<%BANNER%>

Phosphorus Sorption and Desorption in a Brazilian Ultisol: Effects of pH and Organic Anions on Phosphorus Bioavailability


PAGE 1

PHOSPHORUS SORPTION AND DESORP TION IN A BRAZILIAN ULTISOL: EFFECTS OF pH AND ORGANIC ANION S ON PHOSPHORUS BIOAVAILABILITY By SHINJIRO SATO A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2003

PAGE 2

Copyright 2003 by Shinjiro Sato

PAGE 3

This Ph.D. dissertation is dedicated to my life mentor, Dr. Daisaku Ikeda; and to all men, women, and children in the Amazon.

PAGE 4

ACKNOWLEDGMENTS It is my honor to express my sincere gratitude to my chairman, advisor, and friend, Dr. Nicholas Comerford for giving me proper directions for this research, insightful advice, and warm encouragement throughout my Ph.D. program. My special thanks go to Dr. Brian McNeal because it was through his classes that I became fascinated in Soil Science. I would like to thank Dr. Willie Harris for his openness and patience to even my stupid questions, especially regarding soil mineralogy. I would like to acknowledge Dr. Dean Rhue for his keen insights in soil P chemistry and for his kindness in letting me use his calorimeter. Dr. P.K. Nair helped me develop a scientific way of thinking early in my Ph.D. program, for which I am greatly thankful. The financial support for this research was provided by a graduate assistantship in the Soil and Water Science Department. I am extremely grateful to Dr. Randal Brown for this great opportunity. The Researcher Support Scholarship Program of the Soka Alumni Association in Soka University, Japan provided an additional support. I am grateful for all the support that I had from Mary McLeod, Miranda Lucas, and Adriana Chagas in the Forest Soils laboratory; and from Dr. Dean Rhue and Bill Reve with the calorimetry study. I cannot adequately thank Dr. Paulo Gabriel Nacif for collecting soil samples from Bahia, Brazil. Yongsung Joo very generously provided me with statistical help. All of this help and support made my dissertation a reality. I enjoyed all of my old and new friendships (especially those with my fellow graduate students) both on and off campus during this program. I was cheered and iv

PAGE 5

motivated by the spiritual encouragement that I received from all of my beloved SGI family members in Brazil, Japan, and the United States of America. Finally but most importantly, I would like to thank my parents and brother for their infinite love, understanding, and support in every form for so long. Without them, none of this work was possible. v

PAGE 6

TABLE OF CONTENTS page ACKNOWLEDGEMENTS...............................................................................................iv LIST OF TABLES .............................................................................................................ix LIST OF FIGURES .............................................................................................................x ABSTRACT ......................................................................................................................xii CHAPTER 1 GENERAL INTRODUCTION....................................................................................1 Forms of P in Soils.......................................................................................................1 Soil Organic P Solid Phase.................................................................................1 Soil Organic P Solution Phase............................................................................3 Soil Inorganic P Solution Phase.........................................................................3 Soil Inorganic P Solid Phase..............................................................................4 Factors Affecting Inorganic P Sorption and Desorption..............................................5 Phosphorus Desorption Defines P Bioavailability.......................................................8 Classification of Desorbable Inorganic P.....................................................................9 Importance of Phosphorus Bioavailability in Tropical Soils......................................11 2 ASSESSING A QUICK METHOD FOR DEVELOPING PHOSPHORUS DESORPTION ISOTHERMS USING ANION EXCHANGE MEMBRANES.......16 Introduction.................................................................................................................16 Materials and Methods...............................................................................................19 Soil Material........................................................................................................19 Phosphorus Desorption Kinetics Using AEMs...................................................20 Phosphorus Desorption as Affected by AEM Surface Area...............................21 Phosphorus Desorption Isotherms Using the Multiple AEM Method................21 Phosphorus Desorption Isotherms Using the Sequential AEM Method.............22 Statistical Analysis..............................................................................................23 Equation for K d values.........................................................................................24 Results.........................................................................................................................24 Phosphorus Desorption Isotherms: Kinetics and Effect of AEM Surface Area.....................................................................................................24 Phosphorus Desorption Isotherms: Multiple Method vs. Sequential AEM Method...................................................................................................24 vi

PAGE 7

Phosphorus Desorption Isotherms: K d values.....................................................25 Discussion...................................................................................................................25 3 INFLUENCE OF SOIL pH ON INORGANIC PHOSPHORUS SORPTION AND DESORPTION IN A HUMID BRAZILIAN ULTISOL..................................34 Introduction.................................................................................................................34 Effect of Soil pH on P Sorption...........................................................................34 Effect of Soil pH on P Desorption.......................................................................35 Materials and Methods...............................................................................................38 Soil Material and Its Characterization.................................................................38 Phosphorus Sorption Isotherms...........................................................................39 Phosphorus Desorption Isotherms Using the Multiple AEM Method................39 Equation for P Sorption and Desorption Isotherms.............................................41 Statistical Analysis..............................................................................................42 Results.........................................................................................................................42 Phosphorus Sorption pH Envelope......................................................................42 Phosphorus Desorption pH Envelope..................................................................43 Discussion...................................................................................................................45 Soil pH Effect on P Sorption...............................................................................45 Soil pH Effect on P Desorption...........................................................................46 4 INFLUENCE OF LOW MOLECULAR WEIGHT ORGANIC ANIONS ON PHOSPHORUS DESORPTION AND BIOAVAILABILITY IN A HUMID BRAZILIAN ULTISOL.............................................................................................58 Introduction.................................................................................................................58 Materials and Methods...............................................................................................64 Soil Material........................................................................................................64 Standard P Fertilization and LMWOAs Extraction Methods..............................64 Experiments.........................................................................................................66 Experiment 1. Optimum Time for Extraction of Ligand-Desorbable P (Objective 1)...........................................................................................66 Experiment 2. Optimum Anion Concentration for Extraction of Ligand-Desorbable P (Objective 1)........................................................66 Experiment 3. Short Kinetic Extraction of Native Soil with Oxalate (Objective 2)...........................................................................................66 Experiment 4. One-minute Extraction of Native and Fertilized Soils with Oxalate and Citrate at Lower Anion Concentrations (Objective 2)...........................................................................................66 Experiment 5. Seven-hour Extraction of Native Soil with Oxalate and Citrate at Lower Anion Concentrations (Objective 2).....................67 Experiment 6. Differentiation of Desorbable P into Disequilibriaand Ligand-Desorbable P (Objective 3)........................................................67 Experiment 7. Mass Balance Verification for Ligand-Desorbable P (Objective 3)...........................................................................................68 Experiment 8. Investigation of P Immobilization after Sorption.................68 vii

PAGE 8

Statistical Analysis..............................................................................................69 Results and Discussion...............................................................................................69 Ligand-Desorbable P...........................................................................................69 Ligand-Exchangeable P vs. Ligand-Dissolvable P.............................................70 Disequilibria-Desorbable P vs. Ligand-Desorbable P.........................................72 Phosphorus Immobilization/Fixation after Sorption...........................................73 5 SUMMARY, CONCLUSIONS, AND FUTURE RESEARCH.................................85 Summary.....................................................................................................................86 Suitability of Using Multiple AEM Strips to Develop Desorption Isotherms..........................................................................................................86 Comparison of Multiple AEM Method to Sequential AEM Method..................86 Effect of Soil pH on Phosphorus Sorption..........................................................87 Effect of Soil pH on Phosphorus Desorption......................................................87 Method for Measuring Ligand-Desorbable Phosphorus.....................................87 Separating Ligand-Desorbable P into Ligand-Exchangeable P and Ligand-Dissolvable P.......................................................................................88 Differentiating Bioavailable Pools of Applied P into Disequilibria-Desorbable P and Ligand-Desorbable P....................................89 Conclusions.................................................................................................................89 Future Research..........................................................................................................90 APPENDIX A ACID-BASE TITRATION CURVE FOR THE SOIL...............................................92 B DETERMINATION OF SOLUBLE P IN THE PRESENCE OF ORGANIC ANIONS.....................................................................................................................93 C SURFACE REACTIONS OF PHOSPHATE, OXALATE, AND CITRATE MEASURED BY FLOW CALORIMETRY..............................................................94 LIST OF REFERENCES...................................................................................................98 BIOGRAPHICAL SKETCH...........................................................................................111 viii

PAGE 9

LIST OF TABLES Table page 1-1 Types and mechanisms of inorganic P desorption...................................................15 2-1 The amount of P sorbed for each desorption method at 3 different initial P-addition levels.......................................................................................................29 2-2 ANOVA table for the GLM model evaluating the P desorption isotherms.............31 2-3 Estimates for parameters of the GLM model used to test desorption isotherm differences................................................................................................................32 3-1 Surface soil (0 cm) properties for the Atlantic Forest soil, Bahia, Brazil..........50 3-2 Langmuir parameters for P sorption isotherms at 3 different soil pH levels...........52 3-3 Langmuir parameters for P desorption isotherms at 3 different soil pH levels with 3 different initial P-addition levels...................................................................55 3-4 Total amount and ratio to P sorbed of P desorbed at 3 different soil pH levels with 3 different initial P-addition levels...................................................................56 4-1 Disequilibriaand ligand-desorbable pools of P when 3 levels of P are sorbed to the soil..................................................................................................................82 B-1 Effect of varying concentrations of oxalate and citrate on absorbance over a range of standard P concentrations...........................................................................93 ix

PAGE 10

LIST OF FIGURES Figure page 1-1 Influence of pH on the distribution of orthophosphate species in solution..............14 2-1 Kinetics of P desorption using AEMs......................................................................28 2-2 Phosphorus desorption with increasing number of AEM strips using a 2 h shaking time.............................................................................................................28 2-3 Phosphorus desorption isotherms using 2 desorption methods with 3 initial P-addition levels. A) 180 g P g B) 80 g P g C) 40 g P g .........................30 2-4 Relationship between observed C s values and C s values predicted from the P desorption GLM model.........................................................................................33 3-1 Phosphorus sorption isotherms at 3 different soil pH levels with equilibrium concentration of A) 0 g P mL and B) 0.0.0 g P mL ...............................51 3-2 Partition coefficients relative to equilibrium P concentration for P sorption isotherms at 3 different soil pH levels......................................................................53 3-3 Phosphorus desorption isotherms at 3 different soil pH levels with 3 different initial P-addition levels. A) 180 g P g B) 80 g P g C) 40 g P g ...............54 3-4 Partition coefficients relative to equilibrium P concentration for P desorption isotherms at 3 different soil pH levels with 3 different initial P-addition levels. A) 180 g P g B) 80 g P g C) 40 g P g ......................................................57 4-1 Ability of three organic anions to complex Fe as a function of pH as predicted using Geochem-PC (Parker et al. 1995)...................................................................75 4-2 Phosphorus-desorption and Fe-release kinetics by organic anions when added at a loading of 0.02 mmol g soil. A) native soil. B) P-fertilized soil.....................76 4-3 Effect of organic anion loadings for a 7 h interaction time on P desorption and Fe release. A) native soil. B) P-fertilized soil....................................................77 4-4 Effect of organic anion loading on the molar ratio of P desorption to Fe release from native soil.........................................................................................................78 x

PAGE 11

4-5 Iron release by 3 oxalate loadings as a function of shaking time. The horizontal line represents exchangeable Fe level (0.05 mol Fe g soil).................................78 4-6 Effect of low anion loadings for a 1 min interaction time on P desorption and Fe release. A) native soil. B) P-fertilized soil....................................................79 4-7 Effect of low anion loadings for a 7 h interaction time on P desorption and Fe release..................................................................................................................80 4-8 Cumulative P desorbed by AEMs from soil at 3 increasing levels of P. Each desorption step represents one day...........................................................................81 4-9 Distribution of sorbed P between the disequilibriaand ligand-desorbable pools. The first bar in each P-added level shows a combination of disequilibriaand ligand-desorbable pools; and the second bar shows total desorbable P 1 d after P sorption..............................................................................83 4-10 Cumulative P immobilization after P sorption. +S and S are with and without shaking; and +H and H are with and without mercuric chloride for microbial suppression...............................................................................................................84 4-11 Cumulative P immobilization versus the square root of time. Immobilized P is that amount of P immobilized after the effects of microbes and shaking have been removed...................................................................................................84 A-1 Titration curve for the Atlantic Forest soil...............................................................92 C-1 Heats of reaction for phosphate sorption onto the study-site soil............................95 C-2 Heats of reaction for oxalate sorption onto P-saturated soil preceding phosphate sorption onto oxalate-treated soil............................................................96 C-3 Heats of reaction for citrate sorption onto P-saturated soil preceding phosphate sorption onto citrate-treated soil.............................................................97 xi

PAGE 12

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy PHOSPHORUS SORPTION AND DESORPTION IN A BRAZILIAN ULTISOL: EFFECTS OF pH AND ORGANIC ANIONS ON PHOSPHORUS BIOAVAILABILITY By SHINJIRO SATO December, 2003 Chair: Nicholas B. Comerford Major Department: Soil and Water Science Phosphorus bioavailability is a major limitation to plant productivity in weathered Brazilian soils. Soil pH change and organic anions such as oxalate and citrate produce known rhizosphere effects that influence P sorption and desorption. However, few data document their influence on P bioavailability. Quantitative measurement of bioavailable P pools is not well-defined. Our study focused on the surface soil developed under native Atlantic Forest in Bahia, Brazil. Objectives of our study were as follows: To investigate a method of developing P desorption isotherms using multiple strips of anion exchange membranes (AEMs) To evaluate the effect of pH on P sorption and desorption isotherms To separate ligand-desorbable P into ligand-exchangeable and ligand-dissolvable P To estimate the relative contributions of disequilibriaand ligand-desorbable P to the total desorbable P. The Multiple AEM Method was suitable for developing well-defined isotherms and superior to sequential extraction methods over a range of solution concentrations in our xii

PAGE 13

study. Increasing pH from 4.7 to 6 and 7 decreased P sorption (up to 21% and 34%, respectively) and increased P desorption for all P application rates. Using less than 1 mol oxalate or citrate g soil and using 1 min contact time allowed us to desorb P solely by ligand exchange without surface dissolution. Nevertheless, dissolution played the major role in measuring ligand-desorbable P. Although most of P sorbed was found in the ligand-desorbable pool, P was preferentially found in the disequilibria-desorbable pool after P fertilization. The percentage of the total sorbed P that was desorbable was a function of time since fertilization; 13 to 22% on day 1 and 39 to 45% on day 14. This ageing process was due to a shaking artifact (30%), microbial immobilization (57%), and sorption of P onto the soil in a non-oxalate accessible form (13%). Liming and organic anion loading improved P bioavailability in this Brazilian soil. Long-term ageing and continuous exudation of organic anions also influence P bioavailability. Management options that recognize this will improve bioavailability of native and applied P sources in this important agricultural region of northeastern Brazil. xiii

PAGE 14

CHAPTER 1 GENERAL INTRODUCTION The management of tropical soils is crucial to addressing the issues of food security, soil degradation and environmental quality (Lal 2000). Low fertility; high soil acidity; and the role that land-use systems, fertilizers and liming play in mitigating these problems have been areas of interest for the past several decades. These problems have been addressed through the study of soil physical and chemical properties, which are of fundamental value when devising appropriate management approaches; and also through the social constraints that determine what management approaches are possible. Because the soil is highly weathered, many humid, acid, tropical soils are phosphorus (P) deficient. Our study focused on the P chemistry of a particular tropical soil in northeastern Brazil as it relates to P bioavailability. This chapter will summarize some pertinent aspects of P bioavailability as an introduction to the objectives and hypotheses of this dissertation. Forms of P in Soils Soil Organic P Solid Phase Organic soil phosphorus represents, on average, one-half of the total soil P and varies between 15 and 80% in most soils (Stevenson 1986). The amount of soil organic P depends on soil organic matter content; therefore, it tends to be high in the surface horizon and decreases with soil depth. Soil organic P compounds are synthesized by soil microorganisms as esters of orthophosphoric acid (Anderson 1980), with the most stable 1

PAGE 15

2 and abundant compounds being inositol phosphates which make up 10 to 50% of the total soil organic P. Phosphorus held in organic form must first be converted to inorganic form before plant uptake. The process is controlled by P mineralization and immobilization in a manner similar to processes described for release of organic nitrogen and sulfur from soil organic matter. Mineralization of organic P may be both biological and biochemical (Tate 1984). Biological mineralization releases P bound in organic matter as soil microorganisms mineralize soil organic matter. Biochemical mineralization occurs when plant roots and microorganisms produce phosphatase enzymes that can catalyze the hydrolysis of ester bonds between P and C compounds. Phosphatases play a major role in mineralization of soil organic P. Phosphatase activity is generally enhanced in P-stressed soils (Clarholm 1993) and evidently have both spatial and temporal variability (Schneider et al. 2001). Since soil microbial activity is dependent on soil organic carbon content, the C:P ratio is critical in balancing P mineralization and immobilization processes. It has been proposed that critical C:P values for decomposing residues are < 200:1 for net mineralization and > 300:1 for net immobilization (Enwezor 1967, Sharpley 1985). The C:P ratio of the substrate being decomposed may be more important than that of soil organic matter or plant residues, with the critical C:P ratio calculated from microbial growth yield falling in the range of 50 to 70 (White 1981). Besides P fertilization status, mineralization of organic P is influenced by many of the same soil properties that control decomposition of soil organic matter: namely, temperature, moisture, aeration, pH, and cultivation method and/or intensity. Therefore,

PAGE 16

3 in soils of tropical regions where climatic conditions favor organic matter decomposition, P mineralization can supply a portion of the crop P requirement (Anderson 1980). Inorganic P also can be immobilized in organic form as part of the microbial biomass. The quantity of P immobilized varies from 6 to 106 g P g soil (Brookes et al. 1984, Joergensen et al. 1995), and from 38 to 46% of sorbed inorganic P for some variable charge minerals (He and Zhu 1998). The amount of P immobilized by microbes was estimated in one case to be 3 to 5 times greater than the amount of P absorbed by plants in semiarid grasslands (Cole et al. 1978). Soil Organic P Solution Phase Uptake of soluble organic P by plant roots is not well-documented. However, it has been shown that the presence of (endoand ecto-) mycorrhizal fungi stimulates net mineralization of organic P because of phosphatase production (Condron et al. 1996, Joner et al. 2000). Red clover (Trifolium pratense L.) uptake of P increased as the plant and an associated arbuscular mycorrhizal fungus grew. Organic P sources (phytase, lecithin, and RNA) contributed (17 to 31%) to total plant P compatibly with an inorganic P source (KH 2 PO 4 23%) (Feng et al. 2003). Jayachandran et al. (1992) showed a similar effect of arbuscular mycorrhizal fungi associated with big bluestem (Andropogon gerardii Vit.), that promoted a significant increase in plant use of organic P (cytidine 3and 5-phosphate) when added to a soil having low P fertility. Soil Inorganic P Solution Phase Phosphorus concentrations in the soil solution are low compared to other soil macronutrients, normally ranging from 0.001 to about 1 g P mL with an average of about 0.05 g P mL (Barber et al. 1962). Plant roots and mycorrhizal hyphae absorb soil solution P as orthophosphate ions in either the H 2 PO 4 or HPO 4 2 form, depending on

PAGE 17

4 soil pH. The dominant solution inorganic P species below pH 7.2 is the monovalent ion; while HPO 4 2 dominates at pH values between 7.2 and 12.1 (Fig. 1-1). Plant uptake of the divalent ion has been shown to be slower than the monovalent ion (Cresser et al. 1993). Soil Inorganic P Solid Phase If not absorbed by the plant or immobilized by microbes, inorganic P in the soil solution is subject to various chemical reactions. These include precipitation as and dissolution from inorganic P solids, and adsorption to and desorption from soil mineral surfaces. Precipitation is a slow reaction in which phosphate ions react with dissolved Al 3+ Fe 3+ and Mn 3+ in acid soils and Ca 2+ in neutral to calcareous soils to form insoluble hydroxy phosphate precipitates (Sanchez and Uehara 1980). Variscite (AlPO 4 H 2 O) and strengite (FePO 4 H 2 O) are the most common stable minerals found in acid soils; while in alkaline soils, the most stable minerals are calcium phosphates such as octacalcium phosphate [Ca 8 H(PO 4 ) 6 H 2 O] and hydroxyapatite [Ca 5 (PO 4 ) 3 OH]. The solubilities of these Ca-P minerals decrease with increasing pH; whereas Aland Fe-P minerals exhibit increasing solubility with increasing pH (Lindsay and Moreno 1960). In contrast to precipitation, adsorption is a relatively fast initial reaction involving both anion and ligand exchange. Negatively charged phosphate ions in the soil solution are attracted to positively charged soil surfaces, via anion exchange. However, a significant quantity of positively charged surfaces will develop only at low soil pH in highly weathered soils that display variable-charge characteristics. Adsorption by ligand exchange results when an inner-sphere metalOP covalent bond is formed by replacement of hydroxyls on the edges of layer silicate clays and the surfaces of insoluble

PAGE 18

5 oxides of Al, Fe, and/or Mn. This process is referred to as specific adsorption or chemisorption (Schindler and Sposito 1991). Since it is difficult to distinguish adsorption and precipitation reactions from one another, they are usually referred to collectively as P sorption or fixation (Sposito 1984). Sorption of P seems to increase over time, apparently due to slow precipitation processes that are superimposed on the more rapid chemisorption (Van der Zee and Van Riemsdijk 1991). The amount of P in the soil solution appears to determine which reaction dominates. Chemisorption dominates at low solution P; while precipitation proceeds when the concentrations of P and associated cations in the soil solution exceed the solubility products (K sp ) of the relevant P compounds. Therefore, for soils exhibiting low solution P concentrations such as many tropical soils, chemisorption should be a dominant mechanism. Adsorbed or precipitated inorganic P undergoes desorption or dissolution reactions when moving from the solid to the solution phase. Desorption is of particular importance in bioavailability evaluations, because this process controls what the plant can eventually access. Desorption is used here to denote either the precise opposite of adsorption or the counterpart to sorption, which would include a combination of desorption and dissolution. Both soil and plant characteristics affect the extent and rate of desorption and dissolution for inorganic soil P. It is necessary, therefore, to understand these relationships in order to evaluate and predict P bioavailability and the fate of inorganic P. Factors Affecting Inorganic P Sorption and Desorption The soil characteristics that influence P fixation include the amount and type of clay-fraction minerals, soil pH, soil organic matter content, soil temperature, time of reaction, exchangeable Al 3+ soil redox condition (Sanchez and Uehara 1980), and root

PAGE 19

6 exudates. These factors are interactive rather than additive, which makes it difficult to predict inorganic P fixation under a wide range of soil conditions. The higher the Al and Fe oxide contents of soil clay, and the less crystalline (more amorphous) the soil minerals, the greater an acid soils P fixation capacity. This is largely attributed to the greater surface area which these conditions represent (Fernandes and Warren 1994, Gilkes and Hughes 1994, Quintero et al. 1999). Higher clay contents also result in greater P fixation (Sanchez and Uehara 1980, Harris et al. 1996, Voundi et al. 1997). Soil pH, because of its influence on surface charge and the ease with which surface charge can be managed by liming, has been the focus of research on a variety of soils (Awan 1964, Parfitt 1977, Bar-Yosef et al. 1988, Naidu et al. 1990, Pardo et al. 1992, Rupa et al. 2001). In spite of this, results concerning P sorption or solution P concentrations as influenced by liming have been conflicting (Haynes 1982, Anjos and Rowell 1987). Given the P sorption patterns observed at different soil pH values, and the number of explanations used to account for these phenomena, it is clear that the influence of liming on P sorption is not well-understood. In highly weathered and acid soils, existing data support the hypothesis that exchangeable Al 3+ is the main factor determining the pattern of P sorption with changing pH (Lopes-Hernandez and Burnham 1974, Anjos and Rowell 1987, Chen and Barber 1990). Haynes (1984) said that one should expect liming to increase P sorption in soils that are initially high in exchangeable Al 3+ but to decrease P sorption in soils with low exchangeable Al 3+ content. When soils with low exchangeable Al 3+ are limed, the neutralization and precipitation of Al 3+ ion and of hydroxy-Al species to form Al

PAGE 20

7 hydroxide reduces the number of P-sorption sites. Where exchangeable Al 3+ is initially high, the formation of amorphous hydroxyl Al with highly active sorbing surfaces may exceed any decrease in the sorption capacity of the original sorbing surfaces, resulting in increasing P sorption as pH increases. Phosphorus desorption studies are rare compared to the frequency of P sorption studies. Some desorption studies determined bioavailable P by using agriculture-based extraction indices, and correlating them with plant P uptake (Parfitt 1979, Steffens 1994, Martin et al. 2002). Others described the kinetics of P desorption by extracting P using dilution methods or using anion-exchange resins. Such studies examined the suitability of relationships such as first-order, Elovich, and parabolic diffusion equations to describe kinetic P release from soil (Chien and Clayton 1980, Sharpley et al. 1981, Raven and Hossner 1994). Several studies focused on the effect of soil pH on P desorption (Hingston et al. 1974, Rupa et al. 2001). However, the results have been inconsistent; with some workers finding increasing P desorption with increasing pH (Madrid and Posner 1979, Cabrera et al. 1981, De Smet et al. 1998) and others with decreasing pH (Barrow 2002). He et al. (1989) reported that P desorption decreased until pH was raised to about 4.8; and then increased with further pH increases for most of the acid soils from China which contained high Fe and Al oxide and/or kaolinite levels. In contrast, for three representative surface soils of India, both the amount and rate of P desorption initially increased with pH increase from 4.25 to 5.5; and then decreased at higher pH values of 6.75 and 8.0 (Rupa et al. 2001).

PAGE 21

8 No relationship was found between P removal by various extractants and the affinity constant derived from the Langmuir equation for the original sorption of P (Kuo et al. 1988). It was found that the amount of P desorbed by all extractants was significantly related to the fraction of P saturation on soil surfaces, with higher P saturation of the surfaces yielding higher desorbed P values. Kafkafi et al. (1967) found, for kaolinite, that a portion of applied P was non-exchangeable during simple washing using an indifferent electrolyte solution. They concluded that some of the applied P was incorporated into a fixed sink. In another study, Cabrera et al. (1981) postulated that the inability to desorb all P sorbed onto Fe-oxides using strongly alkaline solutions (e.g., 0.1 M NaOH) was due to the existence of micropores in the materials, where some of the applied P could remain occluded. Plant roots are known to exude a range of organic acids including oxalate, malate, and citrate. Their positive effect on enhancing P desorption has been well-recognized (Jones 1998, Ryan et al. 2001). The mechanisms of P desorption by these compounds include ligand exchange and ligand dissolution; yet the evidence for which are active and in what proportion for a range of soils is not available. These processes and their relative contributions to changes in inorganic P bioavailability which can occur for the rhizosphere should vary with plant species, plant nutritional status, and ambient soil conditions (Hinsinger 2001). Phosphorus Desorption Defines P Bioavailability Prior studies on P sorption have outlined trends and provided a foundation for understanding such sorption. However, bioavailability of P is a function of P desorption; not P sorption. Specific information for soils, where needed for agricultural development, is lacking for P desorption and its trends with soil characteristics. Humid,

PAGE 22

9 weathered, tropical soils of Brazil represent one area where an understanding of P desorption dynamics is needed to better manage soils that are inherently P deficient. Understanding and evaluating soil P bioavailability requires that one identify Form(s) of P absorbed by plants Form(s) of P that can be released from the solid to the adjacent and bioavailable solution phase Processes responsible for the release of P from the solid phase to the solution phase How soil/plant/mycorrhiza properties affect these processes Soil properties that affect movement of P in solution to the root/mycorrhiza Plant properties that are important in determining P absorption by the root/mycorrhiza when P arrives at its surface (Comerford 1998). The P desorption process is poorly understood and virtually no such information exists for tropical soils. Phosphorus desorption must be studied separately from or in conjunction with P sorption, because there is a significant hysteresis between the two such that sorption data do not suitably relate to desorption behavior (Barrow 1983). To further confound the situation, the hysteresis that is commonly reported may be partially an artifact of the methodology used, since a recent study suggests that while sorption and desorption isotherms are strongly hysteretic; desorption and resorption isotherms (sorbing P after the desorption step) are not generally hysteretic (Barros et al. in press). Classification of Desorbable Inorganic P For the purpose of this dissertation, the term desorption will be used in the same fashion as the counterpart term sorption. It does not distinguish desorption from adsorption; or distinguish dissolution from precipitation. To clarify the types of desorption that occur in soils, the following categorization is suggested. Two major mechanisms describe inorganic P desorption: disequilibria desorption and ligand

PAGE 23

10 desorption. These can be subcategorized into exchange and dissolution types of desorption (Table 1-1). Disequilibria desorption is consistent with Le Chteliers Principle; and requires that a disequilibrium exist between sorbed and solution P concentrations. Disequilibria can be created during uptake of solution P by plant roots; during leaching; during precipitation as secondary P minerals; during a change in rhizosphere pH that changes surface charge or solubilities; or during immobilization by soil microbes. As the solution P concentration is decreased, P is replenished through desorption from P adsorbed onto the soil surface in exchange with other anions in the solution (disequilibria exchange); or through dissolution of relatively soluble P compounds (disequilibria dissolution) (Wolf and London 1994). Disequilibria-desorbable P can be measured in a specific time frame. It can be extracted by maintaining the solution P concentration at low levels through sequential extractions using anion-exchange resins (Sibbesen 1978, Saggar et al. 1990); anion-exchange membranes (Cooperband and Logan 1994); or iron oxide-impregnated paper strips (Menon et al. 1996/1997); or by shaking with water (Simard et al. 1994); or low-ionic-strength solutions (Rupa et al. 2001). This pool of desorbable P should be bioavailable to plants or soil microorganisms. The nature and pattern of this desorption, particularly when described as a desorption isotherm, is rare in the literature and non-existent for tropical soils. Another desorbable pool of inorganic P can be released by ligands that have a higher affinity for the soil surface than does phosphate. Ligands in solution form inner-sphere complexes with the soil surface that exchange with P on the surface, because of their higher affinity for the soil surface (ligand exchange) (Cambier and Sposito 1991,

PAGE 24

11 Hu et al. 2001). These ligands affect the stability of precipitates and contribute to dissolution of the solid, concomitantly releasing P from the surface (ligand dissolution) (Zinder et al. 1986, Ludwig et al. 1996). Plant roots and soil microbes produce a range of organic acids including acetate, oxalate, citrate, and malate. They are known to exude these acids into the soil solution at concentrations which can promote P desorption (Fox and Comerford 1990, Jones 1998). The problem in defining the pool of ligand-desorbable P is that one does not know with certainty which ligands are in the soil solution; their loading rate onto the solid phase; the volume of soil affected by this loading; and how soil properties affect the effectiveness of P release by these ligands (Comerford 1998). Therefore, the best approach to quantifying this pool is to measure the total amount of P that is desorbable by ligand desorption. Importance of Phosphorus Bioavailability in Tropical Soils Tropical soils are notorious for their low nutrient bioavailability (often related to low exchangeable base contents, high soil acidity, Al toxicity, and low P desorbability). Oxisols and Ultisols of humid, tropical Brazil are no exception. Low P bioavailability is the primary limiting factor for plant growth in many of these soils. The origin of P deficiencies in these soils is threefold (Brady and Weil 1999). First, the total P level of the soils is low, ranging from 0.01 to 0.1% (Chen and Ma 2001). Second, the P compounds commonly found in these soils are highly insoluble. Third, when soluble sources of P, including those in fertilizers and manures, are added to soils they are sorbed to the soil, changed into unavailable forms and, over time, incorporated into highly insoluble compounds.

PAGE 25

12 The high P fixing capacities of many tropical soils (Udo and Uzu 1972, Sanyal et al. 1993, Fontes and Weed 1996) are a function of their mineralogy and are a primary contributor to their P deficiencies. Therefore, it is critical to understand the physical and chemical properties of these soils that control the dynamics of inorganic P tie-up and supply. Better understanding should lead to better nutrient management and better maintenance of soil quality. This is particularly true for the acid, weathered soils of Brazil, where P chemistry is often poorly understood. Based on the above discussion, it is obvious that P desorption is an important topic affecting the nutrient management of P in soils, and particularly many tropical soils. Phosphorus desorption determines P bioavailability; and its interaction with soil pH in the form of desorption isotherms remains poorly documented. Bioavailable inorganic P exists in two major pools, but quantitative estimates of these pools (and methods for deriving these estimates) remain poorly developed. Therefore, our study will address some of these gaps in knowledge, using a humid, highly weathered Ultisol from the Atlantic Forest in northeastern Brazil. The following is a brief description of each chapter. Chapter 2 develops a method for constructing P desorption isotherms using anion exchange membranes (AEMs) in a way that minimizes time spent and experimental artifacts. This method is also compared to a more conventional method. Chapter 3 investigates the effect of soil pH on both inorganic P sorption and desorption isotherms. The hypothesis is that P sorption and desorption isotherms are functions of soil pH, with increasing pH decreasing the amount of P that is sorbed and increasing the amount of P that is desorbed. In both cases, more P would be bioavailable.

PAGE 26

13 Chapter 4 addresses the major types of P desorption (disequilibria desorption and ligand desorption); and attempts to differentiate between ligand-exchangeable desorption and ligand-dissolvable desorption for a Brazilian soil in its native and P-fertilized states. It also investigates the amount of P sorbed that cannot be accessed by the above two desorption mechanisms.

PAGE 27

14 H 2 PO 4 HPO 4 2 H 3 PO 4 100 PO 4 3 Concentration ( % ) 50 0 8 2 4 6 10 12 14 Solution pH Figure 1-1 Influence of pH on the distribution of orthophosphate species in solution

PAGE 28

15 Table 1-1 Types and mechanisms of inorganic P desorption Major type Reaction type Mechanism Disequilibria Follows Le Chteliers Principle. Requires P disequilibria between solution and solid phases, created by plant uptake, leaching, precipitation, rhizosphere pH change, and/or microbial immobilization. Exchange Desorbs P from the soil surface in exchange with other anions in the soil solution. Dissolution Desorbs P from relatively soluble P compounds. Ligand Influenced by anions of low-molecular-weight organic acids such as acetate, oxalate, citrate, and malate, which are exuded by plant roots and soil microbes. Ligands have higher affinity for the soil surface than phosphate and/or destabilize P minerals. Exchange Desorbs P from the soil surface by forming inner-sphere complexes which exchange with P. Dissolution Desorbs P from the soil surface by destabilizing and dissolving the solid itself.

PAGE 29

CHAPTER 2 ASSESSING A QUICK METHOD FOR DEVELOPING PHOSPHORUS DESORPTION ISOTHERMS USING ANION EXCHANGE MEMBRANES Introduction Inorganic phosphorus (P) bioavailability is defined by the combined processes of desorption and dissolution. These processes occur in response to disequilibria established in solution by removal of P from solution (disequilibria desorption); or in response to the action of ligands exchanging with phosphate or dissolving phosphate-bearing compounds (ligand desorption) (Table 1-1). Most commonly, bioavailable desorbable P is indexed by extracting a portion of the labile pool of P with chemical extractants such as Bray 1, Mehlich 1 or 3, or Olsens Solution (Ziadi et al. 2001, Csatho et al. 2002). However, these soil P tests are interpretable only when combined with crop experimental data from the field or greenhouse (Kumar et al. 1992, Kleinman et al. 2001); and do not measure a pool of bioavailable P, but instead act simply as an index to P bioavailability. There is no absolute interpretation for the quantity of P extracted by any of these methods; nor do these methods define the pattern of P release, or the influence of P desorption on the soil solution concentration. Therefore, such techniques have limited utility in the mechanistic modeling of P bioavailability and uptake by plants. Desorption can be described by a desorption isotherm, which is the relationship between P on the solid phase and P in the solution phase. This relationship has been referred to as a quantity (solid P)/intensity (solution P) relationship (Raven and Hossner 1993). The P desorption isotherm is a useful, descriptive characteristic of soil that 16

PAGE 30

17 provides information on the ability of the soil to release P into solution as solution P is removed. This ability is represented by the slope of the desorption isotherm which is termed the partition coefficient (K d ). Partition coefficients are directly related to the buffer power of the soil and are used to calculate soil diffusion coefficients (Van Rees et al. 1990b). Diffusion is important because, in many soils, it is the mechanism governing 90 to 98% of the P supply to roots (Barber 1980). Therefore, K d is a useful parameter in computer-based nutrient uptake models. However, few comprehensive studies address P desorption; and fewer address development of P desorption isotherms. Phosphorus is desorbed from soil, and thus P desorption isotherms can be developed by techniques that use dilution, sequential extraction, or anion exchange resin extraction (Brewster et al. 1975, Bhatti and Comerford 2002). For the dilution method, P is desorbed from the solid phase by shaking a soil sample over a range of soil:solution ratios for a specified equilibration time with single or successive extractions (Madrid and Posner 1979, Sharpley et al. 1981). In the sequential extraction, an extracting solution is added to the soil at a constant soil:solution ratio; and the sample is shaken for a constant equilibration time. This procedure is repeated until P desorption is exhausted or a pattern of P release is established (Fox and Kamprath 1970, Okajima et al. 1983, Bhatti and Comerford 2002). The anion exchange resin extraction is performed in the same manner as the sequential extraction except that bags filled with anion exchange resins (AERs) are used to enhance the removal of P from solution (Sibbesen 1977, Bache and Ireland 1980, Yang and Skogley 1992, Delgado and Torrent 1997). In place of AERs, resin-impregnated membranes such as anion exchange membranes (AEMs) have been used successfully in P desorption studies (Saunders 1965, Saggar et al. 1990, Nuernberg

PAGE 31

18 et al. 1998). Both AERs and AEMs behave as a P sink, desorbing P from the solid phase by causing disequilibrium; and have proven to be useful indices of plant-available P when used as extractants in soil tests (Sibbesen 1978, Schoenau and Huang 1991, Ziadi et al. 2001). The AEM behavior in soil, including recyclability, P sorption kinetics, and adsorbability of solution P by AEMs, are well-documented (Cooperband and Logan 1994, Cooperband et al. 1999). Anion exchange materials have been used to investigate the desorption kinetics of P (Raven and Hossner 1994); to measure P mineralization (Parfitt et al. 1994); and to measure total resin-desorbable P by sequential extraction (He et al. 1994). However, P desorption isotherms have rarely been developed. In one noteworthy case the validity of using AEMs for developing P desorption isotherms was confirmed for Brazilian Cerrado soils with varying clay and organic matter contents (Barros Filho et al. 2001, Barros Filho et al. in press). Brewster et al. (1975) compared P desorption isotherms using AERs with those developed by dilution. Brewster and colleagues concluded that the use of the resins was a more desirable technique than dilution primarily because of the large dilution volumes required. The sequential extraction method has some inherent problems. Since this method can require up to several weeks to develop a complete P desorption isotherm, continuous shaking of samples may increase the soils surface area because of soil particles abrasion (Barros Filho et al. in press). Another concern is that microbial P mineralization or immobilization that may be induced during the prolonged desorption process can influence P desorption patterns (Barros Filho et al. in press). The first objective of our

PAGE 32

19 study was to evaluate the suitability of using multiple strips of AEMs (termed the Multiple AEM Method) to develop P desorption isotherms. Initial work showed that P desorption was sensitive to the amount of AEM surface area in contact with soil. This objective addressed whether the Multiple AEM Method could be used to more rapidly develop a desorption isotherm. The second objective was to compare the Multiple AEM Method with a sequential-extraction approach using AEMs (termed the Sequential AEM Method) to determine if the manner in which AEMs were used would influence the desorption isotherm. Materials and Methods Soil Material The soil material used in our study was the surface horizon (0 to 30 cm depth) of a Kandiudult (Soil Survey Staff 1999), known as an Argissolo vermelho/vermelhoamarelo distrfico in the Brazilian system (Oliveira et al. 1992). It was taken from under a native Atlantic Forest cover type (Morellato and Haddad 2000) located near Una Ecoparque near the city of Una, Bahia, Brazil. Using a shovel, the material was collected from a 1 ha area. Approximately 20 sampling points were combined; and the soil was subsequently air-dried, passed through a 2-mm sieve, and stored in plastic bags. Selected chemical and physical properties of the soil material were determined by standard methods (Page et al. 1982), which are listed in greater detail in Chapter 3. Soil pH was 4.7 (1:2 soil:water), with 2.5% organic carbon and 25% clay content dominated by kaolinite. Mehlich 1-available P and total P were 2.7 g g and 98.3 g g respectively. The iron content was greater than the aluminum content in both the amorphous and crystalline oxide forms.

PAGE 33

20 Phosphorus Desorption Kinetics Using AEMs To maximize effectiveness of AEMs, the kinetics of P desorption in the presence of AEMs was first determined. Triplicate 2 g samples were weighed into 50 mL centrifuge tubes; and then 20 mL of 50 mM KCl containing 25 g P mL as KH 2 PO 4 were added to each tube. Triplicate blanks of 20 mL of 50 mM KCl without soil and without P were included. The tubes were shaken horizontally for 20 h; and then centrifuged for 15 min at 2500 rpm. The supernatant was filtered through Whatman No. 5 filter paper and measured for soluble reactive P by the method of Murphy and Riley (1962), which was used for all P determinations in our study. Soluble reactive P as measured by this method is operationally defined as inorganic P. Sorbed P was calculated as the difference between the initial solution concentration added and the final concentration in the filtrate. After decanting the supernatant, 50 mM KCl was added once more to each tube until it reached its original pre-shaking weight. The AEM strips (1.25 cm 5.00 cm; Type AR204-SZRA-412, Ionics Inc., Watertown, MA) were prepared for use by soaking overnight in 1 M KCl of sufficient volume to provide an anion concentration at least 5 times greater than the total anion exchange capacity of the strips. After rinsing with double-deionized water, one strip was inserted into each centrifuge tube. The tubes then were shaken; and triplicate samples were removed from the shaker at 0.5, 1, 2, 4, 10, 24, and 48 h. Each AEM strip was carefully removed from the tube so as to not concomitantly remove soil particles; was rinsed with double-deionized water; and was placed in another centrifuge tube filled with 15 mL of 1 M KCl. This tube was then shaken for 30 min; and the solution was filtered through Whatman No. 5 filter paper. Phosphorus in this filtrate was determined.

PAGE 34

21 Phosphorus Desorption as Affected by AEM Surface Area Ten sets of triplicate centrifuge tubes were prepared; and the P sorption process was performed in the same manner as described above. After decanting the supernatant, 50 mM KCl were added to the tube until it reached its original pre-shaking weight. After rinsing with double-deionized water, 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 strips (1.25 cm 5.00 cm) of AEMs that had been prepared as described above were inserted into each set of triplicate tubes. The tubes then were shaken for 2 h. The strips were removed from the tubes; were rinsed with double-deionized water; and were placed in another 10 sets of triplicate tubes filled with 15 mL of 3 M KCl for the 1-, 2-, and 3-strip tubes; 25 mL for the 4-, 5-, and 6-strip tubes; 30 mL for the 7and 8-strip tubes; and 40 mL for the 9and 10-strip tubes. These tubes were shaken for 30 min; and the solution was filtered through Whatman No. 5 filter paper. Phosphorus in the filtrate was determined. Phosphorus Desorption Isotherms Using the Multiple AEM Method Phosphorus desorption isotherms then were developed using multiple AEM strips on soil samples with 3 levels of sorbed P. Phosphorus was sorbed onto the soil using initial loading levels of 40, 80 or 180 g P g soil. These levels were chosen because they bounded normal fertilizer additions; and were equivalent to 16, 32, and 72 kg P ha assuming the fertilizer to be broadcast and incorporated into the top 3 cm of soil. Triplicate 2 g soil samples were weighed into 6 sets of triplicate centrifuge tubes; and the initial weight of each tube + soil was recorded. Twenty milliliters of 50 mM KCl containing 4, 8 or 18 g P mL as KH 2 PO 4 were added to the tubes; and the total initial weight of tube + soil + solution was recorded. The tubes then were shaken for 20 h; and centrifuged for 15 min at 2500 rpm. The supernatant solution was filtered through Whatman No. 5 filter paper; and P in the filtrate was determined. Phosphorus sorbed

PAGE 35

22 onto the solid phase was calculated as the difference between the initial or previous P concentration and the P concentration in the filtrate. After decanting the supernatant, tube + soil weights were recorded; and each tube was filled with 50 mM KCl to its initial tube + soil + solution weight. The AEM strips prepared in KCl solution then were rinsed with double-deionized water; and 1, 2, 3, 4, 5, or 6 strips (1.88 cm 7.50 cm) were inserted into each of 6 sets of triplicate tubes. The tubes were shaken for 2 h. The strips then were removed from the tubes; were rinsed with double-deionized water; and were placed in another 6 sets of triplicate tubes filled with 15 mL of 2 M KCl for 1-, 2-, and 3-strip tubes; and 30 mL for 4-, 5-, and 6-strip tubes. These KCl tubes were shaken for 30 min; and then filtered through Whatman No. 5 filter paper. Phosphorus in filtrate was measured. The original tubes with soil samples were shaken for 20 h; and then centrifuged for 15 min at 2500 rpm. The supernatant solution was filtered through Whatman No. 5 filter paper; and P in the filtrate was determined. The first step of this procedure was designed to reduce desorbable P on the solid phase using AEMs. The equilibration step was required to develop a desorption isotherm, because the soil must come into an equilibrium or quasi-equilibrium to measure an equilibrium solution concentration. The entire procedure of the previous paragraph was repeated two more times, resulting in a total of 3 desorption steps and yielding 18 points for each desorption isotherm in 4 d. Phosphorus Desorption Isotherms Using the Sequential AEM Method The Sequential AEM Method for P desorption isotherms was performed using a constant number of AEM strips and multiple extraction steps on soil samples having the same three P-addition levels as in the multiple AEMs study. The procedure was the same

PAGE 36

23 as that using multiple AEM strips, except that only 4 strips were used in all extraction steps. The extraction steps were continued until the concentration of P desorbed by AEMs in the KCl filtrate solution reached the P detection limit (0.005 g P mL ; 0.05 g P g soil). The Sequential AEM Method required 6, 9, and 13 desorption steps and took 7, 10, and 14 d to develop desorption isotherms for samples to which 40, 80, and 180 g P g had been added, respectively. Statistical Analysis Statistical differences in the pattern of P desorption using different methods and initial P loading levels were tested using the General Linear Models (GLM) procedure in the Statistical Analysis System framework (SAS Institute 1988). Observed P desorption was related to the amount of P initially added and the method by which P desorption isotherms were developed. The model used was: EPaddedMethodCCClls|32210 (2-1) where C s is the amount of P sorbed on the solid phase per unit mass of soil in g P g of soil; C l is the equilibrium concentration of P in g P mL ; 0 is the intercept parameter, and 1 2 and 3 are the slope parameters of the regression line; and E is the random error. The Method variable has two levels representing the Multiple and Sequential AEM Methods. The Padded variable has three levels of 40, 80, 180 g P g initially added. Analysis of variance (ANOVA) was used to test differences among regression line parameters; and significance of the differences among means were determined by t-tests at the p = 0.05 level. Means and standard deviations were used to summarize the data, where applicable.

PAGE 37

24 Equation for K d values The tendency of the solid phase to buffer C l can be described numerically using K d Since K d is the rate of change of the ratio of sorbed P to solution P, it can be estimated from the first derivative of equation (2-1) with respect to the solution concentration, expressed as: ldlsCKdCdC212 (2-2) Results Phosphorus Desorption Isotherms: Kinetics and Effect of AEM Surface Area The amount of P desorbed by the AEM strips increased with shaking time up to 2 h, after which it decreased once more (Fig. 2-1). Therefore, 2 h was chosen as the preferred shaking time. When the AEM surface area was increased, P desorption increased with increasing number of AEM strips (Fig. 2-2), with the relationship being linear between 1 and 6 strips (r 2 = 0.997). When more than 6 strips were added, replicate variability increased dramatically; and the linear relationship weakened. Phosphorus Desorption Isotherms: Multiple Method vs. Sequential AEM Method The amount of P initially sorbed at each P-addition level for the two extraction methods is provided in Table 2-1. There were significant differences in the amount of P sorbed between desorption methods for all 3 levels of P addition. Both the Multiple and Sequential AEM Methods resulted in well-defined desorption isotherms for all 3 levels of sorbed P (Fig. 2-3). The P desorption GLM model well described the isotherms for both desorption methods for all three P-addition levels. As shown in Figure 2-4, 99.8% of the variability in observed C s values were accounted for by the model. Type III model ANOVA table and parameter estimates for

PAGE 38

25 the P desorption GLM model (Tables 2-2 and 2-3) confirmed that P desorption isotherms for each desorption method for each P-addition level were numerically different. Statistical analysis also confirmed that the desorption methods and the P-addition levels affected the P desorption isotherms except for those isotherms at the lowest P-addition level (40 g P g ). There was no significant Method*Padded interaction on C s values at this P-addition level. Phosphorus Desorption Isotherms: K d values Calculated K d values derived from the P desorption GLM model were not a function of the desorption method or the amount of P initially sorbed. These K d values were inversely and linearly related to equilibrium P concentration; and were expressed by the equation: K d = 48.1 60.862 C l (2-3) The maximum K d value was 48.1, which was the intercept of the above relationship. Discussion In evaluating phosphorus bioavailability, it is P desorption rather than P sorption which determines the quantity of available P and subsequent ability of the soil to release P. Phosphorus flux through the soil to plant roots depends on the initial solution P concentration, the soils buffer power, and the effective diffusion coefficient (Barber 1984). Buffer power and diffusion coefficients incorporate K d values, which are calculated from P desorption isotherms. Therefore, a well-defined P desorption isotherm is critical, especially when evaluating P bioavailability using computer-based nutrient uptake models (Nye and Tinker 1977, Van Rees et al. 1990a, Smethurst and Comerford 1993, Macariola-See et al. 2003), since the output of the models is dependent on K d values (Bhatti and Comerford 2002).

PAGE 39

26 Our study showed that the Multiple AEM Method was suitable for developing desorption isotherms over a range of P sorption levels, thereby addressing our studys first objective. Regarding the second objective, it was shown that, for all P-addition levels, both methods provided well-defined P desorption isotherms, although they were numerically different. This difference was derived primarily from differences in the amount of P initially sorbed (Table 2-1). The fact that desorption isotherms by the different methods were not developed concurrently might also account for some of the these differences. However, since the ultimate goal of the desorption isotherms is to define K d values with respect to solution P concentrations, both desorption methods gave the same result. Based on the GLM model used in our study, calculated K d values were a function of simply the solution P concentration. The most useful method for developing isotherms should be as rapid as possible in order to reduce the effects of shaking and microbial immobilization and mineralization (Barros Filho et al. in press). In this respect both the Multiple and Sequential AEM Methods are preferable to a dilution method, because the latter requires excessively high dilutions. The Multiple AEM Method may also be superior to sequential desorption methods, since it required about half the time to cover the same range of solution P concentration while providing more data points within that range. The Sequential AEM Method requires weeks to develop desorption isotherms; thus, this prolonged procedure may change the physio-chemical and biological status of P in both the solid and solution phases; and thereby influence the shape of the P desorption isotherm. In the case of highly aggregated soils, extended shaking may induce disaggregation, creating more

PAGE 40

27 surface area. Reduced P desorption may result if more P is sorbed to newly created sorption sites. Microbial P has been reported to reach its maximum level between 3 and 7 d after addition of 30 to 100 kg P ha as triple superphosphate to a Bh horizon from a Florida Spodosol (Bliss 2003). Therefore, the less time required to develop a desorption isotherm, the better the chance that the desorption method will not be influenced by outside factors. It can be concluded that the Multiple AEM Method is superior to the Sequential AEM Method because it requires less time to conclude. Unfortunately, there is no standard methodology for developing desorption isotherms. The Multiple AEM Method gave more desorption data points with fewer extraction steps. This is encouraging enough to suggest that this method be tested on a range of soil types with an objective of evaluating its suitability in developing P desorption isotherms. These data from our study emphasize the need to standardized methodology for the development of desorption isotherms.

PAGE 41

28 Figure 2-1 Kinetics of P desorption using AEMs Figure 2-2 Phosphorus desorption with increasing number of AEM strips using a 2 h shaking time P desorbed by AEMs (g g ) 0.00.51.01.52.02.50102030405 0 Shaking time (hour) P desorbed by AEMs (g g ) y = 0.5871x + 0.069501234560123456789101 Number of AEM strips r 2 = 0.997 1

PAGE 42

29 Table 2-1 The amount of P sorbed for each desorption method at 3 different initial P-addition levels P initially added (g g ) Desorption Method P initially sorbed (g g ) 180 Multiple AEM 152.04b Sequential AEM 144.67a 80 Multiple AEM 74.91b Sequential AEM 72.12a 40 Multiple AEM 37.89a Sequential AEM 38.35b Different letters assigned in a column for each P-addition level denote significant differences between desorption methods at the 0.05 probability level, by Tukeys studentized range test

PAGE 43

30 0204060801001201401600.00.10.20.30.40.50.60.70.8 Multiple 180 Sequential 180 Multiple 80 Sequential 80 Multiple 40 Sequential 40 A P remainin g sorbed (g g ) B C E q uilibrium P concentration (g mL 1 ) Figure 2-3 Phosphorus desorption isotherms using 2 desorption methods with 3 initial P-addition levels. A) 180 g P g B) 80 g P g C) 40 g P g

PAGE 44

31 Table 2-2 ANOVA table for the GLM model evaluating the P desorption isotherms Source Type III Sum of Square df Mean Square F value Pr > F Intercept 79692.83 1 79692.83 24832.61 < 0.0001 C l 206.45 1 206.45 64.33 < 0.0001 C l 2 46.39 1 46.39 14.46 < 0.0003 Method 5372.89 1 5372.89 1674.21 < 0.0001 Padded 22973.08 2 11486.54 3579.25 < 0.0001 Method*Padded 8640.22 2 4320.11 1346.16 < 0.0001 Error 240.69 75 3.21

PAGE 45

32 Table 2-3 Estimates for parameters of the GLM model used to test desorption isotherm differences Parameter Level of parameter Estimate Standard error t value Pr > | t | Intercept 128.00 1.00 127.99 < 0.0001 C l 48.05 5.99 8.02 < 0.0001 C l 2 30.43 8.00 3.80 < 0.0003 Method SEQ 50.46 0.75 67.23 < 0.0001 Method MUL 0.00 Padded 40 92.41 0.97 94.85 < 0.0001 Padded 80 59.10 0.83 71.27 < 0.0001 Padded 180 0.00 Method*Padded SEQ 40 50.72 1.12 45.34 < 0.0001 Method*Padded SEQ 80 43.30 1.01 42.88 < 0.0001 Method*Padded SEQ 180 0.00 Method*Padded MUL 40 0.00 Method*Padded MUL 80 0.00 Method*Padded MUL 180 0.00 SEQ and MUL represent Sequential AEM Method and Multiple AEM Method, respectively Initial P-addition levels in g P g

PAGE 46

33 020406080100120140160020406080100120140160 y = 0.9979x + 0.1649 Predicted C s (g g ) r 2 = 0.998 Observed C s (g g 1 ) Figure 2-4 Relationship between observed C s values and C s values predicted from the P desorption GLM model

PAGE 47

CHAPTER 3 INFLUENCE OF SOIL pH ON INORGANIC PHOSPHORUS SORPTION AND DESORPTION IN A HUMID BRAZILIAN ULTISOL Introduction Effect of Soil pH on P Sorption In tropical regions, liming is frequently used to raise soil pH and increase phosphorus bioavailability (Sanchez and Uehara 1980). However, published results regarding the effect of liming on P sorption and bioavailability are conflicting (Haynes 1982). Phosphorus sorption has been shown to both decrease (Naidu et al. 1990, Holford et al. 1994, Mora et al. 1999); and increase (Geelhoed et al. 1997, Pereira and De Faria 1998, Curtin and Syers 2001) with increasing pH. Still other reports have shown no significant influence of pH (Jones and Fox 1978, Arias and Fernandez 2001). Liming 3 acid soils from southern Brazil increased the amount of P sorbed up to a pH of 5.0, at which point P sorption began to decrease (Anjos and Rowell 1987). Raising the pH of an Oxisol from the Cerrado region of Brazil with an initial pH of 4.5 also reduced P sorption by 18 to 24% (Smyth and Sanchez 1980). The authors attributed the reduced P sorption to increased hydroxyl concentration and increased competition between hydroxyl and phosphate ions for specific adsorption sites on mineral surfaces. Several authors thought that decreased P sorption could be due to the neutralization and precipitation of Al 3+ and hydroxy-Al as Al hydroxide during liming; decreasing the number of P sorption sites (Smyth and Sanchez 1980, Anjos and Rowell 1987, Naidu et al. 1990). Bowden et al. (1980) and Haynes (1982) added that the mineral surface 34

PAGE 48

35 became increasingly negative with increasing pH, which resulted in greater electrostatic repulsion and decreased P sorption. Phosphate is preferentially sorbed by soil mineral surfaces as HPO 4 2 rather than as H 2 PO 4 ; and the concentration of the divalent ion (HPO 4 2 ) increases 10-fold for each unit increase in pH from 2 to 7 (Bowden et al. 1980). This change partially offsets any decrease in electrostatic potential (Haynes 1984). Thus, P sorption may decrease relatively slowly until pH 7; and then, above pH 7, the increase of HPO 4 2 concentration becomes progressively slower; whereas the decrease in surface potential continues, resulting in a more rapid decrease in P sorption (Haynes 1984). In contrast to the studies above, Chen and Barber (1990) showed that for acid weathered soils of pH 4.2 to 4.6 when adjusted up to a pH of 8.3, sorbed P increased up to pH 6.0 to 6.2, and then decreased at higher pH values. The initial increase in P sorption was explained as the formation of amorphous hydroxyl Al with highly active sorbing surfaces. The subsequent decrease in P sorption was attributed to increased competition of hydroxyl with phosphate for sorption sites. Similar conclusions were drawn from liming studies involving Nigerian acid soils (Mokwunye 1975) and Al-organic matter complexes (e.g., Al-peat and Al-humate) (Haynes and Swift 1989). Effect of Soil pH on P Desorption There are few investigations that focused on soil pH and P desorption; and the results have been inconsistent; increased P desorption with increasing pH in some cases (De Smet et al. 1998) and with decreasing pH in others (Barrow 2002). Studies of P desorption kinetics for synthetic goethite as a function of suspension pH from 4 to 10 (Madrid and Posner 1979, Cabrera et al. 1981) showed increased P desorption with increasing suspension pH. This trend could again be explained as above; as due to

PAGE 49

36 competition from hydroxyl ions; and a lessened attraction or an enhanced repulsion caused by increased negative charge on the surface with increasing pH. Rupa et al. (2001) also desorbed P from 3 surface soils using 6 consecutive dilution extractions at pH values varying from 4.2 to 8.0. The surface soils represented an Indian Vertisol, Alfisol, and Oxisol. Phosphorus desorption increased for all soils up to a pH of 5.5, and then decreased at pH values of 6.7 and 8.0. The soils had high levels of exchangeable Ca 2+ plus Mg 2+ (41.2, 8.7, and 4.3 cmol c kg for Vertisol, Alfisol, and Oxisol, respectively). It was thought that decreased desorption at high pH values could be due to P precipitation as a Ca-phosphate. Phosphorus desorption is a key process affecting inorganic P bioavailability; and desorption also can identify specific pools of bioavailable P. However, in order to use the desorption process in nutrient uptake models such as described by Smethurst and Comerford (1993), a P desorption isotherm is required. An isotherm describes the amount of P on the solid phase relative to the solution P concentration with the slope of this relationship being called the partition coefficient, K d The K d is defined as the rate of change of the amount desorbed from the solid phase over the rate of change in solution concentration. The K d values are directly related to soil buffer power as seen in the following equation: b = + K d (3-1) where b is the soil buffer power, is volumetric water content in cm 3 of water per cm 3 of soil, and is soil bulk density in units of g cm Both K d and b are critical parameters for computer-based nutrient uptake models because of their role when calculating diffusion as a nutrient-supply mechanism. Diffusion is important because for many soils,

PAGE 50

37 diffusion is the mechanism supplying 90 to 98% of P absorbed by roots (Barber 1980). Unfortunately, measuring P desorption isotherms is a time-consuming process; and there is no standardized method for their development (see Chapter 2). The overall objective of our study was to investigate the influence of pH on P desorption from an Ultisol from the Atlantic Forest of Brazil. The Atlantic Forest is a major Brazilian ecosystem type, with only about 7.6% of the native forest remaining at present (Morellato and Haddad 2000). Much of the soil in this area has been degraded by deforestation and subsequent extensive management for pasture and row crops. Phosphorus management is a crucial component of rehabilitating these degraded soils. This soil also represents an area considered one of the most biodiverse on the planet (Oliveira and Fontes 2000). Specific objectives of our study were to study this soil and its properties, as related to the effect of soil pH on P sorption and desorption. The first hypothesis was that P sorption is a function of pH, with decreasing sorption with increasing pH between 4.7 and 7.0. This soil is inherently low in exchangeable Al 3+ so the mechanism of amorphous Al-OH creating new surface for sorption was considered relatively unimportant. Understanding the influence of liming on P sorption for these soils should play a role in their management to enhance P bioavailability. A second hypothesis was that P desorption is a function of pH, with increasing desorption as pH increases. The degree to which liming influences P desorption is another important aspect of P management that is virtually unknown; and should also be critical to proper nutrient management.

PAGE 51

38 Materials and Methods Soil Material and Its Characterization The soil material used in our study was the surface horizon (0 to 30 cm depth) of a Kandiudult (Soil Survey Staff 1999), known as an Argissolo vermelho/vermelhoamarelo distrfico in the Brazilian system (Oliveira et al. 1992). It was taken from under a native Atlantic Forest cover type located near Una Ecoparque near the city of Una, Bahia, Brazil. Using a shovel, the material was collected from a 1 ha area. Approximately 20 sampling points were combined; and the soil was air-dried, passed through a 2-mm sieve, and stored in plastic bags. Soil characterization was performed as follows: pH (1:2 soil:deionized H 2 O; 1:2 soil:1 M KCl; and 1:2 soil:0.01 M CaCl 2 ) using a glass-electrode pH meter, Orion Model 720A; organic carbon using a Walkley-Black method (Nelson and Sommers 1982); total N by flash combustion using a Carlo Erba Model NA 1500 apparatus (Dumas 1831); particle-size distribution using a pipette method (Gee and Bauder 1986); exchangeable cations using 1 M ammonium acetate (pH 4) as the displacing solution, with Ca 2+ Mg 2+ K + and Na + measured by atomic absorption spectroscopy; 1 M KCl-exchangeable Al 3+ and 1 M ammonium acetate-exchangeable Fe 3+ measured by atomic absorption spectroscopy; acid ammonium-oxalateand sodium citrate-dithionite-extractable Al 3+ and Fe 3+ determined by the method of McKeague and Day (1966); water-extractable P using 50 mM KCl; Mehlich 1-extractable P (Mehlich 1953); total P by digestion with H 2 SO 4 and 30% H 2 O 2 as described by Grierson et al. (1999); and clay mineralogy and its quantification determined by X-ray diffraction and thermogravimetric analysis, respectively.

PAGE 52

39 Phosphorus Sorption Isotherms Two 100 g sub-samples were taken from the bulk soil sample and adjusted to pH 5.9 and 7.0 from the original pH of 4.7. Each sub-sample received the appropriate volume of NaOH solution to increase soil pH, based on an acid-base titration curve for the soil (appendix A). The sub-samples then were air-dried at room temperature. Phosphorus sorption isotherms were produced for each of the 3 pH levels. Solution pH was not controlled during the sorption process. Triplicate 2 g soil samples were weighed into 50 mL centrifuge tubes; and 20 mL of 50 mM KCl containing 2, 4, 6, 8, 10, 12, 14, 16, 18, or 20 g P mL as KH 2 PO 4 was added to each tube. Triplicate blanks of 20 mL of 50 mM KCl without soil and without P were included as well. The tubes were shaken horizontally for 20 h; and centrifuged for 15 min at 2500 rpm for the pH 4.7 and 5.9 samples, and for 15 min at 10,000 rpm for the pH 7.0 sample in order to clear the supernatant solution. The supernatant was filtered through Whatman No. 5 filter paper; and measured for soluble reactive P by the method of Murphy and Riley (1962), which was used for all P determinations in our study. Soluble reactive P as measured by this method is operationally defined as inorganic P. Sorbed P was calculated based on the difference between the initial solution concentration added and the final concentration in the filtrate. Phosphorus Desorption Isotherms Using the Multiple AEM Method For the Multiple AEM Method, pH of the soil was adjusted as described for the P sorption-isotherms study. The pH adjustment was only at the beginning of the experiments and was not controlled throughout the study. The pH control during P desorption is not expected to have a significant effect on the P desorption process

PAGE 53

40 (Bhatti and Comerford 2002); and pH is not controlled under field conditions following liming. Phosphorus desorption isotherms were initiated from 3 different sorption points at each pH. Phosphorus was sorbed onto the soil using 40, 80, or 180 g P g soil. These levels were chosen because they bounded normal fertilizer additions; and were equivalent to 16, 32, and 72 kg P ha (assuming fertilizer was broadcast and incorporated into the top 3 cm of soil). Triplicate 2 g soil samples from each pH level were weighed into 6 sets of triplicate centrifuge tubes; and the initial weight of tube + soil was recorded. Twenty milliliters of 50 mM KCl containing 4, 8 or 18 g P mL as KH 2 PO 4 were added to each tube; and the total initial weight of tube + soil + solution was recorded. Tubes were shaken for 20 h; and centrifuged for 15 min at 2500 rpm for the pH 4.7 and 6.0 samples, and for 15 min at 10,000 rpm for the pH 7.1 sample in order to clear the supernatant solution. The supernatant solution was filtered through Whatman No. 5 filter paper; and P in the filtrate was determined. After decanting the supernatant, weights of tube + soil were recorded. Each tube then was filled with 50 mM KCl to its initial tube + soil + solution weight. Anion exchange membrane (AEM) strips (1.88 cm 7.50 cm) prepared in KCl solution were rinsed with double-deionized water; and 1, 2, 3, 4, 5, or 6 strips were inserted into each of 6 sets of triplicate tubes. The tubes were shaken for 2 h; then the strips were removed from the tubes; were rinsed with double-deionized water; and were placed in another 6 sets of triplicate tubes filled with 15 mL of 2 M KCl for the 1-, 2-, and 3-strip tubes; and with 30 mL for the 4-, 5-, and 6-strip tubes. These tubes were shaken for 30 min; and

PAGE 54

41 then filtered through Whatman No. 5 filter paper. Phosphorus in filtrate was measured. The original tubes with soil were shaken for 20 h; and then centrifuged for 15 min at 2500 rpm for the pH 4.7 and 6.0 samples, and for 15 min at 10,000 rpm for the pH 7.1 sample. The supernatant solution was filtered through Whatman No. 5 filter paper; and P in the filtrate was determined. The first step of this procedure was to reduce desorbable P on the solid phase using AEMs. The equilibration step was required to develop a desorption isotherm because the soil must come into an equilibrium or quasi-equilibrium to measure an equilibrium solution concentration. The entire procedure of the previous paragraph was repeated two more times, resulting in a total of 3 desorption steps and yielding 18 points on a desorption isotherm for each pH in 4 d. Equation for P Sorption and Desorption Isotherms The Langmuir equation was fitted to all data for both the sorption and desorption isotherms. The following form of the equation was used: llsCKCbKC1 (3-2) where C s is the amount of P sorbed onto the solid phase per unit mass of soil in g P g of soil; C l is the equilibrium concentration of P in g P mL ; K is the affinity index expressed in mL g of P; and b is the sorption maximum with unit of g P g of soil. Both K and b are empirical constants that can be calculated from a linear form of the equation plotted as C l /C s against C l : bKbCCClsl1 (3-3) where 1/b and 1/Kb are the slope and intercept of the linear regression, respectively.

PAGE 55

42 Ability of the solid phase to buffer C l can be described numerically using the partition coefficient, K d This reflects the rate of change of the ratio of sorbed P to solution P; and is the first derivative of equation (3-2) with respect to the solution concentration: 2)1(ldlsCKbKKdCdC (3-4) Statistical Analysis Statistical differences among different linear regression lines were tested using the General Linear Models procedure in the Statistical Analysis System framework (SAS Institute 1988). Analysis of variance (ANOVA) was used to test differences among regression line parameters. Significance of the difference among means was determined by t-tests at the p = 0.05 level. Means and standard deviations were used to summarize the data from P sorption and desorption isotherms, where applicable. Results Selected physical and chemical properties of the soil are summarized in Table 3-1. The soil was acidic with low cation exchange capacity; and with low water-soluble and Mehlich 1-extractable P. Its 25% clay content was dominated by kaolinite. The iron content was greater than the aluminum content in both the amorphous and crystalline oxide forms. Phosphorus Sorption pH Envelope Phosphorus sorption was a function of soil initial pH, with sorption decreasing with increasing pH when equilibrium solutions P concentrations were less than about 3 g P mL (Fig. 3-1a). Phosphorus sorption decreased by up to 21% when pH was

PAGE 56

43 increased from 4.7 to 5.9; and decreased by up to 34% when pH was increased from 4.7 to 7.0. Above 3 g P mL there were no sorption differences between pH treatments. The Langmuir equation fit P sorption data for all 3 pH levels (r 2 = 0.970 to 0.993, Table 3-2), with Langmuir parameters such as K and b values being affected by pH levels. As pH increased, K values significantly decreased; while b values increased significantly only at the highest pH level (Table 3-2). Changes in K d values relative to equilibrium P concentration are provided in Figure 3-2. This relationship proved to be a function of soil pH. The K d values at the higher pH levels were lower than that at the original pH, once the equilibrium concentration was reduced to about 0.5 g P mL The K d values were as much as 22% and 34% lower at pH 5.9 and 7.0, respectively, than at pH 4.7. Phosphorus Desorption pH Envelope Phosphorus desorption exhibited the expected hysteresis between sorption and desorption at all 3 pH levels for each level of P loading (Fig. 3-3). Since P sorption was a function of initial soil pH as mentioned above, P desorption isotherms began from different sorption points which depended on soil pH and the amount of P initially added (Table 3-4). The Langmuir equation fit the majority of the desorption data well (r 2 = 0.338 to 0.867) except for the low P-addition level; and soil pH significantly influenced the values of the Langmuir equations parameters (Table 3-3). For the medium and high levels of added P, K values were significantly decreased only at pH 7.1. At the low level of added P, the K value significantly increased at pH 6.0 but significantly decreased at pH 7.1, compared to values at pH 4.7. The range of K values across pH levels for the medium

PAGE 57

44 and low levels of P addition was approximately 6-fold and 40to 127-fold higher than at the high P-addition level. The b values were significantly lower at pH 6.0 and 7.1 than at pH 4.7, but were not significantly different from each other for the high level of P addition. For the medium P-addition level, b values for both elevated pH levels were significantly higher than at the original pH. For the low P-addition level, b value only at pH 6.0 was significantly higher than at pH levels of 4.7 and 7.1. Phosphorus desorption was a function of soil pH when expressed in terms of either the amount of P desorbed or the ratio of P desorbed to P initially sorbed (Table 3-4). Higher pH levels resulted in more P desorbed and a higher ratio of P desorbed to P initially sorbed. This was true for both the medium and high levels of P addition. However, for the low level of P added, both P desorbed and the corresponding ratio were slightly decreased at pH 6.0 and significantly increased at pH 7.1, compared to values at pH 4.7. The calculated K d values, when plotted against equilibrium P concentrations, evidenced opposite trends to those documented in the sorption study (Fig. 3-4). For the high level of P addition, K d values at pH 6.0 was lower than at pH 4.7 by up to 13%; while K d values at pH 7.1 was up to 38% higher than at pH 4.7. For example, K d values at pH 4.7, 6.0, and 7.1 were 32.9, 28.7, and 43.7 at an equilibrium concentration of 0.2 g P mL respectively. For the medium level of P addition, K d values at pH 6.0 and 7.1 were 13 to 14% and 76 to 85% greater than at pH 4.7, respectively. For the low level of P addition, K d

PAGE 58

45 values were 52 to 53% lower at pH 6.0 and 120 to 126% greater at pH 7.1, respectively, compared to values at pH 4.8. Discussion Soil pH Effect on P Sorption Results from the P sorption study supported the hypothesis that, in these soils, P sorption decreases with increasing pH. This is consistent with what has been shown for other weathered, acid soils with low levels of exchangeable Al 3+ (Naidu et al. 1990). This result has been credited to an increased electrostatic repulsion due to the increased negative surface charge (Bowden et al. 1980, Haynes 1982) that accompanies increasing pH. A contributing factor is that higher hydroxyl concentrations, which can be 10 to 1000 times higher than at lower pH levels, compete effectively with phosphate ions for specific sorption sites on mineral surfaces (Smyth and Sanchez 1980). This result may also be assisted by a reduction in the number of P-sorption sites. At higher pH levels Al hydroxide polymers can neutralize sites where more reactive Al surfaces once were present (Sanchez and Uehara 1980). Soils dominated by kaolinite, goethite, and gibbsite, like the soil used in our study, are particularly susceptible to each of the above-named mechanisms. Empirically, trends for the Langmuir K parameter and for K d support these explanations. The reduced Langmuir K values with increasing pH suggest a reduced P affinity for the sorption surface. The K d represents the rate of change in the amount of P sorbed relative to the amount of P in solution upon continued introduction of P to the solution phase. Therefore, decreasing K d with increasing pH corresponds to decreasing rates of change in the amount of P sorbed, leaving more of the applied P in solution.

PAGE 59

46 At higher equilibrium concentrations ( 3 g P mL ), there were no differences between isotherms at any pH. However, at a commonly practiced P-fertilization rate of 50 kg P ha which corresponds to 125 g P g of soil using the above assumptions, maximum solution P concentration was only 1.4 g P mL Soil solution P concentration is commonly around 0.05 g P mL (Barber et al. 1962), and seldom higher than 0.3 g P mL in most agricultural soils (Fried and Shapiro 1961). It is typically even lower in soils under native forests and forest plantations (Ballard 1980). Therefore, P sorption isotherms for the same 3 pH values were replotted focusing on a lower solution P-concentration range (0.0.0 g P mL Fig. 3-1b). These data show that P sorption decreases with pH increase under common P-fertilization practices. At a solution concentration of 0.2 g P mL sorbed P was 43, 36, and 32 g P g of soil at pH 4.7, 5.9, and 7.0, respectively. These represent 16% and 26% reduction in sorption compared to a pH of 4.7 when the soil is limed to pH 5.9 and 7.0, respectively. These percentages represent 2.8 to 4.5 kg ha of fertilizer P remaining in solution, ready for plant uptake. Soil pH Effect on P Desorption Phosphorus desorption is a primary control on inorganic P bioavailability; yet virtually nothing is known about the nature of P desorption for acid, weathered tropical soils. Our study represents the first investigation of desorption isotherms for soils in this region of Brazil. The results endorse the hypothesis that P desorption increases with increasing soil pH. Both the decreased Langmuir K and the increased amounts of P desorbed at elevated pH levels support the positive relationship between pH and P desorption. These results are consistent with the findings of He et al. (1994). He and

PAGE 60

47 associates determined P desorption from a variety of minerals and ranked the minerals according to their desorption capacity: goethite < kaolinite < amorphous Al-oxide < montmorillonite. Affinity constants also decreased in the same order. The affinity constant, K, is related to the amount of P sorbed. Phosphate is held with its highest affinity to the soil surface when the surface coverage of sorbed P is low. The affinity then decreases with increasing surface coverage (Ryden et al. 1977). The calculated K d for desorption represents the rate of change in the amount of P desorbed relative to the solution P concentration. Therefore, increasing K d is interpreted as increasing the rate of change in P desorption at a particular solution P concentration, as was observed for all levels of P loading when soil pH was raised to 7.1. Similar results can be found among the limited P desorption studies previously reported (Hingston et al. 1974, Madrid and Posner 1979, He et al. 1989, De Smet et al. 1998). General consensus regarding the effect of soil pH on P desorption, drawn from the findings of these authors, is that increased P desorption with increased pH is due both to increased competition with hydroxyl ions and to the change in electrostatic potential of the surface. Although hydroxyl can be a good competitor with phosphate for metal coordination sites on the surface (McBride 1994), this theory is still untested because it has not been convincingly shown that electrostatic repulsion affects the specifically sorbed anions that are directly coordinated to discrete surface metal cations. Given the lack of prior work on P desorption mechanisms, it is suggested that this is a fruitful avenue for further studies. Phosphorus desorption is potentially controlled by the type of P complexation with the surface: monodentate, bidentate mononuclear, and bidentate binuclear. These

PAGE 61

48 complexes can be either non-protonated or protonated, depending on suspension pH (Tejedor-Tejedor and Anderson 1990). Bidentate complexes require more activation energy to break the bond than do monodentate complexes, so can be more difficult for P desorption to take place in environments where the bidentate complexes predominate between phosphate and the soil surface. In spite of precise investigations using infrared spectroscopy (Parfitt and Atkinson 1976) and X-ray photoelectron spectroscopy (Martin and Smart 1987), assigning peaks among possible complexation types remains controversial (Persson et al. 1996, Arai and Sparks 2001, Kreller et al. 2002). Clarification of the effects of pH and P surface coverage on the formation of different surface-complex types may help to resolve these issues. Environmental and socio-economic restrictions such as low pH, high P fixation capacity, and high costs of P fertilizer and liming materials make P management for humid, acid tropical soils problematic. Management strategies that make the most efficient use of P and lime are required. To this end, one objective should be to raise the soil solution P concentration, since it is directly related to plant P uptake. A small increment in solution P concentration increased through liming-induced P desorption can significantly influence plant growth. For example, Jones and Benson (1975) showed that P in sweet corn leaves and equilibrium solution P levels were correlated, with a critical solution P concentration of 0.12 to 0.13 g P mL The relationship between plant P and equilibrium P concentrations was curvilinear over the range of P concentrations from 0.01 to 1.8 g P mL (Fox and Kamprath 1970).

PAGE 62

49 Combining the results from both parts of our current study we conclude that liming has a twofold positive effect. It reduces the amount of P sorbed and increases the amount of P desorbed. Both these changes induce increased solution P concentration and P bioavailability. Translating these results, in combination with the use of non-conventional liming materials found in rural areas (Ohno and Erich 1990, Erich 1991, Chung and Wu 1997, Van den Berghe and Hue 1999, Nkana et al. 2002), should help rural farmers in tropical areas make the most efficient use of limited agricultural resources.

PAGE 63

50 Table 3-1 Surface soil (0 cm) properties for the Atlantic Forest soil, Bahia, Brazil Property Value Physical Soil texture, % Sand 66 Silt 9 Clay 25 Clay mineralogy, % Kaolinite 80 Goethite 7 Chemical pH 1:2 water 4.7 1:2 KCl 3.6 1:2 CaCl2 3.6 Organic carbon, % 2.51 Total N, % 0.20 Exchangeable Ca 2+ cmol c kg 0.53 Exchangeable Mg 2+ cmol c kg 0.47 Exchangeable K + cmol c kg 0.12 Exchangeable Na + cmol c kg 0.06 Water-extractable (50 mM KCl) P, g g 0.02 Mehlich 1-available P, g g 2.68 Total P, g g 98.29 KCl-exchangeable Al, g g 60.52 NH4 acetate-exchangeable Fe, g g 2.85 AAO extractable Al, % 0.16 AAO extractable Fe, % 0.45 SCD extractable Al, % 0.31 SCD extractable Fe, % 1.30 Acid ammonium-oxalate Sodium citrate-dithionite

PAGE 64

51 04080120160012345 pH 4.7 pH 5.9 pH 7.0 02550751000.00.20.40.60.81 A P sorbed (g g ) B .0 Equilibrium P concentration (g mL 1 ) Figure 3-1 Phosphorus sorption isotherms at 3 different soil pH levels with equilibrium concentration of A) 0 g P mL and B) 0.0.0 g P mL

PAGE 65

52 Table 3-2 Langmuir parameters for P sorption isotherms at 3 different soil pH levels Parameters of Langmuir equation Initial soil pH K b r 2 mL g g g 4.7 1.69b 169.04a 0.970 5.9 1.27ab 177.04ab 0.971 7.0 0.98a 192.12b 0.993 Different letters assigned in a column for each parameter denote significant differences among pH levels at the 0.05 probability level, performed by Tukeys studentized range test.

PAGE 66

53 0501001502002503000.00.51.01.52.02 .5 pH 4.7 pH 5.9 pH 7.0 Partition coefficients ( K d ) E q uilibrium P concentration (g mL 1 ) Figure 3-2 Partition coefficients relative to equilibrium P concentration for P sorption isotherms at 3 different soil pH levels

PAGE 67

54 1001201401600.00.40.81.2 pH 4.7 pH 6.0 pH 7.1 60657075800.000.080.160.24 343638400.000.020.040.06 A B P remained sorbed (g g ) C Equilibrium P concentration (g mL 1 ) Figure 3-3 Phosphorus desorption isotherms at 3 different soil pH levels with 3 different initial P-addition levels. A) 180 g P g B) 80 g P g C) 40 g P g

PAGE 68

55 Table 3-3 Langmuir parameters for P desorption isotherms at 3 different soil pH levels with 3 different initial P-addition levels Parameters of Langmuir isotherm P added Initial pH K b r 2 g g mL g g g 180 4.7 100.05b 145.28b 0.800 6.0 102.88b 129.78a 0.651 7.1 62.99a 128.34a 0780 80 4.7 688.70b 73.86c 0.583 6.0 586.92b 72.16b 0.867 7.1 349.89a 70.73a 0.860 40 4.7 6046.62b 37.19a 0.430 6.0 12999.13c 37.72b 0.338 7.1 2668.71a 37.40a 0.524 Different letters assigned in a column for each parameter at each P-addition level denote significant differences among pH levels at the 0.05 probability level, performed by Tukeys studentized range test.

PAGE 69

56 Table 3-4 Total amount and ratio to P sorbed of P desorbed at 3 different soil pH levels with 3 different initial P-addition levels P added Initial pH P sorbed Total P desorbed Ratio of P desorbed to P initially sorbed g g g g g g % 180 4.7 152.04 18.94a 12.5a 6.0 139.47 21.52a 15.4b 7.1 139.28 25.61b 18.4c 80 4.7 74.91 5.02a 6.7a 6.0 73.65 5.59a 7.6a 7.1 72.54 6.68b 9.2b 40 4.7 37.89 1.77ab 4.7ab 6.0 38.44 1.68a 4.4a 7.1 37.86 2.10b 5.5b Different letters assigned in a column at each P-addition level denote significant differences among pH levels at the 0.05 probability level, performed by Tukeys studentized range test.

PAGE 70

57 0501001500.00.20.40.60.8 pH 4.7 pH 6.0 pH 7.1 0501001500.000.050.100.15 0501001500.000.010.020.030.04 A B Partition coefficients (K d ) C Equilibrium P concentration (g mL 1 ) Figure 3-4 Partition coefficients relative to equilibrium P concentration for P desorption isotherms at 3 different soil pH levels with 3 different initial P-addition levels. A) 180 g P g B) 80 g P g C) 40 g P g

PAGE 71

CHAPTER 4 INFLUENCE OF LOW MOLECULAR WEIGHT ORGANIC ANIONS ON PHOSPHORUS DESORPTION AND BIOAVAILABILITY IN A HUMID BRAZILIAN ULTISOL Introduction Natural ecosystems and crops developing on humid, tropical soils often suffer from phosphorus (P) deficiency. Low levels of native bioavailable inorganic and organic P are common and P, when applied as fertilizer, can be sorbed by Al and Fe oxides. Sorbed P can also form a range of minerals in combination with Ca, Al and Fe; whereby little remains bioavailable (McLean 1976, Hsu 1982). However, some of the sorbed P is still bioavailable through the processes of desorption from the soil surface and dissolution of relatively soluble P mineral compounds. These adsorption/desorption and precipitation/dissolution equilibria control the concentration of P in the soil solution, its chemical mobility, and its bioavailability (Hinsinger 2001). To alleviate P deficiency, soil management practices have been studied under field and laboratory conditions. Particularly important among them are liming and the use of organic and inorganic acids (e.g., from animal wastes) to enhance bioavailability of P in both native and applied forms. The effect of liming on inorganic P availability remains controversial (Haynes 1982) and was discussed in Chapter 3. The effect of organic acids on inorganic P bioavailability is discussed below. The term desorption is used here in two ways. First, it is used as above to represent a process that is simply the reverse of adsorption. Second, it is used in the same fashion 58

PAGE 72

59 as sorption, not distinguishing desorption from dissolution. This is considered necessary until proper terminology consistent with that for sorption is accepted. To clarify the types of desorption that occur in soils, the following categorization is suggested. There are two major mechanisms that describe inorganic P desorption: disequilibria desorption and ligand desorption. These can be subcategorized into exchange and dissolution types of desorption (Table 1-1). Following Le Chteliers Principle, disequilibria desorption requires a disequilibria between solid and solution phases. This disequilibria can be created by removal of solution P, mediated by various processes. As the solution P concentration decreases, P is replenished through desorption of P adsorbed onto the soil surface (disequilibria exchange) or dissolution of relatively insoluble P compounds (disequilibria dissolution) (Wolf and London 1994). Ligand desorption is due to reaction of sorbed P with anions of organic acids such as oxalate, citrate, and malate, that are produced by plant roots and soil microorganisms. These ligands have a higher affinity for the soil surface than phosphate; and form inner-sphere complexes with the soil surface; therefore exchange with P, releasing P to the solution (ligand exchange) (Cambier and Sposito 1991, Hu et al. 2001). Ligands also reduce the stability of precipitates and promote dissolution of solids, releasing P (ligand dissolution) (Zinder et al. 1986, Ludwig et al. 1996). Chemical extractants such as Bray 1, Mehlich 1 and 3, or Olsens solution have been commonly used to extract a portion of the soils bioavailable P; however, they have not been shown to quantify any of the pools described above. These soil P tests are useful only when combined with experimental data (Kumar et al. 1992, Kleinman et al. 2001). Phosphorus fractionation schemes, as represented by Hedley et al. (1982),

PAGE 73

60 sequentially extract inorganic and organic P, that are presumably associated with different soil constituents. Bioavailable P defined by these schemes is usually taken as the sum of readily available (resin-extractable), labile, and microbial (NaHCO 3 plus CHCl 3 /NaHCO 3 extractable) pools. These schemes provide an operational definition of bioavailable P, which does little, however, to either explain pools on a mechanistic basis, or indicate the manner by which the plant accesses bioavailable inorganic P. It also is not clear that they actually measure discrete bioavailable pools. Recently anion exchange resins (AERs) and anion exchange membranes (AEMs) have successfully been used to extract a desorbable pool of inorganic P that has proven to be one of the better indices of plant-available P (Sibbesen 1978, Schoenau and Huang 1991, Ziadi et al. 2001). Since AERs and AEMs behave as a P sink and desorb P from the solid phase using disequilibria, P exhaustingly desorbed by AERs or AEMs should constitute the pool of disequilibria-desorbable P, categorized above, at least for the specific time frame of the extraction. This pool of desorbable P should be bioavailable to any plant root or soil microorganism. A description of the nature and pattern of disequilibria desorption, particularly when described by a desorption isotherm, is rare in the literature and non-existent for tropical soils. On the other hand, quantitative measurements of ligand-desorbable P are not well-established for soils of any climatic region. The problem in defining this pool of desorbable P is that one is not certain which ligands are present in the soil solution; their loading rates onto the solid phase; the volume of soil affected by this loading; and how soil properties control the effectiveness of P release by ligand desorption (Comerford 1998).

PAGE 74

61 Plant roots and soil microorganisms are know to exude a range of low molecular weight organic anions (LMWOAs) such as oxalate, malate, and citrate to the soil solution. The composition of these anions is variable and dependent on plant species, plant age, and physiochemical environment (Curl and Trueglove 1986). Exudation of organic anions from plant roots is enhanced by three environmental stimuli: nutrient deficiency (particularly that of P); exposure to toxic cations (particularly Al 3+ ); and anoxia (Ryan et al. 2001). Excretion of organic anions into the rhizosphere occurs essentially continuously for two reasons. First, the total organic anion concentration of roots ranges from 5 to 50 mM, depending on tissue type and nutrient status of the plant; while that in soil solution is about 1000-fold less. This creates a strong gradient for slow passive diffusion from roots to the soil solution. Second, a substantial electric gradient exists across the plasma membrane because of the operation of ATP-driven proton pumps (H + -ATPases) and cytosolic K + diffusion potential. While the H + expelled into the apoplast by these H + -ATPases creates a charge gradient to facilitate the uptake of cations from the soil, it also tends to draw anions (e.g., citrate 3 and malate 2 ) out of the cells and into the external soil solution (Jones 1998). Once exuded into the soil solution, LMWOAs have the capacity to complex metals both in solution and on the soil surface. The degree of complexation depends on the particular organic anion involved; the concentration and type of metal involved; and pH of the soil solution. Generally speaking, organic anions with a higher number of carboxyl groups (e.g., citrate 3 ) chelate the metals more strongly than do those with fewer carboxyl groups (e.g., oxalate 2 and acetate ) (Ryan et al. 2001). The chemical equilibria

PAGE 75

62 speciation program Geochem-PC (Parker et al. 1995) can predict the ability of organic anions to complex metals as a function of soil solution pH. An example of the complexation of malate, oxalate, and citrate with Fe 3+ is diagramed in Fig. 4-1 (Jones 1998). Organic anions in the soil solution have a positive effect on P bioavailability (Sagwal and Kumar 1995, Cajuste et al. 1996). Studies investigated interaction between organic anions and inorganic P sorption/desorption, and particularly the competitive interactions for surface sorption sites on synthetic goethite (Geelhoed et al. 1998); Al-oxides (Violante et al. 1991); kaolinite and montmorillonite (Kafkafi et al. 1988, He et al. 1992); and field soils (He et al. 1992). Organic anions are effective in decreasing P sorption and increasing P desorption, with their effectiveness being most pronounced between pH 3.0 to 7.0 (Violante and Gianfreda 1993, Liu et al. 1999); and when the ligand is sorbed to the soil surface before P (Violante et al. 1991). Violante et al. (1991) concluded that many mineral soil-surface sites were accessible to both P and organic anions during the sorption process; while other surfaces were specific to one or the other. Ligand exchange of P is dependent on the concentration; sorption kinetics; degradation rate of specific anions involved in the reaction (Fox and Comerford 1992); as well as the type of P bonding complexes formed with the surface metals. Sorption of citrate onto the soil surface increases its resistance to microbial decay, positively affecting the mobilization of adsorbed P (Geelhoed et al. 1999). He et al. (1992) postulated that organic anions modify P-metal bonding complexes from bidentate to monodentate form, also making P more exchangeable with organic anions.

PAGE 76

63 Ligand dissolution is based on the ability of organic anions to chelate metal cations both in solution and on the surface. For example, the stability constants of oxalate and citrate with Fe 3+ (1:1:1 metal:ligand:proton stoichiometric ratio and zero ionic strength at 25 C) are 7.74 and 11.50, respectively (Jones 1998). Lan et al. (1995) showed that inorganic P release from spodic horizons in Florida and Georgia was enhanced by organic anions when the Al stability constant with the anion was greater than about 4.4. When organic anions complex metals in solution, they prevent the precipitation of P by the metal and reduce the activity of P in the solution. By forming complexes with metals on soil surfaces, organic anions promote destabilization of the surface and dissolution of the metals, resulting in P release into the solution (Jones et al. 1996). Given the discussion above, it is clear that there are no well-established techniques for quantitatively measuring the desorbable pools of soil inorganic P, with the possible exception of disequilibria-desorbable P. And yet, measuring these pools and understanding mechanisms involved in the release from each pool are fundamental to understanding and predicting P bioavailability to plants. Our study addressed the above issues through 3 objectives. The first is to consider a method with which to measure ligand-desorbable P. The second objective is to attempt separation of the ligand-exchangeable and ligand-dissolvable pools. The hypothesis underlying this objective is that, since ligand exchange and ligand dissolution are thought to occur over different time scales, a kinetic approach might differentiate between them. The final objective is to use exhaustive desorption method with AEMs differentiate disequilibria-desorbable and ligand-desorbable pools, when fertilizer P is applied at

PAGE 77

64 increasing rates. This tested the hypothesis that the majority of P added to a soil should be, in the short term, dominantly in the disequilibria-desorbable pool. Materials and Methods Soil Material The soil material used in this study was the surface horizon (0 to 30 cm depth) of a Kandiudult (Soil Survey Staff 1999), known as an Argissolo vermelho/vermelhoamarelo distrfico in the Brazilian system (Oliveira et al. 1992). It was taken from under a native Atlantic Forest cover type (Morellato and Haddad 2000) located near Una Ecoparque near the city of Una, Bahia, Brazil. Using a shovel, the material was collected from a 1 ha area. Approximately 20 sampling points were combined; and the soil was subsequently air-dried, passed through a 2-mm sieve, and stored in plastic bags. Selected chemical and physical properties of the soil material were determined by standard methods (Page et al. 1982), which are listed in greater detail in Chapter 3. Soil pH was 4.7 (1:2 soil:water), with 2.5% organic carbon and 25% clay content dominated by kaolinite. Water-extractable P performed using 50 mM KCl was 0.65 nmol g soil. Mehlich 1-available P and total P were 87.18 nmol g soil and 3.17 mol g soil, respectively. The iron content was greater than the aluminum content in both the amorphous and crystalline oxide forms. Ammonium acetate-exchangeable Fe was 0.05 mol g soil. Standard P Fertilization and LMWOAs Extraction Methods The following series of studies used standard procedures for adding P at different rates and extracting P and/or Fe with LMWOAs of varying concentrations. To simplify procedure description, each procedure is described below and then referred to by name within the subsequent method descriptions.

PAGE 78

65 The Sorption Procedure was accomplished by adding 2 g of soil sample to a 50 mL centrifuge tube along with 20 mL of a 50 mM KCl containing P as KH 2 PO 4 at varying concentrations. The P concentrations will be described later for each individual study. The centrifuge tubes were horizontally shaken for 20 h; centrifuged for 15 min at 2500 rpm; and the supernatant was filtered through Whatman No. 5 filter paper. Filtrate P was determined by the method of Murphy and Riley (1962), which was used for all P determinations in our studies. Sorbed P was calculated by difference between P initially added and P left in the filtrate. When LMWOAs were added to the soil, oxalate or citrate solutions as potassium oxalate or potassium citrate of varying concentrations were adjusted to an initial pH of 4.0 with HCl. Twenty milliliters of a given anion solution and 2 g soil samples were used for each procedure. Anion Method 1 was performed in a 30 mL syringe tube that was shaken for varying time periods, as listed below for the specific studies. Each suspension then was filtered immediately through 0.22 m nylon membrane filters (MSI Inc.) without centrifugation. Anion Method 2 was performed in 50 mL centrifuge tubes that were shaken for varying time periods on a horizontal shaker, and then centrifuged for 15 min at 2500 rpm. The supernatant was filtered through Whatman No. 5 filter paper. Filtrate P and Fe for all methods were determined by the method of Murphy and Riley (1962) and by atomic absorption spectroscopy, respectively. When using the Murphy and Riley (1962) method in the presence of oxalate and citrate, care must be taken to maintain those anion concentrations below 2 mM, since concentrations above that may inhibit molybdenum blue color development (He et al. 1998, Appendix B).

PAGE 79

66 Experiments Experiment 1. Optimum Time for Extraction of Ligand-Desorbable P (Objective 1) The Sorption Procedure was used by adding 0 or 5.81 mol P g soil to triplicate samples. The Anion Method 2 was applied to these samples to extract P using oxalate and citrate solutions with 0.02 mmol g soil, at shaking times of 1, 2, 3, 4, 5, 10, 30, 60, 180, 360, or 600 min. The objective of this study was to identify the extraction time that released the most P. Experiment 2. Optimum Anion Concentration for Extraction of Ligand-Desorbable P (Objective 1) The Sorption Procedure was used as for Experiment 1, to triplicate samples. The Anion Method 2 was used to extract P using oxalate and citrate solutions with 0.01, 0.05, 0.10, 0.25, 0.50, or 1.00 mmol g soil, except that 0.01 mmol g soil was not used for the fertilized soils; and a shaking time of 7 h. The objective of this study was to define the anion concentration that released the most P at the optimum shaking time. Experiment 3. Short Kinetic Extraction of Native Soil with Oxalate (Objective2) Triplicate samples of native soil were used with the Anion Method 1, where the oxalate solution was 0, 0.02, or 0.05 mmol g soil; and the shaking times were 1, 2, 3, 4, or 5 min. The objective of this study was to determine if using a low range of shaking times as in Experiment 1 would help differentiate ligand-exchangeable from ligand-dissolvable P. Experiment 4. One-minute Extraction of Native and Fertilized Soils with Oxalate and Citrate at Lower Anion Concentrations (Objective 2) Soil samples at two P levels (0 and 5.81 mol P g soil) were prepared using the Sorption Procedure. All samples were prepared in triplicate; and the Anion Method 1

PAGE 80

67 was used with oxalate and citrate concentrations of 0.01, 0.10, 1.00, or 10.00 mol g soil and with shaking time of 1 min. The objective of this study was to determine if, using a short shaking time, a range of low anion concentrations would help differentiate ligand-exchangeable from ligand-dissolvable P. Experiment 5. Seven-hour Extraction of Native Soil with Oxalate and Citrate at Lower Anion Concentrations (Objective 2) The Anion Method 2 was used to triplicate soil samples of native soil, using oxalate and citrate solutions of 0.01, 0.10, 1.00, or 10.00 mol g soil and using a 7 h shaking time. The objective of this study was to determine if the results of Experiment 4 could be reproduced using a longer shaking time, thus demonstrating the importance of anion concentration versus shaking time in differentiating ligand exchange from ligand dissolution. Experiment 6. Differentiation of Desorbable P into Disequilibriaand Ligand-Desorbable P (Objective 3) Four soil levels of P were prepared in triplicate using the Sorption Procedure, at P rates of 0, 1.29, 2.58, and 5.81 mol P g soil. The first P extraction for these samples was accomplished using sequential extraction with AEMs, as described in Chapter 2. The desorption process was continued until incremental P removed by AEMs was lower than 2 ng P mL in the AEMs extracting solution, corresponding to 0.65.61 nmol P g -1 soil. This is the detection limit of the Murphy and Riley (1962) method when using a 5 cm path length. The pool of disequilibria-desorbable P was defined as the difference between the amount of P initially sorbed and the amount of P remaining on the solid phase after the last AEMs extraction step. The objective of this step was to measure the total pool of disequilibria-desorbable P at different P-fertilization levels.

PAGE 81

68 The samples were then extracted with the Anion Method 2 using a solution of 0.50 mmol oxalate g soil and a 7 hr shaking time. The objective of this step was to measure what we defined above as the total ligand-desorbable pool of P, after disequilibria-desorbable P was extracted. Experiment 7. Mass Balance Verification for Ligand-Desorbable P (Objective 3) Triplicate soil samples of the four P levels of Experiment 6 were extracted using the Anion Method 2 of Experiment 6. The objective of this experiment was to confirm that total P extracted in Experiment 6 was equivalent to the amount extracted if only the Anion Method 2 was used. Experiment 8. Investigation of P Immobilization after Sorption Soil samples, in triplicate, had P applied at a rate of 2.58 mol P g soil, using the Sorption Procedure. Enough samples were prepared to establish 3 treatments that could be destructively sampled over time. The treatments were: 1) P extraction by shaking without HgCl 2 added (+SH); 2) P extraction by shaking with HgCl 2 added (+S+H); and 3) P extraction without shaking and with HgCl 2 added (S+H). Each treatment was extracted for 1, 3, 7, 10, or 14 d; and mercuric chloride was added to the appropriate samples (+H) at a rate of 1.84 mol g soil. The shaking treatment (+S) was continuous shaking on a horizontal shaker; while the unshaken samples (S) were kept in a test-tube rack next to the shaker. At each sampling time, P was extracted from each treatment in triplicate using the Anion Method 2 of Experiment 6. It was shown in previous studies that part of the P sorbed by this soil was not extractable by the Anion Method 2. The objective of this experiment was to determine the role of microbial immobilization (+H vs. H); shaking (+S vs. S); and time of

PAGE 82

69 interaction (short-term ageing of 1 to 14 d) on the non-extractability of P in previous experiments. Mercuric chloride was used because it is an effective soil sterilizant; and it reportedly causes minimal change to soil physio-chemical properties (Wolf et al. 1989). Statistical Analysis Analysis of variance (ANOVA) using a One-Way ANOVA design was used to test differences among treatments using STATISTICA (StatSoft 1999). Significance of the difference among means was determined by t-tests of post hoc comparisons using Tukeys studentized range test at the p = 0.05 level. Means and standard deviations were used to summarize the data from P sorption and desorption isotherms, where applicable. Results and Discussion Ligand-Desorbable P In order to develop a method that measures ligand-desorbable P, an optimum extraction time (Experiment 1) and organic anion loading (Experiment 2) must be identified. Over 80% of the P released from both native and fertilized soils by oxalate and citrate was within the first hour and three hours, respectively (Fig. 4-2). For both anions, and for both native and fertilized soils, P desorption reached maximum levels and stabilized beyond 6 h. Seven hours was chosen as the extraction time for subsequent experiments. Oxalate and citrate loading onto native and fertilized samples produced a diminishing return in desorbable P with respect to increasing anion loading (Fig. 4-3). Maximum P release coincided with a loading of 0.5 mmol of oxalate or citrate g soil. Therefore, this loading was chosen for subsequent experiments.

PAGE 83

70 Ligand-Exchangeable P vs. Ligand-Dissolvable P The premise used in evaluating these data was based on release of Fe from the soil. If only ligand-exchangeable P was being released, the amount of Fe released should not be greater than exchangeable Fe, which was 0.05 mol Fe g soil. Once Fe in the extracting solution was greater than exchangeable Fe, dissolution must have occurred; and it would no longer be possible to differentiate ligand-exchangeable P from ligand-dissolvable P. Another background indicator was water-extractable P, which was 0.65 nmol P g soil. Phosphorus levels higher than this suggested P desorption by the action of the organic ions. Iron released from both native and fertilized soils by 0.02 mmol oxalate g soil was up to 40-fold higher than exchangeable Fe, at a shaking time of 600 min (Fig. 4-2). When 1 to 5 min shaking times were tested, Fe release under the influence of no oxalate added ranged from 0.01 to 0.03 mol Fe g soil (less than exchangeable) for all shaking times (Fig. 4-5). When 0.02 and 0.05 mmol oxalate g soil were added, Fe release exceeded exchangeable Fe (0.20 and 0.35 mol Fe g soil, respectively) even for the 1 min shaking time. The next step was to test if lower anion loadings and shorter contact times would distinguish exchangeable from dissolvable P. Figure 4-6 shows the effect of low anion loadings for a 1 min shaking time. Both oxalate and citrate were effective in desorbing P even at the lowest concentration (0.01 mol g soil). Phosphorus release with oxalate remained stable between 0.01 and 1 mol anion g soil. Iron release at these concentrations was below exchangeable Fe, ranging from 0 to 0.04 mol g soil. Beyond 1 mol anion g soil both Fe and P release increased rapidly with increasing Fe far exceeding the exchangeable amount.

PAGE 84

71 Figure 4-7 shows that the separation of ligand-exchangeable P from ligand-dissolvable P in the native sample is more a function of anion loading than of extraction time. When the same loadings were used with a 7 h shaking time, the same pattern and magnitude of P and Fe were released at anion loadings below 1 mol g soil. It was concluded from above studies as follows: Phosphorus was desorbed at all anion concentrations through the action of anions Phosphorus desorbed below 1 mol anion g soil was dominated by ligand-exchangeable P because Fe release was below exchangeable Fe Phosphorus desorbed below 1 mol anion g soil estimated the total amount of ligand-exchangeable P because o It was relatively stable over a 100-fold increase in anion loading o There was a close correspondence in P desorbed by either anion for both native and fertilized soil o Total P desorbed was the same after the 1 min and 7 h shaking times. It was also concluded that additional P release above 1 mol anion g soil involved ligand dissolution because of the high amount of Fe released with the P. Therefore, the second objective of this study was met because it was shown that ligand-exchangeable and ligand-dissolvable P were analytically separable under specific conditions. It was also concluded that the ligand-exchange reaction was more dependent on anion concentration than on reaction kinetics. A notable result was the increase in ligand-dissolvable Fe from the native to fertilized soils (Fig. 4-6). While citrate-extractable Fe increased from 0.16 to 0.20 mol g soil (25%) with P fertilization; oxalate-extractable Fe increased from 0.15 to 0.34 (127%). Wang et al. (1991) previously noted that mineral dissolution by organic anions is influenced by the degree of P saturation of sorption sites. Such sorption

PAGE 85

72 restricts goethite crystallization, leaving the mineral more oxalate-soluble. Conversely, P sorption reduced the solubility of Fe in Fe(OH) 3 in the presence of 500 M citrate (Jones et al. 1996). It is clear that oxalate and citrate differ in their interaction with this kaolinite-dominated soil. This is shown most clearly by the P:Fe molar ratio relative to each anions concentration in the native soil (Fig. 4-4). Flow calorimetry also showed that oxalate and citrate are sorbed differently onto soil surfaces. Flow calorimetry has been successfully used for measuring reactions at the liquid/solid interface (Rhue et al. 2002). Rhue et al. (2002) documented that phosphate sorption onto soil surfaces was exothermic; while precipitation of Al-phosphate was endothermic. In our study, phosphate and oxalate sorption were found to be exothermic; while citrate sorption was endothermic (Appendix C). Unfortunately, this study does not provide enough information to determine why this difference occurs. Disequilibria-Desorbable P vs. Ligand-Desorbable P Disequilibria-desorbable P and ligand-desorbable P are presented in Table 4-1. The P in disequilibriaand ligand-desorbable pools were a function of the amount of P sorbed, with all desorbable pools increasing with increasing levels of sorbed P. However, the percentage of disequilibria-desorbable P to total desorbable P increased, while the ligand-desorbable P decreased. In the short term, P fertilization preferentially puts P in the disequilibria pool, which is a pool that all plants and soil-borne organisms can access. Notably, it was shown that the percentages of total desorbable P to P initially sorbed remained constant, with only a 6% difference (55 to 61%) across all P-sorption levels. Total desorbable P was measured 7 to 14 d after the P application, which was the time required to exhaust the disequilibria-desorbable pool (Fig. 4-8). The shorter time frames represented less P sorbed, so less time was required for the extraction. When total

PAGE 86

73 desorbable P was measured just 1 d after P addition, these percentages increased to 78 to 87% (Fig. 4-9). It was concluded that the amount of desorbable P was a function of time, with decreasing P recovery rates over time. There also appeared to be a pool of sorbed P that oxalate could not access. This pool ranged between 13 and 22% of the P initially sorbed even after just 1 d of equilibration time. Phosphorus Immobilization/Fixation after Sorption The cumulative P that was immobilized (i.e., inaccessible to Anion Method 2) over 14 d is represented in Figure 4-10. Phosphorus immobilized by the +SH treatment is P immobilized by all factors studied under these laboratory conditions. The +S+H treatment eliminated microbial immobilization, therefore the difference between +SH and +S+H was due to microbial immobilization. The treatments +S+H and S+H should have removed the effect of microbes, so their difference would be the effect of shaking the soil. All treatments were similar after 1d, with an average of 0.23 mol P g soil immobilized (about 10% of the sorbed P). Phosphorus immobilization increased between 1 and 14 d. The P immobilized by microbes, expressed as a percentage of the total immobilized, was 19%, 52%, 56%, and 57% on days 3, 7, 10, and 14, respectively. The P immobilized due to shaking (expressed on the same scale) was 76%, 38%, 30%, and 30% of the total P immobilized. Therefore, P immobilized by what might be considered short-term ageing was 5%, 10%, 14%, and 13% for the time periods listed above. In the short run (1 to 3 d), microbes played less of a role than did shaking; but became the major factor at later dates.

PAGE 87

74 Bliss (2003) showed that, for sandy soil material, microbial immobilization of fertilizer P reached its maximum approximately 3 d after fertilization. Khoshmanesh et al. (2002) reported that only a few hours were necessary for sediment bacteria under aerobic conditions to absorb applied P. Microbial P uptake rapidly contributes to P immobilization. The ageing component of P immobilization was plotted against the square root of time to ascertain if a diffusion mechanism might be implicated (Fig. 4-11). The result was a positive linear relationship with high goodness of fit (r 2 = 0.985). This suggests that P diffusion could be occurring along the clay surfaces or, less likely, into the clay structure itself. Assuming a crystal diffusion coefficient of 10 cm 2 s it was calculated that diffusion could have moved P approximately 0.07 nm into the clay structure by 14 d. Phosphate sorption in soil is known to occur rapidly and to precede an slow sorption process. The former is often thought to be controlled by ligand exchange and the latter by precipitation, penetration of the surface, or diffusion into dead-end pores (Parfitt 1978). Slow sorption is a time-dependent reaction and can account for a substantial portion of total P sorption in goethite-rich materials (Torrent et al. 1992). Duration and extent of the slow reaction apparently depend on the ratio between micropore surface area and total surface area (Torrent et al. 1992), and on the crystallinity of the materials (Strauss et al. 1997). Therefore, it is important to consider the effect of slow sorption as it reduces bioavailable P.

PAGE 88

75 100 Figure 4-1 Ability of three organic anions to complex Fe as a function of pH as predicted using Geochem-PC (Parker et al. 1995) 80 60 40 20 0 % Com p lexed with Fe 3+ 3 4 5 6 7 8 Soil solution pH Malate Oxalate Citrate Fe 3+ and organic anion concentrations = 100 M

PAGE 89

76 0.000.020.040.060.080.1001002003004005006000.001.002.003.00 Oxalate P Citrate P Oxalate Fe 0.001.002.003.0001002003004005006000.001.002.003.00 A Fe released (mol g 1) P desorbed (mol g ) B Shaking time (min) Figure 4-2 Phosphorus-desorption and Fe-release kinetics by organic anions when added at a loading of 0.02 mmol g soil. A) native soil. B) P-fertilized soil

PAGE 90

77 0.000.100.200.300.400.000.200.400.600.801.0005101520 Oxalate P Citrate P Oxalate Fe Citrate Fe 0.001.002.003.004.005.000.000.200.400.600.801.00 A Fe released (mol g 1) P desorbed (mol g ) B Anion concentrations (mmol g 1 ) Figure 4-3 Effect of organic anion loadings for a 7 h interaction time on P desorption and Fe release. A) native soil. B) P-fertilized soil

PAGE 91

78 0.000.040.080.120.160.200.000.200.400.600.801.00 Oxalate Citrate P:Fe ( mol: mol ) Anion concentrations (mmol g 1 ) Figure 4-4 Effect of organic anion loading on the molar ratio of P desorption to Fe release from native soil 0.000.200.400.600.801.000123456 0 0.02 0.05 mmol oxalate/g soil Fe released (mol g ) Shaking time (min) Figure 4-5 Iron release by 3 oxalate loadings as a function of shaking time. The horizontal line represents exchangeable Fe level (0.05 mol Fe g soil)

PAGE 92

79 0.0000.0030.0060.0090.010.11100.000.050.100.150.20 Oxalate P Citrate P Oxalate Fe Citrate Fe 0.000.200.400.600.010.11100.000.100.200.300.400.50 A Fe released (mol g 1) P desorbed (mol g ) B Anion concentrations (mol g 1 ) Figure 4-6 Effect of low anion loadings for a 1 min interaction time on P desorption and Fe release. A) native soil. B) P-fertilized soil

PAGE 93

80 0.000.020.040.060.010.11100.000.300.600.901.201.50 Oxalate P Citrate P Oxalate Fe Citrate Fe P desorbed (mol g ) Fe released (mol g 1 ) Anion concentrations (mol g 1 ) Figure 4-7 Effect of low anion loadings for a 7 h interaction time on P desorption and Fe release

PAGE 94

81 0.000.100.200.300.400.501234567891011121314 1.29 2.58 5.81 mol P/g initially added Cumulative P desorbed (mol g ) Number of desorption steps (d) Figure 4-8 Cumulative P desorbed by AEMs from soil at 3 increasing levels of P. Each desorption step represents one day

PAGE 95

Table 4-1 Disequilibriaand ligand-desorbable pools of P when 3 levels of P are sorbed to the soil 82 P initially added P initially sorbed Disequilibria-desorbable P Ligand-desorbable P Total desorbable P mol g mol g (10 )% mol g % mol g % 1.29 1.24 6.4 a 9a 0.63a 91a 0.69a 56a 2.58 2.33 31.0b 24b 0.98b 76b 1.29b 55a 5.81 4.04 83.8c 34c 1.61c 66c 2.44c 61b % of total desorbable P Total desorbable P as a % of P sorbed Net P desorbed subtracting P desorbed at no P addition Different letters assigned in a column denote significant differences among P-added levels at the 0.05 probability level, performed by Tukeys studentized range test

PAGE 96

83 0%20%40%60%80%100%1.292.585.81 Ligand desorption (7-14 d) Disequilibria desorption (7-14 d) Ligand desorption (1 d) Relative P desorbed to P sorbe d P initially added (mol g 1 ) Figure 4-9 Distribution of sorbed P between the disequilibriaand ligand-desorbable pools. The first bar in each P-added level shows a combination of disequilibriaand ligand-desorbable pools; and the second bar shows total desorbable P 1 d after P sorption

PAGE 97

84 Cumulative P immobilized (mol g ) 0.000.200.400.600.800123456789101112131415 +SH +S+H S+H Day Figure 4-10 Cumulative P immobilization after P sorption. +S and S are with and without shaking; and +H and H are with and without mercuric chloride for microbial suppression Cumulative P immobilized (mol g ) y = 0.0308x 0.03640.000.020.040.060.080.101234 Day r 2 = 0.985 Figure 4-11 Cumulative P immobilization versus the square root of time. Immobilized P is that amount of P immobilized after the effects of microbes and shaking have been removed

PAGE 98

CHAPTER 5 SUMMARY, CONCLUSIONS, AND FUTURE RESEARCH Tropical soils are notorious for their low nutrient bioavailability and low agricultural productivity caused by unfavorable soil properties and socio-economic constraints. Expert nutrient management is further complicated by difficulties in the extension of information. Many parts of tropical Brazil fit this scenario. In humid, acid soils of Brazil, phosphorus deficiency can control crop production, yet the soil dynamics of P are not well-understood at a process level. Because of low pH and Al toxicity, liming is a common soil management practice. However, effect of liming on P bioavailability is controversial and appears related to the amount of exchangeable Al present. Phosphorus sorption and, more importantly, P desorption are processes affecting P bioavailability. A more complete understanding of these processes, including the effect of pH change on P processes, is fundamental to establishing P management strategies. Unfortunately, quantitative measurement of bioavailable P pools is not well-defined. Standard methods are required to attain a meaningful database for P bioavailability. Based on the above observations, this dissertation investigated specific aspects of P availability in soils from three points of view: method development; evaluation of the practical consequences of pH adjustment for soils; and fundamental investigations into the forms of inorganic bioavailable soil P. Chapters 2 and 4 each addressed methodology questions. Desorption is a time-consuming and rarely investigated aspect of P dynamics. Chapter 2 investigated two techniques for measuring P desorption isotherms. Chapter 4 resulted in a method for 85

PAGE 99

86 measuring inorganic P that is desorbable by some low molecular weight organic anions (ligand-desorbable P) that are characteristic of root and microbial exudates. This chapter also addressed whether there were experimental artifacts associated with in this method. In Chapter 3, the effect of pH change on P sorption and desorption was investigated and interpreted with a view toward the degree to which this practice may help or hinder P-fertilizer availability. Chapter 4 further addressed potential pools of bioavailable inorganic P by trying to differentiate ligand-exchangeable from ligand-desorbable inorganic P. It also investigated how fertilizer P is partitioned between these two pools. The combined studies of Chapters 2 to 4 provide detailed information for a specific soil type in Brazil; but also provide a path for continued research on P bioavailability in tropical soils. Each chapter was organized around a set of specific objectives, with a summary of results provided below. Summary Suitability of Using Multiple AEM Strips to Develop Desorption Isotherms The Multiple AEM Method was found suitable for developing well-defined P desorption isotherms. Isotherms were developed quickly and precisely over a range of P sorption levels. This method gave a high density of data points over a 3-day period. Comparison of Multiple AEM Method to Sequential AEM Method When the Multiple AEM Method was compared to the Sequential AEM Method, it was found that the desorption isotherms were numerically different. However, partition coefficients (K d values) calculated from the isotherms were not a function of the desorption method or the amount of P initially sorbed; but were inversely and linearly related to equilibrium P concentration. If the intent of developing desorption isotherms was to derive K d values, then either method was appropriate. It was concluded that over a

PAGE 100

87 range of P solution concentration, the Multiple AEM Method was superior to the Sequential AEM Method because Isotherms were developed in less time, minimizing outside and time-related influences on isotherms It yielded more desorption points in a shorter time frame. Effect of Soil pH on Phosphorus Sorption Phosphorus sorption decreased with increasing pH; and calculated K d values decreased with pH increase at low solution P concentrations. At solution concentrations higher than 3 g P mL there was no effect of pH. Liming to pH values of 5.9 and 7.0 resulted in 2.8 to 4.5 kg P fertilizer ha (calculated at 0.2 g P mL solution concentration) that was not sorbed and remained in the soil solution ready for plant uptake. Effect of Soil pH on Phosphorus Desorption P desorption increased as soil pH increased. Total P desorbed, the ratio of P desorbed to P sorbed, and K d values increased as pH increased for all levels of P addition when soil pH was raised to 7. This resulted in an additional 0.2 to 1.0 and 0.7 to 2.7 kg P ha that was desorbable at pH 6 and 7, respectively, which can be attributed to the effect of liming. Combining the results of P sorption and desorption, liming had a twofold positive effect. It reduced the amount of P sorbed and increased the amount of P desorbed, which increased both solution concentration and P bioavailability. Method for Measuring Ligand-Desorbable Phosphorus Oxalate and citrate were used in kinetic and loading studies to define a method that would remove the total amount of P accessible by these anions in the most efficient

PAGE 101

88 manner. By 6 h the maximum P was removed from the soil, so 7 h was chosen as the shaking time. With a 7 h shake, the maximum amount of P was removed at 0.5 mmol anion g soil loading. This method was used for subsequent studies. Separating Ligand-Desorbable P into Ligand-Exchangeable P and Ligand-Dissolvable P Ligand-desorbable P was successfully separated into ligand-exchangeable P and ligand-dissolvable P via an anion loading approach. Phosphorus release was a function of both anion loading and reaction time, with oxalate and citrate anions being used to release P from native and fertilized soil. Ligand-exchangeable P was released at anion loadings 1 mol anion g soil. It was concluded that this represented the total pool of ligand-exchangeable P because The P desorbed was associated with quantities of Fe release that were less than exchangeable Fe The P desorbed was relatively stabile over a 100-fold increase in anion loading There was a close agreement between the P desorbed by either anion under both native and fertilized conditions The same amount of P was desorbed after 1 min and 7 h reaction times. Above 1 mol anion g soil, desorbed P was a combination of ligand-exchangeable and ligand-dissolvable P. It is recommended that a separation of ligand-desorbable P in this soil be made by first extracting with 1 mol of oxalate or citrate g soil for 1 min, and then with the total ligand desorption procedure of Chapter 4, Objective 1. Oxalate and citrate were different in the way they dissolved Fe, even though they removed similar amounts of ligand-desorbable P. This was evidenced by the P:Fe ratio with increasing anion loadings. While it is impossible to determine from this study the mechanisms causing this discrepancy; flow calorimetry showed that oxalate and citrate

PAGE 102

89 have different surface reactions with this soil. Oxalate, like phosphate, sorption is an exothermic reaction; while the sorption of citrate is endothermic. Differentiating Bioavailable Pools of Applied P into Disequilibria-Desorbable P and Ligand-Desorbable P The majority of P sorbed by the soil was found in the ligand-desorbable pool. As P was sorbed to the soil, it was preferentially found in the disequilibria-desorbable pool. The percentage of total desorbable P to sorbed P was independent of the amount of P added. Fifty-five to 61% of the sorbed P was desorbable by oxalate. Thirty-nine to 45% was fixed against oxalate desorption. The recovery of sorbed P was discovered to be a function of time over a 14 d period, with the P fixed in 1 d being 13 to 22% of the sorbed P and P fixed in 14 d represented by the numbers above. This ageing process was demonstrated to be an artifact of shaking (30%), microbial immobilization (57%), and sorption of P onto the soil in non-oxalate-accessible form (13%). The amount fixed by ageing was linear with the square root of time, suggesting a diffusion mechanism. Conclusions Soil pH effects on P sorption and desorption are still controversial for sorption and virtually unstudied for desorption. This is the first study to develop a multiple AEMs procedure for constructing a desorption isotherm, and is also the first study to use isotherms in evaluating the effect of pH on P desorption. The method developed here should promote greater progress in desorption studies; while the data relative to P sorption and desorption give order-of-magnitude effects for enhanced fertilizer availability due to liming.

PAGE 103

90 This dissertation defines bioavailable pools of P via a categorization that is different than that found in the literature. While it is not the first time disequilibria and ligand desorption have been investigated, it is the first time that they have been investigated together, and is the first time that the influence of P fertilization on the distribution of P between forms has been investigated. This is also the first time that ligand-exchangeable P has been individually measured for a soil. This dissertation has increased our database concerning measurement of P forms in soils and interactions of inorganic P with the soil surface. It has also advanced our methodology for measuring P in meaningful bioavailable pools, and has helped to better define bioavailable P and P dynamics under influences of soil pH and organic anions. Not all questions were adequately answered. However other research possibilities have been suggested by the results and problems associated with these experiments. Future Research A variety of future research topics/directions were suggested by the results of the preceding experiments. 1. The Multiple AEM Method should be tested on a wider range of soils over a range of solution P concentrations in order to ensure its usefulness. 2. Since this dissertation did not address mechanisms of P desorption in relation to soil pH, clarification of the effects of pH and P surface coverage on the formation of different surface complexes should help to resolve these issues. 3. Although liming was shown to have positive effects on P bioavailability, high costs of liming materials and P fertilizers in tropical regions remains constraints. Investigations on the interactions of alternative liming materials such as wood ash and composts with soils and plants; and their effect on P bioavailability would be useful. Positive results of these studies would allow rural farmers to manage phosphorus and promote increased cropland productivity. 4. This dissertation showed that oxalate and citrate reacted differently with Fe and were sorbed differently to the soil surface. Although both were effective in desorbing P

PAGE 104

91 from the surface, it would be fruitful to pursue an understanding of surface reaction mechanisms for these anions.

PAGE 105

92 APPENDIX A ACID BASE TITRATION CURVE FOR THE SOIL An acid base titratio n curve was developed for the Atlantic Forest soil using 10 g of soil in 20 mL of double deionized water. Soil pH was measured at 2 h of continuous stirring of the suspension after each increment of 0.083 M NaOH added. Figure A 1 Titration curve for the Atlantic Forest soil y = -0.0025x 2 + 0.1455x + 4.9965 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 0 5 10 15 20 25 30 r 2 = 0.995 pH Amount of NaOH used (mol g 1 )

PAGE 106

APPENDIX B DETERMINATION OF SOLUBLE P IN THE PRESENCE OF ORGANIC ANIONS Table B-1 Effect of varying concentrations of oxalate and citrate on absorbance over a range of standard P concentrations Phosphorus Standard (g mL ) 0.02 0.05 0.20 0.40 Absorbance 0.014 0.036 0.148 0.265 Oxalate % mM 0.0 100.0 100.0 100.0 100.0 0.1 100.0 100.0 98.7 99.3 1.0 100.1 97.3 98.0 99.0 2.0 78.6 94.3 99.3 99.0 4.0 7.0 0.0 0.0 0.0 6.0 0.0 0.0 0.0 0.0 Citrate mM % 0.0 100.0 100.0 00.0 100.0 0.1 100.0 97.3 98.7 99.3 1.0 92.9 91.9 97.3 98.0 2.0 100.0 100.1 100.1 99.3 4.0 14.3 7.1 50.0 68.0 6.0 0.0 2.9 2.7 10.2 Measured at 880 nm wavelength Percentage of P absorbance in the presence of organic anions to P standard absorbance 93

PAGE 107

APPENDIX C SURFACE REACTIONS OF PHOSPHATE, OXALATE, AND CITRATE MEASURED BY FLOW CALORIMETRY Flow calorimetry was used to measure heats of reaction for the sorption of phosphate, oxalate and citrate onto the non-fertilized soil sample described in Chapter 2. The method was based on that described by Rhue et al. (2002). A 67.4 mg soil sample was initially equilibrated with a solution of 50 mM KCl in a glass column under a flow rate of 0.39 mL min Figure C-1 shows the heat of reaction of phosphate sorption with time when the solution was changed to 50 mM KCl containing 5 mM KH 2 PO 4 (KCl-P solution). Phosphate sorption onto the soil was exothermic. The KCl-P solution was leached through the column to P-saturate the samples sorption sites. Curve 1 in Figure C-2 shows subsequent oxalate sorption when the solution was changed to 50 mM KCl plus 10 mM oxalate as potassium oxalate. Oxalate sorption was exothermic. The solution was again changed back to the KCl-P solution; and P was resorbed onto the soil. Phosphate sorption gave the same exothermic reaction (Curve 2 in Fig. C-2). The soil was again saturated with P as above; and the solution then was changed to sorb citrate onto the surface by passing a 50 mM KCl solution containing 10 mM citrate as potassium citrate through the column. Curve 1 in Figure C-3 indicates that the sorption of citrate was endothermic. Finally, the solution was changed back to the KCl-P solution; and P was resorbed again. Phosphate sorption continued to be exothermic (Curve 2 in Fig. C-3). 94

PAGE 108

95 0.000.020.040.060.080.100510152025303540 Phosphate millivolts Time (min) Figure C-1 Heats of reaction for phosphate sorption onto the study-site soil

PAGE 109

96 0.020.040.060.080.100.120510152025303540 Phosphate Oxalate Curve 2 millivolts Curve 1 Time (min) Figure C-2 Heats of reaction for oxalate sorption onto P-saturated soil preceding phosphate sorption onto oxalate-treated soil

PAGE 110

97 0.040.060.080.100.120.140510152025303540 Phosphate Citrate Curve 2 millivolts Curve 1 Time (min) Figure C-3 Heats of reaction for citrate sorption onto P-saturated soil preceding phosphate sorption onto citrate-treated soil

PAGE 111

LIST OF REFERENCES Anderson, G. 1980. Assessing organic phosphorus in soils. p. 411. In F.E. Khasawneh et al. (ed.) The role of phosphorus in agriculture. ASA, Madison, WI. Anjos, J.T., and D.L. Rowell. 1987. The effect of lime on phosphorus adsorption and barley growth in three acid soils. Plant Soil 103:75. Arai, Y., and D.L. Sparks. 2001. ATR-FTIR spectroscopic investigation on phosphate adsorption mechanisms at the ferrihydrite-water interface. J. Colloid Interface Sci. 241:317. Arias, J.S., and P.G. Fernandez. 2001. Changes in phosphorus adsorption in a Palexerult amended with limestone and/or gypsum. Commun. Soil Sci. Plant Anal. 32:751758. Awan, A.B. 1964. Effect of lime on availability of phosphorus in Zamorano soils. Soil Sci. Soc. Am. Proc. 28:672. Bache, B.W., and C. Ireland. 1980. Desorption of phosphate from soils using anion exchange resins. J. Soil Sci. 31:297. Ballard, R. 1980. Phosphorus nutrition and fertilization of forest trees. p. 763. In F.E. Khasawneh et al. (ed.) The role of phosphorus in agriculture. ASA, Madison, WI. Barber, S.A. 1980. Soil-plant interactions in the phosphorus nutrition of plants. p. 591615. In F.E. Khasawneh et al. (ed.) The role of phosphorus in agriculture. ASA, Madison, WI. Barber, S.A. 1984. Soil nutrient bioavailability, a mechanistic approach. A Wiley-Interscience Publication, New York, NY. Barber, S.A., J.M. Walker, and E.H. Vasey. 1962. Principles of ion movement through the soil to the plant root. p. 121. In Proc. Int. Soil Conf. New Zealand. Barros Filho, N.F., N.B. Comerford, and N.F. Barros. Phosphorus sorption, desorption and resorption in soils of the Brazilian Cerrado. Biomass Bioenergy. (in press). 98

PAGE 112

99 Barros Filho, N.F., N.B. Comerford, N.F. Barros, and R.C. Santana. 2001. Dessoro de fsforo do solo por meio de resina aninica em membrana. (In Portuguese) Anais do XXVIII Congresso Brasileiro de Cincia do Solo, Londrina, 16-22 de Julho, 2001. Soc. Brasileira de Ci. do Solo, Viosa, MG, Brasil. Barrow, N.J. 1983. On the reversibility of phosphate sorption by soils. J. Soil Sci. 34:751. Barrow, N.J. 2002. Influence of pH on a secondary effect of phosphate reaction: The decrease in sorption of newly added phosphate. Aust. J. Soil Res. 40:775. Bar-Yosef, B., U. Kafkafi, R. Rosenberg, and G. Sposito. 1988. Phosphorus adsorption by kaolinite and montmorillonite: I. Effect of time, ionic strength, and pH. Soil Sci. Soc. Am. J. 52:1580. Bhatti, J.S., and N.B. Comerford. 2002. Measurement of phosphorus desorption from a spodic horizon using two different desorption methods and pH control. Commun. Soil Sci. Plant Anal. 33:845. Bliss, C.M. 2003. The effect of phosphorus fertilization on the microbial phosphorus pool in a Spodosol under a slash pine plantation. Ph.D. Dissertation. Univ. of Florida, Gainesville, FL. Bowden, J.W., S. Nagarajah, N.J. Barrow, and J.P. Quirk. 1980. Describing the adsorption of phosphate, citrate and selenite on a variable-charge mineral surface. Aust. J. Soil Res. 18:49. Brady, N.C., and R.R. Weil. 1999. The nature and properties of soils, 12th ed. Prentice Hall, Upper Saddle River, NJ. Brewster, J.L., A.N. Gancheva, and P.H. Nye. 1975. The determination of desorption isotherms for soil phosphate using low volumes of solution and an anion exchange resin. J. Soil Sci. 26:364. Brookes, P.C., D.S. Powlson, and D.S. Jenkinson. 1984. Phosphorus in the soil microbial biomass. Soil Biol. Biochem. 16:169. Cabrera, F., P. De Arambarri, L. Madrid, and C.G. Toca. 1981. Desorption of phosphate from iron oxides in relation to equilibrium pH and porosity. Geoderma 26:203. Cajuste, L.J., R.J. Laird, L. Cajuste-B. Jr., and B.G. Cuevas. 1996. Citrate and oxalate influence on phosphate, aluminum, and iron in tropical soils. Commun. Soil Sci. Plant Anal. 27:1377. Cambier, P., and G. Sposito. 1991. Adsorption of citric-acid by synthetic pseudoboehmite. Clays Clay Miner. 39:369.

PAGE 113

100 Chen, J., and S.A. Barber. 1990. Effect of liming and adding phosphate on predicted phosphorus uptake by maize on acid soils of three soil orders. Soil Sci. 150:844850. Chen, M., and L.Q. Ma. 2001. Taxonomic and geographic distribution of total phosphorus in Florida surface soils. Soil Sci. Soc. Am. J. 65:1539. Chien, S.H., and W.R. Clayton. 1980. Application of Elovich equation to the kinetics of phosphate release and sorption in soils. Soil Sci. Soc. Am. J. 44:265. Chung, R.S., and S.H. Wu. 1997. Effect of corncob compost on plant growth in an acid red soil. Commun. Soil Sci. Plant Anal. 28:673. Clarholm, M. 1993. Microbial biomass P, labile P, and acid phosphatase activity in the humus layer of a spruce forest, after repeated additions of fertilizers. Biol. Fertil. Soils 16:287. Cole, C.V., E.T. Elliot, H.W. Hunt, and D.C. Coleman. 1978. Tropic interactions in soils as they affect energy and nutrient dynamics. V. Phosphorus transformations. Microbiol. Ecol. 4:381. Comerford, N.B. 1998. Soil phosphorus bioavailability. p. 136. In J.P. Lynch and J. Deikman (ed.) Phosphorus in plant biology: Regulatory roles in molecular, cellular, organismic, and ecosystem processes. Am. Soc. Plant Physiologists, Rockville, MD. Condron, L.M., M.R. Davis, R.H. Newman, and I.S. Cornforth. 1996. Influence of conifers on the forms of phosphorus in selected New Zealand grassland soils. Biol. Fertil. Soils 21:37. Cooperband, L.R., P.M. Gale, and N.B. Comerford. 1999. Refinement of the anion exchange membrane method for soluble phosphorus measurement. Soil Sci. Soc. Am. J. 63:58. Cooperband, L.R., and T.J. Logan. 1994. Measuring in situ changes in labile soil phosphorus with anion-exchange membranes. Soil Sci. Soc. Am. J. 58:105. Cresser, M.S., K. Killham, and T. Edwards. 1993. Soil chemistry and its applications. Cambridge Univ. Press, Cambridge, UK. Csatho, P., M. Magyar, K. Debreezeni, and K. Sardi. 2002. Correlation between soil P and corn leaf P contents in a network of Hungarian long-term field trials. Commun. Soil Sci. Plant Anal. 33:3085. Curl, E.A., and B. Trueglove. 1986. The rhizosphere. Springer-Verlag. Berlin. Curtin, D., and J.K. Syers. 2001. Lime-induced changes in indices of soil phosphate availability. Soil Sci. Soc. Am. J. 65:147.

PAGE 114

101 Delgado, A., and J. Torrent. 1997. Phosphate-rich soils in the European Union: Estimating total plant-available phosphorus. Eur. J. Agron. 6:205. De Smet, J., J. Vanderdeelen, and G. Hofman. 1998. Effect of soil properties on the kinetics of phosphate release. Commun. Soil Sci. Plant Anal. 29:2135. Dumas, J.B.A. 1831. Procds de l'analyse organique. Ann. Chim. Phys. 247:198. Enwezor, W.O. 1967. Significance of C-organic P ratio in mineralization of soil organic phosphorus. Soil Sci. 103:62. Erich, M.S. 1991. Agronomic effectiveness of wood ash as a source of phosphorus and potassium. J. Environ. Qual. 20:576. Feng, G., Y.C. Song., X.L. Li, and P. Christie. 2003. Contribution of arbuscular mycorrhizal fungi to utilization of organic sources of phosphorus by red clover in a calcareous soil. Appl. Soil Ecol. 22:139. Fernandes, M.L.V., and G.P. Warren. 1994. Exchangeable and non-exchangeable phosphate sorption in Portuguese soils. Fert. Res. 37:23. Fontes, M.P.F., and S.B. Weed. 1996. Phosphate adsorption by clays from Brazilian Oxisols: Relationships with specific surface area and mineralogy. Geoderma 72:37. Fox, R.L., and E.J. Kamprath. 1970. Phosphate sorption isotherms for evaluating the phosphate requirements of soils. Soil Sci. Soc. Am. Proc. 34:902. Fox, T.R., and N.B. Comerford. 1990. Low-molecular-weight organic acids in selected forest soils of the southeastern USA. Soil Sci. Soc. Am. J. 54:1139. Fox, T.R., and N.B. Comerford. 1992. Influence of oxalate loading on phosphorus and aluminum solubility in Spodosols. Soil Sci. Soc. Am. J. 56:290. Fried, M., and R.E. Shapiro. 1961. Soil plant relationships in ion uptake. Annu. Rev. Plant Physiol. 12:91. Gee, G.W., and J.W. Bauder. 1986. Particle size analysis. p. 383. In A. Klute (ed.) Methods of soil analysis, Part 1: Physical and mineralogical methods, 2nd ed. Agronomy Monograph 9. ASA and SSSA, Madison, WI. Geelhoed, J.S., T. Hiemstra, and W.H. Van Riemsdijk. 1998. Competitive interaction between phosphate and citrate on goethite. Environ. Sci. Technol. 32:2119. Geelhoed, J.S., W.H. Van Riemsdijk, and G.R. Findenegg. 1997. Effects of sulfate and pH on the plant-availability of phosphate adsorbed on goethite. Plant Soil 197:241249.

PAGE 115

102 Geelhoed, J.S., W.H. Van Riemsdijk, and G.R. Findenegg. 1999. Simulation of the effect of citrate exudation from roots on the plant availability of phosphate adsorbed on goethite. Eur. J. Soil Sci. 50:379. Gilkes, R.J., and J.C. Hughes. 1994. Sodium-fluoride pH of south-western Australian soils as an indicator of P-sorption. Aust. J. Soil Res. 32:755. Grierson, P.F., N.B. Comerford, and E.J. Jokela. 1999. Phosphorus mineralization and microbial biomass in a Florida Spodosol: Effects of water potential, temperature and fertilizer application. Biol. Fertil. Soils 28:244. Harris, W.G., R.D. Rhue, G. Kidder, R.B. Brown, and R. Littell. 1996. Phosphorus retention as related to morphology of sandy coastal plain soil materials. Soil Sci. Soc. Am. J. 60:1513. Haynes, R.J. 1982. Effects of liming on phosphate availability in acid soils A critical review. Plant Soil 68:289. Haynes, R.J. 1984. Lime and phosphate in the soil-plant system. p. 249. In N.C. Brady (ed.) Advances in Agronomy. Academic Press, NY. Haynes, R.J., and R.S. Swift. 1989. The effects of pH and drying on adsorption of phosphate by aluminium-organic matter associations. J. Soil Sci. 40:773. He, Z.L., V.C. Baligar, D.C. Martens, and K.D. Ritchey. 1998. Determination of soluble phosphorus in the presence of organic ligands or fluoride. Soil Sci. Soc. Am. J. 62:1538. He, Z.L., X. Yang, K.N. Yuan, and Z.X. Zhu. 1994. Desorption and plant-availability of phosphate sorbed by some important minerals. Plant Soil 162:89. He, Z.L., K.N. Yuan, and Z.X. Zhu. 1992. Effects of organic anions on phosphate adsorption and desorption from variable-charge clay minerals and soil. Pedosphere 2:1-11. He, Z.L., and J. Zhu. 1998. Microbial utilization and transformation of phosphate adsorbed by variable charge minerals. Soil Biol. Biochem. 30:917. He, Z.L., Z.X. Zhu, and K.N. Yuan. 1989. Desorption of phosphate from some clay minerals and typical soil groups of China: II. Effect of pH on phosphate desorption. Acta Agriculturae Universitatis Zhejianggensis 15:441. Hedley, M.J., J.W. Stewart, and B.S. Chauhan. 1982. Changes in inorganic and organic soil phosphorus fractions induced by cultivation practices and by laboratory incubations. Soil Sci. Soc. Am. J. 46:970.

PAGE 116

103 Hingston, F.J., A.M. Posner, and J.P. Quirk. 1974. Anion adsorption by goethite and gibbsite. II. Desorption of anions from hydrous oxide surfaces. J. Soil Sci. 25:1626. Hinsinger, P. 2001. Bioavailability of soil inorganic P in the rhizosphere as affected by root-induced chemical changes: A review. Plant Soil 237:173. Holford, I.C.R., B.E. Schweitzer, and G.J. Crocker. 1994. Long-term effects of lime on soil-phosphorus solubility and sorption in 8 acidic soils. Aust. J. Soil Res. 32:795803. Hsu, P.H. 1982. Crystallization of iron(III) phosphate at room temperature. Soil Sci. Soc. Am. J. 46:928. Hu, H.Q., J.Z. He, X.Y. Li, and F. Liu. 2001. Effect of several organic acids on phosphate adsorption by variable charge soils of central China. Environ. Int. 26:353. Jayachandran, K., A.P. Schwab, and B.A.D. Hetrick. 1992. Mineralization of organic phosphorus by vesicular-arbuscular mycorrhizal fungi. Soil Biol. Biochem. 24:897. Joergensen, R.G., H. Kbler, B. Meyer, and V. Wolters. 1995. Microbial biomass phosphorus in soils of beech (Fagus sylvatica L.) forests. Biol. Fertil. Soils 19:215219. Joner, E.J., I.M. Van Aarle, and M. Vosatka. 2000. Phosphatase activity of extra-radical arbuscular mycorrhizal hyphae: A review. Plant Soil 226:199. Jones, D.L. 1998. Organic acids in the rhizosphere A critical review. Plant Soil 205:2544. Jones, D.L., P.R. Darrah, and L.V. Kochian. 1996. Critical evaluation of organic acid mediated iron dissolution in the rhizosphere and its potential role in root iron uptake. Plant Soil 180:57. Jones, J.P., and J.A. Benson. 1975. Phosphate sorption isotherms for fertilizer P needs of sweet corn (Zea mays) grown on a high phosphorus fixing soil. Commun. Soil Sci. Plant Anal. 6:465. Jones, J.P., and R.L. Fox. 1978. Phosphorus nutrition of plants influenced by manganese and aluminum uptake from an Oxisol. Soil Sci. 126:230. Kafkafi, U., B. Bar-Yosef, R. Rosenberg, and G. Sposito. 1988. Phosphorus adsorption by kaolinite and montmorillonite: II. Organic anion competition. Soil Sci. Soc. Am. J. 52:1585.

PAGE 117

104 Kafkafi, U., A.M. Posner, and J.P. Quirk. 1967. Desorption of phosphate from kaolinite. Soil Sci. Soc. Am. Proc. 31:348. Khoshmanesh, A., B.T. Hart., A. Duncan, and R. Beckett. 2002. Luxury uptake of phosphorus by sediment bacteria. Water Res. 36:774. Kleinman, P.J.A., A.N. Sharpley, K. Gartley, W.M. Jarrell, S. Kuo, R.G. Menon, R. Myers, K.R. Reddy, and E.O. Skogley. 2001. Interlaboratory comparison of soil phosphorus extracted by various soil test methods. Commun. Soil Sci. Plant Anal. 32:2325. Kreller, D.I., G. Gibson, G.W. Van Loon, and J.H. Horton. 2002. Chemical force microscopy investigation of phosphate adsorption of the surfaces of iron(III) oxyhydroxide particles. J. Colloid Interface Sci. 254:205. Kumar, V., R.J. Gilkes, and M.D.A. Bolland. 1992. A comparison of 7 soil-P tests for plant-species with different external-P requirements grown on soils containing rock phosphate and superphosphate residues. Fert. Res. 33:35. Kuo, S., E.J. Jellum, and W.L. Pan. 1988. Influence of phosphate sorption parameters of soils on the desorption of phosphate by various extractants. Soil Sci Soc. Am. J. 52:974. Lal, R. 2000. Physical management of soils of the tropics: Priorities for the 21st century. Soil Sci. 165:191. Lan, M., N.B. Comerford, and T.R. Fox. 1995. Organic anions effect on phosphorus release from spodic horizons. Soil Sci. Soc. Am. J. 59:1745. Lindsay, W.L., and E.C. Moreno. 1960. Phosphate phase equilibria in soils. Soil Sci. Soc. Am. Proc. 24:177. Liu, F., J.Z. He., C. Colombo, and A. Violante. 1999. Competitive adsorption of sulfate and oxalate on goethite in the absence or presence of phosphate. Soil Sci. 164:180189. Lopez-Hernandez, D., and C.P. Burnham. 1974. The effect of pH on phosphate adsorption in soils. J. Soil Sci. 25:207. Ludwig, C., J.L. Devidal, and W.H. Casey. 1996. The effect of different functional groups on the ligand-promoted dissolution of NiO and other oxide minerals. Geochim. Cosmochim. Acta 60:213. Macariola-See, N., H.J. Woodard, and T. Schumacher. 2003. Field verification of the Barber-Cushman mechanistic phosphorus uptake model for maize. J. Plant Nutr. 26:139.

PAGE 118

105 Madrid, L., and A.M. Posner. 1979. Desorption of phosphate from goethite. J. Soil Sci. 30:697. Martin, M., L. Celi, and E. Barberis. 2002. Extractability and plant availability of phosphate from P-goethite complexes. Commun. Soil Sci. Plant Anal. 33:143. Martin, R.R., and R.S.C. Smart. 1987. X-ray photoelectron studies of anion adsorption on goethite. Soil Sci. Soc. Am. J. 51:54. McBride, M.B. 1994. Environmental chemistry of soils. Oxford Univ. Press. New York, NY. McKeague, J.A., and J.H. Day. 1966. Dithioniteand oxalate-extractable Fe and Al as aids in differentiating various classes of soils. Can. J. Soil Sci. 46:13. McLean, E.O. 1976. Chemistry of soil aluminum. Commun. Soil Sci. Plant Anal. 7:619636. Mehlich, A. 1953. Determination of P, Ca, Mg, K, Na, and NH 4 by North Carolina soil testing laboratories. North Carolina Soil Test Div., Mimeo, North Carolina Dept. of Agr., Raleigh, NC. Menon, R.G., S.H. Chien, and W.J. Chardon. 1996/1997. Iron oxide impregnated filter paper (Pi test). II. A review of its application. Nutr. Cycl. Agroecosys. 47:7. Mokwunye, U. 1975. The influence of pH on the adsorption of phosphate by soils from the Guinea and Sudan Savannah zones of Nigeria. Soil Sci. Soc. Am. Proc. 39:1100. Mora, M.L., G. Baeza, C. Pizarro, and R. Demanet. 1999. Effect of calcitic and dolomitic lime on physicochemical properties of a Chilean Andisol. Commun. Soil Sci. Plant Anal. 30:427. Morellato, L.P.C., and C.F.B. Haddad. 2000. Introduction: The Brazilian Atlantic Forest. Biotropica 32:786. Murphy, J., and J.P. Riley. 1962. A modified single solution method for the determination of phosphate in natural waters. Anal. Chim. Acta 27:31. Naidu, R., J.K. Syers, R.W. Tillman, and J.H. Kirkman. 1990. Effect of liming on phosphate sorption by acid soils. J. Soil Sci. 41:165. Nelson, S.R., and L.E. Sommers. 1982. Total carbon, organic carbon, and organic matter. p. 539. In A.L. Page et al. (ed.) Methods of soil analysis, Part 2: Chemical and microbiological properties, 2nd ed. Agronomy Monograph 9. ASA and SSSA, Madison, WI.

PAGE 119

106 Nkana, J.C.V., A. Demeyer, and M.G. Verloo. 2002. Phosphorus fertilization in a tropical acid soil under simulated slash and burn conditions: Effects on uptake and adsorption of phosphorus. Agrochimica 46:291. Nuernberg, N.J., J.E. Leal., and M.E. Sumner. 1998. Evaluation of an anion-exchange membrane for extracting plant available phosphorus in soils. Commun. Soil Sci. Plant Anal. 29:467. Nye, P.H., and P.B. Tinker. 1977. Solute movement in the soil-root system. Univ. California Press, Berkeley, CA. Ohno, T., and M.S. Erich. 1990. Effect of wood ash application on soil pH and soil test nutrients levels. Agric. Ecosyst. Environ. 32:223. Okajima, H., H. Kubota, and T. Sakuma. 1983. Hysteresis in the phosphorus sorption and desorption process of soils. Soil Sci. Plant Nutr. 29:271. Oliveira, A.T., and M.A.L. Fontes. 2000. Patterns of floristic differentiation among Atlantic forests in southeastern Brazil and the influence of climate. Biotropica 32:793. Oliveira, J.B., P.K.T. Hacomine, and M.N. Camargo. 1992. Classes gerais de solos do Brasil. (In Portuguese) 2nd ed. FUNEP. Jaboticabal. Page, A.L., R.H. Miller, and D.R. Keeney. 1982. Methods of soil analysis. Part 2: Chemical and microbiological properties, 2nd ed. Agronomy Monograph 9. ASA and SSSA, Madison. WI. Pardo, M.T., M.E. Guadalix, and M.T. Garcia-Gonzalez. 1992. Effect of pH and background electrolyte on P sorption by variable charge soils. Geoderma 54:275284. Parfitt, R.L. 1977. Phosphate adsorption on an Oxisol. Soil Sci. Soc. Am. J. 41:10641067. Parfitt, R.L. 1978. Anion adsorption by soils and soils materials. p. 1. In N.C. Brady (ed.) Advances in Agronomy, Vol. 30. Academic Press, New York, NY. Parfitt, R.L. 1979. The availability of P from phosphate-goethite bridging complexes. Desorption and uptake by ryegrass. Plant Soil 53:55. Parfitt, R.L., and R.J. Atkinson. 1976. Phosphate adsorption on goethite (-FeOOH). Nature 264:740. Parfitt, R.L., K.R. Tate., and R.B. McKercher. 1994. Measurement of phosphorus mineralization using an anion-exchange membrane. Commun. Soil Sci. Plant Anal. 25:3209.

PAGE 120

107 Parker, D.R., R.L. Chaney, and W.A. Norvell. 1995. Chemical equilibria models: Applications to plant research. p. 163. In R.H. Loeppert et al. (ed.) Chemical equilibria and reaction models. Special Publication 42. SSSA-ASA, Madison, WI. Pereira, J.R., and C.M.B. De Faria. 1998. Phosphorus sorption in some soils of the semi-arid region of the Brazilian Northeast. Pesqui. Agropecu. Bras. 33:11791184. Persson, P., N. Nilsson, and S. Sjoberg. 1996. Structure and bonding of orthophosphate ions at the iron oxide aqueous interface. J. Colloid Interface Sci. 177:263. Quintero, C.E., G.N. Boschetti, and R.A. Benavidez. 1999. Phosphorus retention in some soils of the Argentinean Mesopotamia. Commun. Soil Sci. Plant Anal. 30:14491461. Raven, K.P., and L.R. Hossner. 1993. Phosphorus desorption quantity-intensity relationships in soil. Soil Sci. Soc. Am. J. 57:1501. Raven, K.P., and L.R. Hossner. 1994. Soil phosphorus desorption kinetics and its relationship with plant growth. Soil Sci. Soc. Am. J. 58:416. Rhue, R.D., C. Appel, and N. Kabengi. 2002. Measuring surface chemical properties of soil using flow calorimetry. Soil Sci. 167:782. Rupa, T.R., K.P. Tomar, C.H. Srinivasa Rao, and A. Subba Rao. 2001. Kinetics of phosphate sorption-desorption as influenced by soil pH and electrolyte. Agrochimica 45:124. Ryan, P.R., E. Delhaize, and D.L. Jones. 2001. Function and mechanism of organic anion exudation from plant roots. Annu. Rev. Plant Physiol. Plant Mol. Biol. 52:527. Ryden, J.C., J.R. McLaughlin, and J.K. Syers. 1977. Mechanisms of phosphate sorption by soils and hydrous ferric oxide gel. J. Soil Sci. 28:72. Saggar, S., M.J. Hedley, and R.E. White. 1990. A simplified resin membrane technique for extracting phosphorus from soils. Fert. Res. 24:173. Sagwal, O.P., and V. Kumar. 1995. Effect of organic and inorganic anions on availability of phosphorus applied through different phosphorus fertilizers to moong-bean grown in soil. Commun. Soil Sci. Plant. Anal. 26:3181. Sanchez, P.A., and G. Uehara. 1980. Management considerations for acid soils with high phosphorus fixation capacity. p. 471. In F.E. Khasawneh et al. (ed.) The role of phosphorus in agriculture. ASA, Madison, WI. Sanyal, S.K., S.K. De Datta, and P.Y. Chan. 1993. Phosphate sorption-desorption behavior of some acidic soils of south and southeast Asia. Soil Sci. Soc. Am. J. 57:937.

PAGE 121

108 SAS Institute. 1988. SAS/STAT users guide, release 6.03 edition. SAS Institute Inc., Cary, NC. Saunders, W.M.H. 1965. Phosphate retention by New Zealand soils and relationships to free sesquioxides, organic matter and other soil properties. N. Z. J. Agric. Res. 8:30. Schindler, P.W., and G. Sposito. 1991. Surface complexation at (hydr)oxide surfaces. p. 115. In G.H. Bolt et al. (ed.) Interactions at the soil colloid-soil solution interface. Kluwer, Dordrecht, The Netherlands. Schneider, K., M.-B. Turrion, P.G. Grierson, and J.F. Gallardo. 2001. Phosphatase activity, microbial phosphorus, and fine root growth in forest soils in the Sierra de Gata, western central Spain. Biol. Fertil. Soils 34:151. Schoenau, J.J., and W.Z. Huang. 1991. Anion-exchange membrane, water, and sodium bicarbonate extractions as soil tests for phosphorus. Commun. Soil Sci. Plant. Anal. 22:465. Sharpley, A.N. 1985. Phosphorus cycling in unfertilized and fertilized agricultural soils. Soil Sci. Soc. Am. J. 49:905. Sharpley, A.N., L.R. Ahuja, M. Yamamoto, and R.G. Menzel. 1981. The kinetics of phosphorus desorption from soil. Soil Sci. Soc. Am. J. 45:493. Sibbesen, E. 1977. A simple ion exchange resin procedure for extracting plant-available elements from soil. Plant Soil 46:665. Sibbesen, E. 1978. An investigation of the anion-exchange resin method for soil phosphate extraction. Plant Soil 50:305. Simard, R.R., D. Cluis, G. Gangbazo, and A.R. Pesant. 1994. Phosphorus sorption and desorption indices in soil. Commun. Soil Sci. Plant Anal. 25:1483. Smethurst, P.J., and N.B. Comerford. 1993. Potassium and phosphorus uptake by competing pine and grass: Observation and model verification. Soil Sci. Soc. Am. J. 57:1602. Smyth, T.J., and P.A. Sanchez. 1980. Effects of lime, silicate, and phosphorus applications to an Oxisol on phosphorus sorption and ion retention. Soil Sci. Soc. Am. J. 44:500. Soil Survey Staff. 1999. Soil taxonomy. US Government Printing Office. Washington, DC. Sposito, G. 1984. The surface chemistry of soils. Oxford Univ. Press, Oxford. StatSoft. 1999. STATISTICA 5.5 Electric manual. StatSoft, Inc., Tulsa, OK.

PAGE 122

109 Steffens, D. 1994. Phosphorus release kinetics and extractable phosphorus after long-term fertilization. Soil Sci. Soc. Am. J. 58:1702. Stevenson, F.J. 1986. Cycles of soil; Carbon, nitrogen, phosphorus, sulfur, micronutrients. John Wiley and Sons, New York, NY. Strauss, R., G.W. Brummer, and N.J. Barrow. 1997. Effects of crystallinity of goethite. II. Rates of sorption and desorption of phosphate. Eur. J. Soil Sci. 48:101. Tate, K.R. 1984. The biological transformation of phosphorus in soil. Plant Soil 76:245256. Tejedor-Tejedor, M.I., and M.C. Anderson. 1990. Protonation of phosphate on the surface of goethite as studied by CIR-FTIR and electrophoretic mobility. Langmuir 6:602. Torrent, J., U. Schwertmann, and V. Barrn. 1992. Fast and slow phosphate sorption by goethite-rich natural materials. Clays Clay Miner. 40:14. Udo, E.J., and F.O. Uzu. 1972. Characteristics of phosphorus adsorption by some Nigerian soils. Soil Sci. Soc. Am. Proc. 36:879. Van den Berghe, C.H., and N.V. Hue. 1999. Liming potential of composts applied to an acid Oxisol in Burundi. Compost Sci. Util. 7:4046. Van der Zee, S.E.A.T.M., and W.H. Van Riemsdijk. 1991. Model for the reaction kinetics of phosphate with oxides and soil. p. 205. In G.H. Bolt et al. (ed.) Interactions at the soil colloid-soil solution interface. Kluwer, Dordrecht, The Netherlands. Van Rees, K.C.J., N.B. Comerford, and W.W. McFee. 1990a. Modeling potassium uptake by slash pine seedlings from low-potassium-supplying soils. Soil Sci. Soc. Am. J. 54:1413. Van Rees, K.C.J., N.B. Comerford, and P.S.C. Rao. 1990b. Defining soil buffer power: Implications for ion diffusion and nutrient uptake modeling. Soil Sci. Soc. Am. J. 54:1505. Violante, A., C. Colombo, and A. Buondonno. 1991. Competitive adsorption of phosphate and oxalate by aluminum oxides. Soil Sci. Soc. Am. J. 55:65. Violante, A., and L. Gianfreda. 1993. Competition in adsorption between phosphate and oxalate on an aluminum hydroxide montmorillonite complex. Soil Sci. Soc. Am. J. 57:1235. Voundi, N.J.C., A. Demeyer, G. Baert, M.G. Verloo, and E. Van Ranst. 1997. Chemical fertility aspects influenced by the mineralogical composition of some acid tropical soils of the forest zone in Central Cameroon. Agrochimica 41:209.

PAGE 123

110 Wang, H.D., W.G. Harris, and T.L. Yuan. 1991. Relation between phosphorus and iron in Florida phosphatic soils. Soil Sci. Soc. Am. J. 55:554. White, R.E. 1981. Pathways of phosphorus in soil. p. 21. In T.W.G. Hucker (ed.) Proc. of a symposium on phosphorus in sewage sludge and animal waste slurries. Reidel, Dordrecht, The Netherlands. Wolf, D.C., T.H. Dao, H.D. Scott, and T.L. Lavy. 1989. Influence of sterilization methods on selected soil microbiological, physical, and chemical properties. Environ. Qual. 18:39. Wolf, M.B., and D. London. 1994. Apatite dissolution into peraluminous haplogranitic melts An experimental study of solubilities and mechanisms. Geochim. Cosmochim. Acta 58:4127. Yang, J.E., and E.O. Skogley. 1992. Diffusion kinetics of multinutrient accumulation by mixed-bed ion-exchange resin. Soil Sci. Soc. Am. J. 56:408. Ziadi, N., R.R. Simard, T.S. Tran, and G. Allard. 2001. Soil-available phosphorus as evaluated by desorption techniques and chemical extractions. Can. J. Soil Sci. 81:167. Zinder, B., G. Furrer, and W. Stumm. 1986. The coordination chemistry of weathering. II. Dissolution of Fe(III) oxides. Geochim. Cosmochim. Acta 50:1861.

PAGE 124

BIOGRAPHICAL SKETCH Shinjiro Sato was born to Nobuko and Kiyoshi Sato in Osaka-shi, Japan on February 1, 1970. He grew up in his hometown and graduated from Kansai Soka High School in 1988. He then attended Soka University and received a B.A. degree in English Literature in 1993. During his undergraduate work, Sato spent one year (1990) as an exchange student of Portuguese language at the Federal University of Paran in Curitiba, Brazil. He then attended the University of Tsukuba, Ibaraki-ken, Japan, where he earned an M.S. degree in Environmental Science in 1995. He then moved to Manaus, Brazil to work for NGOs as a project consultant (JapanBrasil Network) and research assistant (Centro de Pesquisas Ecolgicas da Amaznia) engaged in community-basis projects for 3 years. In the summer of 1998, Sato joined the University of Florida to pursue a Doctor of Philosophy in agroforestry in the School of Forest Resources and Conservation. As his research progressed, he felt that a more in-depth understanding of soil phosphorus chemistry was required; so he switched to the Soil and Water Science Department at the University of Florida in the spring of 2000. After receiving his doctorate in Soil Science, he plans continuous international research as a soil scientist including work in the Brazilian Amazon. 111


Permanent Link: http://ufdc.ufl.edu/UFE0001359/00001

Material Information

Title: Phosphorus Sorption and Desorption in a Brazilian Ultisol: Effects of pH and Organic Anions on Phosphorus Bioavailability
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0001359:00001

Permanent Link: http://ufdc.ufl.edu/UFE0001359/00001

Material Information

Title: Phosphorus Sorption and Desorption in a Brazilian Ultisol: Effects of pH and Organic Anions on Phosphorus Bioavailability
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0001359:00001


This item has the following downloads:


Full Text












PHOSPHORUS SORPTION AND DESORPTION IN A BRAZILIAN ULTISOL:
EFFECTS OF pH AND ORGANIC ANIONS ON PHOSPHORUS BIOAVAILABILITY

















By

SHINJIRO SATO


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2003

































Copyright 2003

by

Shinjiro Sato



































This Ph.D. dissertation is dedicated to my life mentor,
Dr. Daisaku Ikeda;
and to
all men, women, and children in the Amazon.















ACKNOWLEDGMENTS

It is my honor to express my sincere gratitude to my chairman, advisor, and friend,

Dr. Nicholas Comerford for giving me proper directions for this research, insightful

advice, and warm encouragement throughout my Ph.D. program. My special thanks go

to Dr. Brian McNeal because it was through his classes that I became fascinated in Soil

Science. I would like to thank Dr. Willie Harris for his openness and patience to even my

stupid questions, especially regarding soil mineralogy. I would like to acknowledge Dr.

Dean Rhue for his keen insights in soil P chemistry and for his kindness in letting me use

his calorimeter. Dr. P.K. Nair helped me develop a scientific way of thinking early in my

Ph.D. program, for which I am greatly thankful.

The financial support for this research was provided by a graduate assistantship in

the Soil and Water Science Department. I am extremely grateful to Dr. Randal Brown

for this great opportunity. The Researcher Support Scholarship Program of the Soka

Alumni Association in Soka University, Japan provided an additional support.

I am grateful for all the support that I had from Mary McLeod, Miranda Lucas, and

Adriana Chagas in the Forest Soils laboratory; and from Dr. Dean Rhue and Bill Reve

with the calorimetry study. I cannot adequately thank Dr. Paulo Gabriel Nacif for

collecting soil samples from Bahia, Brazil. Yongsung Joo very generously provided me

with statistical help. All of this help and support made my dissertation a reality.

I enjoyed all of my old and new friendships (especially those with my fellow

graduate students) both on and off campus during this program. I was cheered and









motivated by the spiritual encouragement that I received from all of my beloved SGI

family members in Brazil, Japan, and the United States of America.

Finally but most importantly, I would like to thank my parents and brother for their

infinite love, understanding, and support in every form for so long. Without them, none

of this work was possible.
















TABLE OF CONTENTS
page

A C K N O W L E D G E M EN T S ..................................................................... .................... iv

LIST OF TABLES .......... .................. ............ ........ .... .... .......... .. ix

LIST OF FIGURES ............................... ... ...... ... ................. .x

ABSTRACT ........ .............. ............. .. ...... .......... .......... xii

CHAPTER

1 GENERAL INTRODUCTION ...................................................... .....................

F orm s of P in Soils ......................................... ............. ............... .
Soil O rganic P Solid Phase............................................................ .. ..............
Soil Organic P Solution Phase................................ ......................... ........ 3
Soil Inorganic P Solution Phase ........................................ ...... ............... 3
Soil Inorganic P Solid Phase ............................................ ......................... 4
Factors Affecting Inorganic P Sorption and Desorption ...........................................5
Phosphorus Desorption Defines P Bioavailability ....................................... .......... 8
Classification of Desorbable Inorganic P ................. ................. ........ .............. 9
Importance of Phosphorus Bioavailability in Tropical Soils............................... 11

2 ASSESSING A QUICK METHOD FOR DEVELOPING PHOSPHORUS
DESORPTION ISOTHERMS USING ANION EXCHANGE MEMBRANES .......16

Intro du action ...................................... ................................................ 16
M materials and M methods ....................................................................... .................. 19
S o il M material ..................................................................... .. 19
Phosphorus Desorption Kinetics Using AEMs ................................................20
Phosphorus Desorption as Affected by AEM Surface Area ............................21
Phosphorus Desorption Isotherms Using the Multiple AEM Method ...............21
Phosphorus Desorption Isotherms Using the Sequential AEM Method .............22
Statistical A naly sis ...................................... ................... .... .. ... 23
Equation for Kdvalues ......................................... .................................. 24
Results ............. ....... ........... ..... ... ...... .. ............... .. ..............24
Phosphorus Desorption Isotherms: Kinetics and Effect of AEM
Surface Area ..................................................................... .... 24
Phosphorus Desorption Isotherms: Multiple Method vs. Sequential
AEM M ethod .................................. ... .. ......... .......... .....24









Phosphorus Desorption Isotherms: Kd values................................ ...............25
D discussion ..................................... .......................... ..... ..... ........ 25

3 INFLUENCE OF SOIL pH ON INORGANIC PHOSPHORUS SORPTION
AND DESORPTION IN A HUMID BRAZILIAN ULTISOL..............................34

In tro d u ctio n ................... ....................................................................................... 3 4
Effect of Soil pH on P Sorption...................................... ......................... 34
Effect of Soil pH on P D esorption.................................... ....................... 35
M materials and M ethods ............................................. ................................. 38
Soil M material and Its Characterization..... .......... ....................................... 38
Phosphorus Sorption Isotherm s ..................... .................... ....... ........39
Phosphorus Desorption Isotherms Using the Multiple AEM Method ................39
Equation for P Sorption and Desorption Isotherms .............................................41
Statistical A analysis ...................................... ................... ..... .... 42
R e su lts ................. ................... .......................................................4 2
Phosphorus Sorption pH Envelope ....................................... ...............42
Phosphorus Desorption pH Envelope......................................................43
D iscu ssio n .................. ... .................................. .............. ................ 4 5
Soil pH E effect on P Sorption ................................................................... .....45
Soil pH Effect on P Desorption ................................ ......................... ....... 46

4 INFLUENCE OF LOW MOLECULAR WEIGHT ORGANIC ANIONS ON
PHOSPHORUS DESORPTION AND BIOAVAILABILITY IN A HUMID
B R A Z IL IA N U L T ISO L .................................................................. .....................58

Introduction ............... ... .. ...... ................ .............................58
M materials and M methods ....................................................................... ..................64
Soil M material ......................................... ..................................64
Standard P Fertilization and LMWOAs Extraction Methods............................64
Experim ents ..................... ............................. ................. 66
Experiment 1. Optimum Time for Extraction of Ligand-Desorbable P
(O bjectiv e 1) ...................................... ...... ..... ..... ........ ........... 66
Experiment 2. Optimum Anion Concentration for Extraction of
Ligand-Desorbable P (Objective 1) ......................................................66
Experiment 3. Short Kinetic Extraction of Native Soil with Oxalate
(Objective 2) ............................................ ... .... ............................ 66
Experiment 4. One-minute Extraction of Native and Fertilized Soils
with Oxalate and Citrate at Lower Anion Concentrations
(O objective 2) ............... ..... ............... ......... ................................... 66
Experiment 5. Seven-hour Extraction of Native Soil with Oxalate
and Citrate at Lower Anion Concentrations (Objective 2) ...................67
Experiment 6. Differentiation of Desorbable P into Disequilibria- and
Ligand-Desorbable P (Objective 3) ......................................................67
Experiment 7. Mass Balance Verification for Ligand-Desorbable P
(O bj active 3) ................................ .. .......... ... ... .............. 68
Experiment 8. Investigation of P Immobilization after Sorption ..............68









Statistical A analysis ........................................ .......... ......... .. ...... 69
R results and D discussion ....................... .................. ................... .. ......69
Ligand-D esorbable P ........................... .......................... ............... ...........69
Ligand-Exchangeable P vs. Ligand-Dissolvable P .......................................... 70
Disequilibria-Desorbable P vs. Ligand-Desorbable P......................................72
Phosphorus Immobilization/Fixation after Sorption ....................................... 73

5 SUMMARY, CONCLUSIONS, AND FUTURE RESEARCH ...............................85

S u m m a ry ................................. ......... ... ... ... ... ... ........ ............... 8 6
Suitability of Using Multiple AEM Strips to Develop Desorption
Isotherm s ...................................... ...... ............ ................ ............... 86
Comparison of Multiple AEM Method to Sequential AEM Method ..................86
Effect of Soil pH on Phosphorus Sorption ............... ............ .....................87
Effect of Soil pH on Phosphorus Desorption ............................................... 87
Method for Measuring Ligand-Desorbable Phosphorus .....................................87
Separating Ligand-Desorbable P into Ligand-Exchangeable P and
L igand-D issolvable P ................................................. ... ........... ........ .... 88
Differentiating Bioavailable Pools of Applied P into
Disequilibria-Desorbable P and Ligand-Desorbable P ............. .................89
C o n clu sio n s..................................................... ................ 8 9
F future R research ........................................................................90

APPENDIX

A ACID-BASE TITRATION CURVE FOR THE SOIL ........................ ...............92

B DETERMINATION OF SOLUBLE P IN THE PRESENCE OF ORGANIC
A N IO N S ........................................................... ................ 9 3

C SURFACE REACTIONS OF PHOSPHATE, OXALATE, AND CITRATE
MEASURED BY FLOW CALORIMETRY..................... ................94

LIST OF REFEREN CES ............................................................ .................... 98

BIOGRAPHICAL SKETCH ............................................................. ............... 111
















LIST OF TABLES


Table page

1-1 Types and mechanisms of inorganic P desorption............... .................... 15

2-1 The amount of P sorbed for each desorption method at 3 different initial
P -ad d itio n lev els ........... ................................................ ................ ........ .............. .. 2 9

2-2 ANOVA table for the GLM model evaluating the P desorption isotherms............31

2-3 Estimates for parameters of the GLM model used to test desorption isotherm
different ces ............................................................................ 32

3-1 Surface soil (0-30 cm) properties for the Atlantic Forest soil, Bahia, Brazil..........50

3-2 Langmuir parameters for P sorption isotherms at 3 different soil pH levels ...........52

3-3 Langmuir parameters for P desorption isotherms at 3 different soil pH levels
w ith 3 different initial P-addition levels...................... ......................... ............ 55

3-4 Total amount and ratio to P sorbed ofP desorbed at 3 different soil pH levels
with 3 different initial P-addition levels ............. ................................ ............. 56

4-1 Disequilibria- and ligand-desorbable pools of P when 3 levels of P are sorbed
to th e so il ............................................................................ 82

B-1 Effect of varying concentrations of oxalate and citrate on absorbance over a
range of standard P concentrations..................... ...... ........................... 93















LIST OF FIGURES


Figure p

1-1 Influence of pH on the distribution of orthophosphate species in solution.............14

2-1 Kinetics of P desorption using AEM s....... .. ................................ ............... 28

2-2 Phosphorus desorption with increasing number of AEM strips using a 2 h
sh ak in g tim e ....................................................... ................ 2 8

2-3 Phosphorus desorption isotherms using 2 desorption methods with 3 initial
P-addition levels. A) 180 tg P g B) 80 tg P g C) 40 tg P g ......................... 30

2-4 Relationship between observed C, values and C, values predicted from the
P desorption GLM m odel ........................................................ .............. 33

3-1 Phosphorus sorption isotherms at 3 different soil pH levels with equilibrium
concentration of A) 0-5 tg P mL1 and B) 0.0-1.0 tg P mL1 ............................51

3-2 Partition coefficients relative to equilibrium P concentration for P sorption
isotherm s at 3 different soil pH levels.................................. ........................ 53

3-3 Phosphorus desorption isotherms at 3 different soil pH levels with 3 different
initial P-addition levels. A) 180 Pg P g B) 80 Pg P g C) 40 g P g ...............54

3-4 Partition coefficients relative to equilibrium P concentration for P desorption
isotherms at 3 different soil pH levels with 3 different initial P-addition levels.
A) 180 g P g B) 80 g P g C) 40 g P g1 ...................................... ... 57

4-1 Ability of three organic anions to complex Fe as a function of pH as predicted
using Geochem -PC (Parker et al. 1995)........ ...................................... ............. 75

4-2 Phosphorus-desorption and Fe-release kinetics by organic anions when added
at a loading of 0.02 mmol g 1 soil. A) native soil. B) P-fertilized soil...................76

4-3 Effect of organic anion loadings for a 7 h interaction time on P desorption
and Fe release. A) native soil. B) P-fertilized soil............................................77

4-4 Effect of organic anion loading on the molar ratio of P desorption to Fe release
from native soil ................................................................... ..........78









4-5 Iron release by 3 oxalate loadings as a function of shaking time. The horizontal
line represents exchangeable Fe level (0.05 itmol Fe g 1 soil)..............................78

4-6 Effect of low anion loadings for a 1 min interaction time on P desorption
and Fe release. A) native soil. B) P-fertilized soil..............................................79

4-7 Effect of low anion loadings for a 7 h interaction time on P desorption and
Fe release ........................... ..... ............ .......... .......... 80

4-8 Cumulative P desorbed by AEMs from soil at 3 increasing levels of P. Each
desorption step represents one day..................................................81

4-9 Distribution of sorbed P between the disequilibria- and ligand-desorbable
pools. The first bar in each P-added level shows a combination of
disequilibria- and ligand-desorbable pools; and the second bar shows total
desorbable P 1 d after P sorption........................................ .......................... 83

4-10 Cumulative P immobilization after P sorption. +S and -S are with and without
shaking; and +H and -H are with and without mercuric chloride for microbial
suppression ............... ...... ............. ......... .......... ........ ......... 84

4-11 Cumulative P immobilization versus the square root of time. Immobilized P
is that amount of P immobilized after the effects of microbes and shaking
have been rem oved ................................ ... .......... ........ .... .. .......... .. 84

A-i Titration curve for the Atlantic Forest soil......... ...... ... ...... ................92

C-1 Heats of reaction for phosphate sorption onto the study-site soil ..........................95

C-2 Heats of reaction for oxalate sorption onto P-saturated soil preceding
phosphate sorption onto oxalate-treated soil............. ...............................................96

C-3 Heats of reaction for citrate sorption onto P-saturated soil preceding
phosphate sorption onto citrate-treated soil .................................. .................97















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

PHOSPHORUS SORPTION AND DESORPTION IN A BRAZILIAN ULTISOL:
EFFECTS OF pH AND ORGANIC ANIONS ON PHOSPHORUS BIOAVAILABILITY

By

SHINJIRO SATO

December, 2003

Chair: Nicholas B. Comerford
Major Department: Soil and Water Science

Phosphorus bioavailability is a major limitation to plant productivity in weathered

Brazilian soils. Soil pH change and organic anions such as oxalate and citrate produce

known rhizosphere effects that influence P sorption and desorption. However, few data

document their influence on P bioavailability. Quantitative measurement of bioavailable

P pools is not well-defined. Our study focused on the surface soil developed under native

Atlantic Forest in Bahia, Brazil. Objectives of our study were as follows:

* To investigate a method of developing P desorption isotherms using multiple strips of
anion exchange membranes (AEMs)

* To evaluate the effect of pH on P sorption and desorption isotherms

* To separate ligand-desorbable P into ligand-exchangeable and ligand-dissolvable P

* To estimate the relative contributions of disequilibria- and ligand-desorbable P to the
total desorbable P.

The Multiple AEM Method was suitable for developing well-defined isotherms and

superior to sequential extraction methods over a range of solution concentrations in our









study. Increasing pH from 4.7 to 6 and 7 decreased P sorption (up to 21% and 34%,

respectively) and increased P desorption for all P application rates. Using less than

1 .mol oxalate or citrate g 1 soil and using 1 min contact time allowed us to desorb P

solely by ligand exchange without surface dissolution. Nevertheless, dissolution played

the major role in measuring ligand-desorbable P. Although most of P sorbed was found

in the ligand-desorbable pool, P was preferentially found in the disequilibria-desorbable

pool after P fertilization. The percentage of the total sorbed P that was desorbable was a

function of time since fertilization; 13 to 22% on day 1 and 39 to 45% on day 14. This

ageing process was due to a shaking artifact (30%), microbial immobilization (57%), and

sorption of P onto the soil in a non-oxalate accessible form (13%). Liming and organic

anion loading improved P bioavailability in this Brazilian soil. Long-term ageing and

continuous exudation of organic anions also influence P bioavailability. Management

options that recognize this will improve bioavailability of native and applied P sources in

this important agricultural region of northeastern Brazil.














CHAPTER 1
GENERAL INTRODUCTION

The management of tropical soils is crucial to addressing the issues of food

security, soil degradation and environmental quality (Lal 2000). Low fertility; high soil

acidity; and the role that land-use systems, fertilizers and liming play in mitigating these

problems have been areas of interest for the past several decades. These problems have

been addressed through the study of soil physical and chemical properties, which are of

fundamental value when devising appropriate management approaches; and also through

the social constraints that determine what management approaches are possible.

Because the soil is highly weathered, many humid, acid, tropical soils are

phosphorus (P) deficient. Our study focused on the P chemistry of a particular tropical

soil in northeastern Brazil as it relates to P bioavailability. This chapter will summarize

some pertinent aspects of P bioavailability as an introduction to the objectives and

hypotheses of this dissertation.

Forms of P in Soils

Soil Organic P Solid Phase

Organic soil phosphorus represents, on average, one-half of the total soil P and

varies between 15 and 80% in most soils (Stevenson 1986). The amount of soil organic P

depends on soil organic matter content; therefore, it tends to be high in the surface

horizon and decreases with soil depth. Soil organic P compounds are synthesized by soil

microorganisms as esters of orthophosphoric acid (Anderson 1980), with the most stable









and abundant compounds being inositol phosphates which make up 10 to 50% of the total

soil organic P.

Phosphorus held in organic form must first be converted to inorganic form before

plant uptake. The process is controlled by P mineralization and immobilization in a

manner similar to processes described for release of organic nitrogen and sulfur from soil

organic matter. Mineralization of organic P may be both biological and biochemical

(Tate 1984). Biological mineralization releases P bound in organic matter as soil

microorganisms mineralize soil organic matter.

Biochemical mineralization occurs when plant roots and microorganisms produce

phosphatase enzymes that can catalyze the hydrolysis of ester bonds between P and C

compounds. Phosphatases play a major role in mineralization of soil organic P.

Phosphatase activity is generally enhanced in P-stressed soils (Clarholm 1993) and

evidently have both spatial and temporal variability (Schneider et al. 2001). Since soil

microbial activity is dependent on soil organic carbon content, the C:P ratio is critical in

balancing P mineralization and immobilization processes. It has been proposed that

critical C:P values for decomposing residues are < 200:1 for net mineralization and

> 300:1 for net immobilization (Enwezor 1967, Sharpley 1985). The C:P ratio of the

substrate being decomposed may be more important than that of soil organic matter or

plant residues, with the critical C:P ratio calculated from microbial growth yield falling in

the range of 50 to 70 (White 1981).

Besides P fertilization status, mineralization of organic P is influenced by many of

the same soil properties that control decomposition of soil organic matter: namely,

temperature, moisture, aeration, pH, and cultivation method and/or intensity. Therefore,









in soils of tropical regions where climatic conditions favor organic matter decomposition,

P mineralization can supply a portion of the crop P requirement (Anderson 1980).

Inorganic P also can be immobilized in organic form as part of the microbial

biomass. The quantity of P immobilized varies from 6 to 106 {g P g 1 soil (Brookes et

al. 1984, Joergensen et al. 1995), and from 38 to 46% of sorbed inorganic P for some

variable charge minerals (He and Zhu 1998). The amount of P immobilized by microbes

was estimated in one case to be 3 to 5 times greater than the amount of P absorbed by

plants in semiarid grasslands (Cole et al. 1978).

Soil Organic P Solution Phase

Uptake of soluble organic P by plant roots is not well-documented. However, it

has been shown that the presence of (endo- and ecto-) mycorrhizal fungi stimulates net

mineralization of organic P because of phosphatase production (Condron et al. 1996,

Joner et al. 2000). Red clover (Trifolium pratense L.) uptake of P increased as the plant

and an associated arbuscular mycorrhizal fungus grew. Organic P sources (phytase,

lecithin, and RNA) contributed (17 to 31%) to total plant P compatibly with an inorganic

P source (KH2PO4, 23%) (Feng et al. 2003). Jayachandran et al. (1992) showed a similar

effect of arbuscular mycorrhizal fungi associated with big bluestem (Andropogon

gerardii Vit.), that promoted a significant increase in plant use of organic P (cytidine

3- and 5-phosphate) when added to a soil having low P fertility.

Soil Inorganic P Solution Phase

Phosphorus concentrations in the soil solution are low compared to other soil

macronutrients, normally ranging from 0.001 to about 1 [g P mL 1 with an average of

about 0.05 pg P mL 1 (Barber et al. 1962). Plant roots and mycorrhizal hyphae absorb

soil solution P as orthophosphate ions in either the H2PO4 or HPO42 form, depending on









soil pH. The dominant solution inorganic P species below pH 7.2 is the monovalent ion;

while HPO42 dominates at pH values between 7.2 and 12.1 (Fig. 1-1). Plant uptake of

the divalent ion has been shown to be slower than the monovalent ion (Cresser et al.

1993).

Soil Inorganic P Solid Phase

If not absorbed by the plant or immobilized by microbes, inorganic P in the soil

solution is subject to various chemical reactions. These include precipitation as and

dissolution from inorganic P solids, and adsorption to and desorption from soil mineral

surfaces.

Precipitation is a slow reaction in which phosphate ions react with dissolved Al3+

Fe3+, and Mn3+ in acid soils and Ca2+ in neutral to calcareous soils to form insoluble

hydroxy phosphate precipitates (Sanchez and Uehara 1980). Variscite (A1PO4-2H20) and

strengite (FePO4-2H20) are the most common stable minerals found in acid soils; while

in alkaline soils, the most stable minerals are calcium phosphates such as octacalcium

phosphate [CasH(P04)6-5H20] and hydroxyapatite [Ca5(P04)30H]. The solubilities of

these Ca-P minerals decrease with increasing pH; whereas Al- and Fe-P minerals exhibit

increasing solubility with increasing pH (Lindsay and Moreno 1960).

In contrast to precipitation, adsorption is a relatively fast initial reaction involving

both anion and ligand exchange. Negatively charged phosphate ions in the soil solution

are attracted to positively charged soil surfaces, via anion exchange. However, a

significant quantity of positively charged surfaces will develop only at low soil pH in

highly weathered soils that display variable-charge characteristics. Adsorption by ligand

exchange results when an inner-sphere metal-O-P covalent bond is formed by

replacement of hydroxyls on the edges of layer silicate clays and the surfaces of insoluble









oxides of Al, Fe, and/or Mn. This process is referred to as specific adsorption or

chemisorption (Schindler and Sposito 1991).

Since it is difficult to distinguish adsorption and precipitation reactions from one

another, they are usually referred to collectively as P sorption or fixation (Sposito 1984).

Sorption of P seems to increase over time, apparently due to slow precipitation processes

that are superimposed on the more rapid chemisorption (Van der Zee and Van Riemsdijk

1991). The amount of P in the soil solution appears to determine which reaction

dominates. Chemisorption dominates at low solution P; while precipitation proceeds

when the concentrations of P and associated cations in the soil solution exceed the

solubility products (Kp) of the relevant P compounds. Therefore, for soils exhibiting low

solution P concentrations such as many tropical soils, chemisorption should be a

dominant mechanism.

Adsorbed or precipitated inorganic P undergoes desorption or dissolution reactions

when moving from the solid to the solution phase. Desorption is of particular importance

in bioavailability evaluations, because this process controls what the plant can eventually

access. Desorption is used here to denote either the precise opposite of adsorption or the

counterpart to sorption, which would include a combination of desorption and

dissolution. Both soil and plant characteristics affect the extent and rate of desorption

and dissolution for inorganic soil P. It is necessary, therefore, to understand these

relationships in order to evaluate and predict P bioavailability and the fate of inorganic P.

Factors Affecting Inorganic P Sorption and Desorption

The soil characteristics that influence P fixation include the amount and type of

clay-fraction minerals, soil pH, soil organic matter content, soil temperature, time of

reaction, exchangeable Al3+, soil redox condition (Sanchez and Uehara 1980), and root









exudates. These factors are interactive rather than additive, which makes it difficult to

predict inorganic P fixation under a wide range of soil conditions.

The higher the Al and Fe oxide contents of soil clay, and the less crystalline (more

amorphous) the soil minerals, the greater an acid soil's P fixation capacity. This is

largely attributed to the greater surface area which these conditions represent (Fernandes

and Warren 1994, Gilkes and Hughes 1994, Quintero et al. 1999). Higher clay contents

also result in greater P fixation (Sanchez and Uehara 1980, Harris et al. 1996, Voundi et

al. 1997).

Soil pH, because of its influence on surface charge and the ease with which surface

charge can be managed by liming, has been the focus of research on a variety of soils

(Awan 1964, Parfitt 1977, Bar-Yosef et al. 1988, Naidu et al. 1990, Pardo et al. 1992,

Rupa et al. 2001). In spite of this, results concerning P sorption or solution P

concentrations as influenced by liming have been conflicting (Haynes 1982, Anjos and

Rowell 1987). Given the P sorption patterns observed at different soil pH values, and the

number of explanations used to account for these phenomena, it is clear that the influence

of liming on P sorption is not well-understood.

In highly weathered and acid soils, existing data support the hypothesis that

exchangeable Al3+ is the main factor determining the pattern of P sorption with changing

pH (Lopes-Hernandez and Burnham 1974, Anjos and Rowell 1987, Chen and Barber

1990). Haynes (1984) said that one should expect liming to increase P sorption in soils

that are initially high in exchangeable Al3+, but to decrease P sorption in soils with low

exchangeable Al3+ content. When soils with low exchangeable Al3+ are limed, the

neutralization and precipitation of Al3+ ion and of hydroxy-Al species to form Al









hydroxide reduces the number of P-sorption sites. Where exchangeable Al3+ is initially

high, the formation of amorphous hydroxyl Al with highly active sorbing surfaces may

exceed any decrease in the sorption capacity of the original sorbing surfaces, resulting in

increasing P sorption as pH increases.

Phosphorus desorption studies are rare compared to the frequency of P sorption

studies. Some desorption studies determined bioavailable P by using agriculture-based

extraction indices, and correlating them with plant P uptake (Parfitt 1979, Steffens 1994,

Martin et al. 2002). Others described the kinetics ofP desorption by extracting P using

dilution methods or using anion-exchange resins. Such studies examined the suitability

of relationships such as first-order, Elovich, and parabolic diffusion equations to describe

kinetic P release from soil (Chien and Clayton 1980, Sharpley et al. 1981, Raven and

Hossner 1994).

Several studies focused on the effect of soil pH on P desorption (Hingston et al.

1974, Rupa et al. 2001). However, the results have been inconsistent; with some workers

finding increasing P desorption with increasing pH (Madrid and Posner 1979, Cabrera et

al. 1981, De Smet et al. 1998) and others with decreasing pH (Barrow 2002). He et al.

(1989) reported that P desorption decreased until pH was raised to about 4.8; and then

increased with further pH increases for most of the acid soils from China which contained

high Fe and Al oxide and/or kaolinite levels. In contrast, for three representative surface

soils of India, both the amount and rate of P desorption initially increased with pH

increase from 4.25 to 5.5; and then decreased at higher pH values of 6.75 and 8.0 (Rupa

et al. 2001).









No relationship was found between P removal by various extractants and the

affinity constant derived from the Langmuir equation for the original sorption of P (Kuo

et al. 1988). It was found that the amount ofP desorbed by all extractants was

significantly related to the fraction of P saturation on soil surfaces, with higher

P saturation of the surfaces yielding higher desorbed P values.

Kafkafi et al. (1967) found, for kaolinite, that a portion of applied P was

non-exchangeable during simple washing using an indifferent electrolyte solution. They

concluded that some of the applied P was incorporated into a fixed sink. In another

study, Cabrera et al. (1981) postulated that the inability to desorb all P sorbed onto

Fe-oxides using strongly alkaline solutions (e.g., 0.1 MNaOH) was due to the existence

of micropores in the materials, where some of the applied P could remain occluded.

Plant roots are known to exude a range of organic acids including oxalate, malate,

and citrate. Their positive effect on enhancing P desorption has been well-recognized

(Jones 1998, Ryan et al. 2001). The mechanisms ofP desorption by these compounds

include ligand exchange and ligand dissolution; yet the evidence for which are active and

in what proportion for a range of soils is not available. These processes and their relative

contributions to changes in inorganic P bioavailability which can occur for the

rhizosphere should vary with plant species, plant nutritional status, and ambient soil

conditions (Hinsinger 2001).

Phosphorus Desorption Defines P Bioavailability

Prior studies on P sorption have outlined trends and provided a foundation for

understanding such sorption. However, bioavailability of P is a function of P desorption;

not P sorption. Specific information for soils, where needed for agricultural

development, is lacking for P desorption and its trends with soil characteristics. Humid,









weathered, tropical soils of Brazil represent one area where an understanding of

P desorption dynamics is needed to better manage soils that are inherently P deficient.

Understanding and evaluating soil P bioavailability requires that one identify

* Form(s) ofP absorbed by plants

* Form(s) of P that can be released from the solid to the adjacent and bioavailable
solution phase

* Processes responsible for the release of P from the solid phase to the solution phase

* How soil/plant/mycorrhiza properties affect these processes

* Soil properties that affect movement of P in solution to the root/mycorrhiza

* Plant properties that are important in determining P absorption by the
root/mycorrhiza when P arrives at its surface (Comerford 1998).

The P desorption process is poorly understood and virtually no such information exists

for tropical soils. Phosphorus desorption must be studied separately from or in

conjunction with P sorption, because there is a significant hysteresis between the two

such that sorption data do not suitably relate to desorption behavior (Barrow 1983). To

further confound the situation, the hysteresis that is commonly reported may be partially

an artifact of the methodology used, since a recent study suggests that while sorption and

desorption isotherms are strongly hysteretic; desorption and resorption isotherms sorbingg

P after the desorption step) are not generally hysteretic (Barros et al. in press).

Classification of Desorbable Inorganic P

For the purpose of this dissertation, the term desorption will be used in the same

fashion as the counterpart term sorption. It does not distinguish desorption from

adsorption; or distinguish dissolution from precipitation. To clarify the types of

desorption that occur in soils, the following categorization is suggested. Two major

mechanisms describe inorganic P desorption: disequilibria desorption and ligand









desorption. These can be subcategorized into exchange and dissolution types of

desorption (Table 1-1). Disequilibria desorption is consistent with Le Chdtelier's

Principle; and requires that a disequilibrium exist between sorbed and solution P

concentrations. Disequilibria can be created during uptake of solution P by plant roots;

during leaching; during precipitation as secondary P minerals; during a change in

rhizosphere pH that changes surface charge or solubilities; or during immobilization by

soil microbes. As the solution P concentration is decreased, P is replenished through

desorption from P adsorbed onto the soil surface in exchange with other anions in the

solution (disequilibria exchange); or through dissolution of relatively soluble

P compounds (disequilibria dissolution) (Wolf and London 1994).

Disequilibria-desorbable P can be measured in a specific time frame. It can be

extracted by maintaining the solution P concentration at low levels through sequential

extractions using anion-exchange resins (Sibbesen 1978, Saggar et al. 1990);

anion-exchange membranes (Cooperband and Logan 1994); or iron oxide-impregnated

paper strips (Menon et al. 1996/1997); or by shaking with water (Simard et al. 1994); or

low-ionic-strength solutions (Rupa et al. 2001). This pool of desorbable P should be

bioavailable to plants or soil microorganisms. The nature and pattern of this desorption,

particularly when described as a desorption isotherm, is rare in the literature and

non-existent for tropical soils.

Another desorbable pool of inorganic P can be released by ligands that have a

higher affinity for the soil surface than does phosphate. Ligands in solution form

inner-sphere complexes with the soil surface that exchange with P on the surface, because

of their higher affinity for the soil surface (ligand exchange) (Cambier and Sposito 1991,









Hu et al. 2001). These ligands affect the stability of precipitates and contribute to

dissolution of the solid, concomitantly releasing P from the surface (ligand dissolution)

(Zinder et al. 1986, Ludwig et al. 1996). Plant roots and soil microbes produce a range of

organic acids including acetate, oxalate, citrate, and malate. They are known to exude

these acids into the soil solution at concentrations which can promote P desorption (Fox

and Comerford 1990, Jones 1998).

The problem in defining the pool ofligand-desorbable P is that one does not know

with certainty which ligands are in the soil solution; their loading rate onto the solid

phase; the volume of soil affected by this loading; and how soil properties affect the

effectiveness of P release by these ligands (Comerford 1998). Therefore, the best

approach to quantifying this pool is to measure the total amount of P that is desorbable by

ligand desorption.

Importance of Phosphorus Bioavailability in Tropical Soils

Tropical soils are notorious for their low nutrient bioavailability (often related to

low exchangeable base contents, high soil acidity, Al toxicity, and low P desorbability).

Oxisols and Ultisols of humid, tropical Brazil are no exception. Low P bioavailability is

the primary limiting factor for plant growth in many of these soils.

The origin of P deficiencies in these soils is threefold (Brady and Weil 1999).

First, the total P level of the soils is low, ranging from 0.01 to 0.1% (Chen and Ma 2001).

Second, the P compounds commonly found in these soils are highly insoluble. Third,

when soluble sources of P, including those in fertilizers and manures, are added to soils

they are sorbed to the soil, changed into unavailable forms and, over time, incorporated

into highly insoluble compounds.









The high P fixing capacities of many tropical soils (Udo and Uzu 1972, Sanyal et

al. 1993, Fontes and Weed 1996) are a function of their mineralogy and are a primary

contributor to their P deficiencies. Therefore, it is critical to understand the physical and

chemical properties of these soils that control the dynamics of inorganic P tie-up and

supply. Better understanding should lead to better nutrient management and better

maintenance of soil quality. This is particularly true for the acid, weathered soils of

Brazil, where P chemistry is often poorly understood.

Based on the above discussion, it is obvious that P desorption is an important topic

affecting the nutrient management of P in soils, and particularly many tropical soils.

Phosphorus desorption determines P bioavailability; and its interaction with soil pH in the

form of desorption isotherms remains poorly documented. Bioavailable inorganic P

exists in two major pools, but quantitative estimates of these pools (and methods for

deriving these estimates) remain poorly developed. Therefore, our study will address

some of these gaps in knowledge, using a humid, highly weathered Ultisol from the

Atlantic Forest in northeastern Brazil. The following is a brief description of each

chapter.

Chapter 2 develops a method for constructing P desorption isotherms using anion

exchange membranes (AEMs) in a way that minimizes time spent and experimental

artifacts. This method is also compared to a more conventional method.

Chapter 3 investigates the effect of soil pH on both inorganic P sorption and

desorption isotherms. The hypothesis is that P sorption and desorption isotherms are

functions of soil pH, with increasing pH decreasing the amount of P that is sorbed and

increasing the amount of P that is desorbed. In both cases, more P would be bioavailable.






13


Chapter 4 addresses the major types of P desorption (disequilibria desorption and

ligand desorption); and attempts to differentiate between ligand-exchangeable desorption

and ligand-dissolvable desorption for a Brazilian soil in its native and P-fertilized states.

It also investigates the amount ofP sorbed that cannot be accessed by the above two

desorption mechanisms.





























2 4 6 8 10 12 14
Solution pH


Figure 1-1 Influence of pH on the distribution of orthophosphate species in solution









Table 1-1 Types and mechanisms of inorganic P desorption


Major type

Disequilibria


Reaction type


Exchange


Dissolution


Ligand


Exchange


Dissolution


Mechanism


Follows Le Chitelier's Principle. Requires P
disequilibria between solution and solid phases,
created by plant uptake, leaching, precipitation,
rhizosphere pH change, and/or microbial
immobilization.

Desorbs P from the soil surface in exchange with
other anions in the soil solution.

Desorbs P from relatively soluble P compounds.

Influenced by anions of low-molecular-weight
organic acids such as acetate, oxalate, citrate, and
malate, which are exuded by plant roots and soil
microbes. Ligands have higher affinity for the soil
surface than phosphate and/or destabilize P minerals.

Desorbs P from the soil surface by forming
inner-sphere complexes which exchange with P.

Desorbs P from the soil surface by destabilizing and
dissolving the solid itself.














CHAPTER 2
ASSESSING A QUICK METHOD FOR DEVELOPING PHOSPHORUS
DESORPTION ISOTHERMS USING ANION EXCHANGE MEMBRANES

Introduction

Inorganic phosphorus (P) bioavailability is defined by the combined processes of

desorption and dissolution. These processes occur in response to disequilibria established

in solution by removal of P from solution (disequilibria desorption); or in response to the

action of ligands exchanging with phosphate or dissolving phosphate-bearing compounds

(ligand desorption) (Table 1-1). Most commonly, bioavailable desorbable P is indexed

by extracting a portion of the labile pool of P with chemical extractants such as Bray 1,

Mehlich 1 or 3, or Olsen's Solution (Ziadi et al. 2001, Csatho et al. 2002). However,

these soil P tests are interpretable only when combined with crop experimental data from

the field or greenhouse (Kumar et al. 1992, Kleinman et al. 2001); and do not measure a

pool of bioavailable P, but instead act simply as an index to P bioavailability. There is no

absolute interpretation for the quantity of P extracted by any of these methods; nor do

these methods define the pattern of P release, or the influence ofP desorption on the soil

solution concentration. Therefore, such techniques have limited utility in the mechanistic

modeling of P bioavailability and uptake by plants.

Desorption can be described by a desorption isotherm, which is the relationship

between P on the solid phase and P in the solution phase. This relationship has been

referred to as a quantity (solid P)/intensity (solution P) relationship (Raven and Hossner

1993). The P desorption isotherm is a useful, descriptive characteristic of soil that









provides information on the ability of the soil to release P into solution as solution P is

removed. This ability is represented by the slope of the desorption isotherm which is

termed the partition coefficient (Kd). Partition coefficients are directly related to the

buffer power of the soil and are used to calculate soil diffusion coefficients (Van Rees et

al. 1990b). Diffusion is important because, in many soils, it is the mechanism governing

90 to 98% of the P supply to roots (Barber 1980). Therefore, Kd is a useful parameter in

computer-based nutrient uptake models. However, few comprehensive studies address

P desorption; and fewer address development of P desorption isotherms.

Phosphorus is desorbed from soil, and thus P desorption isotherms can be

developed by techniques that use dilution, sequential extraction, or anion exchange resin

extraction (Brewster et al. 1975, Bhatti and Comerford 2002). For the dilution method, P

is desorbed from the solid phase by shaking a soil sample over a range of soil:solution

ratios for a specified equilibration time with single or successive extractions (Madrid and

Posner 1979, Sharpley et al. 1981). In the sequential extraction, an extracting solution is

added to the soil at a constant soil:solution ratio; and the sample is shaken for a constant

equilibration time. This procedure is repeated until P desorption is exhausted or a pattern

of P release is established (Fox and Kamprath 1970, Okajima et al. 1983, Bhatti and

Comerford 2002). The anion exchange resin extraction is performed in the same manner

as the sequential extraction except that bags filled with anion exchange resins (AERs) are

used to enhance the removal of P from solution (Sibbesen 1977, Bache and Ireland 1980,

Yang and Skogley 1992, Delgado and Torrent 1997). In place of AERs,

resin-impregnated membranes such as anion exchange membranes (AEMs) have been

used successfully in P desorption studies (Saunders 1965, Saggar et al. 1990, Nuernberg









et al. 1998). Both AERs and AEMs behave as a P sink, desorbing P from the solid phase

by causing disequilibrium; and have proven to be useful indices of plant-available P

when used as extractants in soil tests (Sibbesen 1978, Schoenau and Huang 1991, Ziadi et

al. 2001). The AEM behavior in soil, including recyclability, P sorption kinetics, and

adsorbability of solution P by AEMs, are well-documented (Cooperband and Logan 1994,

Cooperband et al. 1999).

Anion exchange materials have been used to investigate the desorption kinetics of

P (Raven and Hossner 1994); to measure P mineralization (Parfitt et al. 1994); and to

measure total resin-desorbable P by sequential extraction (He et al. 1994). However,

P desorption isotherms have rarely been developed. In one noteworthy case the validity

of using AEMs for developing P desorption isotherms was confirmed for Brazilian

Cerrado soils with varying clay and organic matter contents (Barros Filho et al. 2001,

Barros Filho et al. in press).

Brewster et al. (1975) compared P desorption isotherms using AERs with those

developed by dilution. Brewster and colleagues concluded that the use of the resins was

a more desirable technique than dilution primarily because of the large dilution volumes

required.

The sequential extraction method has some inherent problems. Since this method

can require up to several weeks to develop a complete P desorption isotherm, continuous

shaking of samples may increase the soil's surface area because of soil particles' abrasion

(Barros Filho et al. in press). Another concern is that microbial P mineralization or

immobilization that may be induced during the prolonged desorption process can

influence P desorption patterns (Barros Filho et al. in press). The first objective of our









study was to evaluate the suitability of using multiple strips of AEMs (termed the

Multiple AEM Method) to develop P desorption isotherms. Initial work showed that

P desorption was sensitive to the amount of AEM surface area in contact with soil. This

objective addressed whether the Multiple AEM Method could be used to more rapidly

develop a desorption isotherm. The second objective was to compare the Multiple AEM

Method with a sequential-extraction approach using AEMs (termed the Sequential AEM

Method) to determine if the manner in which AEMs were used would influence the

desorption isotherm.

Materials and Methods

Soil Material

The soil material used in our study was the surface horizon (0 to 30 cm depth) of a

Kandiudult (Soil Survey Staff 1999), known as an Argissolo vermelho/vermelho-amarelo

distr6fico in the Brazilian system (Oliveira et al. 1992). It was taken from under a native

Atlantic Forest cover type (Morellato and Haddad 2000) located near Una Ecoparque

near the city of Una, Bahia, Brazil. Using a shovel, the material was collected from a

1 ha area. Approximately 20 sampling points were combined; and the soil was

subsequently air-dried, passed through a 2-mm sieve, and stored in plastic bags. Selected

chemical and physical properties of the soil material were determined by standard

methods (Page et al. 1982), which are listed in greater detail in Chapter 3. Soil pH was

4.7 (1:2 soil:water), with 2.5% organic carbon and 25% clay content dominated by

kaolinite. Mehlich 1-available P and total P were 2.7 {g g 1 and 98.3 pg g1,

respectively. The iron content was greater than the aluminum content in both the

amorphous and crystalline oxide forms.









Phosphorus Desorption Kinetics Using AEMs

To maximize effectiveness of AEMs, the kinetics of P desorption in the presence of

AEMs was first determined. Triplicate 2 g samples were weighed into 50 mL centrifuge

tubes; and then 20 mL of 50 mMKCl containing 25 lg P mL1 as KH2PO4 were added to

each tube. Triplicate blanks of 20 mL of 50 mMKCl without soil and without P were

included. The tubes were shaken horizontally for 20 h; and then centrifuged for 15 min

at 2500 rpm. The supernatant was filtered through Whatman No. 5 filter paper and

measured for soluble reactive P by the method of Murphy and Riley (1962), which was

used for all P determinations in our study. Soluble reactive P as measured by this method

is operationally defined as inorganic P. Sorbed P was calculated as the difference

between the initial solution concentration added and the final concentration in the filtrate.

After decanting the supernatant, 50 mMKCl was added once more to each tube

until it reached its original pre-shaking weight. The AEM strips (1.25 cm x 5.00 cm;

Type AR204-SZRA-412, Ionics Inc., Watertown, MA) were prepared for use by soaking

overnight in 1 MKC1 of sufficient volume to provide an anion concentration at least

5 times greater than the total anion exchange capacity of the strips. After rinsing with

double-deionized water, one strip was inserted into each centrifuge tube. The tubes then

were shaken; and triplicate samples were removed from the shaker at 0.5, 1, 2, 4, 10, 24,

and 48 h. Each AEM strip was carefully removed from the tube so as to not

concomitantly remove soil particles; was rinsed with double-deionized water; and was

placed in another centrifuge tube filled with 15 mL of 1 MKC1. This tube was then

shaken for 30 min; and the solution was filtered through Whatman No. 5 filter paper.

Phosphorus in this filtrate was determined.









Phosphorus Desorption as Affected by AEM Surface Area

Ten sets of triplicate centrifuge tubes were prepared; and the P sorption process

was performed in the same manner as described above. After decanting the supernatant,

50 mMKCl were added to the tube until it reached its original pre-shaking weight. After

rinsing with double-deionized water, 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 strips (1.25 cm x

5.00 cm) of AEMs that had been prepared as described above were inserted into each set

of triplicate tubes. The tubes then were shaken for 2 h. The strips were removed from

the tubes; were rinsed with double-deionized water; and were placed in another 10 sets of

triplicate tubes filled with 15 mL of 3 MKC1 for the 1-, 2-, and 3-strip tubes; 25 mL for

the 4-, 5-, and 6-strip tubes; 30 mL for the 7- and 8-strip tubes; and 40 mL for the 9- and

10-strip tubes. These tubes were shaken for 30 min; and the solution was filtered through

Whatman No. 5 filter paper. Phosphorus in the filtrate was determined.

Phosphorus Desorption Isotherms Using the Multiple AEM Method

Phosphorus desorption isotherms then were developed using multiple AEM strips

on soil samples with 3 levels of sorbed P. Phosphorus was sorbed onto the soil using

initial loading levels of 40, 80 or 180 pg P g 1 soil. These levels were chosen because

they bounded normal fertilizer additions; and were equivalent to 16, 32, and 72 kg P ha1,

assuming the fertilizer to be broadcast and incorporated into the top 3 cm of soil.

Triplicate 2 g soil samples were weighed into 6 sets of triplicate centrifuge tubes;

and the initial weight of each tube + soil was recorded. Twenty milliliters of 50 mMKCl

containing 4, 8 or 18 pg P mL1 as KH2PO4 were added to the tubes; and the total initial

weight of tube + soil + solution was recorded. The tubes then were shaken for 20 h; and

centrifuged for 15 min at 2500 rpm. The supernatant solution was filtered through

Whatman No. 5 filter paper; and P in the filtrate was determined. Phosphorus sorbed









onto the solid phase was calculated as the difference between the initial or previous

P concentration and the P concentration in the filtrate.

After decanting the supernatant, tube + soil weights were recorded; and each tube

was filled with 50 mMKCl to its initial tube + soil + solution weight. The AEM strips

prepared in KC1 solution then were rinsed with double-deionized water; and 1, 2, 3, 4, 5,

or 6 strips (1.88 cm x 7.50 cm) were inserted into each of 6 sets of triplicate tubes. The

tubes were shaken for 2 h. The strips then were removed from the tubes; were rinsed

with double-deionized water; and were placed in another 6 sets of triplicate tubes filled

with 15 mL of2MKCl for 1-, 2-, and 3-strip tubes; and 30 mL for 4-, 5-, and 6-strip

tubes. These KC1 tubes were shaken for 30 min; and then filtered through Whatman

No. 5 filter paper. Phosphorus in filtrate was measured. The original tubes with soil

samples were shaken for 20 h; and then centrifuged for 15 min at 2500 rpm. The

supernatant solution was filtered through Whatman No. 5 filter paper; and P in the filtrate

was determined.

The first step of this procedure was designed to reduce desorbable P on the solid

phase using AEMs. The equilibration step was required to develop a desorption

isotherm, because the soil must come into an equilibrium or quasi-equilibrium to measure

an equilibrium solution concentration. The entire procedure of the previous paragraph

was repeated two more times, resulting in a total of 3 desorption steps and yielding 18

points for each desorption isotherm in 4 d.

Phosphorus Desorption Isotherms Using the Sequential AEM Method

The Sequential AEM Method for P desorption isotherms was performed using a

constant number of AEM strips and multiple extraction steps on soil samples having the

same three P-addition levels as in the multiple AEMs study. The procedure was the same









as that using multiple AEM strips, except that only 4 strips were used in all extraction

steps. The extraction steps were continued until the concentration of P desorbed by

AEMs in the KC1 filtrate solution reached the P detection limit (0.005 ig P mL 1;

0.05 {g P g 1 soil). The Sequential AEM Method required 6, 9, and 13 desorption steps

and took 7, 10, and 14 d to develop desorption isotherms for samples to which 40, 80,

and 180 pg P g1 had been added, respectively.

Statistical Analysis

Statistical differences in the pattern of P desorption using different methods and

initial P loading levels were tested using the General Linear Models (GLM) procedure in

the Statistical Analysis System framework (SAS Institute 1988). Observed P desorption

was related to the amount of P initially added and the method by which P desorption

isotherms were developed. The model used was:

C, = af8 + C, + 2 C12 + af3(Method Padded) + E (2-1)

where C, is the amount of P sorbed on the solid phase per unit mass of soil in {g P g 1 of

soil; C, is the equilibrium concentration ofP in pg P mL 1; flo is the intercept parameter,

and fi, f/2, and /3 are the slope parameters of the regression line; and E is the random

error. The Method variable has two levels representing the Multiple and Sequential AEM

Methods. The Padded variable has three levels of 40, 80, 180 {g P g 1 initially added.

Analysis of variance (ANOVA) was used to test differences among regression line

parameters; and significance of the differences among means were determined by t-tests

at thep = 0.05 level. Means and standard deviations were used to summarize the data,

where applicable.









Equation for Kdvalues

The tendency of the solid phase to buffer C1 can be described numerically using Kd.

Since Kd is the rate of change of the ratio of sorbed P to solution P, it can be estimated

from the first derivative of equation (2-1) with respect to the solution concentration,

expressed as:

dC,
dC1 Kd = 1 + 2/82C1 (2-2)
dC,

Results

Phosphorus Desorption Isotherms: Kinetics and Effect of AEM Surface Area

The amount ofP desorbed by the AEM strips increased with shaking time up to

2 h, after which it decreased once more (Fig. 2-1). Therefore, 2 h was chosen as the

preferred shaking time. When the AEM surface area was increased, P desorption

increased with increasing number of AEM strips (Fig. 2-2), with the relationship being

linear between 1 and 6 strips (r2 = 0.997). When more than 6 strips were added, replicate

variability increased dramatically; and the linear relationship weakened.

Phosphorus Desorption Isotherms: Multiple Method vs. Sequential AEM Method

The amount of P initially sorbed at each P-addition level for the two extraction

methods is provided in Table 2-1. There were significant differences in the amount of

P sorbed between desorption methods for all 3 levels of P addition.

Both the Multiple and Sequential AEM Methods resulted in well-defined

desorption isotherms for all 3 levels of sorbed P (Fig. 2-3). The P desorption GLM

model well described the isotherms for both desorption methods for all three P-addition

levels. As shown in Figure 2-4, 99.8% of the variability in observed C, values were

accounted for by the model. Type III model ANOVA table and parameter estimates for









the P desorption GLM model (Tables 2-2 and 2-3) confirmed that P desorption isotherms

for each desorption method for each P-addition level were numerically different.

Statistical analysis also confirmed that the desorption methods and the P-addition

levels affected the P desorption isotherms except for those isotherms at the lowest

P-addition level (40 pg P g 1). There was no significant Method*Padded interaction on

Cs values at this P-addition level.

Phosphorus Desorption Isotherms: Kd values

Calculated Kd values derived from the P desorption GLM model were not a

function of the desorption method or the amount of P initially sorbed. These Kd values

were inversely and linearly related to equilibrium P concentration; and were expressed by

the equation:

Kd = 48.1 60.862 C1 (2-3)

The maximum Kd value was 48.1, which was the intercept of the above relationship.

Discussion

In evaluating phosphorus bioavailability, it is P desorption rather than P sorption

which determines the quantity of available P and subsequent ability of the soil to release

P. Phosphorus flux through the soil to plant roots depends on the initial solution P

concentration, the soil's buffer power, and the effective diffusion coefficient (Barber

1984). Buffer power and diffusion coefficients incorporate Kd values, which are

calculated from P desorption isotherms. Therefore, a well-defined P desorption isotherm

is critical, especially when evaluating P bioavailability using computer-based nutrient

uptake models (Nye and Tinker 1977, Van Rees et al. 1990a, Smethurst and Comerford

1993, Macariola-See et al. 2003), since the output of the models is dependent on Kd

values (Bhatti and Comerford 2002).









Our study showed that the Multiple AEM Method was suitable for developing

desorption isotherms over a range of P sorption levels, thereby addressing our study's

first objective. Regarding the second objective, it was shown that, for all P-addition

levels, both methods provided well-defined P desorption isotherms, although they were

numerically different. This difference was derived primarily from differences in the

amount of P initially sorbed (Table 2-1). The fact that desorption isotherms by the

different methods were not developed concurrently might also account for some of the

these differences.

However, since the ultimate goal of the desorption isotherms is to define Kd values

with respect to solution P concentrations, both desorption methods gave the same result.

Based on the GLM model used in our study, calculated Kd values were a function of

simply the solution P concentration.

The most useful method for developing isotherms should be as rapid as possible in

order to reduce the effects of shaking and microbial immobilization and mineralization

(Barros Filho et al. in press). In this respect both the Multiple and Sequential AEM

Methods are preferable to a dilution method, because the latter requires excessively high

dilutions. The Multiple AEM Method may also be superior to sequential desorption

methods, since it required about half the time to cover the same range of solution P

concentration while providing more data points within that range. The Sequential AEM

Method requires weeks to develop desorption isotherms; thus, this prolonged procedure

may change the physio-chemical and biological status of P in both the solid and solution

phases; and thereby influence the shape of the P desorption isotherm. In the case of

highly aggregated soils, extended shaking may induce disaggregation, creating more









surface area. Reduced P desorption may result if more P is sorbed to newly created

sorption sites.

Microbial P has been reported to reach its maximum level between 3 and 7 d after

addition of 30 to 100 kg P ha 1 as triple superphosphate to a Bh horizon from a Florida

Spodosol (Bliss 2003). Therefore, the less time required to develop a desorption

isotherm, the better the chance that the desorption method will not be influenced by

outside factors. It can be concluded that the Multiple AEM Method is superior to the

Sequential AEM Method because it requires less time to conclude.

Unfortunately, there is no standard methodology for developing desorption

isotherms. The Multiple AEM Method gave more desorption data points with fewer

extraction steps. This is encouraging enough to suggest that this method be tested on a

range of soil types with an objective of evaluating its suitability in developing P

desorption isotherms. These data from our study emphasize the need to standardized

methodology for the development of desorption isotherms.











2.5



S2.0


1.5

-o
S1.0

0.5
-o
0
- 0.5
c^


Shaking time (hour)

Figure 2-1 Kinetics ofP desorption using AEMs






6


S5


y = 0.5871x + 0.0695
r2 = 0.997


0 1 2 3 4 5 6 7
Number of AEM strips


8 9 10 11


Figure 2-2 Phosphorus desorption with increasing number of AEM strips using a 2 h
shaking time









Table 2-1 The amount of P sorbed for each desorption method at 3 different initial
P-addition levels

P initially added (tg g 1) Desorption Method P initially sorbed (tg g 1)
180 Multiple AEM 152.04bt
Sequential AEM 144.67a
80 Multiple AEM 74.91b
Sequential AEM 72.12a
40 Multiple AEM 37.89a
Sequential AEM 38.35b
t Different letters assigned in a column for each P-addition level denote significant
differences between desorption methods at the 0.05 probability level, by Tukey's
studentized range test











160




140 -


A
120




00




80




S60

Multiple 180

Sequential 180
40 -
4C Multiple 80

Sequential 80

20 Multiple 40

Sequential 40


0 -
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

Equilibrium P concentration (pg mL1)

Figure 2-3 Phosphorus desorption isotherms using 2 desorption methods with 3 initial
P-addition levels. A) 180 gg P g 1. B) 80 gg P g 1. C) 40 gg P g 1











Table 2-2 ANOVA table for the GLM model

Type III Sum
Source of Square df

Intercept 79692.83 1

Ci 206.45 1
C12 46.39 1

Method 5372.89 1

Padded 22973.08 2

Method*Padded 8640.22 2

Error 240.69 75


evaluating the P

Mean
Square

79692.83

206.45

46.39

5372.89

11486.54

4320.11

3.21


desorption isotherms


F value

24832.61

64.33

14.46

1674.21

3579.25

1346.16


Pr> F

< 0.0001

< 0.0001

< 0.0003

< 0.0001

< 0.0001

< 0.0001










Table 2-3 Estimates for parameters of the GLM model used to test desorption isotherm
differences
Level of Standard
Parameter parameter Estimate error t value Pr > I t
Intercept 128.00 1.00 127.99 < 0.0001
C1 48.05 5.99 8.02 < 0.0001
Cl2 -30.43 8.00 -3.80 < 0.0003
Method SEQ -50.46 0.75 -67.23 < 0.0001
Method MUL 0.00
Padded 40 -92.41 0.97 -94.85 < 0.0001
Padded 80 -59.10 0.83 -71.27 < 0.0001
Padded 180 0.00
Method*Padded SEQ 40 50.72 1.12 45.34 < 0.0001
Method*Padded SEQ 80 43.30 1.01 42.88 < 0.0001
Method*Padded SEQ 180 0.00
Method*Padded MUL 40 0.00
Method*Padded MUL 80 0.00
Method*Padded MUL 180 0.00
t SEQ and MUL represent Sequential AEM Method and Multiple AEM Method,
respectively
t Initial P-addition levels in ug P g 1










160

140 -

S120 y = 0.9979x + 0.1649

S100 r2 =0.998

80

60

40

20 -

0
0 20 40 60 80 100 120 140 160
Observed C, (gg g 1)
Figure 2-4 Relationship between observed C, values and C, values predicted from the
P desorption GLM model














CHAPTER 3
INFLUENCE OF SOIL pH ON INORGANIC PHOSPHORUS SORPTION AND
DESORPTION IN A HUMID BRAZILIAN ULTISOL

Introduction

Effect of Soil pH on P Sorption

In tropical regions, liming is frequently used to raise soil pH and increase

phosphorus bioavailability (Sanchez and Uehara 1980). However, published results

regarding the effect of liming on P sorption and bioavailability are conflicting (Haynes

1982). Phosphorus sorption has been shown to both decrease (Naidu et al. 1990, Holford

et al. 1994, Mora et al. 1999); and increase (Geelhoed et al. 1997, Pereira and De Faria

1998, Curtin and Syers 2001) with increasing pH. Still other reports have shown no

significant influence of pH (Jones and Fox 1978, Arias and Fernandez 2001).

Liming 3 acid soils from southern Brazil increased the amount ofP sorbed up to a

pH of 5.0, at which point P sorption began to decrease (Anjos and Rowell 1987). Raising

the pH of an Oxisol from the Cerrado region of Brazil with an initial pH of 4.5 also

reduced P sorption by 18 to 24% (Smyth and Sanchez 1980). The authors attributed the

reduced P sorption to increased hydroxyl concentration and increased competition

between hydroxyl and phosphate ions for specific adsorption sites on mineral surfaces.

Several authors thought that decreased P sorption could be due to the neutralization and

precipitation of Al3+ and hydroxy-Al as Al hydroxide during liming; decreasing the

number ofP sorption sites (Smyth and Sanchez 1980, Anjos and Rowell 1987, Naidu et

al. 1990). Bowden et al. (1980) and Haynes (1982) added that the mineral surface









became increasingly negative with increasing pH, which resulted in greater electrostatic

repulsion and decreased P sorption.

Phosphate is preferentially sorbed by soil mineral surfaces as HPO42 rather than as

H2PO4 ; and the concentration of the divalent ion (HP042 ) increases 10-fold for each

unit increase in pH from 2 to 7 (Bowden et al. 1980). This change partially offsets any

decrease in electrostatic potential (Haynes 1984). Thus, P sorption may decrease

relatively slowly until pH 7; and then, above pH 7, the increase of HPO42 concentration

becomes progressively slower; whereas the decrease in surface potential continues,

resulting in a more rapid decrease in P sorption (Haynes 1984).

In contrast to the studies above, Chen and Barber (1990) showed that for acid

weathered soils of pH 4.2 to 4.6 when adjusted up to a pH of 8.3, sorbed P increased up

to pH 6.0 to 6.2, and then decreased at higher pH values. The initial increase in

P sorption was explained as the formation of amorphous hydroxyl Al with highly active

sorbing surfaces. The subsequent decrease in P sorption was attributed to increased

competition of hydroxyl with phosphate for sorption sites. Similar conclusions were

drawn from liming studies involving Nigerian acid soils (Mokwunye 1975) and

Al-organic matter complexes (e.g., Al-peat and Al-humate) (Haynes and Swift 1989).

Effect of Soil pH on P Desorption

There are few investigations that focused on soil pH and P desorption; and the

results have been inconsistent; increased P desorption with increasing pH in some cases

(De Smet et al. 1998) and with decreasing pH in others (Barrow 2002). Studies of

P desorption kinetics for synthetic goethite as a function of suspension pH from 4 to 10

(Madrid and Posner 1979, Cabrera et al. 1981) showed increased P desorption with

increasing suspension pH. This trend could again be explained as above; as due to









competition from hydroxyl ions; and a lessened attraction or an enhanced repulsion

caused by increased negative charge on the surface with increasing pH.

Rupa et al. (2001) also desorbed P from 3 surface soils using 6 consecutive dilution

extractions at pH values varying from 4.2 to 8.0. The surface soils represented an Indian

Vertisol, Alfisol, and Oxisol. Phosphorus desorption increased for all soils up to a pH of

5.5, and then decreased at pH values of 6.7 and 8.0. The soils had high levels of

exchangeable Ca2+ plus Mg2+ (41.2, 8.7, and 4.3 cmolc kg 1 for Vertisol, Alfisol, and

Oxisol, respectively). It was thought that decreased desorption at high pH values could

be due to P precipitation as a Ca-phosphate.

Phosphorus desorption is a key process affecting inorganic P bioavailability; and

desorption also can identify specific pools of bioavailable P. However, in order to use

the desorption process in nutrient uptake models such as described by Smethurst and

Comerford (1993), a P desorption isotherm is required. An isotherm describes the

amount of P on the solid phase relative to the solution P concentration with the slope of

this relationship being called the partition coefficient, Kd. The Kd is defined as the rate of

change of the amount desorbed from the solid phase over the rate of change in solution

concentration. The Kd values are directly related to soil buffer power as seen in the

following equation:

b = +pKd (3-1)

where b is the soil buffer power, 0 is volumetric water content in cm3 of water per cm3 of

soil, and p is soil bulk density in units of g cm-3. Both Kd and b are critical parameters

for computer-based nutrient uptake models because of their role when calculating

diffusion as a nutrient-supply mechanism. Diffusion is important because for many soils,









diffusion is the mechanism supplying 90 to 98% of P absorbed by roots (Barber 1980).

Unfortunately, measuring P desorption isotherms is a time-consuming process; and there

is no standardized method for their development (see Chapter 2).

The overall objective of our study was to investigate the influence of pH on

P desorption from an Ultisol from the Atlantic Forest of Brazil. The Atlantic Forest is a

major Brazilian ecosystem type, with only about 7.6% of the native forest remaining at

present (Morellato and Haddad 2000). Much of the soil in this area has been degraded by

deforestation and subsequent extensive management for pasture and row crops.

Phosphorus management is a crucial component of rehabilitating these degraded soils.

This soil also represents an area considered one of the most biodiverse on the planet

(Oliveira and Fontes 2000).

Specific objectives of our study were to study this soil and its properties, as related

to the effect of soil pH on P sorption and desorption. The first hypothesis was that

P sorption is a function of pH, with decreasing sorption with increasing pH between 4.7

and 7.0. This soil is inherently low in exchangeable Al3+, so the mechanism of

amorphous Al-OH creating new surface for sorption was considered relatively

unimportant. Understanding the influence of liming on P sorption for these soils should

play a role in their management to enhance P bioavailability. A second hypothesis was

that P desorption is a function of pH, with increasing desorption as pH increases. The

degree to which liming influences P desorption is another important aspect of

P management that is virtually unknown; and should also be critical to proper nutrient

management.









Materials and Methods

Soil Material and Its Characterization

The soil material used in our study was the surface horizon (0 to 30 cm depth) of a

Kandiudult (Soil Survey Staff 1999), known as an Argissolo vermelho/vermelho-amarelo

distr6fico in the Brazilian system (Oliveira et al. 1992). It was taken from under a native

Atlantic Forest cover type located near Una Ecoparque near the city of Una, Bahia,

Brazil. Using a shovel, the material was collected from a 1 ha area. Approximately 20

sampling points were combined; and the soil was air-dried, passed through a 2-mm sieve,

and stored in plastic bags.

Soil characterization was performed as follows: pH (1:2 soil:deionized H20; 1:2

soil:1 MKC1; and 1:2 soil:0.01 M CaC12) using a glass-electrode pH meter, Orion Model

720A; organic carbon using a Walkley-Black method (Nelson and Sommers 1982); total

N by flash combustion using a Carlo Erba Model NA 1500 apparatus (Dumas 1831);

particle-size distribution using a pipette method (Gee and Bauder 1986); exchangeable

cations using 1 M ammonium acetate (pH 4) as the displacing solution, with Ca2+, Mg2+

K+, and Na+ measured by atomic absorption spectroscopy; 1 MKCl-exchangeable Al3+

and 1 M ammonium acetate-exchangeable Fe3+ measured by atomic absorption

spectroscopy; acid ammonium-oxalate- and sodium citrate-dithionite-extractable Al3+ and

Fe3+ determined by the method of McKeague and Day (1966); water-extractable P using

50 mMKCl; Mehlich 1-extractable P (Mehlich 1953); total P by digestion with H2SO4

and 30% H202 as described by Grierson et al. (1999); and clay mineralogy and its

quantification determined by X-ray diffraction and thermogravimetric analysis,

respectively.









Phosphorus Sorption Isotherms

Two 100 g sub-samples were taken from the bulk soil sample and adjusted to

pH 5.9 and 7.0 from the original pH of 4.7. Each sub-sample received the appropriate

volume of NaOH solution to increase soil pH, based on an acid-base titration curve for

the soil (appendix A). The sub-samples then were air-dried at room temperature.

Phosphorus sorption isotherms were produced for each of the 3 pH levels. Solution

pH was not controlled during the sorption process. Triplicate 2 g soil samples were

weighed into 50 mL centrifuge tubes; and 20 mL of 50 mMKCl containing 2, 4, 6, 8, 10,

12, 14, 16, 18, or 20 tg P mL1 as KH2PO4 was added to each tube. Triplicate blanks of

20 mL of 50 mMKCl without soil and without P were included as well. The tubes were

shaken horizontally for 20 h; and centrifuged for 15 min at 2500 rpm for the pH 4.7 and

5.9 samples, and for 15 min at 10,000 rpm for the pH 7.0 sample in order to clear the

supernatant solution. The supernatant was filtered through Whatman No. 5 filter paper;

and measured for soluble reactive P by the method of Murphy and Riley (1962), which

was used for all P determinations in our study. Soluble reactive P as measured by this

method is operationally defined as inorganic P. Sorbed P was calculated based on the

difference between the initial solution concentration added and the final concentration in

the filtrate.

Phosphorus Desorption Isotherms Using the Multiple AEM Method

For the Multiple AEM Method, pH of the soil was adjusted as described for the

P sorption-isotherms study. The pH adjustment was only at the beginning of the

experiments and was not controlled throughout the study. The pH control during

P desorption is not expected to have a significant effect on the P desorption process









(Bhatti and Comerford 2002); and pH is not controlled under field conditions following

liming.

Phosphorus desorption isotherms were initiated from 3 different sorption points at

each pH. Phosphorus was sorbed onto the soil using 40, 80, or 180 {g P g 1 soil. These

levels were chosen because they bounded normal fertilizer additions; and were equivalent

to 16, 32, and 72 kg P ha 1 (assuming fertilizer was broadcast and incorporated into the

top 3 cm of soil).

Triplicate 2 g soil samples from each pH level were weighed into 6 sets of triplicate

centrifuge tubes; and the initial weight of tube + soil was recorded. Twenty milliliters of

50 mMKCl containing 4, 8 or 18 ig P mL1 as KH2PO4 were added to each tube; and the

total initial weight of tube + soil + solution was recorded. Tubes were shaken for 20 h;

and centrifuged for 15 min at 2500 rpm for the pH 4.7 and 6.0 samples, and for 15 min at

10,000 rpm for the pH 7.1 sample in order to clear the supernatant solution. The

supernatant solution was filtered through Whatman No. 5 filter paper; and P in the filtrate

was determined.

After decanting the supernatant, weights of tube + soil were recorded. Each tube

then was filled with 50 mMKCl to its initial tube + soil + solution weight. Anion

exchange membrane (AEM) strips (1.88 cm x 7.50 cm) prepared in KC1 solution were

rinsed with double-deionized water; and 1, 2, 3, 4, 5, or 6 strips were inserted into each of

6 sets of triplicate tubes. The tubes were shaken for 2 h; then the strips were removed

from the tubes; were rinsed with double-deionized water; and were placed in another 6

sets of triplicate tubes filled with 15 mL of 2 MKC1 for the 1-, 2-, and 3-strip tubes; and

with 30 mL for the 4-, 5-, and 6-strip tubes. These tubes were shaken for 30 min; and









then filtered through Whatman No. 5 filter paper. Phosphorus in filtrate was measured.

The original tubes with soil were shaken for 20 h; and then centrifuged for 15 min at

2500 rpm for the pH 4.7 and 6.0 samples, and for 15 min at 10,000 rpm for the pH 7.1

sample. The supernatant solution was filtered through Whatman No. 5 filter paper; and P

in the filtrate was determined.

The first step of this procedure was to reduce desorbable P on the solid phase using

AEMs. The equilibration step was required to develop a desorption isotherm because the

soil must come into an equilibrium or quasi-equilibrium to measure an equilibrium

solution concentration. The entire procedure of the previous paragraph was repeated two

more times, resulting in a total of 3 desorption steps and yielding 18 points on a

desorption isotherm for each pH in 4 d.

Equation for P Sorption and Desorption Isotherms

The Langmuir equation was fitted to all data for both the sorption and desorption

isotherms. The following form of the equation was used:

KbC,
C = (3-2)
1 +KC,1

where C, is the amount of P sorbed onto the solid phase per unit mass of soil in lg P g 1

of soil; C1 is the equilibrium concentration of P in lg P mL 1; K is the affinity index

expressed in mL lg 1 of P; and b is the sorption maximum with unit of lg P g 1 of soil.

Both K and b are empirical constants that can be calculated from a linear form of the

equation plotted as C Cs against C1:

C C 1
+ (3-3)
C, b Kb

where 1/b and 1/Kb are the slope and intercept of the linear regression, respectively.









Ability of the solid phase to buffer C1 can be described numerically using the

partition coefficient, Kd. This reflects the rate of change of the ratio of sorbed P to

solution P; and is the first derivative of equation (3-2) with respect to the solution

concentration:

dC, Kb
Kd (3-4)
dC K (1 + K C)2

Statistical Analysis

Statistical differences among different linear regression lines were tested using the

General Linear Models procedure in the Statistical Analysis System framework (SAS

Institute 1988). Analysis of variance (ANOVA) was used to test differences among

regression line parameters. Significance of the difference among means was determined

by t-tests at thep = 0.05 level. Means and standard deviations were used to summarize

the data from P sorption and desorption isotherms, where applicable.

Results

Selected physical and chemical properties of the soil are summarized in Table 3-1.

The soil was acidic with low cation exchange capacity; and with low water-soluble and

Mehlich 1-extractable P. Its 25% clay content was dominated by kaolinite. The iron

content was greater than the aluminum content in both the amorphous and crystalline

oxide forms.

Phosphorus Sorption pH Envelope

Phosphorus sorption was a function of soil initial pH, with sorption decreasing with

increasing pH when equilibrium solutions P concentrations were less than about

3 pg P mL 1 (Fig. 3-la). Phosphorus sorption decreased by up to 21% when pH was









increased from 4.7 to 5.9; and decreased by up to 34% when pH was increased from 4.7

to 7.0. Above 3 pg P mL1, there were no sorption differences between pH treatments.

The Langmuir equation fit P sorption data for all 3 pH levels (r2 = 0.970 to 0.993,

Table 3-2), with Langmuir parameters such as K and b values being affected by pH

levels. As pH increased, K values significantly decreased; while b values increased

significantly only at the highest pH level (Table 3-2).

Changes in Kd values relative to equilibrium P concentration are provided in Figure

3-2. This relationship proved to be a function of soil pH. The Kd values at the higher pH

levels were lower than that at the original pH, once the equilibrium concentration was

reduced to about 0.5 pg P mL The Kd values were as much as 22% and 34% lower at

pH 5.9 and 7.0, respectively, than at pH 4.7.

Phosphorus Desorption pH Envelope

Phosphorus desorption exhibited the expected hysteresis between sorption and

desorption at all 3 pH levels for each level of P loading (Fig. 3-3). Since P sorption was a

function of initial soil pH as mentioned above, P desorption isotherms began from

different sorption points which depended on soil pH and the amount of P initially added

(Table 3-4).

The Langmuir equation fit the majority of the desorption data well (r2 = 0.338 to

0.867) except for the low P-addition level; and soil pH significantly influenced the values

of the Langmuir equation's parameters (Table 3-3). For the medium and high levels of

added P, K values were significantly decreased only at pH 7.1. At the low level of added

P, the K value significantly increased at pH 6.0 but significantly decreased at pH 7.1,

compared to values at pH 4.7. The range of K values across pH levels for the medium









and low levels of P addition was approximately 6-fold and 40- to 127-fold higher than at

the high P-addition level.

The b values were significantly lower at pH 6.0 and 7.1 than at pH 4.7, but were

not significantly different from each other for the high level of P addition. For the

medium P-addition level, b values for both elevated pH levels were significantly higher

than at the original pH. For the low P-addition level, b value only at pH 6.0 was

significantly higher than at pH levels of 4.7 and 7.1.

Phosphorus desorption was a function of soil pH when expressed in terms of either

the amount ofP desorbed or the ratio ofP desorbed to P initially sorbed (Table 3-4).

Higher pH levels resulted in more P desorbed and a higher ratio of P desorbed to P

initially sorbed. This was true for both the medium and high levels of P addition.

However, for the low level of P added, both P desorbed and the corresponding ratio were

slightly decreased at pH 6.0 and significantly increased at pH 7.1, compared to values at

pH 4.7.

The calculated Kd values, when plotted against equilibrium P concentrations,

evidenced opposite trends to those documented in the sorption study (Fig. 3-4). For the

high level of P addition, Kd values at pH 6.0 was lower than at pH 4.7 by up to 13%;

while Kd values at pH 7.1 was up to 38% higher than at pH 4.7. For example, Kd values

at pH 4.7, 6.0, and 7.1 were 32.9, 28.7, and 43.7 at an equilibrium concentration of

0.2 ig P mL1, respectively.

For the medium level of P addition, Kd values at pH 6.0 and 7.1 were 13 to 14%

and 76 to 85% greater than at pH 4.7, respectively. For the low level of P addition, Kd









values were 52 to 53% lower at pH 6.0 and 120 to 126% greater at pH 7.1, respectively,

compared to values at pH 4.8.

Discussion

Soil pH Effect on P Sorption

Results from the P sorption study supported the hypothesis that, in these soils,

P sorption decreases with increasing pH. This is consistent with what has been shown for

other weathered, acid soils with low levels of exchangeable Al3+ (Naidu et al. 1990).

This result has been credited to an increased electrostatic repulsion due to the increased

negative surface charge (Bowden et al. 1980, Haynes 1982) that accompanies increasing

pH. A contributing factor is that higher hydroxyl concentrations, which can be 10 to

1000 times higher than at lower pH levels, compete effectively with phosphate ions for

specific sorption sites on mineral surfaces (Smyth and Sanchez 1980). This result may

also be assisted by a reduction in the number of P-sorption sites. At higher pH levels Al

hydroxide polymers can neutralize sites where more reactive Al surfaces once were

present (Sanchez and Uehara 1980). Soils dominated by kaolinite, goethite, and gibbsite,

like the soil used in our study, are particularly susceptible to each of the above-named

mechanisms.

Empirically, trends for the Langmuir K parameter and for Kd support these

explanations. The reduced Langmuir K values with increasing pH suggest a reduced P

affinity for the sorption surface. The Kd represents the rate of change in the amount of P

sorbed relative to the amount of P in solution upon continued introduction of P to the

solution phase. Therefore, decreasing Kd with increasing pH corresponds to decreasing

rates of change in the amount of P sorbed, leaving more of the applied P in solution.









At higher equilibrium concentrations (> 3 .ig P mL 1), there were no differences

between isotherms at any pH. However, at a commonly practiced P-fertilization rate of

50 kg P ha 1, which corresponds to 125 {g P g 1 of soil using the above assumptions,

maximum solution P concentration was only 1.4 ig P mL1.

Soil solution P concentration is commonly around 0.05 tg P mL 1 (Barber et al.

1962), and seldom higher than 0.3 tg P mL 1 in most agricultural soils (Fried and

Shapiro 1961). It is typically even lower in soils under native forests and forest

plantations (Ballard 1980). Therefore, P sorption isotherms for the same 3 pH values

were replotted focusing on a lower solution P-concentration range (0.0-1.0 tg P mL1,

Fig. 3-1b).

These data show that P sorption decreases with pH increase under common

P-fertilization practices. At a solution concentration of 0.2 tg P mL1, sorbed P was 43,

36, and 32 tg P g 1 of soil at pH 4.7, 5.9, and 7.0, respectively. These represent 16% and

26% reduction in sorption compared to a pH of 4.7 when the soil is limed to pH 5.9 and

7.0, respectively. These percentages represent 2.8 to 4.5 kg ha1 of fertilizer P remaining

in solution, ready for plant uptake.

Soil pH Effect on P Desorption

Phosphorus desorption is a primary control on inorganic P bioavailability; yet

virtually nothing is known about the nature of P desorption for acid, weathered tropical

soils. Our study represents the first investigation of desorption isotherms for soils in this

region of Brazil. The results endorse the hypothesis that P desorption increases with

increasing soil pH. Both the decreased Langmuir K and the increased amounts of P

desorbed at elevated pH levels support the positive relationship between pH and P

desorption. These results are consistent with the findings of He et al. (1994). He and









associates determined P desorption from a variety of minerals and ranked the minerals

according to their desorption capacity: goethite < kaolinite < amorphous Al-oxide <

montmorillonite. Affinity constants also decreased in the same order.

The affinity constant, K, is related to the amount ofP sorbed. Phosphate is held

with its highest affinity to the soil surface when the surface coverage of sorbed P is low.

The affinity then decreases with increasing surface coverage (Ryden et al. 1977).

The calculated Kd for desorption represents the rate of change in the amount of P

desorbed relative to the solution P concentration. Therefore, increasing Kd is interpreted

as increasing the rate of change in P desorption at a particular solution P concentration, as

was observed for all levels of P loading when soil pH was raised to 7.1. Similar results

can be found among the limited P desorption studies previously reported (Hingston et al.

1974, Madrid and Posner 1979, He et al. 1989, De Smet et al. 1998).

General consensus regarding the effect of soil pH on P desorption, drawn from the

findings of these authors, is that increased P desorption with increased pH is due both to

increased competition with hydroxyl ions and to the change in electrostatic potential of

the surface. Although hydroxyl can be a good competitor with phosphate for metal

coordination sites on the surface (McBride 1994), this theory is still untested because it

has not been convincingly shown that electrostatic repulsion affects the specifically

sorbed anions that are directly coordinated to discrete surface metal cations. Given the

lack of prior work on P desorption mechanisms, it is suggested that this is a fruitful

avenue for further studies.

Phosphorus desorption is potentially controlled by the type of P complexation with

the surface: monodentate, bidentate mononuclear, and bidentate binuclear. These









complexes can be either non-protonated or protonated, depending on suspension pH

(Tejedor-Tejedor and Anderson 1990). Bidentate complexes require more activation

energy to break the bond than do monodentate complexes, so can be more difficult for

P desorption to take place in environments where the bidentate complexes predominate

between phosphate and the soil surface.

In spite of precise investigations using infrared spectroscopy (Parfitt and Atkinson

1976) and X-ray photoelectron spectroscopy (Martin and Smart 1987), assigning peaks

among possible complexation types remains controversial (Persson et al. 1996, Arai and

Sparks 2001, Kreller et al. 2002). Clarification of the effects of pH and P surface

coverage on the formation of different surface-complex types may help to resolve these

issues.

Environmental and socio-economic restrictions such as low pH, high P fixation

capacity, and high costs of P fertilizer and liming materials make P management for

humid, acid tropical soils problematic. Management strategies that make the most

efficient use of P and lime are required. To this end, one objective should be to raise the

soil solution P concentration, since it is directly related to plant P uptake. A small

increment in solution P concentration increased through liming-induced P desorption can

significantly influence plant growth. For example, Jones and Benson (1975) showed that

P in sweet corn leaves and equilibrium solution P levels were correlated, with a critical

solution P concentration of 0.12 to 0.13 pg P mL The relationship between plant P and

equilibrium P concentrations was curvilinear over the range of P concentrations from

0.01 to 1.8 pg P mL 1 (Fox and Kamprath 1970).






49


Combining the results from both parts of our current study we conclude that liming

has a twofold positive effect. It reduces the amount ofP sorbed and increases the amount

ofP desorbed. Both these changes induce increased solution P concentration and

P bioavailability. Translating these results, in combination with the use of

non-conventional liming materials found in rural areas (Ohno and Erich 1990, Erich

1991, Chung and Wu 1997, Van den Berghe and Hue 1999, Nkana et al. 2002), should

help rural farmers in tropical areas make the most efficient use of limited agricultural

resources.









Table 3-1 Surface soil (0-30 cm) properties for the Atlantic Forest soil, Bahia, Brazil

Property Value
Physical
Soil texture, %
Sand 66
Silt 9
Clay 25
Clay mineralogy, %
Kaolinite 80
Goethite 7
Chemical
pH 1:2 water 4.7
1:2 KC1 3.6
1:2 CaC12 3.6
Organic carbon, % 2.51
Total N, % 0.20
Exchangeable Ca2 cmol kg 0.53
Exchangeable Mg2+, cmolc kg 1 0.47
Exchangeable K+, cmolc kg 1 0.12
Exchangeable Na+, cmolc kg 1 0.06
Water-extractable (50 mMKC1) P, pg g 1 0.02
Mehlich 1-available P, tg g 1 2.68
Total P, tg g 1 98.29
KCl-exchangeable Al, tg g 1 60.52
NH4 acetate-exchangeable Fe, [tg g 1 2.85
AAOt extractable Al, % 0.16
AAO extractable Fe, % 0.45
SCD extractable Al, % 0.31
SCD extractable Fe, % 1.30
t Acid ammonium-oxalate
t Sodium citrate-dithionite






51



160 A



120



80

pH 4.7

40 pH 5.9

A pH 7.0

,- 0
bo 0 1 2 3 4 5

S100 B
-o


75



50



25



0
0.0 0.2 0.4 0.6 0.8 1.0
Equilibrium P concentration (pg mL1)

Figure 3-1 Phosphorus sorption isotherms at 3 different soil pH levels with equilibrium
concentration of A) 0-5 pg P mL1 and B) 0.0-1.0 pg P mL 1






52


Table 3-2 Langmuir parameters for P sorption isotherms at 3 different soil pH levels

Parameters of Langmuir equation

Initial soil pH K b r2


mLt g 1

1.69bt

1.27ab


0.98a


-1
Mg g
169.04a

177.04ab

192.12b


0.970

0.971

0.993


SDifferent letters assigned in a column for each parameter denote significant
differences among pH levels at the 0.05 probability level, performed by Tukey's
studentized range test.











300


250 i-pH 4.7


Y 200 pH 5.9













Equilibrium P concentration (gg mL 1)


Figure 3-2 Partition coefficients relative to equilibrium P concentration for P sorption
isotherms at 3 different soil pH levels
100


50


0
0.0 0.5 1.0 1.5 2.0 2.5

Equilibrium P concentration (pg mL )


Figure 3-2 Partition coefficients relative to equilibrium P concentration for P sorption
isotherms at 3 different soil pH levels











*pH 4.7
SpH 6.0
A pH 7.1


0.4


0.08


0.02


UT mm T


rp


0.16


0.04


Equilibrium P concentration (jg mL )

Figure 3-3 Phosphorus desorption isotherms at 3 different soil pH levels with 3 different
initial P-addition levels. A) 180 gg P g B) 80 gg P g 1. C) 40 gg P g 1


0.00


40 C



38



36


0.24


0.00


0.06


L









Table 3-3 Langmuir parameters for P desorption isotherms at 3 different soil pH levels
with 3 different initial P-addition levels

Parameters of Langmuir isotherm


P added

1g g
pgg g


Initial pH


mL g 1
100.05bt
102.88b
62.99a
688.70b

586.92b
349.89a
6046.62b
12999.13c
2668.71a


-1
Mg g
145.28b
129.78a
128.34a
73.86c

72.16b
70.73a
37.19a
37.72b
37.40a


0.800
0.651
0780
0.583

0.867
0.860
0.430
0.338
0.524


T Different letters assigned in a column for each parameter at each P-addition level
denote significant differences among pH levels at the 0.05 probability level,
performed by Tukey's studentized range test.










Table 3-4 Total amount and ratio to P sorbed ofP desorbed at 3 different soil pH levels
with 3 different initial P-addition levels


P added Initial pH


1g g
180


P sorbed
M1
.g g
152.04

139.47
139.28
74.91
73.65

72.54
37.89
38.44
37.86


Total P desorbed
M1
.g g
18.94at

21.52a
25.61b
5.02a
5.59a

6.68b


1.77ab
1.68a
2.10b


Ratio ofP desorbed to
P initially sorbed
%

12.5a

15.4b
18.4c
6.7a
7.6a

9.2b


4.7ab

4.4a
5.5b


SDifferent letters assigned in a column at each P-addition level denote significant
differences among pH levels at the 0.05 probability level, performed by Tukey's
studentized range test.











i-pH 4.7

-pH 6.0

-pH 7.1


0.2 0.4 0.6


0.10


0.15


0.01 0.02 0.03
Equilibrium P concentration (ig mL 1)


0.04


Figure 3-4 Partition coefficients relative to equilibrium P concentration for P desorption
isotherms at 3 different soil pH levels with 3 different initial P-addition levels.
A) 180 g P g 1. B) 80 g P g 1. C) 40 g P g 1


150 A


150 B



100



50



0 -
0.00


0.00


0.05














CHAPTER 4
INFLUENCE OF LOW MOLECULAR WEIGHT ORGANIC ANIONS ON
PHOSPHORUS DESORPTION AND BIOAVAILABILITY IN A HUMID
BRAZILIAN ULTISOL

Introduction

Natural ecosystems and crops developing on humid, tropical soils often suffer from

phosphorus (P) deficiency. Low levels of native bioavailable inorganic and organic P are

common and P, when applied as fertilizer, can be sorbed by Al and Fe oxides. Sorbed P

can also form a range of minerals in combination with Ca, Al and Fe; whereby little

remains bioavailable (McLean 1976, Hsu 1982). However, some of the sorbed P is still

bioavailable through the processes of desorption from the soil surface and dissolution of

relatively soluble P mineral compounds. These adsorption/desorption and

precipitation/dissolution equilibria control the concentration of P in the soil solution, its

chemical mobility, and its bioavailability (Hinsinger 2001).

To alleviate P deficiency, soil management practices have been studied under field

and laboratory conditions. Particularly important among them are liming and the use of

organic and inorganic acids (e.g., from animal wastes) to enhance bioavailability of P in

both native and applied forms. The effect of liming on inorganic P availability remains

controversial (Haynes 1982) and was discussed in Chapter 3. The effect of organic acids

on inorganic P bioavailability is discussed below.

The term desorption is used here in two ways. First, it is used as above to represent

a process that is simply the reverse of adsorption. Second, it is used in the same fashion









as sorption, not distinguishing desorption from dissolution. This is considered necessary

until proper terminology consistent with that for sorption is accepted.

To clarify the types of desorption that occur in soils, the following categorization is

suggested. There are two major mechanisms that describe inorganic P desorption:

disequilibria desorption and ligand desorption. These can be subcategorized into

exchange and dissolution types of desorption (Table 1-1). Following Le Chatelier's

Principle, disequilibria desorption requires a disequilibria between solid and solution

phases. This disequilibria can be created by removal of solution P, mediated by various

processes. As the solution P concentration decreases, P is replenished through desorption

of P adsorbed onto the soil surface (disequilibria exchange) or dissolution of relatively

insoluble P compounds (disequilibria dissolution) (Wolf and London 1994).

Ligand desorption is due to reaction of sorbed P with anions of organic acids such

as oxalate, citrate, and malate, that are produced by plant roots and soil microorganisms.

These ligands have a higher affinity for the soil surface than phosphate; and form

inner-sphere complexes with the soil surface; therefore exchange with P, releasing P to

the solution (ligand exchange) (Cambier and Sposito 1991, Hu et al. 2001). Ligands also

reduce the stability of precipitates and promote dissolution of solids, releasing P

(ligand dissolution) (Zinder et al. 1986, Ludwig et al. 1996).

Chemical extractants such as Bray 1, Mehlich 1 and 3, or Olsen's solution have

been commonly used to extract a portion of the soil's bioavailable P; however, they have

not been shown to quantify any of the pools described above. These soil P tests are

useful only when combined with experimental data (Kumar et al. 1992, Kleinman et al.

2001). Phosphorus fractionation schemes, as represented by Hedley et al. (1982),









sequentially extract inorganic and organic P, that are presumably associated with

different soil constituents. Bioavailable P defined by these schemes is usually taken as

the sum of readily available (resin-extractable), labile, and microbial (NaHCO3 plus

CHCl3/NaHCO3 extractable) pools. These schemes provide an operational definition of

bioavailable P, which does little, however, to either explain pools on a mechanistic basis,

or indicate the manner by which the plant accesses bioavailable inorganic P. It also is not

clear that they actually measure discrete bioavailable pools.

Recently anion exchange resins (AERs) and anion exchange membranes (AEMs)

have successfully been used to extract a desorbable pool of inorganic P that has proven to

be one of the better indices of plant-available P (Sibbesen 1978, Schoenau and Huang

1991, Ziadi et al. 2001). Since AERs and AEMs behave as a P sink and desorb P from

the solid phase using disequilibria, P exhaustingly desorbed by AERs or AEMs should

constitute the pool of disequilibria-desorbable P, categorized above, at least for the

specific time frame of the extraction. This pool of desorbable P should be bioavailable to

any plant root or soil microorganism. A description of the nature and pattern of

disequilibria desorption, particularly when described by a desorption isotherm, is rare in

the literature and non-existent for tropical soils.

On the other hand, quantitative measurements of ligand-desorbable P are not

well-established for soils of any climatic region. The problem in defining this pool of

desorbable P is that one is not certain which ligands are present in the soil solution; their

loading rates onto the solid phase; the volume of soil affected by this loading; and how

soil properties control the effectiveness of P release by ligand desorption (Comerford

1998).









Plant roots and soil microorganisms are know to exude a range of low molecular

weight organic anions (LMWOAs) such as oxalate, malate, and citrate to the soil

solution. The composition of these anions is variable and dependent on plant species,

plant age, and physiochemical environment (Curl and Trueglove 1986). Exudation of

organic anions from plant roots is enhanced by three environmental stimuli: nutrient

deficiency (particularly that of P); exposure to toxic cations (particularly Al3+); and

anoxia (Ryan et al. 2001).

Excretion of organic anions into the rhizosphere occurs essentially continuously for

two reasons. First, the total organic anion concentration of roots ranges from 5 to

50 mM, depending on tissue type and nutrient status of the plant; while that in soil

solution is about 1000-fold less. This creates a strong gradient for slow passive diffusion

from roots to the soil solution. Second, a substantial electric gradient exists across the

plasma membrane because of the operation of ATP-driven proton pumps (H+-ATPases)

and cytosolic K diffusion potential. While the H expelled into the apoplast by these

H+-ATPases creates a charge gradient to facilitate the uptake of cations from the soil, it

also tends to draw anions (e.g., citrate3 and malate2 ) out of the cells and into the

external soil solution (Jones 1998).

Once exuded into the soil solution, LMWOAs have the capacity to complex metals

both in solution and on the soil surface. The degree of complexation depends on the

particular organic anion involved; the concentration and type of metal involved; and pH

of the soil solution. Generally speaking, organic anions with a higher number of carboxyl

groups (e.g., citrate3 ) chelate the metals more strongly than do those with fewer carboxyl

groups (e.g., oxalate2 and acetate ) (Ryan et al. 2001). The chemical equilibria









speciation program Geochem-PC (Parker et al. 1995) can predict the ability of organic

anions to complex metals as a function of soil solution pH. An example of the

complexation of malate, oxalate, and citrate with Fe3 is diagramed in Fig. 4-1 (Jones

1998).

Organic anions in the soil solution have a positive effect on P bioavailability

(Sagwal and Kumar 1995, Cajuste et al. 1996). Studies investigated interaction between

organic anions and inorganic P sorption/desorption, and particularly the competitive

interactions for surface sorption sites on synthetic goethite (Geelhoed et al. 1998);

Al-oxides (Violante et al. 1991); kaolinite and montmorillonite (Kafkafi et al. 1988, He et

al. 1992); and field soils (He et al. 1992). Organic anions are effective in decreasing

P sorption and increasing P desorption, with their effectiveness being most pronounced

between pH 3.0 to 7.0 (Violante and Gianfreda 1993, Liu et al. 1999); and when the

ligand is sorbed to the soil surface before P (Violante et al. 1991). Violante et al. (1991)

concluded that many mineral soil-surface sites were accessible to both P and organic

anions during the sorption process; while other surfaces were specific to one or the other.

Ligand exchange of P is dependent on the concentration; sorption kinetics;

degradation rate of specific anions involved in the reaction (Fox and Comerford 1992); as

well as the type of P bonding complexes formed with the surface metals. Sorption of

citrate onto the soil surface increases its resistance to microbial decay, positively

affecting the mobilization of adsorbed P (Geelhoed et al. 1999). He et al. (1992)

postulated that organic anions modify P-metal bonding complexes from bidentate to

monodentate form, also making P more exchangeable with organic anions.









Ligand dissolution is based on the ability of organic anions to chelate metal cations

both in solution and on the surface. For example, the stability constants of oxalate and

citrate with Fe3+ (1:1:1 metal:ligand:proton stoichiometric ratio and zero ionic strength at

25 C) are 7.74 and 11.50, respectively (Jones 1998). Lan et al. (1995) showed that

inorganic P release from spodic horizons in Florida and Georgia was enhanced by

organic anions when the Al stability constant with the anion was greater than about 4.4.

When organic anions complex metals in solution, they prevent the precipitation of P by

the metal and reduce the activity of P in the solution. By forming complexes with metals

on soil surfaces, organic anions promote destabilization of the surface and dissolution of

the metals, resulting in P release into the solution (Jones et al. 1996).

Given the discussion above, it is clear that there are no well-established techniques

for quantitatively measuring the desorbable pools of soil inorganic P, with the possible

exception of disequilibria-desorbable P. And yet, measuring these pools and

understanding mechanisms involved in the release from each pool are fundamental to

understanding and predicting P bioavailability to plants.

Our study addressed the above issues through 3 objectives. The first is to consider

a method with which to measure ligand-desorbable P. The second objective is to attempt

separation of the ligand-exchangeable and ligand-dissolvable pools. The hypothesis

underlying this objective is that, since ligand exchange and ligand dissolution are thought

to occur over different time scales, a kinetic approach might differentiate between them.

The final objective is to use exhaustive desorption method with AEMs differentiate

disequilibria-desorbable and ligand-desorbable pools, when fertilizer P is applied at









increasing rates. This tested the hypothesis that the majority of P added to a soil should

be, in the short term, dominantly in the disequilibria-desorbable pool.

Materials and Methods

Soil Material

The soil material used in this study was the surface horizon (0 to 30 cm depth) of a

Kandiudult (Soil Survey Staff 1999), known as an Argissolo vermelho/vermelho-amarelo

distr6fico in the Brazilian system (Oliveira et al. 1992). It was taken from under a native

Atlantic Forest cover type (Morellato and Haddad 2000) located near Una Ecoparque

near the city of Una, Bahia, Brazil. Using a shovel, the material was collected from a

1 ha area. Approximately 20 sampling points were combined; and the soil was

subsequently air-dried, passed through a 2-mm sieve, and stored in plastic bags. Selected

chemical and physical properties of the soil material were determined by standard

methods (Page et al. 1982), which are listed in greater detail in Chapter 3. Soil pH was

4.7 (1:2 soil:water), with 2.5% organic carbon and 25% clay content dominated by

kaolinite. Water-extractable P performed using 50 mMKCl was 0.65 nmol g 1 soil.

Mehlich 1-available P and total P were 87.18 nmol g 1 soil and 3.17 .imol g 1 soil,

respectively. The iron content was greater than the aluminum content in both the

amorphous and crystalline oxide forms. Ammonium acetate-exchangeable Fe was

0.05 pmol g 1 soil.

Standard P Fertilization and LMWOAs Extraction Methods

The following series of studies used standard procedures for adding P at different

rates and extracting P and/or Fe with LMWOAs of varying concentrations. To simplify

procedure description, each procedure is described below and then referred to by name

within the subsequent method descriptions.









The Sorption Procedure was accomplished by adding 2 g of soil sample to a 50 mL

centrifuge tube along with 20 mL of a 50 mMKCl containing P as KH2PO4 at varying

concentrations. The P concentrations will be described later for each individual study.

The centrifuge tubes were horizontally shaken for 20 h; centrifuged for 15 min at

2500 rpm; and the supernatant was filtered through Whatman No. 5 filter paper. Filtrate

P was determined by the method of Murphy and Riley (1962), which was used for all P

determinations in our studies. Sorbed P was calculated by difference between P initially

added and P left in the filtrate.

When LMWOAs were added to the soil, oxalate or citrate solutions as potassium

oxalate or potassium citrate of varying concentrations were adjusted to an initial pH of

4.0 with HC1. Twenty milliliters of a given anion solution and 2 g soil samples were used

for each procedure. Anion Method 1 was performed in a 30 mL syringe tube that was

shaken for varying time periods, as listed below for the specific studies. Each suspension

then was filtered immediately through 0.22 [m nylon membrane filters (MSI Inc.)

without centrifugation.

Anion Method 2 was performed in 50 mL centrifuge tubes that were shaken for

varying time periods on a horizontal shaker, and then centrifuged for 15 min at 2500 rpm.

The supernatant was filtered through Whatman No. 5 filter paper.

Filtrate P and Fe for all methods were determined by the method of Murphy and

Riley (1962) and by atomic absorption spectroscopy, respectively. When using the

Murphy and Riley (1962) method in the presence of oxalate and citrate, care must be

taken to maintain those anion concentrations below 2 mM, since concentrations above

that may inhibit molybdenum blue color development (He et al. 1998, Appendix B).









Experiments

Experiment 1. Optimum Time for Extraction of Ligand-Desorbable P
(Objective 1)

The Sorption Procedure was used by adding 0 or 5.81 [mol P g 1 soil to triplicate

samples. The Anion Method 2 was applied to these samples to extract P using oxalate

and citrate solutions with 0.02 mmol g 1 soil, at shaking times of 1, 2, 3, 4, 5, 10, 30, 60,

180, 360, or 600 min. The objective of this study was to identify the extraction time that

released the most P.

Experiment 2. Optimum Anion Concentration for Extraction of
Ligand-Desorbable P (Objective 1)

The Sorption Procedure was used as for Experiment 1, to triplicate samples. The

Anion Method 2 was used to extract P using oxalate and citrate solutions with 0.01, 0.05,

0.10, 0.25, 0.50, or 1.00 mmol g 1 soil, except that 0.01 mmol g 1 soil was not used for

the fertilized soils; and a shaking time of 7 h. The objective of this study was to define

the anion concentration that released the most P at the optimum shaking time.

Experiment 3. Short Kinetic Extraction of Native Soil with Oxalate
(Objective2)

Triplicate samples of native soil were used with the Anion Method 1, where the

oxalate solution was 0, 0.02, or 0.05 mmol g 1 soil; and the shaking times were 1, 2, 3, 4,

or 5 min. The objective of this study was to determine if using a low range of shaking

times as in Experiment 1 would help differentiate ligand-exchangeable from

ligand-dissolvable P.

Experiment 4. One-minute Extraction of Native and Fertilized Soils with
Oxalate and Citrate at Lower Anion Concentrations (Objective 2)

Soil samples at two P levels (0 and 5.81 .imol P g 1 soil) were prepared using the

Sorption Procedure. All samples were prepared in triplicate; and the Anion Method 1









was used with oxalate and citrate concentrations of 0.01, 0.10, 1.00, or 10.00 .imol g 1

soil and with shaking time of 1 min. The objective of this study was to determine if,

using a short shaking time, a range of low anion concentrations would help differentiate

ligand-exchangeable from ligand-dissolvable P.

Experiment 5. Seven-hour Extraction of Native Soil with Oxalate and Citrate
at Lower Anion Concentrations (Objective 2)

The Anion Method 2 was used to triplicate soil samples of native soil, using oxalate

and citrate solutions of 0.01, 0.10, 1.00, or 10.00 imol g 1 soil and using a 7 h shaking

time. The objective of this study was to determine if the results of Experiment 4 could be

reproduced using a longer shaking time, thus demonstrating the importance of anion

concentration versus shaking time in differentiating ligand exchange from ligand

dissolution.

Experiment 6. Differentiation of Desorbable P into Disequilibria- and
Ligand-Desorbable P (Objective 3)

Four soil levels of P were prepared in triplicate using the Sorption Procedure, at P

rates of 0, 1.29, 2.58, and 5.81 .imol P g1 soil.

The first P extraction for these samples was accomplished using sequential

extraction with AEMs, as described in Chapter 2. The desorption process was continued

until incremental P removed by AEMs was lower than 2-5 ng P mL 1 in the AEMs

extracting solution, corresponding to 0.65-1.61 nmol P g-1 soil. This is the detection

limit of the Murphy and Riley (1962) method when using a 5 cm path length. The pool

of disequilibria-desorbable P was defined as the difference between the amount of P

initially sorbed and the amount of P remaining on the solid phase after the last AEMs

extraction step. The objective of this step was to measure the total pool of

disequilibria-desorbable P at different P-fertilization levels.









The samples were then extracted with the Anion Method 2 using a solution of

0.50 mmol oxalate g 1 soil and a 7 hr shaking time. The objective of this step was to

measure what we defined above as the total ligand-desorbable pool of P, after

disequilibria-desorbable P was extracted.

Experiment 7. Mass Balance Verification for Ligand-Desorbable P
(Objective 3)

Triplicate soil samples of the four P levels of Experiment 6 were extracted using

the Anion Method 2 of Experiment 6. The objective of this experiment was to confirm

that total P extracted in Experiment 6 was equivalent to the amount extracted if only the

Anion Method 2 was used.

Experiment 8. Investigation of P Immobilization after Sorption

Soil samples, in triplicate, had P applied at a rate of 2.58 imol P g 1 soil, using the

Sorption Procedure. Enough samples were prepared to establish 3 treatments that could

be destructively sampled over time. The treatments were: 1) P extraction by shaking

without HgCl2 added (+S-H); 2) P extraction by shaking with HgCl2 added (+S+H); and

3) P extraction without shaking and with HgCl2 added (-S+H). Each treatment was

extracted for 1, 3, 7, 10, or 14 d; and mercuric chloride was added to the appropriate

samples (+H) at a rate of 1.84 .imol g 1 soil. The shaking treatment (+S) was continuous

shaking on a horizontal shaker; while the unshaken samples (-S) were kept in a test-tube

rack next to the shaker. At each sampling time, P was extracted from each treatment in

triplicate using the Anion Method 2 of Experiment 6.

It was shown in previous studies that part of the P sorbed by this soil was not

extractable by the Anion Method 2. The objective of this experiment was to determine

the role of microbial immobilization (+H vs. -H); shaking (+S vs. -S); and time of









interaction (short-term ageing of 1 to 14 d) on the non-extractability of P in previous

experiments. Mercuric chloride was used because it is an effective soil sterilizant; and it

reportedly causes minimal change to soil physio-chemical properties (Wolf et al. 1989).

Statistical Analysis

Analysis of variance (ANOVA) using a One-Way ANOVA design was used to test

differences among treatments using STATISTICA (StatSoft 1999). Significance of the

difference among means was determined by t-tests of post hoc comparisons using

Tukey's studentized range test at thep = 0.05 level. Means and standard deviations were

used to summarize the data from P sorption and desorption isotherms, where applicable.

Results and Discussion

Ligand-Desorbable P

In order to develop a method that measures ligand-desorbable P, an optimum

extraction time (Experiment 1) and organic anion loading (Experiment 2) must be

identified. Over 80% of the P released from both native and fertilized soils by oxalate

and citrate was within the first hour and three hours, respectively (Fig. 4-2). For both

anions, and for both native and fertilized soils, P desorption reached maximum levels and

stabilized beyond 6 h. Seven hours was chosen as the extraction time for subsequent

experiments.

Oxalate and citrate loading onto native and fertilized samples produced a

diminishing return in desorbable P with respect to increasing anion loading (Fig. 4-3).

Maximum P release coincided with a loading of 0.5 mmol of oxalate or citrate g 1 soil.

Therefore, this loading was chosen for subsequent experiments.









Ligand-Exchangeable P vs. Ligand-Dissolvable P

The premise used in evaluating these data was based on release of Fe from the soil.

If only ligand-exchangeable P was being released, the amount of Fe released should not

be greater than exchangeable Fe, which was 0.05 .imol Fe g1 soil. Once Fe in the

extracting solution was greater than exchangeable Fe, dissolution must have occurred;

and it would no longer be possible to differentiate ligand-exchangeable P from

ligand-dissolvable P. Another background indicator was water-extractable P, which was

0.65 nmol P g1 soil. Phosphorus levels higher than this suggested P desorption by the

action of the organic ions.

Iron released from both native and fertilized soils by 0.02 mmol oxalate g1 soil

was up to 40-fold higher than exchangeable Fe, at a shaking time of 600 min (Fig. 4-2).

When 1 to 5 min shaking times were tested, Fe release under the influence of no oxalate

added ranged from 0.01 to 0.03 imol Fe g1 soil (less than exchangeable) for all shaking

times (Fig. 4-5). When 0.02 and 0.05 mmol oxalate g1 soil were added, Fe release

exceeded exchangeable Fe (0.20 and 0.35 .imol Fe g1 soil, respectively) even for the

1 min shaking time.

The next step was to test if lower anion loadings and shorter contact times would

distinguish exchangeable from dissolvable P. Figure 4-6 shows the effect of low anion

loadings for a 1 min shaking time. Both oxalate and citrate were effective in desorbing P

even at the lowest concentration (0.01 .imol g1 soil). Phosphorus release with oxalate

remained stable between 0.01 and 1 .imol anion g1 soil. Iron release at these

concentrations was below exchangeable Fe, ranging from 0 to 0.04 imol g1 soil.

Beyond 1 jimol anion g1 soil both Fe and P release increased rapidly with increasing Fe

far exceeding the exchangeable amount.









Figure 4-7 shows that the separation of ligand-exchangeable P from

ligand-dissolvable P in the native sample is more a function of anion loading than of

extraction time. When the same loadings were used with a 7 h shaking time, the same

pattern and magnitude of P and Fe were released at anion loadings below 1 imol g 1 soil.

It was concluded from above studies as follows:

* Phosphorus was desorbed at all anion concentrations through the action of anions

* Phosphorus desorbed below 1 imol anion g 1 soil was dominated by
ligand-exchangeable P because Fe release was below exchangeable Fe

* Phosphorus desorbed below 1 itmol anion g 1 soil estimated the total amount of
ligand-exchangeable P because

o It was relatively stable over a 100-fold increase in anion loading

o There was a close correspondence in P desorbed by either anion for both
native and fertilized soil

o Total P desorbed was the same after the 1 min and 7 h shaking times.

It was also concluded that additional P release above 1 .imol anion g 1 soil involved

ligand dissolution because of the high amount of Fe released with the P.

Therefore, the second objective of this study was met because it was shown that

ligand-exchangeable and ligand-dissolvable P were analytically separable under specific

conditions. It was also concluded that the ligand-exchange reaction was more dependent

on anion concentration than on reaction kinetics.

A notable result was the increase in ligand-dissolvable Fe from the native to

fertilized soils (Fig. 4-6). While citrate-extractable Fe increased from 0.16 to

0.20 .mol g 1 soil (25%) with P fertilization; oxalate-extractable Fe increased from 0.15

to 0.34 (127%). Wang et al. (1991) previously noted that mineral dissolution by organic

anions is influenced by the degree ofP saturation of sorption sites. Such sorption









restricts goethite crystallization, leaving the mineral more oxalate-soluble. Conversely, P

sorption reduced the solubility of Fe in Fe(OH)3 in the presence of 500 WpM citrate (Jones

et al. 1996). It is clear that oxalate and citrate differ in their interaction with this

kaolinite-dominated soil. This is shown most clearly by the P:Fe molar ratio relative to

each anion's concentration in the native soil (Fig. 4-4).

Flow calorimetry also showed that oxalate and citrate are sorbed differently onto

soil surfaces. Flow calorimetry has been successfully used for measuring reactions at the

liquid/solid interface (Rhue et al. 2002). Rhue et al. (2002) documented that phosphate

sorption onto soil surfaces was exothermic; while precipitation of Al-phosphate was

endothermic. In our study, phosphate and oxalate sorption were found to be exothermic;

while citrate sorption was endothermic (Appendix C). Unfortunately, this study does not

provide enough information to determine why this difference occurs.

Disequilibria-Desorbable P vs. Ligand-Desorbable P

Disequilibria-desorbable P and ligand-desorbable P are presented in Table 4-1.

The P in disequilibria- and ligand-desorbable pools were a function of the amount of P

sorbed, with all desorbable pools increasing with increasing levels of sorbed P. However,

the percentage of disequilibria-desorbable P to total desorbable P increased, while the

ligand-desorbable P decreased. In the short term, P fertilization preferentially puts P in

the disequilibria pool, which is a pool that all plants and soil-borne organisms can access.

Notably, it was shown that the percentages of total desorbable P to P initially

sorbed remained constant, with only a 6% difference (55 to 61%) across all P-sorption

levels. Total desorbable P was measured 7 to 14 d after the P application, which was the

time required to exhaust the disequilibria-desorbable pool (Fig. 4-8). The shorter time

frames represented less P sorbed, so less time was required for the extraction. When total









desorbable P was measured just 1 d after P addition, these percentages increased to 78 to

87% (Fig. 4-9).

It was concluded that the amount of desorbable P was a function of time, with

decreasing P recovery rates over time. There also appeared to be a pool of sorbed P that

oxalate could not access. This pool ranged between 13 and 22% of the P initially sorbed

even after just 1 d of equilibration time.

Phosphorus Immobilization/Fixation after Sorption

The cumulative P that was immobilized (i.e., inaccessible to Anion Method 2) over

14 d is represented in Figure 4-10. Phosphorus immobilized by the +S-H treatment is

P immobilized by all factors studied under these laboratory conditions. The +S+H

treatment eliminated microbial immobilization, therefore the difference between +S-H

and +S+H was due to microbial immobilization. The treatments +S+H and -S+H should

have removed the effect of microbes, so their difference would be the effect of shaking

the soil.

All treatments were similar after Id, with an average of 0.23 .imol P g 1 soil

immobilized (about 10% of the sorbed P). Phosphorus immobilization increased between

1 and 14 d. The P immobilized by microbes, expressed as a percentage of the total

immobilized, was 19%, 52%, 56%, and 57% on days 3, 7, 10, and 14, respectively. The

P immobilized due to shaking (expressed on the same scale) was 76%, 38%, 30%, and

30% of the total P immobilized. Therefore, P immobilized by what might be considered

short-term ageing was 5%, 10%, 14%, and 13% for the time periods listed above. In the

short run (1 to 3 d), microbes played less of a role than did shaking; but became the major

factor at later dates.









Bliss (2003) showed that, for sandy soil material, microbial immobilization of

fertilizer P reached its maximum approximately 3 d after fertilization. Khoshmanesh

et al. (2002) reported that only a few hours were necessary for sediment bacteria under

aerobic conditions to absorb applied P. Microbial P uptake rapidly contributes to P

immobilization.

The ageing component of P immobilization was plotted against the square root of

time to ascertain if a diffusion mechanism might be implicated (Fig. 4-11). The result

was a positive linear relationship with high goodness of fit (r2 = 0.985). This suggests

that P diffusion could be occurring along the clay surfaces or, less likely, into the clay

structure itself. Assuming a crystal diffusion coefficient of 10 23 cm2 s 1, it was

calculated that diffusion could have moved P approximately 0.07 nm into the clay

structure by 14 d.

Phosphate sorption in soil is known to occur rapidly and to precede an slow

sorption process. The former is often thought to be controlled by ligand exchange and

the latter by precipitation, penetration of the surface, or diffusion into dead-end pores

(Parfitt 1978). Slow sorption is a time-dependent reaction and can account for a

substantial portion of total P sorption in goethite-rich materials (Torrent et al. 1992).

Duration and extent of the slow reaction apparently depend on the ratio between

micropore surface area and total surface area (Torrent et al. 1992), and on the crystallinity

of the materials (Strauss et al. 1997). Therefore, it is important to consider the effect of

slow sorption as it reduces bioavailable P.









100


S80 -
Fe3+ and organic
anion concentrations Citrate
60 = 100 II

,\ Oxalate
40-

Malate
20 -



3 4 5 6 7 8
Soil solution pH

Figure 4-1 Ability of three organic anions to complex Fe as a function of pH as predicted
using Geochem-PC (Parker et al. 1995)











0.10 A 3.00


0.08

2.00
0.06 2.00


0.04
1.00

0.02 ;- Oxalate P Citrate P
Oxalate Fe

O 0.00 0.00o
0 100 200 300 400 500 600

S3.00 B 3.00
o 0
SJQ
O


2.00 2.00





1.00 1.00





0.00 0.00
0 100 200 300 400 500 600

Shaking time (min)


Figure 4-2 Phosphorus-desorption and Fe-release kinetics by organic anions when added
at a loading of 0.02 mmol g 1 soil. A) native soil. B) P-fertilized soil











- Oxalate P


0.30



0.20



0.10



0 0.00
3
-0.
- .
o 5.00
2
iu


- I Oxalate Fe Citrate Fe



~ m m m- m


Ii
I.'.


).00


L~~~ -) -II m 41


0.20


0.40


0.60


0.80


15 't
C-


10

0O


S0
1.00


4.00


3.00


2.00


1.00


0.00


0.00


0.20


0.40


0.60


0.80


1.00


Anion concentrations (mmol g 1)

Figure 4-3 Effect of organic anion loadings for a 7 h interaction time on P desorption and
Fe release. A) native soil. B) P-fertilized soil


0.40 -1 A


Citrate P










0.20


0.16


0.12


0.08 -


0.04


0.00
0.00


S Oxalate Citrate


0.20 0.40 0.60 0.80


1.00


Anion concentrations (mmol g 1)

Figure 4-4 Effect of organic anion loading on the molar ratio of P desorption to Fe
release from native soil


1.00


0.80

-
0.60

-e
j 0.40


0.20


0.00


- 0 "- 0.02 -i0.05 mmol oxalate/g soil


Shaking time (min)

Figure 4-5 Iron release by 3 oxalate loadings as a function of shaking time. The
horizontal line represents exchangeable Fe level (0.05 imol Fe g 1 soil)


r -~F


i~











0.009 A





0.006





0.003





0.000
- 0.01

o
S0.60 B
-e


SOxalate P

-* Oxalate Fe -


Citrate P

Citrate Fe


0 (
S


0.1 1


0.40


0.20


0.00


,.......=:


0.01 0.1 1 10
Anion concentrations (tmol g 1)

Figure 4-6 Effect of low anion loadings for a 1 min interaction time on P desorption and
Fe release. A) native soil. B) P-fertilized soil


0.20



0.15



0.10



0.05

It








0.40


0.30


, 0.20


0.10


0.00












0.06 -

-




o 0.04

-e

o
0



1 0.02






0.00

0.01


Oxalate P

SOxalate Fe -


Citrate P

Citrate Fe


I,
4,
4,
4,
4,


-1 in 0:1 F


Anion concentrations (imol g 1)

Figure 4-7 Effect of low anion loadings for a 7 h interaction time on P desorption and Fe
release


1.50



1.20








0.60
It
,0..


0.30



0.00


-






81



0.50 1.29 -2.58 5.81 imol P/g initially added


S0.40



0.30


0.20



0.10


0.00
1 2 3 4 5 6 7 8 9 10 11 12 13 14

Number of desorption steps (d)

Figure 4-8 Cumulative P desorbed by AEMs from soil at 3 increasing levels of P. Each
desorption step represents one day












Table 4-1 Disequilibria- and ligand-desorbable pools of P when 3 levels of P are sorbed to the soil


Disequilibria-desorbable P
Lmol g1 (x10 2) %t


6.4 a
31.0b


9a
24b


Ligand-desorbable P
Lmol g1 %T


0.63a
0.98b


91a
76b


Total desorbable P

tmol g1 /
0.69a 51
1.29b 5:


5.81 4.04 83.8c 34c 1.61c 66c 2.44c 61b
% of total desorbable P
Total desorbable P as a % of P sorbed
Net P desorbed subtracting P desorbed at no P addition
Different letters assigned in a column denote significant differences among P-added levels at the 0.05 probability level,
performed by Tukey's studentized range test


P initially
added


P initially
sorbed


Lmol g


1.29
2.58


1.24
2.33










100% OLigand desorption (7-14 d)
SDisequilibria desorption (7-14 d)
O Ligand desorption (1 d)
80% -


60%


40%


20%


1.29


2.58
P initially added (gmol g 1)


5.81


Figure 4-9 Distribution of sorbed P between the disequilibria- and ligand-desorbable
pools. The first bar in each P-added level shows a combination of
disequilibria- and ligand-desorbable pools; and the second bar shows total
desorbable P 1 d after P sorption










0.80


0.60



0.40



0.20


-*-+S-H

-U- +S+H

-S+H


0.00


0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Day

Figure 4-10 Cumulative P immobilization after P sorption. +S and -S are with and
without shaking; and +H and -H are with and without mercuric chloride for
microbial suppression




0.10


0.08


0.06


0.04


0.02


0.00


y= 0.0308x- 0.0364

r2 = 0.985


2 ay3


Figure 4-11 Cumulative P immobilization versus the square root of time. Immobilized P
is that amount of P immobilized after the effects of microbes and shaking
have been removed














CHAPTER 5
SUMMARY, CONCLUSIONS, AND FUTURE RESEARCH

Tropical soils are notorious for their low nutrient bioavailability and low

agricultural productivity caused by unfavorable soil properties and socio-economic

constraints. Expert nutrient management is further complicated by difficulties in the

extension of information. Many parts of tropical Brazil fit this scenario. In humid, acid

soils of Brazil, phosphorus deficiency can control crop production, yet the soil dynamics

of P are not well-understood at a process level. Because of low pH and Al toxicity,

liming is a common soil management practice. However, effect of liming on P

bioavailability is controversial and appears related to the amount of exchangeable Al

present. Phosphorus sorption and, more importantly, P desorption are processes affecting

P bioavailability. A more complete understanding of these processes, including the effect

of pH change on P processes, is fundamental to establishing P management strategies.

Unfortunately, quantitative measurement of bioavailable P pools is not well-defined.

Standard methods are required to attain a meaningful database for P bioavailability.

Based on the above observations, this dissertation investigated specific aspects of

P availability in soils from three points of view: method development; evaluation of the

practical consequences of pH adjustment for soils; and fundamental investigations into

the forms of inorganic bioavailable soil P.

Chapters 2 and 4 each addressed methodology questions. Desorption is a

time-consuming and rarely investigated aspect of P dynamics. Chapter 2 investigated

two techniques for measuring P desorption isotherms. Chapter 4 resulted in a method for









measuring inorganic P that is desorbable by some low molecular weight organic anions

(ligand-desorbable P) that are characteristic of root and microbial exudates. This chapter

also addressed whether there were experimental artifacts associated with in this method.

In Chapter 3, the effect of pH change on P sorption and desorption was investigated

and interpreted with a view toward the degree to which this practice may help or hinder

P-fertilizer availability. Chapter 4 further addressed potential pools of bioavailable

inorganic P by trying to differentiate ligand-exchangeable from ligand-desorbable

inorganic P. It also investigated how fertilizer P is partitioned between these two pools.

The combined studies of Chapters 2 to 4 provide detailed information for a specific

soil type in Brazil; but also provide a path for continued research on P bioavailability in

tropical soils. Each chapter was organized around a set of specific objectives, with a

summary of results provided below.

Summary

Suitability of Using Multiple AEM Strips to Develop Desorption Isotherms

The Multiple AEM Method was found suitable for developing well-defined

P desorption isotherms. Isotherms were developed quickly and precisely over a range of

P sorption levels. This method gave a high density of data points over a 3-day period.

Comparison of Multiple AEM Method to Sequential AEM Method

When the Multiple AEM Method was compared to the Sequential AEM Method, it

was found that the desorption isotherms were numerically different. However, partition

coefficients (Kd values) calculated from the isotherms were not a function of the

desorption method or the amount of P initially sorbed; but were inversely and linearly

related to equilibrium P concentration. If the intent of developing desorption isotherms

was to derive Kd values, then either method was appropriate. It was concluded that over a









range of P solution concentration, the Multiple AEM Method was superior to the

Sequential AEM Method because

* Isotherms were developed in less time, minimizing outside and time-related
influences on isotherms

* It yielded more desorption points in a shorter time frame.

Effect of Soil pH on Phosphorus Sorption

Phosphorus sorption decreased with increasing pH; and calculated Kd values

decreased with pH increase at low solution P concentrations. At solution concentrations

higher than 3 tg P mL1, there was no effect of pH. Liming to pH values of 5.9 and 7.0

resulted in 2.8 to 4.5 kg P fertilizer ha 1 (calculated at 0.2 [tg P mL 1 solution

concentration) that was not sorbed and remained in the soil solution ready for plant

uptake.

Effect of Soil pH on Phosphorus Desorption

P desorption increased as soil pH increased. Total P desorbed, the ratio of P

desorbed to P sorbed, and Kd values increased as pH increased for all levels of P addition

when soil pH was raised to 7. This resulted in an additional 0.2 to 1.0 and 0.7 to

2.7 kg P ha 1 that was desorbable at pH 6 and 7, respectively, which can be attributed to

the effect of liming.

Combining the results of P sorption and desorption, liming had a twofold positive

effect. It reduced the amount of P sorbed and increased the amount of P desorbed, which

increased both solution concentration and P bioavailability.

Method for Measuring Ligand-Desorbable Phosphorus

Oxalate and citrate were used in kinetic and loading studies to define a method that

would remove the total amount of P accessible by these anions in the most efficient