Engineered Struvite Precipitation: Impacts of Component-Ion Molar Ratios and Ph

Missouri University of Science and Technology Scholars' Mine

Civil, Architectural and Environmental Civil, Architectural and Environmental Engineering Faculty Research & Creative Works Engineering

01 Jan 2005

Engineered Struvite Precipitation: Impacts of Component-Ion Molar Ratios and pH

Jun Wang

Joel Gerard Burken Missouri University of Science and Technology, [email protected]

Xiaoqi Zhang

Rao Y. Surampalli

Follow this and additional works at: https://scholarsmine.mst.edu/civarc_enveng_facwork

Part of the Civil Engineering Commons

Recommended Citation J. Wang et al., "Engineered Struvite Precipitation: Impacts of Component-Ion Molar Ratios and pH," Journal of Environmental Engineering, American Society of Civil Engineers (ASCE), Jan 2005. The definitive version is available at https://doi.org/10.1061/(ASCE)0733-9372(2005)131:10(1433)

This Article - Journal is brought to you for free and open access by Scholars' Mine. It has been accepted for inclusion in Civil, Architectural and Environmental Engineering Faculty Research & Creative Works by an authorized administrator of Scholars' Mine. This work is protected by U. S. Copyright Law. Unauthorized use including reproduction for redistribution requires the permission of the copyright holder. For more information, please contact [email protected]. EngineeredStruvite Precipitation: lmpacts of Gomponent-lon MolarRatios and pH

JunWangl;Joel G. Burken,M.ASCE2; Xiaoqi (Jackie) Zhang3; and RaoSurampalli, lU.nSCEo

Abstract: StruviteFecipitation has the potentialfor rcmovingand recovering phosphorus from agriculturalwastewater streams, such as concentratedanimal feeding operations wastewater. However, impacts of anticipatedcomponent-ion molar ratiosand potentiallyinter- fering ions are unknownas arethe compoundingpH relationshipwith respectto all potentialcomplexes. This researchexpe.imentally investigatesand mathematically models these facto.s. Emphasis is placedupon the compositionof formeddeposits and model validation with experimentaldata. Results show that calciumis a rnajorinterfering ion affectingthe depositcomposition, de4reasing struvite puity. X-ray diffraction(XRD) andscanning electron microscopy * energydispersive spectromefy were used to studythe depositstructure and elementalcomposition. Results rcvealed that lhe precipitatesformed at a pH of 8.7 have regularcrystal shape,and XRD analysis confirmedthat the precipitatesare high-purity struvite. Higher pH (> I0) leadsto the formationof amorphousprecipitate and decreases the struvitepurity in the deposits.To maximizestruvite purity, the ratio of Ca to P shouldbe lessthan 0.5 and the pH near8.7. DOf: 10.106U(ASCE)O733-937 2(2ffi5) t3rtt0(1433\ CE Dgtabase subiect headings: Agriculturalwastes; Water pollution; Wastewater; Nutdents; Phosphorus; pH; CalciurniAmmonia,

Introduction At the current utilization rate, the phosphatemineralogy which can be utilized economicallyis estimatedto provide the necessary Phosphorusis generallyregarded as the primary nutrient respon- P supply for as little as 50 years(Driver et al. 1999).Dual benefits sible for eutrophication of rivers and lakes (Momberg and can be achieved if phosphoruscan be recovered and recycled Oellermann 1992). To combat increased eutrophication, the from wastewaterstreams, as not only will nutrient enrichmentof United States and numerous European countries have promul- rivers be prevented,but conservationof a finite resourcewill be gated strict wastewaterdischarge standards on P levels to protect accomplished. water bodies. Unregulatedand nonpoint sourcessuch as agricr.rl- tural wastes still reach surface waters. Of P used in livestock Phosphorus Removal Technologies feeding, an estimated70Va is directly excretedas waste (Greaves The recognizedphosphorus removal technologiesinclude chemi- et al. 1999).While excessP is problematic,phosphorous is also cal precipitation,biological phosphorusremoval, crystallization, a limited resource.Phosphorus is a widely used raw material tertiary filtration, and ion exchange(Morse et al. 1998).Most of in agriculture and industry, primarily as one of the main crop these processesproduce wastes which need to be landfilled or fertilizers but also in animal feeding and food supplements. incinerated, the ethos of sustainability,however, makes these According to an investigationby the EuropeanChemical Industry options unattractive(Durrant et al. 1999). Biological P removal Council in 1998, over 150 million tons of phosphorus are is mainly used for low P concentratedmunicipal wastewater. extractedand processedper year globally, and agricultureuse as Elusive mechanisticunderstanding makes thesebiological meth- fertilizers and feed concentratesaccounts for 857o(CEEP 1998). ods tricky. Crystallization processesstand out becausethey not

'Water only achievehigh P removal, but also recoverP from wastewater and WastewaterEngineer, Crawford, Murphy & Tilly, Inc., as useful products, including struvite, calcium phosphate,and 600 North Dr.. Commons Suite 107. Aurora. IL 60563. E-mail: hydroxyapatite.Chemical makeup and formation stoichiometry [email protected] 'Associate are shown in Table 1 and Eqs. (l)-(3).The recoveryof phos- Professor,Dept. of Civil, Architecturaland Environmental phorus by crystallization has been reported to reduce sludge Engineering, Univ. of Missouri-Rolla, Rolla, MO 65401. E-mail: [email protected] volume under specific conditions by up to 49Vo compared to 3Assistant Professor,Dept. of Civil and Environmental Engineering, chemical phosphorusremoval (Woods et al. 1999). Univ. of Massachusetts Lowell, Lowell, MA 01854. E-mail: -, [email protected] Mgt* * NHo*+ HPO.2-+ 6H2O MgNH+POa.6H2O|+ H* (l) 'USEPA Region VII,901 North 5th Ave., KansasCity, KS 66101. E-mail: [email protected] 3PO43-+ 5Ca2++ OH- --, Cas(POo):OH (2) Note. Discussionopen until March l, 2006. Separatediscussions must be submitted for individual papers.To extend the closing date by one month, a written requestmust be filed with the ASCE Managing Editor. 2PO4?-+ 3Ca2*-- Ca:(PO+)z (3) The manuscript for this paper was submitted for review and possible publication on August 4, 20O4; approved on December 20, 2004. This Struvite and hydroxyapatitecan be used in agricultureas fertiliz- paper is part of the Journal of Environmental Engineering, Vol. 13l, ers; struvite, however, is preferred for numerous reasons.First, No. 10, October l, 2005. OASCE. ISSN 0733-937212005110-1433-14401 nutrients are releasedat a slower rate compared to other fertiliz- s25.00. ers. Plantscan take up the nutrientsbefore being rapidly leached,

JOURNALOF ENVIRONMENTALENGINEERING @ ASCE/ OCTOBER2OO5 / 1433 Table 1 . Characteristicsof Phosphorus-ContainingCrystallization Process Products

Molecular weight Phosphoruscontent Nitrogen content Deposit Molecular formula (g/mol) (% mass) (% mass)

Struvite MgNHaPOa.6H2O 245.41 12.6 5.7 Calcium phosphate Ca3(POa)2 3r0. r8 19.9 N/A Hydroxyapatite Cas(POa)rOH 502.31 18.5 N/A Note: N/A:not available.

and less frequent application is therefore required (Munch and from 8.0 to 10.7 (Stumm and Morgan 1970; Booram et al. l9l5: Barr 2001). Second,the impuritiescaused by heavy metals in the Momberg and Oellermann 19921'Buchanan et al. 1994; Ohlinger recovered strLrviteare two or three orders of magnitude lower et al. 1998;Booker et al. 1999). than that of commercial phosphatefertilizers (Brett et al. 1997). The component-ionmolar ratios also have great impact on the Lastly, the essentialnutrients P, N, and Mg are applied simulta- compositionof the precipitate.Previous researchrevealed that a neously with essentiallyno unnecessarycomponents in the fertil- stoichiometricexcess of ammonium (30-80 ppm) will help drive izer. Struvite contains both magnesium and phosphate,which the reactionto form relatively pure struvite (Stratful et al.200l), plantsrequire in a ratio of 3:2 Mg:P (Taiz and Zeiger l99l). while excess magnesium decreasesstruvite purity (Demeestere The processesalso have physical advantagesin addition to the et al. 2001). Recent research predicts phosphate removal is chemical advantages.The footprint of a struvite crystallization optimal when the molar ratio of magnesiumto total phosphorusis reactor is considerablysmaller than that of biological P removal 1.6 (Burns et al.200l). In a similar research,a magnesiumto infrastructure;and the processhas fewer problematicoperational phosphorusratio of 1.3 was an ideal environment for struvite concerns. Struvite crystallization may have the potential to formation (Munch and Barr 2001). cheaplyremove P and ammonium-nitrogenfrom wastewaterrela- Struvite solubility product is a critical parameterin chemical tive to biological P removal and standard nitrification- equilibrium calculationand processmodeling. Depending on the denitrificationtechniques or causticstripping (Webb et al. 1995). complexes formed in a solution and the chemical speciation The struvite crystallizationprocess has been favorably applied to determined,a range of solubility product constantshave been relatively high phosphateconcentrations ranging from 100 to published (Buchananet al. 1994).A widely used pK"o value is 2OOmglL (Moriyama et al. 2001), and the processcan efficiently 12.6 (Taylor et al. 1963: Stumm and Morgan 1970; Burns and operate at higher concentrations.Some researchershave sug- Finlayson 1982; Loewenthal et al. 19941'Aage et al. 1991; gested that struvite sales could cover the operation cost of Ohlinger et al. 1998).Other researchaccounting for activity have the plant (Unitika-Ltd. 1994). The market value of the struvite found pK.o valuesnear 13.2(Burns and Finlayson1982; Ohlinger precipitateis not currently established. et al. 1998).

Factors in Struvite Crystallization Process Confined Animal Feeding Operations Wastewater

To produce pure struvite deposit, many factors need to be care- Confined animal feeding operations(CAFO) wastewatercontrib- fully controlled. Factors include component-ionratios, potential utes P to surfacewater bodies.Although anaerobiclagoons have interfering ions, and pH values.Calcium is a common cation in been widely applied to treat CAFO wastewater in the United wastewaterand is also an interfering ion to struvite formation. States, the reactive P concentration in the effluent, however, Struvite formation is affected by the interactionof calcium and can be extremelyhigh. Over 2O0 mglL P was measuredin waste- magnesium, depending on their relative concentrationsstruvite waters testedin this study. Lagoon effluent is usually applied as formation can be inhibited as calcium phosphateprecipitates form irrigation water increasing the possibility that excessive P is (Momberg and Oellermann 1992).The reaction of POa with Ca flushed into surface waters during wet seasons(Greaves et al. can form a number of related products depending on reaction 1999). Therefore effective removal of P from CAFO wastewater conditions(Greaves et al. 1999).By modifying the chemicalcom- has direct and substantialvalue to surfacewater quality.Although position of wastewater.increasing Mg/Ca ratios,efficient struvite researchershave studiedthe applicationof struvite crystallization formation can be achieved(Battistoni et al. 2000). Recent work on animal wastewater (Webb and Ho 19921'Wrigley et al. 1992; has also shown that physical parameterscan be optimized to Webb et al. 1995), understandingand modeling of the impact improve struvite precipitation efficiency. Seeding materials and optimal mixing conditions were shown to achieve maximum removal in less than 30 min, with pure struvite crystalsresulting (Wang et al. 2005). For several different calcium phosphates, Table 2. Initial Concentrationsand Molar Ratiosof Four Major Ions the saturationlevel of phosphatefor a given calcium concentra- in SyntheticWastewater Solutions tion depends strongly on the pH (Stumm and Morgan 1970). Concentration(ppm) Molar ratio Numerous researchershave investigatedthe optimum pH values NHo* Mg:Ca:POa:NHo and concluded that the precipitationof hydroxyapatiteoccurs a[ Solution Mgt* Ca2* POot- pH values above 9.5, whereas effective struvite precipitation r 96 40 190 720 2:0.5:l:20 occurs at pH at 8.0 and above. Under different ratios of Mg/Ca, 2 96 80 190 720 2:l:l:20 struvite or hydroxyapatitecould be formed within the pH range 3 96 160 r90 720 2:2:l:20 of 8-10 (Battistoniet aI.2000). The optimum pH for struvite 4 48 40 r90 720 l :0.5:l:20 publishedby a number of researchersdisplays a range of values, 5 24 40 190 720 0.5:0.5:l:20

1434IJOURNALOF ENVIRONMENTALENGINEERING @ASCE / OCTOBER2OO5 Tabfe 3. ParametersUsed in MINEOL Modelins 0.0045

Parameter Value 0.0040 a 0.0035 pH 6-12 pK.n of struvite ^ 0.0030 12.6 3 Ionic strength(M) 0.r0 E o oozs O *u Temperature ("C) 25 ! o.oozo c o t o.oors \.. r of pH, component-ion molar ratios. and interfering ions on 0.0010 provides CAFO wastewateris still lacking. This study a better 0.0005 understandingof theseaspects. l'- -^' 0.0000 - - - ->=:----<-.- E 9 to 11 12 pH

Materialsand Methods Fig. 2. Component-ionconcentration in effluent of Solution 2 (Mg:Ca:PO4:NHa:2:l:l:20, points representexperimental results, Solution Preparation linesrepresent model predictions)

Chemical compositions of the effluents from two anaerobic lagoonstreating piggery wastesin central Missouri were investi- filteredwith 0.45-pm membranefilter paper (Millipore Corp.). gated. According to the measuredvalue ranges, five synthetic The solids were dried in a desiccatorunder room tenlperaturefor wastewatersolutions were then formulatedto mimic typical mag- 24 h, then weighed and stored in capped5-mL glass vials. Each nesium, calcium, ammonium, and phosphate concentrations dried precipitatesample was weighed with an analyticalbalance, that could be expected in the effluent of anaerobic lagoons. dissolvedwith 5 mL of 1.0N HNOj (ACS, FisherScientific), and Four sampleswere preparedfor each solution, each sample was transferredinto a 0.5 L volumetric flask.The volumetric flask was adjusted to a different pH for reaction. The solutions were filled up to 500 mL with deionized water. The solutions were preparedwith magnesium sulfate heptahydrate(MgSOo.7H2O), mixed with stir bars for I h, and analyzed for Mg, Ca, POa, and dehydrate calcium chloride (CaCl2), ammonium chloride NH4-N concentrationswith the analytical methods described (NH4CI), and ammonium phosphatemonobasic [(NH4)H2PO4]. below. To eliminate the effect of adding chemicals to solution ionic strength, 14.0 g of sodium perchlorate(NaClO+) was added to each I L solution to increasethe background ionic strengthto Analytic Methods grade (Fisher 0.1 M. All chemicalswere ACS Scientific).The Magnesium and calcium concentrationsin solutions were mea- prepared were with room samples mixed stir bars at temperature suredby Atomic Absorption SpectrometricMethods 3500-Mg B for 24 h in The before being used the experiment. initial concen- and 3500-Ca B, respectively(Clesceri et al. 1989). Ammonium trations and molar ratios of four ions are listed in Table 2. concentrationwas measuredby HACH HR, Test'N Tube Method (HACH method 10031), phosphatelevel was assessedby the Precipitate Formation and Dissolution HACH amino acid method (HACH Method 8178).The crystalline phasesin the precipitateswere analyzed with x-ray diffrac- The prepared solutions were transferred into five I L Erlenmeyer tion (XRD) in the Material ResearchCenter at the University of flasks. Each solution was adjusted to the designatedpH with Missouri-Rolla. Both crystalline and amorphous phases in the 1.0 N and 0.I N NaOH solution(ACS grade,Fisher Scientific). precipitatesamples were analyzedwith scanningelectron micro- An Orion 2304 plus digital pH meter was used to monitor solu- scopy (SEM) and energy dispersive spectrometry(EDS) in the tion pH values.The solutionswere mixed continuouslyfor l2 h Electron Microscopy Laboratory at the University of Missouri- on a magneticstir plate.After l2h of reaction,the solutionswere Rolla to determinethe elementalcomposition.

0.@45 0.0025 0.0040

0.0035 0.0020 ^ 0.0030 o = .A .Eo.oozs € o.oors o o a g f, o oozo c o o ol 5 0.0010 o.oors o |r,l\ '_'. 0.0010 \I \.lr 0.0005 \t..- 0.000s

0.0000 0.0000 10 ',t't 8 8 9 12 pH pH

Fig. 1. Component-ionconcentration in effluent of Solution I Fig. 3. Component-ionconcentration in effluent of Solution 5 (Mg:Ca:POa:NHa:0.5:0.5:l:20, (Mg:Ca:POa:NHa : 2 :0 .5 : | :20, points representexperi mental results, points represent experimental linesrepresent model predictions) results,lines represent model predictions)

JOURNALOF ENVIRONMENTALENGINEERING @ ASCE/ OCTOBER2OO5I 1435 1.ZO 1.20

1.00 1.00 i9 o g g .$o.ao .4 0.60 g e (L o- .E 0.60 .s 0.60 o tt '€o g o.lo fi o.+o f c c o o 0.20 0.20

0.00 0.00

1.20 1.20

1.00 1.00 g g E E -$ o.eo .$o.ao e p (L (L .s 0.60 .E 0.60 o t o o 5' e o.4o ff o.eo c c .9 -9 0.20 0.20

0.00 0.00 7.8 8.7 9.2 10.5 7.8 8.7 9.2 10.5 pH pH Legend:Z=NH4 I=Ca l=Mg

Fig. 4. Molar ratios of NHa, Ca, and Mg normalized to POa (i.e., POn=1.9; in deposits formed at various initial component-ion molar concentrations

Modeling Software phosphatesCaaH(POa)3.3H2O and CaHPOa.2H2Othat are often not differentiated from struvite. Therefore determination of the MINEQLI version 4.5, an equilibrium speciationmodel devel- optimum pH enables the formation of the maximum amount oped by Environmental ResearchSoftware, was used to model highest purity of struvite, not just targeting P removal. the struvite precipitation process in this study. Compared with and pH other applicationsfor the similar purpose,MINEQL* has a sig- Model results show that the optimum is near 8.7, as the nificant advantage,which is allowing users to create a personal struviteprecipitation potential reaches maximum while precipita- thermodynamicdatabase. Struvite is not included in the supplied tion of undesired by-products is minimized. Combining the internaldatabase. In this study,a customthermodynamic database typical pH range for struvite precipitationand model prediction, was created for struvite by providing equilibrium constants, enthalpy values, and stoichiometriccoefficients to the database (Table 3). With the custom databaseincluded, the program calcu- lates equilibrium concentrationsof ion speciesbased on the input 1.0 of component-ionconcentrations, using the pK.o, pH, tempera- 0.9 rCalMg=Q.512 ture, and ionic strengthas shown in Table 3. MINEQL* doesnot 0.8 lCalMg=ttZ include kinetic aspects,and comparison of model output and c 0.7 LCalMg=V2 experimentsshould considerthat some precipitatesdo form more ou rapidly period i I and experimental is important. a o.5 I

€(! o.+ ( 0.3 Resultsand Discussion o.2 A O 0.1 Optimum pH 0.0 I 9 10 11 Struvite precipitation has been noted to occur over a wide pH pH range (7-l l) but with varying precipitation potential. In addition, other undesiredby-products may be formed, such as Fig. 5. NH4/PO4molar ratios in the depositsformed with various initialcomponent-ion concentrations brucite [Mg(OH)z], hydroxyapatite[Car(POo),OH], and calcium

1436IJOURNALOF ENVIRONMENTALENGINEERING O ASCE / OCTOEER2OO5 l'rltt | l/flft; lrl); . : tdc; ]El0f/ll tl:lt tl:g I l,f!D' (r. Utr; .t;lfl.Si;. lrls: t.Uf - 19,t ':8d.. ttq) lrc{} l|.tc :'.l,ll '|tnJr r'?s lHill r' Gtiltr mil! ffir ErttE ,--'-:'- lttrtrE. r[f H tr .' . f tl tl,' r:- rtt_ ?!, I ll -, SamPle

'!'- nr- I ll lth standard rrg.|l||lt,|,il|,Jl.'.,,l'l'{iffi,,.l. rt0- urt ro-\.i\-*-',il{*pi*,{l-J,1|fiL"[l''ulrr',t$6dl.I|',Un-"{

t;! l.l lLt lrrf l?.f n-r rr-r lt.l FJf lt.l'h.f ll.&r tt.! .l.t tt:.! .?|:t !:t.f }t., tt.r fLi :91,1 tr-t

Eile: {-1, Dr Dctt: fSfflJlt l{:lt SLup; I'F CtL lir; !.lff SGr, thrgr: t,lt . lf.lt lnryl .l$.nDw letf : 1-rl,rrqllri. r|rJ* ,r:, 3 {ttttlil if,if{$lti tffiil*{lti tfJHiln:, .r-ll{Ullt.

Amorphousphase

Fig. 6. X-ray diffraction spectraof depositsamples formed at pH:8.7 (top) and 10.5 (bottom) overlaid with struvite standard.Deposits formed in Solution 2, initial molar ion ratio. MglcalP=2lll1.

JOURNALOF ENVIRONMENTALENGINEERING O ASCE / OCTOBER2OO511437 Deposit Composition Analysis

The elementalanalysis depicted in Fig. 4 revealedthat calcium concentrationhas substantialimpact on the compositionof deposits. Struvite is the only precipitatethat has NHo* in it under the experimental setting, therefore, when precipitates are dissolved and analyzedNHo* molar concentrationis roughly equal to the struvitemolar concentrationin the initial deposit.When the molar ratio of Ca, Mg, and POo3-was 0.5:2:1, the depositsformed at a pH of 7.8,8.J, and 9.2 had nearly I to I ratios for NHo* and P, Fig. 7. Scanningelectron microscopyimage of deposit formed and the calcium concentrationswere below detection,see Fig. 4. at (a) pH:8.7 and (b) pH:10.5. Depositsformed in Solution 2, These resultssuggest that thesethree sampleswere struvite with initial molarion ratio:MglCalP:Zllll. high purity. When pH was at 10.5,the ratio of NHo* to P dropped significantly to 0.22, and the concentration of calcium rose concurrently. The formation of calcium phosphate precipitates was four operationalpH values,J.8,8.J,9.2, and 10.5,were adopted demonstrated.With an increasedcalcium ratio in the synthetic to experimentallyevaluate the purity of the depositsformed under wastewater,NHo* ratios dropped while calcium ratios rose in different conditions. deposits.When the ratio of Ca, Mg, and NHo* was 2:2:l (Solution 3), calcium precipitatewas formed at a pH as low as 7.8. When the pH was increasedto 10.5, NH4* concentrationis Effluent Composition Analysis essentiallyzero suggestingdeposits were void of struvite. The Comparisonof equilibrium ionic concentrationsfrom experimen- results indicate that when molar concentrationsof Mg and NHo tal resultsand model predictionfor Solutions l, 2, and 5 (Table2) are not limiting for struvite formation, Ca concentrationis the are shown in Figs. l-3. MINEQL predictions tracked experimen- major factor that affects the composition of deposits.When the tal results. P concentrationdropped from 190 mg/L to below CalP ratio is less than 0.5, relatively pure struvite can be 5 mg/L at pH- 10.5for SolutionsI and 2; while for the Solution producedif the pH is less than 9.2. When the Mg/P ratio is less 5, P concentrationin the effluentwas 48.7 mg/L at pH- 10.5with than l, both Mg and Ca concentrations impact the deposit removal efficiency of 747o.For the first two solutions,the sum of composition.Even if CalP is low (0.5), a calcium containing Ca and Mg molar concentrationsis much higher than that of P, deposit is formed at a pH of 8.7 when Mg/P is 0.5. A pH of 8.7 2.5:l and 3:1, respectively.While in the last solution,the ratio is is close to published optimum operational pH values. This only l:1, suggestingthat the ratio l:l is adequatelyhigh to suggeststhat Mg is a critical parameterand should be monitored induce maximum P precipitation from solution, but to attain to optimizestruvite precipitation. high P removal, the Ca and Mg molar concentrationmust be in The ratiosof NHa/POo in the depositswere plotted for various stoichiometricexcess than that of P.The differentCalMg ratios in solutions,Fig. 5, as struvite purity can be roughly estimatedby the first two solutions did not significantly affect P removal, the NHo/POa ratio. A clear trend can be observed:struvite purity however,modeling reveaeledthat ratios impact which precipitates decreasedwhen the calcium concentrations increased. When form. According to the MINEQL results, CaaH(PO4)3.3H2O Ca:Mg was 0.5:2, the maximum struvite purity was 85Vo,while and CaHPO4.2H2O will precipitatepreferentially rather than stru- this ratio dropped to 6lVo when Ca:Mg was l:2, and only 38Vo vite. If the CalP ratio is high, P will be removed as calcium when Ca:Mg was 2:2. For all solutions,the optimum removal was precipitatesand struvite formation will be inhibited. obtainedwhen the pH was between8 and 9, showing that the pH pH is a critical parameterin the process.For all the solutions, of near 8.7 is optimal for high purity struviteprecipitation. In real improved P removal was achievedas pH increasedfrom 7.8 to applications, high purity struvite is preferred to increase the 10.5,see Figs. l-3, even thoughstruvite purity is altered.Model market value without the need for subsequentpurification process results show P removal efficiency will decline as pH increases and operationcost. above I I due to two reasons.First, at higher pH, Mg(OH)2 is The collected precipitateswere analyzedwith XRD to confirm formed preferentially with a lower solubility, pK,p= 10.7. the compositionformed under different conditions.Those formed P-containing precipitatesthat may have been formed can be under pH=8.7 match struvitestandard spectra as shown in Fig. 6. dissolveddue to the low Mg*2 concentration.This trend is more Struvite is the only crystalline phase in the deposit, indicating apparentwhen (Ca+Mg)/P is low becauseof the deficiency high purity struvite with no apparentamorphous phase. For the of cations in the solution phase (Fig. 3). Second, higher pH samplesformed at pH- 10.5, a broad hump centeredat approxi- will deprotonateNHo*. a component-ion of struvite, and NH3 mately 30" is apparent,which is due to amorphousmaterials in volatilization will depletethe aqueousNH:/NH.* [Eq. (a)]. The the sample. Existence of both struvite crystals and amorphous appropriateselection of operationalpH is important in achieving calcium precipitatesis clear.The experimentalresults are consis- high P renroval efficiency. However, the optinlum pH value is fent with model prediction.Samples formed at various conditions wastewatercomposition dependent: different wastewaterstreams were further analyzed with SEM and EDS. The SEM images have different optimum pH values primarily as a function of shown in Fig. 7(a) reveal that the depositsformed have typical Ca:Mg:P:NHo* ratios. The composition of CAFO wastewater struvite trapezoidalshape. EDS analysis also matches the stan- varies greatly; therefore model analysis aids to determine the dard struvite pattern,see Fig. 8. Optical and elemental analyses optimum value for each case. further confirm that the deposits formed under this condition are nearlj'pure s'truv'ite. The solidsformcd at pH=10.5arc amorphous when observedvia SEM as shown in Fig. 7(b). No visible NHo* - = (4) + OH- NH-r + HrO, pkn 9.26 crystals wcrc obscrved undcr SEM analysis. EDS arlso dctcctcd

1438IJOURNALOF ENVIRONMENTALENGINEERING @ASCE / OCTOBER2OO5 ca{a3rtr-rt 1-l-*rlrtrn 6 lt rrr qh- "' n!l!E s.atrr+ 2;;;;;;;;;,;;-" +I t- *t ah i***1,6 dH.trr*€r- stqt.t.BFtu tTt H&r,{9Sb b+ibi infir 0 tr tt l"l roa{t tr

Fig.8. Energy dispersivespecfometry analysisof the depositsformed at (a) pH:8.7; (b) pH=10.5; and (c) pure struvite standard.Deposits formed in Sofution 2, initial molar ion fttio: Ms,lcalP:2llll.

a high concentrationof calcium, which is consistentwith the degradation,organic nitrogen and phosphatewill be converted elementalanalysis results and model predictions. to NHo* and reactive P, including H2PO.-, HPO42-,and POo3-, increasingthe struvitecomponent-ion concentrations. The anaerobic process thereby createsan environment promoting struvite E ng i neeri ng I m pl icati on s precipitation.Component-ion ratios were good for struvite forma- Collection and transportof CAFO wastewaterto centralizedtreat- tion in two lagoons in central Missouri, see Table 4. If the sus- ment plants is not economical or feasible.The combination of pended solids concentrationin the lagoon effluent is too high, anaerobic treatment and struvite crystallization processesmay additional solid removal processescan prevent excessivesolids provide an economicaland feasiblesolution to preventexcessive entering the crystallization reactor and decreasingprecipitated P discharge.In anaerobictreatment, most organic P and Poly-P struvite purity. However, traditional coagulants alum, ferrous degrade to reactive P, and most organic N is biologically con- salts, or lime should not be used becausethey will react with verted to ammonium. The subsequentcrystallization reactor can reactiveP in the solutionand could precipitateas aluminum phos- take advantageof these component ions and prodr.rcestruvite. phate (AIPO4). ferrous phosphate(FePOa), and hydroxylapatite precipitation Struvite is better understood with respect to [Ca1e(POa)o(OH)z].Hydraulic solids removal, dissolved air flota- component-ionratios, optimal pH, and competingions for typical tion, or polymer addition are acceptablesolids removal technolo- CAFO wastewatercompositions. From this new understanding, gies. In the struvite crystallization reactor, magnesium chloride potential applicationscan be considered.CAFO wastewatervar- can be added if required to adjust the Mg/P ratio, and sodium ies significantlyin composition,calcium and eachcomponent-ion hydroxide can be added to control the pH inside the reactor to concentrationsshould be measured to determine whether the promotestruvite precipitation. Aeration of the effluent water from wastewater is suitable for the struvite crystallization process an anaerobicprocess has been shown to raise the pH to above 8 directly or additional processesare needed.Mg should also be by striping CO2. Alternatively, magnesium hydroxide could be closely monitored.If the Mg/P ratio is less than I, Mg should be addedto concurrentlyraise pH and increasemagnesium. Treated added to increaseMg concentrationin the solution. If the Ca/P effluent is dischargedfrom the top of the reactor; settled struvite ratio is higher than 0.5, precipitatesformed will not be pure is collected at the bottom of the reactor. struvite,but P will still be removed. Processes to treat CAFO wastewater can be proposed. First wastewateris treated with anaerobicbiological processes Conclusions to decrease BOD and solids. Throush anaerobic biolosical

This researchinvestigated the potentialof struvite crystallization as a method to remove and recover reactive phosphorus from Table 4. Parameters Measured in Two Anaerobic Swine Lasoons in Central Missouri CAFO wastewater,while concurrentlyremoving ammonium and magnesium.Specific conclusions can be drawn from the research Component Lagoon A Lagoon B completed. Calcium is the major competing ion in struvite pH 6.9-7.3 6.1-7.6 formation, and struvite purity can be greatly decreasedwhen the SCOD (mg/L) 600-l,500 l 30-800 CalMg ratio is higher than l. Identificationof Ca as an interfer- Mg (mg/L) 30.0-64.0 30.0-43.5 ence is particularly important as Ca is a common ion in process Ca (mg/L) 100.5-r10.7 105.4-t20.9 water.MINEQL+ 4.5 was also usedto predict optimum pH and P precipitation NHa (mg/L) 29s.4-630.7 t'72.0-480.0 removal efficiency in the struvite process with reasonableparameter settings added to the commercial package. PO+(mg/L) 64.6-181.7 t2s.8-240.3 When considering competitive precipitation with calcium T-N (mg/L) 637.5-867.0 588.2-676.1 phosphates,the optimal pH of 8.7 was lower than what had T-P (mg/L) ll0.2-31t.4 17s.5-351.7 been previously published when not considering competitive TS (mg/L) r,800-4,000 r,500-2,000 precipitationor struvite purity. The higher pH valueslead to pre-

JOURNALOF ENVIRONMENTALENGINEERING @ASCE / OCTOBER2OO5 / 1439 cipitatesother than struvite, reducing the value of the produced Demeestere,K., Smet, E., Van Langenhove,H., and Galbacs,Z. (2001). precipitate. Previously only P removal and precipitation was "Optimalisation of magnesium ammonium phosphate precipitation targeted, disregarding the specific precipitate formed. Finally, and its applicability to the removal of ammonium." Environ. kchnol., if P removal is the targetedgoal and not struvite purity, magne- 2202). 1419-1428. sium and calcium shouldbe in stoichiometricexcess to phosphate Driver, J., Lijmbach, D., and Steen,I. (1999). "Why recover phosphorus to attain highest phosphateremoval. This researchsuggests with for recycling and how?" Environ. Tec:hnol.,20(7), 651-662. Durrant,A.8., Scrimshaw,M. D., Stratful,I., and Lester,J. N.(1999). a design targeted for each specific CAFO wastewater (swine, "Review of the feasibility of recovering phosphate from wastewater dairy, beef, etc.), struvite precipitationcould be an effective and for use as a raw material by the phosphate industry." Environ. Tech- promising nutrient removal technology. nol., 2O(7),749-758. Greaves,J., Hobbs, P.,Chadwick, D., and Hayganh, P. (1999). "Prospects for the recovery of p2 hosphorus from animal manures: A review." Acknowledgments Environ. Technol., 20(7), 697-708. Loewenthal,R. E., Kornmuller, U. R. C., and Vanheerden,E. P. (1994), This researchwas funded by the United States Environmental "Modelling struvite precipitation in anaerobic treatment systems." Protection Agency (XP-99795901-0). The writers also thank Water Sci.Technol., 30(12), 107-116. Dr. Eric Bohannanat the Material ResearchCenter and Dr. Scott Momberg, G. A., and Oellermann,R. A. (1992). "The removal of phos- Miller at the ElectronMicroscopy Laboratoryat the University of phate by hydroxyapatite and struvite crystallisation in South Africa." -996. Missouri-Rolla for technical assistancein analysis.The writers Water Sci. Technol., 26(5-6), 987 are also grateful for the help offered by Dr. Cynthia Henny, Moriyama, K., Kojima, T., Minawa, Y., Matsumoto, S., and Nakamachi, K. (2001). "Development of artificial seed crystal for crystallization Dr. George Qiang, Stephen Homan, and other studentsat the of calcium phosphate." Environ. Technol., 22, 1245-1252. EnvironmentalResearch Center, University of Missouri-Rolla. Morse, G., Brett, S., Guy, J., and Lester,J. (1998). "Review: Phosphorus removal and recovery technologies." Sci. Total Environ., 212(l), 69-8r. References Munch, 8., and Barr, K, (2001). "Controlled struvite crystallisationfor removing phosphorus from anaerobic digester sidestreams." Aage, H. K., Andersen, B. L., Blom, A., and Jensen, L (1997). WaterRes., 35(l), l5l-159. "The solubility of struvite." J. Radioanal. Nucl. Chem., 223(l-2), Ohlinger, K., Young, T., and Schroeder,E. (1998). "Predicting struvite 213-215. formation in digestion;' Water Res., 32(12),3607-3614. Battistoni,P., Pavan, P., Prisciandaro, M., and Cecchi,F. (2000)."Struvite Stratful, I., Scrimshaw, M. D., and Lester, J. N. (2001). "Conditions crystallization: A feasible and reliable way to fix phosphorus in influencing the precipitation of magnesium ammonium phosphate." anaerobicsupernatants ." Water Res., 34(l I ), 3033-3041. Water Res.,35(17), 419l-4199. Booker, N. A., Priestley,A. J., and Fraser,I. H. (1999). "Struvite forma- Stumm, W., and Morgan, J. (1970).Aquatic chemistry,Wiley, New York. tion in wastewater treatment plants: Opportunities for nutrient Taiz, L., and Zeiger, E. (1991). Plant physiology, E. B. Brady, ed., -i recovery." Envi ron. Tec hnol., 20(7), 77 7 82. Benjamin/Cummings,Redwood, Calif. Booram, C. V., Smith, R. J., and Hazen, T. E. (1975). "Crystalline Taylor, A., Frazier, A., and Gurney, E. (1963). "Solubility products of phosphateprecipitation from anaerobic animal waste treatment lagoon magnesium ammonium and magnesium potassium phosphates." liquors." Trans.ASAE, l8(l), 340-343. Trsns.Faraday Soc. 59,1580-1584. Brett, S., Guy, J., Morese, G. K., and Lester, J. N. (1997). Phosphorus Unitika-Ltd . (1994). "Fertilizer produced from wastew ater." Japan Chem. removctland recoverl' technologies, Selper, London. Weekly,35(2). Buchanan,J. R., Mote, C. R., and Robinson,R. B. (1994). "Thermo- Wang, J., Burken, J. G., and Zhang, X. (2005). "Effect of seeding dynamics of struvite formation." Trc,ns.ASAE, 37(2),617-621. materialsand mixing strengthon struvite precipitation." Water Envi- Burns,J. R., and Finlayson,B. (1982)."Solubility product of magnesium ron. Res..in press. ammonium phosphatehexahydrate at various temperatures."J. Urol., Webb, K. M., Bhargava,F. S. K., Priestley, A. J., Booker, N. A., and 128(8), 426-428. Cooney, E. (1995). "Struvite precipitation: Potential for nutrient Bums, R,, Moody, L., Walker, F., and Raman,D. (2001)."Laboratory and removal and re-usefrom wastewater."Chem. Aust., 62(10), 42-44. in-situ reductions of soluble phosphorus in swine waste slurries." Webb, K. M., and Ho, G. E, (1992). "Struvite solubility and its applica- Environ. Technol.,22(l l), 1213-1278. tion to a piggery effluent problem." Water Sci. Technol., 26(9-ll), Centre Europ6end'Etudes des Poyphosphates(CEEP). (1998). Recovery 2229-2232. of phosphates for recycling, European Chemical Industry Council, Woods, N. C., Sock, S. M., and Daigger, G. T. (1999). "Phosphorus Bruxelles,Belgium. recovery technology modeling and feasibility evaluation for munici- Clesceri, L. S., Greegberg,A.8., and Trussell, R. R. (1989). Standard pal wastewater treatment plants." Environ. Technol., 20\7 ), 663-67 9. methods for the examinalion of water and u'astetvnfer, Public Wrigley, T. J., Webb, K. M., and Venkitachalm,H. (1992). "A laboratory Health Association,American Water Works Association, and Water study of struvite precipitation after anaerobic digestion of piggery Environment Federation,Washington, D.C. wastes."Bioresour Technol.. 41. ll7-121.

1440IJOURNAL OF ENVIRONMENTALENGINEERING O ASCE / OCTOBER2OO5