Chemical Reactions of Polyphosphate Fertilisers in Soils and Solutions

Thérèse McBeath

In fulfilment of the requirements for the degree of

I)octor of Philosophy

A thesis submitted to

Soil and Land SYstems

School of Earth and Environmental Sciences

The UniversitY of Adelaide

Australia

July 2006 Table of Gontents

Table of Contents II List of Figures VI List of Tables IX Acknowledgements XI Abbreviations XIII Abstract XIV Declaration XVI

Crr¡,PrNN 1. GNNAN¡T, IXTNONUCUON ÀND LITERÄTURE TIEYIEW 17 1.1. Introduction t7 I.2. Phosphorus in the Environment and Agriculture 18 1.2.1. Phosphorus sPecies in soil t9 1.3. Phosphorus Fertilisers 22 1.4. Reactions of P in Soil 25 1.4.1. Chemical Reactions 25 1.4.2. EnzymaticReactions 29 1.5. Factors Controlling P Availability in Soils 32 1.5.1. Soil pH 32 1.5.2. Iron and Aluminium 33 1.5.3. Calcium 39 1.5.4. Organic Residues and Micro-organisms 45 1.6. Reactions of P Fertiliser in Soils and Solution 48 1.6.I. Orthophosphate FertiliserDissolution 48 1.6.2. Poþhosphate Fertiliser Hydrolysis in Soils and Solutions 51 1.7. Movement of P in Soils 54 1.8. Conclusions and Future Direction 58

CHAPTER 2. POLYPHOSPHÄTE SPNCT¿,TTOU TX SOT,UTTOU 6l 2.1. Introduction 6l 2.2. Materials and Methods 63 2.2.L. On-lineColorimetrY 63 2.2.2. Ion ChromatograPhY 64 2.2.3. PhosphateReagents 65 2.2.4. Performance of the Analytical Techniques 66 2.3. Results and Discussion 69 2.3.1. Performance of the Analytical Methods 69 2.4. Conclusions 74

cu¡prnn 3. Por,ypnospHlTn Fnnrn-rsnn sor-urro¡l sr¡,srr,rrv 75 3.1. Introduction 75 3.2. Materials and Methods 77 3.2.1. Reagents 77 3.2.2. Treatments 77 3.2.3. Speciation and Quantification of P 77 3.2.4. StatisticalAnalysis 78 3.2.5. Calculations 78 3.3. Results and Discussion 79 3.3.1. Change in P Speciation Over One Month 79

il 3.3.2. Hydrolysis Rate Constant 80 3.3.3. Activation EnergY 82 3.3.4. Change In P speciation Over One'Week 83 3.4. Conclusions 85

SOTT-: A CHAPTER 4. HYONOT,YSIS OF PYNOPTTOSPHATE IN A IIIGIil,Y CIT-C^MNOUS sor,ro-Sr¡.Tn 3tP NMR Sru¡Y. 86 4.I. Statement of Contributions 86 4.2. Introduction 87 4.3. Materials and Methods 88 4.3.1. Soil Collection and Chemical Properties 88 4.3.2. Soillncubations 89 4.3.3. Solid-State'lP NMR Spectroscopy 89 4.3.4. Ion ChromatograPhY 9T 4.4. Results and Discussion 92 3tP 4.4.1. Solid-State NMR Spectra of Reference Salts 92 4.4.2. Solid-State'rP NMR Spectra of Soils 94 4.4.3. of orthophosphate and Pyrophosphate contents Quantification 3tP from Solid-State NMR SPectra 97 4.4.4. Quantification of Orthophosphate and Pyrophosphate Concentrations Using Alkaline Extraction and Ion Chromatography 99 4.4.5. Hydrolysis Rate and Pyrophosphate Half-Life 100 4.5. Conclusions 102

103 Crr¡,PrPN 5. SORPTION OF PYROPHOSPIIATE AND ORTHOPHOSPIIATE NI SOTT- 5.1. Introduction 103 5.2. Materials and Methods 104 5.2.I. Reagents 104 5.2.2. Soil Types and Characteristics 104 5.2.3. Soil Analyses 105 5.2.4. Speciation and Quantification of P 106 5.2.5. SorPtionCharacteristics 106 5.2.6. StatisticalAnalysis t07 5.3. Results and Discussion 108 5.3.1. Soil Characteristics 108 5.3.2. SorptionCharacteristics 108 5.3.3. pH 116 5.3.4. CationConcentrations tl7 5.3.5. Organic Carbon t20 5.4. Conclusions t22

Cr¡1¡prBn 6. IS9TOPIC TnCrn¡rQupS ron I¡.IVESTIGATING REACTI9NS OF 123 PoLYPHOSPHATE tr'ERTtr,ISERS IN SOILS 6.1. Introduction t23 6.2. Assumptions and Principles of Isotopic Techniques t24 6.2.1. TracerTechniques t24 6.2.2. Isotopic DilutionTechniques 125 6.3. Double Labelling Techniques 126 6.4. IsotopicExperiments t28 6.4.1. The Processes Under Investigation t28 6.4.2. Chapter 7- A Hydrolysis Study t29

ilI 6.4.3. Chapter 8- A Lability StudY 130 6.4.4. Chapter 9- A Mobilisation Study 130 6.5. Conclusions 131

Culprnn 7. AN ISOroprC SrUoy or PynopnoSPHATE HYDRoLYSIS 132 7.1. Introduction t32 7.2. Materials and Methods t33 7.2.1. Soil Characteristics 133 7.2.2. Reagents 133 7.2.3. ExperimentalDesign t33 7.2.4. AnalyticalMethods t34 7.2.5. Calculations 135 7.2.6. StatisticalAnalysis 138 7.3. Results and Discussion 138 7.3.1. Examination of Isotopic Exchangeability of Phosphorus in Pyrophosphate 138 7.3.2. Effect of Rate of Pyrophosphate Added on Soluble Phosphorus Concentration and Soil Suspension pH 139 7.3.3. Effect of Rate of PP Added on Lability and Hydrolysis of PP t42 7.4. Conclusions 145

CTHPTNN 8. A STUOY OF LABILITY OT P TN SOTT,S TREATED WITH ONTTTOPTTOSPHATE ANDPYROPHOSPHATE 146 8.1. Introduction 146 8.2. Materials and Methods 147 8.2.1. Soil Characteristics 147 8.2.2. Reagents 148 8.2.3. Experimental Design 148 8.2.4. Analytical Methods 149 8.2.5. Calculations 149 8.2.6. Statistical AnalYsis I52 8.3. Results and Discussion 152 8.3.1. Incubation Effects on Soluble Phosphorus, PH, Cations and Organic carbon 152 8.3.2. Incubation Effects on Lability and Partitioning of Orthophosphate Compared to PyroPhosPhate 159 g.4. Conclusions . 163

CHÄPTER 9. MOBTLISATION OF NITTVN PHOSPHORUS NY PYNOPHOSPIIATE 165 9.I. Introduction 165 9.2. Materials andMethods t66 9.2.1. SoilCharacteristics r66 9.2.2. Reagents r66 9.2.3. ExperimentalDesign t66 9.2.4. Analytical Methods t67 9.2.5. Calculations t67 9.3. Results and Discussion 17t 9.3.1. Effect of Pyrophosphate Addition on soluble Phosphorus Concentration and Soil Suspension pH t7I g.3.2. Labile Pyrophosphate, Labile Hydrolysed Pyrophosphate and Mobilised Native PhosPhorus 172

IV 9.4. Conclusions 177

Cn¡,prnn 10. Gnxnnlr, DrscussroN 179 10.1. Summary of Findings r79 10.1.1. Analytical Techniques t79 10.I.2. Hydrolysis in Solution 180 10.1.3. Hydrolysis in Soil 180 10.1.4. Sorption 180 10.1.5. Isotopic Studies of Lability, Hydrolysis and Partitioning 18r 10.1.6. Fate and Effects of PP Reactions in Soils t82 10.2. Future Research 18s

Rrrpnpxcns 187

V List of Figures

Figure 1-1: The P cycle in soils. The boxes represent pools of P forms in the cycle and affows represent movement and transformations between pools (Burns and Slater te82)..... Figure 1-2: Phosphorus rate (ky'ha) vs. grain yield (t/ha) for ammonium @'T:: ::l*i::::T i:::lï:".'lï::l":lï:P :T:*:: ?' :', 7?i?; Figure 1-3: Common ions of P in pollphosphate fertilisers (Rashchi and Finch 2000)

Figure 1-4: Covalent bonding reactions of P (Stevenson and Cole 1999).-.-...... -...... 27 Figure 1-5: Electrostatic attraction reactions of P with Fe (Stevenson and Cole 1999)

Figure 1-6: Effect of pH on enzyme activity (corn-root acid phosphatase and pyrophosphatase activity). Acid phosphatase activity is expressed as pg p- nitrophenol released/l0 mg corn rootslhr, and pyrophosphatase activity is expressed as ¡rg Pi released/l0 mg corn roots/hr (Gilliam and Sample 1968). ...31 Figure 1-7: Ionic forms of P in soil solution as a function of pH (Brady and'Weil, lee6) .32 Figure 1-8: The effect of pH on the P species in soil solution (Lindsay 1979)...... 34 Figure 1-9: Precipitates of Fe and Al (Brady and V/eil 1996)- ....'...... -....36 Figure 1-10: Solubility diagram for Al phosphates (Lindsay et al. 1962)...... 37 Figure 1-11: Solubility of Ca phosphates compared to strengite and variscite (Lindsay 1979). --..-..42 Figure 1-12: OP concentration (0.00lM) as a function of time (hour) for three OP solutions reacted with calcium carbonate with and without Na-PP (Amer and Mostafa 1981). 44 Figure 1-13: Interactions of organic P in soils (Stevenson and Cole 1999)...... 47 Figure 1-14: Hydrolysis of 50ppm P as PP and TP solutions (measured as OP formed) by sterile and non-sterile wheat and pea roots (Savant and Racz I97 2)...... 5 3 Figure 1-15: Movement of phosphate by mass flow and diffusion from a granule of triple or single superphosphate through water-filled and water-lined large micropores in a well aggregated soil. Penetration of P into aggregates is incomplete due to the slow rate of P diffusion in smaller intra- aggregate micropores and discontinuous micropores (Hedley and Mclaugþlin 2005) 56 Figure 2-l: Calibration curves for anaþsis of P species by the on-line colorimetric technique. (A) OP (50.0034x, Rz:l) andby IC (B)OP (52'19x, R2:l), PP (y:2.04x, R':t) utrã tp (y:2.15x,R2:1). *Error bars repres ent -rl- standard effor.. 70 Figure 3-l: Percentage of P species as OP, PP and TP after one day at4"C and after 28 days aI4"C,25"C and 50oC. Measurements were made at solution pH values of 6.4,5.8,4.9 and2.3. .. 80 Figure 3-2 Change in log TP concentration (mg/L) over time (hours) at4"C,25"C and 50'C (-0-50'C -t-25"C -A-4'C) at soh'tion p}J6.4. The equation of the line at 4"C is y: 2xl0-e +3.997 with a R2 of 0'03 I, at 25" C p3x1'0-8+3.99 with a R2 of 0.49, andat 50'C 5-4x10-7+4.04 with an R' of 1..81 Figure 3-3: Change in 1og TP concentration (mg/L) over time (hours) at pH 2.3,4.9, 5.8 and 6.4 (r-pH 2.3 -'-pH 4.9 - L-pH5.8 -r--pH 6.4) at25"C. At pH 2.3 the equation of th.line is 5-1x10ti+3.63 with a R2 of 0.93, at pH

VI pH 6.4 y:- 4.9 y:9x10-ex+3. 86 with R2 of 0. 1 6, at pH 5. 8 y2x 1 0-8x+3.95 and at 7xl0-ex+4.02 with R3 of 0.04. """82 Fi sis of TP at pH 2'3'Ea is the slope of the olysis rate constant) and 1000/T (T is y|-t2.7 0x+36.32 with R3 0.98. ..-...... 83 Figure 3-5: Percentage of P species as OP, PP and TP after 1 day and 7 days , at 4"C, Z5"C and 50oC. Measurements were made at solution pH values of 6'4, 5.8, 4.9 and2.3. ""'84 3rP Figure 4-1: Solid-state direct polanzation (DP) spectra of Na- and Ca- OP and PP salts...... """"""""" 93 31P Figure 4-2: Solid-state cross polarization(CP) and direct polanzation (DP) spectra of unamended soil and soil ãmended with PP supplying 2000 mg P/kg soil and incubated for 1 to 2l days. The vertical scales of the spectra have been set to allow direct comparison within the series of cP and DP spectra. -....96 Figure 4-3: OP and PP concentration determined by solid-state NMR spectroscopy minus and IC on a NaOH P extract. The "undetected" P was calculated as total P the sum of OP and PP. 99 Figure 5-1: Sorption data for OP and PP in eight Australian soils and fitted curves derived.tritrg the Freundlich equation. For soils treated with PP, curves are + plotted for bõth PP in solution vs. PP sorbed and the sum of OP and PP (OP Þe; itr solution vs. OP and PP sorbed (OP treatment -l-, PP treatment -n-, PP treatment with measurement of OP and PP -r-)...... "' 110 Figure 5-2: Relationship between P buffering index (PBI +,on) and (A) the partitioning coeffiòient (Kd L/kg) and (B) the Freundlich sorption constant (Kf). Op treatment -l-, PP treatment -tr-, PP treatment with measurement of OP and pp -r-. For (A) óe 5O.tZ x+87.4í,R2:0.30, PP 55.2lx +689.25, rtr:0.58, PP with OP and PP measurement 51.89x+456.05,R':0'25' For (B), OP y:l.1 lx+33.72,R2:0.96, PP )=0.16x +555.83, R2:0.003, PP with OP and PP measurement 51.1 lx+169.43,n^2:0.40. "" 114 Figure 5-3: Percentage of total P in solution present as PP when 150 mg P/L (150 PP) and 300 mg p/L (300 PP) as PP were added to each soil. 150 PP treatment -¡- 116 and 300 PP treatment -r-' """"" Figure 5-4: Cation concentration in solution (mg/L) in eight soils A-H, as a function of P in solution (mg/L) where P was supplied as OP or PP. Cations measured were calcium (Cà)ãn¿ iron (Fe). OP treatment Ca -l-, PP treatment Ca -r-, OP treatment Fe-0-, PP treatment Fe -¡-. """"" 119 Figure 5-5: Concentrations of DOC in solution (mg/L) as a firnction of P in solution (mdl) in eight soils (A-H) (OP treatment -0-, PP treatment -n-)...... I21 Figurè 6--1:'Distribution of P added as OP and PP in poþhosphate fertiliser amongst ïir'"#tÏ#:iräï:ä'.:Lt""-"igll'",orotio' 32PP radioactive isotopes and "OP'"""" """129 Figure 7-1: Total soluble phosphorus (P) concentration (m_dkg) after treatment with four rates (100, 500, 1000 and 2000 mg Plkg soil) of PP in three soils: a Grey Calcarosol (GC), a Red Calcarosol (RC) and a Sodosol (Sod). PP rate, soil type the and treatment iáieractions were significant ef0.001). Columns appended by same letter were not significantly different (PSO.001). ...'...... 140 PP.added Figure 7 -2: p in solution (PÞ and PP hydrolysed to OP) as a proportion of (%) after treatment with four rates (100, 500, 1000 and 2000 mg P/kg soil) of PP in ihree soils: a Grey Calcarosol (GC), a Red Calcarosol (RC) and a Sodosol

VII (Sod). PP rate by soil type interactions were significant (P S 0.001). Columns àppended by the same létter are not significantly different (P < 0.001)...... 141 Figure ?-3, Mu*i-.t* potential PP recovered and actual PP recovered (1og mdkg) after treatment wilh four rates (100, 500, 1000 and 2000 mg P/kg soil) of PP in three soils: a Grey Calcarosol (GC), a Red Calcarosol (RC) and a Sodosol (Sod). Dotted lines reprãsent nominal addition rates of P. For maximum potential PP recovered in a given soil the columns appended by the same letter are not significantly different (P I 0.001). For acluallabile PP in a given soil, columns appended Uy ttre same letter are not significantly different (P S 0.001)...... --...143 Figure l-+,uyaiolysed pP (1og mg/kg) measured by calculating Eq. 7-5 (Hyd PP calc) and by túe difference between PP recovered (mglkg) and PP remaining (mglkg) (Hyd PP diff) after treatment with four rates (100, 500, 1000 and 2000 t"g flt g àoit¡ of fe in three soils: a Grey Calcarosol (GC), a Red Calcarosol tntl an¿ a Sodosol (Sod).. In a given soil the columns appended by the same letter are not significantly different (P < 0.001). """""144 Figure 7-5:P g) measured by calculating Eq. 7-5 (Hyd pp calc) PP recovered (mg/kg) a.ll{ P! remaining (*df.gj fittedline5I.02x-33'64,R':0'83' 145 Figwe 8--1: ñ,ecovery of soluble phosphorus expressed as a percentage of P added (OP: orthophosphate, PPlyrophosphate), incubation time (l D: 1 day, 3D:3 àays and 7D:7days) and soil; GC : Grey Calcarosol, RC : Red Calcarosol and Sod: Sodosol. The P species by soil type by incubation time interaction was significant (Fpr I 0.001). Columns appended by the same letter are not silnificantly different (P f 0.001) """"""""' 154 nigure 8-2: Change with time in concentrations of calcium (A), iron (B) and aluminium (-C) in solution with different fertiliser treatments (control, orthophosphaie and pyrophosphate) in the Grey Calcarosol (GC), the Red Calcaìosoi (RC) andthe Sodosol (Sod). For (A), (B) and (C) the P species by incubation time by soil type interaction was significant (Fpr I 0.001). Columns appended by the same letter are not significantly different (Fpr < 0.001)...... 157 Figure 8-3' Chuttge with time in concentrations of DOC with different P treatments in the Grey Caicarosol (GC), the Red Calcarosol (RC) and the Sodosol (Sod). The < p speciés by incubation time by soil type interaction was significant (Fpr o.obr¡...... "" 15e Figure 8-4: E-values as a Yo of P added for as oP or PP over time (mglkg) in three soils, a Grey Calcarosol (GC), a Red Calcarosol (RC) and a Sodosol (Sod)' The labile r poól for op (oP) and PP (PP plus oP from PP (hydrolysed). The P species by incubation time interaction was significant (Fpr 5 0.001). Columns appended by the same letter in a given soil were not significantly different.....161 Figure ô-t Co-parison of the P E-value (mg P/kg soil) for native P and for native P with PP' added in a Grey calcarosol, a Ferrosol and a sodosol. .-.---t74 Figure l0-1: The reactions of P added as OP and PP in poþhosphate fertiliser- P will distribute amongst the solution, exchangeable and fixed P pools. In these pools P will react with F1, 41, Ca and DOC. Processes requiring further investigation include reactions over time, diffusion, rhizosphere and plant uptake processes, multi-nutrient fertiliser reactions, and release of native P from DOC.....""""" 183

VIII List of Tables

Table 1-1: Proportion of organic P species in soils (Stevenson and Cole 1999)..------.21 Table 1-2: Global use of common phosphate fertilisers (Hedley and Mclaughlin 2005). ----.--23 Table 1-:: Adsorption maximum and bonding energies of OP and PP in a range of soils (Sutton and Larsen 1964; Al-Kanani and MacKenzie l99l) .....28 Table 1-4: Log equilibrium constants (K) for reactions of phosphates with Fe and A1 ions [uÈAC zoos¡...... """"3s Table 1-5: Solubility products þKsp) for 41, potassium (K) and Fe phosphates (Sample et al.1980)...... ""'38 equilibrium association constants for reactions of phosphates (POa) Table 1-6:-Log t*¡ with calcium ions (Ca at25"c (Gunnar Sillen 1964). """""""""'40 Table 1-7: Reaction products of PP (PzOz) and TP (P¡Oro) with Ca in soil (Philen and Lehr 1967) """""""""43 Table 1-8: Solubility products þKsp) for Ca phosphate compounds (Sample et al' 1es0). Table 1-9: Phosphate compounds commonly found in fertilisers and compositions of their saturated solutions (Sample, 1930)...... """""""""50 Table 1-10: Half-life values of the reactions of TP and more condensed phosphates to both PP and OP. Table 2-1: Operating parameters for the IC system...... '-.... 65 Table 2-2: Gradient profile for the IC A516 column 65 Table 2-3'.Detection limit, quantification limit and minimum working concentration for P (mg/L) using IC and colorimetry """""69 Table 2-4: Cimparison of soluble OP in 1:10 soil water extracts from a Sodosol incubated for I day andT days as determined by IC and colorimetry...... ' """71 Table 2-5: Recovery (%) of standards of OP, PP and TP post-digestion as measured using colorimetry """"71 Table 2-6:Interactions of P species measured by colorimetry. Concentration ratio of each P species 1:1 in every combination, with 0.1 mg P/L of each species added in everycase...... """""""""'72 Table 2-7: OP and PP concentration (mg P/L) determined by IC and colorimetry in saturated solutions of cationic PP complexes... 73 Table 2-8: P, Al, Ca, Fe and Mg concentration (mg/L) measured by ICP-AES. """"74 Table 4-1: Chemical shift (ppm) and percentage of total signal found in the central band for Na- and Ca- OP and PP salts...... """""""""'94 31P Table 4-2: Observability of P in "P CP (P.6.-CP) and DP (Pou.-DP) NMR spectra- results of spin counting. """""""'97 Table 5-1: Soil sites and classification...... "' 105 Table 5-2: Soil characteristics...... 108 Table 5-3: Freundlich sorption parameters where P sorbed (mg/kg): Pi kf with a fit of rtr. Pi is the concentration in solution (mg P/L), Kf and n are Freundlich sorption parameters ...... 1 1 1 Table 5-4:Pafürjioning coefficient between the soil solid and solution phases (Kd) for OP, PP, and OP and PP at a P addition of 10 mg/L.. 1t2 Table 5-5: Correlation matrix of soil characteristics and P sorption parameters Highlighted r values are significant at P < 0.05...... 113

D( Table 5-6: Equilibrium pH values for the lowest (0mg P/L) and highest (150 mg P/L) levels of P applied as OP and PP. lIl Table 7-1: Comparison of relative specific activity (RSA) in the PP fraction (Llmg), 32OP t'PP...... where PP was labelled with or ...138 Table 7-2: Soll suspension pH data in a Grey Calcarosol, Red Calcarosol and Sodosol at 5 rates of PP applied (0, 100, 500, 1000 and 2000 mg P/þ soil) t42 Table 8-1: Change with time in equilibrium soil suspension pH for the Grey Calcarosol, the Red Calcarosol and the Sodosol. The P species by incubation time by soil type interaction was significant (Fpr < 0.001, LSD:0.23) Treatments appended by the same lett different. '...... 155 33OP Table 8-2: Change with time in partiti of for the OP 3tPP treatment and for for the PP he P species by incubation time by soil type interaction was significant (Fpr 10.001). Treatments appended by the same letter were not significantly different ...... 163 Table 9-1: Solution P as a%o of PP added in the Grey Calcarosol, the Ferrosol and the Sodosol. PP treatment results in soluble P as PP and as OP hydrolysed from PP ...t71 Table 9-2: Soil solution pH measurements after a7 day incubation with either no P (control) or PP in the Grey Calcarosol, the Ferrosol and the Sodosol...... '...... 172 Table 9-3: Labile Hydrolysed PP as aYo of P added...... '.--.I73 31OP 32OP Table 9-4: Propagation oierror for Eq 9-7 where the and values were altered by plus or minus the standard error of the observation ...... 115 Table 9-5: Comparison of the specific activity (SA) (Bq/mg P) of hydrolysed P in solution, with the specific activity of the added PP fertiliser. 176

X Acknowled ments

I would firstþ like to acknowledge the enduring support of my supervisors throughout provided my candidature. I feel incredibly lucky for the opportunities I have been with. These opportunities have included: complicated and expensive experiments,

project at extra work opportunities to pay the bills, and representing the fluid fertiliser particular various meetings and conferences. I would like to thank E,nzo andMike in like for their patience and persistence in the earþ days when an IC and LC seemed who different things, and solubility products were my worst enemy! Thanks to Else

practical came on board later in the candidature. Else was always there with a approach to experiments and encouraged me to develop my communication of

the complicated experiments for a broader audience. I would not have made it \¡rithout

me to support of each of my supervisors who put considerable effort into helping make the project work, and who were always available by some form of

communication for advice on pretty much everything'

who Outside of the formal supervisory panel I would like to thank Ron Smernik,

me to helped me with NMR work and several matters of chemistry, and helped

prepare what will probably be my most efficiently created manuscript ever'

of P. I thank him for To Warwick Dougherty , a greatfriend and assistant in all matters

surviving the endless questions, particularly those that come about from being one

month ahead of me in thesis preparation.

Jason Thanks also to Bob Hollow ay, GangaHettiararchchi, Roger Armstrong and

Kirby for their various scientific contributions that have helped me throughout my

phD. For technical assistance through experiments that represented organised chaos, I

thank John Gouzos, Caroline Johnston, Adrian Beech, Sharyn Zma, Colin Rivers,

Mike'Williams, Sean Mason and Anna McBeath'

XI (Pty) I would like to acknowledge the financial support provided by Liquid Fertiliser of Ltd trading as Agrichem, the Plant Nutrition Trust and the Soil Science Society

South Australia for contributions to my scholarship and travel awards' they I feel privileged for the friendships I have made during my PhD' They know who

is very are and I don't want to miss anyone by naming people. Soil and Land Systems lucky to be endowed with such a friendly group of students who are very willing to help each other, and stop for a chat (sometimes very long chats!).

and supply I say a big thank you to my family who always support my endeavours practical advice on matters of agriculture. Special thanks to my sister Anna, who would come to work to keep me company at all hours'

Special mention must be made of my partner Travis, who has given me the varied confidence to keep coming back when times'were tough. His duties have been from formatting chapters that I can't read anymore, to staying on the phone while I made a dash to the car in the midst of a midnight laboratory power failure! *

I would like to dedicate this thesis to Dr Bob Holloway. He has provided me with

inspiration throughout my careef and his confidence in me is the reason that this

the agronomist was able to become a chemist. I am inspired by his passion for

Australian agricultural communify, his sense of humour and courage to try new

things.

XII Abbreviations

A1 aluminium APP ammonium pollPho sPhate Ca calcium CaCO¡ calcium carbonate CP cross polarisation organis ation CSIRO commonwealth s cientific and Industrial Research DCPD dicalcium phosPhate dihYdrate DI deionised water DOC dissolved organic carbon DP direct polarisation EC electrical conductivitY E-value isotopically exchangeable value Fe iron H2O water HzSO¿ sulphuric acid IC ion chromatograPhY ICP-AES inductively coupled plasma atomic-emission spectroscopy Kd partitioning coefficient M molar MWHC maximum water holding caPacitY N nitrogen NaOH sodium hydroxide NMR nuclear magnetic resonance OP orthophosphate PO+ phosphate P phosphorus 3tP counting using NMR Pobs percentage of sample P observable by spin PP pyrophosphate r radioactivity remaining in solution R total radioactivity introduced SE standard error SSB spinning side band TP tripoþhosphate XRD x-ray diffraction

XIII Abstract poþhosphates have been shown to offer substantial agronomic benefits over traditional granular phosphorus (P) fertilisers in highly calcareous soils of southern

Australia. With ongoing field investigations into the efficiency of polyphosphate fertilisers compared to fluid and granular orthophosphate (OP) fertiliser products, a need developed for detailed study of the mechanisms responsible for the enhanced efficiency of polyphosphate fertilisers in Australian soil types. polyphosphates provide an analytical challenge as they contain chemically different forms of phosphate compared to most fertilisers, where P occurs entirely as OP- An investigation was conducted into the most suitable method for the speciation, quantification and separation of the P species supplied in polyphosphate fertilisers.

While the conventionally used colorimetric technique was comparable to ion

chromatography (IC) for quantification of OP, it did not provide the speciation and

separation capabilities of IC.

poþhosphate fertilisers are thermodynamically unstable and hydrolyse to more

simple forms of phosphate: this can be induced both chemically and biologically. A

study was undertaken using IC to ascertain the effects of time, temperature and acidity

on the stability of polyphosphate fertiliser formulations. All of these factors affected

the stability of polyphosphate fertilisers and recommendations on storage of the

product and mixtures with other fluid fertilisers can now be developed. Stability of PP 3lP-nuclear in soils was assessed using solid-state speciation by solid-state magnetic

resonance (NMR). Hydrolysis of both solution and solid-phase PP could be quantified

using the NMR technique, and results using this method were compared to

conventional techniques, which extract OP and PP into aqueous phases and use IC to

assess the extent of PP hydrolysis. The concentrations of OP and PP determined by

XtV the extraction technique were lower than those determined by NMR, and consequently the proportion of undetected P was greater for IC than for NMR'

There is disagreement in the literature as to the differences in partitioning behaviour

(sorption/precipitation) of OP and PP in soil. A partitioning study was undertaken using the IC technique for aqueous P speciation. Retention of PP to soil solid phases

,ù/as much stronger than OP, and addition of PP to soil resulted in greater concentrations of OP in the equilibrating solutions, indicating a possible competition for P sorption sites between PP and OP.

33OP PP A double labelling technique was developed where OP was labelled with and

and was labelled with'2pP. This technique was used to investigate the effects of time

dual concentration on the lability and partitioning of oP and PP in soils. using this

labeling technique it was possible to determine the hydrolysis of added PP, and to

distinguish Op derived from hydrolysis of PP from native OP in soil. The dissolution

of dissolved organic carbon, iron and aluminium and the sorption/precipitation of

the calcium as a result of addition of PP were assessed and related to changes in

lability of p supplied as PP and OP. This double labelling technique was further

PP to developed to assess the possibility of mobilisation of native OP by addition of

soil

The findings of this thesis indicate that the hydrolysis reaction is pivotal to the

behaviour of the P species that constitute a polyphosphate fertiliser in soils.

Investigations of isotopic cxchangeability showed that while native P mobilisation

was not detected, slow reactions of PP in soil including sorption, potentially

desorption, and hydrolysis underpin the potential availability of PP in soil'

XV Declaration

of any other This work contains no material which has been accepted for the award

and, to the best of my degree or diploma in a university or other tertiary institution

or written by another knowledge and belief, contains no material previously published

person, except where due reference has been made in the text'

the university I give consent to this copy of my thesis being made available in

Library

published (as listed components of the research described in this thesis have been

contained below). The author acknowledges that copyright of published material

within this thesis resides with the copyright holders of those works'

Therese M McBeath I m-m

Publication arising from this thesis

McBeath TM, Smernik RJ, Lombi E, Mclaughlin MJ (2006) Hydrolysis_of , soll pyrophosphate in a Highly balcareous soil: aSolid-State 3lP NMR Study' 8 5 6 -862' S c i eic e S o ci ety of Am er i c a J ournal 7 0,

XVI Ghapter 1. General lntroduction and Literature Review

1.1. lntroduction

constant The low efficiency of phosphorus (P) fertilisers in Australian soils is a

account for concern to researchers and growers (Olson et at. l97I). Fertiliser inputs one third of variable input costs on the average Australian farm

(http://www.land.murdoch.edu.aulplantnutrition.html). The major nutrients supplied

second largest in these fertilisers are nitrogen (N) and P. Phosphorus accounts for the tonnage of nutrient sold to Australian grain growers after N. phosphorus inefficiency has been a severe yield limitation, in particular for growing

grain on the highly calcareous soils of southern Australia (Holloway et al. 200I).

in crop P Holloway et al. (2002) showed significant yield advantages and increases

(APP) as uptake when P was supplied in the liquid form as ammonium polyphosphate

of the compared to granular orthophosphate (OP) fertilisers. Previous investigations yield performance of APP fertiliser in North America and Europe have not shown a

fertilisers advantage when using APP as compared to conventional granular OP

et al' (Gilliam 1970; Miner and Kamprath l9/I;Khasawneh et al' 1974; Khasawneh initiated 1979;Sample et al. 1979; Parent et at. 1985a). These contradictory findings

processes the development of experiments to determine the chemical reactions and

that control the degradation and potential availability of APP fertilisers.

This introductory chapter reviews the literature regarding the reactions of

poþhosphate fertiliser in soils that was available to the author prior to initiating the

experimental phase. More recently published research is reviewed in subsequent in soil' chapters. The review outlines the current understanding of P chemistry

previous studies of poþhosphate fertiliser reactions in soils and solutions are

t7 discussed, and the potential effects of these reactions on crop productivity are described. The aim of this review is to demonstrate the key areas of research that need to be addressed in order to develop an understanding ofthe reactions of polypho sphate fertilisers in Australian s oils.

1.2. Phosphorus in the Environment and Agriculture

Phosphorus constitutes approximately 0.12o/o of the earth's crust. It is an essential element for all forms of life and is widely distributed over the surface of the earth, cycling within plants, animals, soil and water. The Australian landscape is commonly described as old and weathered, due to a long geomorphic and weathering history under a range of climatic conditions. Australian soils in their native state are generally

low in total P by world standards with an average soil P content of 0'03o/o P, as

compared to 0.04-0.l YoP for American soils and 0.045% P for English soils (Costin

and Williams 1983).

The optimum soil solution P concentration for plant growth is in the range of 0-003 to

0.3 mg/L dependent on the plant species. Maintaining this optimal P concentration

while restricting P in surface waters to minimise environmental impact is a challenge

to be considered by all soil managers. This challenge is addressed by managing the

soil p cycle and the chemical and biochemical processes within the soil P cycle.

The major processes within the P cycle which determine the distribution and

bioavailability of P in the soil are: dissolution of soil mineral phosphates, retention of

p by inorganic soil constituents, mineralisation of organic P contained in organic

matter, and immobilisation of P by the soil microbial biomass and plant uptake

(Tisdale et al. 1985) (Figure 1-1).

l8 Stabþ Stabþ Plants Organic P lnorganic P

Phnt Ræts Prirary P Resisbnt Minerals Oganic P

Microbial Secmdary P P Minerals Soil Solution lnoryanic P I OqanicP

I Agglegate Ploteded OEanic P Ocduded P

i ta¡¡le : Labile Labile Organic : lnorganic P : P

LEGEND:

; lnoganic P Organic P

and X'igure 1-l: The P cycle in soils. The boxes represent pools of P forms in the cycle arrows represent movement and transformations between pools (Burns and Slater 1982).

1.2.1. Phosphorus species in soil

The p species in soil can be placed into the following classes:

1. Soluble inorganic and organic compounds in soil solution;

2. Weakly adsorbed (labile) inorganic phosphate;

3. Sparingly soluble phosphates of

a. Calcium (Ca) and magnesium (Mg) in calcareous and alkaline soils of arid

and semiarid regions;

b. hon (Fe) and aluminium (41) in acidic soils;

4. Organic forms

a. As part of living soil organisms;

b. As part of the non-living soil organic matter (humus)'

t9 Sp aringly s oluble phosPhates

The bulk of soil P (>g0%) occurs in insoluble or fixed forms as primary phosphate minerals, organic P, insoluble phosphates of Ca, Fe and 41, and P fixed by colloidal oxides and silicate minerals (Stevenson and Cole 1999)'

The native p in soils is derived from soil- forming parent materials. Apatite minerals

be are complex compounds of tricalcium phosphate 3[Ca3(POa)2]'CaX2 where X can

These C1-, F-, OH-, or COI- (chloro-, fluoro-, hydroxy and carbonate-apatites). minerals are highly insoluble in water and the P in the mineral is not readily available to plants. During the processes of weathering and soil development, P in apatite can be released and subsequently either absorbed by plants, be incorporated into organic matter, and/or form sparingly soluble mineral forms, such as secondary Ca-, Fe- and

A1- phosphates.

Labile inorganic phosPhates

Inorganic P in soil primarily occurs as OP compounds of Ca, 41, and Fe with trace

amounts of other cations. The principal water-soluble forms of inorganic P are

thus HzpO¿-, HpOa2-, and PO¿3-. They are interchangeable with variation in pH and

will be discussed in more depth in Section 1.5.1. Only a small fraction of total soil P

occurs in water-soluble forms at any one time (Stevenson and Cole 1999)'

Organic Phosphates

Inorganic P is converted to organic P through a series ofbiological reactions. Plants,

rnicro-organisms and animals require P as an essential nutrient, and convert some P to

organic forms within their cells. The organisms die and decompose, retuming

inorganic and organic P to the soil, initiating a biological P cycle (Anderson 1980).

The studies of Dalal (1977) and Feng et at. (2003) indicated that when organic P

sources such as phytate and lecithin are present in the soil at the same concentration

20 P as the inorganic P added, they make a similar contribution to plant growth and nutrition.

The primary groups of organic P compounds are inositol phosphates (e.g' phytate), phospholipids (e.g. lecithin) and nucleic acids (e.g. RNA) and their degradation products, or derivatives (Dalal 1977; Stevenson and Cole 1999). Organic P

P contributes between 5 and 90Yo of total soil P and is an important potential source of for plants in many agricultural soils. The usual range of the proportion of organic P in

than various forms is shown in Table l-1. The sum of these forms in most soils is less

100%as a significant fraction of soil organic P occurs in undefined forms (Stevenson and Cole 1999) .

Table 1-1: Proportion of organic P species in soils (Stevenson and Cole 1999)'

Organic P species Proportion of total organic P in soil Inositol phosphates 10- 60 Phospholipids 1-5 Nucleic Acids 0.2-2.5 Phosphoproteins Trace Metabolic l.e Trace

lnositol phosphates are esters of hexahydrohexahydroxy cyclohexane, commonly

referred to as inositol. A variety of esters occur, the most conìmon being the

up hexaphosphate ester þhytic acid) existing as a Ca and Mg salt. Phytates constitute

to 60Yoof the organic P content in soils where they accumulate due to their low

solubilities, strong adsorption and slow degradation'

Nucleic acids constitute approximately 5% of the total organic P in soils (Dalal1977)'

Nucleic acids are found in all living cells. Two types are known, ribonucleic acid

(RNA) and deoxyribonucleic acid (DNA), each consisting of a chain of nucleotides. A

very small fraction of organic P in soil consists of simple molecules like nucleotides

2l and the low concentration of these molecules is due to their high degradability

(Stevenson and Cole 1999).

The phospholipids represent a group of biologically important organic compounds

0.5- that are insoluble in water but are lipophilic. Phospholipids account for between

group 7yo (meanvalue of I%) of total organic P in soils. Included in the phospholipid are the phosphoglycerides (40% of phospholipids) such as lecithin, and phosphatidyl serine. As a general rule, between 2 to 5Yo of the organic P of cultivated soils is present in the soil microbial biomass. This includes P contained in various species of bacteria, fungi and protozoa (Dalal l9l7). The P content of microbial cells has been (Dalal reported to range from 1.5 to 2.5Yo for bacteria and as high as 4.8Yo fot fungi

IgjT). The turnover of microbial P is generally ten times gteater at the root surface than in the bulk soil (Tinker 1980), indicating that a plant-soil interaction will ampliff microbial tumover

1.3. PhosPhorus Feñilisers

crop When there are inadequate levels of bioavailable P in the soil to maintain optimal

yields, the addition of P fertiliser is required to increase soil P supply to crop plants

for attainment of optimal yield (Sample et al. 1980). It is estimated that there are 5.7

billion hectares worldwide containing less than optimal levels of available P for crop

production (Hinsinger 2001). Over 800 million hectares of soil that contains less than

optimal levels of P for crop production are calcareous (FAO 2003). Australian soils

are inherently low in P, and the annual plant utilization efficiency of phosphate

fertilisers of between 5-25% is a constant problem to researchers and producers

(Olson et al. l97l). The annual use of P containing fertiliser in Australia is 500 000

tonnes (FIFA, 2005) with the cost of P containing fertiliser to Australian farmers

being approximately $600 million per year (Randall 1997). A variety of P fertilisers

22 afe available to Australian farmers in fluid and granular forms as OP and poþhosphates. These P forms can be combined with a variety of other nutrients,

granules, or within such as N and micronutrients, either within or coated on fertiliser fluid formulations. The majority of P fertilisers applied in Australian broadacre (MAP) cropping are granular ammonium phosphates [mono-ammonium phosphate

a recent and di-ammonium phosphate (DAP)1. The fluid forms of P fertiliser are only

the introduction to dryland agriculture in Australia. The trends in P fertiliser use over

the last thirty years are illustrated in Table 1-2. There has been a shift towards increasing use of high analysis P fertiliser with a single product containing both nitrogen and P.

2005)' Table 1-2: Global use of common phosphate fertilisers (Iledley and Mclaughlin

Tlpical P Proportion of Total P Proportion of Total Analysis Fertiliser Use (%) P Fertiliser Use (%) Product t973174 1998199 r.7 Phosphate rock direct 9-t5% 5.5 20.9 Single superphosphate 7-10% 24.6 6.5 Triple superphosphate 20% 1 1.1 NPIIPK Compounds 34.6 20.9 42.2 Ammonium PhosPhates 20-23% 13.8 4.7 Other Nitrogen PhosPhates 3.2 J Others 7.3 Total 100 100

for crop Southern Australia is a region characterised by less than optimal P available

production. In Victoria for example, S5%o of grain production occurs on

alkalinelcalcareous soils (www.dpi.vic.gov.au/dpi/vro/maps 2005). In South

Australia, Eyre Peninsula is a region of highly calcareous soils, contributing

region approximat ely 41o/oof the state's grain production. The calcareous soils of this

2003). These consist of |9-87%caco: (w/w) (Holloway et al. 2001 Bertrand et al.

reactions levels of carbonate reduce the efficiency of P fertilisers due to rapid fixation

23 The inefficiency of current P fertilisation strategies has initiated the testing of poþhosphates and other liquid fertilisers. poþhosphate fertilisers have been widely used in broadacre agriculture in other parts of the world but their introduction to Australian agriculture is very recent. Preliminary evidence obtained by Minnipa Research centre and cslRo Land and water in

Adelaide suggests that polyphosphates may offer substantial agronomic benefits over traditional granular P fertilisers in calcareous soils. For example, in field trials on

wheat Eyre Peninsula in 2002 ammonium poþhosphate (APP) provided 36Yo greater grain yield than granular P fertiliser applied at arate of 8kg Plha on a red calcareous soil (Figure 1-2) (Holloway et aL.2002).

Emerald Rise 2002' Krichauff Wheat 8.6% GaCOs 38 mg/kg Colwell P

1.4 a (ú 1.2 .C 1.0 o Þ o ...o-"t'6 I 0.8 .JOttt'9"tt-tttttO .9 t-ttDtttttO' 0.6 .= rú 0.4 o APP (, 0.2 o Granular 0.0 024681012141618 P rate (kg/ ha)

polyphosphate tr'igure* 1-2: Phosphorus rate (kg/ha) vs. grain yield (t/ha) for ammonium (ApP) and granulaiP fertiliser (Granular) (Ilolloway et aL 2002).

poþhosphates are chemically different to most other forms of phosphate fertiliser in

which the P occurs entirely as OP. At the point of sale in Australia, polyphosphate

fertilisers commonly contain around 30-40% OP,4}-5}yopyrophosphate (PP) and the

remainder as tripollphosphate (TP) and other more complex phosphates' A

24 polyphosphate is a dehydrated form of OP. The dominant P ions in pollphosphate fertilisers are shown in Figure 1-3

Phosphate Triphosphate (Orthophosphate) (Tripolyphosphate)

Diphosphate Poþhosphate (þrophosphate)

2000)' Figure 1-3: Common ions of P in polyphosphate fertilisers (Rashchi and Finch

to more Polyphosphate fertilisers are thermodynamically unstable and hydrolysis

As PP simple forms of phosphate can be induced both chemically and biologically. most constitutes 10-g0%of the pol¡phosphate in APP (Khasawneh et al' 1974), the

PP to OP common hydrolysis reaction of polyphosphate fertiliser is the conversion of

1.4. Reactions of P in Soil

1.4.1. Ghemical Reactions

correct Throughout the literature there is considerable confusion surrounding the

nomenclature for describing P retention reactions in soil. Several important

definitions are described below: readily Fixøtion-Reactions of P with the solid phase where the reaction product is not

exchangeable with the soil solution pool.

25 or a Absorptìon- Movement of ions and water into an organism (e'g' plant, bacteria) mineral phase.

Adsorption- The attraction of ions or compounds to the surface of a solid phase,

is through ion exchange and electrostatic athaction. The P ion is held in a pool that readily exchangeable with the soil solution pool'

Sorption-The removal of P from soil solution by concentrating it in or on the solid phase, irrespective of the mechanism (Bache 1964; Barrow 1999).

Precipitøtion-Theremoval of two or more components from a solution by their mutual combination into a new solid-phase compound' with Sequestration- Theformation of a coordination complex by certain compounds

ions metallic ions in solution so that the usual precipitation reactions of the metallic

are prevented (Rashchi and Finch 2000).

Fixation and Absorption Reactions

is The addition of p to soil results in a series of reactions. The initial rapidreaction

particles' This followed by a slower continuing reaction between sorbed ions and soil

is a slow reaction means that P fertilisers become less effective with time. Since it

solid phase reaction, it can continue in dry soil (Bramley et al' 1992)' The continuing

in thin reaction occurs in zones of crystal imperfection in interdomainal spaces' or

the loss of layers of silicate between adjacent micro-crystals of Fe oxide. In moist soil, or p from the exterior of the crystal imperfection is replenished by further sorption

redistribution of sorbed P across the crystal surface during continuing reactions

(Barrow 1983; BramleY et al.1992).

Adsorption Reactions

chemical The key elements that determine the potential extent of P sorption are: the

in the soil and physical nature of each sorbing surface, and the ionic environment

26 solution (Holford and Mattin gly 1975). Phosphorus-reactive compounds in soils are clay minerals, Fe/41 oxides, organic matter and calcite (Dalal 1971).

Phosphorus can absorb to surfaces and edges of hydrous oxide, clay minerals and

can be carbonates by replacing HzO or OH-. The bonds which form in these reactions

where monodentate (considered more labile) or bidentate. The first step is adsorption soil p accumulates on the surface. The second step is absorption where P diffuses into constituents (Bramley et al. 1992). covalent bonding is an inner-sphere exchange reaction between phosphate ions and or hydroxyl ions associated with the metal. A phosphate ion is positioned so that one two oxygen atoms are able to lose one pair of electrons in order to fill the outer of Fe/Al electron shell of metal atoms that atebonded with, and exposed at, the edges

in the oxides and minerals (Barrow 1933). The covalent bonding reactions involved

adsorption of P are illustrated in Figure 1-4'

+ \, \oH \" loH2 'o'å M to'l /too ol o o //.o -H^O /',/7 ,27 "off :- / + o OP tonr =+oH Ho/ OH rol ' orfrr" \ \oH \" ,ôH ,o' ¡\+M \+ tor*" oHz oHz ./ ,/" ava¡lable P LaÞ¡te P DiftlculttY

tr'igure 1-4: Covalent bonding reactions of P (stevenson and cole 1999).

PP in several studies have been conducted to compare the sorption affinity of oP and (Blanchar arange of soil types possessing a range of physico-chemical characteristics

1985; Al-Kanani and Hossner 1969a;Hashimoto et at. 1969;Mnkeni and MacKenzie

maximum and MacKenzie l99l). In the study of Hashimoto et al. (1969) the sorption

the and bonding energy of PP were higher than those of OP in soil. In contrast,

findings of Sutton and Larsen (1964) and more recently Al-Kanani and Mackenzie

than (1991) show that the sorption maximum and bonding energy are higher for OP

27 PP (Table 1-3). Experimental conditions that influence the outcome of sorption include: incubation time, temperature, method of shaking, solution to soil ratio,

(Barrow supporting electrolyte and moisture content of the soil prior to equilibration lgTg),and this could partly account for the disagreement in results between authors

(Al-Kanani and MacKenzie l99l).

Table 1-3: Adsorption maximum and bonding energies of OP and PP in a range of soils (SuttonandLarsenlg64;Al.KananiandMacKenziel99l).

Soil Particle Sorption Sorption Bonding Bonding Texture size Maximum Maximum Energy Energy range (mmol P/ (mmolP/ (tn'lg) (-'le) (pm) ke) ke) PP OP PP OP st. 20-2.0 2t.6 38.4 0.s7 0.99 Loam Bernard 2.0-0.2 20.4 39.8 1.1s 1.68 Dalhousie 20-2.0 23.3 35.4 0.7 1.89 Clay 2.0-0.2 27.4 40.1 0.86 2.71 Kaolinite 20-2.0 5.5 10.6 0.65 0.98 Mineral 2.0-0.2 8.9 15.3 0.87 r.64 Goethite 20-2.0 28.2 43.2 0.77 2.40 Mineral 2.0-0.2 30.3 47 1.25 2.89 Davidson 167 64.4 t2.2 3.6 Clay loam Edina 67.3 28.5 5.7 4.3 Silty clay loam Hartsells 16.4 t3.4 t4 4.2 Loam Haysville 88.7 46.6 7.8 5.6 Loam Norfolk 22.8 23.2 2t.6 4.3 Sandy loam Kaolinite 13.3 11.0 8.2 0.8 Mineral Gibbsite t29 r23 3.9 0.7 Mineral Goethite 642 422 1.3 0.5 Mineral

is also A lower affinity sorption reaction is electrostalic attraction (Figure 1-5), which

in the described as an outer-sphere reaction. Phosphorus ions are held electrostatically

diffuse layer of cations and anions that balance the permanent or pH dependent charge

Fe and on the surface of clay minerals, hydrous oxides and oxides of predominantly

Al and organo-metallic complexes (Barrow 1983).

28 + -/o o FëOH Fe /l 2 Hzo P o o OH OH /7 HO + P -oH HO/ \oH FeOH FeOH tr'igure 1-5: Electrostatic attraction reactions of P with tr'e (Stevenson and Cole 1999)

Precipitation Reactions

Amer and Mostafa (1981) suggest that precipitation of a new compound requires nucleation, deposition of an amorphous phase, aging to small crystals and growth of these crystals, and that solubility decreases with each of these steps.

There are some precipitation reactions that are more easily reversed than others. Slow reactions of Fe/Al oxides with P following the initial adsorption reactions usually form insoluble precipitates (Brady and V/eil 1996). The various types of precipitation reactions for P will be described in Sections 1.5'2 and 1'5'3'

Sequestration Reactions

As applied to polyphosphate, sequestration is a mechanism by which poþhosphate

molecules can complex cations on their negatively charged binding sites if

deprotonated (Rashchi and Finch 2000). This mechanism has two major implications

It means that apolyphosphate solution can retain cations in solution such as Zn and

Mn at low concentrations, and can avoid precipitation of OP by ions such as Fe, A1

and Ca by decreasing the activity of these cations in solution.

1.4.2. Enzymatic Reactions

Enzymatic hydrolysis of organic P by the cleavage of P-ester bonds depends on the

stability and accessibility of the ester bonds, concentrations of compatible enzymes

and substrates in soil, and biotic and abiotic factors that inhibit enzymatic activity

(Moghimi et al. 1978; Joner et at.2000). The various forms of organic P differ in

29 > their lability to phosphatase enzymes with the declining order of phospholipids nucleic acids > phytate (Hirsch and Sussman l999;Feng et al. 2003). phosphatases are a group of enzymes found inside all living cells. Active exudation and microbial cell lysis introduces these enzymes into soil where most are rapidly inactivated. A small fraction will stabilise and persist in soil, remaining aclive, and are known as endogenous soil phosphatases. plant roots, fungi and bacteria possess phosphatase activity in the acid through to alkaline pH range. Intracellular phosphatase activity is highest in the alkaline pH range, and except for bacteria, acid phosphatases are mostly extracellular' Acid phosphatases lack substrate specificity as compared to alkaline phosphatases (Joner e/ al.2000). pyrophosphate is the most dominant pollrphosphate in APP fertilisers. It is also widely distributed in nature and has been reported in bacteria, insects, mammalian

tissues and plants (Tabatabai lg82) and in low levels in soils (Turner et aL.2003b).

pyrophosphate hydrolysis in soils can be enzyme catalysed by pyrophosphatase

(parent et at. 1985b). Pyrophosphatase has the ability to catalyse the hydrolysis of PP

to OP using PP as the substrate.

An example of the hydrolysis reaction catalysed by pyrophosphatase is (Lindsay

reTe):

H¿,PzOt*H2O ---+ 2]Hlt.O¿, Eq' 1-1

It is widely suggested in the literature (Sutton et al. 1966; Hashimoto et al, 1969) that

the hydrolysis of polyphosphates is required for the P in polyphosphate fertilisers to

be made readily available for plants as OP. It is also suggested that hydrolysis of

polyphosphates to OP is not a rate limiting step in terms of the ability of plants to take

30 up p from polyphosphate sources (Blanchar and CaldweIl1966; Englestad and Allen reTt). soil The work of Gilliam and Sample (1968) indicated that under sterile conditions, pH did not influence the hydrolysis of PP. The soils were sterilised by autoclaving

and and hydrolysis was measured in soils ranging in pH from 4'8 to 7 '2' Sutton

Larsen (1964) also observed a positive correlation between pH, biological activity

that and the hydrolysis of PP, which surpassed the effect of pH alone' This suggests

and the hydrolysis of PP in soil is affected by an interaction between soil chemical

biological factors. The effects of pH on the activity of corn root phosphatase and

pyrophosphatase activity ate illustrated in Figure 1-6:

200 . ACID PHOSPHATASE O PY?EPHOSPHATASE

;f- 50 ()Ê 4 UJã N r00 z. !l

o 34 5b( 89 aH OF RIIFFFR

Figure 1-6: Effect ofpH on (corn-root acid phosphatase and pyrophosphatas ase activity is expressed as pg p- nitrophenol rele and pyrophosphatase activity is 1968). expressed as pg Pi released/lO mg corn roots/hr (Gilliam and Sample

31 1.5. Factors controtling P Availability rn soils

The ability of soil to provide P to higher plants is determined by the soil solution

and Cole supply of inorganic P. This is influenced by a variety of factors (Stevenson

1999), including:

1. Soil pH; phosphate 2. The solubilities and sorption affinity of Fe- and A1- phosphates and

complexes with hydrous oxides and clay minerals in acid soils;

3. The solubilities of Ca-phosphate and P-minerals in calcareous soils, the rates

at which the minerals are solubilized, and sorption to CaCO¡l and micro- 4. Amount and stage of decomposition of organic residues, and activity of

organlsms

1.5.1. Soil pH

influencing The pH of a soil determines the dominant ions available for sorption by

(Dubus and the extent of ionisation of phosphates and the oxide surface charge

1-7: Becquer 2001). The ionisation of OP as a function of pH is shown in Figure

H3PO4 H)PO4- HPOur- POor- 100 t I t I I ¡ I I , t n t I I 9 I e I Ëso t c I t ô t U I I I t , t I I I t ,I I I

1996)' Figure 1-7: Ionic forms of P in soil solution as a function of pH (Brady and Weil,

is the dominant HpO+2- is the dominant ion in soil solution atpH>1.22 whilst HzPO+-

ion in soil solution at pH <7.zz.There is a ten-fold increase in the divalent HPOa,2-

32 ion with each unit increase in pH from pH 3-6 (Barrow 1999).The dominant anionic form of P in alkaline soils (HPO4'-) it taken up more slowly by plants than the dominant ion in acidic soils (HzPOa-) (Tisdale et ø1. 1985; Cresser et øl' 1993). Soll pH also determines the availability of Ca, Fe and Al ions to react with P' The solubility of Ca, Fe and A1 complexes formed with P will be discussed in further detail in Section 1.5.2.

1.5.2. lron and Aluminium

Reactions of lron and Aluminium with Orthophosphates

Aqueous Reactions solution complexes of oP in soil solution are presented in Figure 1-8. If ca, Fe or A1

the are present in soil solution, complexes with OP will form a small proportion of

total fraction of P in solution (Figure 1-8). While solution complexes of P do not

dominate the total P in solution due to the low solubility of phosphate compounds, it

is apparent from Figure 1-8 that they can occur and that they may influence the

activity of the P in solution (Lindsay 1979).

JJ 1.o

o.a c 9 t õ Ø .s o.6 o- o C .9 o 4 Lo l,!

OJ o s H PO42-

o. 2 H3POa'

o.o 3 4 6 7 a s pH

Figure 1-8: The effect of pII on the P species in soil solution (Lindsay 1979)'

A parameter used to describe reactions between P and cations in aqueous form is the

equilibrium constant which describes the ratio of the products to the reactants.

Therefore, the greater the value the more favoured the reaction is.

The equilibrium constant (K) for the reaction: Eq' l-2 Fe2* + Po¿3- <-+ FePo¿ is: K: (FePoa)l g"3+¡gol-)

The equilibrium constants for reaction products of Fe3* and 413* with OP and PP are

given in Table l-4. The data presented indicate that reactions of P with A1 are more

favoured than those with Fe. While reactions of Fe with PP are favoured over those

with OP, the reverse is true for Al.

34 Al Table l-4:Logequilibrium constants (K) for reactions of phosphates with Fe and ions (IIIPAC 2005).

Orthophosphate Pyrophosphate Cation Species Los K K FeHzL'II/Fe'H¡L 1.33 FeH:L/Fe'H¡L 6.43 FeHLiFe.HL 8.9s FeHzL/Fe'HzL 6.97 FeL.2H20 (s) K.o: -26.1 A13* AIHL/AI.HL 23.25 AlL/41.L 14.3-r7.17 Al}J2LlAI.H.2L 26.18 AlHL/Al.HL 19.2 AIL (s) Ko:-18.24 AIH2L/41 22.79 *(L) ligand, (Kso) solubility product, (s) solid'

Solid Phase Reactions

Sorption Reactions with The addition of granular P fertilisers results in a series of dissolution reactions soil water. Altematively P can be applied as a liquid fertiliser where it is already completely dissolved in solution. Following addition, a series of reactions occurs

that between P, soil constituents and non-P components of the applied fertiliser

reactions remove p from the solution phase and make the P less bioavailable. Sorption

P sorption occur in the zone of soil where the soil P concentration does not exceed the

maximum of the soil, and the soil solution Fe, Al and P concentrations do not exceed

P the solubility product of Fe and/or Al phosphate. The soil characteristics influencing

A1 andCa sorption include: clay content, organic matter content, pH, soil texture, Fe,

the extent content (Hedley and Mclaughlin 2005). An important factor determining

Positive of sorption reactions is the Fe and Al content of the soil (Sample et a\.1980)'

the interactions of Fe and Al content of soil and P sorption have been observed in (1985)' studies of Al-Kanani and MacKenzie (199I), and Mnkeni and MacKenzie

Hydrous metal oxides generally sorb more P than layer silicates and other crystalline

Al is forms of Fe and 41, as the occurrence of the more crystalline forms of Fe and

coatings less common. Hydrous oxides of Fe and Al occur as discrete compounds, as

35 layers on soil particles, and as amorphous Al hydroxyl compounds between of expandable Al silicates.

Precipitation Reactions

reactions Slow reactions of Fe/Al oxides with P following the initial adsorption usually form insoluble precipitates. These compounds are shown in Figure 1-9'

Precipitation reactions ,/oH "Ol AI

I o- o

I I rH,o Alt* + HO-P:o ,-:- HO-P:Q 4 llJ+

I t OH OH Dissolved ions Precipítated form

Figure L-9: Precipitates of tr'e and Al (Brady and Weil 1996)'

The relationship between pH and the complexes of P that dominate the soil

diagram in environment are often described using solubility diagrams. The solubility

Figure 1-8 indicates that the solubility of A1 with phosphates increases with increasing

pH' The pH. Therefore, complexes of Al with phosphates predominantly form at low

behaviour of complexes of Fe with phosphates is quite similar to that of complexes

with A1(Figure 1-10).

36 o

-3 ç o(L I 4 O o

(6) alPÇt4 (7) AIPO4'2Hp -6 (€) H6K3Al5(PO4)O' 18H2O (9) H6(NHa)3415(Poa)g' lBH2O

-7

-8 3 4 5 6 7 I I pH Figure 1-10: Solubitity diagram for Al phosphates (Lindsay et aL 1962)

The solubility product is defined as the ionic product when the system is in

equilibrium (LindsaY t97 9).

The equilibrium is described as follows:

MxAy (solid): xMv+ (aqueous) + yA*-(aqueous) Eq' 1-3

where M represents metal and A represents anion.

The solubility product Ksp is described by the following equation:

Ksp: (Mr).(A)v Eq'r-4

The values for the negative logarithm of the solubility products (pKsp) of some

fertiliser reaction products are shown in Table 1-5. It is important to remember that as

the value of pKsp increases, the solubility of the phosphate compound decreases.

37 Table 1-5: Solubility products (pKsp) for AI, potassium (K) and Fe phosphates (Sample et al.1980).

AIPOq'2HzO 2t.5-22.5 AINI{4(PO q)zOH.2HzO 57 AI2K(PO4)zOH'2HzO 55 A1s(NH4)3H6(POa)s' 1 8H2O 175.5 A15K3H6@O+)g'18H20 178.7 FePO+ 35.35

Reactions of lron and Aluminium with Polyphosphates

Aqueous Reactions

The ability of polyphosphates to coordinate',^/ith metals such as Fe and A1 determines their behaviour in soil by influencing the hydrolysis of poþhosphates, their ability to form complexes and their availability to plants (Hashimoto et al. 1969). Philen and

Lehr (1967) conducted an extensive study on the reaction products of polyphosphates

added to soil minerals. They found that any amount of Fe or A1 that was released by

soil minerals was sequestered by condensed phosphates and remained in solution,

even after considerable hydrolysis of the condensed phosphate to oP.

The work of Lindsay et at.(1979) showed that PP can form complexes with Ca, Mg,

Fe, Al and trace elements in aqueous solution. The filtered PP extract after shaking an

acidic sandy loam with ammonium PP contained no precipitates, which suggested that

soluble Fe- and AI-PP chelates had formed.

Solid Phase Reactions

There is some evidence for the formation of precipitates of Fe and Al in the presence

of pP. Sample et al. (1979) suggested that as the fertiliser P solution moves through

the soil, additional amounts of Al are sequestered until soluble Al exceeds the

solubility product of AI(NH¿) zPzOtOH.2HzO, which then precipitates.

Khasawneh et aI. (1979) suggested that the precipitates of condensed phosphates

occur in well defined zones as compared to the more diffuse precipitates of OP. When

38 polyphosphate and oP were added to a fine sandy loam, the precipitation of PP was

precipitated after 4 almost complete after 1 week. In contrast, oP had not completely

to be weeks. However, in this instance all alkali- extractable PP was considered

Khasawneh precipitated, as it was was not readily hydrolysable which in the view of et al. (1979) suggested that it was not reversibly adsorbed' (1967) contrary to the findings of Khasawneh et al. (1979), Philen and Lehr addition suggested that the reactions of soil with oP begin almost immediatelyupon

days to of p and are completed in amatter of hours whilst those of PP take several initiate and continue for weeks. In these experiments precipitation was measured using X-ray diffraction, infrared spectroscopy, and electron microscopy' The

conclusions differences in methodology may in part be responsible for the different

from these studies

Summary: Iron and Aluminium in The importance of Fe and Al for controlling the lability of P has been demonstrated

OP and PP' the description of sorption and precipitation reactions of Fe and Al with

Discrepancies between sorption studies suggest that an independent investigation of

P is a the partitioning of Op and pp applied to Australian soils in which availability

yield limiting factor is required. Furthermore, better techniques for the investigation

of the lability of oP and PP in a range of soils are required.

1.5.3. Calcium

Reactions of Calcium with Orthophosphate

Aqueous Reactions

or Soluble complexes, in the form of ion pairs or complex ions, such as CaHPO20

of the caPoa can exist at high soil solution pH and can represent a significant amount

p in solution (Figure 1-8, p 34). However, in most soil solution environments,

39 chemical dissociation of these solution complexes rapidly converts them to OP 'When (pierzSmski et a\.2005). dilute P concentrations are added to a soil solution containing high levels of CaCO3, the dominant reaction is one of monolayer

is adsorption of P onto the calcite surface. However, when a concentrated P solution added, the dominant reaction is one of precipitation to a dicalcium phosphate dihydrate (DCPD) compound involving surface reaffangement of amorphous phosphate (Sample et al. 1980; Freeman and Rowell 1981)'

Ca. Table 1-6 shows the log of the equilibrium constant for reactions of P species with

The reaction of PP with Ca has a higher equilibrium constant and is therefore more favoured than that of more condensed phosphates. More condensed phosphates than

PP have a slightly more favoured reaction with Ca than OP'

constants for reactions of phosphates @Oa) with Table 1-6: Log equilibrium association2\ calcium ions (Ca zt25'c (Gunnar Sillen 1964\'

Constant Ca Orthophosphate 2.6 2-)) Pyrophosphate (PzOz 5.0 Tripoþhosphate (P:Oro5- 3.04 Po Àr., 3.0

Solid Phase Reactions

Somtion Reactions

A calcareous soil contains large amounts of exchangeable Ca and limited amounts of

kaolinite or hydrated oxides of Fe/41, allowing Ca to play a dominant role in

determining phosphate solubility (Castro and Torrent 1998). Free CaCO¡ in

calcareous soils can absorb P ions at low solution P concentrations without the

precipitation of ca phosphates (Hedley and Mclaughlin 2005). The specific surface

replaces area of CaCO3 controls P fixation reactions in calcareous soils. Phosphorus

adsorbed water molecules, bicarbonate ions and hydroxyl ions when it is adsorbed by

40 calcite, with adsorbing strength depending upon the solubility of the compound formed with the surface ca ions (sample et al. 1980; celi et al.20o0).

5.yan et al. (1955) studied P-sorption reactions with Fe oxides in several calcareous

the more soils containing 0.1- 66% CaCO3. They observed that Fe oxides, particularly el reactive forms have a dominant influence on P reactions in calcareous soils. Hamad al. (I992)made similar observations in their study where Fe oxides contributed to

et al- (2003) 40yo of p sorption in calcareous soils. However in the work of Berttand

properties in calcareous soils of southern Australia it was shown that the P sorption

the range of 1- were a direct function of CaCOr content. This study included soils in

STY'CaCOz.

Precipitation Reactions

Precipitation reactions of P with ca involve a nucleation process of ca phosphate

phosphate crystals. There is a surface reaffangement of amorphous phosphate into

heteronuclei. At small soil P concentrations, hydroxyapatite forms

(Ca1s(H2PO+)o(OH)z) and at high soil P concentrations, octocalcium phosphate forms

(caa(HzPO¿)o.sHzO). Hydroxyapatite is a less soluble compound than octocalcium

phosphate (Freeman and Rowell 1981).

The activity of P and Ca remaining in solution after precipitation depends on the

solubility product constant of the particular Ca phosphate species formed' The

solubility diagram of Ca compounds is shown in Figure I - 1 1 . This diagram also

Ca shows the lines for strengite and variscite, which illustrates the dominance of

phosphate reactions in the alkaline pH range.

4t o

-1

3 o fL I 4 Lo o (L -5 N T oì I 6

K Kàolinìte -a O Ou ãrtz G Gibbsite

I a I 3 4 5 6 7 pH

(Lindsay Figure 1-11: Solubility of Ca phosphates compared to strengite and variscite te79).

Reactions of C alcium with P olyphosphates

Aqueous Reactions philen and Lehr (1967) showed that PP and TP react with soil minerals at a slower

to this rate than Op and yield markedly different reaction products. The exceptions Philen were carbonates of ca and Mg, which reacted rapidly with all P species. when

is an and Lehr (1g67)reacted calcite with an ammonium TP solution, which

precipitates ingredient of APP fertilisers, they found that ca-NHa-TP and ca- NII+-PP formed.

Solid Phase Reactions

and the A range of compounds form when PP reacts with ca in soils (Table 1-7),

solubility product of some of these compounds have been determined (Table 1-8).

42 Table 1-7: Reaction products of PP (PzOz) and TP (P¡Oro) with ca in soil @hilen and Leh.r 1967).

Time(days) Salt pH when Reaction Products precipitation Started Ended or Minor 33065 2H2O 3HzO Montmorillonite (NHa)2 (P 6HzO (NH¿):HzPzOz 5 30 75 Mg(NH+)zP2O7'4H2O C a3 zO t) z' (NHa)aH2P2O7 8 30 90 Me(NH+)o(PzOt)z'6HzO Ca(NHa)2PzOt'HzO (NH+):HzP¡Oro 3 65 >165 M g(NH+)zH+(P zO t) z' 2IJ2O CazNH¿H¡(PzOt)z'3HzO (NHa)aH2P3O1¡ 4 50 >165 C a3 (NHa)2 (P zO t) z' 6HzO Mg(NH+)zP zOt'4HzO Oro 7.5 10 >165 Calcite Tripolyphosphate 3- <1 Ca(NH¿)¡P zOrc'2HzO Ca(NHa)2P zOt'HzO 7.5

Table 1-8: Solubility products (pKsp) for Ca phosphate compounds (Sample er aL 1980)'

Calcium CaHPO+ 6.66 CaHPO+'2HzO 6.s6 CasHz(PO¿)o.5HzO 93.81 Caro(PO+)o(OH)2 111.82 Caro(PO¿)oFz r20.86 o 39

El-Zahaby et aI. (1982) suggest that dicalcium phosphate dihydrate or hydroxyapatite

do not initially form in a calcareous (7-37% CaCO3) soil when reacted with

ammonium phosphates containing l}%P as PP. According to their study, the

inhibitory effect of PP on Ca-P precipitation increases proportionally with the P ratio (Figure of pp:Op, but it is a short-term inhibition and eventually precipitation occurs

r-r2).

43 r-0 oP,lo-5 N + 0.5 ¿.-4'--o--{ 0.s + 0.8 -+ 1.0 - -+' 0.5 + NôPP o '+' 0,8 + NoPP - 0.6 -+- 1.0 + tloPP J .9 !Þ É o.o !0 U :---*-:- oo- 02

0.0 o gt d o ô¡ ç -e*ñ 3N EHH Ti,n. . hour

OP tr'igure 1-12: OP concentration (0.001M) as a function of time (hour) for three (Amer solutions reacted with calcium carbonate with and without Na-PP and Mostafa 1981). presence It is suggested that the mechanisms for inhibition of precipitation of P in the (CaCO¡) of pp involves nucleation inhibition and distortion of the calcium carbonate pH when surface. In the study ofqI-Zahaby and Chien (19S2) PP lowered the solution oP sorption by caco¡ was inhibited by PP, which suggests that the caco¡ surface

seems likely was blocked by PP so that caco¡ could not contfol the solution pH. It

the that the lower solution pH with PP resulted from the presence of HzPO+- while caco¡ surface was being blocked (El-zahaby and chien 1982). GEOCHEM modelling carried out by Mclaughlin et al. (2003) showed that Ca-PP has a lower solubility than ca-oP species and that at ap[of 8, increasing the concentrations of result polyphosphate ion causes a dissolution of dicalcium phosphate dihydrate' This was likely because the solubility product (pKsp) for Ca-PP in the GEOCHEM

database is l4.7,while pKsp values for the Ca-OP compounds (DCPD, octocalcium

44 phosphate and p-tricalcium phosphate) were much higher. Solution complexation of

Ca by the PP ions resulted in dissolution of DCPD'

The DCpD crystals formed in the presence of PP are smaller than those formed in the (El- absence of PP (found after incubating monocalcium phosphate with CaCOt

Zahaby and Chien lgï2).It appears that PP acts as a crystal growth inhibitor rather than as a nucleation inhibitor when DCPD precipitates as discrete crystals in the solution rather than precipitating on the CaCO¡ surface' The formation of microcrystalline DCPD from monocalcium phosphate with PP results in higher soluble P contents in calcareous soils even after the precipitation inhibition effect disappears (El-Zahaby and Chien l9S2). These findings suggest that the use of ammonium polyphosphates on calcareous soils where P is present both as OP and PP may provide significant benefits in the reduction of P fixation.

Summary: Calcium

A review of the reactions of Ca with OP and PP shows that these reactions are quite

different to those of Fe and Al with OP and PP. Research to date indicates that PP

possibly provides protection of OP from reaction with the solid phase. Therefore, it is

possible that the lability of P is greater when supplied as PP as compared to Op in Ca-

dominated soils. This hypothesis requires further investigation. The development of a

technique for tracing PP and its reaction products is needed to enhance the current

understanding of these processes.

1.5.4. Organic Residues and Micro-organisms

Many studies have shown that the availability of phosphate in soil is enhanced by

additions of organic residues. Several reactions may be involved:

1. Low molecular weight organic acids are exuded from plant roots or micro-

organisms, which lower the rhizosphere pH or accelerate the dissolution of

45 sparingly soluble phosphate minerals by complexing the metal cation of the

mineral (Grinstead et al. 1982; Jones 1998). The mechanisms involved are

depicted in Figure 1-13.

2. Organic acid anions reach an adequate concentration in the rhizosphere

enabling them to compete effectively with OP for adsorption sites on Fe and

Al oxide surfaces (i.e. anions such as citrate andtartrate). Organic anions

reduce the bonding energy of co-adsorbed P and decrease the proportion of

high energy sorption sites, increasing the proportion of labile P on low energy

sites. This will result in an increase in plant availability as there is a rapid

equilibrium between low energy adsorbed P and solution P (Holford and

Mattingly 1975).

3. Humates may form a protective surface over colloidal sesquioxides, with a

reduction in P fìxation.

4. Solubility of Ca- and Mg- phosphates may be increased through production of

carbonic acid from carbon dioxidereleased during mineralisation.

5. Fresh organic matter may have a priming effect on the degradation of native

humus, with mineralisation of organic P (stevenson and cole 1999).

6. Phosphohumate complexes may be formed.

46 Decay cf cgaic Leaclates of Plart Excretion ftom Plalt residues by residæs Pots microogarisms

ORGANIC ACIDS cif ic, oxal ic, 2-ketogl uconi c ard others

lrsolóle Ca, Fe and Al Primay Mirerds phæphaÞs

Mdd clelate comple

Sduble phosphdes

Figure 1-13: Interactions of organic P in soils (stevenson and cole 1999).

the dominant As discussed previously, inositol phosphates, particularly phytates, are

form of organic P in soils. The hydrolysis rate of phytate is largely determinedby Na- solubility. Phytate solubility is pH dependent and generally decreases in the order

phytate salts, the three latter are most common , Mg-, Ca-, Fe-, A1- phytate. Of these

in soil and, apartfrom Ca-phyt ate atpH values< 6, they are sparingly soluble' for More complex organic P in soil organic material becomes increasingly available

C mineralisation as C is mineralised and C: P decreases. The factors that determine lignin mineralisation are the çrality of the organic material (C: N and C: P ratios,

(temperature, water' content etc.) and physical factors influencing microbial activity

oxygen and pH (Dalal 1977).

must be As mentioned above, to become available for plants, organic P compounds

may be of hydrolysed by a group of enzymes called phosphatases or phytases which

47 plant or microbial origin (Feng et at.2003). The work of Giordano et al' (1971) compared the solubilisation of organic matter around a monoammonium phosphate fertiliser band as compared to a triammonium PP fertiliser band. up to 10% of soil organic matter was solubilised. On aveÍage,triammonium PP solubilized twice as much organic matter as MAP. The ability of PP to solubilise organic matter was demonstrated by Bremner and Lees (1949) where they used PP as a neutral extractant for soil organic matter. The potential for PP to solubilise substantially more organic potentially matter than Op is a process that requires further investigation. This process interacts with, and results in, differences in lability of OP and PP.

Sample et al. (1980) suggest that the solubilisation of organic matter may result in three possible outcomes:

1. As the fertiliser solution moves from the band of application, the displaced

metal ions contribute to the concentration of soil solution cations available for

precipitation with P;

2. Removal of the organic matter coatings on soil minerals may expose new

surfaces, which can participate in P adsorption or precipitation reactions; and

3. Solubilised organic matter is carried to a new location in the soil where it may

reprecipitate covering soil mineral surfaces, which may have otherwise been

involved in P retention reactions.

1,6. Reactions of P Fertitiser in soíls and solution

1.6.1. Orthophosphate Fertiliser Dissolution

Ca As shown in Table l-2themost common forms of granular OP are ammonium,

reactions and potassium phosphate. The granular phosphates must undergo dissolution

in order to release nutrients into the soil solution. Solubilisation of the granule is

48 achieved by moisture, which is predominantly supplied by diffusion and capillary flow forming an almost saturated solution. Movement of P from the saturated solution and the soil water is driven by an osmotic potential gradient (Huffrnan and Taylor

1963;Hedley and Mclaughlin 2005). The saturated solution does not remain in the granule, rather it is drawn out by matric suction supplied by the surrounding soil and comes into contact and reacts with soil minerals. The process of inward movement of solution continues until the concentration of solution is decreased by dilution or precipitation of the P such that there is no osmotic potential gradient (Huffman and

Taylor 1963).

Dissolution of the granule can occur atvery low soil water potentials. In most field situations the predominant limiting factor for the dissolution of granules is therefore the precipitation of fertiliser-soil reaction products and the adsorption of P by soil

(1954) surfaces (Hedley and Mclaughlin 2005). In the study of Lawton and Vomocil

soil moisture content was considered to be the most important factor determining the

rate of fertiliser dissolution. However their experiments were conducted on light-

textured (loam and sandy loam) soils of pH 5.0- 5.2 where P sorption is not likely to

dominate soil processes to the same degree as in a large proportion of the world's

arable soils. Furthefinore, comparisons between the effects of soil moisture and P-

sorption capacity were not made.

Benbi and Gilkes (1987) conducted a study of the dissolution of superphosphate in

two soils of differing sorption capacity. They found that after four weeks most P will

occur in a highly P-enriched region of the soil within about 50 mm of the application

point of the P fertiliser granule. Testing of a Laterite and a Podsol showed hhat 45Vo

(Benbi and and j2%orespectively of the P from the fertiliser granule had dissolved

Gilkes 1987).

49 The addition of MAP to a calcareous soil led to an initial rapid dissolution of P, after which the remainingl}-2}Yo of applied P did not diffuse out of the granules. The mineral phase detected in this non-soluble portion included poorly soluble crandallite

[CaAþ(PO4)2(OH)s' (H2O)] (Lomb i et al. 2004b).

The reactions that occur around the granule shell are dependent on both the composition of the soil and the fertiliser granule. The composition of the saturated solution that forms as a result of the dissolution of some common fertiliser compounds is shown in Table 1-9.

Table 1-9: phosphate compounds commonly found in fertilisers and compositions of their saturated solutions (Sample, 1980)'

Composition of saturated solution Compound Formula PHP AccompanYing concentration cation

water-soluble * Monocalcium Ca(HzPO¿)'HzO 1.0 4.5 CaI.3 phosphate (rPS) 4.0 Ca 1.4 1.5* (MrPS) Monoammonium NH¿HzPO+ 3.5 2.9 NH+ 2.9 phosphate Monopotassium KHzPO¿ 4.0 1.7 KI.7 phosphate Triammonium (Nt{¿)tHzPzOt'HzO 6.0 6.8 NH+ 10.2 pyrophosphate Diammonium (NH+)zHPO+ 8.0 3.8 NH+ 7.6 phosphate Dipotassium KzHPO¿ 10.1 6.1 K12.2

Soluble Dicalcium CaHPO+ 6.5 0.002 Ca 0.001 phosphate CaHPO+'2HzO C 6.s 0.00001 Ca 0.001 *Triple point solution (TPS), Metastable triple point solution (MTPS), calcium (Ca), ammonium OrfI4), potassium (K).

50 '1.6.2. Polyphosphate Fertiliser Hydrolysis in Soils and Solutions

Hydrolysis of Polyphosphates in Fertiliser Solutions

While gtanular APP products have been developed, they are more costly than the fluid formulation and therefore polyphosphate fertilisers are predominantly sold in the fluid form. A commercially available APP product commonly contains 50% of P in oP and poþhosphate forms (Mortvedt et al. 1999). Pollphosphates are thermodynamically unstable and hydrolyse in the presence of water.

Due to this instability, the investigation of formulation chemistry and environmental factors that influence stability is an essential component of the poþhosphate research effort.

For linear poþhosphates, stability decreases as the chain length increases up to ten phosphate units (De Jager and Heyns 1998; Rashchi and Finch 2000). Other factors that influence the hydrolysis of pollrphosphate in solution include temperaturo, PH,

enz)¡mes, complexing cations, total P concentration, OP concentration and ionic

environment in the solution (Van Wazer et al. 1952; Gilliam and Sample 1968).

The hydrolysis of pollphosphate chains is catalysed by heavy metal cations, the effect

being most pronounced with cations of high charge and small radius (De Jager and

Heyns 1998; Rashchi and Finch 2000). For example, Li* accelerates the hydrolysis

rate more strongly than Na+, as the ionic radius of Li* is very small compared with

that of Na+. Due to the smaller ionic radius, an inner-sphere complex (or contact ion

pair) forms more easily and promotes the nucleophilic attack of OH on P atoms (Kura

et at. l991).Fruzier and Dillard (1931) found that the presence of Fe and Al increased

the rate of hydrolysis of TP and more condensed phosphates. However, they

suggested that the hydrolysis reaction from PP to OP was in fact reduced by the

presence of Fe or A1 as the metal ions were sequestered by PP. The increase in

51 hydrolysis of TP and more condensed phosphates in the presence of Fe and Al is shown in Table 1-10.

Table 1-10: Half-life values of the reactions of TP and more condensed phosphates to both PP and OP.

Sample Half life (days) of pollphosphates wherePcontains>2 APP + l%oFezOz 29 APP + 2%oFezOt 18 APP + lYo Al2O3 31 APP + 2o/o AlzOt t9 APP + l%oFezOt+ lo/o AlzO: 19 APP 60 * (APP) ammonium poþhosPhate, (Fe2O3) iron oxide, (41203) aluminium oxide.

A kinetic study of the hydrolysis of PP in solution was undertaken by Subbarao

(1975) using laboratory gfade chemicals rather than solutions of commercial poþhosphate fertiliser. Understanding the kinetics of hydrolysis of fertiliser solutions is necessary to develop storage recoÍìmendations for Australian conditions.

This work needs to be carried out in relation to temperature, length of storage and

changes made to fertiliser composition through the addition of phosphoric acid.

Hydrolysß of Polyphosphates in Soils pH Hons e/ al. (1956) conducted a laboratory experiment using six soils ranging in

content, from 4.3 to 7 .9 to investigate the effects of soil moisture, organic matter

PP and TP' temperaturo, PH, texture and incubation time on the hydrolysis of

The results of this study suggest that increasing moisture and presumably biological

activity promoted hydrolysis, as did increasing temperature. As the level of the

the condensation of the phosphate increased (i.e. TP vs. PP), temperature and pH were

most important factors affecting hydrolysis reactions.

The importance of the biological activity component of hydrolysis of poþhosphates

and is supported by the work of Hashimoto et al. (1969), Khasawneh et al' (1979)

and Savant and Racz (1972). This relationship is illustrated in the work of Savant

52 and TP at Racz (1972) where non-sterile plants in non-sterile sand hydrolysed both PP a greater rate than sterile plants in sterile sand (Figure l-14)'

PYROPHOSPHATE 40

WHEAT PEA 30

ROOfS ROOTS STERLÉ E,og a.ol¡J *o +i¿o TRIPOLYPTIOSPHATE Êts o WHEAT €so PEA o-

o RooTs SÍÉ.RILE ROOTS STER\LÉ o o ¿ 3 40 2 34 DAYS

Figure 1-14: Hydrolysis of 50ppm P as PP and TP solutions (measured as OP formed) by sterilõ and non-sterile wheat and pea roots (Savant and Racz1972),

inverse Sutton and Larsen (1964) suggested that in the soil environment there is an

linear relationship between the log of the half-life of PP and both the log of the

biological activity and the pH of the soil. In their study, PP breakdown increased with

soil biological activity and soil pH (only measured up to pH 7'6)'

In disagreement with the findings of Sutton and Larsen (1964), Gilliam and Sample

(1963) showed that at a pH of greater than 6.5, hydrolysis of PP is inhibited, possibly

the due to the changes in the microbial population either slowing the efficiency of

be pyrophosphatases or changing the species responsible for PP hydrolysis' This may

related to other soil components that change with alkalinity i.e. Ca vs. Fe/Al which

influence the microbial ecologY.

53 There are opportunities to build on the current state of knowledge of hydrolysis reactions in soils due to the availability of more advanced speciation techniques.

Halliwell et al. (2001) used ion chromatography to speciate condensed P species and to trace hydrolysis reactions in wastewater samples. This technique could be applied to soil water extracts to monitor poþhosphate hydrolysis. Furthermore, solid-state

3lp nuclear magnetic resonance (NMR) has been used to speciate condensed P species 3lP in soils (Frossard et at. 1994b). Solid-state NMR could be used to track the hydrolysis reaction in a relatively undisturbed sample providing information about the rate of polyphosphate hydrolysis in the total soil volume including labile and non- labile P pools.

1 .7. Movement of P rn Soits

'When roots grow in soil, only small amounts of P willbe found at the root surface

initially, and after this amount is absorbed, the P is replenished by P moving to the

roots by mass flow and diffusion (Barber 1980).

Mass flow is the convective transport of nutrients dissolved in the soil solution from

the bulk of the soil to the root surface. The amount of P moving to the root by mass

flow depends on the P concentration in the soil solution. To determine the amount of

P that must be in soil solution to supply all the P to the plant by mass flow the

transpiration ratio is divided by the P concentration of the plant (Barber 1980;

Marschner 1995;Hinsinger 2001). Marschner (1995) cites the contribution of mass

flow to total crop P demand to average l-4Yo for winter wheat and sugar beet in field

trials.

If the mass flow and initial interception mechanisms do not supply sufficient P for the

plant, the P concentration will decrease at the root due to P absorption by the plant.

This results in a P concentration gradient in the zone directly adjacent to the root and

54 p will diffuse toward the root along this gradient. Diffusion is the main mechanism for the movement of P to the root surface. However, the proportion of P that is supplied by each mechanism will depend on the size of the root system, the P absorption characteristics of the root, the rate of water absorption by the root, and the levels of adsorbed and solution P within the soil (Bolan and Robson 1983; Marschner

1ee5).

P' The sorption characteristics of a soil can have a strong influence on the diffusion of phosphorus that is strongly adsorbed will not diffirse as readily as P that is not strongly adsorbed (Barber 1980). As diffusion is so important to suppþing P to plants, factors that govern the rate of P diffusion to the root and the extent of root growth are important in determining P availability to plants (Barber 1980)'

Dffision Reactions of Orthophosphate F ertilis ers

Following a series of dissolution reactions (section 1.6.1), P is released into the soil

solution and is potentially available for diffusion to the plant. The mechanisms of

diffusion upon the addition of a fertiliser granule are illustrated in (Figure 1-15).

55 soil particles

P ion

POres

c o l= f solution Water fìlm arrived by (Jo) mass flow Lower o U by diffusion (, (d ô- o L Distance from granule surface Precipitation SorPtion zone zone

granule of Figure 1-15: Movement of phosphate by mass flow and diffusion from a triple or single supeiphosphate through water-filled and water-lined large miõro-poreJitt ',n"nnggiegated soil. Penetration of P into aggregates is incomplete due "to the slow rate of P diffusion in smaller intra- aggregate micropores and discontinuous micropores (Iledley and Mclaughlin 2005)'

the One of the most important factors determining the diffusion of P following dissolution of a P fertiliser granule is interception by P sorbing soil factors. Lawton

of variation and Vomocil (1954) conducted a laboratory experiment to test the effect in soil type, soil moisture, compaction and soluble P level in soils on the dissolution

had a and migration of P from granular supefphosphate . They found that texture significant influence on the distance of diffusion of P from superphosphate' Diffusion

period was greater with coafse textured soil, and the maximum movement over a of

and four weeks was 2.5 cm in a loamy sand compared to 1.4 cm in a loam (Lawton

to sorb Vomocil 1954). A soil with a higher clay content usually has a greater ability not p due to the higher surface areaavailable (Sample et al. 1980) and hence will

(1954) was move as far. An advantage of the technique used by Lawton and vomocil that they used an isotopic tracer to detect diffusion of P and therefore these

56 measgrements were not susceptible to the variations caused by heterogeneity in native

P across the sample.

The work of Khasawneh et al. (1979) compared P movement in a sandy loam and a loam soil and showed that DAP was able to diffuse up to a distance of 5cm over a four week incubation, which is considerably further than in the study by Lawton and

Vomocil (1954). Investigations by Benbi and Gilkes (1937) demonstrated the advantages of banding of fertiliser P to improve both the diffusion and availability of p as a plant nutrient. Limited P diffusion restricts the accessibility and therefore uptake of nutrients by plants, and where nutrient supply is insufficient it is a yield limiting factor. Current agronomic recommendations are for banding of P fertiliser close to the seed in order to optimise P diffusion and therefore potential availability.

Dffision Reactions of P olyphosphate Fertilis ers

Hashimoto and Lehr (1973) conducted a study on the mobility of polyphosphate

fertilisers. They determined that in initial reactions after polyphosphate addition to

soil the poþhosphate fraction moves with the advancing front of fertiliser P along

the diffusion concentration gradient showingthatpoþhosphates have significant

mobility. Hashimoto and Lehr (1973) suggest that this is due to the greater length of

time before solid reaction products are formed. Therefore the polyphosphate should

affect a greater volume of soil before being retained by soil components'

ln contrast, the work of Khasawneh et at.(I979) showed that PP and poþhosphate

wcre hydrolysed as fast as, or faster than, the process of diffusion. Therefore virtually

all of the movement of P took place in the OP form. As a result the mobility of

phosphates from DAP, APP or triammonium PP were essentially similar in his work.

Consequently the rate of diffusion away from the fertiliser band was a function of the

rate of hydrolysis to OP. There are soils where the rate of hydrolysis is limited, for

57 example, by low microbial aclivity,high pH and high CaCO¡ content. In these soils proportionally more polyphosphate than OP diffusion will occur, and complexation reactions of polyphosphates will have a greater effect on the distance of diffusion

(Khasawneh et al. 1979; Hedley and Mclaughlin 2005). This will have an important influence on P reaction products and consequentþ P availability to crop plants.

1.8. Conclusrons and Future Direction

Research to date on the reactions of pollrphosphate fertilisers in soils and solutions has not been conducted using Australian soil tlrpes, nor on polyphosphate products commercially available in Australia. The research conducted so far on polyphosphate reactions in soils and solutions has the following limitations:

1. The quantification and speciation of P from pol¡rphosphates has often been

with colorimetric techniques where the polyphosphate concentration is the

difference between total P andameasure of initial OP. For the purpose of this

project a technique was required that had the capability to quantify P with a

high level ofaccuracy, to speciate the range ofcondensed P species in

polyphosphate fertiliser and to separate out the condensed P species.

2. Investigation of fertiliser stability in the past has used pure analytical reagents

rather than commercially available product and has not addressed the extent to

which acidifying the fertiliser product using phosphoric acid (to increase trace

element solubiliry) is detrimental to polyphosphate stability.

3. partitioning studies have not been performed over the range and combination

of soil characteristics that present limitations to P use efficiency in Australian

agriculture. The techniques used have not enabled the description of sorption,

hydrolysis and potentially mobilisation processes that occur simultaneously

where condensed P species are applied.

58 4. The differences in lability of the P species applied in a polyphosphate fertiliser

have not been investigatedusing radioactive labelling of OP and PP. This type

of investigation provides the opporfunity to further understand the chemical

processes that occur when condensed P species are applied to soil.

5. Degradation of poþhosphate fertilisers in soil has not been measured in the

solid state. Therefore all current knowledge about the stability of

poþhosphate in soil is derived from solution state methods.

As a result of highlighting the limitations of prior investigations of polyphosphate reactions, aÍaîge of research gaps or opportunities for further research have been identified. This project will focus on determining the chemical factors that control the behaviour of the P species in a commercially available product and in a variety of

Australian soils. This thesis will address the following research areas:

l. Evaluation of the methods available for the quantification, speciation and

separation of the P species in polyphosphate fertilisers in order to use the best

possible techniques for investigation of polyphosphate chemistry (Chapter 2).

2. Determination of storage factors including temperature, acidity and length of

storage, which influence the stability of polyphosphate fertiliser solutions, to

address concerns of industry and growers about the longevity and integrity of

polyphosphate fertiliser products (Chapter 3)'

3. Investigation of PP and its degradation products in soil in the solid-state using

3lP nuclear magnctic resonance spectroscopy (Chapter 4)'

4. Investigation of the partitioning of OP and PP in a range of Australian soils

(Chapter 5).

59 5. Investigate the ability of a dual labelling technique to trace the isotopic

hydrolysis and recovery of a range of doses of PP supplied to soil (Chapters 6

andT).

6. Study of the potential availability of P supplied as poþhosphates in a variety

of Australian agricultural soils. The techniques developed for the investigation

of potential availability will be used to evaluate the effects of time on the

lability of PP as compared to OP (Chapter 8). i. The possibilþ of mobilisation of native P through the addition of P as PP will

be investigated using the double labelling technique (Chapter 9).

60 Ghapter 2. Po hosphate iation in Solution

2.1. Introduction

Ammonium poþhosphate fertilisers have a unique chemistry compared to OP fertilisers, with up to 600/o of the total P in APP fertiliser present in the form of polyphosphates. The dominant poþhosphate found in APP is PP' The hydrolysis of

OP, condensed phosphates is of great interest as the ultimate breakdown product, does not possess either the sequestering or dispersing properties of the more

speciate condensed P species (Baluyot and Hartford 1996). Therefore the capacity to the dominant P species in poþhosphate fertiliser is required in order to study poþhosphate fertiliser chemistry in soils and solutions'

Many studies that report speciation data for P in polyphosphate fertiliser used extraction methods where the OP concentration is determined colorimetrically.

However, the determination of soil solution P ions in environmental samples can be confounded when using colorimetric methods due to the possible hydrolysis of

soluble organic and condensed forms of P in the acidic matrix (Masson et al.200l).

Total p is determined by acid digestion followed by colorimetric measurement of

total Op, and the polyphosphate concentration is calculated as the difference between

P, and OP (Sutton et al. 1966; Gilliam and Sample 1968; Khasawneh et al' 1979;

is Sample et at. 1979;Parent et al. 1985b). As the poþhosphate concentration

PP, measured by difference, all condensed P species are measured together including

Tp and organic P complexes. Furthermore, the use of acid digestion to determine

total p means that the apparent polyphosphate P concentration may be overestimated

due to inclusion of an unknown amount of organic P'

6l Colorimetric soluble OP determination is sensitive to interferences. These interferences include the presence of labile organic and inorganic P compounds dissolved in solution that arehydrolysed in the acidic matrix of the colorimetric technique. The hydrolysable compounds include polyphosphates, phytates and P

et associated with colloids such as Fe and A1 oxides or humic substances (Masson al.

2001). There are other soil solution components that can interfere with colorimetric determination of phosphate including silicates, arsenate, colloids and nitrites

(Murphy and Riley 1962; Clesceri et al. 1998; Hamon and Mclaughlin 2002).

Ion chromatography (IC) is a widely used method for the analytical separation and determination of condensed phosphates. Ion chromatography has become an important analytical technique in agricultural research which requires minimal sample preparation (Masson et at.200t). Ion chromatography uses differential ion exchangeability of various phosphate species with the stationary phase (Svoboda and

Schmidt lggi),with a gradient elution based on this differential affrnity. The gradient elution improves the resolution and reduces the separation time (Wang and

Li 1999).

Ion chromatography is a technique that can be used to both speciate and separate the

dominant p species in poþhosphate fertilisers (OP, PP, TP and trimetaphosphate).

Capillary zone electrophoresis (CZE) is an alternative method available for

speciation and separation of the P species present in polyphosphate fertilisers (Wang

and Li 1999). In the CSIRO Adelaide laboratory, CZE didnot give reproducible

standard curves after testin g a raîEe of operating conditions and therefore this

technique was not pursued.

Another technique available for speciation of polyphosphate fertiliser solution

extracts is solution NMR. Solution NMR is a powerful technique that is non-invasive

62 and can be used to identiff individual compounds and examine the kinetics of certain reactions (Nanny et at. 1997). While this offers advantages in terms of the range of compounds that can be identified, measurement of solution concentrations below

1mg P/L are generally considered impractical (sample run time greater than24 hours). Also, this technique does not allow for physical separation of P compounds.

The aim of this study was to develop a technique suitable for both speciation and

separation, as the P species needed to be separated for radioassay in isotopic

labelling experiments.

In this chapter, on-line colorimetric techniques were compared with IC as methods

for the speciation and quantification of P species present in poþhosphate fertilisers.

The main aims of this chaPter are:

1. To determine the detection and quantification limits and the minimum working

concentration for colorimetry and IC;

2. To evaluate the strengths and weaknesses of the colorimetric and

chromatographic techniques for speciation of poþhosphates; and

3. To apply these methods to environmental samples and compare the results

obtained.

2.2. Materials and Methods

2.2.1. On-line Golorimetry

on-line colorimetric analysis was performed using a segmented flow analyser

(Burkard Scientific SFA-2). The colour reagents were prepared just prior to analysis

The colour reagent consisted of a solution containingGlL): ammonium molybdate

(40), potassium antimony tartrate (0.14), ascorbic acid (9), and sodium lauryl

sulphate (0.2) in 0.54 M HzSO+. The flow tube length was set so that the colour

63 reagent reacted for 3.5 minutes at room temperature before passing through the spectrophotometer cell where the absorbance was measured at 880nm. The short contact time was chosen to minimise hydrolysis of polyphosphates and organic phosphates to OP. This method has been described in detail by Hamon and

Mclaughlin (2002).

Polyphosphate Digestion for Colorimetric Analys ß

The diluted sample with a volume of 13-15 mL (the exact volume for each sample was determined by weight and recorded for dilution calculations) was evaporated to

2 mL,and the 2 mL of evaporated solution was digested with 2 mL of concentrated sulphuric acid (Hzsoa). The samples were heated for one hour at 160'C and then for four hours at 32}"C,using Tecator digestion blocks, until 2 mL of solution was remaining. Solutions were made up to 25 mL with DI HzO, and pH was adjusted to circumneutral (pH 6.5-7.5) with a further 13.5 mL of 5M NaOH'

2.2.2. lon GhromatograPhY

Deionized distilled water (DI) with a specific resistance of 18.2 milliohms was

prepared in a Millipore (Bedford, MA, usA) Milli-Q@ water purification system.

This DI water was used to prepare all reagents and standards. Any DI water used as

eluent was sparged with helium before use.

Instrument Operating Conditions

(Table A Dionex @ ICS 2500 (Sunnyvale, CA, USA) chromatogtaphy system 2-1')

was equipped with a GP50 gradient pump, an ED50 electrochemical detector and an

AS50 automated sampler. An injection volume of 25 ¡tLwas used with a flow rate of

0.38 ml/min. The eluent generator (EG50) produced a gradient of 20- 80mM

potassium hydroxide (Table 2-2) andwas used in conjunction with internal

was suppression (ASRS 2mmat76mA). The data acquisition and instrument control

64 performed using Chromeleon chromatography management software on a Dell@ personal computer.

The separation was performed using an anion-exchange column (4S16) with an

AG16 guard column. A continuously regenerating anion trap column (CR-TC) was placed between the injection valve and the eluent pump'

Table 2-1: Operating parameters for the IC system

Parameters Conditions Sample loop volume 25 ¡tL Separator column Dionex AS14 (2 x 250 mm) Guard column Dionex AG16 (2 x 50 mm) Eluent 20-80 mM KOH Eluent flow-rate 0.38 ml/min Pump pressure 2000 kPa conductivi 0.639

Table 2-22 Gradient profile for the IC A516 column

Time Potassium Curve Comment (min) Hydroxide

-0.1 20 0 0 20 5 Start 0.2 20 5 Start gradient 8 80 5 10 80 5 10 20 5 Finish gradient t2 20 5 Flush 13 20 5 Finish

2.2.3. Phosphate Reagents

The following reagents used for standards were analytical grade: OP as sodium

dihydrogen orthophosphate (NaH2PO4) @DH), PP as sodium pyrophosphate

decahydrate (Na+PzOz.10H2O) (Sigma) and TP as pentasodium triphosphate

(Na5p3O1¡) (Fluka Chemika). rü/orking solutions for standards were prepared daily

65 2.2.4. Performance of the Analytical Techniques

The techniques were calibrated, and detection limits and the working range were

determined by using the high purity phosphate reagents dissolved in milli-Q water

(0-1 mg P/L for colorimetry and 0-150 mgPlL for chromatography). For the

colorimetric technique only OP was used, but for the chromatographic technique

mixed standards of OP, PP and TP were made.

For all treatments ten replicates were measured.

The detection limit of the analytical procedure was considered to be the lowest

concentration of the analyte that could be distinguished from a blank using the

ruPAC calculation of the limit of detection: Eq'2-l w: T,B* rss,

'Where can be detected, is the mean of 26¡is the smallest measure of response that ?(n

the blank measures, ss is the standard deviation of the blank measures and r is a

numerical constant (Committee 1987). For the detection limit 3 was used as the value

of r. For the quantification limit 5 was used as the value of rc' Currie (1999)

suggested that the minimum working concentration should be calculated with 10 as

the value of r, and this value was also determined. It is important to note that the

detection limit calculated here was a measure of the system performance and did not

use a sample blank but rather a blank of purified Milli-Q water. Therefore the

detection limit of P does not account for any interference effects that may be caused

by other non-P anions.

The standard curves obtained by measuring the standards in the working range of the

two methods were used to compare the linearity of the standard curves between

methods.

66 ent al S amp I es C omp ar i s on of O r th opho sp h at e D et ermin at i on ín E nv ir onm

A comparison was made of the colorimetric and chromatographic measurement of

soluble native P in soil extracts (extract in a soil: water ratio of 1:10) using

colorimetry and IC.

The soil used was a Sodosol from Birchip, Victoria, Australia, which had been

incubated for 1 day andT days with water atï}Yo of maximum water holding

capacity with no P added. Maximum water holding capacity was determined

the according to the method of Jenkinson and Powlson (1976). At the conclusion of

incubation the samples were made to volume (5 mL), and were shaken for 24 hours

on an orbital shaker (17 rpm) and then centrifuged at 2500 tpm (2096 g) for 15

minutes and filtered through 0.2 ¡tmfilters (Schleicher & Schuell) prior to analysis.

Measurements were made of soluble OP in soil solution. The analyses were done in

triplicate for each of the two incubation times.

Measurement of P otyphosphate Concentration using Colorimetric Techniques

The polyphosphate concentration is determined by measurement of initial OP

followed by a digestion procedure to measure total P. Therefore, the effect of

polyphosphate in solution on the initial OP concentration and the effect of the

digestion procedure on accufacy and recovery ofP were evaluated.

Standard curves were analysed containing each of the P species: oP, PP and TP'at

0.1,0.2,0.4, 0.6 and 0.8 mg P/L.

the To test the effects of the interactions of the P species when in solution, each of

possible combinations were made up at a ratio of 1:1 of each species as follows: OP:

PP, OP: TP, PP: TP and OP: PP: TP with every species at a concentration of 0'1 mg

p/L. There were 3 replicates of each treatment. Each of these solutions was measured

using colorimetry to determine the OP concentration prior to digestion. Following

67 digestion (previously described) the samples were analysed colorimetrically to determine total P. Polyphosphate concentration was calculated as the difference between total P and the initial OP measurement.

Comparison of Polphosphate Measurement by Colorimetric and Chromatographic Techniques

A range of PP metal complexes were prepared in order to test if the associated

cations would interfere with the determination of pyrophosphate by colorimetric and

chromato graphic techniques.

Solutions were made of 0.1 M sodium PP with 0.2 M calcium chloride, aluminium

chloride, ferrous chloride, and magnesium chloride respectively. These were stirred

and filtered ('Whatman No. 42 filter paper). The precipitate on top of the filter was

dried at 50'C and resuspended in solution. This solution was stirred for I hour and

the supernatant was filtered using 0.2 pm filters (Schleicher & Schuell). A sub-

sample of 1 mL of this solution was used to measure concentrations of P species by

IC. For colorimetric measurement, a I mL aliquot of the sample was measured for

the Op concentration. A separate 2.5 mL aliquot was digested at 100'C for t hour 'When with 6 drops of concentrated sulphuric acid (Khasawneh et al. 1974). cooled it

was diluted to 10 mL and analysed colorimetrically for concentration of OP.

Solutions were analysed by Inductively Coupled Plasma-Atomic Emission

Spectroscopy (ICP-AES, Spectroflame Modula, Spectro) to measure the total P

concentrations in order to check the recovery obtained by both techniques. Total

concentrations of 41, Ca and Fe were also measured using this technique (Zarcinas et

al. 1996).

68 2.3. Resulfs and Discussion

2.3.1. Performance of the Analytical Methods

Detection and Quantification Limits

The detection limit, quantification limit and minimum working concentration for orthophosphate determined by IC and colorimetry were in close agreement. The detection limits of PP and TP were slightþ higher than those of OP (Table2-3)'

Masson et al. (200t) reported detection limits of soluble P as OP in the range of

0.02-0.03 mg P/L for a Dionex IC system with a 25 ¡tL sample loop, which is in close agreement with the values determined by the equipment used here.

Table 2-3: Detection limit, quantification limit and minimum working concentration for P (mg/L) using IC and colorimetry

P Species Detection Quantification Minimum Working Limit Limit Concentration me P/L mePlL mePlL Ion Orthophosphate 0.02 0.03 0.03 Pyrophosphate 0.03 0.04 0.05 0.05 0.05 0.06

0.02 0.03 0.05

The linearity limits of the calibrations were determined with concentrations

interpolated from peak area for chromatography and peak height for colorimetry. The

standard curves by chromatography for the 3 P species are shown in Figure 2-1 (B),

with the R2 being greater than 0.99 in all cases. The calibration curve for OP

measured by online colorimetry is shown in Figure 2-l (A) with an R2 of greater than

0.99. There was no difference between the methods and P species in the linearity of

the standard curves. Determination of OP in standard solutions by IC and colorimetry

agreed wel[. The working range of P concentrations over which samples can be

69 was far analysed by IC, while conforming to the linearity of the standard curve,

greater (0-120 mg P/L) than colorimetry (0-1 mg P/L)'

Online co orimetric A

J 1.00 (L o) 0.80 E 0.60 co .40 Lõ 0 o 0.20 co oo 0.00 0 100 200 300 Peak height (mm)

lon chromatograPhY B

J 140 È 120 o) E 100 c 80 o O OrthophosPhate (ú 60 L c 40 r flirophosphate o) o 20 o TripolyphosPhate Êo o 0 0 20 40 60 80 area (s)

tr'igure 2-1: Calibration curves for analysis of P spe technique. (A) OP (y=0.0034x, Rf:l) an (y:2.04x,R):1) aod tP (¡r2'15x, nf :1) error.

Comparison of Colorimetric and Chromatographic Orthophosphate Determination in Environmental SamPles

A comparison of the concentrations of native soluble P in soil water extracts

techniques is incubated for 1 rlay andz days using chromatographic and colorimetric

shown in Table 2-4. Only OP was detected in the solutions. The concentrations

and IC' In detected were similar for detection of P concentration using colorimetry

both cases 3 replicates of each measurement were analysed

10 Table 2-4: Comparison of soluble OP in 1:10 soil water extracts from a Sodosol incubated for 1 day and 7 days as determined by IC and colorimetry.

Sodosol 1 day SE 7 day SE incubation incubation IC (mg P/L) 0.r4 0.006 0.054 0.006 PIL o.I7 0.008 0.071 0.005 *SE:standard error

Measurement of Polryhosphate Concentration in Solution Using Colorimetry

Due to the requirement for a measurement of initial OP followed by a digestion

procedure to measure total P, tests were made to investigate the strengths and

weaknesses of the colorimetric technique for estimation of poþhosphate content.

Measurements were made of recovered P following digestion in the concentration

range of the standard curve for colorimetry (0.1- 0.8 mg P/L) (Table2-5). Recovery

of p and measurement of OP from acid hydrolysis of the polyphosphate solutions

was quite low for all P species at the lowest concentration of 0.1 mg P/L, the

remaining samples at 0.2-0.8 mgPlLshowed acceptable recovery levels. The higher

recovery for Op than for PP and TP samples is possibly due to incomplete hydrolysis

in the digestion step. Incomplete digestion was a problem highlighted in the work of

Dick and Tabatabai (1986).

Table 2-5: Recovery (y") of standards of OP, PP and TP post-digestion as measured using colorimetrY.

Phosphorus OP SE PP SE TP SE concentration Recovery recovery recovery PIL 0.1 86 6 84 7 88 6 0.2 101 2 89 4 93 6 0.4 r02 5 90 1 94 9 0.6 105 0 95 4 90 1 0.8 99 1 89 I 89 1 *SE: standard error

The effect of the presence of condensed P species on the detection of OP using

colorimetry is shown in Table 2-6. For most combinations of oP, PP and TP there

7l of was no significant effect of the presence of condensed P species on the detection

Op. However, there was a l3o/o increase in the OP concentration detected where OP was present together with PP and TP. of Table 2-6: Interactions of P species measured by colorimetry. concentration ratio each P species t:t in every combination, with 0.1 mg PIL of each species added in everY casc.

SE P species Initial OP SE Total P SE Pollphosphate P P P 0.006 OP and PP 0.111 0.003 0.206 0.010 0.09s 0.011 OP and TP 0.111 0.006 0.211 0.010 0.100 0.006 PP and TP 0.002 0.001 0.206 0.010 0.203 0.110 OP PP and TP 0.113 0.000 0.4s5 0.100 0342 *SE: standard error

Comparison of Potyphosphate Measurement by Colorimetric and Chromatographic Techniques

Both colorimetry and IC are known to have limited capability for measurement of

non- complexed PP in solution, and most of the measured P is as OP or most likely

complexed PP. While some P complexes occur in solution, particularþ complexes of of Ca phosphate at high pH, this is not a usual situation and the relatively stability

these these complexes in solution is low. Úr most cases chemical dissociation of

soluble complexes rapidty converts the complexes to oP (PierzSmski et a\.2005)

Fe- The results of the analysis by colorimetry (Table 2-7) indicated that P from ca-,

AI-PP. and Mg-PP species had been hydrolysed to oP, but this was not the case for for The requirement for dilution of the sample to fit within the 0-1 mg P/L range

analysis by colorimetry may have had some effect on the stability of the complex

OP' resulting in hydrolysis and therefore measurement of P from a PP complex as

in the Furthermore, the relative strength of the bonding between PP and the cation

complex, and the pH of the analytical matrix would have influenced the stability of

is the p species and therefore the resulting analysis. While detection by colorimetry

72 matrix in an acidic matrix, measurement by chromatography is in a highly alkaline due to the 20-80 mM potassium hydroxide gradient elution. in Ion chromatography appeared to detect PP in A1- and Mg-PP complexes dissolved solution, but not in Ca- and Fe-PP complexes (Table 2-7)'

and colorimetry in Table 2-72 oP and PP concentration (mg P/L) determined by IC saturated solutions of cationic PP complexes'

IC Compound OP PIL PP PIL Total P P CalciumPP (Sigma) 15.10 0.05 15.15 Calcium OP (Sigma) 14.75 0.00 14.75 CalciumPP (lab) 1.97 2.21 4.18 AluminiumPP (ab) 0.01 6.28 6.29 hon PP (lab) 0.36 0.04 0.40 0.29 9.81 10.10 Colorimetry Compound OP P PP Total P P CalciumPP (Sigma) 10.58 8.30 18.88 Calcium OP (Sigma) 10.16 t.79 11.95 CalciumPP (ab) 1.88 5.83 7.70 AluminiumPP (lab) 0.04 10.91 10.95 Iron PP (lab) 0.69 0.65 r.34 PP 1.56 8.49 10.05 *(Sigma) products purchased from Sigma Aldrich and (lab): comPlexes manufactured in the laboratory

Comparison of the total P calculated by the sum of P species detected using and colorimetry and chrom atography with those measured by ICP-AES (Table 2-7

Table 2-8) show that the calculated value for total P as the sum of OP and PP

ICP-AES' measured by IC is in most cases greater than the actual value measured by

This is most likely due to the dilution steps required for analysis causing some

was analyticalerror. The total cation concentration data (Table 2-8) show that there

considerably more Ca in solution when a Ca-PP or Ca-OP complex was measured

than other metals with their corresponding PP complex'

73 Table 2-8: P, Al, Ca, Fe and Mg concentration (mg/L) measured by ICP-AES.

Compound Phosphorus Aluminium Calcium Iron Magnesium

CalciumPP (Sigma) t4.09 0.31 12.27 0.03 0.t7 Calcium OP (Sigma) 13.84 0.31 7.51 0.02 7.03 CalciumPP (ab) 5.23 0.39 r0.64 0.04 0.10 AluminiumPP (ab) 8.8s 8.63 0.31 0.08 0.31 hon PP 0.8 0.62 0.38 0.98 0.44 tMagnesium PP data not available' 2.4, Conclusions

The proposed IC method is a sensitive and precise analytical technique for the

determination and speciation of OP, PP and TP in aqueous solutions'

The performance of colorimetry was comparable to IC in terms of detection limits

and linearity of solutions in the range of the standard curve, and the measurement of

Op at 0.1 mg p/L did not seem to be affected adversely by the presence of condensed

P species.

Ion chromatography does not require the extra digestion step to quanti$r the

polyphosphate concentration, as is the case for colorimetry. Results presented here

indicated that there were some losses of P and therefore inaccuracy in the digestion-

colorimetric technique as evidenced by incomplete recovery. Furthermore IC is able

to quantiff multiple condensed P species and provides the option to physically

separate the P species, which rwas a requirement of experiments presented in

Chapters 7-9.

The ability of IC to speciate and separate several condensed P species in one

analysis, which has shown to be comparable to colorimetry in terms of accuracy of

quantification, provides further flexibility for investigations of the behaviour of

fertilisers based on poþhosphates.

74 Ghapter 3. Polyphosphate Fertiliser Solution Stability

3.1. lntroduction

Due to the increasing popularity of APP fertiliser in Australian agriculture it is

potential important to analyse the stability of the fertiliser solution under araîge of

storage conditions.

the point The p in a polyphosphate fertiliser exists as more than one ionic species. At

is present of sale, approximately 30-40%of the fertiliser P is present as OP, 50-55%

P. proportion as pp and the remainder exists as TP and more condensed forms of The

of each P species does not remain constant due to the occurrence of a hydrolysis

reaction where more condensed P species react with water to form less condensed

forms of P. An example of this reaction is the hydrolysis of pyrophosphate:

Eq' 3-1 P2o+ +n2o -> 2+vol-

This is a chemically and biologically mediated reaction. Factors that control

hydrolysis of polyphosphates include: the poþhosphate concentration, the ionic

2000; environment, temperature and pH (Van Wazet et al. 1952; Rashchi and Finch

promotes Ahmad and Kelso 2001). In particular, an acidifying chemical environment

the hydrolysis of APP, which was demonstrated by Van Wazer et al' (1955) who

showed that hydrolysis continually decreased from pH 1 to 13.

one of the main problems experienced when using APP fcrtilisers has been the

inability to introduce adequate concentrations of micronutrients into the blend'

been Researchers at the South Australian Research and Development Institute have

acidifying App with phosphoric acid in order to introduce a higher concentration of

et al' trace elements such as ZnSO+ and MnSOa into the fertiliser solution (Holloway

75 2002). The solubilities of Zn and Mn in APP ate 1.6%o and}.2Yo w/v (Mortvedt

1991; Van Loon 2001), respectively, while the target concentrations for field applications are up to 6Yo w/v. The acidification of APP in order to supply a greater concentration of trace elements has shown benefits in terms of overall crop nutrition

(Holloway et at.2002), and therefore it is important to investigate the effects this has on the composition and stability of the fertiliser formulation.

Many of the studies investigating hydrolysis reactions of polyphosphates use analytical grade reagents with a very low level of contamination by other P species, (Zinder no initial orthophosphate present , arrd amixture of 1-3 condensed P species et at. 1984).Frazier and Dillard (1931) and Grzmil and Kic (2002) examined the hydrolysis of commercial poþhosphate solutions but their investigations were limited to the effects of cation concentration on fertiliser stability. This chapter describes a study of the effects of the range of storage conditions that a commercially available APP fertiliser would be exposed to in Australia'

V/hile many of the investigations of pollphosphate degradation are based on measuring the rate of formation of OP and extrapolating this to a corresponding rate

for the degradation of the polyphosphate species (Zindet et al. 1984), in the presence

of more than one polyphosphate species this does not supply information about the

hydrolysis of individual condensed P species. Here IC was used to measure changes

in P speciation in poþhosphate fertiliser solutions when treatments of pH,

temperature and time have been imposed. This technique was previously used by

Halliwell et al. (2001) to investigate the hydrolysis of trace levels of polyphosphates

in wastewater.

76 In this chapter the effects of several storage factors þH, temperature and time) on the

stability and composition of commercially available poþhosphate fertilisers are

evaluated.

3.2. Materials and Methods

3.2.1. Reagents

The following analytical grade reagents were used for standards: OP as sodium

dihydrogen orthophosphate (NaH2PO4) @DH), PP as sodium pyrophosphate

decahydrate (Na¿PzOz.10H2O) (sigma) and TP as pentasodium triphosphate

(Na5p3O1e) (Fluka Chemika). Working solutions for standards were prepared daily.

Ammonium polyphosphate as supplied by Agrichem Pty Ltd. was stored af 4"C and

used for the incubation studY.

3.2.2. Treatments

The experimental design was factorial with three temperature treatments (4,25 and

50"C) and four acidity treatments (0, 10, 25 and50% of P supplied as analytical

gradephosphoric acid). Each combination of treatments was incubated fot 1,3,'7,

14, or 28 days with two replicates per combination.

3.2.3. Speciation and Quantification of P

Ion chromatography was used for speciation of the P species supplied in the

poþhosphate fertiliser including OP, PP and TP, as describedin2.2.2.

Concentrations of total P in the fertiliser solutions were determined by ICP-AES

under the operating conditions outlined by zatcinas et al. (1996).

77 3.2.4. Statistical AnalYsis

Anaþsis of variance (ANOVA) was performed using the Genstat 6.0 statistical packagewith the analysis of least significant difference (LSD) atthe 5olo level to discriminate between treatments.

3.2.5. Calculations

The equation used for determination of the hydrolysis rate constant and estimated half life was previously described by Halliwell et al' (2001)'

The hydrolysis rate constant was determined according to the following equation:

Eq.3-2 log[P3 Or o5- lo - log[P3o1s5- ]1 k 2 .3 t

Where: r) k is the hydrolysis rate constant (second t is the time (seconds)

TP at time zero [P¡Ot O5- ] e is the concentration of

[P:OtO5-]1is the concentration of TP at time t.

The half life was calculated as

Eq. 3-3 5 lP:oro 5 l - lP¡oro l tt,r=2'3 t0 tu2 k

Where

tr¿ is the half life (seconds)

[P¡O1O5-]to is the concentration of TP attimezerc

TP at time zero [P¡O1O5-]t172is one half the concentration of

78 The energy of activation was determined by the equation: Eq.3-4 Ea lnK =lîA-- RT

'Where:

K: hydrolysis rate constant (k/second)

A: pre-exponential factor. A is a term which includes factors including the

frequency of collisions and their orientation. It varies slightly with tempeÍature,

although not much, and is often taken as constant across small temperature ranges'

Ea: energ] of activation (kJ/mol) pV: R: gas constant 8.31 J/k/mol. This is a constant which comes from an equation,

nRT, which relates the pressure, volume and temperature of a particular number of

moles of gas.

T: temperature (K")

3.3. Resutfs and Discussion

3.3.1. Change in P Speciation Over One Month

The interaction of acidity, temperature and time had a significant effect on the

proportion (%) of total P as PP in solution (P < 0.001, L.S.D.:3.96) (Figure 3-1)'

on Similarly, the interaction of acidity, temperature and time had a significant effect

the proportion of total P as TP in solution (P < 0.001, L.S.D. 2.aÐ Figu¡re 3-1).

The only temperature treatment, in the absence of a pH treatment, to have a

significant effect on the amount of hydrolysis of condensed P species over the 28 day

incubation period was the 50'C treatment (Figure 3-1). The lowest pH treatment

caused significant hydrolysis of both PP and TP (Figure 3-1) over the 28 day

incubation period. At the lowest pH and highest temperature almost all TP and96o/o

of PP was hydrolYsed after 28 daYs.

79 rT: %OP Y,PP 120 A %TP - 6.4 5.8 H 4.9 2.3 100

80 U' 'õo o ,t 60 È s 40

50"c 0 4"C 25"C 50' "c 25'C 4"C 1 28 days 1 day 28 days 1 day 28 daYs 1 day 28 days day

and after 28 tr'igure 3-1: Percentage of P species as OP' PP and TP after one day ut 4oC days at 4;C,25"Cand 50oC. Measurements wer€ made at solution pH values of 6.41 5.8, 4.9 and 2.3. 3.3.2. Hydrolysis Rate Gonstant

Using Figure 3-2 and,Eq. 3-1, the rate constant for hydrolysis of TP at 50'C was

rate calculated as g.2I x 10-7ls, with a half-life of 20.1 days. As the hydrolysis

at determined was slightly positive at 4"C aîd25"C,no rate constant was determined

these temperatures in the absence of introduced acidity. This suggests that

poþhosphate solutions are very stable where no acidity is introduced at

temperatures less than or equal to 25"C. This is in agreement with the findings of

Van'Wazer et al. (1955) andZindet et al. (1934) who reported the hydrolytic

be very degradation of pol¡phosphates in aqueous solution at room temperature to

80 constants for slow. Halliwell et al (2001) found considerably higher hydrolysis rate

respectively)- trace levels of Tp at 15'C and20"C (2.59 x 10-s/s and 6.47 x 10-5/s, oP zinder et at. (1984) and Dick and Tabatabai (1977) suggested that the initial

the concentration has an effect on the rate of the hydrolysis reaction, causing

considerable range in the hydrolysis rate constants reported in the literature'

45 J o) E 40 L I .9 \; E c. 3.5 -o q) co ì o o I 3.0 a FfL oo) 2.5 0 200 400 600 800 time (hours)

and tr'igure 3-2: Change in log TP concentration (mgil) over time (hours) tt 4"C,25'C 50'C (-i-SO"C 't'25"C -A-4'C) at solution pH6'1' at The equation of the line at ¿'i is Y= 2x10-e *3:Ywith a R2 of 0'031' 2s.c y:sxöi;3ö;iãîn1or0.49, an¿ at 50'c y-4x10-7+4.04 with an RÍ of 1.

for solutions The effect of pH on the hydrolysis rate constant for TP was determined

174 days stored at25"Cusing Eq.3-1. The half-life of TP was 34 days at pH2.3 and

hydrolysis of PP, at pH 5.4 (Figure 3-3). Torres-Dorante et al. (2005) evaluated the

found ttre trati-ffe TP, and trimetaphosphate in a mixed solution at 0.5 mg P/L. They pH of TP to be 15 days at pH 6.4. The increasing rate of hydrolysis with decreasing

was previously observed by Friess (1952) and Subbarao (1975).

81 ^J 5.0 o) E Z +.s .Eo E Ë oo o --aetal oC uu 'a it- q) 9 s.o 0 200 400 600 800 time (hours)

Fígure 3-3: Change in log TP concentration (mg/L) over time (hours) at pH 2.3,4.9'S'8 ana o.Z 1-l-pn 23 -'-pH 4.9'L- 25"c' with a Rf of 0.93, at PH -8x*3.95 and at pH 6.4

The Op concentration in the fertiliser solutions can increase as a result of hydrolysis of both PP and TP. An OP formation constant is commonly determined in experiments where a single condensed P species is hydrolysed. However, in this case the source of the OP could not be determined. Therefore the OP formation constant was not determined.

Similarly the changes in PP concentration in solution could be due to the hydrolysis of Tp increasing PP concentration, or PP hydrolysis reducing the PP concentration'

Due to these simultaneous reactions the hydrolysis rate constant for PP was not

determined.

3.3.3. Activation Energy

The energy of activation (Ea) is the energy barrier that must be overcome before a

reaction can proceed. The activation energy for hydrolysis of TP (I2'7 kJlmol) was

only determined for the lowest pH treatments (Figure 3-4). Busman and Tabatabai

(1935) determined non-enzymatic activation energies for hydrolysis of

the trimetaphosphate in soil as 29-39 kJ/mol. Dick and Tabatabai (1936) evaluated

82 enzymalichydrolysis of PP in soil and determined activation energies in the range of

range. 10.7 to 7l.2ulmo1. The value determined for TP at pH 2.29 was within this

3.0 3.1 3.2 3.3 3.4 3.5 3.6 3'7 ?0 ;-1 o-tc 2-3 o_4o o €Ã(õ-r L .9. -6 c) a E-BÞ E c -'t0 1000/T

the Figure 3-4: Energy of activation for hydrolysis of TP at pH 2.3. Ea is the slope of relationship between ln K (K is the hydrolysis rate constant) and 1000/T (T is temperature in oK). Equation of thl finõy=-12.70x+36.32 with Rf 0.98.

3.3.4. Change ln P speciation Over One Week

The change in P speciation over one week is described here, as it is more representative of timing of fertiliser formulation and application in the field than the changes over 1 month previously described.

Over 7 days, the two lowest pH treatments, at temperatures greater than25"C,

significantly increased hydrolysis of PP (Figure 3-5). Hydrolysis was significant at

50"C for all pH treatments and aÍthe lowest pH also when the solution was stored at

25"C (Figure 3-5).

83 rT\ %OP -%PP%ÎP 120 -

pH 6.4 pH 5.8 pH 4.9 pH2.3 100

BO Ø '6o o È60 fL s 40

0 25"C 50"c 4 c 25"C 4"C 25"C 4"C 25"C 4"C I day 7 days 1 day 7 daYs day 7 daYs I day 7 daYs

4oc, Figure 3-5: Percentage of P species as oP, PP and TP after 1 day and'7 days ,-at 25oC and 50oC. Measurements were made at solution pH values of 6.4, 5.8' 4.9 and2.3.

P in In field experiments, APP has been acidified by supplying3}-45o/o of the total the formulation as phosphoric acid (Holloway et a\.2002).In this experiment, a

a moderate temperature treatment of 25"C,with 50% of P supplied as HIPO4 causing

solution pH of 2.2,resulted in approximately one third of the total P in solution

PP over 7 existing as PP, with a 7o/o dectease in the proportion of total P present as

(6-7% of total days. Only a very small proportion of fertiliser P existed as TP initially

that p) and 55% of this was hydrolysed over 7 days. This experiment demonstrated

acidifying the polyphosphate fertiliser formulation significantly increased the

hydrolysis of condensed P species to oP, and this needs to be considered when

thc formulating mixed ApP blends with acid and trace elements for application in

field.

84 3.4. Conclusions

The data in this chapter details the hydrolysis of condensed P species in APP fertilisers as a function of time, temperature and solution acidity. High temperatures and low acidity promote hydrolysis of condensed P species to oP.

Acidiffing pollphosphate fertilisers therefore has a significant negative impact on their stability. These findings suggest that it is important to avoid storing poþhosphate fertilisers in places likely to be exposed to high temperatures, in order to minimise poþhosphate hydrolysis. Adding phosphoric acid to APP may increase the solubility of trace elements in the short term, but the impact of this blending on hydrolysis and resultant effects of increasing concentrations of OP over time shifting metal solubility need to be considered.

Most studies of this nature suggest that the rate of hydrolysis in the soil-plant system is considerably higher than the rate of hydrolysis in an aqueous system (Hossner and phillips l97l; Subbarao et al.1977;Zinder et al. 1984; Torres-Dorante et aL.2005)-

31P Chapter 4 will examine total hydrolysis of PP in soil using solid-state nuclear magnetic resonance. Studies of the hydrolysis of PP in the labile pool are discussed

in Chapters 7-9.

85 Ghapter 4. Hydrolysis of Pyrophosphate in a Highly Galcareous Soil: A Solid-State "'P NMR Study.

4.1. Statement of Contributions

Chapter 4 was published (McBeath et a|.2006):

McBeath TM, Smernik RI, Lombi E, Mclaughlin MJ (2006) Hydrolysis of Pyrophosphate in a Highly Calcareous Soil: A Solid-State 3lP NMR Study' So;/ Science Society of America Journal70,856-862 and is attached as APPendix B.

The role of R.J. Smemik in this publication was for the analysis of samples by nuclear magnetic resonance (NNR) and assistance with technical writing relating to

NMR methodology. The role of E. Lombi and M.J. Mclaughlin in this publication was the revision of drafts of the manuscript prior to publication. As primary author I was responsible for experimental design and set up, analysis of samples by ion chromatography, assistance with analysis by NMR, data collection, collation and analysis and the primary preparation of the manuscript.

This manuscript is presented in Chapter 4 with only alterations to formatting to suit

that of the thesis. The abstract and general introduction has been removed for

consistency between chaPters. lrûArñ & g) t*År*i,, 4 '

T.M.McBeath R.J. Smernik E.Lombi M.J. Mclaughlin

86 4.2. Introduction

Hydrolysis reactions of poþhosphate fertiliser in soil convert more condensed P

forms species (two or more oP groups linked by oxygen bridges) to less condensed of p (Linds ay 1979; Dick and Tabatabai 1936). Previous data indicate that PP constitutes 70-g0%of the poþhosphate in APP (Khasawneh et al. 1974). Therefore the most common hydrolysis reaction of polyphosphate fertiliser is the conversion of

PP to OP (Sutton and Larsen 1964; Hashimoto et al' 1969)'

The hydrolysis reaction of PP is shown below in Eq' 4-1:

Eq' 4-1 P2o+ +H2o -+ znvol-

non- Several studies have been conducted to compare hydrolysis of PP on a range of

et al' calcareous soil types (Sutton et al. !966; Gilliam and Sample 1968; fkasawneh

1979; Sample et al. 1979;Parent et al. 1985b; Ahmad and Kelso 2001). Most methods previously used to assess the behaviour of PP in soils involved extraction of soil P by aqueous media and subsequent determination of oP by colorimetry.

Following OP determination, PP concentration is measured as the difference between

Op and total P determined by acid digestion and colorimetric measurement' These

the extraction techniques often either ignore extraction efficiency, or assumed that

extraction technique is 100% efficient for solid-phase hydrolysed and non-

hydrolysed PP in the soil. Furthermore, the approach of measuring polyphosphate-P

concentration by difference ignores thc organic-P concentration, effectively

calculating the polyphosphate-P concentration as poþhosphate-P plus organic-P' 3lP This study describes the use of solid-state nuclear magnetic resonance (NMR)

spectroscopy for investigatingPP hydrolysis in a highly calcareous soil. Solid-state

"p NMR was compared to speciation of P by IC, which enables the direct

87 measurement of PP and oP, on a sodium hydroxide (NaoH) extract. while the

NaOH extraction is conventionally used for measurement of organic P it has been

(Shand et al. used with subsequent speciation to measure PP present in soil extracts

2000; Tumer et at.2003a). Solution-state NMR spectroscopy has been used to

investigate the hydrolysis of PP in solution (Subbarao et al. 1977), while solid state

,tp NMR has been used to investigate a poþhosphate-chitosan complex as a source

of P (Fross ard et al. 1994b). However this is the first time solid-state NMR

source in spectroscopy has been used to investigate the hydrolysis of PP as a nutrient

soil.

is Understanding the nature of hydrolysis reactions of PP in Australian soil types

necessary in order to elucidate the mechanisms underlying the superior agronomic

performance of pollphosphate fertiliser as compared to traditional P fertilisers on

calcareous soil tyPes.

4.3. Materials and Methods

4.3.L Soil Gollection and Chemical Properties

The soil used in this study was a grey calcareous sandy loam (Calcixerollic

xerochrept (Soil Survey 1992)),collected from warramboo, upper Eyre Peninsula,

South Australia. The soil was air-dried and sieved to < 1 mm before analysis- Sáil

and water holding capacity was determined according to the method of Jenkinson

powlson (1g76).The pH of a 1:5 soiVsolution extract was 9.1 in water and 8.0 in

0.01 M calcium chloride (CaClz). The electrical conductivity of a 1:5 soiV water

to mixture was 0.12 dS/m. Electrical conductivity and pH were determined according

Rayment and Higginson (1992).The caco3 content was 770 g/kg determined

according to Page (1982).

88 The total P content of the soil was 339 mglkg, as determined by inductively coupled plasma atomic emission spectroscopy (ICP-AES) following digestion in aqua regia

(zarcinas et at. 1996). The resin-exchangeable P content was 7 mg P/kg soil, as determined using anion-exchange resin strips (Mclaughlin et al. 1994)'

4.3.2. Soil lncubations

Soil (50 g) was wetted up to 30% of water holding capactty with a pipette followed

1 week by vigorous shaking to ensure thorough mixing, and allowed to incubate for

P to allow microbially-driven dissolved organic matter flushes to settle before adding

(Jenkinson andPowlsonlgT6).4 solution of sodium (Na) PP (11.82 glL,8'46mL

per 50 g soil) was then added. This resulted in a P addition of 2000 mg P/kg of soil,

This P and also increased the moisture content to 75Yo of water holding capacity.

concentration was chosen as it is similar to that found in close proximity to fertiliser

granules, or in a fluid fertiliser band, in soil (Lombi et a\.,2004b)' Separate soil

replicates of samples were incubated for I , 3 , 7 , 14 and 2l days. There were two

started each treatment. The incubations were staggered, with the longest incubation

under first and all incubations timed to finish on the same day. Soils were incubated

aerobic conditions having constant temperature (20 + 2"C) andhumidity'

4.3.3. Solid-State 3rP NMR Spectroscopy

3tP (MAS) and Solid-state NMR spectra were acquired with magic angle spinning tH high-power decoupling on a Varian Unity INOVA 400 spectrometer with a Doty

Scientific supersonic MAS probe at a3rP frequency of 1 6 1 .9 MHz' Samples were

packed into 7 mm diameter cylindrical zirconia rotors with Kel-F end-caps and spun

at 5 kÍlz at the magic angle of 54.7"

89 3tP polarization Two techniques were used to acquire solid -state NMR spectra-cross

(CP) and direct polanzation (DP). The CP technique involves transfer of lH magnetization (coherence) from to'lP nuclei. The main advantage of the CP lH technique is that it enables more rapid accumulation of scans because nuclei

31P regain equilibrium magnetization more rapidly than do nuclei' The main disadvantage of the CP technique is that the transfer of polaization can be inefficient lH for 31p nuclei remote from nuclei, and hence such nuclei can be underrepresented

direct or absent from CP spectra. The DP technique, as its name suggests, involves polaizarion of 31P nuclei. phosphorus-31 CP NMR spectra of the soils represent the accumulation of 10000

tH and a 4 s scans, and were acquired using a 5.1ps 90' pulse, 1-ms contact time 3tP recycle delay. The total acquisition time for CP NMR spectra was around 11-1 hours. phosphorus-31DP NMR spectra of the soils represent the accumulation of

time 500 scans and were acquired using a 100-s recycle delay. The total acquisition

3lp 31P for Dp NMR spectra was around 13.9 hours. The DP NMR spectra were

corrected for a broad background signal by subtracting the 'lP DP NMR spectrum

acquired for an empty rotor.

1216 Free induction decays were acquired with a sr'veep width of 50 kHz. A total of

12 ms. data points was collected for all spectra, representing an acquisition time of

All spectra rwere zero-filled ro 13 1072 data points and processed with a 100-Hz

Lorentzian line broadening and a 0.010-s Gaussian broadening' Chemical shifts were

extemally referenced to NH¿HzPO¿,at0.72ppm (Frossatd et aL.2002)

phosphorus-31 spin counting experiments were performed using a modification of

13C the spin counting method of Smernik and Oades (2000a,b). Ammonium (i.e' dihydrogen phosphate (NH+HzPOa) was used as an external intensity standard

90 The the NH+HzPO4 spectrum was acquired separately to those of the samples).

NH+HzpO+'lP CP NMR spectrum was acquired in l6 scans, using a l-ms contact

3rP acquired time and a l0-s recycle delay. The NII¿HzPO+ DP NMR spectrum was in one scan after equilibration for 1000 s (16.7 minutes). For cP spin counting

contact experiments, differences in the rate of lH magnetizatiotdecay during the for itme (T1pH) between the sample and the NII+HzPO+ standard were accounted lock using the method of Smernik and Oades (2000), except that a variable spin

ZrpH (Smernik rather than a variable contact time experiment was used to determine

120 ms' et at. 2002). The ZloH relaxation rate for NH+HzPO¿ was found to be

Uncertainty in the precision of Po6, values is estimated to be +l\yo in CP and+l5Vo

in DP (Smemik and Oades 2000).

phosphorus-31 Dp NMR spectra were acquired in a single scan for reference sodium

(Ca-OP) (Na) and ca oP and PP salts. The salts NaHzPO¿.HzO (Na-OP), CaHPO¿

prepared by and Na+Pzoz.l0Hzo (Na-PP) were used as received. calcium PP was mL)' mixing solutions of Na¿PzOz.10H2O (0.1 M, 100 mL) and CaClz (0.2 M, 100 water' The resultant precipitate was isolated by filtration and washing with Milli-Q

suggested that the and dried overnight at 50oC. Anaþsis by x-ray diffraction (XRD)

ca-PP was mostly amorphous with some minor halite (Nacl).

4.3.4. lon GhromatograPhY

of Turner A NaOH extraction was performed using a modified version of the method

(40 and shaken and McKelvie (2002). Soil (4 g) was treated with 1 M NaOH mL),

G) for 30 for 15 minutes (17 rpm), then centrifuged at 3000 rpm (RCF: 2096 filters minutes. The supernatant was filtered using 0.22 ¡tm(Schleicher & Schuell)

was performed and diluted by a factor of 50 with distilled water. Ion chromatography (4516) to using a Dionex ICS 2500 system with an anion-exchange column

91 determine OP and PP concentrations in NaOH extracts. An injection volume of 25

A gradient o120- 80 mM potassium ¡rL was used with a flow rate of 0.38 ml/min. hydroxide (KOH) was used in conjunction with internal suppression (76 mA)'

Orthophosphate and PP were detected using a conductivity detector (BioLC ED50

Electrochemical Detector, Dionex)'

4.4. Resulús and Discussion

4.4.1, Solid-State 3rP NMR Spectra of Reference Salts

31P Figure 4-1 shows the DP NMR spectra of the Na and Ca salts of OP and PP'

Although p was added to the soils as dissolved Na-PP, the high abundance of Ca in this soil, the high pH of the soil solution and the lower solubility of the Ca than Na phosphates, means that the P added to the soils most likely reacts with Ca to form Ca phosphates (Lindsay 1979; Hedley and Mclaughlin 2005)'

3lP Without magic angle spinning, the solid-state NMR spectra of these salts would be very broad, spanning hundreds of ppm. Magic angle spinning results in a series of resonances just a few ppm wide, with a separation that is proportional to the spinning

rate, in this case 5 kflz or 31 ppm. The central band can be identified as the one

whose chemical shift (the "true" or isotopic chemical shift) is independent of the

spinning rate. For the spectra in Figure 4-1 the central band is also the most intense

peak in each spectrum (though this need not always be the case).

The other peaks are artefacts called spinning siclebands (SSBs)' The envelopc of

SSBs approximates the shape and extent of the non-spinning spectrum' Thus, without

magic angle spinning, the resonance of Ca-OP would be the least broad of the 4 salts,

also whereas the resonance of the two PP salts would be considerably broader and

non-symmetric.

92 Na orthophosphate Na pyrophosPhate

Ca orthophosphate Ca pyrophosPhate

-60 100 60 20 -20 -60 PPm 1OO 60 20 '20 PPm

31P PP Figure 4-1: Solid-state direct polarization (DP) spectra of Na- and Ca- OP and salts.

centre The spectra of the four salts differ in both the position (chemical shift) of the

the band (Table 4-1) and in the size and pattern of the SSBs. The chemical shifts of

PP salts are more negative (upfreld) than the corresponding OP salts, and the chemical shifts of the Ca salts are more negative than the corresponding Na salts.

The size and pattern of SSBs are similar for both PP salts (Figure 4-1)' The centre

and band represents 27 to 36Yo of total signal for the PP salts (Table 4-1). The size

pattern of SSBs are quite different for the Na- and Ca-OP salts (Figure 4-1)' The

signal for centre band represents 650lo of total signal for ca-oP, but only 35%o of total

Na-Op (Table 4-1). The greater line width of the Ca-PP resonances most likely

reflects the amorphous (poorþ crystalline) nature of this salt.

93 Table 4-1: Chemical shift (ppm) and percentage of total signal found in the central band for Na- and Ca- OP and PP salts'

P compound Chemical shift Signal in central band of Na orthophosphate 1.81 35 Na pyrophosPhate -t.93 27 Ca orthophosphate -r.69 65 Ca -6.88 36

to be the These results demonstrated that Ca-OP and Ca-PP, which are expected

of both dominant species present in the soil samples can be distinguished on the basis

chemical shift and SSB Pattern.

3rP 4.4.2. Solid-State NMR Spectra of Soils

3rP Figure 4-2 shows the CP and DP spectra for the unamended soil and soil samples

The 31P DP amended with PP (2000 mg P/kg soil) and incubated for one to 21 days- at l'7 NMR spectrum of the unamended soil (Figure 4-2) contains a sharp resonance

is ppm, along with associated low intensity SSBs. The size and pattern of the SSBs is slightly similar to that for the ca-oP salt (Figure 4-1), although the chemical shift

different (-1.69 ppm for ca-oP, Table 4-1). Chemical shift is sensitive to factors

shift between such as pH and crystal packing, so a 3.4 ppm difference in chemical

OP) is the pure Ca-OP salt and the soil resonance (which is presumably also mostþ

CP not unreasonable. The SSBs are stronger and broader in the corresponding 'lP

on NMR spectrum of the unamended soil (Figure 4-2), and there is a broad shoulder

can be assigned the upfield side of the strongest peak. The broad shoulder and SSBs

spectrum to organic P (Dougherly et a\.,2005). They are more prominent in the CP

spectrum; the because the inorganic P is observed with very low sensitivity in the cP

DP (see observability of P for the unamended soil was 20%by cP and 78%by

below).

94 31p for one The Cp and Dp NMR spectra of the soil amended with PP and incubated

to PP, day (Figure 4-2) contain a strong new set of resonances that can be assigned pattem of based on the upfield shift of the central band (-8.5 ppm) and the distinctive

of the SSBs (compare Ca-PP spectrum in Figure 4-1). The'1P DP NMR spectrum

resonance at soil amended with PP and incubated for one day also contains a small

present in 2.1 ppmthat canbe assigned to oP. The quantity of oP is similar to that

been set to the unamended soil (the vertical scales of the spectra in Figure 4-2have

incubations allow direct comparison within the series of CP and DP spectra)' Longer

in the result in a decrease in the intensity of the PP resonances and an increase intensity of the OP resonances (Figure 4-2).

95 CP spectra DP spectra unamended

1 day

3 days

7 days

1 5 days

21 days

14O 80 20 -40 PPm 140 80 20 -40 PPm

31P X,igure 4-2: Solid-state cross polarization (CP) and direct potarization @P)_spectra of unamended soil and soil amended with PP supplying 2000 mg P/kg soil and incubated for I to 2l days. The vertical scales of the spectra have been set to allow direct comparison within the series of CP and DP spectra.

96 4.4.3. Quantification of orthophosphate and Pyrophosphate 31P Gontents from Solid-State NMR Spectra

to OP' The spectra presented in Figure 4-2 cleatly indicate a conversion of added PP but they also show a decrease in overall signal intensity, especially for the CP 3tP spectra. To obtain quantitativ e datafrom the solid-state NMR spectra, we first

needed to establish the sensitivity of NMR detection. This was achieved by spin

per unit P counting, which is simply a calibration of total signal intensity in a sample

biases against that of a standard. Spin counting has proven very useful for identifying

t3c 2000 in signal distribution in solid-state NNß. spectra (Smernik and oades 2000; ttN b), and has been applied recently to solid-state lsmernik and Baldock 2005;

3rP Smernik and Baldock 2005 b) and lDougherty et at.2005) NMR spectra. The

each results of spin counting, presented in terms of the relative observability of P in

sample versus that of the standard (Pou.), are shown inTable 4-2.

3tP 3rP Table 4-2: Observability of P in CP (P,',-CP) and DP (Pob,-DP) NMR spectra- results of sPin counting.

Incubation time Pou-CP Po6r-DP

Unamended soil 20 78 1 24 88 3 13 85 7 8 77 15 10 7l 2l 6 72

The NMR observability of P in the unamended soil was 78%o ror the DP technique

the DP and2¡%ofor the cP technique (Table 4-2). From these values, it is clear that

technique detects P more quantitatively. The most likely cause of the low CP lH observability is that most P nuclei are not close (within a few bonds) to nuclei.

P is The most likely reason that the DP observability was less than 100% is that some

rendered.1.{MR-invisible" through close association with paramagnetic species in

the soil, especially Fe (Doughefi et a1.,2005).

97 Dougherty et Both the Cp and Dp observabilities are higher than those reported by

Po6, (cP) to raîge at. (2005),who found Pou (DP) to range rtom l7o/o to 28Yo and (2005) fromlyo to l2yo. The low NMR observability of P in the Dougherty et al

content was 0'86-1'64% soils was attributed to the effect of paramagnetic Fe (soil Fe

this study has a much lower Fe content (0'33% [\^//w]). The calcareous soil used in

[wiw]). amended with PP (Table The NMR observability of P was higher in the soil freshly

PP was more 4-2) thanin the unamended soil, indicating that the freshly added

decreased readily observed than the native P. Phosphorus-31 NMR observability (Table 4-2)' with increasing time after PP addition, especially for the cP technique

the potential to The low and variablo Pob* values for the CP technique compromised

31P the Po6. values quantiff oP and PP from the cP NMR spectra. on the other hand,

quantification of oP and for the DP technique are high enough (71-SS%) to attempt

the "NMR invisible P" PP from the DP spectra, although the potential for bias due to

must remain a consideration.

of OP and PP to the'lP NMR spectra was Quantification of the relative contributions of the unamended soil achieved by spectral subtraction of the "P DP NMR spectrum 3tP soil' (which contains only OP) from the DP NMR spectra of the unamended

were subtracted Increasing proportions of NMR spectrum of the unamended soil

of PP was until the central0P resonance was cancelled out. The contribution

determined by integration of the resultant "difference" spectfa'

PP, 1660 mg P/kg soil Figure 4-3 shows that 1 day after adding 2000 mg P/kg soil as

870 mg P/kg soil as PP as PP was detected by NMR, and after 2l days of incubation,

the concentration of was still detected. On the other hand , aftet 1 day of incubation

native oP to 390 mg oP detected by NMR increased from the 260 mgPlkg soil as

98 p/kg soil and continued to increase to 810 mg P/kg soil after 2l days of incubatron

the The concentration of undetected P also more than doubled during the course of incubation (Figure 4-3).

NMR NaOH Extract 1 800 I 600 'o, 1400 -!¿ o) 1200 Orthophosphate E I 1 000 tr PyrophosPhate c 800 o E Undetected C 600 o fL 400 200 0 1 3 7 1521 1 3 7 1521 lncubation time (daYs)

and trigure 4-3: OP and PP concentration determined by solid-state NMR spectroscopy P IC on a NaOH P extract. The "undetected" P was calculated as total minus the sum of OP and PP.

4.4.4. Quantification of Orthophosphate and Pyrophosphate Goncentrations using Alkaline Extraction and lon GhromatographY

The concentration of OP and PP were determined by IC after a I M NaOH in this extraction. Figure 4-3 shows that the concentrations of OP and PP determined

proportion of way were lower than those determined by NMR, and consequently the

than for undetected P (P which was not extracted by lM NaOH) was greater for IC

NMR. However, both methods showed similar trends in OP and PP concentrations

through the incubation (Figure 4-3), as well as increases in the proportion of

undetected P.

99 4.4.5. Hydrolysis Rate and Pyrophosphate Half-Life

rate of decay The decay of NMR observable PP did not appear to be exponential, the (nonJinear) being proportionately faster earlier in the incubation. Nonetheless, the

3tP days half-life of added PP can be estimated from NMR (Figure 4-3) as 15 to 21 from the time of application. The (non-linear) half-life of extractable PP can be

shorter estimated as 3 to 7 days from the Ic data shown in Figure 4-3. This is much much of than the value indicated by the NMR data(15-21days). This indicates that

longer the non-extractable P (i.e. undetected by NaOH extraction followed by IC) at

incubation times is PP.

the results of The halflife of PP determined by NMR (15-21days) is comparable to

was Ahmad and Kelso (2001) who found that approximately 50% of PP added

a pH 8'49' In hydrolysed 2l days after addition in a soil with 14% CaCO¡ w/w and of that only comparison they tested non-calcareous soils þH 4.51 and 5.93) and found

30-35%of PP hydrolysed over 2l days (Ahmad and Kelso 2001). Dick and (7.1% Tabatabai(1986) measured the hydrolysis of PP in a slightly calcareous soil

as comparedto 40-45o/o of PP CaCO¡ [\4//w]; pH 7.8) and found a half-life of 7 days, of Dick hydrotysed in 7 days in three non-calcareous soils (pH 5.3-6.4). In the work

has the and Tabatabai (1986), P was extracted using 2.5 M sulphuric acid, which

ability to hydrolyse PP in the extraction step (De Jager and Heyns 1998)' Khasawneh

8-16 days et al. (1979) found the halfJife for PP in a fine sandy loam to be from

(soil pH not reported)' using a NaOH extract and colorimetric measurcment of P

(pH and found Parent et at. (1985b) measured hydrolysis of PP in a low pH soil 5'1)

to the that the halflife of PP was approximately 40 days. The pH of PP was adjusted

after soil pH of 5.1 before the experiment was conducted, and it was extracted

addition to soil with 2.5 M sulphuric acid.

100 to In all of the aforementioned experiments, the hydrolysis reaction was considered

rate, which is have an initial rapid hydrolysis phase, followed by a slower hydrolysis most likely due to the PP being adsorbed or precipitated over time (Dick and

Tabatabai 1986; Ahmad and Kelso 2001).

and the Comparison of our data with the literature shows that both soil characteristics

P sorption extractability of P species in soils with a range of pH, CaCO¡ content and

PP half-life' capacities has a significant influence on the resulting measurement of

by IC with Our comparison between a NaOH extraction and subsequent P speciation 3tP proportion of non-invasive solid-state NMR spectroscopy suggests that a greater

PP added is measured by the NMR method.

solid-state The findings of this study suggest it is well worth investigating the use of

3tp under a range of NMR as a technique for specation of PP in a range of soil types,

method. Used in chemical and biological conditions to ascertain the limitations of the potential for conjunction with speciation of PP in soluble P and total P pools there is

of P species solid state'lp NMR to provide useful information about the distribution

amongst pools. to Hydrolysis of PP is considered to be the rate limiting step for PP to be equivalent

PP added to a OP as a P source (Khasawneh et al. lgTg). This study shows that

P nutrition highly calcareous soil has a half-life of 14 to 2l days after addition. Early that of cereal crops is essential in dryland agriculture and is one of the reasons

of PP fertiliser is placed in close proximity to seeds at sowing. Delays in hydrolysis

the final could adversely affect nutrition of young seedlings, which may compromise

yield potential of crops. One of the concerns regarding the use of polyphosphate

fertilisers in dryland alkaline soils was that hydrolysis of the polyphosphate ions Our might be delayed by low soil moisture, high soil pH and low biological activity.

101 studied, and data show that pp hydrolysis was relatively rapid in the calcareous soil

is hence availability of OP to plants should not be compromised. Further work

types' required to assess PP hydrolysis in a wider range of Australian soil

4.5. Conclusions

fate of Solid-state'lP NMR spectroscopy was successfully employed to follow the

on added PP in a highly calcareous soil. Orthophosphate and PP were differentiated

spinning the basis of differences in chemical shift and the size and distribution of

than the cP side bands. The NMR observability of P was much higher for the DP

P in these technique. Solid-state'lP DP NMR detected a greater proportion of total

that PP soils than did alkaline extraction followed by IC. Both NMR and IC indicated

a higher was hydrolysed to OP over the duration of the incubation, but IC indicated

of PP at rate of hydrolysis. However, this may be due to decreased extractability

longer incubation times, and suggests that the pool of P from which the hydrolysis

rate is being determined needs to be clearþ defined'

The hydrolysis of PP, as measured by NMR and compared with NaOH-extractable

of this values suggests that there are araîgeof reactions occurring beyond the scope

where PP is study which have important implications for the P availability processes

(Chapters 6- supplied as fertiliser. These reactions are currentþ under investigation

e).

102 Ghapter 5. Sorption of Pyrophosphate and OrthophosPhate in Soil

5.1. lntroduction

(P) This chapter discusses the partitioning behaviour of the dominant phosphorus

using the IC species of polyphosphate fertiliser in a number of Australian soil types,

P speciation technique described in Chapters 2 and3. Sorption curves for condensed

previous studies did species in Australian soil types have not been investigated, and

not use ion chromatography (IC) for speciation'

was In most previous studies, the concentration of condensed P species in solution

as the determined using colorimetry, with the polyphosphate concentration calculated

difference between total P and initial orthophosphate (OP). Due to the digestion

P would procedure required to determine total P, some weakly adsorbed and organic

of P be included in the solution P measurement, resulting in an underestimation

sorption.

of OP and Most of the studies conducted to compare the soil sorption characteristics

pyrophosphate (PP), the most abundant condensed P species in polyphosphate

for OP fertiliser, concluded that the sorption capacity of soils for PP was gleater than

1985; (Blanchar and Hossnet 1969a;Hashimoto et al. 1969; Mnkeni and MacKenzie

Al-Kanani and MacKenzie l99l). Blanchar and Hossner (1969b) used P speciation

more (by anion exchangc resin) and found that PP generally sorbed quantitatively

by the than OP, but that the sorption value determined for PP may be confounded only by hydrolysis of PP to oP. As PP hydrolyses to OP in soil, measuring sorption

Mnkeni and disappearance of pP from solution leads to an overestimate of sorption. to Mackenzie (1985) also found a greatet level of adsorption of PP than oP applied

103 organic matter surface soils. They attributed this to the capacity of PP to solubilise exposing more mineral surfaces for the sorption of PP'

than philen and Lehr (1967) suggested that PP reacts with soil minerals more slowly

P to sequester OP and yields different reaction products. The ability of condensed soil' In this metals such as Fe and Al influences their behaviour and speciation in

soil types' chapter the sorption of oP and PP was investigated in eight Australian

cation This chapter will also discuss the influence of the added P species on solution' concentrations, organic carbon concentrations and pH in the equilibrium

5.2. Materials and Methods

5.2.1. Reagents

grade The following reagents used for standards and sorption curves were analytical

PP decahydrate OP as sodium dihydrogen OP (NaHzPO4) (BDH) and PP as sodium

(Na¿PzOr 10H2O) (Sigma). Working solutions for standards were prepared daily'

5.2.2. Soil Types and Gharacteristics

sites Surface soil samples (0-10cm deptÐ were collected from eight agricultural

classified according across the southern cropping region of Australia. The soils were

in Table 5-1' to .The Australian Soil Classification' (Isbell 199'7) and are presented room All soils were dried at 40oC for 3 days, sieved (< 2 mm) and stored at

temperature before experimentation.

t04 Table 5-1: Soil sites and classification

Soil Australian State Location name Kandosol Western Australia Wongan Chromosol Victoria Hamilton Ferrosol Tasmania Ulverstone Calcic Calcarosol Victoria Walpeup Vertosol Victoria Minyip Grey Calcarosol South Australia Warramboo Red Calcarosol South Australia Emerald Rise Sodosol Victoria agricultural soils in Calcarosols are one of the most widespread and important victoria had caco¡ Australia. The calcic calcarosol used in this study from western

have a in the soil profrle, but not in the surface soil used in this study' Chromosols red subsoil and are strong texture contrast between the A and B horizon, often with

Ferrosols contain a Bz amongst the most widespread agricultural soils in Australia. and Tasmania' horizon high in Fe oxides and are found mostly in Queensland

large areas in some Kandosols range in their distribution across Australia and in very

of any country in the States. Australia has the greatest area and diversity of vertosols

world. They are clay soils with shrink-swell properties (Isbell 1997)'

5.2.3. Soil AnalYses

(Rayment and Soil pH (HzO) wase measured in a 1:5 soil:solution suspension, coupled Higginson lgg2). Total P, 41, Fe and ca were determined by inductively Spectro) plasma atomic emission spectroscopy (ICP-AES, Spectroflame Modula,

1996)' after digestion of soil sample s in aqua regia (zarcinas et aI. method of Colwell Bicarbonate-extractable P was determined according to the P was (1963) using 0.5 M NaHCO¡ (pH S.5) as the extractant. Resin-exchangeable 1994)' Soil determined using anion-exchange resin strips (Mclaughlin et al'

strips were removed and solutions were shaken with the resin strip for 16 hours. The

eluant with trvo hours P was extracted from the strip using a 0.1 M NaCl: 0'1 M HCl

105 index of shaking. Orthophosphate was measured colorimetrically. The P buffering Burkilt et al' for oP was determined using the single-point P sorption as described by

(2002).The P buffering index (PBI+corr) is described by the following equation:

Eq' 5-1 pBI + cop: (Ps + Colwell P)/C0'41 1000 mg Where P. is the amount of P sorbed (mg P/kg) from a single addition of

PBI + P wfls P/þ and C is the resulting solution P concentration (mg P/L)' The col

dissolved organic determined independently of the sorption curve anaþsis. Soluble TOC/TN carbon (DOC) was measured in a 1:5 soil: water extract using a Formacs@ method of the Analyser (2000). Particle size anaþsis was conducted according to the

usDA (19S2). calcium carbonate content was determined according to the

procedure of Martin and Reeve (1955)'

5.2.4. Speciation and Quantification of P

of the IC Ion chromatography was used for speciation of oP and PP. Details

operating system are given in section 2'2'2'

5.2.5. Sorption Gharacteristics

Table 5-1' sorption curves were derived for each of the eight soils presented in

OP or Phosphorus sorption curves were determined using solutions of sodium with two drops of sodium pp in 0.01M KCl at 8 concentrations from 0 to 300 mg P/L

to reduce the toluene to inhibit microbial activity. Microbial activity was inhibited

potential for enzymatic hydrolysis of PP to oP. The pH of these solutions was

polyphosphate adjusted to 7 .25,which is equivalent to the pH of a commercial

This also fertiliser solution containing 100 mg P/L (Mnkeni and MacKenzie 1985)'

ratio was 1:10 (4 g served to reduce PP hydrolysis (Chapter 3). The soil to solution

shaker soil: 40 mL solution). The suspensions were equilibrated in an end-over-end

106 then for 24hours (17 rpm) at room temperature (= 25"C)' The samples were centrifuged (1000 g) for 30 minutes. After centrifugation, 5 mL of the suspension

pm & ìvas removed with a plastic syringe and filtered through a0.2 Schleicher by IC' Schuell membrane. The concentrations of OP and PP were determined

the equation: Phosphorus sorbed (PJ (mg P/kg) for each species was calculated by

Eq' 5-2 Ps (mg P/kg) : (P added - P in solution) x (Jl) m

(mg/L) is the amount Where P added is the amount of P added (mglL),P in solution

(mL) and m of P remaining in solution after equilibration, v is the volume of solution

is the mass of soil (g).

equation in The P sorption data were then summarised using the Freundlich sorption

the following form:

Eq' 5-3 P,(mg P/kg): KrCo of P in Where Kr and n are empirical constants with n< 1 and C is the concentration

solution (mg P/L).

The partitioning coefficient (Kd) was determined using the equation:

5-4 Kd (L/ke): P./C Eq'

and The concentrations of Al, ca and P in the samples following filtration of DOC centrifugation were measured using ICP-AES. The pH and concentrations

organic were measured in the supernatant of the equilibrium solution. Dissolved

(2000)' carbon rwas measufed using a Formacs "t TOC/TN Analyser

5.2.6. Statistical AnalYsis

matrix' Using The Genstat@ 6 statistical package was used to produce a correlation

value of the r-value from the correlation matrix, and the Pearson's test for the critical

the correlation coefficient, the significance of each interaction was determined'

l07 Sigma Plot@ 9.0 was used for curve fitting of the Freundlich equation.

5.3. Resulfs and Discussíon

5.3.1. Soil Characteristics

(Table 5-2) across The ten measured soil characteristics showed a considerable range

to 8'7, while the soil types tested. The pH of the equilibrium solution ranged from 5'2 from 0'02 total Al varied from 0% to 5.4Yo,total Fe from 0.01% to 8.9o/o and total Ca

the acid to to 24.5 Yo. TheGrey Calcarosol had a CaCO3 content of 68Vo while

ranged from neutral soils did not contain any detectable CaCO¡. The DOC content

In terms of P 0.3%o ínthe Calcic Calcarosol up to nearly 5o/o inthe acidic Chromosol'

mg P/kg availability, resin P ranged from 0.8 to 6 mg P/kg, Colwell P from 1 to 63

and the P buffering index (PBI+c"n) from 16 to 655'

Table 5-2: Soil characteristics.

ColP PBI Soil Tlpe pH Total Total Total CaCO¡ ClaY DOC Resin A1 Fe Ca P colP HzO % % %%% % mgP mgP

5.5 I6 Kandosol 5.2 0.1 0.1 0.1 0.0 11.0 0.5 0.77 63.0 158 Chromosol 5.4 5.4 8.9 0.2 0.0 25.0 4.7 6.43 29.0 655 Ferrosol s.6 0.5 0.4 0.0 0.0 24.0 0.5 8.48 1.0 70 Calcic 7.2 1.0 0.0 0.1 0.0 2.5 0.3 0.96 Calcarosol 2.08 1.0 t24 Vertosol 8.1 3.4 0.1 1.3 1.8 51.9 1.5 25.0 2l Grey 8.4 2.3 0.0 24.5 67.5 4.6 1.0 r.94 Calcarosol 113 Red 8.5 0.7 0.0 2.0 4.8 s.l 0.4 3-23 t2'0 Calcarosol 123 Sodosol 8.7 3.3 2.4 0.7 0.7 32.9 t.2 5.79 19.2 *ColP: Colwell P

5.3.2. Sorption Characteristics

(version 9'0) with Sorption cgrves were draw¡ for each soil type using Sigma Plot

(Figure 5-1) were curves following the Freundlich equation. The resulting curves

108 stronger reflective of the wide range of soil types tested but consistently showed a

all cases the P in sorption affinity where P was supplied as PP as compared to OP' In

P was supplied as solution in unamended soils was negligible (< 0.1 mg P/L). Where

and the sum of OP + PP PP, curves were drawn for both PP in solution vs. PP sorbed in solution in solution vs. oP rL PP sorbed. It is important to note that oP measured

and from when P is supplied as PP could be derived from PP by hydrolysis

the processes mobilisation of native P, and it is not possible to distinguish between

was without isotopic labelling. In the Kandosol the sum of OP and PP in solution

greater than 150 mg greater than the amount of PP added at PP application rates of

application P/L. This highlights the possibility of mobilisation of native P at high

rates of PP, which will be investigated further in chapter 9.

precipitation of solid- The shape of the sorption curves suggests that there \ilas no

(1979) who suggest phase PP species, contrary to the findings of Khasawneh et al.

to the soil' If that PP undergoes rapid precipitation reactions upon addition

have precipitation had occurred, an upward inflection in the sorption curve would

exceeded. If the pÇn been observed at high rates of PP addition, where the pÇn was in solution with was exceeded, there would be little or no change in P concentration

indicative of a increasing rates of PP added, but increasing amounts of P sorbed,

increading precipitation reaction. In the sorption curves presented (Figure 5-1), with

concentration rate of pp added the amount of P sorbed plateaued, as the equilibrium

the solid phase were of P in solution increased, suggesting that the P sorption sites on

was diminishing approaching saturation, or that the surface potential on the surfaces

(Barrow 1987).

109 800 2000

600 1500

Èr 400 E E 't 000 o à zoo o L 500

-200 '100 150 200 100 200 300 50 in solution (mg/ L) P in solution (mg/L) P

3000

2500 150

2000 È, E E ! 1500 a 100 o o o o 1000 È

500

1@ '100 150 200 s P in solution (mg/L) P in solution (mg/L)

3000 3000

2500 2500

2000 2000 xo E E 1500 1500 o €o o G È 1000 1000

500 500

0 0 50 100 150 200 50 '100 150 200 P ¡n solution (mg/L) P in solutlon (mg/L)

2000 2000

1 500 1 500 o o o E E ! 1000 ! 1000 ¡o € o È ù

600 500

0 100 150 s 100 1S æ0 s P solution (mg/L) P ¡n solution (mg/L) in

curves tr'igure 5-1: Sorption data for oP and PP in eight Australian soils and fitted are derived using the tr'reundlich equation. For soils treated with PP' curves (OP plotted for uãtn PP in solution vs. PP sorbed and the sum of oP and PP -tr-' + rr¡ in solution vs. oP and PP sofbed (oP treatment -l-' PP treatment PP treatment wÍth measurement of OP and PP -r-)'

110 presented in The Freundlich sorption parameters derived from the sorption curves are

than for OP' Table 5-3. The sorption affinity of soil for PP was consistently stronger

its value due even though the Freundlich sorption constant (Kf) was wide ranging in

1.4 for the to the range of soil types tested. For OP the Kf value ranged from

the calcic Kandosol to 556 for the Ferrosol. For PP the Kf value ranged from 0 for

than calcarosol to 2123 for the vertosol. The Kf values for PP+OP data were 10wer

that of PP alone and ranged from 0 for the Kandosol to 948 for the Vertosol'

A poor fit of the Langmuir equation was observed on several of the soils. Therefore,

(1978) only a Freundlich equation was fitted to the data. FurtheÍnore Barrow

than a more suggests the use of simple equations in describing adsorption is better

the data and complex approach, and that equations should only be used to summarise

not to infer P bonding mechanisms.

(mgkg) : Pi kfn with a fit of Tabte 5-3:- Freundlich sorption parameters where P sorbed R1 pi is the coìcentiation in solution (mg PiL), Kf and n are Freundlich sorPtion Parameters

OP PP OP andPP Soil Kf n R2 Kf n R2 KfnR' 0 r.7l 0.35 Kandosol 1 1.03 0.95 129 0.27 0.98 Chromosol 80 0.47 1.00 340 0.37 0.97 29s 0.22 0.98 Ferrosol 5s6 0.28 1.00 1233 0.28 0.95 863 0.29 0.92 Calcic 11 0.20 0.85 35 0.22 0.95 2I 0.30 0.95 Calcarosol Vertosol 113 0.46 1.00 2123 0.08 0.91 948 0.35 0.96 Grey 53 0.52 0.99 671 0.34 0.98 322 0.49 1.00 Calcarosol Red 56 0.3s 1.00 480 0.28 1.00 219 0.47 0.98 Calcarosol 0.64 0.99 Sodosol 39 0.47 1.00 318 0.40 0.98 111 soil A single-point partitioning coefficient (Kd) was derived for OP and PP in each

was using the treatment where 10 mg P/L was added (Table 5-4)' This treatment

point on the chosen, as it was the lowest level of application and therefore the

111 expected the Kd for PP sorption curve most likely to be in the initial linear phase. As was greater than that of OP in all soils'

phases (Kd) for Table 5-4: partitioning coefficient between the soil solid and solution OP, PP, Op and PP at tP addition of 10 mg/L' "tt¿ Kd (L/ke) Soil OP PP OP and PP Kandosol 6 135 22 Chromosol t6l 1662 t67r Ferrosol t534 4231 1680 Calcic Calcarosol 3 62 39 Vertosol 78 1677 825 Grey Calcarosol 1539 2747 1558 Red Calcarosol 97 835 2r7 Sodosol 99 836 55

41, Ca, Fe, clay Soil factors that influence the sorption of P in soil include pH and

regression was tested and organic matter content (Celi et al. 2000). A multiple linear

the sorption to determine if a single or several soil characteristics influenced

\¡/ere determined due to parameters determined. However, no successful predictors

soil characteristics' an inadequate number of data points and co-alignment between

parameters and The best available statistical test of the relationship between sorption with significance soil characteristics was the use of a correlation matrix (Table 5-5)

correlation coefficient' determined by the pearson's test of the critical value of the

Dubus and Becquer This approach is similar to that used by Zhot andli (2001) and coefficient (2001). At the 5% significance level the critical value of the correlation

> between the Kd and Kf was 0.71. There was significant correlation with an r 0-71

as PP was expressed of OP ancl PP, regardless whether the Kd and Kf of P supplied

Kd of PP and Kf of oP as measured PP or as the sum of measured oP + PP' The

was correlated with were positively correlated with PBI +colP, while the Kf of OP

parameters has been resin P. The correlation between the PBI+"otp and sorption

observed previously (Bolland et al. 1996)'

tt2 The correlation matrix demonstrated the relationships between soil characteristics that discounted the use of multiple linear regression due to co-alignment, including

C, total the relationships between total Al and Fe, total Al and org C' total Fe and org

P' Fe and Colwell P, total Ca andpH, total Ca and CaCO¡ and organic C and Colwell

Surprisingly the PBI+c6¡, and resin P were positively correlated.

Table 5-5: Correlation matrix of soil characteristics and P sorption parameters' Ilighlighted r values are significant at P S 0'05'

OeY OS C ø\dlP FB+dP }6CP KICP+FP IøFP KCP KCP{FP KFP Èr Tdd A TddFe TcHø @ bnP r€cP 1.m KJCP+FP aß 1.m ICFP 0¡9 0s 1.m KCP 066 058 ot!4 1.o KCPTFP 004 0â 0.æ 015 1.m KFP 0.43 060 r74 A7 otr 1.æ 'lm pl-l 46 4.Ð 414 {36 034 {o1 TdA 4.17 0.40 06 {22 0.o 015 011 1(D Tdd Fe {20 040 o(P 46 419 {.10 038 tig 1.O TcH G 4'3 o6 417 4â 047 014 076 06 4A 1.m OF 1.m c4 064 040 03 415 0ß {@ 038 004 4Z ht 416 018 02. aa o74 06 00/ 054 02. o16 45 l.co 1.O ogc 417 0.s oa 411 0.07 0ü/ 431 088 6 413 410 036 0.3 1.m FHnP o5t 03+ ad o72 4æ 0¿10 {2ô OO 0s o18 {2ô 03 0P 1.æ ondP 0æ o72 oß o24 48 06 04 0@ 0g 48 0æ om 068 {ræ 021 (ìfi) 0.æ 10 FB+odP 05r 051 trTB os 06 063 438 {.,l8 o(2 48 {.6

parameters Due to the observation of a correlation between PBI a.o1p and the sorption

(Figure 5- of Kd pp and Kf Op, these relationships were plotted to test their linearity

2). Theresulting plots suggest that the Kf for OP has a strong 6'z: O'lO¡ relationship

withthePBl+colP,whereasthisisnotthecaseforPP(Figure5-28).TheplotofKd

(Figure 5-2/t) is more against the PBI as61p sttggosts that PP partitioning (R2:0.5S)

that these dependent on PBI a"oç, than is OP. However it is important to note,

Ferrosol, relationships were strongly leveraged by the data points resulting from the

thc analysis. and these relationships do not hold in the absence of this soil type from

This finding was confirmed by using a log transformation to normalize the variance that of the PBI+colP including the Ferrosol data (graph not shown), which showed

the relationship was not significant.

113 A B 2500 4500

4000 tr 2000 3500

3000 1500 o 2500 J :< € 2000 :¿ 1000 I 1500 I

1 000 -.ø l 500 I 500 I o 0 0 0 100 200 300 400 500 600 700 0 200 400 600 800

PBI + col P PBI+colP

(A) partitioning Figure 5-2: Relationship between P buffering index (PBI *dP) and the constant (KÐ' OP coefficient 1kA V¡*¡ and (B) the Freundlich sorption P treatment with measurement of OP and , Rf=0.58, (B), OP OP and PP measurement y=1'11** 169'43,Rf:0'40'

and Sutton and Larsen (Ig64),Blanchar and Hossner (1969a) and Al-Kanani

MacKenzie (1991) also showed that PP had a greater sorption affinity than OP.

regard However, when comparing all available studies there are conflicting data with

and to the extent and bonding energy of OP and PP adsorption reactions (Sutton

Larsen 1964;Hashimoto et at. 1969; Savant and Tambe 1979). The wide range of

for the large methods and soil types used to test this relationship could partly account

variance in results between authors (Al-Kanani and MacKenzie l99I). Experimental

conditions that influence this difference include incubation time, temperature,

content method of shaking, solution to soil ratio, supporting electrolyte and moisture

that the of the soil prior to equilibration (Barrow 1973). Morel et al. (1996) suggest

time dependence of conventional sorption curves makes comparisons between

that the experiments not performed uncler identical conditions invalid' They suggest

time use of isotopic dilution techniques to investigate partitioning remove this

TT4 dependence and enables comparison between experiments. This is only partly true as isotope exchange is also time-dependent.

Amer and Mostafa (1981) and Guan et at. (2005) concluded that the ratio of PP to

Op plays a very important role in the extent and type of P retention reactions in soil.

This is related to the process of competitive adsorption where the species with the greatest affrnity is preferentially adsorbed(EI-Zahaby and Chien 1982). Guan et al-

(2005) showed that the competitive adsorption between different phosphate species modifies the surface properties of the adsorbent differently and therefore potentially affects the interaction between the adsorbent and cations andl or organic matter in soil. A comprehensive investigation of these effects is beyond the scope of this study.

However, the proportion of P in solution as PP was determined at the intermediate

(150 mg p/L) and highest (300 mg P/L) levels of P application. The data (Figure 5-3) indicate that the higher the level of P application, the greater the proportion of P that was present as PP. Therefore PP was more stable when added at higher concentrations. The sorption curves suggest that PP had a greater sorption affinity than Op. The increased stability of PP with increasing concentration may explain why the sorption curves do not appear to approach an adsorption maximum in this working range.

115 100 90 o 80 (u E 70 60 (to o 50 ! o. 40 Lo 30 o- 20 .t) 10 (! 0 fL s

Soil

tr'igure 5-3: Percentage of total P in solution present as PP when 150 mg P/L (150 PP) and 300 mg P/L (300 PP) as PP were added to each soil. 150 PP treatment - ¡- and 300 PP treatment -r-.

5.3.3. pH

The pH was measured for all equilibrium solutions. The resulting pH values at the lowest (0 mg p/L) and highest (300 mg P/L) level of P application and the resulting difference are presented in Table 5-6. For OP addition there was an increase in pH for acidic soil types, but the change decreased as soil pH approached neutrality. The pH decreased with OP addition to alkaline soil types and the degree of this change

decreased with increasing alkalinity.

pyrophosphate addition similarþ increased soil pH in acidic soil types, while it

tended to decrease pH in alkaline soil types, but these changes were less consistent

than they were for OP. The difference in soil pH between the addition of the

equivalent amount of P as OP or PP was less than one unit exccpt in the Grey

Calcarosol. As the pH data were not consistent, this suggests that pH was not the

only soil factor controlling the differences in P sorption between the P species, and

that there is most likety an interactive relationship between pH and other soil

116 characteristics. The pH changes are consistent with the known ability of solution P

species to act as pH buffers.

Table 5-6: Equilibrium pII values for the lowest (0mg P/L) and highest (150 mg PiL) levels of P aPPlied as OP and PP.

OP PP Soil 0 me PiL 300 mg P/L Change 0 PIL 300 PIL Kandosol 5.95 6.8s 0.90 5.63 6.79 t.t6 Chromosol 5.22 6.12 0.90 5.12 6.03 0.91 Ferrosol 5.32 6.1s 0.83 s.40 6.21 0.81 Calcic 6.96 7.08 0.12 7.08 6.87 -0.21 Calcarosol Vertosol 7.63 6.85 -0.78 7.74 6.87 -0.87 Grey 8.2s 7.29 -0.96 8.15 8.55 0.40 Calcarosol Red 8.06 7.45 -0.61 8.21 7.87 -0.34 Calcarosol Sodosol 7.60 7.t7 -0.43 7.s8 7.26 -0.32

5.3.4. Gation Goncentrat¡ons

Changes in cation concentrations differed between the P species due to differences in

the sequestering capability and the hydrolysis reactions of PP (Philen and Lehr

lg67).Across all eight soil types considerably more Fe and less Ca was generally

detected in solution with P supplied as PP (Figure 5-4). Pyrophosphate is able to

sequester Fe, which results in a release of Fe into soil solution with increasing rates

of pp supply (Philen and Lehr Ig67).In the study of Philen and Lehr (1967), no solid

Fe compounds were formed in the reaction of condensed phosphates with clay

minerals or their impurities. The Fe was sequestered by the condensed phosphate,

ancl remained in solution even after considerable hydrolysis to OP.

The concentration of Ca was depressed in soil solution with the addition of either P

source (Op or PP). In all except for 2 of the soils the depression of Ca in soil solution

was greater with the addition of PP than with OP. As PP has a greater sorption

affinity for these soils than OP, there is a greater negative charge on the soil surface

tt7 when PP is added to react with Ca, removing it from solution. However, as previously mentioned it is unlikely Ihatthere was precipitation of Ca due to the shape

ofthe sorption curves.

118 A B Kandosol (pH 5.2) Chromosol (pH 5.a) 35 80 9 80 70 I 70 30 I j J 7 60è ô03 o) Þzsts ç E E 50e o 50- E c C É,o 5 o F +o 'P .9 4oE= l f J 4 õ15 U) aoE Ø 30t 3 c .= .E to 20o (o 20 1) (ú 2 tr IL os 't0 o 10 1 0 0 0 0 o 10 50 100 150 200 250 300 o 10 50 100 150 200 250 300 P added (mg/L) P added (mg/L)

Ferosol (pH 5.6) Calcic Calcarosol (PH 7.2) 60 250

20 30 2oo ^50J a 18 è, È, lro 25j 940 E o) È, rso !' E-14 zoE c o E o F c12 E30 f o 15=o õ roo Ø S õ .c 2o .c =10EB 'to o (ú o) .=ô .= () 50 l! (5¿ o '10 o 5l! 2 0 0 U 0 o lo 50 100 150 200 250 300 'to 50 100 150 200 250 300 P added (mg/L) P added (mg/L) E F Vertosol (pH 8.1) Grey Calcarosol (PH La) 5 160 5 50 45 140 4j i40 4 J j È, è) 120 d) E E E E* 100 3c c30 3 c -C o o 'a E25 l = 2 õ =80 zl Ezo Ø E60 .= c .E .g 1s '. (¡) o 40 rf 8ro IL 20 5 0 0 0 0 50 loo 150 200 250 300 0l0 50 t00 150 200 250 300 010 P added (mg/L) P added (mg/L) H Red Galcarosol (pH 8.5) Sodosol (pH 8.7)

1.0 60 120 1.0 0.9 0.9 150 08 Q I 100 o.s e È, 07 o 07 €+o P Eao P c 0.6 ; o.ô c o o oE o 'E 30 0.5 E E60 0.5 'E- õ 0.4 õ 0.4 õ Ø õ Ø 0.3 40 .Z,o .E .? 0.3 .E 8ro 02 fl 0.2 I 0.1 8zo 0.1 0.0 0 0 00 100 150 200 250 300 o 10 50 o l0 50 100 150 200 250 300 P added (mg/L) P added (mg/L)

Figure 5-4: Cation concentration in solution (mg/L) in eight soils A-II, as a function of p in solution (mg/L) where P was supplied as OP or PP. Cations measured were calciun (C;) and iron (Fe). OP treatment Ca -l-, PP treatment Ca -r-, OP treatment Fe'0-, PP treatment Fe -¡-.

It9 5.3.5. Organic Garbon

For most of the soils tested, PP addition resulted in a greater concentration of total

DOC in the equilibrium solution (Figure 5-5). Exceptions to this were the Sodosol

(pH S.7) and the Calcic Calcarosol (f,lH7.2). The reason for this difference is

unknown, as these soils do not possess similar soil characteristics (Table 5-1). The

most likely reason is the heterogeneity in the soil samples, as the DOC for 0 mg P/ L

is already considerably lower for the PP treatment than for the OP treatment. For all

other soils, DOC is the same for the OP and PP 0 mg P/L treatment. Pyrophosphate

is capable of solubilizing organic matter, which plays an important role in the

sorption and hydrolysis reactions of PP (Mnkeni and MacKenzie 1985). The organic

matter solubilising capability of PP is widely accepted, and PP is routinely used in

soil testing laboratories as an extractant for organic carbon (Dijkstra and Fitzhugh

2003; Ussiri and Johnson2004).

t20 A B Kandosol (5.2) Chromosol (pH 5.4) 200 600

500 150 ^ ¿oo : 3ct o, a e 19 too 5 300 O C) o a o 200 oso o o a 100

0 200 0 50 100 150 200 250 300 0 50 100 150 P in solution (mg/L) P in solution (mg/L)

C D Fenosol (pH 5.6) Ca c c Ca caroso (PH 7 2) 600 200

500 J 150 Q +oo cn o) o q E 100 I 300 r) O () o o 200 o o o 50 100

0 0 200 250 0 50 100 150 200 250 0 50 100 150 P in solution (mg/L) P in solution (mg/L)

E F Vertosol (pH 8.1) Grey Calcarosol (PH La)

200 600

500 150 ^I Qg) +oo o C -9 roo I 300 () () o 200 oo50 o '100

0 0 100 150 200 0 50 100 150 200 250 0 50 P in solution (mg/L) P in solution (mg/L)

G H Sodosol (pH 8.7) Red Ca caroso (pH 8 5) 200 200

150 J 150 I o) q) c E 100 5 100 () O ¡) o o o 50 o50

0 0 200 250 0 50 100 150 200 0 50 100 150 P in solution (mg/L) P in solution (mg/L)

tr'igure 5-5: Concentrations of DOC in solution (mg/L) as a function of P in solution (mg/L) in eight soils (A-Ð (OP treatment -l-, PP treatment -r-)

t2t 5.4, Conclusions

This study has shown that PP has a stronger sorption affinity than OP in Australian soil types. In general, the addition of PP to soil resulted in a decrease in Ca concentration, an increase in Fe concentration and an increase in DOC in soil solution as compared to OP. The different effects of OP and PP addition to soil on pH, organic carbon and cation concentrations suggest that the investigation of the mechanisms causing these changes may add to our understanding of pollphosphate fertiliser chemistry in soil. Comparison with previously published results is difficult due to the different methodologies and materials used. Part of the uncertainty with regard to sorption data for PP is the unknown amount of hydrolysis, which may occur during the sorption measurement, thereby overestimating PP sorption- Isotopic labelling of PP potentially overcomes this limitation and will be described in Chapter

Measurement of P speciation using ion chromatography suggested that in the case of the Kandosol there was more P in solution than that added as PP. The possibility of mobilisation of native P reserves through the addition of PP to soil will be further

investigated in Chapter 9.

t22 Ghapter 6. lsotopic Techniques for lnvestigating Reactions of Pol hosphate Fertilisers in Soils

6.1. lntroduction

Conventional techniques for the determination of soil P availability have made use of chemical extraction procedures to estimate the concentrations of P in pools that supply P to plants. However, finding an extractant that has the capability to represent the bioavailable P fraction in all soil conditions is a continuing effort.

A different approach to the problem of estimating soil P availability was made possible by the introduction of isotopic labelling techniques. In this case, an isotope of P is added to soil and equilibrates with soil solution and surface exchangeable solid phase P. This allows the determination of readily exchangeable and 'chemically reactive' P in soil, which will be described in further detail below. However, as the reactions of P in soil are continuous, a true equilibrium is not achieved, and this

experimental equilibrium represents only a semi-stable state (Amer et al. 1955).

Isotopically exchangeable P therefore does not directly represent bioavailable P, but

rather provides information about the processes that contribute to bioavailability.

An advantage of using isotopic techniques for estimating soil-P availability is that

the experimental conditions are similar to field conditions, except for adding an.

excess of water. In contrast, chemical extraction procedures rely on the use of

extreme shifts in pH or complexation to estimate P pools (Amer et al' 1955).

This chapter will describe aïaîge of isotopic techniques available for the

32PP 33OP investigation of the contribution of PP and OP to the labile pool using and

both in single and dual labelling systems. This chapter provides an introduction to

the isotopic studies described in chapters 6-8 and details the principles used to

develop this work.

123 6.2. Assumptions and Principles of Isotopic Techniques

6.2.1. Tracer Techniques

Tracer techniques are characterised by the lactÍhatthe isotope used in the study is

applied to the soil together with the fertiliser and therefore the fertiliser is

radiolabelled before it is introduced into the system. In the tracer system the isotope 3rP added will behave and react in the soil-plant system in the same way in which

does. By knowing the specif,rc activity (SA) of the P in the labelled fertiliser it is

possible to quantitatively trace the distribution of the fertiliser P in different

compartments of the system such as plants and soil pore water. The SA (Bq/mg P) in

a given pool is determined by the equation:

Eq' 6-1 SA: r Psolution

'Where 3rP r is the radioactivity remaining in the pool of study (Bq/L), and P is the

concentration (mg P/L) in the pool of study.

The amount of P derived from a labelled P addition in a given pooVcompartment (q

in mg P/kg soil) is calculated as:

Eq.6-2 q = (#)x Psolution x I

'Where SA, and SAn are the specific activities (kBq/mg P) of the pool and the P

addition, respectively. The volume (v) of solution in mL divided by the mass (m) of

soil in g represents the soil:solution ratio.

(Fardeau 1996; Bunemann et aL.2004)

The following assumptions are made when using tracer techniques.

1. The dose of the tracer does not influence the behaviour of the tracee'

2. Lab elled and unlabelled components are treated identically.

t24 6.2.2. lsotopic Dilution Techniques

The isotopic dilution method uses isotopes to determine the size of the exchangeable or potentially available pool without extracting it from the system (Fatdeau et al. l9g6).In this case the isotope is often added "cafüer free" which means with virtgally no other isotope of the same element. Isotopic dilution makes an assessment of the exchangeable or labile pool of P, termed the E-value. This technique differs from a tracer technique mainly for the factthatin this case the isotope is introduced in a system at equilibrium, whereas in the previous case it is introduced together with the fertiliser (which will change the equilibrium of the element of interest in the system).

When a small amount (that will not significantly change the pre-existing equilibrium) of isotope of the element to be studied is introduced to the soil it will distribute amongst the solution and exchangeable phase on surfaces in the same way as other isotopes of the same element. The exchangeable phase is in equilibrium with the

solution phase.

Eq' 6-3 r P soln - soln r exch P exch

where r.o6 (Bq/L) represents the activity of the isotope in the solution phase, and r"*.¡

(Bq/kg) the exchangeably adsorbed isotope on the soil solid-phase, respectively. P.o¡'

3lP 3rP"'.r' (mgll) is the concentration of unlabelled in solution, and (mg/kg) is the

concentration of isotopically exchangeable P on the soil solid phase. The sum of rro6

and r.*"¡ will be equal to the total radioactivity introduced (R) and the sum of P.om

and P.*"¡ will be equal to the labile P pool (Talibudeen 1957)'

Sampling and analysis of the solution phase allows determination of the partitioning

solution and solid phases. The soil to solution ratio (I) of isotope between the m

t25 requires consideration in the following calculation where the labile P pool is

described.

The labile pool is termed the E-value (mglkg) and can be calculated according to the

following:

Eq.6-4 P soln E value (P P = xRxI - = soh+ exch) r soln m

Where P.on is the concentration of soluble P remaining in solution (mglL), r.os is the

radioactivity remaining in solution (Bq/L), R is the radioactivity introduced (Bq/L)' v

is the volume of solution (L) and m is the mass of soil (kg)'

the value If a plant is used to 'sample' the ratio between P and r in solution, then

obtained is called an L-value (Larsen 1952).If the plant does not mobilise the

element of interest in the rhizosphere (Grinstead et al. 1982;Di et al. 1997) and if the

seed contribution is considered (Bertrand et al.2006), then the E and L values should

be identical so long as equivalent equilibration periods are used (Frossatd et al.

1994a). The assumptions underpinning isotopic dilution techniques are that:

1. The isotope introduced has not changed the equilibrium of the element under study

Z. A1lunlabelled element measured in solution is isotopically exchangeable

3. The isotope has physically mixed with the entire labile element pool'

(Hamon et a|.2002)

The equations above demonstrate that the determination of an E-value involves a

number of independent measurements. Therefore, by combining these measurements

for the calculation the propagation of error is an important consideration.

6.3. Double Labelling Techniques

32OP Traditionally labelling of OP with o.3'OP for determination of potential

availability of P in soils and plants in the form of E and L values has provided

126 sufficient information on P fertiliser: soil reactions. The diffrculty with a polyphosphate fertiliser is the fact that phosphate exists in several species in the product and the more complex P species hydrolyse in certain conditions to form more simple (less phosphate group) P species.

A double labelling technique involves using two isotopes of the same element to

label two chemical species of that given element. The dominant P species in a

polyphosphate fertiliser are OP and PP. Ilr order to meet the requirements of

assumption 2 of the isotopic dilution principles, in which all P measured is

isotopically exchangeable, it is necessary to ensure that both OP and PP lability are

32OP to measurable with the added isotope/s. There is uncertainty about the ability of

3rPP. 32OP OP measure the lability of It is probable that the only exchanges with the

portion of the fefülizer,giving erroneous values of potentially available P from PP' 3rOP 33OP 3rPP 32PP Double labelling by measuring the lability of with and with can

ensnre that assumption2 is upheld.

In the double labelling experiments, both the isotope introduced and the species

under investigation need to be determined. Therefore, in the case of double labelling

of Op and PP both the radio and stable isotopes of P need to be speciated and

determined. This requires coupling of the isotopic techniques with a speciation and

separation procedure such as ion chromatography.

The two P isotopes for both OP and PP can be used in tracer experiments and

isotopic dilution experiments. Flexibility in the approach and design of experiments

is required in order to provide information on various aspects of the PP dynamics in

the system, including hydrolysis, exchangeability and mobilisation of native P'

t27 6.4. lsotopic ExPerimenfs

6.4.1. The Processes Under lnvestigation

The dominant P species in a polyphosphate fertiliser are OP and PP. When polyphosphate fertiliser is added to soil (Figure 6-1) it moves into the solution pool as PP or can hydrolyse to form OP. Also present in the solution pool is native OP.

The P species in the solution pool undergo sorption reactions with the exchangeable phase, and the exchangeable phase, like the solution phase, will contain PP and hydrolysed and native OP. Phosphorus in the exchangeable phase can participate in desorption reactions and return to the solution phase or precipitation reactions and move to the fixed pool. Phosphorus in the fixed pool could be present as PP and./or both hydrolysed and native OP. Dissolution or diffusion reactions from the fixed pool to the exchangeable pool generally occur on a much smaller scale than the sorption and precipitation reactions.

To study the potential availability of polyphosphate fertiliser in soil it is necessary to

31OP, 33OP develop isotopic techniques that will measure the distribution of "PP,

32PP and amongst the solution and exchangeable phases. Considering the reactions that PP undergoes when added to soil it was necessary to develop a method that is

capable of measuring the potentially available P as OP from native P, PP and OP

derived from the hydrolYsis of PP.

128 Fertiliser lnputs

FertiliserP 1"oP, "PP¡

Fertiliser Band

Fixed SoilS hangeable P pool -g P pool P pool 31oP -o(ú atç_ ægp srgp_ srgp ofL (native or fertiliser OP) coo)- (ndive or fstlliser (native or fertiliser OP) op oP) 31PP ço 31PP_32PP .xõ Hþrolysis TO 31PP-32PP Ityûqtys¡s oc. I z + t .roP_rroP 31oP

Radol$elled Pod

Figure 6-1: Distribution of P added as OP and PP in polyphosphate fertiliser amongst the solution, exchangeable and fixed P pools. The measure of potential availability of OP and PP occurs in the radiolabelled pool containing the solution and exchangeable pool using the radioactive isotopes "PP and "oP. 6.4.2. Ghapter 7- A Hydrolysis Study

32OP Initially atracertechnique was used to determine the ability of to exchange 3lpp 32PP with and then a comparison was made of the ability of to exchange with

.tpe to address the concerns lchapter 7) in a single labelled system. This work was 32OP of Lombi et al (2004a) and Bertran d et al. (2006) that is not suitable to

t29 3lPP, effectively measure the lability of and therefore would provide an effoneous estimation of the lability of P from polyphosphate fertiliser'

This tracer technique was then applied in a dual-labelled system where "PP was

32PP, 33OP labelled with and was added carrier free. The effect of increasing additions of PP on PP hydrolysis and recovery, over 24hours, was investigated in a range of soils (Chapter 7).

6.4.3. Ghapter 8- A LabilitY StudY

Using the isotopic dilution technique the differences in lability or isotopic exchangeability of OP and PP in araîge of soils over time (1, 3 and 7 days) were investigated (Chapter 8). Due to the requirement to measure the lability of PP, OP derived from hydrolysed PP, and native OP in a single system, a dual labelling system was used. Pyrophosphate isotopic dilution was measured at the end of the

3tPP. incubation period with carrier-free Lubility of native OP (in the PP treatment)

and OP derived from the fertiliser (in the OP treatment) were measured at the same

33OP. time with carrier-free

6.4.4. Ghapter 9- A Mobilisation Study

32PP In a dual labelling system the hydrolysis and distribution of PP added with

tracer was determined over one week while simultaneously measuring the labiliiy of 3'Ot native P by using a carrier-free technique with lchapter 9). The aim of this

work was to test the hypothesis that PP may in fact be able to mobilise native soil P.

The mobilisation of native P by PP addition was previously inferred by Torres-

Dorante et al. (2005).

130 6.5, Conclusions

32pp 33OP The use of and to investigate the fate and lability of PP and OP using both

Iracer and isotopic dilution principles enabled a comprehensive investigation of the reactions of polyphosphate fertilisers in soils and the resulting processes of these reactions will be described in Chaptersl-9.

The objectives of the isotopic component of this PhD study were:

32oP 1. To evaluate whether equilibrates with the PP supplied in an APP

fertiliser (Chapter 7);

2. To evaluate the hydrolysis reactions of PP supplied at a range of

concentrations (ChaPter 7);

3. To evaluate the effects of incubation time on the lability of P supplied as OP

and PP (Chapter 8);

4. To investigate the chemical changes in soil caused by PP addition as

compared to OP addition, in particular changes in pH, concentrations of

DOC, Ca, Fe and A1 (ChaPter 8);

5. To determine if these chemical changes in soil were the cause of changes in

lability of P from PP as compared to OP (Chapter 8); and

6. To determine if the addition of P as PP was able to release native P into the

exchangeable pool (ChaPter 9).

131 Ghapter 7. An lsotopic Study of Pyrophosphate Hydrolysis

7,1. Introduction

Tracer techniques with radiolabelled P are often used to measure plant uptake of P from fertilisers (Dean et al. 1947; Morel and Fardeau 1989), and have previously been used to trace P uptake from a PP fertiliser (Sutton et al. 1966;Leikam et al.

1e83).

Tracer techniques also allow the measurement of the mineralisation or hydrolysis of condensed P species including organic P (Harrison 1982) and P from polyphosphate fertilisers (Sutton and Larsen 1964). Furthermore, if the fertiliser is spiked with isotope a recovery measurement can be made. 32OP This chapter outlines the results of experiments designed to test the ability of to equilibrate with PP. Lombi et at. (2004a) and Bertrand et al. (2006) both compared the lability or isotopic exchangeability (E-values) of aÍarLge of fluid and granular P fertiliser products including polyphosphate fertilisers. In both cases lability values were not determined for ammonium poþhosphate (APP) fertiliser due to

32OP 3rPP. uncertainty about the ability of to measure the lability of They suggested

32OP that it is probable that only equilibrates with OP and therefore only measures

the lability of the OP portion of the fertiliser giving erroneous values of potentially

available p from APP. The aim of this work was to determine if these concerns were

warranted.

This chapter also presents the findings of a hydrolysis study where PP spiked with

3tpp was applied at arange of concentrations to soil. Sutton and Larsen (1966) used

3'PP as atracer of PP hydrolysis and recovery. However, in the present study the

t32 33OP, 32pp-spiked fertiliser was combined with the addition of so that changes in the

32PP native OP pool were identified and accounted for in the calculation of hydrolysis.

7.2. Materials and Methods

7.2.1. Soil Gharacteristics

Surface soil samples (0-10cm deptÐ were collected from three agricultural sites in the southern cropping region of Australia. The soils were classified according to 'The

Australian Soil Classification'(IsbeIl 1997), and a range of soil characteristics were measured, and are presented in 4.2.2. The soils used in the experiments described in this chapter are the Grey Calcarosol, the Red Calcarosol and the Sodosol. All soil analyses were conducted after drying soil at 40oC for 3 days. The soils were sieved

(< 2mm) and stored at room temperature.

7.2.2. Reagents

The following analytical grade reagents were used for standards and to supply P as oP or PP: OP was applied as sodium dihydrogen oP (NaH2POÐ (BDH), and PP as

sodium PP decahydrate (Na+PzO7.10H2O) (Sigma). Radiolabelled OP was purchased

32P- as either "P o. "P-labelled H3PO4 while radiolabelled PP was purchased as

labelled sodium PP (Perkin-Elmer).

7.2.3. Experimental Design

Examination of Isotopic Exchangeability of Phosphorus Applied as Pyrophosphate

The Grey Calcarosol (0.5 g) was weighed into 10 mL centrifuge tubes and suspended

in 4 mL distilled deionised water with2 drops of toluene to inhibit microbial activity

133 and shaken for 24 hours. Each soil suspension was spiked with 1 mL of solution containing PP at a rate equivalent to 1000 mg P/þ soil and either 22}lcBq"OP o'

2201

3lPP, 32oP 32PP. and activities of and

Effect of Rate of PP Added on Lability of PP

In this experiment there were five treatments (0, 100, 500, 1000 and 2000 mg P/kg supplied as PP) applied to three soils (Sodosol, Red Calcarosol, and Grey Calcarosol) with three replicates.

Each soil (0.5 g) was suspended in 4 mL of distilled deionised water with 2 drops of toluene to inhibit microbial activity and shaken for 24 hours. All fertiliser solutions

32PPlsample were spiked with isotope (1a0 kBq and 191 kBq "OP/sample) prior to 33OP,t'PP 31PP addition to soil. The soil suspensions \ryere spiked with 1 mL of and as a mixed solution, and shaken fot 24 hours. After shaking fot 24 hours, the

solutions were centrifuged at2500 rpm for 15 minutes, filtered (<0.22 prm) and 3lOP 3lPP, 32OP 32PP. analysed for concentrations of and and activities of and

7.2.4. Analytical Methods

Determination of Soluble Phosphorus and pH

For each sample spiked with radioisotope, a separate duplicate, non-radiolabelled,

sample was prepared to allow measurement of pH and concentration of soluble P-

The unlabelled samples were treated in exactly the same way as labelled samples.

3tP Soluble was measured both in labelled and unlabelled samples. V/ith three

replicates each for labelled and unlabelled treatments, this gave six replicates for

measurement of cold soluble P, and three replicates for measurement of solution pH.

t34 3tP Increased replication of measurements of soluble was used, as the determination of soluble P in solution is often the factor most limiting accuracy of isotopic determinations especially in highly P-fixing soils (Hamon and Mclaughlin2}02)-

Ion chromatography (IC) was used for speciation of the OP and PP species as described in2.2.2. Soil suspension pH was measured in the supernatant solutions following incubation and centrifugation but prior to filtering.

Radioassay

3lP Following analysis for by IC, peak fractions were collected to enable the

33OP 32OP 32PP radioassay of and in the OP fraction and in the PP fraction using a

'I'he Wallac Winspectral o/B counter (Wallac 19S9). analytical methods for the fraction collection and radioassay are described in Appendix A.

7.2.5, Calculations

Relative Sp ecific Activity Calcul ation

The relative specific activity was used to examine the isotopic exchangeability of P

32Op 3'PP. in pp with and This was used as the amount of radioactivity introduced

32OP 3'PP (R), was not exactly the same in each of the and as the solutions were

required to be prepared separately, and the relative specific activity normalises for

this difference. This is similar to the approach used by Bunemann et al' (2004)'

The relative specific activity (RSA) inLlmg, of a given treatment is determined by

the equation: ¡ Eq' 7-1

RSA = =l,,P

Where:

r: the radioactivity remaining in solution (Bq/L),

135 R: the radioactivity introduced (Bq/L), and

3lP: the P concentration in the solution phase (mg/L).

Maximum Potential Labile PP and Actual Labile PP Calculations

The equations for the determination of maximum potential labile PP (mg P/kg soil) and actual labile PP (non-hydrolysed PP) are: 31pp Eq'7-2

: >< lL Maximum Potential Labile PP (mg P/lrg soil) 32,o R"PP "

Actual Labile PP (mg P/kg soil): Eq.7-3 31pp x - (32oPx (l+Kd * JL a* 1R32lr "oe¡¡ m

'Where:

"PP: PP concentration in the soil solution extract (mglL), peaþ (Bq/L), "PP: radioactivity remaining in solution 132PP in the PP

R32PP: radioactivity introduced as labelled PP (Bq/L),

32OP t'OP : activity of 32PP hydrolysed to and measured in the OP peak (Bq/L), and I : the volume of solution (L) to mass of soil (kg). m

Both equations assume that no precipitation of PP occurred after the labelled

fertiliser was applied to the soil. This assumption can be an oversimplification of the

system, and more refined procedures have been developed in the following chapters.

The component of this equation in brackets accounts for the hydrolysis of PP on the 32RPP. solid and solution phase which effectively reduces the value for The Kd of

3tOP ttOP. is used as it is assumed that "OP will partition in the same way as

33OP The partitioning coefficient (Kd) of (L/kg) was determined according to:

136 (33oP]- t ra 33on1LtkÐ_ * F,q.7-4 33oP solution m

Where:

33Op 33OP Kd represents the partitioning of OP added as between solid and solution phase (Llkg), 33oPR: "oP introduced (Bq/L), 33oP solution (Bq/L): "oP remaining in solution after incubation, and I : the volume of solution (L) to mass of soil (kg) m

Hy drolys ß C al cul ations

Equations 7-2 andJ-3 canbe combined to give OP hydrolysed from PP (mg P/kg soil):

OP hydrolysed from PP: maximum potential labile PP - actual labile PP Eq.7-5

The labile OP hydrolysed from PP (mg P/kg soil) can also be calculated by the following equation:

Labile OP hydrolysed from PP: Eq'7-6

31oP-31oP control v x(32 OP(l+ Kd33 oP) x- 32 m OP

'Where labile OP hydrolysed from PP is the amount of isotopically exchangeable OP

derived from PP (mg P/kg soil),

31OP (mg/L)- OP "Op- control: OP concentration in the soil solution extract

concentration in the soil solution extract of the control treatment (no P added)

(mglL),

32PP 32OP in the OP peak (Bq/mL) '2OP : activity of hydrolysed to and measured 33OP Kd 33Op : partitioning of OP added as between solid and solution phase (L/kg),

137 and il : the volume of solution (mL) to mass of soil (g) m

When using this equation the assumption is that all hydrolysed OP is labile, that there is no precipitation and that there is no mobilisation of native OP (that is all OP in solution, once the soluble OP from the control is subtracted from total OP, is derived from hydrolysis of PP). Theoretically equations 7-5 andT-6 should give identical results and any differences will be discussed with the analysis of results.

7.2.6. Statistical AnalYsis

Analysis of variance (ANOVA) was undertaken using the Genstat@ 6 statistical package, including determination of the least significant difference (LSD) between treatments at the 5% level.

7.3. Resulfs and Discussion

7.3.1. Examination of lsotopic Exchangeab¡l¡ty of Phosphorus ¡n Pyrophosphate

3lPP 32OP. There was insignificant exchange of with Comparison of the relative

32OP 32PP specific activity of and in the PP fraction (Table 7-1) shows a minimal

32OP value for in the PP fraction.

specific activity (RSA) in the PP fraction (L/mg)' Table 7-1: Comparison of relative ttop wheie PP was labelled witn or "PP.

PP labelled RSA in PP fraction Standard with Error 0.06 0.01 32PP 8.39 0.03

The findings of this small study suggest that the concerns of Lombi et al. (2004a)

and Bertrand et al. (2006) were warranted. The radiolabel "OP do"t not exchange

138 32OP with 3rpp. Therefore if was used to assess the lability of APP fertiliser, which

31OP 3tPP, contains and an effoneous E-value would be determined.

7.3.2. Effect of Rate of Pyrophosphate Added on soluble Phosphorus Goncentration and Soil Suspension pH

Soluble Phosphorus

The concentration of soluble P in solution increased with dose of PP supplied and the

hydrolysed OP made a significant contribution to the soluble P pool (Figure 7-1). 3tOP, Hydrolysed OP as presented in Figure 7-1 accounts for all hydrolysed since

3top measured in the control treatment was subtracted from the total3lOP measured.

The OP concentration increased with rate of P applied up to 1000 mg P/kg soil, while

the concentrations of soluble PP increased markedly at all rates of PP applied. This

would suggest that PP added at doses above 1000 mg P/kg soil is more stable in

these soils over the 24hotx period and therefore less susceptible to hydrolysis.

139 I hydrolysed OP 1400 r----- PP â'õ Ø 1200 o) -v fL 1 000 o) E (t) f 800 o _c o- c oat, 600 -cÈ o 400 -o) õ f Ø 200

h 0 o o o o o o o o o o o o o o o o o o O o o a o o lr) a o lr) o O ro o a N N E E N o o o o O o o E E o o o o É É o CJ) Ø o o o (, É. É U) C) Treatment

Figure 7-1: Total soluble phosphorus @) concentration (mg/kg) after treatment with four rates (100, SOO, 1OOO and 2000 mg P/kg soil) of PP in three soils: a Grey Calcarosol (GC), a Red Calcarosol @C) and a Sodosol (Sod). PP rate, soil type and treatment interactions were significant (PS0.001). Columns appended by the same letter were not significantty different @50'001)'

The percentage (%) of added PP measured as soluble P increased with increasing

'r.-2). dose in the Calcarosols but not in the Sodosol (Figure An increasing recovery

of soluble PP suggests that these soils were reaching a saturation of surface binding

sites for PP. In all soils the proportion of PP added present as hydrolysed PP (OP)

decreased with increasing rate of PP application (Figure 7-2), confffming the

previous suggestion that PP is more stable in soil when applied at higher

concentrations or that the hydrolysis rate of PP is limited and not directly

proportional to the total PP added.

140 At all rates above 100 mg P/kg soil, soluble P was greater in the Calcarosols than in the Sodosol. This would suggest that the Sodosol has a gteater sorption capacity for

PP than the two Calcarosols, in contrast to the data presented in Chapter 4'

70 I hydrolysed OP t-----l PP a 860Þ ! b50 s U'(Ú40 U' J b å30 oU' -c

t20_o -8f 10

0 o o o o o o o o o o o o o o o o o o o o o rl) o o rO o o rl) o o c{ ôt E C\I o O O o E o E E c) É. É. O o o o o o É. U) U) o o o (, É. CJ) Ø Treatment x'igure 7-2zP in solution (PP and PP hydrolysed to oP) as a proportion of PP added (%) after treatment with four rates (100, 500, 1000 and 2000 mg P/kg soil) of PP in three soils: a Grey Calcarosol (Gc), a Red calcarosol (RC) and a Sodosol (Sod). PP rate by soil type interactions were significant @ S 0.001). < Columns appended by the same letter are not significantly different @ 0.001).

Soil Suspension pH

The pH data (Table 7 -2) were consistent with previous experiments (e.g. Chapter 4)

where solution pH increased with increasing supply of PP and was greatest in the

Grey Calcarosol. Pyrophosphate addition is expected to raise soil solution PH,

particularly in soils containing appreciable quantities of Ca (Lindsay 1979)'

r4t Table 7-2: Soil suspension pH data in a Grey Calcarosol, Red Calcarosol and Sodosol at 5 rates of PP applied (0, 100,500, 1000 and 2000 mg P/kg soil)'

PP rate Grey Calcarosol Red Calcarosol Sodosol soil 0 9.00 9.04 8.19 100 9.r2 9.03 8.18 500 9.94 9.71 8.58 1000 r0.37 t0.17 8.82 2000 10.63 t0.26 8.92

7.3.3. Effect of Rate of PP Added on Lability and Hydrolysis of PP

Lability

The isotopic lability measurements were log transformed to normalize the distribution of variance across the wide range of PP concentrations added. Maximum potential labile PP was not 100% of that added for most treatments except the higher dose treatments of 1000 and 2000 mg P/kg soil in the Sodosol (Figure 7-3). This small difference between the maximum potential labile PP and the nominal application rate is most likely due to analytical error and./or biological variability.

There is a missing value for the Grey Calcarosol at 1000 mg P/kg soil added, due to a

counting error in the PP chromatography peak'

In all cases the maximum potential labile PP was greater than the actual labile PP,

indicating that there was some hydrolysis of PP to OP over the 24 hours incubation

period, despite the inhibition of microbial activity. However, toluene does not

provide a complete inhibition of microbial activity as it has been shown to act as a

source of carbon for some micro-organisms, and therefore is not suitable as a

microbial inhibitor of extracellular enzymes such as phosphatases (Kaplan and

Hartenstein 1979).

r42 4.0 I Actuallabile PP PP -'õ tT l Maximum potential labile PP U) <') 3.5 t¿ 2000 fL a <') 1000 E 3.0 o) '-c 500 '(Ú 2. 5 E Lo E o 100 o 2 .0 o o fL 5 o)o

1.0 o o o o o o o o o o o o o o o o o o o o o o rf) o o r.() o o rr) o o ôt C\¡ E ! N o o o o o o E o o (, o o É. É. o O Ø Ø o o (, o É. É U) U) Treatment

Figure 7-3: Maximum potential PP recovered and actual PP recovered Qog mg/kg) after treatment with four rates (100, 500, 1000 and 2000 mg P/kg soil) of PP in three soils: a Grey calcarosol (GC), a Red calcarosol (RC) and a sodosol (Sod). Dotted Hnes represent nominal addition rates of P. For maximum potential PP recovered in a given soil the columns appended by the same letter are not significantly different @ S 0.001). tr'or actual labile PP in a given soil, columns appended by the same letter are not significantly different (P S 0.001).

Hydrolysß

Data for pp hydrolysis to OP were log transformed to normalize the distribution of

variance. The concentrations of added PP hydrolysed to OP were calculated in two

ways, firstþ using F;q7-5. (hyd PP calc) and secondly by the difference between

maximum potential labile PP and actual labile PP remaining (hyd PP diff) as

determined in the previous section (Figure 7-3).The size of the labile hydrolysed PP

pool was dependent on rate of PP added up to 1000 mg P/kg, but thereafter rate of PP

t43 added had no effect on the concentration of PP hydrolysed (Figure 7-4), which confirms the findings of the soluble P data (Figure 7-I andT-2).

3.0 I Hyd PP calc r::-l Hyd PP diff -'õ 2.5 c U, (') t¿ ò 2.0 d o) c E c È 1.s ! (to

1.0 E T o) o 0.5

o o o o o o o o o o o o o o o o o o o o r.() o o ro o Õ ro o o (\¡ N € ! c! o o () o () o o E ! o (, o o É. É o U) U) o o (, o É. É. U) U) Treatment

Figure 7-4: Hydrolysed PP (log mg/kg) measured by calculating Eq.7-5 (Hyd PP calc) - and by the difference between PP recovered (mg/kg) and PP remai{ng (me/kg) (Hyd PP diff) after treatment with four rates (100, 500' 1000 and zooo rng P/kg soil) of PP in three soils: a Grey Calcarosol (Gc), a Red Calcarosol (nC) an¿ a Sodosol (Sod)..In a given soil the columns appended by the same letter are not significantly different (P S 0.001)'

The resulting datafrom the two calculation methods were plotted against each other

(Figure 7-5) giving a linear regression with an R2 of 0.83 and a slope of 1.0

suggesting that the two methods provided similar results. The equation of the line

had a negative y-intercept suggesting that in most cases the hydrolysed PP

calculation gave a slightly higher value than the estimate of hydrolysed PP by

difference.

The hydrolysed PP calculated by difference had a greater variability than the

hydrolysed PP calculated by Eq. 7-5. This greater variability in the hydrolysed PP by

t44 difference may be due to the greater number of independent measurements required to estimate the value, increasing the likelihood of variability in the data.

It should be noted that Eq 7-5 assumes that all OP hydrolysed from PP remains in the labile pool and there is no precipitation of hydrolysed PP or mobilisation of native

OP, but at this point these assumptions have not been tested'

a 3 ¿oo a o v a fL Co I goo oc) c a oE E aa ! 200 a o IL a aa fL E E roo ->o ! I e 0 0 100 200 300 400 Hydrolysed PP calculation (mg P/kg soil)

Figure 7-5: Plot of llydrolysed PP Qog mgikg) measured by calculating Eq' 7-5 (IIyd PP calc) and by the difference between PP recovered (mgikg) and PP remaining (møkg) (Hyd PP diff). Equation of the fitted line 5 1.02x - 33.64,R1=0.83. 7.4. Conclusions

The use of isotopes for measurement of the reactions of poþhosphate fertilisers in 3'OP soil requires the labelling of both the OP and PP components of the fertiliser, as

does not equilibrate with labile PP'

32pp 3'OP The use of a tracer,in conjunction with for assessment of the native P

31PP pool, makes it possible to evaluate the lability of PP and the hydrolysis of to

(relative to PP added) "Op. As the rate of PP applied increased, the lability of PP

increased, and the percentage of PP added which hydrolysed to oP decreased. A

relatively high proportion of the PP added remained in the labile pool, either as labile

PP or as hydrolysed OP.

t45 Chapter 8. A Study of Lability of P in Soils Treated with Orthophos hate and phos hate

8.1. Introduction

32OP Traditionally the single labelling of OP with for determination of potential availability of P in soils and plants in the form of E and L values has provided sufficient information on P fertiliser: soil reactions. The difficulty with a polyphosphate fertiliser is the fact that P exists as several species in the product, and

(less the more complex P species hydrolyse in certain conditions to form more simple phosphate groups) P species. The dominant P species in a polyphosphate fertiliser are

Op and PP. Therefore it is necessary to develop a method capable of measuring the potentially available P as OP, PP, and the amount of OP derived from the hydrolysis

ofPP

A double labelling technique involves using two isotopes of the same element to label

two chemical species of that given element. Leikam et al. (1983) used a dual labelling

32PP system in a tracer experiment where PP was labelled with and OP was labelled

33Op with in an APP fertiliser solution. This solution was then equilibrated with soil

and plants were grown. The harvested plants were digested for counting, and the

contribution of PP and OP from the APP solution to total plant P was determined.

However, using this technique it was only possible to assess the contribution of OP

and pp to plant nutrition, but neither the lability of OP and PP nor the hydrolysis of

PP prior to plant uptake could be determined.

Further examples of dual P labelling include the work of Mclaughlin et al. (1986;

33OP 32OP 19SS) where organic and inorganic P sources \ryere labelled with and

respectively to trace P uptake in a wheat crop, and more recently'Waser and Bacon

32OP 33OP' (1995) measured soil deposition of cosmogenic and

r46 All of the above examples of dual labelling experiments were with P isotopes used as tracers. Chapter 7 demonstrated that a double labelling system could be used as a tracer to measure the recovery and hydrolysis of PP added to soil. In this chapter the

3'PPI33OP double labelling technique was extended to a carrier-free procedure where the radioactive isotopes were added at the end of an equilibration period in order to compare the lability or isotopic exchangeability of different P species from polyphosphate fertilisers (containing mixtures of OP and PP). Measurement of P lability using a double labelling system is potentially a more accurate means for predicting the contribution of PP to P nutrition, as it is possible to measure PP lability, hydrolysed PP lability and native OP lability simultaneously. It should be emphasised that this work was conducted using analytical reagents of OP and PP and the purpose

was not to compare commercial fertiliser products but rather to develop an

understanding of the chemical reactions and processes of the P species supplied in

APP fertiliser.

Furthermore, the effects of OP and PP addition on soil parameters including DOC, Fe,

Al and Ca content were assessed and compared with the solution and labile P pools.

8.2. Materials and Methods

8.2.1. Soil Characteristics

Surface soil samples (0-10cm depth) were collected from three agricultural sites in the

southern cropping region of Australia. The soils were classified according to 'The

Australian Soil Classification' (Isbell 1997), and a range of soil characteristics were

measured, and are presented in 4.2.2. The soils used in the experiments described in

this chapter are the Grey Calcarosol, the Red Calcarosol and the Sodosol. All soil

t47 analyses were conducted after dryittg soil at 40oC for 3 days. The soils were sieved (<

2mm) and stored at room temperature.

8.2.2. Reagents

The following analytical grade reagents were used for standards and to supply P as oP or PP: OP was applied as sodium dihydrogen oP (NaH2PO4) @DH), and PP as sodium PP decahydrate (Na+PzO7.10H2O) (Sigma). Radiolabelled OP was purchased

33p-labelled 3'P-1abe11ed as HlpO+ while radiolabelled PP was purchased as sodium

PP (Perkin-Elmer).

8.2.3. Experimental Design

Incubation Effects on Lability of Orthophosphate Compared to Pyrophosphate

The incubation study had three time treatments (1 day' 3 days and 7 days), 3 soil treatments (a Sodosol, a Red Calcarosol , and a Grey Calcarosol) and 3 P treatments

(control of no P, OP and PP at 1000 mg P/

replicates for each treatment.

The fertilisers were made up to an appropriate concentration in order to deliver 1000

mg P/kg for the 3 and7 day treatments, and to wet soils to 80% maximum water

holding capacity (MV/HC). The soils were spiked with fertiliser and incubated for 3

or 7 days atS1YoMWHC. After each incubation period the samples were made up to

a soil:water ratio of 1:10 (0.5 g soil:5 mL), two drops of toluene were added and the

suspensions were shaken end-over-end for 24 hours. These treatments were then

32PP 33 spiked with (240 r

32PP 33OP, 3lP The 1 day treatment was spiked with similar activities of and and was

added at the same time to the suspensions to make the soil concentration equivalent to

1000 mg P/kg. All samples were shaken for a further 24 hours. After shaking for 24

148 hours, the solutions \¡/ere centrifuged at2500 rpm for 15 minutes, filtered (<0.22pm) 3lOP 3lPP, 32OP 32PP. and analysed for concentrations of and and activities of and

8.2.4. Analytical Methods

Determination of Soluble Phosphorus, Cations, Dissolved Organic Carbon and pH

For each sample spiked with radioisotope, a separate duplicate, but not radio-labelled, sample rwas prepared to allow measurement of pH, and concentrations of DOC, 3lP cations, and total P. This gave six replicates for measurement of soluble and three replicates for measurement of pH, and concentrations of DOC, cations, and P'

Ion chromatography was used for speciation of the OP and PP species as described in

2.2.2. Concentrations of total P and cations were determined by ICP-AES under the

operating conditions outlined by Zatcinas et al. (1996). Soil solution pH was

measured in the supernatant solutions following incubation and centrifugation but

prior to filtering. The concentrations of DOC were measured in the centrifuged and

filtered supernatant solutions using a Formacs@ TOC/TN Analyser (2000)'

Radioassay

3lP Following analysis for by IC, peak fractions were collected to enable the

33Op 32OP 32PP radioassay of and in the OP fraction and in the PP fraction. The

analytical methods for the fraction collection and radioassay are described in

Appendix A.

8.2.5. Galculations

Lability Calculations

Labile OP (mg P/kg soil) was determined according to the conventional isotopic

dilution calculation:

t49 31oP Eq.8-1 Labile oP : * R"op * I 33oP m

Where Labile OP is the amount of isotopically exchangeable OP (mg P/kg soil),

"OP: OP concentration in the soil extract (mglL),

"OP: radioactivity remaining in solution (Bq/L),

R33OP: radioactivity introduced (Bq/L), and I : volume of solution (L) to mass of soil (g). m

The equation for the determination of labile PP (mg Plkg soil) was:

3lpp Eq.8-2 I Labile PP: 32pp x 1R32re - ("op (l+Kd "oe¡¡ *

Where labile PP is the amount of isotopically exchangeable PP (mg P/kg soil),

"PP: PP concentration in the soil solution extract (mg P/L)' peaþ, "PP: radioactivity remaining in solution (Bq/L) 132PP in the PP

R3tPP: radioactivity introduced as labelled PP (Bq/L)'

32PP 32OP peak (Bq/L), "OP : activity of hydrolysed to and measured in the OP

and I : volume of solution (L) to mass of soil (g). m

32PP The component of this equation in brackets accounts for the hydrolysis of to

32OP t'Op,and the subsequent redistribution of between solution and solid phases. 3'PP This is required to determine an accurate measure of R for atthe end of the

isotopic equilibration period.

33OP The partitioning coefficient (Kd) for was determined according to:

8-3 JJ 33 Eq. ( OPR - OP solution) v Kd 33 OP: JJ m OP solution

1s0 Where:

Kd 33Op represents the partitioning of OP added as "OP between solid and solution phases (Llke), t'oP '3oPR: introduced (Bq/L), 33oP solution (Bq/L): "oP remaining in solution after incubation, and I : volume of solution (L) to mass of soil (kg). m

33OP 32OP The Kd of was used, as it was assumed that would partition identically to

33oP.

Hydrolys is Calculation

The labile OP hydrolysed from PP was calculated as:

31oP Eq'!Y' 8-4 '?) ^-" -- '33 ^-" v ' - * 132or1t + rd33oP) xlL Labile oP hydrolysed from PP 3%rp

Where E OP hydrolysed from PP (mg P/kg soil) is the amount of isotopically

exchangeable OP derived from PP,

"OP: OP concentration in the soil solution extract (mdl), 32OP 3'Op : activity of 32PP hydrolysed to and measured in the OP peak (Bq/L),

and il : volume of solution (L) to mass of soil (kg). m

33OP 33OP The term Kd (L/kg) represents the partitioning coefficient for between'

solid and solution phase (Eq 1-3).

P art itioning C al cul at ion 3tet 3tPP The partitioning coefficient for gcd itt Llkg) is calculated using the

following equation:

Eq.8-5 32 32 _(32 33 32 R PP_ PP OP x (1+ Kd oP) Kd PP (Llkets 32 PP

151 Where: 3,pp: radioactivity remaining in solution (Bq/L) 132PP in the PP peak),

R"PP: radioactivity introduced as labelled PP (Bq/L),

32OP 3top : activity of 32pP hydrolysed to and measured in the OP peak (Bq/L) and

33OP Kd33Op: partitioning coefficient for between solid and solution phase (Eq 1-3)

8.2.6. Statistical AnalYsis

Analysis of variance (ANOVA) was undertaken using the Genstat@ 6 statistical package, including determination of the least significant difference (LSD) between treatments at the S%olevel.

8.3. Resulfs and Díscussion

8.3.1. lncubation Effects on Soluble Phosphorus, pH' Gations and Organic Carbon

Soluble Phosphorus

There was a significant increase in the percentage of soluble PP recovered (the sum of

3lPP solution phase and "Op) from 1 to 3 days, but no significant difference between

3 and 7 daysin the percentage of added PP recovered in the 3 soils (Figure 8-1). It

was evident that the PP added was continually hydrolysing as the OP from PP as a

proportion of PP added increased over time in all 3 soils'

The percentage of added OP recovered as soluble OP, while initially significantly

greater than the soluble P added as PP, declined over time. After 1 week there was a

significantly greater percentage of P added recovered in solution when PP was added

in all 3 soils as compared to when OP was added (Figure 8-1). Most studies

investigating the potential of PP as a fertiliser suggest that it is equivalent to OP as a P

source (Gilliam 1970;Miner and Kamprathlg7l; Khasawneh et al. 1974; Khasawneh

ts2 et al. 1979; Sample et al. 1979; Parent et al. 1985a). The results of this study would suggest that PP is an equivalent or better source of P than OP for supplying P to the solution pool over time when tested in some highly P-fixing soils (refer to Chapter 5 for P sorption capacity).

The capacity of added PP to supply solution P over time is supported by the findings of Chapters 4 and 5. In Chapter 4 asolid-state "P NMR study showed that PP was capable of hydrolysing in the non-extractable solid-phase while Chapter 5 showed that the Australian soil types used in this study have a considerably higher P sorption capacity for PP than for OP. Therefore this resupply of solution phase P is most likely coming from adsorbed PP which is hydrolysing to resupply the solution phase'

Sutton and Larsen (1964) concluded that the amount of P available to crops when supplied as PP was essentially controlled by the partitioning of PP onto the solid phase prior to hydrolysis. Any sorbed PP was assumed to be non-available, and any

solution phase PP was presumed to rapidly hydrolyse to OP. However, as the solution

p as a percentage of P added increased over time, predominantly as OP, this would

suggest that the sorbed PP is able to hydrolyse and contribute OP to the soil solution

phase over time.

Comparison of equivalent treatments between Chapter 7 andChapter 8 (1000 mg P/kg

soil applied as PP for time 1 day treatment) showed that the recovery of soluble P was

similar for all3 soils.

1s3 rl as % of Added ö 8.8 ä g 3 oa-. Soluble Phosphorus ?qqa¡.=*, L æ ã:.=.ä.4c.) o oooNÞA o (94ÈÈ ã Èoaã:Iv1.% R ERõÐi5S õ.*€ r "E GC 1D OP 8'Èiqi?,.F.Ðvu #.É.FÞ tr e.Ét¡.ê Ë 3 ã I ^rjp.(DÞS ä Ei l ss ? GC 1D PP O!-Il o qãds-!É\:ì.Èqc\ ! 1' å 9 ä I á¡ GC 3D OP õ ä ? É ¿ãä A gE q 3 õâalã o.j'9

The highly calcareous soils used in this study have a high buffering against any acidifying effects of OP addition (Lindsay 1979), and the Sodosol has previously showed a high buffering capacity against pH change with increasing additions of P

(Chapter 5).

Table 8-1: Change with time in equilibrium soil suspension pH for the Grey Calcarosol, the Red Calcarosol uttã th" Sodosol. The P species by incubation time by soil fype interaction was significant (Fpr s 0.001, LSD:0.23) Treatments appended by the same letter were not significantly different.

Orthophosphate Pyrophosphate Soil Time pH LSD PH LSD

Grey 1 7.86 J 8.78 cd Calcarosol J 8.36 fg 9.33 a 7 8.25 gh 9.14 ab Red I 8.46 efg 8.66 de Calcarosol 3 7.85 j 8.92 bc 7 7.85 j 8.58 def Sodosol 1 7.32 k 6.90 I a J 7.3r k 8.10 hi 7 7.41 k 8.01

Cation Concentrations

The pH of a soil determines the dominant ions (including Ca, Al and Fe) available for

sorption by influencing the extent of ionisation of phosphates and the oxide surface

charge (Dubus and Becquer 2001). Generally the solubility of Ca phosphates

decreases with increasing pH while the solubility of Fe and Al phosphates increases

with increasing pH (Lindsay Ig79).In a study of Lindsay et al. (1962) precipitates of

CaNH¿pzO t.HzO formed when PP was added to a calcareous soil. It would therefore

be expected that the PP added in the calcareous soils would result in a decrease in Ca

in solution (Lindsay L979;Mclaughlin et al. 2003), which in this study occurred over

time in all 3 soils (Figure 8-24). However the sorption curves presented in Chapter 5

did not suggest precipitation, but rather saturation of sorption sites (section 5.3 -2.)-

155 çs I I4 H 3 3 F å;; F ä [ãóí sF äå 3Ë q ã' Calcium in solution (mg/L) "' ('r o) o oo ttåB [-ãFä o B8à 3.fã 3 á3.ã E GCIDC GC.ID OP =ÊeãZgsi'?Ë' llo öc) GC ID PP lt! a ù g GC3DC õ : E ã ä Ê E GC 3D OP Ei ä i gs F& GC 3D PP õ.99"õÊgrDc, GCTDC )¡3ÞO5'-'O GC 7D OP E € 6 Z Eñ A Ë GC 7D PP âig å RCIDC ÊgfipFqgHos(DR s RC.ID OP ä -{ RC.ID PP î i d ã ñ z Ë õ RCSDC o j q 3 RC 3D OP Hfr äå i 3 --Jõ;Þã93'i o RC 3D PP r/.ä'X F = RCTDC Ë. I ã g RC 7D OP q RC 7D PP 1E ? ä I i Sod lD C 8'E ä + s 3 ? Sod ID OP ÊÈg.o< Sod lD PP Sod 3D C g : Sod 3D OP äs € l* Sod 3D PP á:3'gFSÈ Sod 7D c HB'Ñiàu) Sod 7D OP õ.:.9)aEê,'t:ûaãA. Sod 7D PP g FSqf Ë'E Ë ã'g 3 Fai.fBõ+[s ã *FËåE L¡r I F +.f o\ .F L ã ? ì:, in solution (mg/L) Aluminium in solution (mg/L) lron N oo 99 o *EË Ë09 o o Ào) @ o b (j o ol b bñ GCIDC Ë?F,ãíË GCIDC GC 1D OP Wø gÊs+ GC.ID OP !o Ëã u-D o GC.ID PP T9q GC ID PP f GC3DC o GC3DC o GC 3D OP GC 3D OP ËãåËiF(e ê:-ä E'-. GC 3D PP ã 9.Xi7- ll GC 3D PP GCTDC GCTDC GC 7D OP GC 7D OP GC 7D PP GC 7D PP RC tDC s:E+rÊË 3.0 ^ a tr RCIDC RC ID OP D PP =¡9.>v,+El' RC 1D OP -.1 RC.f E E € å. o- RC3DC i.e r RC 1D PP o RC 3D OP RC3DC 3 o RC 30 PP RC 3D OP ) ËHËsäiä RCTDC RC 3D PP RC 7D OP RCTDC g'Eoägå- RC 7D PP RC 7D OP Sod lD C RC 7D PP Sod ID OP ä'å iÊ A !. Sod'lD C Sod 1D PP Sod 'l D OP Sod 3D C lFg Þä Ë Sod 1D PP Sod 3D OP Sod 3D C Sod 3D PP ã e 3 ã Éå Sod 3D OP Sod 7D C Sod 3D PP Sod 7D oP - sod 7D PP ã;.çÞ: Sod 7D C o @ É sEã3 äåäËE a \ìL¡r €'äç'(D Dissolved Organic Carbon

The concentrations of DOC in solution were generally unaffected by the addition of

OP, but markedly increased by the addition of PP (Figure 8-3.). Pyrophosphate addition to soil has the capacity to solubilise organic carbon in soil (Mnkeni and

MacKenzie 1985), and is routinely used in soil testing laboratories for the extraction of organic matter from soil (Bremner and Lees 1949; Dijkstra and Fitzhugh2Û03;

Ussiri and Johnson 2004). A possible mechanism for this solubilisation of DOC is due to the ability of PP to sequester Al and F'e in solution, causing an increase in pH and a consequent decrease in Ca in solution. A decrease in Ca in solution increases the solubility of DOC and result in a subsequent increase in DOC concentration in

solution (Romkens and Dolfing 1998).

Solubilisation of DOC at high soil solution pH can provide competition with P for

binding sites due to the negative charge on the DOC (Dubus and Becquer 2001;

Guppy et at.2005). This may explain why there was more soluble P when PP was

added as compared to OP, as PP addition resulted in an increase in DOC in solution.

However, it has been recommended by Guppy et al. (2005) that further work is

required on the mechanisms of sorption competition between P and DOC to ensure

that any P that is released from solubilised DOC is considered. In the case of these

experiments it is difficult to evaluate the contribution of P from the solubilised DOC.

This is because the agent solubilising DOC contains is PP, which hydrolyses to OP,

and it is not possible to distinguish between PP hydrolysed to OP and solubilised P in

this experimental design.

158 o- ø'$ÈF d1 Dissolved Orgalic Carbon (mgt) o90 oa-. N R E ã ã$ A 9.t (¡o('l o l'¡ (!Fa o öoo o ã æ ä'qg!'.Y tt- I -l€=s¡- !t' o9DÊDãa\-Fu'r=\ oo I^=EÈÔ GClDC (t) Þ FF tr -r- o= !r ct Pã iil GC 1D OP e?q äS 5Ð E 3r.ã GC 1D PP i9s,-J =(rÉ.o-È 0¡ = o + 8 o;9 $ =1 Er e'u GC3DC o ;i-TOvo¡e% d9 ' õ Å{ ìÁ=)Âù-!. À-_ á GC 3D OP Ë 3 å €F 5'n g' GC 3D PP SR\¡a¡È. sHr^9 -- 5(D o GCTDC q S E S tto ='ú 5'tr

The problem with the separation was most likely a lack of separation between OP and pp or a problem with the elution time resulting in effors in the timing of the fraction collection.

The total labile P pool where PP was added was significantly less than the labile P pool of Op on the day the P was added, but by the end of the 7 day incubation period there was no significant difference between the labile P pools between fertiliser fonns

(Figure 8-4). For both OP and PP treatments, labile P as a proportion of P added

P decreased with time. In the PP treatments, a significant portion of the total labile pool remained present as PP.

160 1 OQ Ë-E ä L Phosphorus E-value (% of P added) Ë s i * ã NÞo)eo N F^ 3 10 o õoooo o þ É Ë ä Ì È g Ë Èq rE ä gäF E'õ'.?iS3ä *9Ë 1D OP È ìi É GC € ì ò I É rÁ c E= u,¡'<='.Þ.d E Fi A ã'* ô Uí GC ID PP rË.\oEAeoSã' Ìäã'ËÉi g Ë ã aäf GC 3D OP Ë 3 F äÞ Z E Éç'ã S .ãã E v ã É * Þ"B.ozg GC 3D PP ? Ë r Ei ã +'= I9Iâ'c GC 7D OP '<+ ¡+.Oc.)P É äåË Ë + å 5 (D(ÞÞFeoH9ìv -v GC 7D PP ÞÅajè;+S.é ..AJãØ-.--- ÐXêfoT Ë F RC 1D OP à(D+oEu(D Ë ö=.ê.¡D 6 lI -Êx 'l,r^i RC 1D PP r.? 8I < ãã:Vã -.{ hO.=vOÊl.5 g - * Ë ts Ê o RC 3D OP g g A # t i'E 0) 3 íÉEei o RC 3D PP äa{FEäãiERgõ' qg. J 9. o'Ð 'É áE'sã? RC 7D OP s{\Hîq¿FFFs : Ë3;q RC 7D PP ÊË 3 äl' F * È * A ä $ f -Ep ã Sod 'lD OP (D'j v út)^ gâË EEEÊË Sod 1D PP È + ã H ãrneZÈ Ë âä 1 ^ 3 gq9 Sod 3D OP zll 5*är ç ãã'e Sod 3D PP o!o ã +5 ã.C Ëi !-u! F ø 3 F€ F ã'. ìlì- Fr Sod 7D OP ã' 2 Ë E Ë Ë å A Þ 4io-;8 3 =. tl(Þ9. c7-t! SodTD PP 1' åeÞ+ãq8 'Y o\ Ø Ìl presented in this chapter also found no significant difference in the total labile P pool where P was supplied as PP as compared to OP. previous studies often concluded that the reason that PP was equivalent to OP as a P source for crops is because the biological and chemical environment in soil caused PP to rapidly hydrolyse to OP. This meant that all P, whether added as PP or OP, essentially behaved as OP after avery short length of time (Khasawneh et al. 1979).

The results presented in this chapter do not support this hypothesis, as the lability of non-hydrolysed PP made a significant contribution to the total labile P pool in soils treated with pP. The distinguishing feature of this study is that it had the capability to chemically assess the importance of the contribution of both hydrolysed and non- hydrolysed PP to the total labile P pool. Even after 7 days of incubation, there was a

considerable proportion of P added as labile non-hydrolysed PP contributing to the

total labile P pool in the PP treatment.

The determination of an E-value involves the contribution of a number of independent

measurements. The E-value for OP requires 3 independent measurements: solution

33OP the E- "Op, solution and the activity of isotope added ß"p). In comparison,

value for PP requires 6 independent measurements and the E-value for OP derived

from pp requires 4 independent measurements (See Eqn 8-2 and 8-4). Accounting for

the independent measurements used in both E-PP and E-OP from PP, the total labile P

pool from PP addition involves 7 independent measurements. While considerable

effort was made to ensure that the analytical techniques used were both accurate and

precise, the more independent measurements required for a calculation, the greater the

error in the final value due to propagation from the error involved in each individual

measurement.

162 Orthopho sphate and Pyrophosphate P artiti oning 3'PP The Kd calculation for PP uses Eq 8-5 to subtract hydrolysed from the equation and the Kd derived is for PP only (Table 8-2). The advantage of determining the Kd in this way is that the hydrolysed PP was accurately subtracted from the total solution

OP and therefore solution OP is not overestimated by including native OP'

For all treatments the Kd for PP was significantly greater than that for OP (Table 8-2)'

These findings support those of Chapter 5 and several previous studies (Blanchar and

Hossner l969a;Hashimoto et al. 1969; Mnkeni and MacKenzie 1985; Al-Kanani and

MacKenzie lggl),which suggest that soils have a gleater sorption affinity for PP than

32PP 33OP for Op. The change in Kd for and was not consistent over time in any of the

3 soils (Table 8-2).

33OP Table 8-2: Change with time i4_partitioning coefficient (Kd) (L/kg) of for_the OP t'ÞP treatment and for for the PP treatment in the three soils. The P species by incubation time by soil type interaction was significant @pr S-. 0-.001). Treatments appended by the same letter were not significantly different.

Orthophosphate Pyrophosphate 33oP 32PP Soil Time Kd Kd

Grey 1 5.5 ef 20.3 cd Calcarosol 3 7.0 ef na 7 9.0 ef 77.r a Red 1 5.9 ef 28.5 fe Calcarosol 3 2.5 f 15.2 de 7 4.r f t4.5 de Sodosol 1 2.6 f 42.7 b J 5.0 ef 67.7 a 7 na rt.4 def *na: data not available 8.4. Conclusíons

The data from this incubation study suggests that while the labile P pool of soil

treated with PP was initially smaller than that of soil treated with OP, after 7 days

and there was no significant difference. The lability of P from PP increased between I

3 days suggesting the possibility of a slow release mechanism, which is potentially

r63 desorption of added PP or mobilisation of native OP. The potential for PP to mobilise native P into the labile pool willbe investigated in Chapter 9.

The soluble P data suggests that there was significantly greater recovery of the P added as PP after I week. Part of the discrepancy between the soluble P data and the lability study is related to the propagation of error due to the complexity of the double labelling technique requiring a series of independent measurements contributing to the final calculation. This propagation of error resulted in a LSD that was a considerably larger proportion of the mean than for the comparable soluble P data. However, the double labelling technique is an improved methodology for the determination of the labile p pool and it's components due to the addition of PP, as it has the capacity to account for PP, hydrolysed PP and native oP in one treatment.

The addition of PP to soil resulted in the complexation of Fe and Al in soil solution, increasing the concentration of these metals in solution. Furthermore the addition of

pp resulted in a decrease in Ca in solution due to the strong binding capacity of Ca for

pp. Dissolved organic carbon was released into solution with the addition of PP and

this release is related to the release of Fe and Al in solution, and suppression of Ca

concentration in solution.

r64 Ghapter 9. Mobilisation of Native Phosphorus by phos hate

9.1. Introduction

The pp molecule has the capability to bind and complex several cations in soil involved in complexation/precipitation of OP including 41, Ca and Fe as discussed in

Chapter 8. If the complexin g capacity of PP for these elements is sufficient, it is possible that sparingly soluble solid-phase P compounds containing these elements

could be solubilised, and non-labile P could enter the labile P pool. A test calculation

of the capability of PP to solubilise DCPD was performed using GEOCHEM-PC'

Increasing concentrations of PP in solution caused DCPD to dissolve, due to

complexation of Ca by the PP molecule reducing C** activity in solution, thus

promoting dissolution of the solid phase DCPD. However, it should be noted that the

GEOCHEM model does not account for precipitation of Ca-PP solid phases, nor does

it take into account the potential sorption reactions of OP and PP with the soil

(Mclaughlin et a\.2003; Hedley and Mclaughlin 2005).

The aim of the study presented in this chapter was to establish if the addition of PP to

soil had the effect of mobilising native OP. A double isotopic labelling system was 32PP, used where PP hydrotysis was measured using atracer system (carrier) with and

the lability or isotopic exchangeability of native OP was measured in a carrier free

33oP. system with

165 9.2. Materials and Methods

9.2.1. Soil Characteristics

the Surface soil samples (0-10cm depth) were collected from three agricultural sites in 'The southern cropping region of Australia. The soils were classified according to

Australian Soil Classification'(Isbell 1997), and a range of soil characteristics were measured, and are presented in 4.2.2. The soils used in the experiments described in this chapter are the Grey Calcarosol, the Ferrosol and the Sodosol.

All soil analyses were conducted after drying soil at 40oc for 3 days. The soils were sieved (< 2mm) and stored at room temperature.

9.2.2. Reagents

The following analytical grade reagents were used for standards and to supply P as

oP or PP: Orthophosphate was applied as sodium dihydrogen oP (NaH2PO4) @DH),

was and PP as sodium PP decahydrate (Na+PzO?.10H2O) (Sigma). Radiolabelled OP

33p-labelled 3tP-labelled purchased as H¡pO+ while radiolabelled PP was purchased as

sodium PP (Perkin-Elmer).

9.2.3. Experimental Design

Examination of Hydrolysis of Pyt'ophosphate and Mobilisation of Native OP in Soil

3 soils The hydrolysis and mobilisation study had2 P treatments (control and PP) and

(Grey Calcarosol, Ferrosol and Sodosol) with three replicates.

and The PP treatment was made up to deliver a rate of P equivalent to 800 mg Plkg

32pp spiked with (183 Bq/g) before addition to the soil. The PP solution was delivered

and in a volume that wet soils Io S}Yomaximum water holding capacity (MWHC)

then shaken vigorously to ensure even distribution of the PP solution. The soils were

t66 incubated for 7 days at8}%oMWHC. After incubation, the samples were made up to a soil:water ratio of 1:10 (0.5g soil: 4.95 mL), two drops of toluene were added and the 33OP suspensions \Mere shaken for 24hours. The samples were then spiked with 1280

KBq/sample). The final volume of all solutions was 5 mL. The samples were shaken for a further 24 hours. After shaking for 24 hours the solutions were centrifuged at 3lOP 2500 rpm for 15 minutes, filtered (< 0.22 ¡rm) and analysed for concentrations of

31PP, 32oP 32PP. and and activities of and

9.2.4. Analytical Methods

Concentration of Solution Phosphorus and Soil Suspension pH'

Ion chromatography (IC) was used for speciation of the OP and PP species in solution

as describe din2.2.2. Soil solution pH was measured in the supernatant solutions

following incubation and centrifugation but prior to filtering.

Radioassay

3lP Following analysis for by IC, peak fractions were collected to enable the

33Op 32OP 32PP radioassay of and in the OP fraction and in the PP fraction. The

analytical methods for the fraction collection and radioassay are described in

Appendix A.

9.2.5. Galculations

Maximum Potential Labile PP and Actual Labile PP Calculations

The equations for the determination of maximum potential labile PP (mg P/kg soil)

and actual labile PP (non-hydrolysed PP) (mg P/kg soil) are: 31pp Eq' 9-1 fL Maximum potential labile PP : -* t R3'PP *

t67 31pp Eq.9-2 Actual labile tt: x (f2pp - (32oPt (l+Kd * I ¡z----YY "or¡¡; m

'Where:

"PP: PP concentration in the soil solution extract (mg/L), 32pp: radioactivity remaining in solution 132PP in the PP peak) (Bq/L),

R3'PP: radioactivity introduced as labelled PP (Bq/L)'

32OP t'OP : activity of 32PP hydrolysed to and measured in the OP peak (Bq/L), and I : volume of solution (L) to mass of soil (kg). m

Both equations assume that no precipitation of PP occurred after the labelled fertiliser was applied to the soil. The component of equation 9-2 inbrackets accounts for the hydrolysis component of PP on the solid and solution phase which reduces the value

for R32Pp. Therefore, the actual labile PP is derived from the maximum potential

labile PP when hydrolysis of PP during theT day incubation was taken into account.

33OP 32OP The Kd of is used as it is assumed that will partition in the same way'

33Op t3OP The term Kd (L/kg) represents the partitioning coefficient for OP added as

between solid and solution phase (Eq 9-3).

Eq.9-3 33 33 OPR - OP:) v Kd JJ OP= 33 OP m

Where:

33oPR : 33oP introduced (Bq/L),

"OP:33OP remaining in solution after incubation (Bq/L),

and I : volume of solution (L) to mass of soil (kg). m

Hy droly s ß C al cul at ions

168 Equations 9-2 andg-3 canbe combined to give the OP hydrolysed from PP (mg Pikg soil):

OP hydrolysed from PP: maximum potential labile PP - actual labile PP Eq.94

The Labile OP hydrolysed from PP (mg P/kg soil) can be calculated as:

Labile OP hydrolysed from PP Eq. 9-5 loP 31oP-3 control jl x 132or1t+ Kd33 oP))x 32 m OP

Where labile OP hydrolysed from PP is the amount of isotopically exchangeable OP derived from PP (mg P/kg soil),

3top- 31OP control: OP concentration in the soil solution extract (my'L)- OP concentration in the soil solution extract of the control treatment (no P added) (my'L),

32OP 3'Op : activity of 32PP hydrolysed to and measured in the OP peak (Bq/L),

and il : volume of solution (L) to mass of soil (kg). m

When using this equation the assumption is that all hydrolysed OP is labile, that there

is no precipitation and that there is no mobilisation of native OP (that is all OP in

solution, once the soluble OP from the control is subtracted from total OP, is derived

from hydrolysis of PP). Theoretically equations 9-4 and 9-5 should give identical

results.

Mo b ilis ation C al cul ations

The mobilisation of native OP was assessed by comparing the lability of native OP in

the control and in the PP treatment.

Labile native OP (mg P/kg soil) in the control treatment was determined according to

the conventional isotopic dilution calculation: 31oP t Eq.9-6 Labile native oP: r R33oP 33op " m

r69 Where labile native OP is the amount of isotopically available OP (mg P/kg soil),

3tOP: OP concentration in the soil extract (mg/L),

33OP: radioactivity remaining in solution (Bq/L),

R3'OP: radioactivity introduced (Bq/L),

and L: volume of solution (L) to mass of soil (kg)' m

The labile native OP where PP was added (mg P/kg) was determined according to:

Eq.9-7 ("oP -{"oe4ffi¡¡ E-native OP with PP added: ,, xR330P xfL OP m

Where labile native OP with PP added is the amount of isotopically exchangeable native OP where PP has been added (mg P/kg soil),

3tOP: OP concentration in the soil solution extract (mg/L)'

32oP 3'oP:'2PP hyd.olysed to and measured in the oP peak (Bq/L), 3lpp '''î*: 32PP 31P/Bq;, inverse specific activity of the spiked PP fertiliser 19

33OP: radioactivity remaining in solution (Bq/L) ("OP in the OP peaþ,

R3'OP: radioactivity introduced as labelled OP (Bq/L)'

and I : volume of solution (L) to mass of soil (kg). m

t70 9,3. Resulús and DrccussÍon

9.3.1. Effect of Pyrophosphate Addition on Soluble Phosphorus Goncentration and Soil Suspension pH

Soluble Phosphorus

There was some variability in the actual spiking concentration of PP between the 3 soils, therefore solution P data were noÍnalised by presenting soluble P recovered as a

a %o ofP added (Table 9-1). The recovery of P added as soluble P, expressed as percentage of P added was similar for the Grey Calcarosol and the Sodosol (26-l%

and25.2o/o,respectively). The soluble P in the Ferrosol was considerably lower at

3.3% suggesting a significantþ higher sorption capacity for P in this soil, which is

supported by the sorption curves described in chapter 5. Less P was added and

recovered in this study than in the study reported in chapter 8'

Table 9-1: Solution P as aVo of PP added in the Grey Calcarosol, the Ferrosol and the Sodosol. PP treatment results in soluble P as PP and as OP hydrolysed from PP (OP from PP).

Solution P asYo of PP added Standard Standard Soil PP Op from pp t11TOT Error Grey Calcarosol 0.1 0.0 26.7 0.9 Ferrosol 0.1 0.1 3.3 0.3 Sodosol 0.0 0.0 2s.2 7.0

Soil Suspension pH

In the Grey Calcarosol and the Ferrosol the soil suspension pH in the PP treatment

was greater than that of the control soil at the end of the incubation period (Table 9-

2).ThepH in the soil suspension of the Sodosol was 0.2 units lower with the addition

of pp compared to the control, which is a minimal shift in pH suggesting that this soil

solution is quite well buffered against pH change as compared to the other soils.

These data generally match those of Chapters 5, 7 and 8.

t7t Table 9-2: Soil solution pH measurements after a 7 day incubation with either no P (control) or Þp in the Grey Calcarosol, the Ferrosol and the Sodosol.

Standard Soil P treatment PH Error Grey Calcarosol Control 8.5 0.0 Pyrophosphate 8.9 0.0 Ferrosol Control 6.1 0.0 Pyrophosphate 6.8 0.0 Sodosol Control 7.7 0.0 7.5 0.1

9.3.2, Labile Pyrophosphate, Labile Hydrolysed Pyrophosphate and Mobilised Native PhosPhorus

Maximum Potential Labile Pyrophosphate and Actual Labile Pyrophosphate

In Chapter 7 the terms maximum potential labile PP and actual labile PP (Eq 9-1 and

3lPP: 32PP 9-2) wereused to describe the measurable ratio of lmaximum potential labile PP), and this ratio when accounting for hydrolysed PP (actual labile PP)'

However, when -800 mg P/þ soil was added as PP and incubated for 1 week, there

was no detectable PP remaining in solution. An important assumption of tracer

techniques is that the resultant specific activity of the fertiliser in soil solution will not

differ from that of the added fertiliser (Aten 1948;Di et a\.2000). If the concentration

31pp 3'PP of is at or below detection whereas is quantifiable, it is impossible to

confirm this assumption experimentally due to the relatively poorer detection limits

32PP for 3rPP of ion chromatography in comparison to the detection of by beta

counting. Therefore, the data for maximum potential labile PP and actual labile PP

could not be calculated using equation 9-l and9-2.

Labil e Hy drolys ed PYr oPho sPhat e

Due to the inability to calculate maximum labile PP and actual labile PP, hydrolysed

pp was not calculated according to the difference between these values (Eq 9-4).

However, the labile hydrolysed PP was determined using Eq. 9-5, which does not rely

172 3lPP: 32PP. on the ratio of There was a significant percentage of PP added present as labile hydrolysed PP after 1 week of incubation in all 3 soils (Table 9-3). These values are slightly lower than those found in Chapter 8, and are in closer agreement with the values determined inChapter 7.

Table 9-3: Labile Hydrolysed PP as a %" of P added.

Soil Hydrolysed PP Standard Error ofP ofP Grey Calcarosol 36.3 0.8 Ferrosol 13.1 0.6 Sodosol 32.8 4.7

Potential Mobilised PP

E-native OP with PP Added

The following equation was described in the materials and methods but is also

displayed below, as it is pivotal to the discussion of the results attained.

E-native OP with PP added (mg Plkg soil):

31oo (31 oP 32 oPx (==1)) -( J¿PP xR 33oP * (I) 33 m OP

This equation was developed to test whether the addition of PP to soil results in the 3'PP mobilisation of native OP. As already mentioned was used as attacet of the

3'OP added pP and was added carrier free to measure the lability of the native OP. As

32pp was applied as a tracer, the assumption used in this equation is that the specific

activiry of the added fertiliser ,#, will be the same as that of the hydrolysed PP.

Therefore to determine the concentration of OP in solution that is derived from the

hydrolysis of pp the component 132ot",H, is used. The total oP in solution

t73 minus the hydrolysed PP in solution would give the native OP in solution. However, when the data obtained in this experiment was entered into this equation the value for native Op was in most cases slightly negative. By entering the slightly negative value into the complex equation a perpetuation of this negative value by multiplication factors resulted in a negative E-value for native OP with PP added. The values obtained using this approach are reported in Figure 9-1'

100

BO -'õ (t, 60 o) l¿ fL o, 40 E o20f Elo LrJ U, o -20 -c= o- 8 -40 -c fL -60 I E-native P E-native P (PP added) -80 Grey Calcarosol Ferrosol Sodosol Soil

Figure 9-1: Comparison of the P E-value (mg P/kg soil) for native P and for native P with PP added in a Grey Calcarosol, a tr'errosol and a sodosol.

Prooasation of Error

propagation of error when making isotopic measurements can be an issue due to the

number of independent measurements required. The error in each single measurement

is propagated through calculations incorporating a number of independent

measurements. Frossard et al. (1996) briefly touched on the implications of

174 propagation of elïor in complex equations describing the mineralisation of organic P in soils. Frossard et at. (1996) showed that ashift in input data equivalent to twice the standard deviation of an observation could result in a vastly different conclusion for their data.

While theoretically Eq 9-7 should solve for the E-value of native OP with PP added, one of the problems with this methodology is the number of independent measurements required to solve the equation. A total of 8 independent measurements are required to solve this equation. Every independent measurement contains error associated with both the analytical process and with biological variability-

Some simple calculations demonstrate the effect of inserting input data in the range of the standard error of the measurement on the overall determination of the E-value of native Op with PP added. These shifts have the potential to result in vastly different

conclusions from a positive to negative results for the E-value of native P with PP 3tOP 32OP added. These examples are provided in Table 9-4, where either or values

are altered plus or minus the standard error of the observations. This is a less dramatic

shift in the input data using the observation plus or minus the standard error,

compared to Frossard et al. (1996) who used twice the standard deviation'

31OP 32OP Table 9-4: propagation of error for Eq9-7 where the and values were altered by plus or minus the standard error of the observation'

Standard Observation- Observation Observation* Error Standard Standard Error Error OP (mg P/L) 1.04 22.57 23.61 24.65 OP (PP added) E- Native -28.18 -2.53 23.72 (mg Pikg soil) 3toe lortwmr¡ s35 17104 18239 18774 (PP added) E-Native OP 1s.01 -2.53 -0.27 soil

175 It can be therefore concluded that, even if the methods and equations described above can in theory be used to predict variations in native OP lability, in practice this approach is so complex that obtaining reliable data is difficult if not impossible to achieve.

Specific Activitv- An Alternative Approach

A much simpler altemative approach to assess whether mobilisation of native OP is caused by addition of PP makes use of the specific activity of PP and OP in solution.

after The specific activity of PP of the hydrolysed P 132OP: "OP) in the soil solutions

1 week of incubation was compared with the specific activity of the added fertiliser 3tre¡ 132PP: ltable 9-5).

Table 9-5: Comparison of the specific activity (sA) (Bq/mg P) of hydrolysed P in soluiion, with the specific activity of the added PP fertiliser.

Soil SA of Standard SA ofPP hydrolysed P Error fertiliser

Grey Calcarosol 77s 2I 768 Ferrosol 9s4 28 790 Sodosol 7s0 85 672 3lOP If there was mobilisation of native P the value would have increased, making the

the added specific activity of the hydrolysed P 132OP: "OP) smaller than that of

fertiliser. This approach is similar to that used by Walbridge and Vitousek (1987) and

Lopez-Hern andez and Nino (1g93),but addresses the concerns of Frossard e/ ø/.

(1996) by measuring values in soil suspensions in the compartment to which the

fertiliser was added, and by considering the multiple compartments across which the

P distributed.

In not one of the three soils was the specific activity of hydrolysed P in solution

smaller than the specific activity of the added fertiliser, and when the boundaries of

the standard error were considered there was no difference between the specific

176 activity of the hydrolysed P in solution and of the added fertiliser, except for the

Ferrosol where the specific activity of hydrolysed P was 2OYo gteater than that of the added fertiliser. Indeed, the data obtained in this study suggest that there was no

most of the OP in substantial , if any,mobilisation of native OP. It also suggests that solution was derived only from the PP fertiliser'

The high sorption capacity of the Ferrosol for PP with only 3-4o/o ofPP added remaining in solution after 1 week, could have contributed to the higher specific

31P activity of hydrolysed P, due to problems with detection of in the highly P sorbing soil. Isotopic studies in the past have illustrated the erroneous effect that highly P sorbing soils þarticularly those high in Fe oxides) have on isotopic studies of P due to 31P 32P diffrcuþ with the determination of very low concentrations of as compared to

(Amer et al. 1969; Barrow l99l; Fardeau 1996; Hamon and Mclaughlin 2002)'

Isotopic studies (Frossard et al. 1994a; Morel and Plenchette 1994) have shown that

the mineralisation of P from organic P occurs over a time course of three months and

in some cases longer, compared to the 1 week incubation of this study. It is possible

that the solubilisation of native OP through the addition of PP is a slow reaction and

therefore not observable within 1 week.

9.4. Conclusions

In conclusion, this double labelling method was unable to detect mobilisation of

native Op where PP was added to soil. The E-value equation for native OP with PP

added should provide a measure for the ability of PP to mobilise native OP. However,

due to issues related to the complexity of the equation and corresponding propagation

of error, sensible values were not derived. Evaluation of the specific activity of the

hydrolysed P in solution, which did not significantly differ from the specific activity

of the added PP fertiliser, suggests that there was no mobilisation of native P in the 7

t77 day period in the soils tested in this experiment. Evaluation of the specific activity in this way may be the best currently available evaluation of the potential for the addition of PP to mobilise native OP. However, it is recofllmended that this work is extended to further assess the possibility that PP solubilises native OP most likely associated with DOC over time, in a greater range of soils and with a higher concentration of PP applied to represent fertiliser band concentrations.

178 Ghapter 10. General Discussion

Poþhosphate fertilisers have demonstrated great potential for enhanced P uptake efficiency in cereal crops as compared to granular OP fertilisers in calcareous soil types of Southern Australia. These outcomes have been recorded in both field and glasshouse evaluations, with yield increases of up to 40Yo in wheat where pollphosphate fertiliser was applied compared to granular OP fertiliset at an equivalent P rate.

The literature review outlined the potential contribution that a greater understanding of the chemical reactions of poþhosphate fertilisers in Australian soil types would have to the appropriate development of a fluid fertiliser industry in Australia. The

experimental component of this thesis has investigated sorption, hydrolysis and

lability or isotopic exchangeability reactions of the P species in a polyphosphate

fertiliser using a combination of sophisticated ana$ical techniques.

10,1. Summary of Findings

10.1.1. Analytical Techniques

To successfully address the research gaps described in the literature review, it was

necessary to find an analytical technique to speciate the chemical P species in a

poþhosphate fertiliser, and with the ability to make a separation of these P species

for collection and further analysis in isotopic experiments. Following araîge of tests,

IC was deemed to be the most appropriate technique with the dual purpose of

speciation and separation.

r79 10.1.2. HYdrolYsis in Solution

The effect of storage factors on the stability of polyphosphate fertilisers is an important industry issue, and astudy was made of the effects of pH, temperature and

the time on the stability of polyphosphate fertiliser. All three of these factors influence stability of polyphosphate fertiliser, the most marked of these effects being pH'

Therefore the industry recom,mendation for those wishing to modiff the fertiliser formulation is to do so immediately prior to application, or run the risk of compromising the stability and composition of the pollphosphate product'

10.1.3. HYdrolYsis in Soil

Following an assessment of the stability of the poþhosphate fertiliser under storage

made' conditions an assessment of the stability of polyphosphate in a soil system was

was In this case PP, the dominant condensed P species in polyphosphate fertiliser,

applied and the degradation over an incubation period of 1 to 2I days was analysed

measured the using solid-state '1P NMR. The NMR technique successfully

degradation process, which is a unique application of this technique. The total P

applied that was measurable was considerably greater in the solid-state as compared

PP to a solution extraction measured by IC. The results suggest that there is some

hydrolysing in the solid-state portion, which is not extractable by lM NaOH for

measurement by IC.

10.1.4. SorPtion

There have been sorption studies comparing the sorption capacity of soil for OP

compared to PP but not on Australian soil types. Eight different soil types were

pH and evaluated using eight point sorption curves. Simultaneous analysis of DOC,

have a cations was made. The results of this study suggested that Australian soil types

180 greater sorption capacity for PP than for OP. The addition of PP to soil results in a release of DOC in solution, an increase in soil solution pH and Fe content and a reduction in soil solution Ca content.

10.1.5. lsotopic Studies of Lability, Hydrolysis and Partitioning

Over the course of four chapters, the use of isotopic techniques for the investigation of the partitioning, lability or isotopic exchangeability and hydrolysis of PP was discussed. It was necessary to first clearly outline the principles of isotopic dilution, and the core assumptions that come with the use of these techniques.

32PP The use of a tracer technique where PP was labelled with and OP was labelled with 33OP enabled the measurement of the isotopic recovery and hydrolysis of PP in soil over a24hovrperiod. The recovery increased and the proportion of PP added which hydrolysed decreased with increasing dose of PP added'

32PP In Chapter 9 this tracer technique was extended such that was used to trace PP

hydrolysis over a 7 day period, while "OP was used to measure the lability of native

Op. The aim of this technique was to determine if the addition of PP to soil resulted in

the mobilisation of native OP. The calculation developed for assessing the E-value of

native p when PP was added was susceptible to the effects of the propagation of error.

With 8 independent measurements entered into the calculation, variation in one of the

input data by plus or minus the standard error of the observation was enough to

significantly alter the result. However, the simple comparison of the specific activity

of the added fertiliser with the specific activity of the OP in soil solution suggested

that there was no mobilisation of native OP. If there was mobilisation of native OP the

3lop value would increase, making the specific activity of OP in solution decrease

and be less than that of the added fertiliser.

181 In Chapter 8 a study compared the lability of OP and PP in 3 Australian soil types over an incubation of l-7 days. The study found that initially there was significantly more labile P from oP, but after 7 days there was no difference between P supplied as

OP or PP.

There was significantly greater recovery of the fertiliser P added after I week when added as pp. Part of the discrepancy between the soluble P data and the lability study was related to the propagation of error due to the complexity of the double labelling technique requiring a series of independent measurements contributing to the final

calculation.

The double labelling technique is an improved methodology for the determination of

the labile p pool and it's components due to the addition of PP. The previously used

method overestimated the labile P pool due to PP addition and did not uphold some

important isotopic assumPtions.

10.1.6. Fate and Effects of PP Reactions in soils

Based on the findings of this thesis, a conceptual model of our current understanding

of the processes and reactions of PP are demonstrated in Figure 10-1, which also

highlights potential areas for further research.

182 Po Storage Area of potential new Fertiliser ln rsseardt

Plant P Uptake

FertiliserP (OP + PP)

Fertilis€r Band Zn applied togethêr

Diffusion of PP rh2osphore of PP

Fixed SoilSolution angeable P pool P pool P pool OP OP OP (native or fertiliser OP) (native or fertiliser (native or fertiliser OP) oP) Dissolution PP PP Hydro ys s PP HWro¡ysis Hydrolys s + + I V OP OP OP

PP PP Releêsês Binds Fe/Al/ Ca prêcipitate DOC PP may Ca over

DoC release may mobilis native P

Ueasurement techn¡que8: tc NMR lsotopb Labeling tr'igure 10-1: The reactions of P added as OP and PP in polyphosphate fertiliser.l will distribute amongst the solution, exchangeable and fixed P pools. In these pools p wü react with Fe, Al, Ca and DOC. Processes requiring further investigation include reactions over time, diffusion, rhizosphere and plant uptake processes, multi-nutrient fertiliser reactions, and release of native P from DOC.

Fertiliser Storage and Addition

The dominant P species in a polyphosphate fertiliser are OP and PP, with small amounts of TP. Under storage condensed P species in pollrphosphate fertilisers can

183 undergo hydrolysis reactions to more simple P compounds. These reactions are favoured in conditions of low pH, high temperature and increasing time' polyphosphate fertiliser is generally added to the soil in a band below the seed and the fertiliser predominantly contains OP and PP. In addition to P, polyphosphate fertilisers contain N and some trace elements. The interactions of multi-nutrient fertilisers aîe an arcathat requires further research.

Solution Pool

'When poþhosphate fertiliser is added to soil it moves into the solution pool as OP or pp, which can hydrolyse to form OP. Also present in the solution pool is native OP.

The research presented in this thesis shows that the addition of PP to soil results in the release of Fe, Al and DOC into soil solution. It is possible that the solubilisation of

DOC releases P from organic matter to the solution pool. This process was iclentified

and while initial experiments were undertaken, the potential for PP to mobilise native

P reserves requires further research'

Plant and Rhizosphere Processes

plants access P by mass flow and diffusion from the solution pool. Micro-organisms

in the rhizosphere and plant root properties influence the rate and species of P that are

taken up by the plant. These are aII areas for further research.

Exchangeable Pool

The p species in the solution pool undergo sorption reactions with the exchangeable

phase, and the exchangeable phase, like the solution phase, will contain fertiliser OP,

and pp which can hydrolyse to OP. The exchangeable pool also contains native OP.

The addition of PP to soil results in a decrease in the Ca concentration in solution

suggesting a fixation reaction. While over the short time scale of the experiments

184 presented here there was no evidence of precipitation, this may occur over longer time periods and requires further investigation.

Fixed Pool

Over time P in the exchangeable phase can participate in desorption reactions and return to the solution phase or participate in precipitation reactions and move to the fixed pool. Phosphorus in the fixed pool could be present as OP, PP and hydrolysed

OP from fertiliser, and native OP from the soil.

Measurement of Reactions of PP in Soils

The techniques that have been identified as the most appropriate for the measurement of these processes include ion chromatography, nuclear magnetic resonance and isotopic techniques.

10.2. Future Research

While the findings of this thesis have made a significant contribution to the

understanding of polyphosphate fertiliser in Australian soil types, there have been

several new research questions identified and there are areas ofresearch presented in

this thesis that could be further developed'

Dffision

There is potential to investigate the interactions between diffusion, hydrolysis and

lability reactions of PP using the double labelling isotopic technique. A Petri dish

system was previously used to compare the diffusion and lability of granular OP

fertilisers as compared to fluid OP fertilisers (Lombi et a\.2004a). A further extension

of this technique would be to adapt the diffusive gel in thin film method (Mason et al.

2005) to make a msasure of the diffusive flux of the P species away from the point of

application. This system could be adapted so that measurements could be made of the

diffusion, lability and hydrolysis reactions of PP.

185 Hydrolysß

Hydrolysis reactions were successfully measured in a highly calcareous soil using

NMR. There is an opportunity for the testing of the NMR technique for a wider range of soil types. The capability of NMR to measure the degradation of condensed P species has a range of applications beyond nutrition research.

Organic Carbon Solubilisation and Interactions with P Chemistry

Due to the observation of dissolution of DOC with the addition of PP, there is an opportunity to investigate the behaviour of organic carbon with the addition of PP. ft is possible that this dissolution reaction extracts P-bearing organic matter and it is still unclear whether native P, especially organic P, is released by PP application'

This would require further development of the isotopic techniques.

Rhizosphere and Plant Processes

an The ability of a plant to take up the PP ion prior to hydrolysis largely remains

gnanswered question. There is potential to build on the double labelling isotopic

techniques developed here to investigate the fate of PP in the rhizosphere and to

determine if plants take up P as the PP ion. Plant studies could assess whether roots

influence pP hydrolysis reactions. The effects of rhizosphere micro-organisms on the

hydrolysis and lability reactions of PP also warrant investigation.

186 References

Ahmad F. and Kelso W.I. (2001) Pyrophosphate as a Source of Phosphorus: Hydrolysis Under Different Conditions . Journal of Research (Science) 12,130-139

Al-Kanani T. and MacKenzie A.F. (1991) Sorption and Desorption of orthophosphate and Pyrophosphate by Mineral Fractions of Soils, Goethite and Kaolinite. Canadian Journal of Soil Science 71, 327 -338 '

Amer F., Bouldin D.R., Black C.A. and Duke F.R. (1955) Char actenzation of S oil Phosphorus by Anion Exchange Resin Adsorption and P"-Equilibration. Plant and Soil6,391-408.

Amer F., Mahdi S. and Alradi A. (1969) Limitations in Isotopic Measurements of Labile Phosphate in Soils. Journal of Soil Science 20'

Amer F. and Mostafa H.E. (1931) Effect of Pyrophosphate on Orthophosphate Reactions in Calcareous Soils. Soil Science Society of America Journal 45,842-847 .

Anderson G. (1980) Assessing Organic Phosphorus in Soils. In'The Role of pp' Phosphorus in Agriculture'. (Èds F.E. Khasawneh,E.C. Sample andE, Kamprath) 4tl-^28. (Amerióan Societies of Agronomy, Crop Science and Soil Science: Madison)

Anon (2000) 'Formacs HT TOC/TN Analyser User Manual'' (US)

Aten A.H.W. (194S) Isotopes as Tracers in Analytical Chemistry. Analytica Chimico Acta2,492-500.

Bache B.V/. (1964) Aluminium and hon Phosphate Studies Relating to Soils tr. Reactions BetweenPhosphate andHydrous Oxides. Journal of Soil Science 15, 110- 1 16.

Baluyot E.S. and Hartford C.G. (1996) Comparison of Poþhosphate Anaþsis by Ion Chromatography and by Modified End-Group Titration. Journal of Chromatography A739,217-222.

Barber S.A. (1980) Soil-Plant Interactions in the Phosphorus Nutrition of Plants. In (American 'The Role of Phosphorus in Agriculture'. (Ed. E. Kamprath) pp. 591-616. Society of AgronomY: Madison)

Barrow N.J. (1973) The Description of Phosphate Adsorption Curves. Journal of Soil Science 29, 441-462.

Barrow N.J. (1933) A Mechanistic Model for Describing the Sorption and Desorption of Phosphate by Soil. Journal of Soil Science 34,733-750'

Barrow N.J. (1937) 'Reactions With Variable Charge Soils.' (Martinus Nijhoff Publishers : Dordrecht, The Netherlands.)

187 Barrow N.J. (1991) Testing a Mechanistic Model.XI. The Effects of Time and of Level of Appìication on Isotopically Exchangeable Phosphate. Journal of Soil Science 42,271-288.

Barrow N.J. (1999) The Four Laws of Soil Chemistry: The Leeper Lecture 1998- Australian Journal of Soil Research 37' 7 87 -829.

Benbi D.K. and Gilkes R.J. (1937) The Movement into Soil of P from Superphosphate Grains and its Availability to Plants. Fertilizer Research 12,2I-36.

Bertrand I., Holloway R.E., Armstrong R.D. and Mclaughlin M.J. (2003) Chemical Characteristics of Phosphorus in Alkaline Soils from Southern Australia. Australian Journal of Soil Research 41,6I-76-

Bertrand I., Mclaughlin M.J., Holloway R.E., Armstrong R.D. and McBeath T'M' (2006) Changes in P Bioavailability Induced by the Application of Liquid and Powder òo*".5 of p, N andZnin Alkaline Soils. Nutrient Cycling in Agroecosystems 74,27- 40.

Blanchar R.W. and Caldwell A.C. (1966) Phosphate-Ammonium-Moisture Relationships in Soils: II. Ion Concentrations in Leached Fertilizer Zones and Effects on Plants. Soil Science Society of America Proceedings 30,43-47.

Blanchar R.W. and Hossner L.R. (1969a) Hydrolysis and Sorption Reactions of Orthophosphate, Pyrophosphate, Tripolyphosphate, and Trimetaphosphate Anions Addeå to an Elliot Soil. ,Soil Science Society of America Proceedings 33,l4l-144-

Blanchar R.\ù/. and Hossner L.R. (1969b) Hydrolysis and Sorption of Ortho-, Pyro-, Tripoly-, and Trimetaphosphate in 32 Midwestern Soils. Soil Science Society of America Proceedings 33, 622-625.

Bolan N.S. and Robson A.D. (1933) Plant and Soil Factors Including Mycorrhizal Infection Causing Sigmoidal Response of Plants to Applied Phosphorus. Plant and Soil73, t87-201.

Bolland M.D.A., Gilkes R.J., Brennan R.F. and Allen D.G. (1996) Comparison of Seven Phosphorus Sorption Indices. Australian Journal of Soil Research 34, 81-89.

Brady N.C. and Weil R.R. (1996) 'The Nature and Properties of Soils.' (Prentice-Hall Inc.:New Jersey)

Bramley R.G.V., Barrow N.J. and Shaw T.C. (1992) Ther Reaction between phosphâte and Dry Soil I. The Effect of Time, Temperature and Dryness. Journal of Soil Science 43, 7 49-l 58.

Bremner J.M. and Lees H. (1949) Studies on Soil Organic Matter Part II' The Extraction of Organic Matter from Soil by Neutral Reagents. Journal of Agricultural Science 39,274-279.

188 Bunemann E., Steinebrunner F., Smithson P.C., Frossard E. and Oberson A' (2004) phosphorus Dynamics in a Highly \Weathered Soil as Revealed by Isotopic Labeling Techniques . Soil Science Society of America Journal6S,1645-1655.

Burkitt L.L., Moody P.w., Gourley C.J.P. and Hannah M.c. (2002) A Simple phosphorus Buffering Index for Australian Soils. Australian Journal of Soil Research 40,497-513.

Burns R.G. and Slater J.H. (1932) 'Experimental Microbial Ecology.' (Blackwell Scientific Publications)

Busman L.M. and TabatabaiM.A. (1935) Hydrolysis of Trimetaphosphate in Soils. Soil Science Society of America Journal49,630-636'

Castro B. and Torrent J. (1993) Phosphate Sorption by Calcareous Vertisols and Inceptisols as Evaluated àom Extended P-sorption Curves. European Journal of Soil Science 49,661-667.

Celi L., Lamacchias. and Barberis E. (2000) Interaction of Inositol Phosphate with Calcite. Nutrient Cycling in Agroecosystems 57,211-277 '

clesceri L.s., Greenberg A.E. and Eaton A.D. (1993) 'standard Methods for the Examination of Water and'Wastewater.' (American Publio Health Association, American Water Works Association, Water Environment Federation: Washington, D.C.)

Colwell J.D. (1963) The Estimation of Phosphorus Fertilizer Requirements of Wheat in Southern New Sôuth Wales by Soil Analysis. Australian Journal of Experimental Agriculture 3, 190-197 .

Committee A.M. (1937) Recommendations for the Definition, Estimation and Use of the Detection Limit. Analyst ll2, t99-204.

costin A.B. and v/illiams c.H. (1983) 'Phosphorus in Australia.' (centre for Resource and Environmental Studies: Canberra, A.C.T.)

Cresser M., Killham K. and Edwards T. (1993) 'Soil Chemistry and it's Applications.' (Cambridge University Press : Melbourne)

Currie L.A. (1999) Detection and Quantification Limits: Origins and Historical Overview. Analytica Chimica Acta 391, l2l -134.

DalalR.C. (1977) Soil Organic Phosphorus. Advances in Agronomy 29,83-ll7 '

De Jager H.J. and Heyns A.M. (1998) Kinetics of Acid- CatalyzedHydrolysis of a Polphosphate in Water. Journal of Physical Chemistry 102,2838-2841- \ù/.H. Dean L.4., Nelson W.L., Mackenzie 4.J., Armiger and Hill W.L. (1947) Application of Radioactive Tracer Technique to Studies of Phosphatic Fertilizer

189 Utilization by Crops L Greenhouse Experiments. Soil Science Society of America Proceedings 1947, 107 -112.

Di H.J., Condron L.M. and Frossard E. (1991) Isotope Techniques to Study phosphorus Cycling in Agricultural and Forest Soils: A Review. Biologt and Fertility of Soils 24,1-12.

Di H.J., Cameron K.C. and Mclaren R.G. (2000) Isotopic Dilution Methods to Determine the Gross Transformation Rates ofNitrogen, Phosphorus, and sulfur in Soil: A Review of the Theory, Methodologies, and Limitations. Australian Journal of Soil Res earch 38, 213 -230.

Dick R.p. and Tabatabai M.A. (1936) Hydrolysis of Pollphosphates in Soils. Soil Science 142,132-t40.

Dick W.A. and Tabatabai M.A. (1977) Determination of Onhophosphate in Aqueous Solutions Containing Labile Organic and Inorganic Phosphorus Compounds. Journal of Environmental Quality 6,82-85.

Dijkstra E.A. and Fitzhugh R.D. (2003) Aluminium Solubility and Mobility in Relation to Organic Carbon in Surface Soils Affected by Six Tree Species of the Northeastem United States. Geoderma 114,33-47 .

Dougherty W.D., Smernik R.J. and Chittleborough D.J. (2005) IdentificatioX,?ttd quaitifrcátion of Soil Organic and Inorganic Phosphorus Using Solid-State "P NMR' Soil Science Society of America Journal69,2058-2070'

Dubus I.G. and Becquer T. (2001) Phosphorus Sorption and Desorption in Oxide-rich Ferrasols of New Caiedonia. Australian Journal of Soil Research 39, 403-414.

E¡-Zahaby E.M. and Chien S.H. (1982) Effect of Small Amounts of Pyrophosphate on Orthophosphate Sorption by Calcium Carbonate and Calcareous Soils. Soil Science Society of America Journal 46.

Englestad O.P. and Allen S.E. (1971) Ammonium Pyrophosphate and Arnmonium On-hophosphate as Phosphorus Sources: Effects of Soil Temperature, Placement, and mcubàtion. Soil Science Society of America Proceedings 35, 1002-1004.

FAO (2003) High P Fixing Soils. In. (Food and Agriculture Organization of the UnitedNations)

Fardeau J.C. (1996) Dynamics of Phosphate in Soils. An Isotopic Outlook. Fertilizer Research 45,91-100.

Fardeau J.C., Guiraud G. and Marol C. (1996) The Role of Isotopic Techniques on the Evaluation of the Agronomic Effectiveness of P Fertilizers. Fertilizer Research 45, 101-109.

190 Feng G., Song Y.C., Li X.L. and Christie P. (2003) Contribution of Arbuscular Myónhizal Fungi totJtllization of organic Sources of Phosphorus by Red Clover in a Calcareous Soil. Applied Soil Ecologt 22, 139-148'

FruzierA.W. and Dillard E.F. (1981) Ionic Effects on the Hydrolysis of poll,phosphates in Liquid Fertilizers. Journal of Agricultural and Food Chemistry 29, 69s-698.

Freeman J.S. and Rowell (1981) The Adsorption and Precipitation of Phosphate onto Calcite. Journal of Soil Science32,75-84.

and Friess S.L. (1952) The Energetics of Acid,-catalyzed Hydrolysis of Triphosphoric Pyrophosphoric Acids. Journal of the American Chemical Society 74,4027-4029.

Frossard E., Fardeau J.C., Brossard M. and Morel J.L. (1994a) Soil lsotopically Exchangeable Phosphorus: A Comparison between E and L Values. Soil Science Society of America Journal 58, 846-851.

Frossard E., Tekely P. and Morel J.L. (1994b) Chemical chnacteization and Agronomic Effectiveness of Phosphorus Appilied as a Poþhosphate- chitosan Complex. Fertilizer Research 37, 151-158.

Frossard E., Lopez-HernandezD. and Brossard M. (1996) Can Isotopic Exchange Kinetics give Vãluable Information on the Rate of Mineralizationof Organic Phosphorus in Soils? Soil Biolog,t and Biochemistry 28' 857- 864'

Frossard E,., Skrabal P., Sinaj S., Bangerter F. and Traore O. (2002) Forms and Exchangeability of Inorganic Phosphate in Composted Solid Organic'Wastes. Nutrient Cy cl ing in Agroeco sys t ems 62, I03 -l 13'

Gilliam J.W. and Sample E.C. (1963) Hydrolysis of Pyrophosphate in Soils: pH and Biological Effects. Soil Science 106,352-357 .

Gilliam J.V/. (1970) Hydrolysis and Uptake of Pyrophosphate by Plant Roots. Soil Science Society of America Proceedings 34, 83-86'

Giordano p.M., Sample E.C. and Mortvedt J.J. (1971) Effect of Ammonium Ortho- and Pyrophosphate on Znand P in Soil Solution. Soil Science lll, 101-106'

Grinstead M.J., Hedley M.J., white R.E. and Nye P.H. (1932) Plant-Induced changes in the Rhizosphere of h.ape (Brassica Napus var' Emerald) Seedlings I'p! Change and the Increãse in P Concentration in the Soil Solution. New Phytologist 91,19-29.

Grzmil B.U. and Kic B. (2002) Pyro- and Tripolyphosphates and Citrates as the Complexing Agents for Micronutrients in Liquid Fertilizers. Industrial and Engineering Chemistry Res earch 41, 139 -1 44.

Guan X.-H., Chen G.-H. and Shang C. (2005) Competitive Adsorption Between Orthophosphate and Other Phosphates on Aluminium Hydroxide. Soil Science 170, 340-349.

191 Gunnar Sillen L. (1964) 'Stability Constants of Metal lon Complexes' Part 1- Inorganic Ligands.' (The Chemical Society: London)

Guppy C.N., Menzies N.w., Moody P.W. and Blamey F.P.C. (2005) competitive Sotpìío" Reactions Between Phosphorus and Organic Matter in Soil: A Review' Australian Journal of Soil Research 43,189-202.

HalliwellD.J., McKelvie I.D., Hart B.T. and Dunhill R.H' (2001) Hydrolysis of Triphosphate from Detergents in a Rural Vy'aste Water System. W'ater Research 32, 448-4s4.

Hamad M.E., Rimmer D.L. and Syers J.K. (1992) Effect of Iron Oxide on Phosphate Sorption by Calcite and Calcareous Soils. Journal of Soil Science 43,273-281'

Hamon R.E., Bertrand L and Mclaughlin M.J. (2002) Use and Abuse of Isotopic Exchange Data in Soil Chem ístry. Australian Journal of Soil Research 40, 1- 1 1 '

Hamon R.E. and Mclaughlin M.J. (2002) Interferences in the Determination of Isotopically Exchangeable P in Soils and a Method to Minimise them. Australian Journal of Soil Research 40, 1383-1397 .

32P-Method Harrison A.F. (1932¡ to Compare Rates of Mineralization of Labile Organic PhospÌrorusin Woodland Soils. Soil Biologt and Biochemistry 14,337-341'

Hashimoto I., Hughes J.D. and Philen O.D.J. (1969) Reactions of Triammonium Pyrophosphate with Soils and Soil Minerals. ,Soll Science Society of America Proceedings 33, 401-405.

Hashimoto I. and Lehr J.R. (Lg73)Mobility of Poþhosphates in Soil. Soil Science Society of America Proceedings 37,36-41-

Hedley M.J. and Mclaughlin M.J. (2005) Reactions of Phosphate Fertilizers qd By- Produôts in Soils. In'Phosphorus: Agriculture and the Environment'' (Ed' A'N' Sharpley) pp. 181-252. (American Society of Agronomy, Crop Science Society of America, Soil Science Society of America: Madison,'WI)

Hinsinger p. (2001) Bioavailability of Soil Inorganic P in the Rhizosphere as Affected by Roo--t-Induced Chemical Changes: A Review. Plant and Soil237,173-195'

Hirsch R.E. and Sussman M.R. (1999) Improving Nutrient Capture from Soil by the Genetic Manipulation of Crop Plants. TIBTECH 17,356-360'

Holford I.C.R. and Mattingly G.E.G. (1975) The High- and Low- Energy Phosphate Adsorbing Surfaces in Calcareous Soils. Journal of Soil Science 26,407-4I7 .

Holloway R.E., Bertrand I., Frischke 4.J., Brace D.M., Mclaughlin M'J. and Soils Shepperá W. (2001) Improving Fertiliser Efficiency on Calcareous and Alkaline wittr rmid sources of P, N andzn. Plant and soi|236,209-219.

t92 Holloway R.E., Frischke A.J. and Brace D. (2002) 'How Much Fluid P Fertiliser is EnoughiEyre Peninsula Farming Systems Summary.' South Australian Research and Development Institute, Pt. Lincoln, South Australia.

Hons F.M., Stewart W.M. and Hossner L.R. (1936) Factor Interactions and their Influence on Hydrolysis of Condensed Phosphates in Soils. Soil Science 141,408- 4t6.

Hossner L.R. and Phillips D.P. (1971) Pyrophosphate Hydrolysis in Flooded Soil. Soil Science Society of America Proceedings 35,379-383-

Huffrnan E.O. and Taylor A.W. (1963) The Behaviour of Water-soluble Phosphate in Soils. Journal of Agricultural and Food chemistry ll,182-187.

Isbell R.F. (1 997)'The Australian Soil Classifi cation.' (CSIRO Publishing: Melbourne) rupAC (2005) Equilibrium Constants. In. (International Union of Pure and Applied Chemistry)

Jenkinson D.S. and Powlson D.S. (1976) The Effects of Biocidal Treatments on Metabolism in Soil- V.A Method for Measuring Soil Biomass. Soll Biologt and B iochemistry 8, 209 -2 13 .

Joner E.J., van Aarle I.M. and Vosatka M. (2000) Phosphatase Acitivity of Extra- radical Arbuscular Mycorrhizal Hlphae: A Review. Plant and Soil226,199-210.

Jones D.L. (1998) Organic Acids in the Rhizosphere- A Critical Review. Plant and 5oi1205,25-44.

Kaplan D.L. and Hartenstein R. (1979) Problems with Toluene and the Determination of Extracellular Enzyme Activity in Soils. Soil Biology and Biochemistry 11, 335-338

Khasawneh F.E., Sample E.C. and Hashimotol. (1974) Reaction of Ammonium Ortho- and PolyphospLate Fertilizers in Soil: I. Mobility of Phosphorus. Soil Science Society of America Proceedings 38,446-45I.

Khasawneh F.E., Hashimoto I. and Sample E.C. (1979) Reactions of Ammonium Ortho- and Poþhosphate Fertilizers in Soil: II. Hydrolysis and Reactions with Soil. Soil Science Society of America Journal43,52-57.

Kura G., ¡1¡¡iyazal

Larsen S. (1952) The Use of P32in Studies on the Uptake of Phosphorus by Plants- Plant and SoiI1, 1-10.

193 Lawton K. and Vomocil J.A. (1954) The Dissolution and Migration of Phosphorus from Granular Superphosphate in Some Michigan Soils. Soil Science Society of America Proceedings 18, 26-32.

Leikam D.F., Murphy L.S., Kissel D.E., Whitney D.A. and Moser H.C. (1983) Effects of Nitrogen and Phosphorus Application Method and Nitrogen Source on Winter Wheat Grain Yield and Leaf Tissue Phosphorus. Soil Science Society of America Journal 47,530-535.

Lindsay V/.L., Frazier A.W. and Stephenson H.F. (1962) Identification of Reaction products from Phosphate Fertilizers in Soils. Soil Science Society of America Proceedings 26, 446-452.

Lindsay w.L. (1979) Phosphates. In 'chemical Equilibria in soils'. (Ed. v/.L- Lindsay) pp.162-205. (John Wiley and Sons: New Yorþ

Lombi E., Mclaughlin M.J., Johnston c., Armstrong R.A. and Holloway R-8. (2004a)Mobility, Solubility and Lability of Fluid and Granular Forms of P Fertiliser in Cul"ur"o.rs and Non-Calcareous Soils under Laboratory Conditions - Plant and Soil 269,25-34.

Lombi E., Mclaughlin M.J., Johnston c., Armstrong R.D. and Holloway R.E. (2004b) Mobility and Lability of Phosphorus from Granular and Fluid Monoammonium Phosphate Differs in a Calcareous Soil. Soil Science Society of America Journal 68, 682-689.

Lopez-HernandezD. and Nino M. (1993) Phosphorus Mineralization During Latoratory Incubation in Soils Derived from Different Textured Parent Materials. Geoderma 56,527-537.

Marschner H. (1995) 'Mineral Nutrition of Higher Plants.' (Academic Press Ltd.: London)

Martin A.E. and Reeve R. (1955) A Rapid Manometric Method for Determination of Soil Carbonate. Soil Science 79,187-197.

Mason S., Hamon R.8., Nolan A.L.,Zhang H. and Davison V/. (2005) Performance Using of a Mixed Binding Layer for Measuring Anions and Cations in a Single Assay the Diffusive Gradients in Thin Films Technique . Analytical Chemistry 77,6339- 6346.

Masson P., Morel C., Martin E., Oberson A. and Friesen D. (2001) Comparison of Soluble p in Soil Water Extracts Determined by Ion ChromatograPhY, Colorimetric, and Inductively Coupled Plasma Techniques in PPB Range. Communications in Soil Science and Plant Analysis 32,2241-2253-

McBeath T.M., Smernik R.J., Lombi E. and Mclaughlin M-J. (2006) Hydrolysis of Pyrophosphate in a Highly Calcareous Soil: A Solid-State Phosphorus-3l NMR Study. Soil Science Society of America Journal70,856-862'

t94 Mclaughlin M.J. and Alston A.M. (1986) The Relative Contribution of Plant Residues and Fertiliser to the Phosphorus Nutrition of Wheat in a PasturelCeteal System. Australian Journal of Soil Research24,517-26.

Mclaughlin M.J., Alston A.M. and Martin J.K. (19SS) Phosphorus Cycling in Wheat- Pasture Rotations. I. The Source of Phosphorus Taken up by V/heat. Australian Journal of Soil Research 26, 323-331.

Mclaughlin M.J., Lancaster P.4., Sale P.G., Uren N.C. and Peverill K.I. (1994) Compaãson of Cation/ Anion Exchange Resin Methods for Multi-Element Testing of Acidic Soils. Plant and 5oi132,229-240-

Mclaughlin M.J., Lombi E., HollowaY B., Johnston C., Brace D. and Armstrong R' (2003) Reactions of Granular and Fluid Products in Calcareous Soils. 'Fluid Forum Proceedings'. Phoenix, 2003 . (Fluid F ertilizer Foundation)

Miner G. and Kamprath E. (1971) Reactions and Availability of Banded Polyphosphate in Field Studies. Soll Science Society of America Proceedings 35,927- 930.

Mrkeni P.N.S. and MacKenzie A.F. (1935) Retention of Ortho- and Polyphosphates in Some Quebec Soils as Affected by Added Organic Residues and Calcium Carbonate. Canadian Journal oJ' Soi I Science 65, 575-5 85.

Moghimi 4., Lewis D.G. and Oades J.M. (1978) Release of Phosphate from Calcium Phosphates by Rhizosphere Products. ,Soil Biologt and Biochemistry 10,217-28I.

Morel C. and Fardeau J.C. (1939) Native Soil and Fresh Fertilizer Phoshorus Uptake as Affected by Rate of Application and P Fertilizers. Plant and Soil l1.5,l23-I28.

Morel C. and Plenchette C. (1994) Is the Isotopically Exchangeable Phosphate of a Loamy Soil the Plant-Available P? Plant and Soil158,287-297 '

Morel C., Tiessen H. and Stewart J.W.B. (1996) Sensitivity to Equilibration Periods of phosphate Sorption and Isotopic Exchange Methods Assessing Q/I Relationships in Soils. Soil Science 161,459-467.

Mortvedt J.J. (1991) Micronutrient fertilizer technology. In'Micronutrients in Agriculture'. (Ed. S.H. Mickelson) pp. 523-548. (Soil science Society of America, Inc.: Madison, WI)

Mortvedt J.J., Murphy L.S. and Follett R.H. (1999) 'Fertilizer Technology and Application.' (Meister Publishing C ompany: Ohio)

Murphy J. and Riley J.P. (1962) A Modified Single Solution Method for the Determination of Phosphate in Natural Waters. Analytica Chimica Acta 27 ,3I-36.

Nanny M.4., Minear R.A. and Leenheer J.A. (1997) Nuclear Magnetic Resonance Spectroscopy in Environmental Chemistry. In'Topics in Environmental Chemistry' (Ed. J.V/. Birks). (Oxford University Press: New Yorþ

195 Olson R.4., Army T.J., Hanway J.J. and Kilmer V.J. (1971)'Fertilizer Technology and Use.'(Soil Science Society of America: New YorÐ

Page A.L. (1982) 'Methods of Soil Anaþsis- Chemical and Microbiological Properties.' (ASA, SSSA. : Madison, V/I)

Parent L.E., Mackenzie A.F. and Perron Y. (1985a) Rate of P Uptake by Onions Compared With Rate of Pyrophosphate Hydrolysis in Organic Soils. Canadian Journal ofPlant Science 65, 1091-1095.

Parent L.E., Mackenzie A.F. and Perron Y. (19S5b) Influence of Four Levels of pyrophosphatase Activity on the Rate of Pyrophosphate Hydrolysis in Organic Soils. communications in soil science and Plant Analysis 16, 1095-1105. perkin-Elmer'Wallac 1414 Winspectral Liquid Scintillation Counter Manual. In. (Helsinki, Finland) philen O.D.J. and Lehr J.R. (1967) Reactions of Ammonium Polyphosphates with Soil Minerals. Soil Science Society of America Proceedings 31,196-199. pierzynski G.M., McDowellR.W. and Sims J.T. (2005) chemistry, cycling and potential Movement of Inorganic Phosphorus in Soils. In'Phosphorus: Agriculture and the Environment'. (Ed. A.N. Sharpley) pp. 53-86. (American Society of Agronomy, Crop Science Society of America, Soil Science Society of America: Madison, WI)

Randall p. (1997)'Root Activity and its Role inNutrient Aquisition and Cycling in Rotations- a Strategic Initiative.' GRDC.

Rashchi F. and Finch J.A. (2000) Poþhosphates: A Review. Their Chemistry and Application with Particular Reference to Mineral Processing. Minerals Engineering 13.

Rayment G.E. and Higginson F.R. (1992)'Australian Laboratory Handbook of Soil and Water Chemical Methods.'(Inkata Press: Melboume)

Romkens P.F.A.M. and Dolfing J. (1998) Effect of Ca on the Solubility and Molecular Size Distribution of DOC and Cu Binding in Soil Solution Samples. Env ir onm ent al S c i enc e & T e chno I o gy 32, 3 63 -3 69 -

Ryan J., Curtin D. and Cheema M.A. (1935) Significance of Iron Oxides and Calcium Cârbonate Particle Size on Phosphate Sorption by Calcareous Soils' Soil Science Society of America Journal 49, 7 4-7 6.

Sample E.C., Khasawneh F.E. and Hashimoto I. (1979) Reactions of Ammonium Ofth;- and Pollphosphate Fertilizers in Soil: III. Effects of Associated Cations- Soil Science Society of America Journal43, 58-65.

196 Sample E.C., Soper R.J. and Racz G.J. (1930) Reactions of Phosphate Fertilizers in soils. In'The Role of Phosphorus in Agriculture'. (Ed. F.E. Khasawneh, Sample, E.C. and Kamprath,E.J.). (American Society of Agronomy Inc., Crop Science Society of America Inc., Soil science society of America Inc.:'wisconsin, u.S.A.)

Savant N.K. and Racz G.J. (1972) Hydrolysis of Sodium Pyrophosphate and Tripollphosphate by Plant Roots. Journal of soil science ll3, 18-22.

Savant N.K. and Tambe K.N. (1979) The Langmuir Parameters of Orthophosphate and pyrophosphate Sorption for Ammoniated Tropical Soils. Communications in Soil Science and Plant Analysis 10, 503-51 1.

Shand C.4., Smith S., Edwards A.C. and Fraser A.R. (2000) Distribution of phosphorus in Particulate Colloidal and Molecular-Sized Fractions of Soil Solution' Water Res earch 34, I27 8-1284.

Smernik R.J. and Oades J.M. (2000) The Use of Spin Counting for Determining Matter l.Model euantitation in Solid State 13C NMR Spectra of Natural Organic Systems and the Effects of Paramagnetic Impurities. Geoderma 96' l0t-129'

Smernik R.J. and Oades J.M. (2000 b) The Use of Spin Counting for Determining Matter 2.rIF- Quantitation in Solid State 13C NMR Spectra of Natural Organic Treated Soil Fractions. Geoderma 96, 159-17 I -

Smemik R.J., Baldock J.4., Oades J.M. and Whittaker A.K. (2002) Determination of TlpH Relaxation Rates in Charred and Uncharred Wood and Consequences for NMR 50 -7 0. Quãntitatio n. Solid St at e Nucl ear Magnetic Res onance 22, ttN 15N- Smernik R.J. and Baldock J.A. (2005) Solid-State NMR Anaþsis of Highly Enriched Plant Matenals. Plant and Soil275,21l-283. ItN Smernik R.J. and Baldock J.A. (2005 b) Does Solid-State NMR Spectroscopy Detect All Soil Organic Nitrogen? Biogeochemistry 75,507-528'

Soil Survey S. (1992)'Keys to Soil Taxonomy.'(Pocahontas Press: Blacksburg, Virginia)

Stevenson F.J. and Cole M.A. (1999) The Phosphorus Cycle. In'Cycles of Soil- carbon, Nitrogen, Phosphorus, sulfur, Micronutrients'. (Eds F.J. Stevenson and M.A. Cole) pp.279-329. (John Wiley and Sons: New Yorþ

Subbarao 1'.V. (1975) 'Poþhosphate Hydrolysis and Reaction Products in Soils and S olutions.' (Kansas State University: Kansas)

Subbarao Y.V., Ellis R.J., Paulsen G.M. and Pakstelis J.V. (1977) Kinetics of Pyro- and Tripoly,phosphate Hydrolyses in the Presence of Com and Soybean Roots as Determlned Uy ¡Un Spectroscopy. Soil Science Society of America Journal 41,316- 318.

t97 Sutton C.D. and Larsen S. (1964) Pyrophosphate as a Source of Phosphorus for Plants. Soil Science 99,196-201.

Sutton C.D., Gunary D. and Larsen S. (1966) Pyrophosphate as a Source of Phosphorus for Plants: tr. Hydrolysis and Initial Uptake by a Barley Crop' Soil Science l0l,199-204.

Svoboda L. and Schmidt U. (1997) Single-column Ion Chromatography of Linear Polyphosph ates. Journal of Chromatography 767 , 107 -lt3 '

TabatabaiM.A. (1932) Soil Enzyrnes. In'Methods of Soil Analysis- Chemical and Microbiological Properties'. (Eds A.L. Page,R.H. Miller andD.R. Keeney) pp' 903- 943. (American Society of Agronomy Inc.: Madison)

Talibudeen O. (1957) Isotopically Exchangeable Phophorus in Soils. Part tr. Factors Influencing the Estimation of 'labile'Phosphorus. Journal of Soil Science 8, 86-96.

Tinker p.B. (1980) Role of Rhizosphere Microorganisms in Phosphorus Uptake by plants. In'The Role of Phosphorus in Agriculture'. (Eds F.E. Khasawneh,E.C. Sample andE. Kamprath). (American Society of Agronomy: Madison)

Tisdale S.L., Nelson W.L. and Beaton J.D. (1985) 'Soil Fertility and Fertilizers-' (MacMillan Publishing Co.: New York, U.S.A.)

Torres-Dorante L.U., Claassen N., Steingrobe B. and Olfs H.-W. (2005) Hydrolysis Rates of Inorganic Poþhosphates in Aqueous Solution as well as in Soils and Effects on P Availability. Journal of Plant Nutrition and Soil Science 168,352-358'

Tumer B.L. and McKelvie I.D. (2002) A Novel Technique for the Pre-Concentration and Extraction of Inositol Hexakisphosphate from Soil Extracts with Determination by phosphorus-31 Nuclear Magnetic Resonance. Journal of Environmental Quality 31,466-470.

Turner B.L., Cade Menun B.J. and Westermann D'T. (2003a) Organic Phosphorus Composition and Potential Bioavailability in Semi-Arid Arable Soils of the'Western united states. soil science society of America Journal 67,1168-1179.

N. and Condron L.M. (2003b) The Phosphorus Composition of Turner B.L., Mahieu 3rP Temperate pasture Soils determined by NaCiff-POTA Extractìon and Solution NMR Spectroscop y. Organic Geochemistry 34, I 199 -l2IO'

USDA (1932) Particle size anaþse. In'Procedures for Collecting Soil Samples and Methods of Analysis for Soil Survey'. (Ed. S.C. Service). (Washington, D.C')

Ussiri D.A. and Johnson C.E. (2004) Sorption of Organic Carbon Fractions by Spodsol Mineral Horizons. Soil Science Society of America Journal68,253-262-

Van Loon T. (2001) 'A Study on Liquid Technology and its Future Role in Influencing the Nutritional Outcomes on the Calcareous Soils in Southern Australia.' Australian Nuffreld Farming Scholarship Trust, Adelaide'

198 Van Wazer J.R., Griffith E.J. and McCullough J.F. (1952) Hydrolysis of Condensed Phosphates . Journal of the American chemical society 74,4977-4978.

Van'Wazer J.R., Griffith E.J. and McCullough J.F. (1955) Structure and Properties of the condensed Phosphates. vII. Hydrolytic Degradation of Pyro- and Tripolyphosphate. Jòurnal of the American Chemical Society 77,287-291.

Walbridge M.R. and Vitousek. p.M. (19SJ) Phosphorus Mineralisation Potentials in 32POq3- Acid orlanic Soils: processes Affecting Isotope Dilution Measurements. Soll Biologt & Biochemistry 19, 7 09 -7 17 .

Wallac (1939) Wallac 1414 Winspectral Liquid Scintillation Counter Manual. In- (Helsinki, Finland)

Wang T. and Li S.F.Y. (1999) Separation of Inorganic Phosphorus-containing Anions by cãpi[ary Electrophorcsis. Journal of chromatography A 834,233-241.

32P and Waser N.A.D. and Bacon M.P. (1995) Wet Deposition Fluxes of Cosmogenic 33Pf2P 33p and Variations in the Ratios at Bermuda. Earth and Planetary Science Letters 133, 71-80.

ZarcinasB.A., Mclaughlin M.J. and Smart M.K. (1996) The Effect of Acid Digestion Technique on the Perfõrmance ofNebulizationsystems Used in Inductively Coupled plasma Spectrometry. Communications in Soil Science and Plant Analysis 27,l33l- t3s4.

Soils Zhou M. and Li Y. (2001) Phosphorus-Sorption Characteristics of Calcareous and Limestone from the Southern Everglades and Adjacent Farmlands. Soil Science Society of America Journal 65, 1404-1412 -

zinderB.,HertzJ. and oswald H.R. (1984) Kinetic Studies on the Hydrolysis of Sodium Tripollphosphate in Sterile Solution. Water Research 18, 509-512'

t99 Appe ndix A- Rad Methodolo

Fraction Collection prior to the radioassay it was necessary to separate out the filtered soil solutions into orthophosphate (OP) and pyrophosphate (PP) fractions using ion chromatography

32OP (IC) with fraction collection. The measured in the OP fraction was assumed to

32PP, be hydrolysed from and the33oP in the OP fraction was assumed to be from 33OP native Op. In the pP fraction'2PP was measured, and counts for were negligible.

to A Pharmacia LKV Super Frac@ (Michigan, USA) fraction collector was set up collect the post- IC column fractions. Fractions were collected every 3'55 minutes with 4 fractions collected per sample. The second fraction of each sample contained the solution from the OP peak and the third fraction of each sample contained the

that solution from the PP peak. Fractions 1 and 4 were periodically checked to ensure

1.33 mL no counts were carried over into the subsequent fraction. Each fraction was

in volume.

Radioss ay of 32 P and 33 P

and Where possible disposable plastic equipment was used to minimise contamination 32P adsorption of the radioisotopes to glass surfaces. In all cases and33P activities

were measured by tiquid scintillation counting using a'Wallac Winspectral alþ l4l4

Liquid Scintillation Counter with Ecoscint A@ (chiral phenyl alkane) as a scintillation

liquid.

Radioactivity lntroduced

The radioactivity introduced (R) was determined by the measurement of the

exactly radioactivity of a spiked blank (water) solution. These samples were treated in

the same r'vay aS treaûnent samples and there were three replicates'

200 Quenching Gorrections

240 kBq Ten quenching calibration solutions were prepared. Each solution contained

33oP/sample. 2 32pplsampl e or 203 kBq The volume of solution was made to exactly

provide mL with 0, 50, 100, 150, 200,250,300, 350, 400 and 450 1tL of humic acid to

Ecoscint@ A, a colour (quench) gradient. Ten mL of scintillant (National Diagnostics

32P 33P was Georgia, USA) was added to this solution. Activity of and in each solution By determined in a Wallac'Winspectral alþ t4l4 Liquid Scintillation Counter.

(where specifying separate energy counting windows (channels) for each isotope of P

32P an 33P is channel 1 and is channel2), andusing a spectrum quench parameter for external standard (SQPE), the quench calibration curve correction or finetuning library was created (Perkin-Elmer)'

there The radioactivity in each of the quench curve solutions was measured where

33P containing was a quench curve series containing only (Figure 1-A) and a series

33P only 32p (Figure 1-B). Spillover for the channel (channel 1) is shown in channel2

32P (Figure 1-A), while spillover for (channel2) is shown in channel 1 (Figure 1-B).

32P to While there was significant spillover for at low levels of SQPE, it is important

note that this is atthehighest concentrations of humic acid added, and the colour of

the these solutions was much darker than any of the treatment samples. Therefore

same spillover would not have occurred in the experimental samples'

20r A- 33P a channel 1 o channel 2 2500 2000 o >= .O 1 500 O::(5> t .9d 1000 EE a É. 500 0 400 500 600 700 800 900 SQPE

channel I B- 32P a tr channel 2 5000 >= 4000 tfo od tr 3000 o> tr EÈ 2000 or.l(E= a É. 1000 0 4oo 5oo 600 700 800 900 SQPE

33P 2: Measurement of spillover from channel 1 into channel 2 for radioassy of Figure 32P. (A) and from chãnnel 2 into channel I for radioassay of Increasing concentration of humic acid for colour quenching resulted in decreasing spectrum quench parameter for an external standard (SQPE) with the colour range much greater than that of experinental samples.

Radioassay of Experimental Samples

A subsample of 1 mL of the OP or PP fraction with 1 mL of water and 10 mL of

National Diagnostics Ecoscint@ A (Georgia, USA) scintillant were placed in a plastic

32P macro scintillation vial (Kartell). The and33P activity in the filtrates were measured using a'Wallac Winspectral ulþ l4l4 Liquid Scintillation Counter. All counts were coffected for decay and dilution before being entered into calculations.

Recovery

The theoretical post-column recovery of the pre-column radioactivity is 13.7% due to the dilutions that occur in the IC analysis. The pre-column radioactivity had a dilution of 100 uL in 2000 uL: 0.05 (spike dilution). The post column counts had a 25 tL

202 : injection measured in a 1350 uL fraction collected post-analysis 0.0185, where

1000 uL of this fraction was removed for radioassay (0.7 of 13S0¡: 0.0137 with a further I in2 dilution for counting : 0.00685, so 0.00685/ 0'05:0 '137 , giving a theoretical post-column recovery of l3.7Yo. The measured post-column counts as 33OP compared to the pre-column counts had arecovery of 20.26Yo to 23.47Yo for and

32PP, 3.4o/a a recovery of I4.28o/oto 24.72o/o for with a coefficient of variation of fot

32PP. 33oP and 15.4% for

203 pendix B- Publication

McBeath T.M., Smernik R.J., Lombi E., and Mclaughlin M.J. (2006) Hydrolysis of Pyrophosphate in a Highly Calcareous Soil: A Solid-State Phosphorus-31 NMR Study. Soil Science Society of America Journal 70, 856-862.

204