buffer capacity based monitoring of algal wastewater treatment pH LievenVanVooren ,PaulLessard ,Jean-PierreOttoy andPeterA.Vanrolleghem 1(cid:3) 2 1 1 DepartmentofAppliedMathematics,BiometricsandProcessControl(BIOMATH), 1 UniversityGent,CoupureLinks653,B-9000Gent,Belgium De´partementdeGe´nieCivil,PavilionPouliot,Universite´ Laval,Que´bec,CanadaG1K7P4 2 Abstract A buffercapacitybasedmethod formultivariatemonitoring of tertiaryalgalwastewater treatment pH systemsispresented.Apilotplantwithalgalbiomassreceivingtheeffluentofacarbonremovingactivated sludge municipal wastewater treatment plant was used for the experimental work. Two types of buffer capacityprofilesareconsidered,firstresultingfromatitrationfromtheactual to 2.5(downtitration), pH pH and second asa result of atitration from 2.5to 11 (uptitration). Theexperiments wereconducted pH pH withalaboratorytitratorandthebuffercapacityprofileswereprocessedbyfittingmathematicalmodelsof buffersystems to them. This allowed to quantify theinorganic carbon (IC)bufferfromthe down titration profiles. IC is of major importance, because it is the only carbon source used by algae. It is shown that the IC concentration derived from the buffer capacity profile gives different process information than a standardalkalinitymeasurement. Theuptitration profilesweresuccessfullyusedforthequantificationof and - , although an exact comparison of the laboratory results and the buffer capacity based + NH4 o PO4 results is difficult, because of the filtration step preceding laboratory analyses and possible interferences caused by unmodeled buffer systems. No filter device was used, which makes it possible to implement thismeasurementmethodinarobustandfieldapplicablesensor. Thecompletemeasurementcycle(down titration followed by up titration and datainterpretation) canbe performedin less than 1 hour, with little manipulation,thatcouldeasilybeautomated. Keywords buffercapacity,algalwastewatertreatment,alkalinity,on-linemonitoring,mathematical modeling pH (cid:3)Authortowhomallcorrespondenceshouldbeaddressed.[Fax:+3292234941,E-mail:[email protected]] 1 Introduction Algal wastewater treatment: Traditionally, the aims of wastewater treatment have been to reduce the con- centrationoforganicmatterandthenumberofpathogens.However,inmanytreatmentplantstypicalprimary andsecondarytreatmentscannotmeettheincreasinglystringentrequirementsofwaterpollutioncontrol. Al- though secondary effluents may contain low levels of and , they still contain high levels of in- BOD5 COD organic nutrients ( , , ). These nutrients are primarily responsible for eutrophication of the + (cid:0) 3(cid:0) NH4 NO3 PO4 receivingwaters[19][27]. Nowadaysmoreandmoreplantsadaptand/orextendtheirsecondarytreatmentfacilitiestowards“nutri- ent removal capacity” [2] [17] [25], but even effluents with low concentrations of thesepollutants can have a negativeimpactontheenvironmentiftheyaredischargedinhighvolume,especiallyifthereceivingwateris arathersmallriver. Tertiarytreatmentisbecominganecessitytominimizetheharmfulimpactofeffluentson theenvironment. Besidesbiologicalnitrogenandphosphorusremoval,numerousphysico-chemicaltreatmentssuchaschem- ical coagulation, breakpoint chlorination, -stripping, reverse osmosis and filtration have been proposed NH3 for removing nitrogen and phosphorus. However, due to costs and operational practices, thesetechnologies arenotimplementableformunicipalwastewatertreatmentofsmallcommunities[20]. As an alternative, tertiary treatment involving micro-algae or cyanobacteria is shown to be effective in nitrogenandphosphorusremoval[20][19]. Theprocessisquiteattractivebecauseofitscapacitytotransform wastes into useful biomass using sunlight as an energy source. However these algal treatment systems also havesomedrawbacks: Thegenerationtimeoftheorganismsinvolvedislongcomparedtobacteria;thereisa necessityforsunlight;theactivepartofthebiomassissmallandharvestingofbiomassisdifficult[27][7]. Inrecentyears,researchhasbeenconductedtoovercometheproblemsrelatedtoalgaltertiarywastewater treatment, and to optimize the treatment process [20] [18] [34] [38] [8]. The harvesting problem of the algal biomasshasbeensolvedbyusingorganismslikethecyanobacteriumPhormidiumbohneri,whichhastheability toformflocksthatsettle,thusfacilitating harvest[30]. Algal wastewater treatment processes often are discontinuous, of SBR (sequencing batch reactor) type, where one treatment cycle consists of a filling phase, an aeration phase, a settling phase and an effluent dis- charge phase. Such processes need continuous monitoring of the most important process variables, in order toobtainanoptimalnutrientremovalcapacityatminimalcosts. Especiallywhenonerealizesthatnutrientre- moval processesareoftenrelatedtophasesofluxuryuptake(e.g. phosphates)andphasesofnutrientrelease, 2 theusefulnessofon-linemeasurementsisobvious. Alkalinity related to algal processes: The alkalinity of a water sample is defined as the amount of acid necessarytodecreasethe toapredefinedendvalue. Twoendvaluesareconsidered[14]: pH 1. End value 8.3 (only relevant if the sample is higher than 8.3): this is called the phenolphtalein or pH carbonate alkalinity (symbol C), and is a measure of the amount of strong bases, carbonates and alkali presentinthesample. 2. Endvalue4.5(onlyrelevantifthesample ishigherthan4.5): thisiscalledthetotalalkalinity(symbol pH T),andisameasureoftheamountofstrongandweakbases(likebicarbonates). Thealkalinitycanbemeasuredinthelaboratoryusingarecipientofknownvolumeofsampleandaburette containingastrongacid(e.g. HCl0.02N).Itcanbeexpressedin orin . (cid:0)1 (cid:0)1 meql mgCaCO3l Alkalinityisapopularmeasurementusedintheprocesscontrolofdiversewatertreatmentprocesses,e.g. in activated sludgetreatment[28] [40] [29] or in anaerobic wastewatertreatment[9] [5]. In practice alkalinity isoftenusedasapracticalmeasureoftheamountofcarbonatesandbicarbonatesinthewater. Therearesome particularremarksconcerningtheuseofalkalinityas“(bi)carbonate” estimator: In cases the sample is loaded with weak acid buffering systems other than the subsystem (like (cid:15) CO2 ammonium,phosphatesandVFA’s),thesecomponentswillbeincludedinthealkalinityandwillsuggest thatmore(bi)carbonatesarepresentthaninreality[6]. For the meter method, it is advised to choose the end value in function of the alkalinity (end-value (cid:15) pH rangingfrom 4.3 to 4.9 forhightolowalkalinity respectively)[39]. Forcolorindicator methods, pH pH likethebromcresolmethod,the value4.5correspondswiththeinflectionpointonthetitrationcurve pH where the color indicator will switch to its other state. In case the inflection point is not so sharp, or the solution contains strong coloring substances, the color switch will be difficult to observe. For this reasons,the metermethodismostlypreferredabovetheoldercolorindicatormethods. pH Intheapplication area ofwastewatertreatment,modifiedorextendedmethodsofthealkalinity measure- menthavebeendeveloped.Examplesofextendedmethodswheremorethan1endvalueinthetitrationprofile areconsideredareathree-pointmethod[1]andafive-pointmethod[22]forthedeterminationofbicarbonate and VFA. Methods using more than 5 points for bicarbonate and VFA estimation have also been developed 3 [3] [26]. Toeliminate theeffect ofinterfering bufferswhenoneis onlyinterestedin bicarbonate content,a pH method was developed where the is strippedfrom the sample by addition of strong acid. The stripped CO2 can be directly measured by the volumetric [16] [15] or pressure [9] build-up of . The described CO2 CO2 methods have been used in the application area of anaerobic treatment, where high concentrations of both bicarbonate and VFA’s are present[5] [6]. Anoverview of measurementmethodsfor and bicarbonate is CO2 givenin[10][24]. Alkalinity related techniques for bicarbonate estimation are well developedin applications where is CO2 continuously generated by micro-organisms degrading organics. In these areas possible interferences with other buffering components are negligible. In the area of algal treatment, on the other hand, is con- pH CO2 sumed by the algae, resulting in relatively low or even limiting concentrations of bicarbonate in the water samples. Inorganic carbon (IC) is of major importance, because it is the only carbon source used by algae. AmongthedifferentformsofIC( , and ),onlythetwofirstonesaretakenupbythealgal (cid:0) 2(cid:0) CO2(aq) HCO3 CO3 biomass, and thus useful as a carbon substrate [11] [13] [36]. The preference for or uptake (cid:0) CO2(aq) HCO3 is dependent [42] and an adaptation period to switch between the and transport system (cid:0) pH CO2(aq) HCO3 may be required [11]. The importance of the IC buffer system in relation to photosynthetic algal growth is widely describedin literature [31] [35] [23] [37]. Therefore, the usefulnessof quantifying the IC buffer for al- gal wastewater treatment is twofold. First, a precise knowledgeof the inorganic carbon content can indicate whether the IC becomes limiting [23], and whether IC supply is needed [31] [37]. Second, because IC is the only carbon source used by the algae, the rate of IC consumption under light conditions can be used as a processandcontrolvariable,e.g. indicatingapossiblereactorfailureiftheICconsumptionistoolow. Theoretical Model for Buffer Capacity Curves pH In generalthe buffer action of a systemis definedby itsresistanceagainst changeswhenacid ( ) or pH pH CA base( )areadded. Thesymbolforbuffercapacityis or (bufferindex)[33],anditsunitis . (cid:0)1 (cid:0)1 CB (cid:12) (cid:25) meql pH dCB dCA (i) (cid:12) = =(cid:0) dpH dpH The origin of the buffer capacity are conjugated acid/base systems. A weak monoprotic acid ( ) (cid:0)1 Ca mol l dissolvedinwaterwilldissociateuntilanequilibriumbetweentheacid( )andthebase( )formisestab- (cid:0) HA A lished: (ii) + (cid:0) HA*) H +A 4 Amathematical representationofthischemical equilibriumisbasedonthreeequations: amassbalance,a dissociationequationandanelectro-neutralityequation. First,amassbalancefor canbewritten: HA (iii) (cid:0) Ca =[HA]+[A ] Second,theequilibriumisdrivenbyitsdissociationconstant ordissociationexponent [33]: Ka pKa [H+][A(cid:0)] and (iv) Ka = pKa =(cid:0)logKa [HA] The valueisafunctionofthethermodynamic ,thetemperatureandtheionicstrength[10][32]. Water 0 pKa pKa itself can be consideredtobe a weakmonoproticacid, with concentration = 55.5 and = 15.74 (cid:0)1 Cw moll pKw (atT= ),thusasimilarmassbalanceandequilibriumequationcanbewritten: o 25 C (v) (cid:0) Cw =[H2O]+[OH ] [H+][OH(cid:0)] and (vi) Kw = pKw =(cid:0)logKw [H2O] The influence of the water buffer on the experimental buffer capacity curve is only significant at or pH < 5 [43]. In a water sample containing several buffering components, one can write that the total buffer pH > 9 capacity is equal to the buffer capacity of the water itself summed with the buffer capacity of the (cid:12) (cid:12)w (cid:12)i differentcomponents inthesample: i n (vii) (cid:12) = (cid:12)w + (cid:12)i i=1 Consider a water sample with 1 monoprotic buffering coXmponent (HA) at low , to which a small volume pH ofstrongbase(e.g. NaOH) witha concentration ( )isadded. Third,theelectro-neutralityhastobe (cid:0)1 CB moll maintained: (viii) + + (cid:0) (cid:0) [Na ]+[H ]=[OH ]+[A ] Basedonthiselectro-neutralityequationonecanwriteforaninfinitesimaladditionof : CB (ix) dCB =(cid:0)d[H+]+d[OH(cid:0)]+d[A(cid:0)] orintermsofbuffercapacity : (cid:12) dCB d[H+] d[OH(cid:0)] d[A(cid:0)] (x) (cid:12) = =(cid:0) + + dpH dpH dpH dpH After substitutionof by , followed by differentiation (using ) and using the pH (cid:0)log[H+] d(logu) = u1 loge du massbalances(iii)and(v),togetherwiththeequilibriumequations(iv)and(vi),onecanrewrite(x): + [H+] [H+] (xi) (cid:12) =2:303 [H ]+CwKw +CaKa + 2 + 2 ([H ]+Kw) ([H ]+Ka) ! 5 If differentmonoproticweakacids are presentin thesample,theequationfor can beextendedwithsimilar (cid:12) additionalterms. Forpolyprotonicacidstheadditionaltermsaresomewhatmorecomplex. Foradiproticacid withconcentration inwater: Cb + [H+] [H+]2+4Kb2[H+]+Kb1Kb2 (xii) (cid:12) =2:303 [H ]+CwKw + 2 +CbKb1 + 2 + 2 ([H ]+Kw) ([H ] +Kb1[H ]+Kb1Kb2) ! Andforatriproticacidwithconcentration inwater: Cc + + [H ] (cid:12) =2:303 [H ]+CwKw + 2 ([H ]+Kw) [H+]4+4Kc2[H+]3 +(Kc1 +9Kc3)Kc2[H+]2 +(4[H+]+Kc2)Kc1Kc2Kc3 (xiii) +CcKc1 + 3 + 2 + 2 ([H ] +Kc1[H ] +Kc1Kc2[H ]+Kc1Kc2Kc3) ! Ageneralequationcanbewrittenforthebuffercapacityofasamplecontaining monoprotic, diproticand l m triproticweakacids: n + [H+] l m n (xiv) (cid:12) = 2:3030[H ]+CwKw([H+]+Kw)2 + (term1)i+ (term2)j + (term3)k1 i j k X X X Theobjectiveofthi@spaperistodevelopa buffercapacitybasedsensor,capableofextractiAngconcentra- pH tions of buffering components from on-line measured titration curves coming from the influent, effluent pH and reactor content of an algal pilot plant. The information that can be extracted from the buffer capacity profiles using mathematical modeling is evaluated towards its usefulness related to tertiary algal treatment processes. Therelationshipbetweenthealkalinity andtheIC buffer capacity will bediscussed,aswell as the ammonium andortho-phosphateassessmentfromthebuffercapacityprofile. Inthelattercase,IC, and + NH4 - will be incorporated in equation (xiv) as diprotic, monoprotic and triprotic weak acids respectively. o PO4 The results obtained with the buffer capacity sensor seem promising as input to develop a control strategy, althoughthisisnottheobjectiveofthispaper. Materials and Methods Thealgalpilotplant: Apilotreactorwithalgal biomass,installedattheoutletoftheactivatedsludgeplant ValcartiernearQuebeccity(Canada),wasusedforthesamplecollectionbetweenJuly22andAugust12,1997. The activated sludgeplant was treating domestic wastewaterof the military base Valcartier, and received an influent flow of around 2500 . Theplant was operatingin partially nitrifying mode and the secondary 3 (cid:0)1 m d effluentnitrogenwaspresentintwoforms, and . + (cid:0) NH4 NO3 6 The algal bioreactor with a triangular cross-section had a volume of 6 and combined aeration and 3 m mixing was obtained by means of a perforated flexible tube at the bottom of the reactor. This pilot reactor was in use for 2 years and its function was to further decrease the nutrient concentrations of the secondary effluent. The reactor was usedin batch mode, in cycles of 1 or 2 days. A cycle consistedof a filling phase (in themorning),anaerationphaseof12or36hours,asedimentationphase(atnight)toallowthealgaetosettle down, and a decantation phase (in the morning) to remove an upperliquid layer of around 75 % of the total reactorvolume. Thebiomassconcentrationinthereactorwaskeptbetween100and600 . (cid:0)1 mgDW l Sampling and laboratory measurements: Light, , DO and temperaturewere monitoredcontinuouslyin pH thealgalpilotplant. Threedifferentkindsofsamplesweretakenforfurtheranalysisinthelaboratory: TheEffluentoftheValCartierplant(samplecodeEVC),beingtheinfluentofthealgalpilotreactor; (cid:15) ThecontentoftheAlgalPilot(samplecodeAP),3hoursafterthereactorwasfilledandcompletelymixed (cid:15) bytheaeration; TheEffluentoftheAlgalPilotplant(samplecodeEAP). (cid:15) The EVC and EAP samples were taken with a peristaltic pump in the inlet and outlet of the pilot plant during filling and decantation phasesrespectively. The AP samples were taken as manual grab samples. All sampleswerestoredimmediatelyinthefridge( )andprocessedinthelaboratorywithin1day. o 4 C Prior to the laboratory analyses ( , , - ), all samples were filtered on a Whatman 934 AH + (cid:0) NH4 NO3 o PO4 filter, with a pore size of 1.5 . The analyses were conducted according to Standard Methods [14]. For the (cid:22)m totalalkalinitymeasurementsinthelaboratory,the metermethodwasused[14]. pH Titration curves: Titrations were performed with a Titrino 716 titrator (Metrohm, Switzerland). Data ac- quisition was performed with a 386 PC connected to the titrator. The titrated sample volume was 100 , ml and titration took place in a completely closed magnetically stirred titration vessel containing a scrub- CO2 ber (bowed glass tube with soda lime pellets) to prevent entrance of from the air. The headspace of the CO2 titrationvesselwas150 . Priortotitration,therefrigeratedsampleswereequilibratedto withawarm o ml 25 C waterbath,andtheywereimmediatelytitratedatroomtemperature,whichvariedbetween20and . Two o 25 C differenttypesoftitrationprofileswerecollected: 1. Downtitrationprofile: Thesampleassuchwastitratedwith0.1NHClfromtheactual to 2.5using pH pH amonotonicendpointtitration(MET)method[21],withfixedstepsof0.05 . Thistypeoftitrationwas ml 7 usedforthedeterminationoftheICbuffer. 2. Uptitrationprofile: Thedowntitratedsamplewas stronglyagitatedwitha magnetic stirrerfor15 min- utes to remove all , while the titration vessel was open to the air. Next, the vessel was closed and CO2 the sample was titrated with 0.1 N NaOH to 11 using a dynamic endpoint titration (DET) method pH [21]. ThedifferencewiththeMETmethodisthattheDETmethoddoesnottakefixedvolumesteps,but variablevolumesteps(smallversusbigstepswhenthebuffercapacityisloworhighrespectively). This typeoftitrationwasusedforthedeterminationofammoniumandortho-phosphate. Data processing software: Software developed in C++ was used for the calculation of the buffer capacity profiles. Such a profile is the inverse of the first derivative of the titration profile. The calculation of the derivative was performed using a parabolic regression in a moving window of 5 experimental data points of the titration curve. No further smoothing algorithms were necessary to obtain a smooth buffer capacity profile. The same software was used to fit several mathematical models to the calculated buffer capacity curve. For the parameter optimization, the PRAXIS algorithm [4], which is freely available on the internet, wasimplementedinthesoftware. Themathematical modelspecificationsweredefinedinaspecialinputfile, read by the program. The software was running on an INDIGO workstation(Silicon Graphics, U.S.) and the dataprocessingof1experimentaldatafileneededapproximately1minute. Mathematicalmodels: Themathematicalmodelsusedwereallbasedonthegeneralbuffercapacityequation (xiv). Foreachmodelused,thefollowingconsiderationsweretakenintoaccount: 1. Whatisthe intervalusedfordataprocessing. pH 2. Which bufferingcomponentshavetobeincludedinthemodel. pH 3. Whataretheinitialguessesfortheconcentrationsand valuesofthedifferentbufferingcomponents. pKa 4. Which concentrations and values are to be adjusted in order to fit the simulation model to the pKa experimental data. The parameters that are to be estimated are specified with a lower and upper limit (constrained optimization). The reason for optimizing values instead of keeping them fixed at the pKa oranothervalueistotakeintoaccounttheresidualeffectsoftemperature,ionicstrengthandpossi- 0 pKa bleelectrodeerrorsonthepositionofthe valuesintheexperimentalbuffercapacitycurve. pKa 8 Forthedowntitrationsasimulationintervalbetween 4and 7.3wastaken. Eveniftheinitial of pH pH pH thesamplewasmuchhigherthan7.3,itwasnotworthwhiletoextendthesimulationintervalandincorporate ortho-phosphateorammoniuminthemodel,becausethemagnitudeoftheICbufferwouldinterferewiththe othersmallerbuffersystems. Fortheuptitrations,asimulationintervalbetween 4and 10waschosen. pH pH In the software, concentrations can be specified in or . More detailed model specifications for (cid:0)1 (cid:0)1 moll mg l thedownanduptitrationsareshowninTable1andTable2respectively. [Table1abouthere.] [Table2abouthere.] The “soap” term in the models stands for a range of components that all have a buffer capacity in the range of 4 to5.5, e.g. numerousorganic acids. The “blank1” termstandsfor an unknownbuffering component pH around 10,foundtobepresentinalmostall samplesanalyzed. Thereasonforincorporatingthe“carbon” pH componentin theuptitrationmodel,despitethefact thatthesampleis priorlymade free,isthatduring CO2 up titration the initially present in the headspace above the vessel can enter the sample and typically CO2 accountsfor0.07 . (cid:0)1 meqIC l Furthermore,oncethetotalconcentration (whichis )isdeterminedusing 2(cid:0) (cid:0) CIC [CO3 ]+[HCO3 ]+[CO2(aq)] the simulation model, the partitioning of these 3 forms can be calculated as function of the actual [39]. pH Using the mass balance and the 3 equilibrium equations for IC, the concentrations , (cid:0) and [CO2(aq)] [HCO3 ] 2(cid:0) are given by equations (xv), (xvi) and (xvii) respectively. Details on obtaining these equations can be [CO3 ] foundin[33][39]. (cid:0)2pH (xv) 10 [CO2(aq)]= (cid:0)2pH (cid:0)(pH+pKa1) (cid:0)(pKa1+pKa2)CIC 10 +10 +10 (cid:0) 10(cid:0)2pH +10(cid:0)(pKa1+pKa2) (xvi) [HCO3 ]= 1(cid:0) (cid:0)2pH (cid:0)(pH+pKa1) (cid:0)(pKa1+pKa2) CIC 10 +10 +10 ! (cid:0)2pH (cid:0)(pH+pKa1) (xvii) [CO32(cid:0)]= 1(cid:0) (cid:0)2pH 10 (cid:0)(pH++p1K0a1) (cid:0)(pKa1+pKa2) CIC 10 +10 +10 ! Importanttoremarkisthatthelatter3equationsareonlyvalidinacompletelyclosedandequilibratedsystem. Asthealgalreactorisanopensystem,theconcentrationof isalsodrivenbyHenry’slaw[32]: CO2(aq) (xviii) [CO2(aq)]= KH (cid:2)pCO2(air) With (at )and ,equation(xviii)leadstoanair/liquid (cid:0)2 (cid:0)1 o (cid:0)2 KH =3:410 atm M T =25 C pCO2(air) =310 atm equilibriumconcentrationof or0.012 ,whichisaverylowvalue. Furthermore, (cid:0)5 (cid:0)1 [CO2(aq)] =1:210 M meql 9 onehastotakeintoaccountthatabiologicalactivesystemlikeandalgalsystemisquiteseldominequilibrium withtheatmosphere,socarefuluseofequation(xviii) isadvised. Results and Discussion Alkalinity and IC buffer capacity: A random sample from the algal pilot reactor AP, taken in the morning ofa sunnydaywasanalyzed withthetitratorandinthelaboratory. Afilteredandanunfilteredsamplewere titratedfrom theactual to 2.5 (Figure1). Bothsamples werealso analyzed onthetotal alkalinity with pH pH twostandardmethods. [Figure1abouthere.] From the titration curves, the correspondingbuffer capacity curves (Figure 2) were calculated and the simu- lation model(equation (xiv) and Table 1) was usedtoquantify theconcentration of theIC buffer. Thebestfit simulationresultfortheunfilteredandfilteredsampleisgiveninFigure2. [Figure2abouthere.] Forthetotalalkalinity determination,theend-point valueforthe metermethodwassetat 4.5, pH pH pH accordingtothestandardmethod[14]. ThismethodwasalsoappliedtothetitrationcurvesofFigure1,where aninterpolationbetween2successivedatapointswasusedtomakeareadingoftheamountofacidnecessary to bring the to 8.3 and 4.5 respectively. For comparison reasons, the simulation results to quantify the IC pH bufferwerealsoexpressedas . (cid:0)1 meql TheresultsofthealkalinitymeasurementsandthesimulationresultsforICarepresentedinTable3. [Table3abouthere.] The precision of the total alkalinity measurements (both the meter method and from the titration curve), pH expressed as relative standard deviation (r.s.d.), was between 1 and 2 %. The r.s.d. of the simulated IC con- centrationbasedondowntitrationprofileswas2%. Onecannoticethattheresultsforthefilteredsampleare onlyabout7%lowerthantheunfilteredsample. ThismeansthatthealkalinityorICbuffercapacityismainly relatedtothesolublephaseandnottothealgalbiomassphase. Thetotalalkalinitywiththestandard meter pH methodwasnotsignificantlydifferentfromthetotalalkalinitycalculatedfromthetitrationcurvesinFigure1 (t-test; =5%). AninterestingfeatureofthisexperimentisfoundwhencomparingtheICcontentfoundwith (cid:11) 10
Description: