Manuscript May 2010

For Review Purposes Only/Aux fins d'examen seulement

1

Application of Polymath Chemical Equilibrium Model for Struvite Precipitation in Soils1

Michael K. Miyittah1, Sachin Gadekar2, Pratap Pullammanappallil2, Craig D. Stanley1, Jean-2Claude Bonzongo3, Jack E. Rechcigl13

4

1Soil and Water Science Department, University of Florida, IFAS, Gulf Coast Research and 5Education Center, Wimauma, FL. 33598. USA. 2Department of Agricultural and Biological 6

Engineering, University of Florida, Gainesville, FL. 32611. USA. 3Department of Environmental 7Engineering Sciences, University of Florida, Black Hall, P.O. Box, 116450, Gainesville, FL. 8

32611. USA.91011

Environmental Context. Excess loss of phosphorus is a major concern for many areas laden 12with manure. It is confirmed that combination of Al-based with Ca-Mg-based materials may help 13reduce offsite release of phosphorus. The formation of phosphorus solid phases in soil may 14prevent phosphorus runoff. Notably, solid phases difficult to predict in soil is struvite. Struvite is 15considered as a slow release fertilizer. Polymath model predicted struvite formation well, but 16prediction was relatively poor in Visual-Minteq. Struvite prediction in soil suggests that the 17immobilized phosphorus can be re-used. The slow kinetics to release phosphorus by struvite may18be suitable for plants gradual uptake. The processes of fixing and slowly releasing phosphorus 19for re-use, due to struvite gradual dissolution may help in promoting the sustainability of 20environmental nutrient resource. 21

22Abstract23

24A new speciation model is developed and implemented in Polymath was found to be successful 25in predicting struvite precipitation in soils. Struvite (NH4MgPO4) has been identified as a mineral 26for the recovery of nitrogen and phosphorus. Predicting struvite precipitation potential in soil is 27important for optimal quantification of nutrient species. Polymath and visual-minteq model were 28used for prediction of several solid-phases in the soil. One approach to immobilize P for solid 29phase formation is by co-blending. Immobilization was achieved through the blending of an Al-30based water treatment residual (Al-WTR) and with Ca-Mg-based materials (slag and magnesium 31oxide). The results suggest that Polymath model revealed solid phases of Dicalcium phosphate 32pentahydrate (DCPP), Magnesium hydroxide (MHO), Magnesium orthophosphate (v) 33docosahydrate (MP22), Magnesium orthophosphate (v) octahydrate MP8, and struvite, which 34were lacking in the modeling from visual-minteq. Residual leachate from the co-blended 35amendments; Soil+WTR+Slag, Soil+WTR+MgO, Soil+MgO, Soil+Slag, Soil+WTR, and the 36control (without amendment) had struvite of 353, 199, 119, 90, 37, and 12 mg L-1 respectively. 37This implies that struvite a phosphate mineral can be precipitated in the soil and could be 38released as nutrients for plant uptake. Struvite precipitation in soil and for re-use may reduce cost 39and may be a safe practice for sustainable environmental nutrient management. 40

41Keywords. Struvite, manure-impacted soil, phosphorus, modeling, 42

43


2

1Corresponding address:2Gulf Coast Research & Education Center314625 CR 672, University of Florida4Wimauma, Fl. 33598.5Phone: (1) - 813-633-41116Fax: (1) -813-634-00017*Corresponding email:[email protected]

1011

Introduction1213

Phosphorus (P) pollution from agriculture sources are primary contributor to more than14

40% of impaired lakes in the United States.[1] Phosphorus from dairy manure-laden soils has 15

also been identified as non-point sources leading to eutrophication and water quality 16

degradation.[2] Significant amount of research has been carried out in recent times to understand 17

P dynamics in animal manure and the interactions of the manure P components with soils.[3]18

Soil P has been found to be strongly associated with iron (Fe)-aluminum (Al) and 19

calcium (Ca)-magnesium (Mg) metal hydr(oxides) depending on the prevailing pH. In acid 20

soils, Fe-Al hydr(oxides) and other metals react with P to form solid phases. On the other hand, 21

under alkaline condition, Ca and Mg activities tended to control the solubility of P in soils.[3,4]22

In Florida, Spodosols are usually characterized by low contents of Fe-Al and Ca-Mg due 23

to coarse texture sandy soils. In addition, the prevailing high rainfall also leads to leaching of 24

basic cations and significant loss of applied P through runoff.[5] Various management practices 25

have been suggested to reduce potential P losses and subsequent impact on water bodies. One 26

approach is the use of amendments to increase the soil sorption capacity for P retention. 27

Successful soil amendments include industrial by-products and municipal wastes that contain 28


3

metals (e.g. Fe, Ca, Al) for P retention. Metals salts (e.g. Al2(SO4)3, FeCl3) have also been used 1

to mitigate P solubility and mobility in poultry manure and poultry manure amended soils.[6,7,8]2

The reaction of these metal cations of Fe-Al and Ca-Mg in soils containing high P has 3

been a subject of recent investigations. The importance of the metal cations is to form a solid 4

phase of P in order to prevent future release. However, phosphate solid phases vary greatly in 5

soils depending on the metal cations. It is therefore of great advantage to understand solubility of6

P if metal cations are co-blended in various ratios. By co-blending we refer to the mixing or 7

blending of two or more by-products containing aluminum water treatment residual (Al-WTR)8

with Ca-Mg based materials (slag and low activity magnesium oxide).[9] 9

Geochemical model (MINTEQA2) has been used as a predictive tool for solid phase 10

speciation of P in manure-impacted soils using Al-WTR as sorbent.[5,10] However, the model 11

used failed to identify struvite (MgNH4PO4.6H2O) as one of the phosphate minerals in the soils. 12

The availability of struvite, a sparingly soluble mineral if present, suggests slow release of P for 13

plants uptake. Consequently, the slow release of nutrients from struvite may prevent the 14

likelihood of adverse risk of excess P into water bodies. In addition, the available struvite in the 15

soil if quantified could provide insight on how much P, N and Mg could be released.16

In animal manure study (pig, cattle, sheep and poultry liter), struvite has been identified 17

using qualitative scanning electron microscopy/energy dispersive X-ray spectroscopy 18

(SEM/EDX), X-ray powder diffraction (XRD) and solid-state 31P NMR techniques.[11,12] In 19

addition, sophisticated equipment like XANES has been used to evaluate the Mg phosphates 20

(struvite) in animal manure.[3] However, the use of these methods in investigating the presence 21

of struvite is labor intensive, expensive and quantitative assessment becomes difficult. To our 22


4

knowledge, to predict struvite in soil is a daunting task. However, polymath model has been used 1

to predict struvite in wastewater.[13]2

The Polymath model was developed for predicting magnesium and calcium related 3

minerals. It is a mathematical model for precipitation, which uses physicochemical equilibrium 4

expressions, mass balance equation for N, P, Mg, and charge balance. The equations derived 5

from the expressions are solved using polymath education version 6.1.[13]6

The objective of this study is to compare prediction of struvite in soil with polymath and 7

that of Visual-Minteq model, resulting from co-blending of aluminum residual and Ca-Mg-based 8

materials.9

10

Experimental methods11

Soil and Amendments Characterizations12

A manure-impacted sample of Immokalee fine sand (sandy, siliceous, hyperthermic 13

Arenic Haplaquod) was obtained from a field site on a dairy cattle ranch located in the Lake 14

Okeechobee County watershed. This soil was selected for its high total P due to long-term 15

manure application and/or fecal deposition, as well as the soil’s geographic extent in South 16

Florida. Random samples were collected from A horizons (0-15 cm), and were thoroughly 17

mixed to yield a composite sample. The soil was air-dried and sieved through a 2 mm stainless18

steel screen. 19

The Al-WTR was collected from a drinking water treatment plant in Bradenton Florida. 20

In characterizing the amendments, pH was measured in a 1:2 (w: v) aqueous suspension as 21

reported elsewhere.[5] Redox potential (Eh as mV) was also measured using two electrodes 22


5

connected to the pH meter (Accument XL60, Fisher Scientific, USA). The reference electrode 1

and temperature sensor were connected to an interface multi-channel (pH-meter/redox meter),2

which enabled simultaneous readings of pH and Eh. After every 10 samples, Eh monitoring was 3

constantly evaluated to avoid drifts in voltage through quality control/assurance maintenance.[14]4

The Al-WTR samples were air-dried and sieved through < 850 µm screen to minimize 5

slaking and increase reactivity.[15] Two Ca and/or Mg products, low activity magnesium oxide 6

(MgO) and slag were characterized in a similar manner as the Al-WTR. Total recoverable Fe, P, 7

Al, Ca and Mg were determined using inductively coupled plasma optimal-emission 8

spectrometry (ICP-OES, Perkin-Elmer Plasma 2100DV, Perkin-Elmer Inc, Waltham, MA., 9

USA.), following digestion according to EPA Method 3050B.[16]10

11

Soil Extractions, Co-blending and Leaching12

The soil was sequentially extracted for P according to Chang et al.[17], using 1:20 13

soil/solution ratio, with modification as in Silveira et al.[5] The sequential extraction procedure is 14

an operational defined scheme of fractionating the soil P. The extraction procedure utilized: 1) 15

1M KCl to determine the soluble or exchangeable P. This P fraction is easily leachable and most 16

responsible for algal growth. 2) 0.1M NaOH to determine the chemisorbed P onto oxides and 17

hydroxides of Fe, Al and Mn. 3) 0.5M HCl to determine P associated with Ca-and Mg18

complexes.[18, 19, 20] However, at the end of the analysis, the recalcitrant P was not analyzed. 19

Amendments containing Ca-, Mg- and Al-based were co-blended as described elsewhere.[9]20

In brief, amendments were applied at 0%, 1%, and 2% by mass to the weight (d.w.) of the soil. 21


6

Each of the soil mixes amended with Al-WTR, Al-WTR+MgO, and Al-WTR+Slag had rate of1

(20 gkg-1 or 2%, (10+10 g kg-1, or 1%+1%), or (10+10 g kg-1, or 1% +1%), respectively. In all, 2

the soil mixes had the same rates of 20 g kg-1 amendment for comparison, with the exception of 3

the control (0%). All treatments were replicated three times in a completely randomized design. 4

Soil column set up was similar as previously reported in Silveira et al.[5] One hundred 5

milliliters of DDI water (adjusted to pH 5, to mimic the pH of rainfall in South Florida) were 6

added to each column weekly. Each leaching event corresponded to approximately 1 pore 7

volume. Leachate of the column studies were collected and major anions (PO4, SO4, NO3, and 8

Cl) and cations (Al, Fe, Ca, Mg, K, Na, and NH4), pH plus EC were analyzed. Cations with 9

exception of ammonium were analyzed using Inductively Coupled Plasma Optical Emission 10

Spectrometers (ICP-OES). Ammonium was measured with ion selective electrode, Fisher 11

scientific, USA. Phosphate, SO4, Cl and NO3 were analyzed using ion chromatograph, Metrohm 12

USA, Inc. 13

Quality assurance and quality control (QA/QC) protocols were followed, including the use 14

of 5% repeats and 5% spikes and blanks for each procedure.[22] Standard calibration curves, as 15

well as quality check standards, were prepared for each procedure. Repeats were within 10% 16

relative standard deviation. In situation where samples standard deviation falls outside the 10% 17

range, repeats were conducted to cross check any anomaly. Spike recovery varied with 18

procedure, but all fell within acceptable levels (95-110%). External quality control checks 19

achieved recoveries within (98-110%). Statistical tests were performed to validate the model 20

fitness. A root mean square error (RMSE), coefficient of determination (R2) and d-statistics or 21

index of agreement was used in the model evaluation. 22

23


7

Chemical Equilibrium Speciation Modeling1

Phosphorus chemical speciation was calculated for the leachates from the control (without 2

amendment), and for 2% application rate of Al-WTR, Slag, and MgO, respectively to the 3

manure-impacted soil. In addition, speciations were calculated for manure-impacted soils co-4

blended with (1%+1%) each of Al-WTR+MgO, and Al-WTR+Slag, respectively. Previous work 5

from Silveira et al.[5] using several weeks of leachates suggested that, number of weeks of 6

selected leachate used did not have any significance difference in chemical speciation. Based on 7

that, the 1st week of data was selected for the chemical speciation equilibrium modeling. In 8

addition, after 2nd to 4th week, due to P immobilizations by the amendments application, PO49

measured was almost zero. The low observed PO4 limits the prediction of possible phosphate 10

minerals in the leachates. Leachates anions analyzed were NO3, Cl, SO4, ammonium, and PO4. 11

Cations analyzed were Al, Mg, Ca, K, and Na. The pH, EC (electrical conductivity) and 12

alkalinity data were also analyzed as inputs for the chemical speciation. Ionic strength (I), as mol13

L-1 of the extracts was calculated from the EC as 0.013*EC.14

The use of Visual-Minteq has a wide application in water quality assessments. The model 15

has borrowed and incorporated wide thermodynamic databases from MINTEQA2. Visual-16

Minteq utilizes the activities of dissolved chemical species in calculating saturation index (SI) of 17

different mineral phases. The SI of each solid is defined as:18

spKIAPSI log (1)19

where, IAP is the ion activity product of the respective chemical species, and Ksp is the solubility 20

product of the estimated solid phase. For any particular mineral solid, if the calculated SI > 0, it 21

implies that solution/leachate is saturated with the respect to that mineral. Similarly, if calculated 22


8

SI < 0, it means that the leachate is under saturated with the mineral phase. On the other hand, if 1

SI = 0, it suggests that the solid has reached equilibrium.[22]2

3

Chemical Equilibrium Modeling with Polymath4

A chemical equilibrium model was developed recently with the aid of polymath and 5

reported in details for predicting magnesium related minerals such as struvite.[13] It is a 6

mathematical model which uses physicochemical equilibrium expressions, charge balance, 7

solubility equilibrium expressions, and mass balance equation for N, P, and Mg. Values of 8

equilibrium constants and solubility products used in the model were reported in Gadekar and 9

Pratap[13]. The model is comprehensive in terms of number of solid species considered, wide 10

range of flexibility and applications. The model can be applied to reactors as in soil systems, to 11

explicitly determine the concentrations of all species (dissolved, ionic and solid) for 12

quantification of purity and yield of struvite under various conditions of pH, and initial 13

concentrations of NH4+, Mg2+, and PO4. Initial conditions which include pH, total concentrations 14

of nitrogen, magnesium, phosphorus, inorganic carbon and calcium are inputs along with 15

equilibrium constants. Initial guesses for Mg2+, NH4+, PO4

3-, CO32+, and Ca2+ were provided. The 16

Polymath Educational Version 6.1 program solved the expressions and gave concentrations of 17

dissolved, ionic species and concentrations of solid components. These expressions included as 18

mass balance equations for total magnesium (Mgt), calcium (Cat), ammonia-N (Nt), inorganic 19

carbon (TIC), and phosphorus (Pt). The expressions were written as functions of corresponding 20

ionic species concentrations in molar quantities. In all, the model predicted the formation of 15 21

different solid phases when solutions containing ammonium, calcium, phosphate, and carbonates 22

species are mixed. Struvite is one of the solids predicted. Using the charge balance equation, the 23


9

appropriate [Exion] concentration was determined. This gives the acid or base requirement to 1

maintain the pH. The following charge balance equation applies to the system:2

[NH4+]+[K+]+2[Mg2+]+2[Ca+]2+[Na+]+[H+]+[MgH2PO4

+]+[MgOH+]=3[PO4-3]+2[HPO4

-2] 3

+[H2PO4-]+2[CO3

2-]+[HCO3-]+[H2PO4

-]+[MgPO4-]+[Cl-]+[OH-]+[Exions] (2) 4

[Exions] = ∑[other cations]-∑[other anions]5

[Exion] was calculated from the charge balance, a positive value indicated addition of alkali, and 6

a negative value indicated addition of an acid to maintain pH. The resulting outcome from the 7

Polymath computation suggests all possible Mg, Ca, K, N, P, and carbonates based on minerals 8

availability. 9

10

Results and discussions11

Amendments and Soil12

The physicochemical properties of Al-WTR taken from Bradenton Florida, showed 86 g 13

kg-1 Al content, and 2.3 g kg–1 Fe content (Table 1). The relative high Al content in Al-WTR 14

suggests that the material might reasonably achieve high P sorption.[23, 24] The slag contained Ca, 15

Mg and Al (598, 106, 156 g kg-1), respectively. In comparison, MgO had a greater Mg content 16

(897 g kg-1) than slag. The pH of Al-WTR was 6.3, which was consistent with values reported 17

elsewhere (6.0-8.4).[26] The Al-WTR had P concentration of 0.02 g P kg-1, which was slightly 18

below the range reported for Al-WTRs (0.3-4.0 g P kg-1).[26, 27] The MgO and slag amendments 19

had alkaline pH values of 10.9 and 11.5, respectively, consistent with values reported by 20

Cucarella and Renman[28]. The soil pH was 7.1. Oxalate-extractable Fe and Al are usually 21

associated with amorphous phases of oxides, suggest that P will be attached to the sesquioxides 22


10

and other forms of hydr(oxides) present in the soil (Table 1). Total P in the Okeechobee test soil 1

was approximately 2800 mg kg-1 indicating a high P load.[2] In the soil, most of the reactive P 2

was associated with Ca-Mg representing, 62% of total reactive P, whereas P associated with Al-3

Fe was 23% and easily leachable P ~15% (Table 1). The above fractionation data are consistent 4

with results from Nair et al.[2] and Sharpley et al.[25], showing that Ca-Mg bound P dominated the5

forms in a manure-impacted soil. The total P concentration of MgO was below the instrument’s 6

detection limits, and slag P content was approximately 0.03 g P kg-1. 7

8

Simulation vs. Experimental data using Polymath model and Visual-Minteq9

(1) Solid Phase Equilibria from the literature10

Experimental, synthetic, and actual waste water data taken from the literature were 11

simulated using Visual-Minteq and Polymath model.[13] The inputs data for simulations utilized 12

the concentrations of total NH4-N, Mg, and PO4-P, Ca, total inorganic carbon and pH. The 13

current version of Visual-Minteq used did not include struvite in its database. Thermodynamic 14

parameters for struvite were added before running the computer model in Visual-Minteq for the 15

prediction of solid phases. The results are presented in Table 2 and 3. A graphical representation 16

of the amount of struvite predicted versus that of the experimental values is plotted (data not 17

shown). The results showed clearly, the superior strength of Polymath model (R2 = 0.57) over 18

that of the Visual-Minteq (R2 = 0.33) in predicting struvite (Table 3). However, to validate the 19

accuracy of the R2 values, the use of Root Mean Square Error (RMSE) and index of agreement 20

(d-statistics) were obtained. The RMSE was used to differentiate the model having the most 21

significant influence, since R2 may indicate inaccuracy when used as a measure of accuracy.[29].22

A lower value of 293.2 RMSE for Polymath suggests that the model fits the data well in terms of 23


11

struvite predictions (Table 3). Further, index of agreement or d-statistics was also evaluated. The 1

“d” of 0.78 was greater for Polymath compared to 0.69 for Visual-Minteq, suggesting that the 2

model is better in agreement for struvite prediction. The inability of Visual-Minteq to predict 3

struvite well may be due to greater stability of calcium ions such as calcium phosphate,4

hydroxyapatite or due to Ca ions which compete with Mg for phosphate ions in solution.[30, 31] In 5

addition, struvite formation is affected by precipitation kinetics and wide degree of variabilities 6

such as chemical species, pH, EC and ions activity products.[36] These factors may have 7

accounted for the large amount of RMSE observed. A polymath chemical equilibrium model, 8

however, accounted for saturation indexes of phosphate minerals, and other amorphous forms 9

when Ca ions are greater than Mg into its formulations. Polymath model accounting for 10

phosphate saturation indexes might have led to the lower reduction of errors observed in the 11

model predictions. In further agreement, the polymath model was specifically designed for Mg-P 12

solid phases, thus the quality of the observed predictions. 13

14

(2) Solid phase Equilibria of Co-blended soil with Visual-Minteq15

Simulation results of soil (control) and co-blended samples were calculated using Visual-16

Minteq and Polymath model and are presented in Table 4 and 5, respectively. Although co-17

blended samples were assumed to be thoroughly mixed, we recognized the complexities and the 18

uncertainties associated with non-equilibrium nature of soil samples, as well as the inherent 19

kinetic effects in chemical speciation. Bearing the uncertainties in mind, we restrict our 20

interpretation to major significant trends in the modeling results.21

Phosphorus chemical speciation was calculated for the leachates collected from the samples 22

soil [Soil+MgO (2%), Soil+Slag (2%), Soil+Al-WTR (2%), Soil+Al-WTR+MgO (1+1%), 23


12

Soil+Al-WTR+Slag (1+1%)]. Visual-minteq was used to predict the saturation indices for 1

possible solid minerals formation. Table 2 indicates the soil solution saturation indices for 2

phosphate minerals whose dissolution-precipitation reactions may control P activity of the soil.3

For the control, major fraction of P was associated with Ca-Mg ~60%, followed by HPO42-4

~30%, and H2PO4- ~4% of negatively charged complexes. The control soil solution was5

undersaturated with respect to Mg-P minerals of farringtonite [Mg3(PO4)2] and newberryite 6

[MgHPO4.3H2O], but were saturated with Ca-P minerals of hydroxyl apatite [Ca5(PO4)3OH], β-7

tricalcium phosphate [β- Ca3(PO4)2], and octacalcium phosphate [Ca4H(PO4)3.3H2O], except that 8

of dicalcium phosphate dihydrate [CaHPO4.2H2O] and dicalcium phosphate [CaHPO4] whose 9

mineral phases were almost at equilibrium with the soil. The results suggest that phosphate 10

solubility and activity in soil solutions from manure-impacted soils, without any amendment11

were dominated mainly by the dissolution-precipitation reactions of under-saturated Ca- and Mg-12

P solid phases (SI< 0). The above observation with the control soil without any amendment is 13

consistent with results from Immokalee soil, a spodosol.[5, 10]14

Co-blending the soil with Slag, Al-WTR and MgO at 2% rate, had similar mineral 15

predictions under visual-minteq modeling. All treatments were under saturated with newberryite16

[MgHPO4.3H2O], farringtonite [Mg3(PO4)2] except that of Soil+MgO, indicating super-17

saturation of newberryite. The super saturation with respect to newberryite supports the fact that 18

Mg-P bearing minerals are responsible for controlling P solubility in manure-impacted soils. On 19

the other hand, addition of Slag and Al-WTR co-blended soil had similar solid phases with Mg-P 20

minerals but with undersaturation indices (Table 4). In addition, undersaturation was also 21

observed with some calcium mineral phases of [CaHPO4.2H2O] and [CaHPO4], thus indicating22

the role of Ca and Mg in controlling P solubility, which is consistent with the conclusions of 23


13

other studies involving manure-impacted soils.[5, 10, 32] Co-blended samples with Soil+WTR+Slag 1

and Soil+WTR+MgO, at 1% +1% amendments had solid phase similar to either Soil+MgO or to 2

Soil+Slag at 2%, respectively. Although, the solid solution of the co-blended samples were 3

thoroughly mixed, it was suggested that the chemical activity of solid phases in the soil system 4

dissolve nonstoichiometrically, thereby creating in-homogeneity and disequilibrium between the 5

solid’s surface and that of the interior.[33] Consequently, the true equilibrium in soil suspensions 6

maybe extremely difficult, and applying strict thermodynamics may be less effective for co-7

blended samples. In all treatments, the most thermodynamically stable solid phase observed 8

appeared to be that of hydroxyapatite with SI > 0.9

10

(3) Solid phase Equilibria of Co-blended soil with Polymath model11

The types of solid phases identified by polymath model were identical to those predicted by 12

visual-minteq (Table 5). However, major differences revealed from Polymath were solid phases 13

of Dicalcium phosphate pentahydrate (DCPP), Magnesium hydroxide (MHO), Magnesium 14

orthophosphate (v) docosahydrate (MP22), Magnesium orthophosphate (v) octahydrate MP8, 15

and struvite. Polymath model shows clearly, the prediction of magnesium-based solids. Struvite 16

precipitation was observed to follow the following order: Soil+WTR+Slag > Soil+Slag > 17

Soil+WTR+MgO > Soil+MgO > Soil+WTR > control. The order of struvite precipitation might 18

suggest a combined effect of magnesium availability and pH. This is because in struvite 19

formation, magnesium ion and pH play a major role. The importance of struvite cannot be 20

overemphasized. This is because struvite as a sparingly soluble mineral can slowly release its 21

nutrients in the long run. The slow release of struvite may offer great advantage to forage 22

growers who are faced with excess laden of manure. The convectional school of thought in 23


14

addressing soil P pollution is to permanently fix the P with a sorbent in the soil. However, with 1

the identification of struvite in the amendment application which is aimed to immobilize the P 2

suggests that the unconventional thinking should be fixing the P and releases it later for future 3

use. This unconventional school of thought is beneficial in cost terms, as well as for sustainable 4

environmental management. Struvite is produced at industrial scale and used as a slow release 5

fertilizer. It can release the P and N at a slower rate which can be controlled by adjusting the pH 6

if needed. The released N and P can be beneficial to plant growth. However, struvite is not used 7

alone, but rather used by mixing with other forms of inorganic and organic fertilizers.[34] The 8

presence of struvite as revealed by the Polymath model suggests that cost will be saved9

compared to synthetic/industrial struvite applied to the soil. This is due to the fact that the 10

kinetics of P release by struvite is so slow and a function of prevailing pH.[35] The slow release 11

of struvite may lead to saving cost to ranchers whose soils are heavily laden with manure. The 12

dwindling phosphate reserve in Florida is a concern thus, suggesting that cycling and re-use of 13

excess P in soil through co-blending techniques of amendments may be a safe practice.14

15

Conclusions16

Loss of phosphorus (P) is a major concern for many areas laden with manure. It is 17

confirmed that co-blending of Al-based materials with Ca-Mg-based materials may help to 18

reduce offsite release of P. The use of chemical equilibrium models of Visual-Minteq and 19

Polymath model revealed various solid phases’ formed in the soil. Polymath model however was 20

observed to add new information to that revealed by Visual-Minteq. The results suggest that 21

Polymath model revealed solid phases of Dicalcium phosphate pentahydrate (DCPP), 22

Magnesium hydroxide (MHO), Magnesium orthophosphate (v) docosahydrate (MP22), 23


15

Magnesium orthophosphate (v) octahydrate MP8, and struvite, which were lacking in the 1

modeling from visual-minteq. Notably, the solid phase that is difficult to predict in soils is 2

struvite. Struvite precipitation in soil suggests that the immobilized P can be re-used. Residual 3

leachate from the co-blended amendments; Soil+WTR+Slag, Soil+WTR+MgO, Soil+MgO, 4

Soil+Slag, Soil+WTR, and the control (without amendment) had struvite of 353, 199, 119, 90, 5

37, and 12 mg L-1 respectively. The slow kinetics of struvite to release P may be harnessed for 6

future plants uptake. Such a process of P immobilization and release may help in promoting the 7

sustainability of environmental nutrient. 8

9

References10

[1] USEPA, National Water Quality Inventory 2000 Report. USEPA, Washington, D.C. 11

Available at http://www.epa.gov/305b/2000report/. 2000, (Accessed July 22nd, 2009).12

[2] V.D. Nair, D.A. Graetz, K.M. Portier, Forms of phosphorus in soil profiles from dairies of 13

South Florida. Soil Sci. Soc. Am. J. 1995, 59, 1244.14

[3] K. Güngör, A. Jürgensen, K.G. Karthikeyan, Determination of phosphorus speciation of 15

dairy manure using XRD and XANES spectroscopy. J. Env. Qual. 2007, 36, 1856. 16

[4] W.L. Lindsay, Chemical equilibria in soils. The Blackburn Press. Caldwell, New Jersey. 17

USA 2001.18

[5] M.L. Silveira, M.K. Miyittah, G.A. O’Connor, Phosphorus release from a manure-impacted 19

spodosol: effects of a water treatment residual. J. Env. Qual. 2006, 35, 529. 20


16

[6] P.A. Moore, D.M. Miller, Decreased phosphorus solubility in poultry litter with aluminum, 1

calcium and iron amendments. J. Env. Qual. 1994, 23, 325. 2

[7] P.A. Moore, T.C. Daniel, D.R. Edwards, Reducing phosphorous runoff and improving 3

poultry production with alum. Poultry Sci. 1999, 78, 692.4

[8] L.M. Malecki-Brown, J.R. White, Effect of aluminum-containing amendments on 5

phosphorus sequestration of wastewater treatment wetland soil. Soil Sci. Soc. Am. J. 2009, 6

73, 852.7

[9] M.K. Miyittah, C.D. Stanley, C. Mackowiak, J. E. Rechcigl, Developing a remediation 8

strategy for phosphorus immobilization: Effect of co-blending Al-residual and Ca-Mg 9

amendments in a manure-impacted spodosol. Soil Sediment Contam. 2010 in Press.10

[10] S. Agyin-Birikorang, G.A. O’Connor, J-C. Bonzongo, Modeling solid phase control of 11

solubility of drinking water treatment residuals-immobilized phosphorus in soils. 12

Commun. Soil Sci. Plant Analysis, 2009, 40, 1747.13

[11] K. Güngör, K.G. Karthikeyan, Probable phosphorus solid phases and their stability in 14

anaerobically digested dairy manure. Trans ASABE, 2005, 48, 1509.15

[12] G.S. Toor, S. Hunger, J.D. Peak, J.T. Sims, D.L. Sparks, in Advances in Agronomy (Ed. 16

D.L. Sparks) 2006, p.1-72, Vol. 89. (Academic Press, Boston). 17

[13] S. Gadekar, P. Pullammanappallil, Validation and applications of chemical equilibrium 18

model for struvite precipitation. Environ. Modeling Assess. 2009. doi:10.1007/s10666-19

009-9193-720


17

[14] M.C. Rabenhorst, Making soil oxidation-reduction potential measurements using 1

multimeters. Soil Sci. Soc. Am. J. 2009, 73, 2198.2

[15] E. A. Dayton, N.T. Basta, A method for determining the phosphorus sorption capacity and 3

amorphous aluminum of aluminum-based drinking water treatment residuals. J. Env. Qual.4

2005, 34, 1112.5

[16] USEPA, Acid digestion of sediments, sludges, and soils [Online], Available at 6

http://www.epa.gov/epaoswer/hazwaste/test/pdfs/3050b.pdf (verified December 6th, 7

2009). USEPA, Cincinnati, OH. 1996. 8

[17] A.C. Chang, A. L. Page, F.H. Sutherland, E. Grgurevic, Fractionation of phosphorus in 9

sludge affected soils. J. Env. Qual. 1983, 12, 286.10

[18] A.H. M. Hieltjes, L. Lijklema, Fractionation of inorganic phosphates in calcareous 11

sediments. J. Env. Qual. 1980, 9, 405.12

[19] K.C. Ruttenberg, Development of a sequential extraction method for different forms of 13

phosphorus in marine sediments. Limnol. Oceanogr. 1992, 37, 1460. 14

[20] W. Stumm, J.J. Morgan, Aquatic chemistry: chemical Equilibria and rates in natural waters 15

3rd edition, 1996. (John Wiley & Sons, Inc. New York). 16

[21] V.H. Kennedy, A.P. Rowland, J. Parrington, Quality assurance for soil nutrient analysis: a 17

case study. Commun. Soil Sci. Plant Anal. 1994, 25, 1605.18

[22] C. Zhu, G. Anderson, Environmental applications of geochemical modeling. 200219

(Cambridge University Press. Cambridge). 20


18

[23] D.G. Grubb, M.S. Guimaraes, R. Valencia, Phosphate immobilization using an acidic type 1

F fly ash. J. Hazard. Mater. 2000, 76, 217.2

[24] M. Arias, J. Da Silva-Carballal, L. Garcia-Rio, J. Mejuto, A. Nunez, Retention of 3

phosphorus by iron and aluminum oxide-coated quartz particles. J. Coll. Interf. Sci. 2006, 4

295, 65. 5

[25] A. N. Sharpley, R.W. McDowell, P.J. A. Kleinman, Amounts, forms and solubility of 6

phosphorus in soils receiving manure. Soil Sci. Soc. Am. J. 2004, 68, 2048. 7

[26] K.C. Makris, G.A. O’Connor, in Currents perspectives in environmental geochemistry, 8

(Eds D. Sakar, R. Datta, R. Hannigan), 2007, (Geological society of America Press, 9

Denver Col.)10

[27] Dayton, E.A., N.T. Basta. C.A. Jakober, and J.A. Hattey, Using treatment residuals to 11

reduce phosphorus in agricultural runoff. J. Am. Water Works Assoc. 2003, 95,151. 12

[28] V. Cucarella, G. Renman, Phosphorus sorption capacity of filter materials used for on-site 13

wastewater treatment determined in batch experiments-a comparative study. J. Env. Qual.14

2009, 38, 381. 15

[29] C.J. Willmott, Some comments on the evaluation of model performance. Bull. Amer. 16

Meteor. Soc. 1982, 63, 1309. 17

[30] S. Kristell, C. Le, E. Valsami-Jones, P. Hobbs, S.A. Parsons, Impact of calcium on struvite 18

crystal size, shape and purity. J. Crys. Growth. 2005, 283, 514.19


19

[31] N. Ch. Bouropoulos, P.G. Koutsoukos, Spontaneous precipitation of struvite from aqueous 1

solutions. J. Crys. Growth. 2000, 213, 381. 2

[32] M.S. Josan, V.D. Nair, W.G. Harris, D. Herrera, Associated release of magnesium and 3

phosphorus from active and abandoned dairy soils. J. Env. Qual. 2005, 34, 181. 4

[33] H.L. Bohn, R.K. Bohn, Solid activities of trace elements in soils. Soil Sci. 1987, 143, 398.5

[34] Y. Ueno, M. Fujii. Three years operating experience selling recovered struvite from full 6

scale plant. Environ. Sci. Technol. 2001, 22, 1373.7

[35] M.I.H. Bhuiyan, D.S. Mavinic, R.D. Beckie, Dissolution kinetics of struvite pellets grown 8

in a pilot-scale crystallizer. Can. J. Civ. Eng. 2009, 36, 550.9

[36] I. Celen, J.R. Buchanan, R.T. Burns, R.B Robison, D.R. Raman, Using chemical 10

equilibrium model to predict amendments required to precipitate phosphorus as struvite in 11

liquid swine manure. Water Res. 2007, 41, 1689. 12

13

14

15

16

17

18


20

Table 11

General properties of amendments and soil2

Properties P Fe Ca Mg Al pH

Amendments …………………………….g kg-1…………..

Al-WTR 0.02 2.3 4.6 0.5 86 6.3

Slag 0.03 14 598 106 156 11.5

MgO bBDL 2.0 9.0 879 2.0 10.9

Sequentially extracted P Oxalate extractable

...…………………. mg kg-1………………………..

#Soil pH cKCl dNaOH fHCl Fe Al

A-horizon 7.1 400 640 1720 348 210

3

a Values are means of triplicate. b BDL (below detection limits of 1*10-3 mg L-1). c,d,f Soil 4

fractionations (KCl-P (soluble or exchangeable P), NaOH-P (extractable P bound to Al-Fe) and 5

HCl-P (extractable P bound to Ca-Mg)) determined using the method of Chang et al., 1983.6

#Okeechobee soil. 7

8

9


21

1

Table 22Dataa taken from literature to simulate model predictions of Polymath and Visual-Minteq.3

Initial concentrations (mM)

Model PredictionsStruvite (mg L-1)

Type of wastewater pH

MgT PT NT

Exptl Struvite (mg L-1) Visual-

MinteqPolymath

Solutions prepared by adding NH4Cl, KHPO4, MgCl, carbonate and acetate

6.8 8.3 12.9 21.43 601 0 223.6

Synthetic urine containing PO4, NH4, Na, Mg, K, Ca, Cl, citrate, carbonate

8.0 20.0 13.45 20.18 1685 0 1253.3


9.4 7.42 14.83 18.70 1045 1006 987.07


9.4 14.8 14.83 18.70 2011 1087 1845.0

Liquid manure 8.5 2.39 5.51 80.00 338 319 322.2Supernatant from anaerobically digested sludge dewatering centrifuge

8.5 1.51 1.97 43.88 195 171.05 200.9

Anaerobic digester effluent supernatant 8.5 7.03 6.38 24.5 805 825 818.2Synthetic samples prepared by using MgCl2, NaH2PO4, NH4Cl

9.0 14.3 14.26 14.26 1714 1581 452.4

Domestic wastewater + 2 % landfill leachate

9.2 7.79 7.79 7.79 1036.9 681.7 207.4

Supernatant from sludge centrifuges in a biological nutrient removal plant

8.1 1.54 2.00 44.5 210.98 143.6 198.2

Swine waste 9.0 9.74 6.09 12 758.6 713.5 705.5aData taken from Gadekar and Pullammanappallil[13] .4


22

1

Table 32

Statistical parameters for model prediction with coefficient of determination (R2), root mean 3square error (RMSE), and index of agreement (d).4

5

Where,n

OPRMSE

n

iii

1

2)(, n = number of observations, Pi = predicted value for the ith6

measurement, and Oi = observed value for the ith measurement. 7

8

10,)(

)(1

1

2

1

2

dOP

OPd n

iii

n

iii

, n = number of observations, Pi = predicted value for the ith 9

measurement, Oi = observed value for the ith measurement, Oi = observed value for the ith10

measurement, O = the overall mean of observed values, OPP ii and OOO ii .[29]11

12

13

Statistics Polymath Visual-Minteq

R2 0.57 0.33

RMSE 293.2 355.7

d 0.78 0.69


23

1

Table 42

Saturation indices calculated using Visual-Minteq for treatments with and without co-blending.3

Predicted minerals

Treatments (β-TCP) OCP DCP DCPD HA FNT NBT

Soil (Control) 4.17 3.99 0.40 0.10 13.57 -0.48 -0.39

Soil+MgO (2%) 3.13 1.03 -1.52 -1.81 13.42 1.89 -1.18

Soil+Slag (2%) 2.05 0.19 -1.28 -1.57 11.01 -2.33 -1.98

Soil+Al-WTR (2%) 3.22 2.59 -0.05 -0.35 12.13 -1.81 -0.97

Soil+WTR+MgO (1+1%) 2.37 0.14 -1.64 -1.94 12.01 0.16 -1.62

Soil+WTR+Slag (1+1%) 1.15 -0.62 -1.52 -1.80 9.58 -3.46 -2.28

Saturation index = log ion activity product (IAP)- log solubility product (Ks).4

(β)-Ca3(PO4)2 = β-Tricalcium phosphate (β-TCP)5

Ca4H(PO4)3.3H2O = Octacalcium phosphate (OCP)6

CaHPO4 = Dicalcium phosphate (DCP)7

CaHPO4.2H2O = Dicalcium phosphate dihydrate (DCPD)8

Ca5(PO4)3OH = Hydroxyapatite (HA)9


24

Mg3(PO)2= Farringtonite (FNT)1

MgHPO4. 3H2O = Newberryite (NBT)2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17


25

1

Table 52

Calculated values of solid (mg) found per litter of residual leachate as determined by Polymath model3

4

(β)-Ca3(PO4)2 = β-Tricalcium phosphate (β-TCP)5

Ca5(PO4)3OH = Hydroxyapatite (HA)6

CaHPO4.5H2O = Dicalcium phosphate pentahydrate (DCPP)7

CaHPO4 = Dicalcium phosphate (DCP)8

Mg(OH)2 = Magnesium hydroxide (MHO)9

Predicted mineral

Treatments β-TCP

HA DCPP DCP MHO NBT MP22 MP8 Struvite (%Purity)

Total (Solid)

Control 26.9 1.35E-9 622.3 280.9 22.5 23.2 0.15 1.135 12.24(1.2) 989.3

Soil+MgO (2%) 377.6 9.60E-7 173.4 78.3 22295.9 3.64 0.37 2.86 119.0(3.9) 3051.2

Soil+Slag (2%) 14.3 4.95E-10 484.4 218.6 96.7 87.9 9.2 70.2 90.29(8.4) 1071.7

Soil+Al-WTR (2%) 48.1 4.96E-9 544.4 245.7 96.0 36.9 1.61 12.28 37.18(3.6) 1022.2

Soil+WTR+MgO (1+1%) 9.2 6.58E-08 150.3 67.8 2080.7 8.79 1.98 15.1 199(7.4) 2610.9

Soil+WTR+Slag (1+1%) 8.8 8.23E-11 1092.6 493.0 19.0 324.1 24.6 187.4 353.1(14.1) 2502.6


26

MgHPO4. 3H2O = Newberryite (NBT)1

Mg3(PO4)2. 22H2O = Magnesium orthophosphate (v) docosahydrate (MP22)2

Mg3(PO4)2. 8H2O = Magnesium orthophosphate (v) octahydrate (MP8)3

MgNH4PO4.6H2O = Struvite4

5

6


27

1

Documents

Manuscript May 2010