution of the oxides of nitrogen, particularly at higher temperatures [11]. In some studies [5,12,13] the decomposition temperature has been reported as the temperature at which oxygen, nitrogen or nitrous oxide is detected in the gas phase. In the work of Gordon and Campbell [12] the single salts alkali metal nitrates were observed to undergo a thermal reaction at temperatures 100–300 ◦ C above their melting points as indicated by bubbling of the molten salt. The evolution of nitrous fumes observed by Gordon and Campbell [12], occurred at temperatures ranging from about 200–350 ◦ C above
R.I. Olivares, W. Edwards / Thermochimica Acta 560 (2013) 34–42 Table 1 Eutectic LiNO3 –NaNO3 –KNO3 formulation. Compound
wt%
mol%
LiNO3 NaNO3 KNO3
30 18 52
37 14 49
Note: Compositions were determined using phase diagrams calculated with FactSage software (Fig. 1).
the initial bubbling reaction. This was claimed to be as a result of the decomposition of the nitrate and/or nitrite to oxide. The later stage of the decomposition of these nitrates was still occurring at temperatures as high as 900 ◦ C which was the upper limit of the apparatus employed [12]. Kust and Burke [14] conducted experiments with a pure equimolar sodium potassium nitrate melt in an oxygen rich atmosphere (pO2 at 0.85 atm) and measured high oxide ion concentration (O2− ) in the molten salt which corresponded to the dissociation constant for nitrite ion (NO2 − ); based on the equilibrium NO3 − → NO2 − + O2− , (nitrite ion concentration reached equilibrium within 7–16 h at temperatures between 295 ◦ C and 340 ◦ C in Kust and Burke experiments [14]) demonstrating that nitrate ion (NO3 − ) could decompose to a measurable extent at temperatures as low as 295 ◦ C. The thermal stability of binary mixes of alkali metal nitrates as investigated by Abe et al. [15], by means of differential thermal analysis (DTA) and evolved gas analysis (EGA), concluded that the nitrate melts are thermally unstable. The evolved gas analysis measurements [15] indicated that the decomposition of nitrates took place independently and followed the ranking order of thermodynamic stability as determined by the reactions sequence MNO3 → MNO2 + 1/2O2 followed by 2MNO2 → M2 O + NO + NO2 . On this basis the decomposition of a mixture of LiNO3 –NaNO3 –KNO3 would be determined by the least thermodynamically stable species, in this case LiNO3 followed by NaNO3 and then KNO3 , a possibility if the solution is close to ideal. Most recently the thermal stability of the eutectic composition in LiNO3 –NaNO3 –KNO3 ternary salt was reported by Wang et al. [16], after studies where temperature was cycled between 75 ◦ C and 500 ◦ C in argon gas atmosphere, using a TG/DTA method. Wang et al. [16] observed that once the salt had lost its water content after the first melting cycle it reached a stable condition. In long term stability tests at 500 ◦ C though, there was a weight change of ∼8.6 wt% and [16] concluded that the salt was unstable beyond this point. Given that weight loss alone as a measure of stability may be compromised due to some salt evaporation, further analysis is required before the practical limit of salt stability can be ascertained. In the present work the thermo-chemical behaviour and thermal stability of the LiNO3 –NaNO3 –KNO3 eutectic was studied from at least three different points of view: (i) the temperature at which the rapid evolution of gases, NO and/or NO2 and O2 , are detected by means of mass spectrometry (MS), (ii) the temperature of the melt at which an irreversible endothermic peak of decomposition is resolved by differential scanning calorimetry (DSC), and (iii) the temperature at which rapid weight loss is observed in a thermogravimetric curve (TG). A combination of these three criteria is used for elucidation of the limiting temperature for operation of the salt as would apply in a TES installation. The chemistry and equilibrium reactions by which molten nitrites/nitrates may interact with the atmosphere can be found in Refs. [4,17,18,26]. 2. Experimental The eutectic salt composition was prepared by mixing the individual components in the weight proportions given in Table 1.
35
Analytical reagent grade chemicals LiNO3 –NaNO3 –KNO3 supplied by Ajax (UNIVAR) were dried overnight in vacuum prior to weighing and mixing. 2.1. Experimental procedure A 20 g mixture of the pre-dried pure chemicals was melted inside an alumina crucible in a muffle furnace and allowed to equilibrate in air at 250 ◦ C for 2 h before casting into a stainless steel mould to quench and solidify. The solidified salt was kept under vacuum at 65 ◦ C at all times from which a small sub-sample was taken and used for simultaneous DSC/TG–MS experimentation as needed. About 15–25 mg were ground and loaded into a 100 L Al2 O3 crucible in a SETARAM DSC/TG SETSYS Evolution analyser, coupled with a Pfeiffer QUADSTAR-422 mass spectrometer. Due to the presence of LiNO3 , the sample was highly hygroscopic and it was inevitable that upon handling the sub-sample, moisture was absorbed. The data acquisition was done by programming the MS to scan for masses from 1 to 60 (amu) to cover the range of possible gaseous species reporting to the gas phase. Identification of gaseous species was done by matching the relative Ion Current Intensities measured by the MS, with the corresponding spectra available in the QUADSTAR-422 MS library database. The experiments were carried out from room temperature up to 1000 ◦ C at a heating-rate of 10 ◦ C/min in atmospheres of argon, nitrogen, air and oxygen respectively. Once the atmosphere around the sample was set, the gas was allowed to flow at a rate of 30 ml/min. Five melting–solidification cycles were initially carried out between 50 ◦ C and 400 ◦ C before the full decomposition up to 1000 ◦ C was completed. At all times during heating and cooling the carrier gas was maintained at 30 ml/min and the heating rate was at 10 ◦ C/min. A weight loss in the first heating cycle of between 1 and 2 wt% corresponded to the evolution of water absorbed during sample handling; this was detected by the MS, and was not considered in the analysis. The subsequent four remaining cycles did not show any weight loss or water evolution and were used for determination of the melting and solidification points and to check reproducibility of the measurement. At the end of each decomposition run, contamination of the DSC platinum sensor was evident by a dark stain around the sensor that had to be thoroughly cleaned between runs by first soaking in hot water and then heating to 1200 ◦ C in oxygen to remove any residue. 2.2. Experimental error and uncertainty Temperature and heat flow calibration of the DSC was carried out as per the standard method for this type of analysis [19,20] using certified reference materials (CRM) listed in Table 2 to cover the temperature range 100–1050 ◦ C. All calibration curve parameters were estimated by a least-squares optimisation routine (build in Calisto Software–SETARAM). The uncertainty on the temperature and heat flow measurements were calculated from the standard error of the estimate, s, according to: s2 =
1 N − (n + 1)
(Ycorr − Ystd )2
where s is the standard error of the estimate; N, number of data points; n, number of independent variables; Ycorr , corrected temperature of melting, Tcorr ; or, corrected energy of melting, Hcorr ; and Ystd , melting temperature of standard, Tstd ; or, energy of melting of standard, Hstd . From these and using the calculation procedure described, the uncertainty for the temperature is ±2.7 ◦ C and the uncertainty for the energy is ±2.3 J/g. The uncertainty of the standard (see Table 2)
36
R.I. Olivares, W. Edwards / Thermochimica Acta 560 (2013) 34–42
Table 2 Standard certified reference materials (CRM)–DSC temperature and energy calibration. CRM melting, Tstd (◦ C) [21]
CRM
Heating rate (◦ C/min)
In
10
156.59 ± 0.01
Pb
10
327.47 ± 0.02
Al
10
660.33 ± 0.05
Au Pb Pb
10 5 20
1064.18 327.47 ± 0.02
Measured melting, Tmeas (◦ C)
Corrected melting, Tcorr (◦ C)
156.46 156.39 326.52 326.74 660.69 659.99 1065.11 325.56 326.99
154.08 154.01 325.06 325.28 661.05 660.35 1067.67 323.62 326.49
and measured temperatures (noise level ±5 mK) are negligible compared to the standard error of the estimate, s. 3. Results and discussion The present work concerns the low melting point eutectic LiNO3 –NaNO3 –KNO3 prepared in the laboratory from high purity chemicals. This formulation was studied over the temperature range up to 1000 ◦ C in atmospheres of argon, nitrogen, oxygen and air using simultaneous DSC/TG and MS analysis (Fig. 1). 3.1. Melting and solidification point determination The melting and solidification temperatures respectively were determined as per conventional DSC analysis at the point where the baseline intercepts the onset of the endothermic melting event or the exothermic solidification event. An example of a heating–cooling cycle is shown in Fig. 2. Repeatability of the measurements in four heating–cooling cycles between 50 ◦ C and 400 ◦ C is given in Table 3; the reproducibility of the results was excellent. Prior to melting, the eutectic underwent an endothermic (␣/) solid–solid type transformation starting at 87 ◦ C. The melting point was 121 ◦ C, and the solidification point 98 ◦ C. Significant
CRM energy of melting, Hstd (J/g) [22]
Corrected energy of melting, Hcorr (J/g)
28.51 ± 0.19
28.49
23.00 ± 0.06
23.09
401.30 ± 1.6 64.50
404.64 63.51
under-cooling of the salt was evident, possibly due to delayed nucleation because of the relatively fast cooling rate of 10 ◦ C/min. Considering that at the eutectic composition the liquidus and solidus temperatures should be the same, the under-cooling observed here is most likely a function of the cooling rate.
3.2. Salt decomposition and evolved gas analysis The results for the simultaneous DSC/TG–MS experiments with the salt in atmospheres of argon, nitrogen, oxygen and air are shown in Figs. 3–6 respectively. The figures show superimposed in the same graph: (i) the ion current intensity (A) for the detection by the MS of the evolved gaseous species, (ii) the TG (mg) curve to monitor the weight change, and (iii) the DSC (mW) trace to identify the temperature at which heat events (endothermic/exothermic) occur. In all cases the figures illustrate a wide endothermic heat flow through which coincides with a sharp drop in weight once a critical temperature is reached, reflecting a rapid rate of decomposition of the salt. Independent measurements indicated that the reproducibility of the results was excellent. In an inert atmosphere of argon, Fig. 3, the gaseous species detected by the MS were NO, O2 , O, NO2 , N, N2 and also N2 O. The evolution of nitrous oxide (NO) is first noticed at 325 ◦ C and the evolution of nitrogen species (N, N2 ) at the much higher
Fig. 1. Calculated LiNO3 –NaNO3 –KNO3 eutectic composition–FactSage6.3.
R.I. Olivares, W. Edwards / Thermochimica Acta 560 (2013) 34–42
37
Fig. 2. Melting point and solidification point determination of eutectic LiNO3 –NaNO3 –KNO3 .
temperatures of 584 ◦ C and 693 ◦ C respectively. The O2 and NO2 species evolved at temperatures between these two. The evolution of N2 O was apparent at 895 ◦ C. Irreversible decomposition of the eutectic LiNO3 –NaNO3 –KNO3 in this case can be considered to occur at a temperature between the point at which rapid weight loss begins on approaching 602 ◦ C and the bulk of decomposition is identified by the broad endothermic heat event at ∼670 ◦ C in the DSC trace. In an inert atmosphere of nitrogen, Fig. 4 shows that the evolution of gaseous species NO started at 425 ◦ C followed by O2 at 474 ◦ C with NO2 and mono-atomic oxygen (O) at about 620 ◦ C; evolution of N2 O was apparent at 825 ◦ C. Irreversible decomposition in this case can be considered to occur at a temperature between the point at which rapid weight loss begins approaching 584 ◦ C and the bulk of decomposition identified by the broad endothermic heat event at ∼655 ◦ C in the DSC trace. Because the blanket atmosphere was nitrogen, the background ion current intensity in the MS was too large and the species N and N2 were not detected in this experiment. For the experiment in an atmosphere of oxygen, Fig. 5, the gaseous species detected evolving from the melt were NO at 540 ◦ C
and NO2 at between 422 and 540 ◦ C, with N2 O at the higher temperature of 855 ◦ C; whereas for the experiment in air, Fig. 6, only NO and NO2 gases were detected evolving at 475 ◦ C and 550 ◦ C respectively. The MS background ion current intensities for oxygen and nitrogen species under these atmospheres (oxygen and air) are too large and prevent the detection by the MS of the possible evolution of N, N2 O, O2 from the melt. In oxygen (Fig. 5) the critical temperature after which rapid weight loss began was 577 ◦ C, and the bulk of decomposition as per the large endothermic event in the DSC trace occurred at ∼608 ◦ C. In air (Fig. 6) the critical temperature after which rapid weight loss occurred was 602 ◦ C, and the bulk of decomposition as per the large endothermic event in the DSC trace at ∼625 ◦ C. The data taken in Figs. 3–6 are further summarised in Table 4; the criteria used to select the maximum temperature in the TG curve was the point at which a 1 wt% weight loss is reached. The TG curves in Figs. 3–6 suggest that the salt could be taken to temperatures higher than 550 ◦ C and even to 600 ◦ C in air. TGA alone has been used by others [23,24] to evaluate thermal stability of molten salts for TES. Raade and Padowitz [24] defined the high temperature limit of the salt as that where the salt is observed to rapidly begin to lose weight beyond a maximum acceptable
Table 3 Repeatability of melting and solidification point measurement. Melting
␣/ (Solid–solid) transformation ◦
Cycle 2 Cycle 3 Cycle 4 Cycle 5 Average (4 cycles) Solidification
Cycle 2 Cycle 3 Cycle 4 Cycle 5 Average (4 cycles)
Melting process ◦
Onset, T ( C)
Max, T ( C)
H (J/g)
Onset, T (◦ C)
Max, T (◦ C)
H (J/g)
86 86 86 88 87
91 94 93 95 93
12.5 6.9 6.1 4.1 7.4
121 122 120 122 121
128 128 127 128 128
141.5 140.5 147.2 135.5 141.2
Solidification process Onset, T (◦ C)
Max, T (◦ C)
H (J/g)
98 98 98 98 98
93 93 93 94 93
−92.9 −97.4 −97.7 −95.5 −95.9
38
R.I. Olivares, W. Edwards / Thermochimica Acta 560 (2013) 34–42
Fig. 3. DSC/TG–MS analysis for thermal decomposition of LiNO3 –NaNO3 –KNO3 salt in argon atmosphere.
Fig. 4. DSC/TG–MS analysis for thermal decomposition of LiNO3 –NaNO3 –KNO3 salt in nitrogen atmosphere.
R.I. Olivares, W. Edwards / Thermochimica Acta 560 (2013) 34–42
39
Fig. 5. DSC/TG–MS analysis for thermal decomposition of LiNO3 –NaNO3 –KNO3 salt in oxygen atmosphere.
weight loss of 3 wt%. Long term stability experiments for the eutectic LiNO3 –NaNO3 –KNO3 salt prepared for the present work were carried out in an air atmosphere at fixed temperatures of 400 ◦ C and 500 ◦ C over 4 h. These indicated that at 400 ◦ C a small weight loss over 4 h was consistent with some loss due to salt evaporation from the open crucible (supported by the fact that at the end of each experiment the crucible was stuck to the base of the sensor due to condensed vapours). At 500 ◦ C very rapid weight loss took place as shown in Fig. 7. It is noted that the carrier gas flow rate was 30 ml/min and this may have also contributed to some losses by salt carry over. The condition of the salt was subsequently checked by DSC and this is shown in Figs. 8 and 9. DSC analysis in Fig. 8, for the sample after 4 h at 400 ◦ C, showed that no change in thermo chemical behaviour occurred and the curve reproduced well the DSC analysis of the fresh sample (see Fig. 2). DSC analysis in Fig. 9
showed that for the sample after 4 h at 500 ◦ C, the salt presented a significantly different thermo chemical behaviour. It is recognised [25] that in general the thermal decomposition of nitrates can be divided at least in three zones: (i) formation of little nitrite soon after melting, (ii) nitrate–nitrite equilibrium in the range 400 ◦ C to 550 ◦ C (MNO3 → MNO2 + 1/2O2 ) dependent on pO2 , and (iii) decomposition with the release of nitrogen oxides at greater than 550 ◦ C (2MNO2 → M2 O + NO + NO2 ), where M can represent either Li, Na or K. A survey of the literature provided by Stern [26] list the gaseous products from decomposition of nitrates and nitrites as follows; O2 , N2 , NO, NO2 , N2 O3 , N2 O4 , N2 O5 . Which of these are produced, depend on the salt, the atmosphere, the temperature, and the experimental conditions. From the thermodynamic data compiled by Stern [26], it can be seen that N2 O5 is unstable with respect to dissociation into
Table 4 Summary for the effect of atmosphere on thermal decomposition of LiNO3 –NaNO3 –KNO3 salt. Atmosphere
Argon Nitrogen Oxygen Air
Melting, T (◦ C)
121 122 120 122
DSC–event endothermic, T (◦ C)
88 87 86 88
Irreversible DSC event, T (◦ C) decomposition
671 655 608 625
TG rapid wt loss after, T (◦ C)
602 584 577 602
All temperatures reported are rounded figures and have an uncertainty of ±2.7 ◦ C; and nd = not detected.
Evolved gaseous species detected by MS–temperature of detection (◦ C)
NO
O2
N
O
NO2
N2
N2 O
325 425 540 475
418 474 nd nd
584 nd nd nd
588 619 nd nd
659 622 422–540 550
693 nd nd nd
895 825 855 nd
40
R.I. Olivares, W. Edwards / Thermochimica Acta 560 (2013) 34–42
Fig. 6. DSC/TG–MS analysis for thermal decomposition of LiNO3 –NaNO3 –KNO3 salt in air atmosphere.
N2 O4 or N2 O3 and the equilibrium shifts markedly to NO2 with increasing temperature, such that above 127 ◦ C the concentration of N2 O4 is negligible and N2 O3 is unstable with respect to both NO2 and NO. The equilibrium NO + 1/2O2 = NO2 is very temperature dependent, where up to 227 ◦ C NO2 predominates and above this temperature both species are significant. N2 and O2 are very stable with respect to NO and NO2 ; if N2 is formed during salt decomposition it will not react further, but O2 may react with NO and nitrites. If all equilibriums during salt decomposition were established quickly, N2 and O2 would be the only gaseous species. The reaction NO2 = 1/2N2 + O2 is very slow below 727 ◦ C and if NO2 is the primary product of salt decomposition, it will not decompose appreciably so only the equilibrium NO + 1/2O2 = NO2 needs to be considered [26]. Likewise the decomposition of NO to its elements N and O can be neglected. The combined DSC/TG–MS analysis used in the experiments described here showed that for the ternary eutectic LiNO3 –NaNO3 –KNO3 , the blanket gas atmosphere had a significant role in determining the mode of decomposition, predominantly due to its effect on pO2 and the reversible equilibrium nitrite/nitrate. The stability of the salt, as measured by the gases evolving from the melt, indicated that the main gaseous species NO was detected at 325 ◦ C in an atmosphere of argon, at 425 ◦ C in an atmosphere of nitrogen, at 475 ◦ C in an atmosphere of air and at 540 ◦ C in an atmosphere of oxygen, albeit without measurable weight loss. By two assessment criteria, TG and DSC, the salt could be heated up to 600 ◦ C in air with negligible decomposition. For long term stability, given the changes seen in the eutectic after 4 h at 500 ◦ C (Figs. 7 and 9), it is unlikely that this system could
be used at a temperature of 600 ◦ C for an extended period given that the relative concentration of constituents would have changed having an impact on the constitution of the eutectic. Wang et al. [16], from characterisation by XRD–SEM/EDS analysis of the eutectic LiNO3 –NaNO3 –KNO3 before and after long term thermal stability experiments in an argon atmosphere, showed that the least stable lithium nitrate compound was decomposing to its oxides at the high temperature of 500 ◦ C. In Wang et al.’s [16] experiments, the limit of stability was found to be 435 ◦ C in an argon gas atmosphere and this was based on thermo-gravimetric (TG) analysis alone. The incorporation of evolved gas analysis (EGA) as used in this work, allows for a more accurate evaluation.
Fig. 7. Weight variation of eutectic LiNO3 –NaNO3 –KNO3 eutectic at 400 ◦ C and 500 ◦ C.
R.I. Olivares, W. Edwards / Thermochimica Acta 560 (2013) 34–42
41
Fig. 8. DSC analysis of LiNO3 –NaNO3 –KNO3 eutectic after sample was held in air at 400 ◦ C for 4 h.
Fig. 9. DSC analysis of LiNO3 –NaNO3 –KNO3 eutectic after sample was held in air at 500 ◦ C for 4 h.
4. Conclusions
Acknowledgements
The thermo-chemical behaviour and thermal stability of LiNO3 –NaNO3 –KNO3 eutectic was studied by simultaneous DSC/TG–MS analysis in atmospheres of argon, nitrogen, oxygen and air. The results showed that evolution of the main gaseous species NO was detected at 325 ◦ C in an atmosphere of argon, at 425 ◦ C in an atmosphere of nitrogen, at 475 ◦ C in an atmosphere of air and at 540 ◦ C in an atmosphere of oxygen, demonstrating the favourable effect of pO2 on the reversible nitrite/nitrate equilibrium. Prior to melting, the eutectic underwent endothermic (␣/) solid–solid type transformation at 87 ◦ C. The melting point was 121 ◦ C, and the solidification point 98 ◦ C. Under-cooling of the salt coincided with the onset of the (␣/) solid–solid transformation upon heating and this is likely due to the delayed nucleation upon cooling at the relatively fast rate of 10 ◦ C/min in these experiments. At a temperature of 500 ◦ C in air, TG analysis showed that the long term stability of the salt was limited and this was confirmed by DSC.
This project has been supported by the NSW Government through the NSW Science Leveraging Fund and by the Australian Government through the Australian Solar Institute (ASI). The Australian Government, through ASI, is supporting Australian research and development in solar photovoltaic and concentrating solar power technologies to help solar power become cost competitive with other energy sources. References [1] A. Gil, M. Medrano, I. Martorell, A. Lazaro, P. Dolado, B. Zalba, L.F. Cabeza, State of the art on high temperature thermal energy storage for power generation. Part 1 – Concepts, materials and modellization, Renew. Sustain. Energy Rev. 14 (2010) 31–55. [2] M. Medrano, A. Gil, I. Martorell, X. Potau, L.F. Cabeza, State of the art on high temperature thermal energy storage for power generation. Part 2 – Case studies, Renew. Sustain. Energy Rev. 14 (2010) 56–72. [3] R.I. Dunn, P.J. Hearps, M.N. Wright, Molten-salt power towers: newly commercial concentrating solar storage, Proc. IEEE 100 (2) (2012) 504–515.
42
R.I. Olivares, W. Edwards / Thermochimica Acta 560 (2013) 34–42
[4] R. Olivares, The thermal stability of molten nitrite/nitrates salt for solar thermal energy storage in different atmospheres, Solar Energy 86 (2012) 2576–2583. [5] E.S. Freeman, The kinetics of the thermal decomposition of sodium nitrate and of the reaction between sodium nitrite and oxygen, J. Phys. Chem. 60 (11) (1956) 1471–1600. [6] C.M. Kramer, DSC of sodium and potassium nitrates and nitrites, Thermochim. Acta 55 (1982) 11–17. [7] C.M. Kramer, Z. Munir, Z.A. Volponi, Simultaneous dynamic thermogravimetry and mass spectrometry of the evaporation of alkali metal nitrates and nitrites, J. Therm. Anal. 27 (1983) 401–408. [8] R.W. Bradshaw, D.E. Meeker, High-temperature stability of ternary nitrate molten salts for solar thermal energy systems, Solar Energy Mater. 21 (1990) 51–60. [9] G.D. Sirotkin, Equilibrium in melts of the nitrates and nitrites of sodium potassium, Russ. J. Inorg. Chem. 4 (11) (1959) 1180–1184. [10] A. Buchler, J. Stauffer, Gaseous alkali–nitrogen–oxygen alkali–phosphorous–oxygen compounds, J. Phys. Chem. 70 (1966) 4092–4095. [11] R.F. Bartholomew, A study of the equilibrium KNO3 (l) = KNO3 (l) + 1/202 (g) over the temperature range 550–750 ◦ C, J. Phys. Chem. 70 (11) (1966) 3442–3446. [12] S. Gordon, C. Campbell, Differential thermal analysis of inorganic compounds nitrates and perchlorates of the alkali and alkaline earth groups and their subgroups, Anal. Chem. 27 (7) (1955) 1102–1109. [13] B. Bond, P. Jacobs, The thermal decomposition of sodium nitrate, J. Chem. Soc. A (1966) 1265–1269. [14] R.N. Kust, J.D. Burke, Thermal decomposition in alkali metal nitrate melts, Inorg. Nucl. Chem. Lett. 6 (1970) 333–335. [15] O. Abe, T. Utsunomiya, T. Hoshino, The thermal stability of binary alkali metal nitrates, Thermochim. Acta 78 (1984) 251–260.
[16] T. Wang, D. Mantha, G. Reddy, Thermal stability of the eutectic composition in LiNO3 –NaNO3 –KNO3 ternary system used for thermal energy storage, Solar Energy Mater. Solar Cells 100 (2012) 162–168. [17] D.A. Nissen, D.E. Meeker, Nitrate/nitrite chemistry in sodium nitrate–potassium nitrate melts, Inorg. Chem. 22 (5) (1983) 716–721. [18] R.W. Bradshaw, N.P. Siegel, Molten nitrate salt development for thermal energy storage in parabolic trough solar power systems, in: Energy Sustainability ES (2008) Jacksonville, Florida, 4 p., 2008. [19] G.W.H. Höhne, Remarks on the calibration of DSC, J. Therm. Anal. 37 (1991) 1987–2000. [20] R. Sabbah, A. Xu-wu, J.S. Chickos, M.L.P. Leitao, M.V. Roux, L.A. Torres, Reference materials for calorimetry and differential thermal analysis, Thermochim. Acta 331 (2) (1999) 93–204. [21] H.K. Cammenga, W. Eysel, E. Gmelin, W. Hemminger, G.W.H. Höhne, S.M. Sarge, The temperature calibration of scanning calorimeters. Part 2 – Calibration substances, Thermochim. Acta 219 (1993) 333–342. [22] S.M. Sarge, E. Gmelin, G.W.H. Höhne, H.K. Cammenga, W. Hemminger, W. Eysel, The caloric calibration of scanning calorimeters, Thermochim. Acta 247 (1994) 129–168. [23] Q. Peng, X. Wei, J. Ding, J. Yang, X. Yang, High-temperature thermal stability of molten salt materials, Int. J. Energy Res. 32 (2008) 1164–1174. [24] J. Raade, D. Padowitz, Development of molten salt heat transfer fluid with low melting point and high thermal stability, J. Solar Eng. 33 (2011) 0310131–031013-6. [25] T. Bauer, D. Laing, U. Kroner, R. Tamme, Sodium nitrate for high temperature latent heat storage, in: The 11th International Conference on Thermal Energy Storage – Effstock, 14–17 June, Stockholm, Sweden, 8 pp., 2009. [26] K.H. Stern, High temperature properties and decomposition of inorganic salts. Part 3 – Nitrates and nitrites, J. Phys. Chem. Ref. Data 1 (3) (1972) 747–772.
Contents lists available at SciVerse ScienceDirect
Thermochimica Acta journal homepage: www.elsevier.com/locate/tca
LiNO3 –NaNO3 –KNO3 salt for thermal energy storage: Thermal stability evaluation in different atmospheres Rene I. Olivares ∗ , William Edwards CSIRO Energy Centre, Newcastle, Australia
a r t i c l e
i n f o
Article history: Received 20 November 2012 Received in revised form 20 February 2013 Accepted 26 February 2013 Available online 16 March 2013 Keywords: Thermal energy storage (TES) Heat transfer fluid (HTF) Concentrating solar power (CSP) Eutectic LiNO3 –NaNO3 –KNO3 Thermal stability
a b s t r a c t The thermal stability of the eutectic LiNO3 –NaNO3 –KNO3 salt was investigated by simultaneous differential scanning calorimetry, thermogravimetry and mass spectrometry (DSC/TG–MS). The work was carried out between room temperature and 1000 ◦ C in blanket gas atmospheres of argon, nitrogen, oxygen and air. The stability of the salt, as measured by the gases evolving from the melt, was influenced by the atmosphere. Evolution of the main gaseous species NO was detected at 325 ◦ C in an atmosphere of argon, at 425 ◦ C in an atmosphere of nitrogen, at 475 ◦ C in an atmosphere of air and at 540 ◦ C in an atmosphere of oxygen. Prior to melting, the eutectic underwent endothermic (␣/) solid–solid type transformation at 87 ◦ C. The melting point was 121 ◦ C, and the solidification point 98 ◦ C. Under-cooling of the salt coincided with the onset of the (␣/) solid–solid transformation upon heating. At a temperature of 500 ◦ C in air, TG analysis showed that the long-term stability of the salt was limited and this was confirmed by DSC. Uncertainty analysis indicated that the measurement of temperature is accurate to ±2.7 ◦ C. © 2013 Elsevier B.V. All rights reserved.
1. Introduction Solar thermal plants which concentrate the Sun’s energy to produce steam and electricity often use molten salt mixtures as a heat transfer fluid (HTF) and/or as thermal energy storage (TES) medium [1–3]. It is desirable that such a salt should have a low melting point in addition to being stable at the highest possible temperature. Nitrates based salts are generally used at temperatures below 500 ◦ C. Operating such salts at temperatures higher than 500 ◦ C may be possible by rigorous control of the atmosphere. In previous contribution [4], using the same methodology described in this work, it was shown that atmosphere had an important effect on the thermal stability of the nitrite/nitrates based ternary salt (NaNO2 –NaNO3 –KNO3 ) allowing for the maximum operating range of the salt to be extended at high pO2 . Oxidising atmosphere however changed the chemistry of the salt by converting nitrite (NO2 − ) to nitrate (NO3 − ) increasing the melting point. In this work, we revise the low melting point nitrates only, eutectic LiNO3 –NaNO3 –KNO3 , under the influence of different gas atmospheres. A low melting point salt that is stable at temperatures higher than 700 ◦ C is needed to increase the heat-to-electricity conversion efficiency for Rankine cycle systems, for use in Brayton cycle
∗ Corresponding author at: 10 Murray Dwyer Circuit, Steel River Estate, Mayfield West, NSW 2304, Australia. Tel.: +61 2 4960 6253. E-mail address: [email protected] (R.I. Olivares). 0040-6031/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.tca.2013.02.029
systems and to supply high grade heat to industrial processes. The high temperature thermal stability of the salt is then an important aspect that requires consideration given that it will determine the long term viability of its use in a high temperature storage system. Numerous reports addressing thermal stability for nitrate based systems are available [5–8], albeit that some ambiguous conclusions have been made as well. It is therefore essential that the thermo-chemical behaviour, thermal stability, and mode of decomposition of the salt are studied in some detail. For this, the low melting point ternary eutectic system LiNO3 –NaNO3 –KNO3 was characterised in the present work between room temperature and 1000 ◦ C under various gas atmospheres. 1.1. Literature The principal mode of thermal decomposition in nitrate salts has been agreed amongst researchers [5–10]; to correspond to the main reaction: NO3 − ↔ NO2 − + 1/2 O2 . Further decomposition also takes place with the evolution of the oxides of nitrogen, particularly at higher temperatures [11]. In some studies [5,12,13] the decomposition temperature has been reported as the temperature at which oxygen, nitrogen or nitrous oxide is detected in the gas phase. In the work of Gordon and Campbell [12] the single salts alkali metal nitrates were observed to undergo a thermal reaction at temperatures 100–300 ◦ C above their melting points as indicated by bubbling of the molten salt. The evolution of nitrous fumes observed by Gordon and Campbell [12], occurred at temperatures ranging from about 200–350 ◦ C above
R.I. Olivares, W. Edwards / Thermochimica Acta 560 (2013) 34–42 Table 1 Eutectic LiNO3 –NaNO3 –KNO3 formulation. Compound
wt%
mol%
LiNO3 NaNO3 KNO3
30 18 52
37 14 49
Note: Compositions were determined using phase diagrams calculated with FactSage software (Fig. 1).
the initial bubbling reaction. This was claimed to be as a result of the decomposition of the nitrate and/or nitrite to oxide. The later stage of the decomposition of these nitrates was still occurring at temperatures as high as 900 ◦ C which was the upper limit of the apparatus employed [12]. Kust and Burke [14] conducted experiments with a pure equimolar sodium potassium nitrate melt in an oxygen rich atmosphere (pO2 at 0.85 atm) and measured high oxide ion concentration (O2− ) in the molten salt which corresponded to the dissociation constant for nitrite ion (NO2 − ); based on the equilibrium NO3 − → NO2 − + O2− , (nitrite ion concentration reached equilibrium within 7–16 h at temperatures between 295 ◦ C and 340 ◦ C in Kust and Burke experiments [14]) demonstrating that nitrate ion (NO3 − ) could decompose to a measurable extent at temperatures as low as 295 ◦ C. The thermal stability of binary mixes of alkali metal nitrates as investigated by Abe et al. [15], by means of differential thermal analysis (DTA) and evolved gas analysis (EGA), concluded that the nitrate melts are thermally unstable. The evolved gas analysis measurements [15] indicated that the decomposition of nitrates took place independently and followed the ranking order of thermodynamic stability as determined by the reactions sequence MNO3 → MNO2 + 1/2O2 followed by 2MNO2 → M2 O + NO + NO2 . On this basis the decomposition of a mixture of LiNO3 –NaNO3 –KNO3 would be determined by the least thermodynamically stable species, in this case LiNO3 followed by NaNO3 and then KNO3 , a possibility if the solution is close to ideal. Most recently the thermal stability of the eutectic composition in LiNO3 –NaNO3 –KNO3 ternary salt was reported by Wang et al. [16], after studies where temperature was cycled between 75 ◦ C and 500 ◦ C in argon gas atmosphere, using a TG/DTA method. Wang et al. [16] observed that once the salt had lost its water content after the first melting cycle it reached a stable condition. In long term stability tests at 500 ◦ C though, there was a weight change of ∼8.6 wt% and [16] concluded that the salt was unstable beyond this point. Given that weight loss alone as a measure of stability may be compromised due to some salt evaporation, further analysis is required before the practical limit of salt stability can be ascertained. In the present work the thermo-chemical behaviour and thermal stability of the LiNO3 –NaNO3 –KNO3 eutectic was studied from at least three different points of view: (i) the temperature at which the rapid evolution of gases, NO and/or NO2 and O2 , are detected by means of mass spectrometry (MS), (ii) the temperature of the melt at which an irreversible endothermic peak of decomposition is resolved by differential scanning calorimetry (DSC), and (iii) the temperature at which rapid weight loss is observed in a thermogravimetric curve (TG). A combination of these three criteria is used for elucidation of the limiting temperature for operation of the salt as would apply in a TES installation. The chemistry and equilibrium reactions by which molten nitrites/nitrates may interact with the atmosphere can be found in Refs. [4,17,18,26]. 2. Experimental The eutectic salt composition was prepared by mixing the individual components in the weight proportions given in Table 1.
35
Analytical reagent grade chemicals LiNO3 –NaNO3 –KNO3 supplied by Ajax (UNIVAR) were dried overnight in vacuum prior to weighing and mixing. 2.1. Experimental procedure A 20 g mixture of the pre-dried pure chemicals was melted inside an alumina crucible in a muffle furnace and allowed to equilibrate in air at 250 ◦ C for 2 h before casting into a stainless steel mould to quench and solidify. The solidified salt was kept under vacuum at 65 ◦ C at all times from which a small sub-sample was taken and used for simultaneous DSC/TG–MS experimentation as needed. About 15–25 mg were ground and loaded into a 100 L Al2 O3 crucible in a SETARAM DSC/TG SETSYS Evolution analyser, coupled with a Pfeiffer QUADSTAR-422 mass spectrometer. Due to the presence of LiNO3 , the sample was highly hygroscopic and it was inevitable that upon handling the sub-sample, moisture was absorbed. The data acquisition was done by programming the MS to scan for masses from 1 to 60 (amu) to cover the range of possible gaseous species reporting to the gas phase. Identification of gaseous species was done by matching the relative Ion Current Intensities measured by the MS, with the corresponding spectra available in the QUADSTAR-422 MS library database. The experiments were carried out from room temperature up to 1000 ◦ C at a heating-rate of 10 ◦ C/min in atmospheres of argon, nitrogen, air and oxygen respectively. Once the atmosphere around the sample was set, the gas was allowed to flow at a rate of 30 ml/min. Five melting–solidification cycles were initially carried out between 50 ◦ C and 400 ◦ C before the full decomposition up to 1000 ◦ C was completed. At all times during heating and cooling the carrier gas was maintained at 30 ml/min and the heating rate was at 10 ◦ C/min. A weight loss in the first heating cycle of between 1 and 2 wt% corresponded to the evolution of water absorbed during sample handling; this was detected by the MS, and was not considered in the analysis. The subsequent four remaining cycles did not show any weight loss or water evolution and were used for determination of the melting and solidification points and to check reproducibility of the measurement. At the end of each decomposition run, contamination of the DSC platinum sensor was evident by a dark stain around the sensor that had to be thoroughly cleaned between runs by first soaking in hot water and then heating to 1200 ◦ C in oxygen to remove any residue. 2.2. Experimental error and uncertainty Temperature and heat flow calibration of the DSC was carried out as per the standard method for this type of analysis [19,20] using certified reference materials (CRM) listed in Table 2 to cover the temperature range 100–1050 ◦ C. All calibration curve parameters were estimated by a least-squares optimisation routine (build in Calisto Software–SETARAM). The uncertainty on the temperature and heat flow measurements were calculated from the standard error of the estimate, s, according to: s2 =
1 N − (n + 1)
(Ycorr − Ystd )2
where s is the standard error of the estimate; N, number of data points; n, number of independent variables; Ycorr , corrected temperature of melting, Tcorr ; or, corrected energy of melting, Hcorr ; and Ystd , melting temperature of standard, Tstd ; or, energy of melting of standard, Hstd . From these and using the calculation procedure described, the uncertainty for the temperature is ±2.7 ◦ C and the uncertainty for the energy is ±2.3 J/g. The uncertainty of the standard (see Table 2)
36
R.I. Olivares, W. Edwards / Thermochimica Acta 560 (2013) 34–42
Table 2 Standard certified reference materials (CRM)–DSC temperature and energy calibration. CRM melting, Tstd (◦ C) [21]
CRM
Heating rate (◦ C/min)
In
10
156.59 ± 0.01
Pb
10
327.47 ± 0.02
Al
10
660.33 ± 0.05
Au Pb Pb
10 5 20
1064.18 327.47 ± 0.02
Measured melting, Tmeas (◦ C)
Corrected melting, Tcorr (◦ C)
156.46 156.39 326.52 326.74 660.69 659.99 1065.11 325.56 326.99
154.08 154.01 325.06 325.28 661.05 660.35 1067.67 323.62 326.49
and measured temperatures (noise level ±5 mK) are negligible compared to the standard error of the estimate, s. 3. Results and discussion The present work concerns the low melting point eutectic LiNO3 –NaNO3 –KNO3 prepared in the laboratory from high purity chemicals. This formulation was studied over the temperature range up to 1000 ◦ C in atmospheres of argon, nitrogen, oxygen and air using simultaneous DSC/TG and MS analysis (Fig. 1). 3.1. Melting and solidification point determination The melting and solidification temperatures respectively were determined as per conventional DSC analysis at the point where the baseline intercepts the onset of the endothermic melting event or the exothermic solidification event. An example of a heating–cooling cycle is shown in Fig. 2. Repeatability of the measurements in four heating–cooling cycles between 50 ◦ C and 400 ◦ C is given in Table 3; the reproducibility of the results was excellent. Prior to melting, the eutectic underwent an endothermic (␣/) solid–solid type transformation starting at 87 ◦ C. The melting point was 121 ◦ C, and the solidification point 98 ◦ C. Significant
CRM energy of melting, Hstd (J/g) [22]
Corrected energy of melting, Hcorr (J/g)
28.51 ± 0.19
28.49
23.00 ± 0.06
23.09
401.30 ± 1.6 64.50
404.64 63.51
under-cooling of the salt was evident, possibly due to delayed nucleation because of the relatively fast cooling rate of 10 ◦ C/min. Considering that at the eutectic composition the liquidus and solidus temperatures should be the same, the under-cooling observed here is most likely a function of the cooling rate.
3.2. Salt decomposition and evolved gas analysis The results for the simultaneous DSC/TG–MS experiments with the salt in atmospheres of argon, nitrogen, oxygen and air are shown in Figs. 3–6 respectively. The figures show superimposed in the same graph: (i) the ion current intensity (A) for the detection by the MS of the evolved gaseous species, (ii) the TG (mg) curve to monitor the weight change, and (iii) the DSC (mW) trace to identify the temperature at which heat events (endothermic/exothermic) occur. In all cases the figures illustrate a wide endothermic heat flow through which coincides with a sharp drop in weight once a critical temperature is reached, reflecting a rapid rate of decomposition of the salt. Independent measurements indicated that the reproducibility of the results was excellent. In an inert atmosphere of argon, Fig. 3, the gaseous species detected by the MS were NO, O2 , O, NO2 , N, N2 and also N2 O. The evolution of nitrous oxide (NO) is first noticed at 325 ◦ C and the evolution of nitrogen species (N, N2 ) at the much higher
Fig. 1. Calculated LiNO3 –NaNO3 –KNO3 eutectic composition–FactSage6.3.
R.I. Olivares, W. Edwards / Thermochimica Acta 560 (2013) 34–42
37
Fig. 2. Melting point and solidification point determination of eutectic LiNO3 –NaNO3 –KNO3 .
temperatures of 584 ◦ C and 693 ◦ C respectively. The O2 and NO2 species evolved at temperatures between these two. The evolution of N2 O was apparent at 895 ◦ C. Irreversible decomposition of the eutectic LiNO3 –NaNO3 –KNO3 in this case can be considered to occur at a temperature between the point at which rapid weight loss begins on approaching 602 ◦ C and the bulk of decomposition is identified by the broad endothermic heat event at ∼670 ◦ C in the DSC trace. In an inert atmosphere of nitrogen, Fig. 4 shows that the evolution of gaseous species NO started at 425 ◦ C followed by O2 at 474 ◦ C with NO2 and mono-atomic oxygen (O) at about 620 ◦ C; evolution of N2 O was apparent at 825 ◦ C. Irreversible decomposition in this case can be considered to occur at a temperature between the point at which rapid weight loss begins approaching 584 ◦ C and the bulk of decomposition identified by the broad endothermic heat event at ∼655 ◦ C in the DSC trace. Because the blanket atmosphere was nitrogen, the background ion current intensity in the MS was too large and the species N and N2 were not detected in this experiment. For the experiment in an atmosphere of oxygen, Fig. 5, the gaseous species detected evolving from the melt were NO at 540 ◦ C
and NO2 at between 422 and 540 ◦ C, with N2 O at the higher temperature of 855 ◦ C; whereas for the experiment in air, Fig. 6, only NO and NO2 gases were detected evolving at 475 ◦ C and 550 ◦ C respectively. The MS background ion current intensities for oxygen and nitrogen species under these atmospheres (oxygen and air) are too large and prevent the detection by the MS of the possible evolution of N, N2 O, O2 from the melt. In oxygen (Fig. 5) the critical temperature after which rapid weight loss began was 577 ◦ C, and the bulk of decomposition as per the large endothermic event in the DSC trace occurred at ∼608 ◦ C. In air (Fig. 6) the critical temperature after which rapid weight loss occurred was 602 ◦ C, and the bulk of decomposition as per the large endothermic event in the DSC trace at ∼625 ◦ C. The data taken in Figs. 3–6 are further summarised in Table 4; the criteria used to select the maximum temperature in the TG curve was the point at which a 1 wt% weight loss is reached. The TG curves in Figs. 3–6 suggest that the salt could be taken to temperatures higher than 550 ◦ C and even to 600 ◦ C in air. TGA alone has been used by others [23,24] to evaluate thermal stability of molten salts for TES. Raade and Padowitz [24] defined the high temperature limit of the salt as that where the salt is observed to rapidly begin to lose weight beyond a maximum acceptable
Table 3 Repeatability of melting and solidification point measurement. Melting
␣/ (Solid–solid) transformation ◦
Cycle 2 Cycle 3 Cycle 4 Cycle 5 Average (4 cycles) Solidification
Cycle 2 Cycle 3 Cycle 4 Cycle 5 Average (4 cycles)
Melting process ◦
Onset, T ( C)
Max, T ( C)
H (J/g)
Onset, T (◦ C)
Max, T (◦ C)
H (J/g)
86 86 86 88 87
91 94 93 95 93
12.5 6.9 6.1 4.1 7.4
121 122 120 122 121
128 128 127 128 128
141.5 140.5 147.2 135.5 141.2
Solidification process Onset, T (◦ C)
Max, T (◦ C)
H (J/g)
98 98 98 98 98
93 93 93 94 93
−92.9 −97.4 −97.7 −95.5 −95.9
38
R.I. Olivares, W. Edwards / Thermochimica Acta 560 (2013) 34–42
Fig. 3. DSC/TG–MS analysis for thermal decomposition of LiNO3 –NaNO3 –KNO3 salt in argon atmosphere.
Fig. 4. DSC/TG–MS analysis for thermal decomposition of LiNO3 –NaNO3 –KNO3 salt in nitrogen atmosphere.
R.I. Olivares, W. Edwards / Thermochimica Acta 560 (2013) 34–42
39
Fig. 5. DSC/TG–MS analysis for thermal decomposition of LiNO3 –NaNO3 –KNO3 salt in oxygen atmosphere.
weight loss of 3 wt%. Long term stability experiments for the eutectic LiNO3 –NaNO3 –KNO3 salt prepared for the present work were carried out in an air atmosphere at fixed temperatures of 400 ◦ C and 500 ◦ C over 4 h. These indicated that at 400 ◦ C a small weight loss over 4 h was consistent with some loss due to salt evaporation from the open crucible (supported by the fact that at the end of each experiment the crucible was stuck to the base of the sensor due to condensed vapours). At 500 ◦ C very rapid weight loss took place as shown in Fig. 7. It is noted that the carrier gas flow rate was 30 ml/min and this may have also contributed to some losses by salt carry over. The condition of the salt was subsequently checked by DSC and this is shown in Figs. 8 and 9. DSC analysis in Fig. 8, for the sample after 4 h at 400 ◦ C, showed that no change in thermo chemical behaviour occurred and the curve reproduced well the DSC analysis of the fresh sample (see Fig. 2). DSC analysis in Fig. 9
showed that for the sample after 4 h at 500 ◦ C, the salt presented a significantly different thermo chemical behaviour. It is recognised [25] that in general the thermal decomposition of nitrates can be divided at least in three zones: (i) formation of little nitrite soon after melting, (ii) nitrate–nitrite equilibrium in the range 400 ◦ C to 550 ◦ C (MNO3 → MNO2 + 1/2O2 ) dependent on pO2 , and (iii) decomposition with the release of nitrogen oxides at greater than 550 ◦ C (2MNO2 → M2 O + NO + NO2 ), where M can represent either Li, Na or K. A survey of the literature provided by Stern [26] list the gaseous products from decomposition of nitrates and nitrites as follows; O2 , N2 , NO, NO2 , N2 O3 , N2 O4 , N2 O5 . Which of these are produced, depend on the salt, the atmosphere, the temperature, and the experimental conditions. From the thermodynamic data compiled by Stern [26], it can be seen that N2 O5 is unstable with respect to dissociation into
Table 4 Summary for the effect of atmosphere on thermal decomposition of LiNO3 –NaNO3 –KNO3 salt. Atmosphere
Argon Nitrogen Oxygen Air
Melting, T (◦ C)
121 122 120 122
DSC–event endothermic, T (◦ C)
88 87 86 88
Irreversible DSC event, T (◦ C) decomposition
671 655 608 625
TG rapid wt loss after, T (◦ C)
602 584 577 602
All temperatures reported are rounded figures and have an uncertainty of ±2.7 ◦ C; and nd = not detected.
Evolved gaseous species detected by MS–temperature of detection (◦ C)
NO
O2
N
O
NO2
N2
N2 O
325 425 540 475
418 474 nd nd
584 nd nd nd
588 619 nd nd
659 622 422–540 550
693 nd nd nd
895 825 855 nd
40
R.I. Olivares, W. Edwards / Thermochimica Acta 560 (2013) 34–42
Fig. 6. DSC/TG–MS analysis for thermal decomposition of LiNO3 –NaNO3 –KNO3 salt in air atmosphere.
N2 O4 or N2 O3 and the equilibrium shifts markedly to NO2 with increasing temperature, such that above 127 ◦ C the concentration of N2 O4 is negligible and N2 O3 is unstable with respect to both NO2 and NO. The equilibrium NO + 1/2O2 = NO2 is very temperature dependent, where up to 227 ◦ C NO2 predominates and above this temperature both species are significant. N2 and O2 are very stable with respect to NO and NO2 ; if N2 is formed during salt decomposition it will not react further, but O2 may react with NO and nitrites. If all equilibriums during salt decomposition were established quickly, N2 and O2 would be the only gaseous species. The reaction NO2 = 1/2N2 + O2 is very slow below 727 ◦ C and if NO2 is the primary product of salt decomposition, it will not decompose appreciably so only the equilibrium NO + 1/2O2 = NO2 needs to be considered [26]. Likewise the decomposition of NO to its elements N and O can be neglected. The combined DSC/TG–MS analysis used in the experiments described here showed that for the ternary eutectic LiNO3 –NaNO3 –KNO3 , the blanket gas atmosphere had a significant role in determining the mode of decomposition, predominantly due to its effect on pO2 and the reversible equilibrium nitrite/nitrate. The stability of the salt, as measured by the gases evolving from the melt, indicated that the main gaseous species NO was detected at 325 ◦ C in an atmosphere of argon, at 425 ◦ C in an atmosphere of nitrogen, at 475 ◦ C in an atmosphere of air and at 540 ◦ C in an atmosphere of oxygen, albeit without measurable weight loss. By two assessment criteria, TG and DSC, the salt could be heated up to 600 ◦ C in air with negligible decomposition. For long term stability, given the changes seen in the eutectic after 4 h at 500 ◦ C (Figs. 7 and 9), it is unlikely that this system could
be used at a temperature of 600 ◦ C for an extended period given that the relative concentration of constituents would have changed having an impact on the constitution of the eutectic. Wang et al. [16], from characterisation by XRD–SEM/EDS analysis of the eutectic LiNO3 –NaNO3 –KNO3 before and after long term thermal stability experiments in an argon atmosphere, showed that the least stable lithium nitrate compound was decomposing to its oxides at the high temperature of 500 ◦ C. In Wang et al.’s [16] experiments, the limit of stability was found to be 435 ◦ C in an argon gas atmosphere and this was based on thermo-gravimetric (TG) analysis alone. The incorporation of evolved gas analysis (EGA) as used in this work, allows for a more accurate evaluation.
Fig. 7. Weight variation of eutectic LiNO3 –NaNO3 –KNO3 eutectic at 400 ◦ C and 500 ◦ C.
R.I. Olivares, W. Edwards / Thermochimica Acta 560 (2013) 34–42
41
Fig. 8. DSC analysis of LiNO3 –NaNO3 –KNO3 eutectic after sample was held in air at 400 ◦ C for 4 h.
Fig. 9. DSC analysis of LiNO3 –NaNO3 –KNO3 eutectic after sample was held in air at 500 ◦ C for 4 h.
4. Conclusions
Acknowledgements
The thermo-chemical behaviour and thermal stability of LiNO3 –NaNO3 –KNO3 eutectic was studied by simultaneous DSC/TG–MS analysis in atmospheres of argon, nitrogen, oxygen and air. The results showed that evolution of the main gaseous species NO was detected at 325 ◦ C in an atmosphere of argon, at 425 ◦ C in an atmosphere of nitrogen, at 475 ◦ C in an atmosphere of air and at 540 ◦ C in an atmosphere of oxygen, demonstrating the favourable effect of pO2 on the reversible nitrite/nitrate equilibrium. Prior to melting, the eutectic underwent endothermic (␣/) solid–solid type transformation at 87 ◦ C. The melting point was 121 ◦ C, and the solidification point 98 ◦ C. Under-cooling of the salt coincided with the onset of the (␣/) solid–solid transformation upon heating and this is likely due to the delayed nucleation upon cooling at the relatively fast rate of 10 ◦ C/min in these experiments. At a temperature of 500 ◦ C in air, TG analysis showed that the long term stability of the salt was limited and this was confirmed by DSC.
This project has been supported by the NSW Government through the NSW Science Leveraging Fund and by the Australian Government through the Australian Solar Institute (ASI). The Australian Government, through ASI, is supporting Australian research and development in solar photovoltaic and concentrating solar power technologies to help solar power become cost competitive with other energy sources. References [1] A. Gil, M. Medrano, I. Martorell, A. Lazaro, P. Dolado, B. Zalba, L.F. Cabeza, State of the art on high temperature thermal energy storage for power generation. Part 1 – Concepts, materials and modellization, Renew. Sustain. Energy Rev. 14 (2010) 31–55. [2] M. Medrano, A. Gil, I. Martorell, X. Potau, L.F. Cabeza, State of the art on high temperature thermal energy storage for power generation. Part 2 – Case studies, Renew. Sustain. Energy Rev. 14 (2010) 56–72. [3] R.I. Dunn, P.J. Hearps, M.N. Wright, Molten-salt power towers: newly commercial concentrating solar storage, Proc. IEEE 100 (2) (2012) 504–515.
42
R.I. Olivares, W. Edwards / Thermochimica Acta 560 (2013) 34–42
[4] R. Olivares, The thermal stability of molten nitrite/nitrates salt for solar thermal energy storage in different atmospheres, Solar Energy 86 (2012) 2576–2583. [5] E.S. Freeman, The kinetics of the thermal decomposition of sodium nitrate and of the reaction between sodium nitrite and oxygen, J. Phys. Chem. 60 (11) (1956) 1471–1600. [6] C.M. Kramer, DSC of sodium and potassium nitrates and nitrites, Thermochim. Acta 55 (1982) 11–17. [7] C.M. Kramer, Z. Munir, Z.A. Volponi, Simultaneous dynamic thermogravimetry and mass spectrometry of the evaporation of alkali metal nitrates and nitrites, J. Therm. Anal. 27 (1983) 401–408. [8] R.W. Bradshaw, D.E. Meeker, High-temperature stability of ternary nitrate molten salts for solar thermal energy systems, Solar Energy Mater. 21 (1990) 51–60. [9] G.D. Sirotkin, Equilibrium in melts of the nitrates and nitrites of sodium potassium, Russ. J. Inorg. Chem. 4 (11) (1959) 1180–1184. [10] A. Buchler, J. Stauffer, Gaseous alkali–nitrogen–oxygen alkali–phosphorous–oxygen compounds, J. Phys. Chem. 70 (1966) 4092–4095. [11] R.F. Bartholomew, A study of the equilibrium KNO3 (l) = KNO3 (l) + 1/202 (g) over the temperature range 550–750 ◦ C, J. Phys. Chem. 70 (11) (1966) 3442–3446. [12] S. Gordon, C. Campbell, Differential thermal analysis of inorganic compounds nitrates and perchlorates of the alkali and alkaline earth groups and their subgroups, Anal. Chem. 27 (7) (1955) 1102–1109. [13] B. Bond, P. Jacobs, The thermal decomposition of sodium nitrate, J. Chem. Soc. A (1966) 1265–1269. [14] R.N. Kust, J.D. Burke, Thermal decomposition in alkali metal nitrate melts, Inorg. Nucl. Chem. Lett. 6 (1970) 333–335. [15] O. Abe, T. Utsunomiya, T. Hoshino, The thermal stability of binary alkali metal nitrates, Thermochim. Acta 78 (1984) 251–260.
[16] T. Wang, D. Mantha, G. Reddy, Thermal stability of the eutectic composition in LiNO3 –NaNO3 –KNO3 ternary system used for thermal energy storage, Solar Energy Mater. Solar Cells 100 (2012) 162–168. [17] D.A. Nissen, D.E. Meeker, Nitrate/nitrite chemistry in sodium nitrate–potassium nitrate melts, Inorg. Chem. 22 (5) (1983) 716–721. [18] R.W. Bradshaw, N.P. Siegel, Molten nitrate salt development for thermal energy storage in parabolic trough solar power systems, in: Energy Sustainability ES (2008) Jacksonville, Florida, 4 p., 2008. [19] G.W.H. Höhne, Remarks on the calibration of DSC, J. Therm. Anal. 37 (1991) 1987–2000. [20] R. Sabbah, A. Xu-wu, J.S. Chickos, M.L.P. Leitao, M.V. Roux, L.A. Torres, Reference materials for calorimetry and differential thermal analysis, Thermochim. Acta 331 (2) (1999) 93–204. [21] H.K. Cammenga, W. Eysel, E. Gmelin, W. Hemminger, G.W.H. Höhne, S.M. Sarge, The temperature calibration of scanning calorimeters. Part 2 – Calibration substances, Thermochim. Acta 219 (1993) 333–342. [22] S.M. Sarge, E. Gmelin, G.W.H. Höhne, H.K. Cammenga, W. Hemminger, W. Eysel, The caloric calibration of scanning calorimeters, Thermochim. Acta 247 (1994) 129–168. [23] Q. Peng, X. Wei, J. Ding, J. Yang, X. Yang, High-temperature thermal stability of molten salt materials, Int. J. Energy Res. 32 (2008) 1164–1174. [24] J. Raade, D. Padowitz, Development of molten salt heat transfer fluid with low melting point and high thermal stability, J. Solar Eng. 33 (2011) 0310131–031013-6. [25] T. Bauer, D. Laing, U. Kroner, R. Tamme, Sodium nitrate for high temperature latent heat storage, in: The 11th International Conference on Thermal Energy Storage – Effstock, 14–17 June, Stockholm, Sweden, 8 pp., 2009. [26] K.H. Stern, High temperature properties and decomposition of inorganic salts. Part 3 – Nitrates and nitrites, J. Phys. Chem. Ref. Data 1 (3) (1972) 747–772.