Thermal and Stability Investigation of Phase Change Material Dispersions for Thermal Energy Storage by T-History and Optical Methods

Article Thermal and Stability Investigation of Phase Change Material Dispersions for Thermal Energy Storage by T-History and Optical Methods

Maria Gabriela De Paola 1, Natale Arcuri 1, Vincenza Calabrò 2,* and Marilena De Simone 1 1 Department of Mechanical, Energy and Management Engineering (DIMEG), University of Calabria, Via P. Bucci 46/C, 87036 Rende, Italy; [email protected] (M.G.D.P.); [email protected] (N.A.); [email protected] (M.D.S.) 2 Department of Informatics, Modelling, Electronics and System Engineering (DIMES), University of Calabria, Via P. Bucci 39/C, 87036 Rende, Italy * Correspondence: [email protected]; Tel.: +39-0984496703

Academic Editor: Francesco Asdrubali Received: 29 September 2016; Accepted: 2 March 2017; Published: 13 March 2017

Abstract: Glauber’s salt (sodium sulphate decahydrate) is a promising phase change material (PCM) for use in the building sector, thanks to its high enthalpy of fusion associated with a proper phase transition temperature. It also offers economic and environmental advantages because it can be obtained as a byproduct from the disposal process of lead batteries. However, due to phenomena of phase segregation and supercooling, Glauber’s salt cannot be used in its pure state and requires the addition of rheological modiﬁers and nucleating agents. In this work, the initial thermal performances of mixtures based on Glauber’s salt with different compositions are compared by using the T-history method and adopting sonication for mixing, and following the same preparation procedure for all the samples. With ﬁxed composition, the effects of the addition sequence of the reagents are also examined. The analysis carried out by optical methods based on light scattering (Turbiscan equipment) allowed us to identify the kinetics of destabilization for each sample and revealed the need to specify in detail the preparation stages of PCMs, in order to make the composition reproducible in the laboratory and on a wider scale.

Keywords: thermal energy storage (TES); phase change material (PCM); Glauber’s salt; T-history; stability

1. Introduction Phase change materials (PCMs) are suitable products for thermal management solutions, because they store and release thermal energy during the process of melting and freezing. They are also very effective in providing thermal barriers or insulation, for example in temperature controlled transport. PCM dispersions based on salts can be very interesting for applications in the building sector, thanks to the high latent heat of fusion and the temperature at which this transition occurs. In such applications, they are enclosed in boards for the thermoregulation of internal environments. Panels can cover walls, ﬂoors, or ceilings, therefore large quantities of such materials are required. This need imposes limits in the choice of materials with regards to economic aspects, making it useful to address the recycling of waste materials, for example the Glauber’s salt that is obtained as a byproduct from the disposal process of lead batteries. The recovery is carried out by the crystallization of sulphate, which also allows for the elimination of traces of heavy metals, toxic to the environment and damaging to health.

Energies 2017, 10, 354; doi:10.3390/en10030354 www.mdpi.com/journal/energies Energies 2017, 10, 354 2 of 15

In particular, in the phase of the pastel desulfurization in which the lead sulphate is treated by using the carbonate or hydroxide of sodium, the following reactions take place [1,2]:

PbSO4 + Na2CO3 → PbCO3 + Na2SO4 (1)

PbSO4 + 4NaOH → Pb(OH)2 + Na2SO4 (2) Due to its thermal properties, the Glauber’s salt has been one of the first salt hydrates to be considered suitable for applications in the building sector [3]. Glauber salt’s transition is a peritectic phase change where the solid Glauber salt produces a liquid phase consisting of a sodium sulphate solution and a solid phase consisting of Na2SO4 [4]. In reference [5], data for the pure Glauber’s salt are available: the enthalpy of fusion is equal to 251.2 kJ/kg, the phase transition temperature is 32.4 ◦C, and the specific heat of the solid and liquid phase are equal to 2.09 kJ/(kg·K) and 3.35 kJ/(kg·K), respectively. However, during initial investigations, problems of stability have been highlighted and, as a consequence, difficulties in long term performances. These problems were mainly due to two phenomena common to many other hydrated salts: supercooling and phase segregation. The first prevents the release of heat at a defined temperature, while the second, due to the incongruent melting, limits the stability of the structure which is manifested by a progressive reduction of the enthalpy of fusion. The solutions proposed in the literature involve adding a nucleating substance [6] and a rheological modifier [7], in addition to an excess of water whose function is to counter the low solubility of the salt [8]. When additives are mixed with pure PCMs, a heterogeneous mixture occurs. The chemical, physical, and thermal properties are completely different with respect to the pure compounds, and they have to be evaluated experimentally for such heterogeneous mixtures. The study presented in this paper is based on the characterization and the thermal analysis of two compositions with different percentages of Glauber’s salt, obtained by using bentonite as a thickener and borax as a nucleating agent. This paper presents results related to the compositional weight percentage of sodium sulphate of 25% and 31%, and of 67% and 60% of water chosen after a selection of different compositions, previously tested with the aim of achieving interesting PCM thermal properties. The objective is to analyze their effect on the enlargement of the solidification interval, on the lowering of the solidification temperature, and on the fusion enthalpy and stability in comparison to the pure salt. Since the samples are dispersions, it was necessary to identify the most appropriate type of mixing, opting for sonication as the more effective method as opposed to the magnetic stirring technique in reducing the size of the aggregates [9]. Two samples were prepared for each composition by varying the sequence of addition of the reagents. Each sample was analyzed successively by means of the T-history method for the determination of their thermal properties, and by using optical techniques to obtain information on their stability. The purpose of the paper was to introduce a methodology able to predict the stability of Glauber’s salt based PCMs in a relatively short time with a significant amount of samples, in order to guarantee that any inhomogeneity will be taken into account. By such a method, it is possible determine how both compositions and methods of preparation affect the properties of PCMs. The results have also been used in order to obtain further information about the thermal performance of the studied PCMs.

2. Materials and Methods

2.1. Reagents and Preparation of the Samples For the preparation of the samples the following reagents were utilized: distilled water, sodium sulfate 99.5% purity, borax 99.5% purity, potassium chloride 99% purity, purchased from Carlo Erba (Milan, Italy), and bentonite, provided by Sigma-Aldrich Italia (Milan, Italy). Table1 shows the two analyzed compositional weight percentages. Energies 2017, 10, 354 3 of 15

Table 1. Compositional weight percentage of the tested mixtures.

Composition Water (%) Bentonite (%) Sodium Sulphate (%) Borax (%) 1 67 4 25 4 2 60 5 31 4

The compositional weight of bentonite and borax were selected thanks to preliminary experimental activity, carried out with the aim of investigating the optimal range of composition of the additives able to guarantee that the phase change occurs in the temperature range of interest for applications in building ﬁelds. As a consequence, the ratios between sodium sulphate and water are equal to 0.37 for sample 1 and 0.52 for sample 2. The samples have been made using a sonicator UP200S provided by Hielscher (Teltow, Germany); sonotrode S14 with a diameter of 14 mm), proceeding with 0.8 cycles of impulse followed by 0.2 s of pause at 100% of the length amplitude. Each composition (in total 100 g) has been obtained by the application of two different modalities: Modality (1): water and bentonite were sonicated for 10 min. After the addition of salts, the dispersion sonicated for a further 30 min. Modality (2): water and salts are sonicated for 30 min, and in a second step the bentonite was added and sonication was applied for 10 min. The samples will be identiﬁed in the paper following the nomenclature shown in Table2.

Table 2. Identiﬁcation of the four samples in relation to the composition and modality of preparation.

Sample Composition Modality of Preparation C1(1) 1 1 C1(2) 1 2 C2(1) 2 1 C2(2) 2 2

During sonication an increase of temperature to 70 ◦C was observed in all modalities. For this reason, all the samples were cooled to 40 ◦C in the thermostable bath, in order to apply the T-history method and optical analysis. It was possible to observe that all the samples appeared solid at temperatures lower than 19 ◦C and liquid for temperature higher than 31 ◦C. Sample preparation, as well as thermal and stability analyses, were replicated more than two times in order to guarantee a good reproducibility of the results. The data scattering was calculated for the thermal analysis and the optical stability analysis in terms of mean percentage error between the sets of data.

2.2. Thermal Analysis: T-History Method The thermal analysis was carried out by using the T-history method [10] in the temperature range from 10 ◦C to 40 ◦C. The original setup has been improved by using a refrigerated thermostat FOC 215 l from Velp Scientiﬁca, (Usmate Velate (Monza Brianza), Italy)), as also suggested by Stankovic et al. [11], coupled in our case with the Software TEMPSoft™ supplied from Velp Scientiﬁca, (Usmate Velate (Monza Brianza), Italy)) to program the heating and cooling ramps continuously, in order to obtain an adequate control of the environmental conditions, following the assigned function. Duran® test tubes (height of 16 cm and diameter of 1 cm) were used to store the PCM and reference material during the T-history analysis. Water was used as reference substance, because it does not show phase change in the investigated temperature range. The distance between test tubes was large enough to avoid any interference during heat transport. Energies 2017, 10, 354 4 of 15

The T-history method was applied by decreasing the temperature from 40 ◦C to 10 ◦C with a step of 1◦/min, then the samples stayed at 10 ◦C for two hours, and after they were heated to 40 ◦C (step of 1◦/min). The cycle was repeated four times. Temperatures inside the PCM test tube as well as inside the reference (water) test tubes were continuously measured with resistance temperature detectors (RTDs) PT100, a miniature sensor RTD with high accuracy, with a diameter of 0.5 mm and length of 100 mm (able to guarantee fast and reproducible results). Two tests were performed for each sample. PCM mixture samples in the liquid phase, after sonication and cooled to 40 ◦C, were immediately placed in the refrigerated thermostat and RTD probes were inserted in the test tubes. Simultaneously, test tubes with reference substances were heated at 40 ◦C and the RTD probes were lodged in the refrigerated thermostat that will represent the environment at which the temperature is called the ambient temperature. Temperature profiles during the heating and cooling cycles were continuously measured, following the method proposed from Zhang and Jiang [10] and T-history curves were prepared. As reported in the literatures [10–12], cooling curves were used for the evaluation of ∆H. The results are therefore presented with reference to the method of Marin et al. [12], in which in addition to the enthalpy of fusion, the representation of the enthalpy-temperature trend is shown in the examined range. Such representation takes into account that PCMs do not have homogeneous compositions and, therefore, they have a rather wide solidification range. The enthalpy–temperature curve has been obtained by considering the changes in enthalpy in small temperature ranges according to Equation (3): mw·cpw(Ti) + mt·cpt(Ti) Ii mt ∆hP(Ti) = · ·∆Ti − ·cpt(Ti)·∆Ti (3) mP Ii’ mP where mw, mt, and mp are the mass of water, test tube, and PCM, respectively (kg); cpw and cpt are the specific heat of water and the test material (kJ/(kg·K)); for the ith step, ∆Ti is the temperature ◦ range ( C) with average value Ti, to which the enthalpy variation ∆hp(Ti) (kJ/kg) corresponds; Ii is the integral with respect to time of the difference in temperature between the PCM and environment; Ii’ is the integral with respect to time of the difference in temperature between the water and environment. The latent heat of solidification was obtained as [∆hp(TiS) − ∆hp(TfS)] where TiS is the initial temperature of solidification calculated as the maximum value reached after sub-cooling and TfS is the final temperature, calculated according to as a point of inflection on the cooling curve of PCM [13]. The specific heat of PCM in the liquid (cpl) and solid state (cps) was calculated in reference to the article of Marinas, where the gradients of the two straight lines are observable in the enthalpy versus temperature trend [12,13].

2.3. Kinetic of Destabilization PCMs made of Glauber’s salt and additives are dispersions. For this reason, they are subject to destabilization phenomena during the time that could affect their thermal properties. Generally, these phenomena concern particle migration that determines sedimentation and creaming, or variation of their dimensions (flocculation and coalescence). Destabilization could affect the thermal performance of PCMs, and consequently it is important to predict these phenomena from the beginning. When sedimentation, clarification, or flocculation occur, the texture of the materials is changed, and as a consequence the thermal conductivity, thermal capacity, heat of fusion or solidification, and density change: the system could be compared to a side or series resistance system, in terms of heat exchange. Fusion or solidification phenomena occur in different ways in the material and that could affect the effectiveness of the thermal behaviour of the PCM. Destabilization phenomena are not visible to the human eye when they are characterized by low kinetics, and only some sophisticated optical methods and instruments, such as the Turbiscan, are able to detect them from the initial stages [14–16]. The operating principle is based on multiple light scattering. The instrument scans samples with a volume of about 25 mL and draws transmission (T) and backscattering (BS) profiles along the height EnergiesEnergies 2017, 201710, 354, 10 , 354 5 of 145 of 14 EnergiesEnergies 2017 2017, 10,, 35410, 354 5 of 514 of 14 EnergiesEnergies 2017 2017, 10, ,354 10, 354 5 of 514 of 14 EnergiesThe2017 Theoperating, 10 ,operating 354 principle principle is based is based on multipleon multiple light light scattering. scattering. The Theinstrument instrument scans scans samples5 samples of 15 TheThe operating operating principle principle is basedis based on onmultiple multiple light light scattering. scattering. The The instrument instrument scans scans samples samples withwith Thea volumeThe aoperating volume operating of about of principle about principle25 mL 25 isandmL basedis anddraws based drawson transmissiononmultiple transmissionmultiple light (lightT )scattering. and (T scattering.) andbackscattering backscattering The The instrument instrument (BS) ( profilesBS )scans profiles scans along samples alongsamples the the withwith a volume a volume of aboutof about 25 mL25 mL and and draws draws transmission transmission (T) (andT) and backscattering backscattering (BS ()BS profiles) profiles along along the the heightwithwithheight a volumeof a thevolume of testthe of abouttest cell.of about cell.The 25 ThemLscan25 mL andscan can and draws canbe draws repeated be transmission repeated transmission for afor long ( Ta) (longand Ttime,) and backscatteringtime, and backscattering andat fixed at fixed (intervalsBS () BS intervalsprofiles) profiles of time.along of alongtime. The the theThe heightof theheight testof theof cell. the test The test cell. scan cell. The can The scan be scan repeatedcan can be berepeated for repeated a long for time, fora long a andlong time, at time, fixed and and intervalsat fixedat fixed intervals of time.intervals The of time.of more time. The the The moreheightheightmore the of profiles theof theprofilestest test arecell. cell.distinguishedare The distinguishedThe scan scan can fromcan be repeatedfrombe the repeated initialthe initialfor one, fora long one,thea long moretime,the time, more unstableand and unstableat fixedat the fixed dispersion.intervalsthe intervals dispersion. of time.of time. The The profilesmoremore the arethe profiles distinguished profiles are are distinguished distinguished from the initialfrom from the one, the initial the initial more one, one, unstablethe the more more the unstable dispersion.unstable the the dispersion. dispersion. moremoreIn the the In theprofiles casethe profiles case of are opaque of aredistinguished opaque distinguished dispersions, dispersions, from from as the those as theinitial those initialconsidered one, considered one, the the morein morethis in unstable thisinvestigation, unstable investigation, the thedispersion. dispersion.the backscatteringthe backscattering In theIn the case case of opaqueopaqueof opaque dispersions,dispersions, dispersions, as thosethoseas those consideredconsidered considered in thisthisin this investigation,investigation, investigation, the the backscatteringbackscattering backscattering profileprofileIn istheIn more theiscase more caseeffective of opaqueeffectiveof opaque than dispersions, than the dispersions, transmissionthe transmission as thoseas thoseprofile considered profile considered that thatis null.in is this innull. this investigation, investigation, the thebackscattering backscattering profileprofileprofile is more is more effectiveeffective effective thanthan than thethe the transmissiontransmission transmission profileprofile profile thatthat that isis null.null. is null. profileprofileThe is The BSmore is percentagemoreBS effective percentage effective thanalong than along the the thetransmission testthe transmission testcell cellprovides profile provides profile information that thatinformation is null. is null. regarding regarding the homogeneitythe homogeneity of the of the TheThe BS BSpercentage percentage along along the the test test cell cell provides provides information information regarding regarding the the homogeneity homogeneity of theof the sample,sample,The Thebut BS inbutBSpercentage order percentagein order to easily along to alongeasily identify the theidentifytest test thecell cellphenom theprovides phenomprovidesena information enain information the in samples,the samples,regarding regarding it is itrecommended theis recommended thehomogeneity homogeneity to use ofto usetheof thethe sample,sample, but but but in in orderin order order to to easilyto easily easily identify identify identify the the the phenom phenomena phenomenaena in inthein thethe samples, samples,samples, it is it recommended is recommended recommended to useto to use usethe the Deltasample,sample,Delta mode. but mode. but inOne orderin One of order the toof easilyprofilestheto easily profiles identify (by identify default,(by the default, thephenom the phenom firsttheena firstone)ena in one) theisin used thesamples, is usedsamples, as a as “blank”it ais it “blank”recommended is recommendedcorresponding corresponding to use toto use the to the the DeltatheDelta Delta mode. mode. mode. One One Oneof theof ofthe profiles the profiles profiles (by (by default, (by default, default, the the first the first one) first one) is one) used is used is as used aas “blank” a as “blank” a “blank” corresponding corresponding corresponding to theto the initialDeltaDeltainitial mode.dispersion mode. dispersion One One ofstate theof state of theprofiles the profilesof sample.the (by sample. (by default, This default, This profile the theprofilefirst isfirst one)then is one) thenis subtracted used is subtractedused as aas “blank”from a “blank” from the corresponding otherthe corresponding other profiles, profiles, to thus theto thus the initialto theinitial initialdispersion dispersion dispersion state state of state theof the ofsample. thesample. sample. This This profile This profile profile is thenis then is subtracted then subtracted subtracted from from the from the other the other otherprofiles, profiles, profiles, thus thus emphasizinginitialinitialemphasizing dispersion dispersion the variations thestate variationsstate of theof [17]. thesample. [17]. sample. This This profile profile is thenis then subtracted subtracted from from the theother other profiles, profiles, thus thus emphasizingthusemphasizing emphasizing the the variations the variations variations [17]. [17]. [17 ]. emphasizingemphasizingIn TableIn Table the 3, thevariationsexamples 3, variations examples [17].of [17]. the of theresults results in termsin terms of deltaof delta backscattering backscattering (∆BS ()∆ BSobtained) obtained with with In TableInTable Table3 ,3, examples examples3, examples of theof resultsofthe the results inresults terms in ofinterms deltaterms of backscattering ofdelta delta backscattering backscattering ( ∆BS) obtained (∆ BS(∆)BS withobtained) obtained Turbiscan with with TurbiscanTurbiscanIn InTable areTable presentedare3, examples3,presented examples with ofwith reference theof reference theresults resultsto the into mostthetermsin mostterms important of important deltaof delta destabilizationbackscattering destabilizationbackscattering phenomena (∆ BSphenomena(∆)BS obtained) obtained that that occurwith occurwith TurbiscanareTurbiscan presented are are presented with presented reference with with reference to reference the most to theto important the most most important important destabilization destabilization destabilization phenomena phenomena phenomena that occur that that occur with occur withTurbiscanTurbiscanwith inhomogeneous inhomogeneous are arepresented presented PCMs. with PCMs. with reference reference to the to themost most important important destabilization destabilization phenomena phenomena that that occur occur withinhomogeneouswith inhomogeneous inhomogeneous PCMs. PCMs. PCMs. withwith inhomogeneous inhomogeneous PCMs. PCMs. TableTable 3. Destabilization 3. Destabilization phenomena phenomena occurring occurring in inhomogeneous in inhomogeneous dispersions dispersions and theand ∆ theBS ∆analysisBS analysis TableTable 3.3. DestabilizationDestabilization 3. Destabilization phenomenaphenomena phenomena occurring occurring in inhomogeneous inhomogeneousin inhomogeneous dispersions dispersions and and the the∆BS ∆ analysisBS analysis carriedTableTablecarried 3. out Destabilization 3. by outDestabilization Turbiscan by Turbiscan phenomenainstrument phenomena instrument [17]. occurring occurring[17].BS: backscattering. BS in: backscattering.inhomogeneous in inhomogeneous dispersions dispersions and and the the∆BS ∆ analysisBS analysis carriedcarried out outby Turbiscan by Turbiscan instrumentinstrument instrument [[17].17]. [17]. BS:: BS backscattering.backscattering.: backscattering. carriedcarried out outby Turbiscanby Turbiscan instrument instrumentParticle [17]. [17].Migration BS: BSbackscattering.: backscattering. Particle Migration Δ Δ DestabilizationDestabilization Phenomena Phenomena ParticleParticle Migration Migration BS BS DestabilizationDestabilization Phenomena Phenomena or ParticleChangeor Change Migrationorin Size in Size ΔBSΔ BS Destabilization Phenomena orParticle ChangeParticle Migration in Migration Size ∆ΔBSΔ DestabilizationDestabilization Phenomena Phenomena orChange Change in in Size Size BS BS The densityThe density of the of dispersed the dispersed or Changeor Change in Size in Size ThephaseThe density is density greater of the of than thedispersed thedispersed TheThe densityphase density is of greater the of thedispersed than dispersed the phasedensityphase is ofgreater is the greater continuous than than the the phaseThephasedensity densityis greater is greaterof of the thethan continuous dispersedthan the the phase densityphasedensity so of the the of volume thecontinuous continuous fraction densityisdensity greaterphase of theso thanof thethecontinuous the volumecontinuous density fraction of the phaseΦ continuouswillphase sodecrease the so thevolume phase atvolume the so fraction top the fraction of volume phasephaseΦ sowill the so decrease thevolume volume at fraction the fraction top of Φthe fractionwill Φcell,the will decrease andcell, decreaseΦ willincreaseand at decreaseincrease the at at thetop the topof at the of top Φ willΦ will decrease decrease at the at thetop topof of thebottom.of thecell, the cell, cell,and and andincrease increase increase at the at at the the thecell,bottom. cell, and and increase increase at the at the bottom.thebottom. bottom. bottom.bottom.

The dispersedThe dispersed phase phase has a has a ThelowerThe dispersedlower density dispersed density thanphase phase thanthe has thehas a a lowerTheTheThe dispersed disperseddensity dispersed than phase phase phase the has has hasa a lowera continuouslowercontinuous density phase phase than so the theso the lowerdensitylower density thandensity than the than continuous the the phase continuousvolumecontinuousvolume fraction phase fraction phase Φ so will the Φso will the continuoussocontinuous the volume phase phase fraction so the so Φthewill volumeincreasevolume fraction at thefraction top Φ willof Φ the will cell, volumeincreasevolumeincrease fraction at fraction theat the topΦ willtop Φ of will theof the cell, cell, increaseandincrease decrease at the at at thetop the topof bottom. the of thecell, cell, increaseandincreaseand decrease atdecrease the at thetop at the at topof the bottom.the of bottom. thecell, cell, andand decrease decrease at the at thebottom. bottom. andand decrease decrease at the at thebottom. bottom. This Thisphenomenon phenomenon will affectwill affect Thisthe Thisaverage phenomenon phenomenon diameter will ofwill affect the affect ThisThisThis thephenomenon phenomenon averagephenomenon diameter will will will affect affect of affect the the theparticles theaverage average in thediameter sample.diameter of It the of will the theaverage theaverageparticles average diameter diameter in diameterthe ofsample. of the the of particles theIt will particlesincreaseparticles overin the in the thesample. whole sample. It will It will particlesinparticles theincrease sample. in the inover thesample. It willthesample. whole increase It will It will over increaseheightincrease of over the over cellthe theaswhole whole increasetheincreaseheight whole over of heightover the thewholecell of whole theas cell as heightcoalescencecoalescenceheight of the of or thecell or flocculation flocculationcellas as are heightheightcoalescence of the of thecell or cellas flocculation as coalescenceareglobal coalescenceglobal phenomena, phenomena, or flocculation or flocculation taking place coalescencecoalescenceare global or flocculationphenomena, or flocculation aretakingeverywhere areglobaltaking placeglobal phenomena, place everywhere phenomena, ina everywhere sample. in a in a takingare areglobal place global phenomena, everywhere phenomena, in a sample.takingsample. place everywhere in a takingtaking place place everywhere everywhere in a in a sample.sample. sample.sample. By meansmeansBy means ofof comparing comparing of comparing the the profiles theprofiles profiles in time, in time, thein time, instrument the theinstrument instrument determines determines determines the destabilization the thedestabilization destabilization kinetics By Bymeans means of comparingof comparing the the profiles profiles in time,in time, the the instrument instrument determines determines the the destabilization destabilization basedkineticskineticsBy on Bybasedmeans the meansbased computation on of the oncomparingof computationthecomparing computation of thethe Turbiscan the profilesof theprofiles of Turbiscanthe in Stability Turbiscantime,in time, Stabilitythe Index the Stabilityinstrument (instrumentTSI Index). Index The (TSIdeterminesTSI determines().TSI Theis). a The computationTSI the isTSI the adestabilization iscomputation destabilization a computation directly kineticskinetics based based on onthe the computation computation of theof the Turbiscan Turbiscan Stability Stability Index Index (TSI (TSI). The). The TSI TSI is ais computation a computation baseddirectlykineticskineticsdirectly on basedbased the basedbased raw onon the dataontheon therawthecomputation of computation therawdata transmissiondata of the of transmissionthe of thetransmissionTurbiscan and Turbiscan backscattering and Stability andbackscatteringStability backscattering Index signals. Index (TSI signal( TheTSI). Thesignal). superpositions. The TSIThes. TSI isThesuperposition a is computationsuperposition a computation of the BSof of directlydirectly based based on onthe the raw raw data data of theof the transmission transmission and and backscattering backscattering signal signals. Thes. The superposition superposition of of profilesthedirectly directlyBSthe profiles BS isbased indicativeprofiles based ison indicative theon is indicative theofraw stability raw data of datastability of while stability theof thetransmission while their transmission while their deviation their deviation and deviation and acrossbackscattering backscattering across time across time increases timesignal increases signal increasess. with Thes. with The thesuperposition with thesuperposition instability instabilitythe instability of of thethe BS BSprofiles profiles is indicative is indicative of stabilityof stability while while their their deviation deviation across across time time increases increases with with the the instability instability theofthe the theBS dispersions.of dispersions. theprofilesBS profilesdispersions. is Theindicative is The indicativeTSI TheTSIcan canTSIof bestability of becan determinedstability determined be determinedwhile while their by by their using usingdeviation by deviation usingthe the comparison comparisonacrossthe acrosscomparison time time ofincreasesof thetheincreases of scansthe with scans atwith thea atgiven theinstability a given instability time. time. time. of theof the dispersions. dispersions. The The TSI TSI can can be determinedbe determined by byusing using the the comparison comparison of theof the scans scans at aat given a given time. time. Inof particular,theofIn the dispersions.particular, dispersions. thethe TSI TSIthe The isisTSI The determineddeterminedTSI is TSI candetermined can be determinedbe forfor determined thethe for i iththeth stepstep byith usingby step [[17]:17 using]: [17]: the thecomparison comparison of the of thescans scans at a at given a given time. time. In particular,In particular, the the TSI TSI is determined is determined for forthe the ith istepth step [17]: [17]: In particular,In particular, the theTSI TSI is determined is determined for forthe theith stepith step [17]: [17]: (1) (1) by comparing by comparing every every scan iscan(h) ofi(h )a ofmeasurement a measurement to the to previousthe previous one onescan iscan−1(h),i− 1on(h), theon selectedthe selected (1) (1) by bycomparing comparing every every scan scani(h)i (ofh) aof measurement a measurement to theto the previous previous one one scan scani−1(hi−),1( hon), onthe the selected selected (1) (1)height by bycomparingheight comparingh; h; every every scan scani(h) i(ofh) aof measurement a measurement to theto theprevious previous one one scan scani−1(hi),−1 (hon), onthe theselected selected heightheight h; h; heightheight h; h;

Energies 2017, 10, 354 6 of 15

(1) by comparing every scani(h) of a measurement to the previous one scani−1(h), on the selected height h; Energies 2017, 10, 354 6 of 14 (2) by dividing the result by the total selected height (H) in order to obtain a result which does not depend on the quantity of product in the measured cell: (2) by dividing the result by the total selected height (H) in order to obtain a result which does not depend on the quantity of product in the measured cell: ∑ [scan (h) − scan − (h)] TSI = ∑ h i i 1 (4) i H ∑ (ℎ) −(ℎ) = (4) The TSI computation sums up all the variations in the sample, to give as a result a unique number reﬂectingThe TSI the destabilizationcomputation sums occurring up all in athe sample. variations The higherin the thesample,TSI, the to strongergive as thea result destabilization a unique innumber the sample. reflecting Consequently, the destabilization an asymptotic occurring value in of aTSI sample.could meanThe higher two things: the TSI, the stronger the destabilization in the sample. Consequently, an asymptotic value of TSI could mean two things: TSI (1)(1) TheThe material is stable: in in this case, TSI has a constant value from the initial time and during the entireentire analysis. Usually, Usually, this happenshappens for homogeneous materials. (2)(2) TheThe material material is unstable at the beginning and becomes stable after a long time (more than 24 h): inin thisthis case,case, the the asymptotic asymptotic value value could could be observedbe observed after after a long a timelong becausetime because the original the original sample samplechanged changed its texture. its texture. As a consequence, As a consequence, the material the material probably probably became became stable stable but with but awith very a verydifferent different texture, texture, with differentwith different overall overall thermal thermal properties, properties, like conductivity, like conductivity, thermal capacity,thermal capacity,density, or density, other physical or other properties: physical properties: It became It a differentbecame a material different with material respect with to therespect initial to onethe initialand its one thermal and its behavior thermal (i.e., behavior as a PCM) (i.e., as needs a PCM) to be needs recalculated. to be recalculated. In thethe investigationinvestigation proposedproposed in in this this paper, paper, the the analyses analyses were were conducted conducted considering considering a perioda period of 24of h24 and h and the the temperature temperature was was ﬁxed fixed at the at maximum the maximum value value in the in observation the observation range (40range◦C) (40 in which°C) in thewhich sample the sample remains remains in the liquid in the phase.liquid phase.

3. Results and Discussion By preparingpreparing the the sample sample following following modalities modalities 1 or 1 2,or it 2, is possibleit is possible to observe to observe signiﬁcant significant and visible and differencesvisible differences in their in textures. their textures. In particular, In particul the ﬁrstar, samplethe first (both sample for C1(both and for C2) C1 has and an appearanceC2) has an moreappearance viscous more than viscous the second. than the Figure second.1 shows Figure the 1 comparison shows the comparison between the between compositions the compositions C1(1) and C1(2).C1(1) and C1(2).

Figure 1. Sample with composition C1(1) in comparison to the sample with composition C1(2). Figure 1. Sample with composition C1(1) in comparison to the sample with composition C1(2).

Sample C1(1) appears like a dough, with a rheological behavior intermediate between a liquid and aSample solid system. C1(1) appearsOn the contrary, like a dough, sample with C1(2) a rheological appears as behavior a liquid. intermediateTo the naked betweeneye, it appears a liquid to andhave a a solid more system. homogeneous On the contrary,texture than sample C1(1), C1(2) where appears the inhomogeneity as a liquid. To theis more naked evident. eye, it appears to have a more homogeneous texture than C1(1), where the inhomogeneity is more evident. 3.1. Thermal Analysis of the Samples In Figure 2, the results of the T-history method applied at each sample Ci(j), are reported. They depict the heating and cooling cycles obtained in the refrigerated thermostat (the environment) where the temperature was changed from 40° to 10 °C following the imposed function, as described in Section 2.2. In the figure, the temperature evolution of the refrigerated thermostat where the samples are lodged (ambient temperature) as well as the temperature evolution in the PCM and water (reference substance) are reported.

Energies 2017, 10, 354 7 of 15

3.1. Thermal Analysis of the Samples In Figure2, the results of the T-history method applied at each sample C i(j), are reported. They depict the heating and cooling cycles obtained in the refrigerated thermostat (the environment) where the temperature was changed from 40◦ to 10 ◦C following the imposed function, as described in Section 2.2. In the ﬁgure, the temperature evolution of the refrigerated thermostat where the samples are lodged (ambient temperature) as well as the temperature evolution in the PCM and water (referenceEnergies 2017 substance), 10, 354 are reported. 7 of 14

(a) (b)

Figure 2.2. HeatingHeating and and cooling cooling cycles cycles for for all theall the samples, samples, during during the T-history the T-history method. method.(a): Sample (a): Sample C1(1); (C1(1);b): sample (b): sample C1(2); (C1(2);c): sample (c): sample C2(1); C2(1); and (d and): sample (d): sample C2(2). C2(2).

It is possible to see the supercooling limited to 2–3 °C and the repeatability of the cycles. As It is possible to see the supercooling limited to 2–3 ◦C and the repeatability of the cycles. reported in the literatures [10–13], cooling curves have been used for the second part of the method, As reported in the literatures [10–13], cooling curves have been used for the second part of the then by considering the cooling curves and Equation (3), the enthalpy versus temperature curves method, then by considering the cooling curves and Equation (3), the enthalpy versus temperature were determined for all the samples. The results obtained for sample C1 are shown in Figure 3. curves were determined for all the samples. The results obtained for sample C1 are shown in Figure3. The thermal properties of the first sample, estimated by considering the method described in Section 2.2, are summarized in Table4. The first composition has a percentage of sodium sulphate equal to 25%. In both modalities of preparation, a decrement of the temperature range of solidification with respect to the value of the pure salt is observed. This phenomenon could be correlated to the presence of additives that also determines an extension of the solidification interval. Such range of temperature is greater for the first composition and marks a major degree of inhomogeneity. Furthermore, it need to be taken into account that for sample C1(1), the water and bentonite were first mixed. Thus, it can be predicted that part of the water will be trapped by the bentonite silicate layers, thus reducing the water available to mix with the sodium sulphate, affecting melting temperature range.

Figure 3. Sample C1. Enthalpy vs. temperature curves with variation of the modality of preparation.

The thermal properties of the first sample, estimated by considering the method described in Section 2.2, are summarized in Table 4. The first composition has a percentage of sodium sulphate equal to 25%. In both modalities of preparation, a decrement of the temperature range of solidification with respect to the value of the pure salt is observed. This phenomenon could be correlated to the presence of additives that also determines an extension of the solidification interval. Such range of temperature is greater for the first composition and marks a major degree of inhomogeneity. Furthermore, it need to be taken into account that for sample C1(1), the water and bentonite were first mixed. Thus, it can be predicted

Energies 2017, 10, 354 7 of 14

(a) (b)

Figure 2. Heating and cooling cycles for all the samples, during the T-history method. (a): Sample C1(1); (b): sample C1(2); (c): sample C2(1); and (d): sample C2(2).

It is possible to see the supercooling limited to 2–3 °C and the repeatability of the cycles. As reported in the literatures [10–13], cooling curves have been used for the second part of the method,

Energiesthen by2017 considering, 10, 354 the cooling curves and Equation (3), the enthalpy versus temperature curves8 of 15 were determined for all the samples. The results obtained for sample C1 are shown in Figure 3.

Energies 2017, 10, 354 8 of 14 that partFigure of 3. theSample water C1. willEnthalpy be trapped vs. temper temperature byature the curves bentonite with variation silicate layers,of the modality thus reducing of preparation. the water available to mix with the sodium sulphate, affecting melting temperature range. TheTable thermal 4. Thermal properties properties of ofthe the first first sample, sample C1esti preparedmated by by meansconsidering of the modalitiesthe method (1) anddescribed (2). in SectionTable 2.2, 4.are Thermal summarized properties in Tableof the firs4. t sample C1 prepared by means of the modalities (1) and (2). ◦ ◦ The first compositionSample T iShas( C)a percentageTfS ( C) of∆ sodiumH (kJ/kg) sulphatecps (kJ/kg equal·K) to 25%.cpl (kJ/kg In both·K) modalities of preparation, a decrementSample ofT iSthe (°C) temperature TfS (°C) ∆rangeH (kJ /ofkg) solidificationcps (kJ/kg·K) withcpl (krespectJ/kg·K) to the value of the C1(1)C1(1) 29.7529.75 19.2619.26 122.28 122.28 4.10 4.10 2.70 2.70 pure salt is observed.C1(2) This 25.63phenomenon 21.07 could be 87.99 correlated to 3.24 the presence 3.63of additives that also C1(2) 25.63 21.07 87.99 3.24 3.63 determines an extension of the solidification interval. Such range of temperature is greater for the first composition and marks a major degree of inhomogeneity. Furthermore, it need to be taken into The specificspecific heats,heats, forfor the the solid solid and and liquid liquid phase, phase, are are larger larger than than those those for for the the pure pure Glauber’s Glauber’s salt account that for sample C1(1), the water and bentonite were first mixed. Thus, it can be predicted reportedsalt reported in the in introductionthe introduction [5]. In[5]. fact In thefact excessthe excess of water of water is known is known to elevate to elevate the specificthe specific heat. heat. In Figure 44,, thethe enthalpiesenthalpies of fusion of the two two samples samples with with composition composition 2 2 are are compared compared and and a asignificant significant difference difference in in the the values values is is noticed. noticed.

Figure 4. Sample C2. Enthalpy vs. temper temperatureature curves with variation of the modality of preparation.

The thermal properties of the second sample are presented in Table 5. The thermal properties of the second sample are presented in Table5.

Table 5. Thermal properties of the second sample C2 prepared by means of the modalities (1) and (2). Table 5. Thermal properties of the second sample C2 prepared by means of the modalities (1) and (2).

Sample TiS (°C) TfS (°C) ∆H (kJ/kg) cps (kJ/kg·K) cpl (kJ/kg·K) ◦ ◦ SampleC2(1) TiS 28.06( C) TfS21.15( C) ∆H 72.65(kJ/kg) cps 4.64(kJ/kg ·K) c 2.43pl (kJ/kg ·K) C2(1)C2(2) 28.0630.97 21.1524.45 171.9 72.65 3.62 4.64 3.80 2.43 C2(2) 30.97 24.45 171.9 3.62 3.80 Both the C2 compositions contain a percentage by weight of sodium sulphate equal to 31%, however the sample obtained by modality 1, C2(1), presents a thermal performance less effective Both the C2 compositions contain a percentage by weight of sodium sulphate equal to 31%, than that of sample C2(2) and also than that of the C1 samples that contain a minor percentage of however the sample obtained by modality 1, C2(1), presents a thermal performance less effective than sodium sulphate (25%). This result is probably due to the high viscosity of C2(1) that reduces the that of sample C2(2) and also than that of the C1 samples that contain a minor percentage of sodium efficacy of the sonication process: the same sonication power is not able to effectively break up the anhydrous sodium sulphate, reducing the formation of the decahydrate form (Glauber’s salt) and therefore the final thermal performance of the sample. In all the analyzed compositions, the size of supercooling is 2–3 °C and it can be considered very reduced in comparison to the values higher than 15 °C indicated in the literature for the pure salt [18], revealing the efficacy of the adopted nucleant. Further consideration could be performed taking into account the phase diagram of the sodium sulphate–water system, as proposed by Biswas [19]. In order to apply such a diagram, the amount of Na2SO4 has been calculated and the corresponding theoretical melting temperature values have been estimated for each sample. It is important to notice that the phase diagram is relative to a system without additives, therefore the amount of sodium sulphate has been recalculated assuming a system without bentonite and borax. For sample 1 where the amount of Sodium sulphate is 25%, the corresponding value for the phase diagram has been estimated at 27.17%. That corresponds to a condition in which the

Energies 2017, 10, 354 9 of 15 sulphate (25%). This result is probably due to the high viscosity of C2(1) that reduces the efficacy of the sonication process: the same sonication power is not able to effectively break up the anhydrous sodium sulphate, reducing the formation of the decahydrate form (Glauber’s salt) and therefore the final thermal performance of the sample. In all the analyzed compositions, the size of supercooling is 2–3 ◦C and it can be considered very reduced in comparison to the values higher than 15 ◦C indicated in the literature for the pure salt [18], revealing the efficacy of the adopted nucleant. Further consideration could be performed taking into account the phase diagram of the sodium sulphate–water system, as proposed by Biswas [19]. In order to apply such a diagram, the amount of Na2SO4 has been calculated and the corresponding theoretical melting temperature values have been estimated for each sample. It is important to notice that the phase diagram is relative to a system without additives, therefore the amount of sodium sulphate has been recalculated assuming a system without bentonite and borax. For sample 1 where the amount of Sodium sulphate is 25%, the corresponding value for the phase diagram has been estimated at 27.17%. That corresponds to a condition in which the anhydrous sodium sulphate crystals are not expected in the studied range of temperature (10–40 ◦C). Furthermore, the theoretical transition temperature is 25 ◦C, and the value is within the experimental range reported in Table4 for both C1(1) and C1(2). For sample 2, the amount of charged sodium sulphate is 31% and, assuming a system without additives, the corresponding value is 34.07%. In the phase diagram, this value corresponds to a limit condition and, beyond this value, the formation of anhydrous sodium sulphate is expected. With this value in the phase diagram, the theoretical transition temperature could be expected to be 32 ◦C. This value is more similar to that obtained in Table5, because the additives play a more significant role during the activity of Glauber’s salt as a PCM. With the amount of sodium sulphate and water reported in Table1, it is possible to expect that the availability of water is greater for sample 1 than for sample 2. Bentonite and sodium sulphate compete for water. Sample 1 has enough water for both, and consequently, the theoretical melting point is in the observed range of temperature Tis and Tfs for both samples C1(1) and C1(2). On the other hand, for sample 2 where less water is available, the modality plays a more significant role. Modality 1 gives an immediate possibility for bentonite to hydrate itself with water: less water remains for the Glauber’s salt formation. Consequently, the temperature range where melting occurs is closer to a system where a smaller amount of hydrated salt is present. With modality 2, Glauber’s salt could be easily obtained during the first sonication. Residual water becomes available for the hydration of bentonite. As a consequence, thermal performance closer to Glauber’s salt could be reached, as confirmed for the temperature range shown in Table5. Finally, the reproducibility of the data was confirmed from the evaluation of the mean error percentage between the data sets of each sample. With reference to the data shown in Figures3 and4, mean errors range from 4.45% and 5.56%.

3.2. Stability Analysis of the Samples The BS profiles of the samples C1(1) and C1(2) are shown in Figure5. The composition C1(1) shows in each scan a percentage of the BS variable along the height of the cell, revealing the inhomogeneity of the sample. The composition C1(2) is rather homogeneus until 36 mm height; from this point, sedimentation and creaming phenomena become evident. Thus the sonication is more effective on the second dispersion that is less viscous. As a consequence, its solidification range (shown in Figure3) is more limited (the first one is 10.49 ◦C, the second one is limited to 4.56 ◦C) while the total ∆H of the process is comparable. ∆BS profiles are presented in Figure6 in order to underline the nature of these phenomena. Energies 2017, 10, 354 9 of 14

anhydrous sodium sulphate crystals are not expected in the studied range of temperature (10–40 °C). Furthermore, the theoretical transition temperature is 25 °C, and the value is within the experimental range reported in Table 4 for both C1(1) and C1(2). For sample 2, the amount of charged sodium sulphate is 31% and, assuming a system without additives, the corresponding value is 34.07%. In the phase diagram, this value corresponds to a limit condition and, beyond this value, the formation of anhydrous sodium sulphate is expected. With this value in the phase diagram, the theoretical transition temperature could be expected to be 32 °C. This value is more similar to that obtained in Table 5, because the additives play a more significant role during the activity of Glauber’s salt as a PCM. With the amount of sodium sulphate and water reported in Table 1, it is possible to expect that the availability of water is greater for sample 1 than for sample 2. Bentonite and sodium sulphate compete for water. Sample 1 has enough water for both, and consequently, the theoretical melting point is in the observed range of temperature Tis and Tfs for both samples C1(1) and C1(2). On the other hand, for sample 2 where less water is available, the modality plays a more significant role. Modality 1 gives an immediate possibility for bentonite to hydrate itself with water: less water remains for the Glauber’s salt formation. Consequently, the temperature range where melting occurs is closer to a system where a smaller amount of hydrated salt is present. With modality 2, Glauber’s salt could be easily obtained during the first sonication. Residual water becomes available for the hydration of bentonite. As a consequence, thermal performance closer to Glauber’s salt could be reached, as confirmed for the temperature range shown in Table 5. Finally, the reproducibility of the data was confirmed from the evaluation of the mean error percentage between the data sets of each sample. With reference to the data shown in Figures 3 and 4, mean errors range from 4.45% and 5.56%.

3.2. Stability Analysis of the Samples Energies 2017, 10, 354 10 of 15 The BS profiles of the samples C1(1) and C1(2) are shown in Figure 5.

(a)

Energies 2017, 10, 354 10 of 14

The composition C1(1) shows in each scan a percentage of the BS variable along the height of the cell, revealing the inhomogeneity of the sample. The composition C1(2) is rather homogeneus until 36 mm height; from this point, sedimentation and creaming phenomena become evident. Thus the sonication is more effective on the second dispersion that is less viscous. As a consequence, its solidification range (shown in Figure(b) 3) is more limited (the first one is 10.49 °C, the second one isFigure limited 5. ( ato) Backscattering4.56 °C) while (BS the) percentage total ΔH of of theC1(1) process sample; is and comparable. (b) of C1(2) sample. Figure 5. (a) Backscattering (BS) percentage of C1(1) sample; and (b) of C1(2) sample. ∆BS profiles are presented in Figure 6 in order to underline the nature of these phenomena.

(a)

(b)

FigureFigure 6. 6.( a(a))∆ ∆BSBS ofof C1(1)C1(1) sample; and and ( (bb)) ∆∆BSBS ofof C1(2) C1(2) sample. sample.

In sample C1(1), the flocculation phenomena are dominant and the shift of the trends up and down during the time depends on the particle diameter, smaller or larger than the incident radiation [17]. In sample C1(2), sedimentation, creaming, and flocculation are observed. The presence of sedimentation and creaming is probably due to the lower viscosity of the dispersion. Also for the composition C2, the second method of preparation makes the dispersions more homogeneous, as shown in Figure 7. Unlike the previous case, a minor effect on the solidification temperature interval is observed, as shown in Figures 3 and 4. However for this composition, in both sequences of reagent additions, only the flocculation is reported. The absence of sedimentation and creaming is probably correlated to a higher viscosity of composition C2(2) in comparison to C1(2) thanks to the higher contents of bentonite and the lower percentage of water. ∆BS percentages, obtained for composition C2, are presented in Figure 8 for both modalities of preparation.

Energies 2017, 10, 354 11 of 15

In sample C1(1), the ﬂocculation phenomena are dominant and the shift of the trends up and down during the time depends on the particle diameter, smaller or larger than the incident radiation [17]. In sample C1(2), sedimentation, creaming, and ﬂocculation are observed. The presence of sedimentation and creaming is probably due to the lower viscosity of the dispersion. Also for the composition C2, the second method of preparation makes the dispersions more Energieshomogeneous, 2017, 10, 354 as shown in Figure7. 11 of 14

(a)

(b)

FigureFigure 7. 7. ((aa)) BSBS percentagepercentage of of C2(1) C2(1) sample; sample; and ( b)) of of C2(2) C2(2) sample. sample.

Unlike the previous case, a minor effect on the solidification temperature interval is observed, as shown in Figures3 and4. However for this composition, in both sequences of reagent additions, only the flocculation is reported. The absence of sedimentation and creaming is probably correlated to a higher viscosity of composition C2(2) in comparison to C1(2) thanks to the higher contents of bentonite and the lower percentage of water. ∆BS percentages, obtained for composition C2, are presented in Figure8 for both modalities of preparation. The stability analysis has been quantified by considering the TSI behavior and comparing the trends obtained in 24 h. As already stated, such an index could be used as a measure of the progressive destabilization of the system. It is important to notice(a) that the optical testing obtained by Turbiscan is disruptive and non-repeatable. Thus, it is possible to test each sample one time during the 24 h period, with a scan of 5 min. It seems interesting to do the comparison between the repeated samples in terms of the TSI data scattering: the mean values of the errors in terms of TSI behaviour has been found, ranging from 6.25% to 10.2%. In Figure9, the TSI values are reported with reference to sample C1 for the two different preparation modalities. It appears evident that modality 2 is not adequate for

(b)

Figure 8. (a) ∆BS of C2(1) sample; and (b) ∆BS of C2(2) sample.

Energies 2017, 10, 354 11 of 14

(a)

Energies 2017, 10, 354 12 of 15 the preparation of Glauber’s salt, as discussed by observing the features of the samples in Figure2. TSI values increase progressively until 15, for sample C1(2). This trend indicates that the sonication of (b) water and bentonite is preferable in the ﬁrst step of preparation of such composition.. That allows us to achieve more homogeneousFigure 7. (a) BS samples. percentage of C2(1) sample; and (b) of C2(2) sample.

(a) Energies 2017, 10, 354 12 of 14

The stability analysis has been quantified by considering the TSI behavior and comparing the trends obtained in 24 h. As already stated, such an index could be used as a measure of the progressive destabilization of the system. It is important to notice that the optical testing obtained by Turbiscan is disruptive and non-repeatable. Thus, it is possible to test each sample one time during the 24 h period, with a scan of 5 min. It seems interesting to do the comparison between the repeated samples in terms of the TSI data scattering: the mean values of the errors in terms of TSI behaviour has been found, ranging from 6.25% to 10.2%. In Figure 9, the TSI values are reported with reference to sample C1 for the two different preparation modalities. It appears evident that modality 2 is not adequate for the preparation of Glauber’s salt, as discussed by observing the features of the samples in Figure 2. TSI values increase progressively until(b )15, for sample C1(2). This trend indicates that the sonication of water and bentonite is preferable in the first step of preparation of such composition.. Figure 8. (a) ∆BS of C2(1) sample; and (b) ∆BS of C2(2) sample. That allows us to achieveFigure more 8. (a homogeneous) ∆BS of C2(1) sample; samples. and (b) ∆BS of C2(2) sample.

Figure 9. Sample C1. Comparison of the time evolution of the Turbiscan Stability Index (TSI) as a Figure 9. Sample C1. Comparison of the time evolution of the Turbiscan Stability Index (TSI) as functiona function of of the the modality modality of of preparation. preparation.

In order to evaluate the influence of composition, TSI behavior has been reported in Figure 10 In order to evaluate the influence of composition, TSI behavior has been reported in Figure 10 for for samples C1(1), C2(1), and C2(2). That allows us to compare the influence of composition on samples C1(1), C2(1), and C2(2). That allows us to compare the influence of composition on stability stability (samples C1(1) and C2(1)) and the influence of modality in the cases where more (samples C1(1) and C2(1)) and the influence of modality in the cases where more concentrated solutions concentrated solutions were used (samples C2(1) and C2(2)). It is interesting to note that the TSI for were used (samples C2(1) and C2(2)). It is interesting to note that the TSI for C1(1) is almost double C1(1) is almost double the TSI of C2(1). No significant difference can be observed for C2(1) and C2(2). That could be due to the fact that the sample C2 is characterized by a higher thickener (as bentonite)/water ratio, able to guarantee a greater stability of the solution.

Figure 10. Comparison of the time evolution of TSI as a function of the composition and modality of preparation.

Energies 2017, 10, 354 12 of 14

The stability analysis has been quantified by considering the TSI behavior and comparing the trends obtained in 24 h. As already stated, such an index could be used as a measure of the progressive destabilization of the system. It is important to notice that the optical testing obtained by Turbiscan is disruptive and non-repeatable. Thus, it is possible to test each sample one time during the 24 h period, with a scan of 5 min. It seems interesting to do the comparison between the repeated samples in terms of the TSI data scattering: the mean values of the errors in terms of TSI behaviour has been found, ranging from 6.25% to 10.2%. In Figure 9, the TSI values are reported with reference to sample C1 for the two different preparation modalities. It appears evident that modality 2 is not adequate for the preparation of Glauber’s salt, as discussed by observing the features of the samples in Figure 2. TSI values increase progressively until 15, for sample C1(2). This trend indicates that the sonication of water and bentonite is preferable in the first step of preparation of such composition.. That allows us to achieve more homogeneous samples.

Figure 9. Sample C1. Comparison of the time evolution of the Turbiscan Stability Index (TSI) as a function of the modality of preparation.

In order to evaluate the influence of composition, TSI behavior has been reported in Figure 10 for samples C1(1), C2(1), and C2(2). That allows us to compare the influence of composition on Energies 2017, 10, 354 13 of 15 stability (samples C1(1) and C2(1)) and the influence of modality in the cases where more concentrated solutions were used (samples C2(1) and C2(2)). It is interesting to note that the TSI for theC1(1)TSI isof almost C2(1). double No signiﬁcant the TSI difference of C2(1). canNo besignificant observed difference for C2(1) andcan C2(2).be observed That could for C2(1) be due and to theC2(2). fact That that could the sample be due C2 to is the characterized fact that the by sample a higher C2 thickener is characterized (as bentonite)/water by a higher thickener ratio, able (as to guaranteebentonite)/water a greater ratio, stability able to of guarantee the solution. a greater stability of the solution.

FigureFigure 10.10. Comparison of of the the time time evolution evolution of of TSITSI asas a afunction function of of the the composition composition and and modality modality of ofpreparation. preparation.

In any case, samples C2(1) and C2(2) are more stable than C1(1). However, sample C2(2) offers better thermal performances in comparison to C2(1), and consequently it should be preferred for further studies.

4. Conclusions In this paper, samples of PCMs based on Glauber’s salt dispersions have been studied, by changing their composition and modality of preparation with the aim of estimating, by the T-history method, their thermal properties, and by means of optical methods, their stability. It has been observed that both compositions and modalities of preparation significantly affect the phase change enthalpy, melting temperature range, and structural stability. This is due to how the components are distributed in the PCM solutions at different thermal conditions, affecting both dispersion and viscosity as also verified for other families of heterogeneous materials [20,21]. This is more relevant when PCMs are subject to numerous thermal cycles or when crystallization or supercooling are involved. In particular, the nucleant selected in order to limit the supercooling has produced satisfactory results in all the analyzed cases, as it is able to reduce this phenomenon to a few degrees celcius. With reference to the thickener, it was observed that the stability of the dispersion improves for the samples that present a greater bentonite/water ratio, thanks to the higher viscosity. In defining the optimal composition of Glauber’s salt and additives, the results underlined that the thermal performances and the stability of the dispersion cannot be defined by considering only the percentage of the composition. The final properties depend on the preparation method and in particular, on the sequence used for adding the reagents and on the viscosity of the mixture. In any case, the aim of this paper is to provide an example of a procedure useful for testing the quality of inhomogeneous samples by considering both thermal performance and stability. The different compositions and modalities of preparation have been tested in order to underline the different behaviors of PCMs in relation to these variables. We suggest the use of this method as a way to characterize PCMs and, consequently, as a method of preparation and composition optimization. Of course, further studies should be carried out at different lifetimes, depending on the purpose of the investigation. However, it is important to verify how the preparation methodology can affect the initial thermal properties and the stability of samples. Energies 2017, 10, 354 14 of 15

Acknowledgments: This publication is part of a project that has received funding from the European Union’s Horizon 2020 research and innovation program under grant agreement No. 657466 (INPATH-TES). Author Contributions: Maria Gabriela De Paola carried out the experimental activity. She realized the T-history system and she introduced the use of Turbiscan as method for the characterization of PCM dispersions. Furthermore she carried out, with other authors for the analysis and elaboration of data. Finally, she was responsible of the reference selection. Natale Arcuri cooperated with the acquisition of experimental instrument and during the preliminary part of the research. He was also involved in the study of literature. Vincenza Calabrò, corresponding author, followed and coordinated the experimental activity. She was mainly involved in the analysis and elaboration of experimental data in order to describe the phenomena that occurred. Furthermore she has worked as in the research setup phases, as for writing and revision of the manuscript. Marilena De Simone, responsible of the European Union’s Horizon 2020 research and innovation program under grant agreement No. 657466 (INPATH-TES), was involved in the analysis and elaboration of experimental data. She plays a significant role for manuscript preparation and for writing and revising it until the publication. Conflicts of Interest: The authors declare no conflict of interest.

References

1. Atanasova, D.; Lyakov, N.; Vassilev, V.; Angelov, G.; Haralampiev, G. Desulphurization of lead cake by sodium carbonate and sodium hydroxide. J. Univ. Chem. Technol. Metall. 2008, 43, 267–272. 2. Atanasova, D.A. Hydrometallurgical processing of dumped lead paste for lead acid batteries. Bulg. Chem. Commun. 2009, 41, 285–296. 3. Telkes, M. Thixotropic Mixture and Method of Making Same. U.S. Patent 3,986,969 A, 19 October 1976. 4. Derluyn, H.; Saidov, T.A.; Espinosa-Marzal, R.M.; Pel, L.; Scherer, G.W. Sodium sulfate heptahydrate I: The growth of single crystals. J. Cryst. Growth 2011, 329, 44–51. [CrossRef] 5. Eckoff, S.; Okos, M. Solar Energy Heat Storage for Home, Farm and Small Business: Suggestions on Selecting and Using Thermal Storage Materials and Facilities. Available online: https://www.extension.purdue.edu/ extmedia/ae/ae-89.html (accessed on 22 September 2016). 6. Mehling, H.; Cabeza, L.F. Heat and Cold Storage with PCM: An up to date Introduction into Basics and Applications; Springer: Berlin/Heidelberg, Germany, 2008.

7. Feng, G.; He, N.; Li, G.; Chen, Q.; Li, H.; Huang, K. Experimental research of anhydrous Na2SO4+H2O phase change energy storage system. In Proceedings of the APEC Conference on Low-carbon Towns and Physical Energy Storage, Changsha, China, 25–26 May 2013. 8. Prabhu, P.A.; Shinde, N.N.; Patil, P.S. Review of phase change materials for thermal energy storage applications. Int. J. Eng. Appl. 2012, 2, 871–875. 9. Ultrasonic Production of Nano-Size Dispersions and Emulsions. Available online: http://www. euronanoforum2011.eu/wp-content/uploads/2011/10/enf2011_production-nanostructures_hielscher_fin.pdf (accessed on 22 September 2016). 10. Zhang, Y.; Jiang, Y. A simple method, the T–history method, of determining the heat of fusion, speciﬁc heat and thermal conductivity of phase-change materials. Meas. Sci. Technol. 1999, 10, 201–205. 11. Stankovic, S.B.; Kyriacou, P.A. Improved measurement technique for the characterization of organic and inorganic phase change materials using the T-history method. Appl. Energy 2013, 109, 433–440. [CrossRef] 12. Marín, J.M.; Zalba, B.; Cabeza, L.F.; Mehling, H. Determination of enthalpy–temperature curves of phase change materials with the temperature-history method: Improvement to temperature dependent properties. Meas. Sci. Technol. 2013, 14, 184–189. [CrossRef] 13. Hong, H.; Kim, S.K.; Kim, Y.S. Accuracy improvement of T-history method for measuring heat of fusion of various materials. Int. J. Refrig. 2004, 27, 360–366. [CrossRef] 14. Mengual, O.; Meunier, G.; Cayre, I.; Puech, K.; Snabre, P. Characterisation of instability of concentrated dispersions by a new optical analyser: The TURBISCAN MA 1000. Colloids Surf. A 1999, 152, 111–123. [CrossRef] 15. Mengual, O.; Meunier, G.; Cayre, I.; Puech, K.; Snabre, P. Turbiscan MA 2000: Multiple light scattering measurement for concentrated emulsion and suspension instability analysis. Talanta 1999, 50, 445–456. [CrossRef] 16. De Paola, M.G.; De Simone, M.; Arcuri, N.; Calabrò, V. Crossed analysis by T-history and Turbiscan for the characterization of PCM with Glauber salt. In Proceedings of the INNOSTORAGE Conference, Beer-Sheva, Israel, 16–18 February 2016. Energies 2017, 10, 354 15 of 15

17. Data Interpretation for Turbiscan Analysis. Available online: http://files.instrument.com.cn/FilesCenter/ 20061226/SH101155-36019.pdf (accessed on 2 March 2017). 18. García-Romero, A.; Diarce, G.; Ibarretxe, J.; Urresti, A.; Sala, J.M. Influence of the experimental conditions on the subcooling of Glauber’s salt when used as PCM. Sol. Energy Mater. Sol. Cells 2012, 102, 189–195. [CrossRef] 19. Biswas, D.R. Thermal Energy Storage using sodium sulphate decahydrate and water. Sol. Energy 1977, 19, 99–100. [CrossRef] 20. Lupi, F.R.; Gabriele, D.; Baldino, N.; Seta, L.; de Cindio, B.; De Rose, C. Stabilization of meat suspensions by organogelation: A rheological approach. Eur. J. Lipid Sci. Technol. 2012, 114, 1381–1389. [CrossRef] 21. Seta, L.; Baldino, N.; Gabriele, D.; Lupi, F.R.; de Cindio, B. Rheology and adsorption behaviour of β-casein and β-lactoglobulin mixed layers at the sunflower oil/water interface. Colloids Surf. A Physicochem. Eng. Asp. 2014, 441, 669–677. [CrossRef]

© 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).