Chemically improved Terminalia catappa L. oil: A possible renewable substitute for conventional mineral transformer oil
M.C. Menkiti
a,b,*, C.M. Agu
b, P.M. Ejikeme
c, O.E. Onyelucheya
daCivilandEnvironmentalEngineeringDepartment,WaterResourcesCenter,TexasTechUniversity,Lubbock,TX,USA
bChemicalEngineeringDepartment,NnamdiAzikiweUniversity,Awka,Nigeria
cDepartmentofPureandIndustrialChemistry,UniversityofNigeria,410001Nsukka,Nigeria
dChemicalEngineeringDepartment,FederalUniversityofTechnology,Owerri,Nigeria
ARTICLE INFO Articlehistory:
Received9September2016
Receivedinrevisedform22January2017 Accepted24January2017
Availableonline27January2017 Keywords:
Transformeroil Terminaliacatappa Solventextraction Transesterification Tukey’sposthocanalyses Analysisofvariance
ABSTRACT
ThisworkfocusesonthechemicalmodificationandcharacterizationofTerminaliacatappakerneloil (TCKO)forpossibleuseassubstituteformineraltransformeroil.TheTCKOwasextractedbysolvent extractionmethod.Directpurificationandtransesterificationmethodswereindividuallyappliedforthe modificationofthekerneloil.ModifiedTerminaliacatappakerneloil(MTCKO)andTCKOcharacteristics weredeterminedusingstandardmethods.Fouriertransforminfrared(FTIR)spectrometrywasusedto determinetheprevalentfunctionalgroupsintheMTCKOsamples.Atprocessconditionsof55C,150min and0.5mmparticlesize,kerneloilyieldwas60.45%(byweight).ForMTCKOblendedwithantioxidant (AceticacidAA,CitricacidCA),dielectricstrengthof46.36kVand48.55kVwereobtainedwhendirect purificationandtransesterificationmethodswere,respectively,used.TheANOVAandtheTukey’spost hocanalysesindicatedthattimeandparticlesizeeffectsweresignificantwhiletemperaturewasnot.
PhysicochemicalpropertiesofTCKOandMTCKOindicateditspotentialforuseastransformerfluid.
©2017ElsevierLtd.Allrightsreserved.
1.Introduction
The major energy supply of the world is substantially dependentonpetroleumproducts,whichinfluencethefunction- alityofdifferentsupplyoptionssuchaselectricity.Amongthese majorinputsincludebutnotlimitedtooils.Higherproportionsof theworld’soilneedsaremetthroughmineraloilproducedfrom petroleumofwhichtransformeroil(TO)isnotleftout.Mineraloil basedtransformerfluidsareextensivelyusedinbothpowerand distributiontransformers[1–3].Anestimated30–40billionliters ofmineraloilarepresentlyinuseintransformersglobally[4,5].
These sources are not renewable and at the current rate of consumption,demand willincrease withyears tocome.Mean- while,both governmentandenvironmental regulatoryagencies arepresentlyenacting strictenvironmental lawsandguidelines aimed at minimizing the accidental and non-accidental risks associatedwithmineralTOusage[4–6].
The limitations associated with utilization of mineral TO emphasize the critical challenges facing its continual use in transformers.Inthelightofthis,TOlikeanyotherindustrialoil, needs to meet international regulatory standard that would envisagestableenvironmentalstatusineventofspillage.Hence, thereisneedfortheprospectiveoiltobehighlybiodegradableand exhibitsvitalTOpropertiessuchasdielectricstrength,flashpoint, viscosityand pour point. Achievingthese requirementscall for viable alternative bio-degradable sources of TO using locally available seedsinNigeria (Terminalia catappain thisstudy) and elsewhere[7].
Over theyears, successfulresultshad been obtainedonthe potentialapplicationofoilseeds/nutsthatincludebutnotlimited tococonut[8,9],soybeanoil[8],sunflower[10,11],Castoroil[12], rapeseed(canola)[10],palmkernel[8,13],Jatrophacurcas[14,15]
and palm oil [9,16] for TO production. However, the major challengethatlimitstheuseofoilfromsomeoftheseseed/nuts asTOistheirhighpourpoint[17].Itisinthelightofthischallenge thatthisworkalternativelyexploitedthenovelutilizationofTCKO asastartingmaterialforTOproduction[18].
Terminalia catappa (TC) is of Combretaceae family, [19] and distributedthroughoutthetropics[20].Theworld’sproductionof Terminaliacatappafruit/seedisestimatedatabout700,000tons perannum[21].InNigeria,Terminaliacatappaiswellpopulatedin
* Correspondingauthorat:ChemicalEngineeringDepartment,NnamdiAzikiwe University,Awka,Nigeria.
E-mailaddresses:cmenkiti@yahoo.com,matthew.menkiti@ttu.edu (M.C.Menkiti).
http://dx.doi.org/10.1016/j.jece.2017.01.037 2213-3437/©2017ElsevierLtd.Allrightsreserved.
ContentslistsavailableatScienceDirect
Journal of Environmental Chemical Engineering
j o u r n a l h o m e p a g e : w w w . e l s ev i er . c o m / l o c a te / j e c e
thesouth-east[18,21,22].Thekernelisnon-foodcompetingand containsaminoacids(22–25%),lipids(35–52%),unsaturatedfatty acids(oleicacidandlinoleicacid),minerals[21,23]andrelatively highoil contentthat providedfocusfor this study[24–30].For instance,TCKOobtainedbyOliveriaetal.[31]was583.0g/kgdry matter or about 49–60% [18,32]. These results are relatively comparabletothose fromsunflower,peanutandrapeseed[31].
Consequently,investigationshavebeenmainlyconductedonthe useTCKOandoilsfrommentioned plantsfortheproductionof biodiesel [20,32]. More specifically, TC hasattracted significant investigationonthebiologicaland phytochemical studiesofits leaves,barkanditsfruitsextractsformedicinalpurposes[33–35].
Inthiscurrentwork,therefore,theauthorsextendedtheuseof TCKO(duetoitsavailability,highoilyield,andbio-degradability) fornovel applicationasTO,sincenospecificstudieshavebeen directedinthisdirection.Itisthisgapthatthisstudyseekstoclose.
Manystudieshavereportedthemodificationofvegetableoils intopotentialTOusingdifferentapproaches:esterification[36], transesterification[37–39]anddirectpurification[10,18]methods.
Oneofthemajorsetbacksforstand-aloneesterificationofoilsisits low purity level due toacid contamination of the product, as against no or low acid content for direct purification and transesterificationmethods[10,18].
Themajorchallengethatwouldassociatewithmodifyingoils (TCKO in this case) for possible use as TO using alkali (KOH) catalyzedtransesterificationwouldbelowyieldofalkylesterand difficultyinproductseparationduetosoapformation[40,41].Asa betteralternative,acid-catalyzed(H2SO4)transesterificationcould beemployedtodirectlyconverttheFFAintheoilintoalkyester first. However, in order to eliminate these disadvantages (for higher yield and alky ester purity), current study adopted sequentialtwo-stepconversionprocess:anacid-catalyzedesteri- fication (to lower the FFA content) and alkali-catalyzed trans- esterification (to improve on the alky ester purity and yield) [41,42].
Inthecaseofdirectpurification,modificationofoilsforpossible useasTOwouldbeachievedusingactivatedclayandaluminaas reported by Sundin [10], for a number of vegetable oils. This procedure was extended to the modification of TCKO, with advantageofremovingsimplepolarmolecules(byactivatedclay)
andFFA(byactivatedalumina)forimprovedstableoilwithbetter electricalcharacteristics[10].
Thiswork therefore,elicits interest intothe modification of TCKO, using direct purification and sequential acid-catalyzed esterification-alkali-catalyzed transesterification methods, for possibleuseastransformerfluid.Theevaluationoftheproducts qualityfromthestudiedprocesseswouldalsobeconsidered,and better process route identified. Furthermore, physicochemical characteristicsof theTCKOand MTCKOsamplesobtainedwere carried out using standard methods. The prevalent functional groups present in the MTCKOsamples were determined using FourierTransformInfrared(FTIR).Finally,sensitivityanalysiswas alsocarriedout, whilestatisticalanalysis(usingSPSSstatistical package, version 21) was conducted to determine the process parameters whose variationshad statisticalsignificance onthe processyield.
2.Materialsandmethods
2.1.Materials
Terminalia catappa kernels (TCK)were picked from Nsukka, EnuguStateNigeria.Activatedclayandaluminawerealsoobtained fromNsukka.AnalyticalgradeH2SO4(purity>99%),KOH,n-hexane andanhydrousmethanolwerepurchasedfromlaboratorychemi- cal vendor in Enugu. All reagents were used without further purification.
2.2.ExtractionofTCKO
TheTCKwereoven-driedat60Cfor12h,milledandsievedto obtainvariousparticlesizes.15gof milledkernels ofa specific average particle size was packed in a thimble of the soxhlet extractorandtheextractorwasfilledwith150mlofn-hexane.The oilextractionprocesswascarriedoutattemperaturesof35,40,45, 50, and 55C using n-hexane and five average particle sizes (0.5mm,1.0mm,1.5mm,2.0mmand2.5mm).Ateverytempera- ture,extractionwascarriedoutfor30,60,90, 120,and150min.The oil yieldobtainedat theend ofeach extraction time, forevery extractionconditionwascalculated andrecorded.Thesoluteto solventratiousedfortheentireextractionwas1:10(15g:150ml), mass-by-volume.Theentireextractionprocesswascarriedoutin triplicate and the average values reported. The necessary conditionsforadequateextractionprocedurewerealsoobserved fortheprocesses.Extraction(usingn-hexane)ofTCKOfromthe milledkernels,andevaluationofpercentageoilyieldweredone accordingtotheAssociationofOfficialAnalyticalChemists(AOAC) standardmethod[43].
2.3.PhysicochemicalpropertiesofTCKO
Thephysicochemicalcharacterizationsweredoneusingcrude TCKOextract.Theoildensity(AOAC985.19),iodinevalue(AOAC 993.20) and acidity/acidvalue(AOAC 969.17), weredetermined according to AOAC approved techniques [44]. However, TCKO viscosityanddielectricstrengthweremeasuredfollowingASTM D445[45]andIEC60156[46]standardmethods,respectively.Each physicochemical property was measured three times, and the averagevaluesofthepropertieswererecorded.
2.4.FattyacidcompositionofTCKO
ThefattyacidprofilewasdeterminedaccordingtoAOAC996.06 [44].Inthemethod,thequantitativedeterminationoffattyacidsin theTCKOwasperformedinagaschromatograph(ShimadzuGC– 14B,Model910),equippedwitha flameionizationdetectorand Nomenclature
AA Aceticacid AAA Aceticacidadditive
ASTM AmericanSocietyforTestingandMaterial AV Acidvalue
A.O.A.C. AssociationofOfficialAnalyticalChemist A.O.C.S. AmericanOilChemist’sSociety
CA Citricacid CAA Citricacidadditive GC Gaschromatography FTIR Fouriertransforminfrared IV Iodinevalue
TC Terminaliacatappa
TCKO Terminaliacatappakerneloil
MTCKO ModifiedTerminaliacatappakerneloil
MTCKOd ModifiedTerminaliacatappakerneloilobtainedby directpurificationmethod
MTCKOt ModifiedTerminaliacatappakerneloilobtainedby transesterificationmethod
DS Dielectricstrength SFA Saturatedfattyacids
integrator,usingaHP88capillarycolumn(0.25mmi.d.100m, filmthickness0.25
m
m–ShimadzuCorporation,Tokyo,Japan).In othertoachievethis,theinjectoranddetectortemperaturevalues werethesameat250C.Theoventemperaturewasmaintainedat 190Cfor15min.Itwasthensubjectedtoriseupto230Catthe rate of 5C per min. Thereafter, it was controlled at this temperaturefor the sametime interval as theinitial step.The carriergasusedwasnitrogenatapressureof500kPa.Thefatty acidswereidentifiedand compared withstandard compounds.Thequantityofeachfattyacidwasestimatedfromthepercentage areaoftheindividualfattyacid[47,48].Theanalysiswasconducted threetimes.
2.5.Directpurificationmethod
DirectpurificationmethodwasusedtoproduceMTCKOfrom TCKO.Inthismethod,0.756lofTCKOwasheatedto90C.90.8gof adsorbentmixture(60%activatedclayand40%activatedalumina) wasintroducedintotheTCKOsampleandstirredcontinuouslyfor 70min.Attheend ofmixing,theadsorbentwasremovedusing plate and frame filters. The residual clay particles, dissolved moistureandgaseswereremovedbysequentiallypassingthefluid throughaparticulatefilter(5micronssize)andgasandmoisture removingdegassifier.Additives(Aceticacid(AA),Citricacid(CA)), at0.2%(bymassofmodifiedoil)wereafterwardsaddedtoprevent theoxidationofthemodifiedoil.TheresultingoilwastheMTCKO.
2.6.Transesterificationexperiment 2.6.1.Acidcatalyzedesterification
TCKOsamplewasheatedto60Cinareactionglassvessel.A solutioncomposed ofcatalytic H2SO4(1.0%w/w) and 99%pure methanol(30%v/v)wasalsoheatedat30Cfor4h. Theheated solutionwas added tothe reactionglass vessel containing the TCKOtoinitiatetheesterificationreaction.Themixturewasstirred withamagneticstirreratastirringspeedof600revolutionsper minute(rpm)for1h,afterwhichthecontentwaspouredintoa separating funneland allowed tosettle for 2h. The methanol- waterfraction(thetoplayer)wasremovedandtheoilphasewas usedforbase-catalyzedtransesterification.
2.6.2.Basecatalyzedtransesterification
50ml of theTCKO sample obtainedfromtheacidcatalyzed esterificationwaspouredinto150mlconicalflaskandheatedto 60C usinga waterbath.Solutionofpotassiummethoxidewas preparedbydissolving0.25gofKOHpelletsinanagitated250ml beakercontaining10.5mlofanhydrousmethanol.ThePotassium methoxidesolutionwasthentransferredintothewarmesterified TCKOsampleatamethanolto-oilratioof6:1andstirredvigorously for 90min, usingmagnetic stirrer. Themixture was left undis- turbedfor24hinaseparatingfunneltoensurepropersettling.
Aftersettling,theupperlayer,whichwasmethylester[modified Terminalia catappa kernels oil (MTCKO)] was decanted into a beakerand washedwithdistilledwater. Thiswas toensurethe removalof residualmethanol,glycerin,catalyst,soapand other impurities. It was then demoisturized by heating slowly to constanttemperatureof100C,whilethelowerlayercomprising ofglycerol,andsoapwascollectedfromthebottomofthefunnel.
Acetic acid (AA) and Citric acid (CA) additives were then introducedtopreventtheoxidationoftheMTCKO.
2.7.FourierTransformInfraredSpectroscopy(FT-IR)analysis TheFTIRanalysisoftheMTCKOsampleswascarriedoutusing BUCKScientificInfraredSpectrophotometerModel530.
2.8.Statisticalanalyses
One-way analysis of variance (ANOVA) and Tukey’s honest significant difference (HSD) test were performed using SPSS statisticalpackage(version21).TheTukey’sHSDtestwasusedto test for significantdifferences betweentwogroups. Thistest is used to define a principal value called the honest significant difference (HSD).HSDis theminimum distancethat mustexist betweentwogroupsmeans,beforethedifferencebetweenthem could be considered as being statistically significant [49].
Significantlevelsweretestedatp<0.05.
2.8.1.Analysisofvariance(ANOVA)
Inthisstudy,temperature,particlesize,andtime,whichwere independentvariables,werecomparedwiththeresponsevariable (%yield),bymeansofANOVAsoastoestimatevariancewithinand betweengroups.Thecorrelationcoefficientsandtheirprobability levelswereobtainedfromquadraticregressionanalysis.
2.8.2.Tukey’sHSDtest
Thistestisbasedondefiningaprincipalvaluecalledthehonest significant difference (HSD). Like previously stated, HSD is a representation of the minimum distance between two group means that must exist before the difference between the two groupsistobeconsideredstatisticallysignificant[49].
3.Resultsanddiscussion
3.1.OilpropertiesofTCKO
TheoilyieldfromTCKwasfoundtobe60.45%(bydrymass), which was higher than thevalue reportedfor cottonseed[50], soyabean[51]andothercommercialoilsources[52].Thus,thereis apossibilityofitseconomicutilization,industrially.However,Iha etal.[32]reported50%yield(bydrymass)fortheTCKobtained from Brazil’s coastal region. This value was lower than that obtainedin this work.This differencein theoil yield couldbe attributedtofactorssuchasgeographicallocation,seedvarietyand harvestperiod[53].
SomeimportantphysicochemicalpropertiesofTCKOshownin Table1,weredeterminedandcomparedwiththatreportedbyIha etal.[32].ItcouldbeobservedfromTable1thattheviscosityand acidityofTCKOadiffergreatlyfromthatofTCKOb,withthelater havinghigherviscosityandaciditythantheformer.Itwouldbe notedthatthehighoilyieldandthelowacidvalueoftheTCKO couldbe attributedtothespecieof Terminalia catappaL.Thus, improvedbreedsofTerminaliacatappakernelsenhanceditsyield andoilproperties.Theiodinevalue(IV)oftheTCKOinthiswork (Table 1), was higher than 83.92g/I2/100g oil reported byDos santosetal.[20].Thiswasanindicationofthemoderatelevelof unsaturation intheoil ascouldbeobservedinTable2.Theoil thereforeissemi-dryinginnature.Finally,thedielectricstrengthof TCKOwashigherthanthatofpalmkerneloil(25KV)[8].
Table1
PhysicochemicalpropertiesoftheTCKO.
Oilproperty TCKa TCb StandardMethod
Oilyield(%) 60.45 50 ASTMD445
Dielectricstrength(KV) 30.61 – Viscosity(mm2s1) 20.29 36.8
Acidvalue(mgKOH/goil) 4.73 10.5 AOCSCD3d–63
Densityat20C(gm3) 890 913 ASTMD1298
IodineValue(g/I2/100goil) 101.86 – AOCSCD1c–85 –Notreported.
aExperimentalvalues.
bIhaetal.[32].
The TCKO fatty acid composition, determined by GC, is presentedin Table 2,and compared withthat reported byIha etal.[32].FromtheresultsinTable2,itcouldbeobservedthat morethan40%oftheTCKOawascomposedofsaturatedfattyacids (SFAs),whileTCKObreportedbyIhaetal.[32]containnearly10%
lessSFAs.Theoilhadover55%ofitscompositionasunsaturated fatty acid. The result of the fatty acid composition for TCKO obtainedinthisworkwasincloseagreementwiththosereported forTCKOinBrazil[20,32].Furthermore,itcouldbeobservedthat TCKOahadlauricacidpresenceaswellashighlevelofsaturation, unliketheTCKOb.Thiscouldbemostlikelyduetofactorssuchas geographicallocationandvariety[53].
3.2.PhysicochemicalcharacteristicsofMTCKOobtainedbydirect purification(MTCKOd)
3.2.1.Dielectricstrength
Dielectric strength(DS) ofTO isthemaximum electric field strengththattransformeroilcanwithstandcontinuouslywithout breakingdown or experiencing failureof its properties [15]. It couldbeobservedfromTable3,thatthevalueofDSofTCKOwas 30.61kV.Furthermore,theDSofMTCKOd,MTCKOd+0.2%AAA,and MTCKOd+0.15%AAA&CAAwere32.88kV,39.46kVand46.36kV, respectively.ItisevidentfromTable3,thattheDSvalueofthe MTCKO was higher than that of TCKO. Furthermore, MTCKOd samplesblendedwithadditiveshadhigherDSthantheunblended ones.ThisenhancingeffectofadditiveontheDSofMTCKOdwas duetolessformationofcarbonwhichcausedlessertendenciesfor theformationofgases[54].ThesevalueswerelowerthantheDSof coconutoil(60kV),butcomparabletothatofsoyabeanoil(39kV) [8].However,MTCKOdhadDSvaluesthatweremuchhigherthan thatofpalmkerneloil(25kV)[8].Table3indicatesthatthehighest
DS for the blend of MTCKOd+0.2% AAA and MTCKOd+0.15%
AAA&CAAwere39.46kVand46.36kV,respectively,andbothfall withinthestipulatedrangeforconventionalmineraltransformer oil.
3.2.2.Moisturecontent
Moisturecontentrepresentstheamountofwatercontainedina substancesuchasTO,anditisdirectlyproportionaltoTOaging [55].Table3indicatesthatthemoisturecontenthadlimitingeffect onthedielectricstrengthoftheMTCKOdsamples.Aswatercontent increased,thecoolingpropertiesandDSoftheTOdecreased,given thathigherwatercontent(andotherimpurities)wouldlowerthe temperatureat which theoil couldstartconducting electricity.
Furthermore,moisturepresencewouldresultinoildecomposition duetooxidation[55].Table3indicatesthatthemoisturecontentof theMTCKOdrangedfrom1.3–1.4mg/kg.Thesevaluesareinclose agreementwiththosereportedbyUsmanetal.[8]forcoconutoil (1.0mg/kg),soyabeanoil(2.0mg/kg)andpalmkerneloil(1.9mg/
kg)andlessthan20mg/kgrecommendedmaximumbyIEC60156 testmethod[56].
3.2.3.Flashandpourpoints
Flashpointofavolatilematerialisthelowesttemperatureat which itcanvaporizeand ignitewhenexposedtoa flameora spark. Conversely, pour point is the temperatureat which the amountofwaxoutofsolutionandtheoilcanstillflow[55].Table3 showsthatthepourandflashpointsofTCKOdecreasedaftertheoil wasmodifiedusingdirectpurificationmethod.However,theflash pointofMTCKOdincreasedwiththeintroductionof additive(s).
MTCKOdblendedwithbothAA andACadditiveswereofhigher qualitycomparedtoAAblendedsample.BecauseAAandACcould notfunctionasdepressant,theyhadlittleornoeffectonthepour pointof MTCKOd. Meanwhile,thepourpointsofMTCKOdwere higherthantheconventionalTO.ThepourpointsofMTCKOdin Table3werelowerthanonesforsoyabeanoil(1C),palmkernel oil(15C)andcoconutoil(20C),whileflashpointswerehigher thanthoseofcoconutoil(225C),soyabeanoil(234C)andpalm kerneloil(242C)[8].
3.2.4.Viscosity
Viscositymeasurestheresistanceoffluids(e.g.TO)toflow.Itis thereforeameasureoftheresistanceofTOwithreferencetoits gradualdeformationbyshearstressortensile stress[55].From Table3,it couldbenoted thattheviscosityofMTCKOsamples decreasedsignificantlywhencomparedwiththatoftheTCKO.It couldalsobe observedfromTable 3 that theviscositiesof the MTCKO samples decreased when blended with antioxidant additives.It wasalsoaffectedbythenumberof additivesused, asitwasevidentthattheMTCKOsampleblendedwithbothAAand AC additives had lower viscosity than that blended with AA Table2
FattyacidcompositionoftheTCKO.
Fattyacid TCKOa TCKOb
C12:0(Lauricacid) 0.94 –
C14:0(Myristicacid) 0.54 0.10
C16:0(Palmiticacid) 36.01 28.30
C16:1(Palmitoleicacid) – 0.90
C18:0(Stearicacid) 6.4 4.90
C18:1(Oleicacid) 33.25 30.00
C18:2(Linoleicacid) 22.26 32.80
C18:3(Linolenicacid) 0.59 1.70
Saturatedfattyaccids(%) 43.89 34.20
Mono-unsaturatedfattyacid(%) 33.25 30.00
Poly-unsaturatedfattyacid(%) 22.85 34.50
Unsaturatedfattyacids(%) 56.10 74.50
–Notreported.
aExperimentalvaluesofthiswork.
b Ihaetal.[32].
Table3
PhysicochemicalpropertiesofTCKO,MTCKOobtainedfromdirectpurificationmethod.
Property Unit TOa TCKOb MTCKOc MTCKOd+0.2%AAA MTCKOe+0.15%AAA&CAA Standardmethod
Dielectricstrength KV 40–60 30.61 32.88 39.46 46.36 IEC60156
Moisturecontent mg/kg <20 2.1 1.4 1.3 1.3 ASTME203
Pourpoint C 48 3 3 3 3 ASTMD97
Flashpoint C 152 260 256 266 276 ASTMD93
Density,20C g/cm3 870 890 850 840 840 ASTMD1298
Viscosity,40C mm2/s 10 20.29 15.22 12.48 10.96 ASTMD445
Acidvalue mgKOH/g <0.01 4.73 0.87 0.861 0.856 AOCSCD3d-63
aConventionalmineraltransformeroil.
b Terminaliacatappakernelsoil.
cModifiedTerminaliacatappakernelsoil.
d ModifiedTerminaliacatappakernelsoilblendedwith0.2%(aceticacidadditive).
eModifiedTerminaliacatappakernelsoilblendedwith0.15%(aceticacidandcitricacidadditives).
additivesonly.TheviscositiesoftheMTCKOsamplesinTable3 werelowerthanthoseof coconutoil(29mm2/s),soya beanoil (34.5mm2/s) and palm kernel oil (29.2mm2/s) at the same temperatureof40C[8].Inaddition,theviscositiesoftheMTCKO sampleswerelowerthanthatofTerminaliacatappa(36.8mm2/s) reportedbyIhaetal.,[32].Lowviscosityoftransformeroilishighly recommendedsincelowerviscosityenhancestheoverallperfor- manceoftheoilduringtransformeroperation[57].
3.2.5.Density
ThedensitiesofMTCKOdsamplesinTable3werewithinthe rangeof840to850g/cm3.Itcouldbeobservedthattheirdensities were slightly lower than that of TO produce from soybean (1462.4g/cm3), coconut (917g/cm3), and palm kernel oil (1462.4g/cm3)[8].Similarly,theirdensitiesweresimilartothat ofTerminaliacatappaseedoilbiodiesel(879g/cm3)[32].
3.2.6.Acidvalue
Acidvaluereferstothemeasurementofaciditythatisobtained bytheamountofpotassiumhydroxide(expressedinmilligrams) requiredtoneutralizethehydrogenions(H+(aq))inonegramofoil.
FromTable3,itcouldbeobservedthattheacidvaluesofMTCKOd werelowerthanthatofTCKO.Theiracidvaluesrangedfrom0.856 to 0.87mg KOH/g. These values were higher than the 0.01mg KOH/g for the conventional TO required by IEC standard.
Furthermore,MTCKOd and TCKO acid values were higher than 0.01mgKOH/gforAfricanbushmangonutoilbiodiesel[58]and 0.5mgKOH/gforTerminaliacatappaseedoilbiodiesels[32].Acid valueshouldbereducedasmuchaspossiblesinceitisdetrimental totheperformanceofthetransformer.
3.3.PhysicochemicalcharacteristicsofMTCKOobtainedby transesterification(MTCKOt)
3.3.1.Dielectricstrength
From Table 4, the DS of MTCKOt, MTCKOt+0.2% AAA, and MTCKOt+0.15%AAA&CAAwere33.95KV,41.08KVand48.55KV, respectively. Table 4 shows that the DS value of the MTCKOt samplesproducedbythismethodwashigherthanthatofTCKO.
Furthermore,MTCKOtsamplesblendedwithadditiveshadhigher DSthantheunblendedones.Thisenhancingeffectofadditiveon theDSofMTCKOcouldbeattributedtolessformationofcarbon which caused lesser tendencies to formation of gases [54].
However,MTCKOtsampleshadDSvaluesthatweremuchhigher thanthatofpalmkerneloil(25KV)[8].TheDSvaluesofMTCKOt samplesblendedwithboth0.2%AAAand0.15%AAA&CAAwere 41.08KVand48.55KV,respectively,andwasinagreementwiththe DSvalueofconventionalTO.
3.3.2.Moisturecontent
Considering Table 4, as the watercontent increased theDS decreased. This was because, water tended to increase the potential of the oil toconductelectricity, therebyreducing the coolingpropertiesandvoltage(DSvalue)atwhichtheoilstartsto conduct electricity. Secondly, moistureled to oxidation, hence, increased the amount of acid and water that could cause oil decomposition[55].FromTable4,MTCKOtsamplesexhibitslower moisture content(0.9–1.0mg/kg) which are in agreementwith reportsofUsmanetal.[8]forcoconutoil(1.0mg/kg),soyabeanoil (2.0mg/kg)andpalmkerneloil(1.9mg/kg)butlessthan20mg/kg recommendedmaximumIECstandard[56].
3.3.3.Flashandpourpoints
Table4showsthatthepourandflashpointsofTCKOdecreased after the MTCKO was produced by transesterification method (MTCKOt).Table4showsthatadditiveshadenhancingeffectonthe flashpointofMTCKOtsamples.Flashpointvaluewasaffectedby thenumberofadditivesused,asitwasevidentthattheMTCKOt sampleblendedwithbothAAandACadditivesshoweddifference, whencomparedtothatblendedwithAAadditiveonly.AAandAC additiveshadlittleornoeffectonthepourpointofMTCKOt.The pour points of the MTCKOt samples were higher than the conventional mineral TO, with exception of MTCKOt sample blended with 0.15% AAA & CAA. However, a possible suitable solution/techniques to overcome the differences in pour point betweenTOandTCKOwouldbetheuseofpourpointdepressants [59].ThepourpointsofMTCKOtsamplesinTable4,werelower thanthosereportedbyUsmanetal.[8]forsoyabeanoil(1C), palm kerneloil (15C) andcoconutoil (20C)while theirflash pointswerehigherthanthoseofcoconutoil(225C),soyabeanoil (234C)andpalmkerneloil(242C)withouttheuseofAAAand CAAadditives.
3.3.4.Viscosity
FromTable4,theviscosityoftheMTCKOtsamplesdecreased significantly when compared with that of the TCKO. Also, the viscositiesofthesamplesdecreasedwhenblendedwithantioxi- dantadditives.Thenumberofadditivesusedhadeffect,giventhat the MTCKOt sample blended with both AA and AC had lower viscosity than that blended with AA alone. MTCKOt and TCKO viscositieswerelowerthanthoseofcoconutoil(29mm2/s),soya beanoil(34.5mm2/s),palmkerneloil(29.2mm2/s)at40C[8]and TC(36.8mm2/s)[32].
3.3.5.Density
ThedensitiesofMTCKOtsamplesinTable4werewithinthe rangeof840and852g/cm3.Thedensitieswereslightlylowerthan those of transformer oil produced from soybean, coconut, and palmkerneloil[8]andclosetothatofTerminaliacatappaseedoil
Table4
PhysicochemicalpropertiesofTCKOandMTCKOobtainedfromtransesterification.
Property Unit TOa TCKOb MTCKOc MTCKOd+0.2%AAA MTCKOe+0.15%AAA&CAA Standardmethod
Dielectricstrength KV 40–60 30.61 33.95 41.08 48.55 IEC60156
Moisturecontent mg/kg <20 2.1 1 0.9 0.9 ASTME203
Pourpoint C 48 3 5 5 5 ASTMD97
Flashpoint C 152 260 255 265 275 ASTMD93
Density,20C g/cm3 870 890 852 840 840 ASTMD1298
Viscosity,40C mm2/s 10 20.29 14.1 11.84 10.29 ASTMD445
Acidity/Acidvalue mgKOH/g <0.01 4.73 0.857 0.845 0.844 AOCSCD3d-63
aConventionalmineraltransformeroil.
b Terminaliacatappakernelsoil.
cModifiedTerminaliacatappakernelsoil.
d ModifiedTerminaliacatappakernelsoilblendedwith0.2%(aceticacidadditive).
eModifiedTerminaliacatappakernelsoilblendedwith0.15%(aceticacidandcitricacidadditives).
biodiesel(879g/cm3)[32].Thedensities of MTCKOt samplesin Table4werecloseto870g/cm3stipulatedforconventionalTO.
3.3.6.Acidvalue
FromTable4,theacidvaluesofMTCKOtsampleswerelower thanthatof TCKO.Theacidvalues(0.844–0.857mgKOH/g)are significantlyhigherthanthe0.01mgKOH/gspecifiedformineral TObyIECstandard.0.844–0.857mgKOH/gwasalsohigherthan the values for African bush mango nut oil biodiesel [58] and Terminaliacatappabiodiesels[32].Therelativelyhighacidvalue reported in this work could be attributed to the single stage transesterification processusedasagainsttwostagetransester- ificationbyIhaetal.[32].Acidvalueshouldbereducedtopossible minimum since it is detrimental to the performance of the transformer.Thiscould beachievedby seriesof acidcatalyzed esterificationpriortothetwostagetransesterificationprocess.
3.4.Briefcomparisonoftheimportantcharacteristicsresults ThephysicochemicalpropertiesoftheMTCKOsamplesmodi- fied by direct purification and transesterification methods are summarized in Tables 3 and 4, respectively. MTCKO samples modified by both methods (MTCKOd and MTCKOt), exhibit TO propertiesthatwasinagreementwiththeconventionalmineral TO.FromTables3and4,theDSoftheoilsamples(blendedwith 0.15%AAA+CAAinTable3,MTCKOd,and0.2%AAAand0.15%AAA&
CAAinTable4,MTCKOt)wereinaccordancewiththestipulated standard.WiththeexceptionofpourpointandDSofMTCKOand MTCKO+0.2%AAAproducedbydirectpurification,otherproper- ties of the modified Terminalia catappa kernel oils samples producedbythetwomethodswereinaccordancewithstipulated standard[56].
Furthermore, thepresence of additives(antioxidants) inthe MTCKOsampleshadenhancingeffectonsomephysicochemical propertiesoftheMTCKOproducedbybothmethods.Someofsuch propertiesareviscosityandDS.Theeffectsofantioxidantsonthe DSofMTCKOsamplesareshowninTable5andFig. 1.TheDSvalues of the samples increased with the introduction of antioxidant
additives.TheenhancingeffectoftheantioxidantontheDSwas morepronouncedwiththecombinationoftwoantioxidants.This effectwasinagreementwiththatreportedbyRaymonetal.[54]
forsunfloweroil, ricebranoil,soyabeanoiland cornoilbased transformerfluid.
3.5.FT-IRanalysesofthemodifiedTerminaliacatappakerneloil samples
TheFTIRspectraldataareshowninTable6whilethespectral plotsareshowninSupplementaryFigs.S1andS2.TheFTIRspectral patterns of MTCKOd (samples obtained by direct purification method)areshowninFig.S1a–c.Theresultswerecomparedwith knownsignatureofidentifiedmaterialsintheFTIRlibrary[60].The spectraof theMTCKOd,MTCKOd+0.15%AAA/CCAand MTCKOd+ 0.2%AAA,producedbydirectpurificationexhibitedsevenmajor discernablepeaksatfrequencyrangeof4000–400cm1(Table6).
Theyhadpeakrangesthatcenteredaround3413–3417cm1.These peaks are characteristics of OH stretching, which was an indicationofthepresenceofwater.Thepeakrangesthatcentered around2914cm1 arecharacteristicsof CH stretching,which wasanindicationofthepresenceoffatsandcarbohydrates.Peaks around 2281–2285cm1 for MTCKOd, MTCKOd+0.15%AAA/CCA and MTCKOd+0.2% AAAwere characteristicsof combination of C¼H and NH stretching’s, depicting thepresence of organic/
aminespecies. The peaksat 2061.9cm1 and 2053.71cm1,for MTCKOdandMTCKOd+0.15%AAA/CCA,respectively,arecharacter- istics of OH stretching. Furthermore, the peaks around 1705.91cm1 were assigned to first overtone CH stretching and C¼O stretching linkedto esters. Finally,the peaks around 1180cm1indicatedpresenceofcarbohydrates.Thesefunctional groupspresentinMTCKOdsamplesweresimilartothosereported byMenkitietal.[61]fortherawunmodifiedTerminaliacatappa kerneloil.SimilarfunctionalgroupswerealsoobservedintheFTIR spectraofTerminaliacatappaoilreportedbyAdewuyietal.[62].
The FTIRresultsof MTCKOd(producebydirect purification) showed that MTCKOd basicallyhad commonfunctional groups irrespective of type and quantity of antioxidants content.
Table5
EffectsofantioxidantsonthedielectricstrengthvalueofMTCKOsamples.
Antioxidant Quantityofantioxidant(%) DielectricstrengthValue(KV)
DirectPurificationmethod Transesterificationmethod
AA 0.2 39.46 41.08
AA+CA 0.15 46.36 48.55
Fig.1.EffectofAntioxidantsonDielectricStrengthofMTCKO.
Furthermore, the presence of CH, C¼H, NH, and OH functional groupsindicatesbiodegradability, a majoradvantage forhealthyenvironment.[18].
TheFTIRspectraldataforMTCKOobtainedbytransesterifica- tion(MTCKOt)areshowninTable6whilethespectralplotsare shownintheSupplementFig.S2a–c.ThespectraoftheMTCKOt, MTCKOt+0.15%AAA/CCA and MTCKOt+0.2%AAAexhibited5, 7 Table6
FTIRspectrabandsincm1forthevarioussamplesprepared.
Samples/S/No. MTCKOd MTCKOd+0.15%AAA/CCA MTCKOd+0.2%AAA MTCKOt MTCKOt+0.15%AAA/CCA MTCKOt+0.2%AAA
1 3413.2798 3417.1711 3415.8074 3274.4526 3414.5013 3387.6645
2 2914.8405 2914.8017 2914.4658 2629.494 2914.1616 2920.209
3 2285.3545 2291.2034 2280.3988 – 2287.719 2295.692
4 2061.9439 2053.7058 2083.3333 – 2064.4639 2091.0856
5 1704.1497 1705.91 1706.594 1687.5759 1705.2147 1699.225
6 1180.9139 1190.0853 1184.6423 1291.8592 1187.5649 1182.717
7 708.342 700.2732 698.6434 778.7528 699.547 705.8223
Fig.2. EffectoftimeandparticlesizeontheyieldofTCoilat:(a)45C,(b)50Cand(c)55C.
and6majordiscernablepeaks,respectively,atfrequencyrangeof 4000–400cm1.Thepeaksaround3274–3414cm1arecharacter- isticsofOHstretching.Thepeakat2785.89cm1,forMTCKOtis characteristic of aldehyde CH stretching while the peak at 2629.49cm1,forMTCKOtischaracteristicofphosphorusacidand esterOHstretching.Thepeaksat2914.16and2920.21cm1,for MTCKOt+0.15%AAA/CCAandMTCKOt+0.2%AAA,respectively,are characteristics of CH (indication of fats and carbohydrates).
2287.72,and2064.46cm1peaks,forMTCKOt+0.15%AAA/CCAare characteristicsofC¼H/NHstretching,indicatingthepresenceof organic/aminespecies. Thepeak at 2292.37cm1,for MTCKOt+ 0.2%AAAindicatesC¼Hstretching.Thepeaksat1687.58,1677.22 and 1705.21cm1, for MTCKOt, MTCKOt+0.15% AAA/CCA and MTCKOt+0.2% AAA, respectively, are characteristics of first overtoneCHstretchingforesters.Peaksaround1182.72cm1, arecharacteristicsofcarbohydrateswhilethepeakat778.75cm1, forMTCKOtischaracteristicofthirdovertoneNHstretching.In general,thefunctional groupspresentinMTCKOt sampleswere mostlysimilartothosereportedforMTCKOdsamplesinthiswork andTCKOreportedelsewhere[61,62].
MTCKOt(bytransesterification)blendedandunblended with additives had basically common functional groups in them.
However,MTCKOt (unblended)had aldehyde C H stretching, whileitisnotpresentinMTCKOt+0.15%AAA/CCAandMTCKOt+ 0.2% AAA (blended), hence eliminating oxidation causing sub- stancein the later (blended) samples. The presence of C H, C¼H,NH,andOHfunctionalgroups,wereofgreatadvantage tothebiodegradabilityoftheoil,due tothepresence ofwater, organicmoleculesandcompounds[18,61].
3.6.EffectsetractionparametersonTCKoilyield 3.6.1.Effectofparticlesize
Theeffectof fivedifferentparticlesizes(2.5–0.5mm)onoil yieldispresentedinFig.2(a–c)fortemperaturesof45,50and55C respectively.Itcouldbeobservedthat thesmallerparticlesizes between 1.0 and 1.5mm extracted more oil by 1–2% when compared to its immediate preceding bigger particle size of 2.0mm.Conversely,thesmallestparticlesizeof0.5mmextracted more oil by 5% when compared with its immediate preceding bigger particle size of 1.0mm. This increased oil yield with decreasedparticlesizecouldbeattributedtothebiggerinterfacial areaofthefinerparticles,shorterdiffusionpathoffinerparticles.
However,inthecaseoflargerparticles,smallerquantityofoilyield was obtaineddue tominimal contact surface areaand solvent entrainmentdifficulty,aswellaslimitedoildiffusionfrominside thelargerparticletothesolution[63–65].Reduceddiffusionpath wouldleadtoincreasedmasstransferrateandsubsequenthigh rateofoildissolutionfrommilledsampleintothesolvent.
3.6.2.Effectoftemperature
TheTCKoilyieldat35,40,45,50,and55Candtimeintervalsof 30, 150minfor particlesizesof0.5 and 1.0mm areshownin Fig. 3a and b. It was evident that oil yield increased with temperatureincreaseduetoincreaseinoildiffusionanddecreased viscosity [65,66]. Observed temperatureincrease couldcause a marginaldecrease inthefluiddensityand solutesolubilityand subsquentlywouldenhancemasstransfercofficientforincreased extractionyield[65,67].
Fig.3.Effectoftemperatureandtimeonyieldatparticlesizesof:(a)0.5mmand(b)1.0mm.
3.6.3.Effectoftime
Figs.2(aandc)and3(aandb)indicatedoilyieldsincreased with increase in extraction time for varying particle size and temperature,respectively.Oilyieldwasinitiallyrapid(between30 and90min)duetoinitialinternalrapiddiffusionasfreeoilonthe surfaceofthemilledTCKwasexposedtofreshsolvent.Oilyield graduallyslowedbetween90and150minduetodepletionofoil contentofthemilledkernel[18,63,65].
In this study, the yield generally increased with increasing temperature/timeanddecreasingparticlesize,hence,thehighest yields of 60.45% and 49.0% wereobtained at 55C/150min for particle sizesof 0.5mm and 1.0mm, respectively. Higher yield would be expected with higher temperature due to higher available kinetic energy of the solutesfor faster diffusion into thesolvent.
3.7.EsterificationandtransesterificationofTCKO
The removal of FFA relies on acid-catalyzed esterification reactionusingtetraoxosulphate(IV)acidascatalyst.Theesterifi- cationtemperatureusedwas60C,withmorethan94%conversion ofFFAtomethylester(ME).60Cwasusedduetoitsclosenessto theboilingpointofmethanolatatmosphericpressure[68].94%
conversionwasincloseagreementtothe92%obtainedbyJansri etal.[69]formethylesterproductionfrompalmoil.TheFFAwas reduced toless than 4wt.% in 60min at 60C using 1wt.%/wt H2SO4 ascatalyst.However,Prateepchaukuletal. [70]obtained similarresultof2wt.%in90minat60Cusing3wt.%/wtH2SO4for crude palmoil esterification.Darnokoetal. [71]reportedmore than 95% conversion of palm oil to ME at 60C in 30min at methanoltooilratioof6:1using1wt.%/wtofKOH.Furthermore Sanchezetal.[37];Lopesetal.[38];BeneckeandGarbark[72]had reportedsuccessfulproductionofTOusingvarietiesofvegetable oilsat60C,indicatingsuitability60Cforthisrangeofprocess.
3.8.Statisticalanalyses
3.8.1.One-wayanalysesofvariancefortimeeffect
Table 7 indicates that at 150min, the highest % oil yield obtained was 60.45%, as depicted in Fig. 4. Conversely, the minimumoilyieldwasrecordedat30min.ANOVAwas usedto describetheincreaseinoilyieldwithtime.FromTable8,thenull hypothesis would be excluded (p<0.05), an indication of the statistical difference between the various times. The statistical meanyieldsweresubjectedtoTukey’sposthocHSDanalysis.It wouldsubstantiatethattimevariationhadsignificanteffectonthe oil yield (Table 9). This analysis indicated which possible comparisons between the performance statistical means that were indeed statisticallysignificant. Those having asterisks on their mean difference value were adjudged significant ones (p<0.05)(Table9).Ontheotherhand,those withoutasterisks Table8
ANOVAfortimeeffectonthe%oilyieldofTCK.
SumofSquares Df MeanSquare F Sig.
BetweenGroups 2206.34 4 551.584 50.580 0.000
WithinGroups 218.102 20 10.905
Total 2424.44 24
Table9
Tukeypost-hocanalysisfortimeeffectonthe%oilyieldTCK.
(I)time (J)time Meandifference(I-J) Std.error Sig. 95%ConfidenceInterval
Lowerbound Upperbound
30min 60min 10.97* 2.09 0.00 17.22 4.72
90min 15.90* 2.09 0.00 22.15 9.65
120min 23.89* 2.09 0.00 30.14 17.64
150min 25.99* 2.09 0.00 32.24 19.74
60min 30min 10.97* 2.09 0.00 4.72 17.22
90min 4.93 2.09 0.17 11.18 1.32
120min 12.92* 2.09 0.00 19.17 6.67
150min 15.02* 2.09 0.00 21.27 8.77
90min 30min 15.90* 2.09 0.00 9.65 22.15
60min 4.93 2.09 0.17 1.32 11.18
120min 7.99* 2.09 0.01 14.24 1.74
150min 10.09* 2.09 0.00 16.34 3.84
120min 30min 23.89* 2.09 0.00 17.64 30.14
60min 12.92* 2.09 0.00 6.67 19.17
90min 7.99* 2.09 0.01 1.74 14.24
150min 2.11 2.09 0.85 8.36 4.14
150min 30min 25.99* 2.09 0.00 19.74 32.24
60min 15.02* 2.09 0.00 8.77 21.27
90min 10.09* 2.09 0.00 3.84 16.34
120min 2.11 2.09 0.85 4.14 8.36
Table7
FactorsAffectingTCKSeedOilYield.
Parameter Values
Particlesize(mm)at150min 0.5 1.0 1.5 2.0 2.5
%oilyield 60.45 49.00 46.50 44.00 42.00
Timeat0.5mmparticlesize 30 60 90 120 150
%oilyield 34.90 45.90 50.50 58.30 60.45
Fig.4.Temporalvariationofparticlesizesonyieldattemperatureof55C.
ontheirmeandifferencevalueweretheinsignificantones(Since p0.05). The maximum time for the extraction process could thereforebe chosen to be 120min,for economic reasons.This wouldbeinconsiderationoftheextraenergythatwillbespentfor 30moreminutes(to150min)withmarginalincreaseinyield.
3.8.2.One-wayanalysesofvarianceforparticlesizeeffect
The performance variations in Fig. 5 were also tested for significant difference using ANOVA (Table 10), to confirm the validityoftheresult.TheANOVAindicatesthatthenullhypothesis would beexcluded (p<0.05) asan indication of thestatistical differencebetweenthevariousparticlessizes.Table7showsthat 150minhadthehighest%oilyieldof60.45%fortime-particlesize variations (Fig. 5). The statistical mean yields were subjected subsequentlytoTukey’sposthocHSDanalysis(Table11).Those having asterisks on their mean difference value were the
significantoneswhilethosewithoutasteriskwerethestatistically insignificantones(sincep0.05)asshowninTable11.
For the temperature variations (Fig. 6), one-way ANOVA indicatesthattheeffectoftemperatureonoilyieldwasstatistically insignificance.
4.Conclusion
At the conditions of the experiment, MTCKO that exhibited similar physicochemical properties with conventional mineral transformer fluid could be produced from TCKO using direct purification and transesterification methods. The maximum percentage oilyield fromTCK was60.45%. Theoil yield ofTCK extractedusingn-hexanewasinfluencedbytime,particlesizeand temperature. MTCKO modified by transesterification method relatively conformed better to the stipulated standard, when comparedtothatmodifiedbydirectpurification.However,both havepotentialsfor applicationasdielectric fluidin distribution transformers. The FTIR results indicated the presence of CH, C¼H,NH, and OH functional groups in theMTCKO, which wouldbeexpectedtopromotethebiodegradabilityofthemodified TerminaliacatappakerneloilobtainedfromTCKO.Thetimeand particlesizeseffectswerestatisticallysignificantaccordingtothe ANOVA and Tukey’s posthocHSDanalyses, while temperature effectwasnotstatisticallysignificant.
Fig.5.Temporalvariationoftimeonyieldattemperatureof55C.
Table10
ANOVAforparticlessizeseffectontheoilyieldfromTCK.
SumofSquares Df MeanSquare F Sig.
BetweenGroups 1019.395 4 254.849 26.954 0.000
WithinGroups 189.096 20 9.455
Total 1208.491 24
Table11
Tukeypost-hocanalysisforparticlessizeseffectonthe%oilyieldTCK.
(I)Particlesize (J)Particlesize Meandifference(I-J) Std.error Sig. 95%ConfidenceInterval
LowerBound UpperBound
0.5mm 1.0mm 11.66* 1.94 0.00 5.84 17.48
1.5mm 14.09* 1.94 0.00 8.27 19.91
2.0mm 16.22* 1.94 0.00 10.40 22.04
2.5mm 18.13* 1.94 0.00 12.31 23.95
1.0mm 0.5mm 11.66* 1.94 0.00 17.48 5.84
1.5mm 2.43 1.94 0.72 3.39 8.25
2.0mm 4.57 1.94 0.17 1.25 10.39
2.5mm 6.47* 1.94 0.03 0.65 12.29
1.5mm 0.5mm 14.09* 1.94 0.00 19.91 8.27
1.0mm 2.43 1.94 0.72 8.25 3.39
2.0mm 2.13 1.94 0.81 3.69 7.95
2.5mm 4.04 1.94 0.27 1.78 9.86
2.0mm 0.5mm 16.22* 1.94 0.00 22.04 10.40
1.0mm 4.57 1.94 0.17 10.39 1.25
1.5mm 2.13 1.94 0.81 7.95 3.69
2.5mm 1.90 1.94 0.86 3.92 7.72
2.5mm 0.5mm 18.13* 1.94 0.00 23.95 12.31
1.0mm 6.47* 1.94 0.03 12.29 0.65
1.5mm 4.04 1.94 0.27 9.86 1.78
2.0mm 1.90 1.94 0.86 7.72 3.91
Fig.6.Temporalvariationoftemperatureonyieldatparticlesizeof0.5mm.