先進(jìn)的太陽能熱電廠的碳酸鹽熔融鹽外文文獻(xiàn)

Contents lists available at ScienceDirect Solar Energy Materials and Solar Cells journal homepage Molten carbonate salts for advanced solar thermal energy power plants Cover gas eff ect on fl uid thermal stability Sonia Fereresa Cristina Prietoa b Pau Gim nez Gavarrellb Alfonso Rodr gueza Pedro Enrique S nchez Jim nezc Luis A P rez Maquedac aAbengoa C Energ a Solar 1 41014 Sevilla Spain bDept Ingenier a Energ tica ETSI Universidad de Sevilla Spain cInstituto de Ciencia de Materiales de Sevilla C S I C Universidad de Sevilla C Am rico Vespucio no 49 41092 Sevilla Spain A R T I C L E I N F O Keywords Molten salts Carbonates Thermal energy storage Heat transfer fl uid Thermal stability Decomposition A B S T R A C T The eutectic mixture Li2CO3 Na2CO3 K2CO3 is investigated as a high temperature heat transfer fl uid and storage medium alternative for molten salt solar thermal power plants This salt has an operating temperature range of 400 700 C enabling the use of higher temperature effi ciency power cycles However this carbonate mixture is known to thermally decompose in air This study evaluates the thermal stability of the salt mixture under diff erent cover gases air nitrogen carbon dioxide and an 80 20 carbon dioxide air mixture Initial char acterization is performed through thermogravimetric analysis TGA followed by larger scale testing in a custom made reactor to simulate conditions closer to its practical use The results show improved thermal stability with a CO2 atmosphere The decomposition kinetics under diff erent cover gases are estimated from TGA data However larger scale longer duration experiments show much slower decomposition rates compared to the classical TGA approach These fi ndings indicate that the main contribution to mass loss in TGA is due to vaporization rather than thermal decomposition Thus a proper evaluation of the molten salt s thermal stability can only be obtained from reactor experiments where vaporization is inhibited Very long induction periods of the order of days are observed suggesting that the kinetic decomposition mechanism is a nucleation and growth type Other considerations for future plants incorporating these high temperature salts are discussed 1 Introduction The competitive advantage of Solar Thermal Energy STE plants is the capability of overcoming the natural intermittencies of the sun with Thermal Energy Storage TES to produce electricity continuously be yond daylight hours Current STE power plants use nitrate molten salts as their storage media and depending of the technology also as their Heat Transfer Fluid HTF in Molten Salt Tower MST designs There is a constant search in the engineering community for novel HTF and TES media working at higher temperatures to be used with advanced ther modynamic cycles to improve the overall solar to electric conversion effi ciency 1 3 Novel MST plants with higher temperature HTF could feed both subcritical and supercritical cycles increasing the effi ciency of current MST with nitrates 4 Traditional nitrate MST power plants work in the 240 565 C tem perature range using solar salt a 60 40 wt mixture of NaNO3and KNO3 5 Through appropriate mitigation strategies on the container materials and operation procedures their upper temperature limit could be extended by roughly 50 C above the classical practical values However reaching maximum operating temperatures above 700 C is only possible with new salt compositions Carbonates are the natural fi rst option to explore higher temperature molten salts since their corrosion behavior is notoriously better than other high temperature molten salts such as hydroxides fl uorides or sulfi des 6 The ternary carbonate mixture Li2CO3 Na2CO3 K2CO3 LiNaK car bonate has been previously proposed in the literature as a potential candidate for high temperature STE applications despite its high melting temperature Wu et al 6 investigated over thirty six diff erent compositions of LiNaK carbonate salts for sensible heat storage in STE looking to replace molten nitrates with peak salt temperatures from 565 C to higher values between 700 C and 850 C fi nding melting temperatures around 400 C Table 1 summarizes the literature reported data for a LiNaK carbonate mixture compared to the baseline solar salt The evaluation criteria for new salt compositions focuses on their melting point energetic cost thermal stability and corrosion compat ibility A lower melting point Tm is desirable as it reduces the risk of https doi org 10 1016 j solmat 2018 08 028 Received 25 June 2018 Received in revised form 27 August 2018 Accepted 29 August 2018 Corresponding author E mail address sonia fereres S Fereres Solar Energy Materials and Solar Cells 188 2018 119 126 Available online 05 September 2018 0927 0248 2018 Elsevier B V All rights reserved T frozen blockages in pipes A higher decomposition temperature or thermal stability Td increases the operational temperature with possible coupling to higher effi ciency thermodynamic cycles working at higher temperatures The energetic cost is calculated with the mate rial s specifi c heat capacity over the operational temperature range The main handicaps associated with using this carbonate salt as a HTF in advanced MST plants are 1 High melting point in the range of 400 C 7 8 In practice the working temperature for carbonate salts should be even slightly higher to avoid solidifi cation risk However mixing additives such as NaNO3 KCl or NaOH has been proposed to reduce the melting point of LiNaK carbonates by as much as 75 C 8 11 2 High cost 2 4 kg Despite having a higher specifi c heat and a higher operating temperature range than solar salt its increased energetic cost is essentially due to the cost of lithium carbonate 6 kg 3 Low thermal stability in air 9 4 Corrosion performance of this mixture at temperatures near or above 700 C is relatively unknown Recent publications are ana lyzing container materials for this salt to be used in solar thermal power plants 12 13 The viscosity of commercial grade LiNaK carbonate is nearly 40 mPa s at 400 C just above its melting point and 6 2mPa s at 700 C following an exponential dependency 14 This viscosity is comparable yet still higher than the viscosity of commercial solar salt which has a viscosity of approximately 4 6mPa s at 250 C above its melting point and 1 8mPa s at 400 C 14 It is known that carbonate salts are unstable in air 8 9 but they can be stabilized in a carbon dioxide atmosphere by establishing an equilibrium with the CO2gas LiNaK carbonates are being widely in vestigated for other low carbon or carbon sequestration energy related technologies Waste disposal gas cleaning processes CO2concentration and fuel cells are some of the most common applications 6 For ex ample Deng et al 15 investigated the carbonation of Li2O in this ternary carbonate mixture for CO2capture and electrochemical con version via electrolysis with molten salts Others such as Ren et al 16 have recently synthesized carbon nanofi bers by electrolysis of the CO2 dissolved in Li2CO3molten carbonates In all these cases the molten carbonate is in certain equilibrium with dissolved atmospheric CO2 Thus it should not come as a surprise that gaseous CO2can stabilize this molten salt Olivares et al 9 investigated the thermal stability of the LiNaK 32 1 33 4 34 5 wt mixture under three diff erent cover gases CO2 argon and air using conventional thermogravimetric analysis The cover gas was found to have very little eff ect on the melting tempera tures which were between 400 and 405 C However the onset tem perature of the decomposition reaction did change signifi cantly with cover gas reaching values of 601 C 715 C and 1000 C with air argon and CO2respectively The main objective of the present study is to quantify the thermal decomposition of ternary carbonates under conditions similar to those in a MST plant studying the eff ect of dif ferent cover gases air nitrogen carbon dioxide and evaluating if an air leak in a CO2tank is an assumable risk Thus the LiNaK carbonate salt is also tested with cover gas mixtures of diff erent CO2 air pro portions comparing it to its behavior under pure CO2atmospheres Thermogravimetric analysis TGA experiments provide a good fi rst approach to understand the thermal stability of new salt mixtures In these experiments a small sample of the salt is heated in an open crucible at ambient pressure with overfl ow of the cover gas to measure the mass evolution with time and temperature These tests are usually employed to estimate the kinetics of the thermal decomposition reac tions and from there estimate a reliable maximum operation tem perature for a given process However they are not very representative of the actual working conditions in a molten salt TES tank in a solar plant which operates with several tones of salt over 10 15 years Consequently the TGA results presented here are complemented with experiments performed with more relevant custom built tests Larger size samples over longer duration tests are designed to evaluate the salt mixture behavior under a static atmosphere avoiding mass loss due to evaporation and cover gas fl ow renewal inside the TGA equipment Also longer duration tests allow the development of slower decom position reactions appearing over long term high temperature exposure To the best of the authors knowledge it is the fi rst time that the lim itations from standard laboratory scale TGA for molten carbonate salts systems are compared with long term decomposition tests indicating important implications in practice A detailed characterization of this ternary carbonate mixture has been carried out to assess its performance in a large scale pre commercial demonstrator The following properties have been ana lyzed a heat capacity heat of fusion and melting point and b thermal stability and cover gas dependency in both a standard commercial TGA equipment and in a large scale custom made reactor These results lead the path for new engineering considerations for future MST and other energy related systems using molten carbonates 2 Materials and methods Theternaryeutecticcarbonatesalt Li2CO3 Na2CO3 K2CO3 43 5 31 5 25mol was prepared using technical grade Li2CO3 Na2CO3and K2CO3 99 95 purity by Panreac Qu mica S L U Barcelona Spain The salt mixture was weighed dry melted in an oven at 450 C and grinded in a manual mortar to ensure a homogenous powder mixture This mixing procedure was compared to dissolving the dry components in the appropriate proportion in water in a glass vial sonicating for 200 min and evaporating the solvent water over a hot plate at 150 C inside a laboratory hood for 5h to obtain a powder form salt mixture following the process described in detail in 17 2 1 Diff erential scanning calorimeter AQ2000 Diff erentialScanningCalorimeter DSC fromTA Instruments is used to measure the melting point latent heat and spe cifi c heat capacity of the LiNaK carbonate The calibration procedure is performed by using the melting temperatures and latent heat of stan dard certifi ed reference materials In Zn Salt samples weighing 7 12mg are introduced in Tzero standard aluminum crucibles The melting temperature and latent heat are determined using a standard linear DSC procedure at a heating rate of 5 C min pre melting the samples to ensure a good contact between the sample and the crucible Good contact between the bottom of the crucible and the sample is necessary for reliable results 18 This fi rst pre melt was performed at 20 C min from 80 C to 560 C and discarded for the analysis The samples are kept isothermally for 2min then cooled down to 300 C at the same rate followed by another isothermal segment Table 1 Candidate salt properties for high temperature MST solar salt compared to the salt mixture presented in this study FluidComposition wt Cp J g K Tm C Td C Energetic Cost kWht Refs Solar salt NaNO3 KNO36040 1 52406008 8 7 Ternary carbonate Li2CO3 Na2CO3 K2CO33233351 739766226 1 8 10 S Fereres et al Solar Energy Materials and Solar Cells 188 2018 119 126 120 Finally a heating ramp at 5 C min was performed from 300 C to 560 C The melting onset temperatures were estimated by the tangent at the point of largest slope on the DSC curve The latent heats of phase change were determined by numerical integration of the area under the peaks Modulated DSC is used for the specifi c heat capacity measurement The following thermal cycle is applied to each sample they are kept isothermally at 80 C for 10min then they are heated up from 80 C to 560 C at 20 C min to pre melt the sample Then the sample is cooled down to 150 C at the same heating rate At this point the modulation starts with the temperature amplitude of 0 5 C every minute at a heating rate of 5 C min up to 560 C recommended by the equipment manufacturer Two minutes isothermal segments are added before each dynamic segment Nitrogen is used as inert gas during the thermal program 50 ml min 2 2 TGA description Thermogravimetric analysis was performed to measure the LiNaK carbonate thermal decomposition under diff erent cover gases using a commercial TGA DSC2 equipment from Mettler Toledo The calibra tion procedure was performed by using the melting temperatures of standard certifi ed reference materials CRMs In Al Au at a heating rate of 10 C min resulting within the limits specifi ed by the equipment manufacturer Samples of 6 8mg were introduced in 70 l poly crystalline alumina oxide PCA crucibles The cover gas corresponding to each test was continuously fed around the pan at a volumetric fl ow rate of 150ml min Dynamic tests were performed with an initial 15 min isothermal segment at 120 C to evaporate any possible moisture from the sample followed by a 10 C min heating ramp from 120 C to 1000 C For the isothermal testing the samples were heated at 20 C min from 40 C to isothermal test temperature 650 C 700 C and 750 C and kept at the test temperature during 5h 2 3 X ray diff raction X ray diff raction patterns were obtained with a Rigaku MiniFlex600 system that has a D teX Ultra 1D silicon strip detector Both original LiNaK carbonate mixture and salt residues collected in the cold parts of the TGA instrument were analyzed 2 4 Custom made salt reactor A laboratory scale reactor Fig 1 was fabricated to measure the thermal stability of larger mass samples compared to those tested in commercial TGA equipment In these experiments about 3g of sample were placed inside a sealed alumina reactor After fi lling the reactor it was evacuated with a vacuum pump and subsequently fi lled with a 1bar pressure of the desired gas The reactor had a check valve with a cracking pressure of 0 005 MPa that guarantees a constant pressure inside the reactor even if there is a change in temperature or if some gases are evolved due to decomposition The alumina reactor was placed inside a tubular furnace and the temperature was maintained constant for diff erent periods of time while the evolution of the sample was monitored by gravimetric measurements As with commercial TGA equipment the reactor has two sources of experimental errors tem perature and mass measurements The temperature measurement error is considered to be that of the thermocouple itself 1 C The bal ance employed has an error of 1mg Tests repeatability was found acceptable with deviation in mass measurements below 4 3 Results and discussion 3 1 Melting point and latent heat DSC results were performed to confi rm the LiNaK carbonate salt melting point latent heat of fusion and specifi c heat capacity pre viously reported in the literature The results average of at least three tests and comparison with literature data are shown in Table 2 Although onset temperatures are relatively invariant the heating rates will aff ect the peak temperatures and latent heat values measured 19 For very precise measurements e g phase diagram property characterization several tests should be performed at low heating rates and then a melting point value is extrapolated for 0 C min 19 20 Here a heating rate of 5 C min was considered a good compromise between the standard engineering practice of using 10 C min and minimizing the test duration Thus peak temperatures increase with faster heating rates explaining the results shown in Table 2 Janz et al s 10 handbook data should be close to the theoretical value and the measurements presented here should be in between Janz et al and those measured at 10 C min by Olivares et al 9 3 2 TGA experiments 3 2 1 Dynamic tests at 10 C min Small samples of LiNaK ternary carbonate salt mixtures were heated under diff erent cover gases air 100 CO2 and the 80 20 CO2 air mixture to evaluate their thermal stability by traditional TGA The samples were preheated at 120 C for 15 min to evaporate any possible moisture and then heated up until 1000 C at a heating rate of 10 C min Sample curves for air and pure CO2can be seen in Fig 2 Table 3 shows the results for the decomposition temperature at which the sample mass changed 1 and 2 Td 1 Td 2 and the percent total mass loss at 1000 C as an average of three diff erent tests Olivares et al 9 reported the temperatures after which rapid weight loss was observed onset of decomposition in the TGA experiments corre sponding to Td 673 C in air and Td 788 C in pure CO2 showing signifi cantly higher stability in CO2atmosphere than air like the results presented here Table 3 These values for the onset of decomposition are in between the values for 1 and 2 mass change determined here The total measured mass loss at 1000 C is slightly higher but also following the same trend as the study by Olivares et al 9 where 6 7 mass was lost in air compared to 10 here in Tables 3 and 1 4 in CO2 compared to 2 6 here in Table 3 The decomposition of carbonates is associated to the following chemistry X COX OCO 2322 1 where X is either component Li Na or K A CO2cover gas shifts the equilibrium and slows down the de composition rate as shown