TRANSFU‘RMATION OF ADDED PHOSPHORUS IN THREE MICHIGAN SOILS Thesis for the Degree of M. S. MECHIGAN STATE UNNERSITY Gary Darwin Rinkanberger 198666 ' (may: I -— z in L [B RA R Y ‘§ Michigan State University (J1 thliifl $335 0. ABSTRACT TRANSFORMATION OF ADDED PHOSPHORUS IN THREE MICHIGAN SOILS by Gary Darwin Rinkenberger Laboratory studies were performed to trace the changes in dicalcium phosphate upon its addition to three Michigan soils. Inorganic P fractions (Chang and Jackson's procedure) were compared with phase diagrams and P po- tentials for the untreated soils and soils to which 200 and 500 ppm P were added. The soils were incubated for two months after P additions. Periodic samples were taken for analysis throughout the incubation. Dicalcium phosphate was found to convert readily to Al—P in both Iron River silt loam and Warsaw loam. A small quantity of the newly formed Al-P was apparently transformed to Fe-P in the Iron River soil during a two month incubation. 4 + 1/3 pAl vs. pH—l/3 pAl) indicated that Al-P or Fe-P Phosphorus phase diagrams (pHZPO vs. pH) and P potentials (pH2P04 compounds were controlling the activity of P in these soils which- is consistant with P fractionation data. The P level was found to be somewhat higher in the Warsaw soil than in the Iron River soil. Gary Darwin Rinkenberger Dicalcium phosphate was not completely removed from the calcareous Wisner silty clay loam by one extraction with NH4Cl. Thus, when Chang and Jackson's fractionation pro- cedure was followed precisely, the added CaHPO4 appeared in the Al-P fraction-—a conclusion contrary to expectation and the data from P solubility diagrams. But after the CaHPO4 was removed by successive extractions with NH4C1, little change in the Al—P fraction was noted when CaHPO4 was added. This result was consistant with phase diagram and P po- tential (pH2P0 + % pCa vs. pH-% pCa) data, indicating that 4 Ca-P compounds control P activity. TRANSFORMATION OF ADDED PHOSPHORUS IN THREE MICHIGAN SOILS BY Gary Darwin Rinkenberger A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Soil Science 1966 ACKNOWLEDGMENTS The author expresses his gratitude to Dr. B. G. Ellis for guidance throughout this investigation and in the preparation of the thesis. Dr. Ellis's willingness to dis— cuss with enthusiasm problems related to this research at almost anytime is sincerely appreciated. Grateful ac— knowledgment is extended to Anthony S. R. Juo for his sug— gestions about laboratory procedures, and for his ideas on interpretation of the data received. Also, appreciation is extended to his wife, Kathy, for typing the rough draft and for her support through his Master of Science program. ii TABLE OF CONTENTS INTRODUCTION REVIEW OF LITERATURE Fixation of Inorganic Phosphorus in Soils Phosphorus Fixation in Acid Soils Phosphorus Fixation in Alkaline Soils Solubility Product Principle Phosphate Phase Equilibrium Phosphate Potentials Limitations in the Use of the Solubility Product Principle . . . . . . . Fractionation of Inorganic Soil Phosphates .Modified Fractionation Procedure MATERIALS AND METHODS RESULTS AND DISCUSSION Fractionation of Inorganic Phosphates Solubility Product Diagrams SUMMARY AND CONCLUSIONS BIBLIOGRAPHY iii Page pg: 12 16 21 22 24 28 32 32 43 61 65 Table Interaction fractions and time Interaction fractions Interaction fractions time of in of in of in LIST OF TABLES added P, inorganic P a Wisner silty clay loam added P, inorganic P a Warsaw loam and time added P, inorganic P a Iron River silt loam and Successive l_N NH4C1, pH 7.0, extractions of Wisner silty treated with bated two months clay loam, untreated and 500 ppm P as CaHPO4, incu— Successive extractions with NH4C1 of treated (200 ppm P as CaHPO4) and untreated Warsaw loam . . . . . . iv Page 33 34 35 4O 44 Figure LIST OF FIGURES Phosphorus removed from Wisner silty clay loam by successive extraction with 1.5 NH4C1 Theoretical Al-P isotherms as influenced by increased solubility of amorphous variscite and/or gibbsite Theoretical aluminum and phosphate po- tentials as influenced by increased solu- bility of amorphous variscite Solubility-phase diagram for Iron River silt loam . . Solubility-phase diagram for Warsaw loam Solubility-phase diagram for Wisner silty clay loam . . . . . . . . . Lime and phosphate potential diagram for Wisner silty clay loam Aluminum and phosphate potential diagram for Iron River silt loam and warsaw loam Page 41 48 48 50 52 54 56 57 INTRODUCTION Phosphorus was one of the first elements to be recog— nized as essential for plant growth. In addition, many agronomists have considered it to be the most limiting ele- ment in food production throughout the world. Consequently, much effort has been expended in studying the chemistry of soil P. But the nature of soil P has been elusive. Volumes of literature have reported a "loss" of soluble P when water soluble forms such as Ca(H2PO are added to soils. And 4)2 hypotheses have been formulated to explain this disappear- ance; many of which have failed to weather the test of time. A procedure for fractionating inorganic soil P into water soluble (NH4C1 extractable), Al-P, Fe—P and Ca-P fractions was developed by Chang and Jackson (14). This gave promise of describing the chemical transformations of P when added to soil. Volumes of data concerning P fraction- ation in soils have been reported. General trends in trans— formations of P added to soils have been summarized from these data. The fractionation procedure has not been as pre- cise as desired, causing many researchers to propose modifi— cations. Therefore, interpretation of the inclusion of specific P minerals in the various fractions may be approximate. Thurlow (93) reported a major increase in the Al-P fraction of a calcareous Wisner soil when treated with P as NH4(H2PO4) in greenhouse studies. But this should not be expected in a calcareous soil. Sanchez (82) studied the fractions of P in Conover and Pewamo acid soils in a labora- tory incubation experiment. When the soils were treated with various rates of KHPO4 solution, fractionation indi- cated that most of the increase as a result of added P oc— curred in the Al-P and Fe-P fractions, with significant amounts in the water soluble fraction at the higher P levels. While these results are more plausible for the acid soils, questions may still be raised concerning their absolute accuracy. Phase equilibrium diagrams have been used in an at— tempt to indicate the type of soil minerals that control P activity in solution (1,45,46,47,48,49,50,51,58,102,103). Lack of knowledge about Specific P minerals in soils and possible chemical changes that may occur in the dynamic soil system has hampered interpretation of solubility data. Al- though this method yields only indirect evidence about the nature of soil P, it does allow for interpretation of the interactions of P and pH that have been observed in soil fertility. This investigation was designed to study the change in soil inorganic P when P was added as CaHPO4. The ob- jectives were: To compare P fractionation data with phase equi- librium data evaluating each method as it is used to trace added P in soils. To study methods of removing CaHPO4 from soils prior to P fractionation. REVIEW OF LITERATURE Fixation of Inorganic Phos- rphorus in Soils Retention of P by soils has been noted for more than 100 years. According to Wild (99), Way first demonstrated this phenomenon. Since that time volumes of data concern- ing P in soils have been collected. Several articles summarizing the literature con- cerned with the chemistry of inorganic P in soils have been published. In more recent times, reviews by Wild (99) and Dean (24) covered the retention and fixation of P in soils. Hemwall (35) termed inorganic P literature as confusion. Smith (84) presented a recent review of A1 and Fe phosphates in soils, covering their reactions and availability. The term "P fixation" will refer here to the re— tention of added P in soil by means of adsorption and/or precipitation. Fe and A1 phosphates have been found in abundance in acid soils and Ca phosphates in abundance in alkaline soils (9,12,15,16,l9,36,52,54,56,58,64,79,82,85,93,94,98). A general relationship between synthetic soil minerals was suggested by Rathje (75): _ , A1 . Al Ca P s + Fe hydrox1de.E==E Fe phosphates + Ca(OH)2 The equilibrium would shift to the right in acid media, and to the left in alkaline media. Russell (78) stated that in acid soils P occurred in association with Fe and Al compounds. Fixation in Acid Soils Smith (84) suggests that four approaches have been used to study the fixation reactions of Fe and Al phosphate compounds in acid soils. First, the correlation of Fe and Al oxide content of soils with the amount of P fixed has been reported by many workers. Toth (92) demonstrated that removal of free Fe and Al oxides in soils resulted in re— duced P fixation. Swenson,_g£_§l. (86) reported significant contribution to P fixation by Fe and Al oxides. Perkins, §£;§l. (73) related high concentrations of A1 and Fe oxides to high P fixation. The rate of reaction of P with the soil was greatly reduced when the extractable Fe oxides were re— moved, Kittrich and Jackson (42). Upon the addition of soluble P to acid soils, in— creased amounts of Fe and Al phosphates have been recovered. Chang and Jackson (15) reported these increases. Saeki and Okamoto (80) recovered 60—70 percent in the form of Fe and Al phosphates. Volk and McLean (94) reported recovering almost all of the added P as Al and Fe phosphates when treat— ing acid soil with soluble P and fractionating according to the procedure of Chang and Jackson (14). Following the addition of Al or Fe compounds to the soil a P fixation increase was demonstrated by Larsen et a1. (50). A fourth approach has been the removal or inacti- vation of Fe and A1 in solution which results in reduced P fixation. Using ferrocyanide, fulvic acid, and other agents which form compounds with Fe and Al, Leaver and Russel (53) demonstrated reduced P fixing power in soils. Both precipitation and adsorption mechanisms have been considered in P fixation. A simplified representation of the mechanism in acid soils was proposed by Smith (84): Adsorption Al(OH)3 + H2POZ-Vzi‘ A1(OH)2H2PO4 + OH— Precipitation +3 - + Al + 2H20 + H2P04 ‘__———> A1(OH)2H2PO4 + 2H If this relationship occurs in a pure system it is easy to ascertain which mechanism is proceeding by following the concentration of H+ or OH- ions. But in the complex soil system other simultaneous reactions prevent such simple studies. Rennie and McKercher (77) used the Langmuir ad- sorption isotherm to calculate an adsorption maximum for acid soils and found that it was a reliable indicator of the soils ability to supply P in plants. Close agreement be- tween P fixation and Langmuir adsorption isotherm may not necessarily mean that an adsorption reaction has taken place, Hsu and Rennie (40). Hydroxide exchange sites were reported to be associ— ated with P fixation possibly indicating adsorption, Baker (5). Bache (4) and also Raupach (76) using pure clay and oxide mineral systems concluded that free energy surface re- actions may be a better explanation for fixation than pre— cipitation and the use of solubility products of soil minerals. Kittrick and Jackson (42) observed under an electron microscope that Fe and Al hydroxide minerals in contact with P solution showed in a few minutes the formation of separate phase, phosphate crystals by solution precipitation. Hemwall (37) hypothesized that P is fixed by clay minerals by reacting with soluble Al which originates from the ex— change sites or from lattice dissociation of the clay minerals to form a highly insoluble Al phosphate. Evidence for this occurrance is that clays support an appreciable Al concentration in solution, and there is a solubility product relationship between the Al and P concentration in clay—P suspensions. Bache (4) reported that the ion activities in so- lution during sorption of P on gibbsite and hydrous ferric oxide takes place in three stages of reaction: (a) A high energy chemisorption of small amounts of P; (b) Precipitation of a separate P phase; (c) A low energy sorption of P onto precipitates. Fixation in Alkaline Soils Eisenberger, Eggal, (26) reported that there exists a continuous series of solid compounds between CaO and CaHPO4 in the calcareous soil system. Hemwall (36) stated that P fixation in alkaline and calcareous soils is usually at- tributed to the formation of Ca—phosphates. Fe and Al com- pounds may be responsible for some fixation in soils of higher pH; however, Ca phosphates predominate. Ellis and Troug (27) demonstrated that Ca saturated montmorillonite treated to remove Fe and A1 oxides fixed large amounts of phosphorus against water extraction. But the P was recovered with a weak acid extraction. Fixation of P‘by clay before and after removal of Fe and Al oxides indicated that the free oxides accounted for most of the fixation. Evidence also indicated that P fixed by Ca satu- rated clay is fixed as a Ca complex and not by a H2PO4-Ca- clay bonding as hypothesized by some workers. Clays satu- rated with Na, K, or Mg fixed only small amounts of P against water extraction. Pratt and Thorne (74) reported that Ca clays fixed ITKDre P in alkaline systems than did Na clays; however, no eaatplanation of possible mechanisms was given. Boischott, g£_al. (6) reported on the importance of CZEBCO in the fixation of P in calcareous soils. The initial 3 Jreaaction was indicated to be an adsorption onto the CaCO3, Iraather than a precipitation reaction of an insoluble sepa— Iraate phase. The amount of P adsorbed was a function of the jfzineness of the CaCO material. When the P concentration 3 was raised to 5 X 10-.5 M at pH 8.3 to 8.5, precipitation oc- <2Lxrred. Olsen (72) also indicated that P was adsorbed on 4 CaCO3 until a concentration of 2 X 10- M K2HPO4 was reached; 'tduen precipitation took place. When the P concentration was 4 greater than 2 x 10" M KZHPO4, ‘ECIUilibrium concentration dropped below that found before precipitation began and the EDIHecipitation began. Calcium phosphate crystals were noted growing on the CaCO3 surface. Fried and Shapiro (32) suggest that the two ap— EDINoaches to P fixation, mineralogical precipitation and ad- Sc>rption, are not necessarily incompatible. Since the soil j-SS a dynamic system, the possibility of both processes oc- <2Llrring is plausible, Smith (84). ~s§531ubility Product Principle The lack of reliable solubility information necessary 13C>r thermodynamic interpretation of soil P reactions lO benevented advancement for many years, Lindsay and Moreno (£58). More recently much work has been done in this area. In order to use the solubility product principle in :Lr1terpreting the P status of soils, the solubility of vari— C>L1s phosphates that occur or have been suggested to be E31resent in the soil must be known. The following list con- tZEiinS many of the compounds studied and their szp values as <>13tained by different workers. Compound Chemical formula szp Author .1J Gibbsite Al(OH)3 33.8 Lindsay and Moreno (58) 2. Variscite Al(OH) 2H 2PO4 30.5 Lindsay, et a1. (58) Variscite Al(OH) 2H 2PO4 29.5 Cole and Jackson (21) Variscite Al(OH)2H 2 m4 30.5 Bache (3) Variscite Al (OH)2H 2 m4 30.6- Wright and 31.7 Peech (103) Variscite Al(OleHZPO4 28.4— Kittrick and 27.7 Jackson (42) Variscite AlPO4.2H20 22.5 Taylor and Gurney (88) 13. Strengite Fe(OH)2H2PO4 33.6— Chang and 35.0 Jackson (13) Strengite Fe(OH)2H 2 m4 34.3 Bache (3) Strengite Fe(OH)2H 2 m4 34.7 Wright and Peech (103) Strengite Fe(OH)2H2PO4 35.3 Huffman and Taylor (39) ll Compound Chemical formula ,szp Author strengite Fe(OH) H P0 34.6 Egan, et al. 2 2 4 -—-- (25) 4. Colloidal iron 33.3 Egan, et a1. phosphate (25) 5. Ferric FeOOH 38.1 Lamb and hydroxide Jacques (24) 6. Dicalcium phosphate CaHPO4'2H20 6.56 Moreno, et a1. dihydrate (68) 7. Dicalcium phosphate CaHPO4 6.66 Lindsay and anhydrate Moreno (58) 8. Octacalcium Ca4H(PO4)3'3H20 46.9 Moreno, et al. phosphate (68) 9. Hydroxyapatite Ca (OH)2 113.7 Lindsay and (PO ) 10 4 6 Moreno (58) 3.0. Fluorapatite Ca (PO ) (F) 118.4 Lindsay and 10 4 6 2 Moreno (58) The pK values for variscite were redetermined by UDaylor and Gurney (88), who showed that the szp for varis- Ckite was 22.5 when equilibrated with dilute HCl and 21.5 3P04. The dissociation Iteeaction for variscite was given to be AlPO4 = A1 + PO4 with c3'C>rresponding szp = pAl + pH2P04-2pH + pK2K3, rather than EDIKsp = pAl + 2POH + szPO4 as reported by earlier workers. ablle difference in szp determined in the dilute HCl system ‘Vllen equilibrated with dilute H QCDmpared to the dilute H3PO4 system was probably due to CIEBZLculations of ionic activity affecting one set of data 12 more than another. A lack of information about complex ions in dilute H3PO4 solutions and resultant dissociation has prevented the determination of a true solubility product for variscite. The szp values for P minerals which represents the various minerals evolved in P reactions may be only approxi- mate. Methods used to determine these values must be con— sidered when interpreting accurate solubility product measures, Smith (84). Phosphate Phase Equilibrium Lindsay and Moreno (58) presented the activity iso- therms for Al(OH)2H2PO4 (Variscite), Fe(OH)2H2PO4 (strengite), CalO(PO4)6F2 (fluorapatite), CalO(PO (OH)2 (hydroxyapatite), 4)6 4)3 3H20 (octacalc1um phosphate), and CaHPO4'2H20 (dicalcium phosphate dihydrate) on a single solubility dia- Ca4H(PO gram in which HZPO; activity was plotted as a function of pH. The diagram was proposed for predicting the formation and stability of various Al, Fe and Ca phosphate compounds that may be present in natural soils or form in soils upon fertilization. Generally, only the most soluble phosphate compounds dissolve or precipitate fast enough in soils to govern phosphate activity in solution. Dicalcium phosphate, possibly octacalcium phosphate, and the more soluble alkali or ammonium Fe and Al phosphates slowly dissolve as insoluble forms like hydroxyapatite, fluorapatite, strengite, and l3 variscite precipitate. Amorphous, freshly precipitated phosphates of Fe and Al will have greater activities than crystalline strengite or variscite. An example of use of the solubility diagram was pre- sented by Lindsay and Moreno (58): If lime is added to an acid soil in equilibrium with gibbsite and variscite at pH 4.0 (pH2PO =6.7) until pH 7.0 4 is obtained, there would be a tendency to precipitate gibbsite and dissolve variscite until pH2PO4=3.7 is reached. The resulting increase in phosphorus in solution could also result in supersaturation of calcium phosphates, and thus the precipitation of CaHPO4-—-—) Octa-Ca-P with possibly eventual formation of hydroxyapatite. If this occurred pH2P04 would go from 3.7 to 8.8. Lehr and Brown (54) placed granules of CaHPO4°2H20 in Hartsell fine sandy loam limed to pH 7.2 and 7.8, and cropped the soil with ryegrass. They observed by X—ray an- alysis a mixture of octacalcium phosphate and smaller amounts of apatite occurring in the residue of the root zone. Only a few crystals of CaHPO4 remained under the alkaline cone ditions. In three acid Hartsell soils studied, 95 percent of the residue was a mixture of CaHPO4 and CaHPO4°2H20. Lindsay and Stephenson (59) studied solutions of Ca(H2P04)2°H20 and the reactions of bands of Ca(H2PO4)2°H20 with soil. Excess Ca(H2PO4)2°H20 dissolved in water, 56 T? ry \ A\H pa ta Ce l4 attained metastable—triple—point-solution (4.50 M P, 1.35M Ca, pH 1.00) equilibrium in one hour and remained near that point for 24 hours. The solution then approached a triple— point—solution (3.98M P, 1.44M Ca,.pH 1.48) at three days until it was attained at 17 days. The reactions are repre- sented below. (1) Ca(H2PO HO-—->CaHPO +HPO +HO 4)2 2 CaHPo '2HO+MTPS (2) ca(H2PO4)2 2 2 <:-- 4 2 °HO+XH (3) ca(312130432 2 2 __:s O-<:— CaHPo4 + TPS The Ca(H2PO 'H 0 band in the Hartsell soil attained a compo- 4)2 2 sition in the soil solution similar to metastable-triple-point solution. Upon dissolution of Ca(H2PO4)2'HZO in the band a very low pH was attained. This highly concentrated solution dissolved Fe, Al, Mn and other soil constituents. As the pH increased gradually, precipitation of Fe, Al, and Ca com- pounds followed. Dicalcium phosphate was noted to precipi— tate most abundantly at the advancing concentration front. Ca(H2PO4)2°H20 had completely dissolved after eight days. It was noted that the greater the Ca content, the greater possibility that Ca will have a dominating role in precipi- tation of fertilizer P. Fe and Al were readily dissolved from the acid Hartsell soil, and will likely predominate in initial precipitation of P leaving the granule. Bouldin, et a1. (7) noted that the fraction of added P that remained as a residue at the granule site varied from 15 92 percent when Ca(H2PO4)2'H20 was mixed with CaCO3, to two percent when mixed with NH4SO4. Moreno, et a1. (69) leached a column of CaHPO4°2H20 ‘with six liters of water and noted the formation of octa— calcium phosphate in the upper column as determined by X—ray and petrographic analysis, indicating possible occurrence in soil. Studies of the reactions of P solution with acid soils, with Al and Fe oxides, and with silicate clays (38, 42), show that when crystalline products are detected they are usually simple Al, Fe phosphates related to variscite or alkali Al and Fe phosphates related to taranakite (H6K3A15(PO4)'18H20). Fairly concentrated acid solutions Inust be used in order to detect crystals. Therefore alter— ations of oxide and clay minerals may occur. Bache (3) studying the effects of time of reaction, EJH of solution and solid to solution ratio concluded that \fariscite can be in equilibrium with the surface and solu— tZion phase only in the most acid soils, and strengite is rlever likely to be in equilibrium. Further, thermodynamic (Zonstants for variscite and strengite are only maintained at -143w pH. At higher pH's surface hydrolysis reactions occur 1T€eleasing P ions into solution and forming more basic in— £3<>luble metal P. Taylor and Gurney (88) also noted incon— §3"1?uent dissolution of variscite as the pH was increased, a 1:”Eésult which could not be explained as precipitation of l6 Al(OH)3 as their system was undersaturated with respect to gibbsite. Wada (95) indicated that solubility equilibrium of variscite and gibbsite might be established below pH 5.2, but apparently not higher on the ando, alluvial, and red yellow podzolic soils studied. Taylor,_g§_al. (90) using a Ca-P solution in the pressence of gibbstie and goethite noted the principle precipitation was from Al(OH)3 and con— cluded that unless large amounts of reactive Fe are present, most P will precipitate as Al phosphate. jhosphate Potentials Schofield (83) suggested CaCl2 as an electrolyte to determine the chemical potentials of soil P. Phosphate po- tential is measured as an approximation to chemical po- ‘tential (100). The chemical potential of P in the solid Eahase is a measure of its ease of removal. Since nearly all ILabile inorganic soil P is held by the solid portion of ssoil, the availability and release must be determined by the eamount and form held by the solid phase. Schofield (83) 1first used P potential (%pCa + pH2PO4) versus lime potential (1§H-%pCa) as proposed by Aslying (l) to identify mineral 15C>rms of P in calcareous soils. The solubilities of Ca Ebllosphates may be represented on a single solubility diagram -i~11 which the functions of chemical potentials for Ca(OH)2 £3 rid Ca(H2PO4)2 are used as coordinates. Balanced equations 17 can be written and a solubility line represented on the diagram. Equations for CaHPO4 and hydroxyapatite are given below. dicalcium phosphate Ca(OH)2 + Ca(H2PO4)2 ? CaHPO4 + 2H20 hydroxyapatite 7 Ca(OH)2 + 3 Ca(H2PO4_)2 : ca10(PO4)6(OH)2 + 12 H20 Similar basic solubility relationships were used in Lindsay and Moreno's phase diagrams (58). For the general case Clark and Peech (l9) repre- sented the formation of a hypothetical calcium phosphate as: M Ca(OH)2 + N Ca(H2PO4)2 : solid + Z H20 ‘where M and N represent reacting moles of Ca(OH)2 and Ca (H2PO4)2 forming one mole of solid + Z H20. Based on the assumption that the reactants are completely dissociated the .linear plot of %p Ca + pH2PO4(P potential) versus pH —2p Ca (lime potential) should have a slope of M/N characteristic (of the solid species. For example, octacalcium phosphate <3a4H(PO would be 5/3 = 1.67 slope. 4)3 Similar solubility diagrams may be constructed for 231(OH)3-A1PO4, and Fe(OH)3-FePO4 systems. Taylor and Gurney (89) equilibrated acid soil in dilute CaCl2 and I;Hlotted pH + pH2P04 as a function of pH —l/3pAl. Points for illne untreated soil fell on the variscite line. The variscite ‘3L- (pH+pH 2P04) -3 (pH-l/3pAl) Taylor and Gurney concluded that the results do not neces— sarily mean that variscite controls the composition of the solution. Chakrovarti and Tablibudeen (10) examined 54 British and Indian soils plotting Fe potentials (pH -l/3Fe+3) and Al potentials (pH -l/3Al+3) as a function of P potential [1/3p (Al+3 Fe+3) +pH2PO4—1. For the determination 5 gram samples of soil were shaken in 100 ml of 0.02 M KCl for eight days. Results were interpreted as related to soils grouped according to pH. The temperate soils in pH range 3.8 - 4.2 corresponded to compounds approximating to the composition of strengite. Tropical soils pH 3.8 - 6.7 indicated strengite coexisting with hydrated Fe oxides. Variscite type com- pounds controlled P concentration in temperate climate soils up to pH 4.7. Above 4.7 non—stoichiometric P: hydroxide adsorption complexes may be controlling. Tropical soils in the range of pH 4.3 - 5.8 indicated that the system was con- trolled by compounds similar to variscite. Taylor and Gurney (88) reported that upon acidifi— Cation of a potential system, the points moved parallel to ‘the variscite isotherm and not toward it as if variscite were (iissolving. The movement was due almost entirely to pH and -1\l content change, probably associated with clays. Wright 'Eind Peech (103) made five successive extractions with 200 ml 19 0.01 M CaCl2 plus 2g soil shaking for 72 hours. The con- stantcy of ion product of variscite was maintained. Moreno, _§E_§1. (70) placed one gram of Ca(H2PO4)2'H20 in 50 g of Hartsell acid soil shaking in 100 ml of water. After one hour and at 30 days the system, represented by lime po— tential versus P potential, was still on the CaHPO4'2H20 isotherm. In a similar experiment with the acid Hartsell, after 20 days of shaking in 100 m1 of water, 0.1 grams of CaHPO '2H20 in soil resulted in points below the apatite iso- 4 therm, 0.3 grams of CaHPO4'2H20 in soil was on the apatite isotherm, 1.2 grams of CaHPO4'2H20 in soil was on CaHPO4'2H20 isotherm, and 5.0 grams of CaHPO4°2H20 in soil was above the CaHPO4'2H20 isotherm. Weir and Soper (98) studied P and lime potentials in calcareous Manitoba soils. Using 20 grams of soil in 50 ml 0.01 M CaCl2 and shaking 12 hours they noted that all soils xwere supersaturated with respect to hydroxapatite. Ferti— lized soils fell in a region closer to the octacalcium phos- jEMate isotherm, with P activity seemingly governed by octa— <:alcium phosphate and CaHPO4°2H20. Withee and Ellis (102) 'treated two calcareous soils with 200 and 500 ppm P as H3PO4 E3nd noted equilibrium at the 200 ppm level with octacalcium Eihosphate, and equilibrium at 500 ppm level similar to Camp4 2H20. 20 Procedures for measuring phosphate potentials have received much attention recently. Nethsinghe (71) noted that air drying of soils caused a change in chemical po- tentials of soil P not reversable upon rewetting. Larsen and Court (48) reported that P enrichment has an effect on potential varying markedly with soil to solution ratio. Several possible reasons for this variance were proposed. (a) (b) (C) (d) (e) (f) (9) An energy effect in which mean bonding energies change as P is removed; The mechanism of adsorption may be dependent on a unit weight of soil; Incongruent dissolution of solids may occur; Solutions of ions related to a definite solubility product may cause a shift in the equilibrium; P of indefinite forms may be unexplainable; Soil organic matter reactions are probably not a factor in 0.01 CaCl2 solution; Variation in the cation balance in the system is not probable, as a 10% change in 0.01 CaCl2 so— lution results in only 1% change in potential. Lindsay et a1. (57) concluded that pH—1/3pAl (aluminum Emotential) for 100 grams of soil in 200 ml of CaCl2 varying 15rom.0.01 M to 0.10 M remained constant. Frink and Peech (.33) reported that equilibrium of gibbsite and Al(Cl)3 so- -J—ution was attained slowly (i.e., 1—3 months). Limed acid Soils, one gram in 200 ml 0.01 M CaCl 2., were not in 21 equilibrium in one month; however, unlimed soils were in equilibrium. Larsen and Widdowson (51) studied the production of CO2 at different soil to solution ratios concluding that it seemed to explain the variation in pH. A reduction in pH and an increase in P concentration (decreased P potential) was noted with increasing weight of soil in the suspension. Ten grams of four soils were shaken for 16 hours in 50 ml of CaCl Larsen (45). The lime potential was independent of 2, the CaCl2 concentration between 0.002 to 0.05 M. The P potential of neutral and alkaline soils decreased with in- creased molarity of CaC12. If allowance was made for the formation of a CaHPO4 complex, P potential was also inde- pendent of CaCl molarity in the pH range 5 to 8. Larsen 2 (46) noted higher lime potentials when calcareous soils were shaken in a CaCl2 solution in stoppered bottles as a result of CO2 pressure not present in bottles open to the atmos- phere. White and Beckett (99) also reported that aeration was critical in determining potentials. Limitation in the use of the Solubility Product Principle Lewin (55) pointed out limitations in the use of the solubility product principle. There is a neglect of any possible variations in the activities of solvent molecules and the solid phase in determining activity products. 22 Equilibrium conditions may not be present in most precipi- tation reactions. The equilibrium of a soil with an ex— tracting solution requires equilibrium between difficulty soluble compounds and its constituent ions adsorbed by the clay. Adsorbed ions obtain equilibrium rapidly. But diffi- culty soluble compounds are probably rate controlling, Frink and Peech (33). This implies that assessments of P ion activity in soil solution based on the solubility product principle are bound to be approximate. No exact definite re— lationship between pH and H2PO _ was noted in arable soils 4 studied by Wada (95), as would exist if H PO - activity in 2 4 the soil solution was governed by simple phosphates and re- lated compounds. Results indicated that operative ranges of simple solubility are rather limited for interpreting P transformations. P fertilizer application is small in com- parison with amounts of aluminosilicates, iron oxide minerals, and calcium compounds present in soils, and there- fore P fertilizers may not increase P in soil solution very much, Kittrick and Jackson (42). Fractionation of Inorganic Soil Phosphates Chang and Jackson (14) proposed a fractionation pro- cedure to separate soil inorganic P into water soluble Ca, Al, Fe and reductant soluble P removed after the first four forms. Since its develOpment the procedure has been widely 23 used. Fractionation of fertilized acid soils, Wright and Peech (103), showed an increase mainly as Fe-P, although two soils increased in Al-P. The Ca-P fraction was affected little. In acid Taiwan soils after three days incubation most of the 200 ppm P added was fixed in the Al-P fraction. Some Fe-P and Ca-P were indicated, Chang and Chu (12). After one-hundred days however, the percent recovery was more variable with a noticeable increase in Fe-P and a decrease in Al-P. In two latosols studied there was more of an increase in the Fe—P fraction in comparison to the Al—P fraction. Eight soils from Ontario, Canada, (six alkaline, one acid, one slightly acid) were equilibrated with p32 and 0.01 M CaCl2 for 120 hours, MacKenzie (63). A comparison of P solubility with the various fractions removed by Chang and Jackson's procedure indicated that upon removal of Al—P the solubility (micrograms P per gram of soil) dropped con— siderably except in the case of an Fe—P, Ca-P dominant soil. Al—P appeared to supply most of the soluble fraction. Spe— cific activities of P32 for Fe-P were one—third those for Al-P, while Ca-P were very low. Chang and Juo (16) studied twenty-six Tiawan soils (latosols, sandstone, shale, slate, mudstone, and alluvial) finding eleven soils dominant in Fe-P, eight soils dominant in Fe—P and Ca—P, and seven soils dominant in Ca—P. Volk and McLean (94) studied four Ohio acid soils pH 4.5—5.3 and concluded that if high P fixing acid soils 24 were treated with P there was a decrease in the availability of native P, and a tendency to recover more than half as Fe—P. In the low P fixing soils added P increased available native P, and there was a tendency to recover more than half as Al—P. Almost all the added P was recovered as Al—P and Fe-P. The forms of P in a Bridgehampton silt loam (Brown Podzolic) after 65 years of superphosphate and rock phos— phate treatments resulted in recovery of most P in the rock phosphate treatment as Ca-P, while the superphosphate treat- ment resulted in a greater increase in the Al—P fraction, Manning and Solomon(66). In a recent review of soil P literature Smith (84) stated that many workers have noted an equal if not greater importance of Fe compared to Al in P fixation. A number of workers have reported the reverse. Many factors such as soil type, type of fertilizer used, method of application and sampling, and time between application and sampling may affect the importance of interpretation of the role of Fe and Al in P fixation. _Modified Fractionation Procedure The fractionation procedure described by Chang and ~Jackson (14) for soil inorganic P has been criticized for -lack of a good separation between P forms. The procedure VVill be described here briefly. One gram of soil is placed i4) 50 ml of l_N NH Cl shaking 30 minutes to remove 4 25 exchangeable Ca and readily soluble P such as Ca(H2PO4)2 and CaHPO4 in the soil. One-half normal NH4F, pH 7.0 is then employed (one hour shaking time) to remove Al—P (Fe—P dis— solves slightly). Chang and Jackson (14) prOposed a second one hour NH4F extraction to determine the amount of Fe-P re— moved by the NH4F. Tenth normal NaOH is then employed with a seventeen hour shaking period to remove Fe-P. Next, Ca—P is removed by shaking one hour in 50 ml of 0.5 N H2804. Re- ductant soluble and occluded P may then be removed if desired. Fife (28,29,30,31) studied the separation of Al-P and Fe—P and suggested use of lug NH4F pH 8.5 for 16 hours, to reduce the amount of Fe—P removed. Chang (11) states that a longer extraction time may completely dissolve Al-P whose solubility decreases above pH 7.0; however, alkaline hydrolysis of Fe-P must also be considered. Dependent on the soil it would be difficult to predict whether an in— crease or decrease of P extracted would result from a longer extraction period at pH 8.5. Chang and Liaw (17), studied representative Tiawan soils and concluded that the pH of l _§'NH F ranging from 7.0 to 8.5, made little difference in 4 dissolution of iron phosphate. Khin and Leeper (43) sug— ‘gested that P may be adsorbed by ferric oxide during the (Extraction period and thus recorded as Fe—P. The enhanced £3Olubility of Al—P in NH F may result in the shift from Al—P 4 t<> Fe—P of some P, a process which occurs slowly in nature. 26 Extending the reaction time may enhance such a shift. Chang and Liaw (17) noted that in neutral NH4F solutions a repre- cipitation of some P released from Al-P occurred on the Fe- oxides. Thus the Fe—P fraction increased at the expense of the Al—P fraction. The dissolution of Fe—P by NH4F has re— verse effect, increasing the Al-P fraction. A correction may not be necessary if these two amounts are about the same. Chang and Jackson (14) noted that some reductant soluble Fe-P which is dissolved in sodium dithionite—citrate, and occuluded Al—P may be dissolved in 0.5N_H2804. Khin and Leeper (43) suggested dissolved amounts of occluded P could be considerable. Glen, et al., cited by Chang (ll) sug— gested the extraction of occluded Fe and Al-P be done after the 0.1 N NaOH, before O.5_N H2804. Questions as to the solubility of Ca(HPO4)°2H20 in NH4F and NaOH have been raised. Igawa, cited by Chang (11), suggested the use of 2.5 percent acetic acid to separate Ca-P'before separation of Al—P and Fe-P. No mention as to the acids affect on the other fractions was made. Saeki (79) used 2_N (NH4)ZSO leaching soil placed on filter paper 4 with four 50 m1 portions. Dicalcium phosphate was added to the soils at 250 and 500 ppm P. Recovery of P added as (CaHPO was 18 percent for the 500 ppm level and 28 percent 4 :for the 250 ppm level when leaching with two 50 ml portions c>f (NH4)2SO solution immediately after adding the material. 4 27 Forty-three percent of the P added as CaHPO4 was recovered when four 50 ml leachings were employed. Additional incre- ments of leaching recovered only small amounts of P. Chang (11) states that, "Since the concentration of P in soil so— lutions is usually only a fraction of a ppm, there should not exist any appreciable amount of the more soluble P such as CaHPO in common soils." 4 Chang (11) proposed a modified procedure for the fractionation of soil inorganic phosphates. He suggested the use of pH 7, 0.5 N NH F shake one hour for paddy soils, 4 and pH 8.2, NH F shake one hour for upland soils for removal 4 of Al—P. Chang also suggested removing reductable soluble Fe-P and removal of occluded Fe-P and Al-P after the 0.1_N NaOH Fe-P fraction, and prior to the 0.5_N H2504 Ca—P fraction. MATERIALS AND METHODS Three Michigan surface soils were selected on the basis of their dominant inorganic phosphate fraction. Iron River was selected because of its' iron phosphate dominance. warsaw was selected as a dominant aluminum phosphate soil, and Wisner as a calcium phosphate dominant calcareous soil. The locations from which the soils were sampled are listed below. Warsaw Loam — Surface — Lee Rhoda, RFD, Schoolcraft, Michigan. One mile south of Schoolcraft on west side of road. SE % swk Sec. 19 T 48 R 11 w. Samples taken in old garden area directly south of house. Iron River Silt.Loam — Surface - Raiko Petroff, RFD 3, Iron River. SE % NE % Sec. 14 R 35W T 42N. Upper peninsula, Michigan. Wisner Silty Clay Loam — Surface — No Zn, No P treatment, tier B, Monitor Sugar Co. plots, Bay City, Michigan, 1965. One—thousand gram portions of air dry soil that had been passed through a .7 mm sieve were placed in a tumbling 28 29 mixer. Treatments of 200 and 500 ppm P as CaHPO4 were mixed with the soils, tumbling for fifteen minutes. Preliminary studies indicated that adequate uniform mixtures were ob- tained in fifteen minutes of mixing. Six polyethelene bags of each soil; two control, and two each containing 200 and 500 ppm P as CaHPO4 respectively, were incubated at eighty— five percent of the one—third atmosphere pressure moisture content. Incubation was at 30 degrees centigrade. The moisture content was maintained throughout the incubation by addition of distilled water to a given weight. Periodic samples were taken for fractionation and phosphate potential determinations. Samples were taken initially at three days, one week, two weeks, one month and two months. Samples of five to ten grams of moist soil were dried in a vacuum oven at 40 degrees centigrade under 30 pounds ‘vacuum. One gram samples of dry soil were fractionated into JNH C1 (water soluble) Al-P, Fe—P, and Ca-P fractions accord— 4 :ing to the procedure of Chang and Jackson (14). Potentials were determined by sampling a weight of IrlOist soil equivalent to fifty grams of oven dry soil. The Schils were shaken four days in 100 ml. of 0.01 M CaCl2 plus 011€3 ml. of chloroform. The chloroform was employed to limit miczrobial activity. The pH of the soil suspension was de- tealnnined with a Beckman (Model G) glass electrode pH meter. The soil suspension was then centrifuged at 2200 rpm for 15 Inllilltes and Ca, P, Al, and Fe concentrations were determined 30 in the clear supernatant liquid. Calcium concentration was determined using a Perkin—Elmer 303 atomic absorption spectrophotometer. Phosphorus concentration was determined by the colorimetric sulfomolybdic, stannous chloride method described by Chang and Jackson (14). Aluminum concentration was determined by the "Aluminon method" as described by Jackson (41). The Orthophenantholine method was employed to determine Fe concentration colorimetrically (41). Total P was determined in all soil samples, after two months incu— bation by sodium carbonate fusion,Jackson p. 175 (41). Calcium, P, and Al activities were calculated using the Debye-Huckel Theory of interionic attraction (67). The szp values used for the various minerals as represented on potential diagrams and on phosphate phase diagrams were taken from Lindsay and Moreno (58), and from Taylor and Gurney (88). The szp values used included strengite 35.0, variscite 30.5 and 22.5, CaHPO '2H 4 2 calcium phosphate 46.91, and fluorapatite 118.4. 0 6.56, CaHPO 6.66, octa— 4 Successive extractions of one—gram samples of Wisner and Warsaw treated with CaHPO and also untreated samples 4 were made with 1_N NH Cl pH 7.0. Each increment envolved 4 shaking the soil with 50 m1. of solution for one—half hour. Phosphorus concentration was determined by the sulfomolybdic acid colorimetric method (14). Following the NH4C1 31 extractions, the samples were fractionated (14) into in- organic Al-P, Fe—P, and Ca-P. The successive extractions of the Wisner soil were done on the two month incubated sample. For the Warsaw soil, 200 ppm P as CaHPO4 was added and im- mediately extracted with successive extractions on the un— incubated samples. RESULTS AND DISCUSSION Fractionation of Inorganic Phosphates The results of fractionation of three incubated soils at two phosphorus levels and a control are presented in tables 1, 2, and 3. The untreated, acid Iron River soil showed essentially no change in the four fractions during two months of incubation. The untreated soil contained ap- proximately 120 ppm P in the Fe-P fraction, 30 ppm P in Al-P fraction and 50 ppm P in Ca—P fraction, with only a trace of P in the NH Cl (water soluble) fraction. The total P con- 4 tent of the soil was 740 ppm P as determined from a Na C03 2 fusion. With the addition of 200 ppm P as CaHPO4 increases from 31 to 155 ppm P in Al-P fraction and 119 to 178 ppm P in Fe-P fraction were noted in comparison of the initial un- treated to 200 ppm P treated soil. No change in the water soluble and Ca-P fractions were noted during the two months incubation. However, a small increase in the Fe-P fraction at the expense of the Al—P fraction was indicated during the two month period. The Ca—P fraction was the only fraction of the iron river soil unaffected at the 500 ppm P level. The initial 32 33 Table 1. Interaction of added P, inorganic P fractions in a Wisner silty clay loam and time. Time (days) p addedl p fraction2 0 3 7 14 28 56 ---------- ppm P ———---—--—————-—— None water soluble 5 6 5 5 5 5 Al-P 70 68 63 87 61 56 Fe—P 36 36 33 41 37 45 Ca—P 170 119 172 157 149 167 200 ppm P water soluble 51 58 66 64 65 65 Al—P 163 165 141 149 130 133 Fe—P 43 41 43 44 47 46 Ca—P 178 200 196 181 182 179 500 ppm P water soluble 133 130 143 158 141 141 Al—P 323 255 294 254 270 278 Fe-P 36 44 48 51 49 37 Ca—P 193 206 198 165 204 220 1P was added as CaHPO4 at the initiation of the experiment. 2 Chang and Jackson phosphorus Fractionation procedure (14)- 3Each value reported is a mean of two replications. 34 Table 2. Interaction of added P, inorganic P fractions in a Warsaw loam and time. Time (days) p addedl p fraction2 0 3 7 14 28 56 ----------- ppm P3———----—————--—- None water soluble 3 4 3 3 2 5 Al-P 239 230 289 247 273 225 Fe-P 138 150 130 135 141 129 Ca—P 38 43 39 35 43 40 200 ppm P water soluble 25 19 14 16 13 13 Al—P 315 275 268 310 322 289 Fe—P 156 169 153 151 164 163 Ca-P 48 48 42 37 45 42 500 ppm P water soluble 115 103 74 66 54 49 Al—P 438 413 351 453 350 407 Fe-P 166 175 168 161 174 173 Ca—P 43 48 52 46 54 52 1P was added as CaHPO at the initiation of the . 4 experiment. 2Chang and Jackson phosphorus Fractionation procedure (14). 3Each value reported is a mean of two replications. 35 Table 3. Interaction of added P, inorganic P fractions in an Iron River silt loam and time. Time (days) p added1 9 fraction2 0 3 7 14 28 56 ——————————— ppm P -—————-———————-- None water soluble 2 3 0 0 1 0 Al-P 31 30 35 31 31 28 Fe—P 119 122 125 117 130 122 Ca—P 53 51 46 51 53 47 200 ppm P water soluble 4 5 3 O 1 3 Al—P 155 138 146 — 139 131 Fe—P 178 200 211 218 225 234 Ca—P 51 54 51 50 50 52 500 ppm P water soluble 15 14 11 1 7 6 Al—P 310 290 328 308 276 293 Fe-P 218 262 271 258 283 317 Ca-P 54 54 49 52 58 53 1 P was added as CaHPO4 at the initiation of the experiment. 2Chang and Jackson phosphorus Fractionation procedure (14). 3Each value reported is a mean of two replications. 36 samples showed an increase from a trace to 15 ppm P in the water soluble fraction, 31 to 310 ppm P in Al-P, and 119 to 219 ppm P in Fe—P compared to the untreated soil. A de— crease in the water soluble and Al—P fractions and a simul- taneous increase in the Fe—P fraction over the incubation period was also noted at the 500 ppm P level. Calculation of the percent recovery of the added P showed 95 percent at the 200 ppm P level and 88 percent at the 500 ppm P level. This is well within experimental error. The acid Warsaw soil is dominant in Al—P. Untreated samples of the Warsaw soil revealed essentially no change in the water soluble, Fe-P, and Ca—P fractions throughout the incubation; however, somewhat variable results were ob— tained for the Al-P fraction. This fraction was unchanged except for the one week and one month sampling dates yield- ing slightly higher ppm P values. Total analysis indicated 905 ppm P in the untreated soil. A slight increase in Fe—P at the 200 ppm and 500 ppm P levels was generally noted in the Warsaw soil. The initial samples increased from 137 to 156 to 166 ppm P as Fe—P from the untreated to the 200 and 500 ppm P levels re— spectively. A very slight increase in Ca-P was noted at both 200 and 500 ppm P levels. Initial samples indicated the water soluble fraction to be 115 ppm P at the 500 ppm P level, 25 ppm at the 200 ppm level and 3 ppm for the untreated soil. After two months of incubation, the untreated Warsaw 37 soil was unchanged, but the P content of the water soluble fractions in the 200 and 500 ppm P levels had decreased by half. The percent recovery of added P showed 44 percent re— covered from the 200 ppm P treated soil, and 54 percent re— covered from the 500 ppm P level. Other workers have re— ported low recovery, with no explanation as to possible reasons (15,94). The Wisner calcareous soil is dominant in Ca-P. The P content in all four fractions of the untreated soil was unchanged at the various sampling dates. The addition of 200 and 500 ppm P caused the Al—P fraction to increase from 70 to 163 to 323 ppm P respectively. Some increases were noted in Fe—P and Ca-P fractions over the complete incu- bation period. But the most significant increases were in the water soluble and Al-P fractions at both 200 and 500 ppm P rates. The total P content indicated 905 ppm P to be present in the Warsaw soil. The percent recovery was 85 per- cent at the 200 ppm P level, and 79 percent at the 500 ppm P level. Upon the addition of CaHPO4 to the acid Iron River soil, initial rapid increases in the Fe-P and the Al-P fractions were noted at both 200 and 500 ppm P levels. An indication of a shift of some P from the Al-P fraction to the Fe-P fraction throughout the incubation period may be explained by a shift from a more soluble compound (Al—P such as variscite) to a less soluble phosphate (an Fe—P like 38 strengite) adhering to the solubility product principle. Chang and Chu (12) reported the greatest fixation of P after three days to be in the Al—P fraction of P enriched, acid, Taiwan soils. However, after one-hundred days, most of the added P was in the Fe—P fraction. This type of shift was not noted in the acid Warsaw after two months. An increase in the Al-P fraction accounted for most of the P added as CaHPO4 to the acid Warsaw soil. Little change in the Ca—P fraction was noted in the three soils. The first of two reasons why this may occur is the fact that Ca-P's detectable in the 0.1_N NaOH extracted fraction (i.e. Hydroxyapatite) are not probable to form under very acid conditions. Secondly, the forms of Ca—P that may appear in this fraction have been noted to form very slowly, possibly taking more than a year. This con— sideration is especially important in interpreting calcareous soil P fractions where Ca—P formation is more probable. The fact that a given inorganic P fraction (i.e. Al-P) indicates an increase or decrease may not necessarily mean that the forms included in the fraction are actually P of the specified type (Al-P's). It is realized that Chang and Jackson's fractionation procedure has not been acclaimed to be precise but rather a good indicator of inorganic P forms. When it was devised, the method was tested on acid soils that had not received recent application of P ferti— lizer. The application to calcareous soils or soils that 39 have received recent applications of P fertilizer may result in misinterpretation. The question as to whether CaHPO4 changes into Al-P and Fe—P as rapidly as indicated is paramount. Igawa, cited by Chang (11), suggested the use of 2.5 percent acetic acid for the removal of CaHPO4 before extract— ing Al-P, Fe—P and Ca-P. No reference was made as to the acetic acids effect on the other forms of inorganic P in the soil. An evaluation of the use of 2.5 percent acetic acid for this purpose indicated that it was unsuccessful in re- moval of 200 ppm P as CaHPO4 when added and immediately ex— tracted from the acid Iron River and Warsaw soils. Succes— sive NH4C1 extractions of CaHPO4 treated soils was used in an attempt to explain the indicated rapid change of CaHPO4 to an Al—P as revealed by the fractionation procedures. The untreated and 500 ppm P treated Wisner samples were analyzed after two months incubation. Five, thirty—minute extractions with 50 m1 of 1.0 N NH Cl pH 7.0, followed by one 50 m1 ex— 4 traction for one hour shaking with 0.5 N NH F, pH 7.0 were 4 carried out. The results (given in Table 4 and Figure 1) for this procedure are compared with the Chang and Jackson procedure of one NH4C1 extraction followed by the NH4F ex- traction. The second increment of NH4C1 extraction of the 500 ppm P treated Wisner removed about twice as much P as did the first. The following increments removed considerably less P until increment five which contained a level about 40 Table 4. Successive 1_N NH4C1, pH 7.0, extractions of Wisner silty clay loam, untreated and treated with 500 ppm P as CaHPO4, incubated two months. NH Cl Extractionl No. 4 p Added ppm 1 2 3 4 5 A1-p2 —————————————————— ppm P -——————-——-—--———- o 11 14 12 10 10 55 o 5 — - - — 56 500 115 209 42 21 14 91 500 140 - — - — 278 1 Extracted with NH Cl for one-half hour shaking period. 4 2Extracted with NH Jackson procedure (14). 3Each'value reported is a mean of two duplicates. 4F as described by Chang and P removed (mg/1000 g soil) 260 220 180 140 100 60 41 Q -_ 500 ppm P as CaHPO4 L. X — No P added Extraction number Figure 1. Phosphorus removed from Wisner silty clay loam by successive extraction with 1 .N NH Cl. P removed vs. extraction number. 42 equivalent to that in each of the five extractions of the un— treated soil. The NH4F extracted Al-P fraction seemed to be unaffected following five extractions with NH4C1, having about the same concentration of P in the Al—P fraction after one extraction as it did after five NH4C1 extractions. The great increase in P in the second increment of NH4C1 indi— cates that one 50 m1 NH Cl extraction may be inadequate to 4 remove all the CaHPO4 present in treated calcareous soils. If only one extraction is made, some CaHPO4 may be reported as Al-P since CaHPO4 is soluble in NH4F. A probable reason for not removing more in the first increment may be the high concentration of Ca in that first extraction, and the re— sultant common ion effect keeping much of the CaHPO4 in the solid state. With much of this Ca ion removed, more CaHPO4 would go into solution during the second increment of ex- traction. The possibility of some of the P in the five increments being from other sources is recognized; however, freshly precipitated Fe and Al phosphates are probably not abundant in this alkaline Wisner soil. Therefore CaHPO4 or some Ca-P more soluble than octacalcium phosphate is probable. Further, successive extractions with NH4C1 did not affect the Al—P fraction in the untreated soil, and amorphous forms of AlPO4 probably would be associated with the NH4F fraction and therefore be unaffected. One gram samples of untreated acid Warsaw SOil were extracted with 50 m1 portions of 1.N NH4C1 pH 7.0 for 43 two, three, five, and seven increments, followed by fraction— ation into Al—P, Fe—P, and Ca-P. Only trace amounts of P were detected in each of the increments, and the following extractions revealed essentially no change in the Al-P, Fe—P, or Ca-P fractions due to a number of NH4C1 extractions (see Table 5). Samples of the Warsaw soil treated with 200 ppm P as CaHPO4 were immediately subjected to increments of NH4C1 extraction followed byfractionation into the remaining forms. One extraction removed about one third of the added P. About 20 ppm P was removed in each of the additional incre— ments. The forms of P indicated by immediate fractionation of the CaHPO4 treated acid Warsaw soil must be considered from the standpoint of the solution phase. Probably much of the CaHPO4 added was rapidly transformed into freshly pre- cipitated Al-P, and/6r fixed by clay minerals. These pro— cesses would be speeded up when shaking in the NH4C1 so— lution. Avnimelech (2) reported that P applied to soils in solution was fixed within hours to about the same degree as that applied as powder and incubated three months in moist soil. Solubility Product Diagrams Solubility data for thermodynamic interpretations of phosphate reactions is presented on two types of diagrams; phosphate phase diagram as described by Lindsay and Moreno (58), and phosphate potential diagrams as described by 44 Table 5. Successive extractions with NH4C1 of treated (200 ppm P as CaHPO4) and untreated Warsaw loam. NH4C1 Extraction2 No. p addedl ppm 1 2 3 4 5 6 7 Total Al—P Fe—P Ca—P ——————————————————————— ppm Ph—-—-——————-——---——-—--------——— 0 3 2 5 202 116 36 0 2 3 4 9 206 116 36 0 3 2 2 T T 7 204 122 44 0 3 3 5 T T T T 11 203 116 39 200 70 70 279 149 74 200 72 19 91 248 135 60 200 72 28 19 118 238 144 49 200 68 33 20 8 10 138 264 133 58 T indicates trace. 1Each value reported is a mean of two duplicates. The CaHOP was added to unincubated soil and immediately ex- tracted W1th NH Cl, and then fractionated by Chang and Jackson procedure (14). 2 Each NH4C1 extraction involved shaking one—half hour. 45 Aslying (1), and also by Wright and Peech (103). The solu— bility products (szp) for the various components of the phase equilibrium systems used in the calculation of solu— bility isotherms were taken from Lindsay and Moreno (58), and from Taylor and Gurney (88). Before interpreting the results obtained on the basis of solubility diagrams, a consideration of the values used to calculate solubility products of minerals is im— portant. Lindsay and Moreno (58) represented the dissoci— ation of variscite in the following way: Al+++ + 20H" + H PO ' ————\ Al(OH)2HzPO4 :— 2 4 pKSp(V) = pAl + 2pOH + pH2PO4 30.5 pKSp(V) More recently Taylor and Gurney describe the relationship as: . +++ = 1411204 21120 4:3 A1 + P04 szp(v) = pAl + pPO4 pAl + pH2PO or szp(v) 4 — 2pH + pK2K3 szp(v) 22.52 Both isotherms have the same slope on a phase diagram or a potential diagram, but differ in intercept, with Taylor and Gurney's relationship being shifted downward from the iso- therm calculated by Lindsay and Moreno. These relationships have been calculated and represented on phase diagrams 46 assuming that the A1+3 activity in solution is controlled by gibbsite with a szp(G) of 33.8. Solubility diagrams are only approximate indicators of controlling solid phases and therefore ion activity in solution. A lack of knowledge about P minerals, their solubilities, dissociation products, and modes of formation in a dynamic soil system renders in— terpretations difficult. For example, an amorphous form of Al(OH)3 that is more soluble than gibbsite may be con— trolling A1+3 activity in the solution phase. Also, con- sider an amorphous AlPO4 more soluble than variscite af— fecting P activity in solution. 1. szp(G) = pAl + 3(14—pH) 2. szp(AG)= pAl + 3(14-pH) where: szp(G) = solubility products of gibbsite szp(AG) = solubility products of amorphous Al(OH)3 and szp(G) > szp(AG) 3. szp(V) = pAl + PO 4 4. szp(AV)= pAl + PO4 where: szp(V) = solubility product for variscite szp(AV) = solubility product for amorphous AlPO and szp(V) > szp(AV) 4 Consider the representation of isotherms on a phase diagram where: pH + pH P0 = pK(v) — pK(G) + pKw 2 4 pKw = szp for water 47 If an amorphous form of Al(OH)3 more soluble than gibbsite is controlling A1+3 activity, the AlPO4 isotherm would shift downward from the isotherm where gibbsite and variscite are controlling the system (Figure 2). Phosphorus activity could be controlled by an amorphous AlPO more soluble than 4 variscite, and if gibbsite controls Al+3 activity the AlPO4 isotherm would shift upward. If both an amorphous AlPO4 and amorphous Al(OH)3 are present, or the presence of any com- bined participation of crystalline and amorphous forms of both variscite, or variscite-like compounds and gibbsite, the resulting isotherms may shift up or down dependent on the relative amounts of each form present. Similar relationships should be considered for a strengite isotherm when assuming that goethite controls Fe+3 activity in solution. Representing an AlPO isotherm on a P and Al po— 4 tential diagram will also be approximate. Consider the following: AlPO4 : Al + P04 szp(V) = pAl + pK —pH + pK -pH + pH PO 3 2 2 4 rewritten: pH2P04+l/3 pAl = (szp(V) - pK2K3)+2(pH—1/3pAl) intercept slope where: pK2 = pH + pHPO4—pH2PO4 = 7.20 pK3 = pH + pPO4-pHPO4 = 12.32 48 PH % Figure 2. Theoretical Al—p isotherms as in- fluenced by increased solubilities of amorphous variscite and/or gibbsite. 0 «57‘: ”F9 ’85 +— pH2PO4 + 1/3 pAl pH— 1/3 pAl ———). Figure 3. Theoretical aluminum vs. phosphate potentials as influenced by in- creased solubility of amorphous variscite. 49 The slope of the variscite, or variscite—like compound iso— therm will be two when P potential is plotted against Al po— tential. The intercept will vary according to the values of szp(V) used. If an amorphous form of variscite is con— trolling the system, experimental points may be above the isotherm calculated on the basis of crystalline variscite. Data for treated and untreated Iron River, Warsaw and Wisner soils at various sampling dates throughout the incubation period are presented on phase diagrams. The un— treated Iron River appeared to indicate that variscite or strengite like compounds were controlling P solubility (Figure 4). Upon incubation the “equilibrium" points moved parallel to the variscite and strengite isotherms, with a resultant decrease in P activity in solution and a decrease in pH. With the addition of 200 ppm P as CaHPO4 to the system slight increases in P activity with the points moving approximately parallel to the variscite and strengite lines occurred. At the 500 ppm P level the points moved up verti- cally with increased P activity in the system. Slight in— crease in pH were noted when the Iron River soil was treated at both P levels. The 500 ppm P level resulted in a greater increase in pH than did the 200 ppm P level. This increase in pH may be due to the dissociation of CaPHO according to 4 the following equations. pH2P04 50 28 (acidified) 56 (acidified) ,- Sample No. Time (days) O'-- No P __ l 7 I: 2 14 ‘_ 200 ppm P as CaHPO4 _ 3 28 .A-— 500 ppm P as CaHOP4 4 56 5 6 Figure 4. Solubility phase diagram, pHZPO vs. pH, for Iron River silt loam. Each point is an average of duplicate samples. 51 CaHpo4 # Ca++ + HPO4= HPO _ 4 PO + OH + H 2 4 20:;252 H The rise in pH may be off set somewhat since microbial activity during the shaking period produces CO At both P 2. treatment levels, the pH of the Iron River soil decreased with no apparent change in P activity. This decrease oc— curred continuously from the initial samples until the end of the two month incubation. When the pH of the system was re- duced by acidification of the two month incubated samples, the P concentration was essentially unchanged for the treated and untreated soil. After the four day shaking period, the system may not be in equilibrium, therefore P activity would not have changed. It has been suggested by Chakravarti and Talibudeen (10) that an eight day shaking period was necessary for equilibrium in acid soils. In the Iron River acid soil it seems that variscite-like and/or strengite-like compounds control the untreated system, with possibly amorphous forms participating in the P treated systems. The phase diagram for the acid Warsaw soil (see Figure 5) indicates compounds more soluble than those in the Iron River soil control P activity in solution. Generally, no change in treated or untreated systems were noted upon incubation. But the two month sampling of the 500 ppm P treatment indicated a lower P concentration than all the l \lmU'I-hOJNH 52 Sample No. Time_(days) O— No P 3 7 14 28 56 28 (acidified) 56 (acidified) U- 200 ppm as CaHPO 4 A“ 500 ppm as CaHPO 4 Figure 5. Solubility phase diagram, pH2P04 vs. pH, for Warsaw loam. Each point is an average of duplicate samples. 53 previous samplings. In the acidified untreated soil, a slight increase in P concentration accompanied the decrease in pH. The acidification of the one month incubated 200 and 500 ppm P soils resulted in little change in P concentration. The 500 ppm P treated two month incubated samples had a more noticeable decrease in P activity accompanying the decreased pH. Amorphous forms of AlPO are likely to be controlling 4 this system. The CaHPO4 probably accounts for the points being far above the variscite isotherm for the treated soils, as generally the most soluble compounds control P activity in solution. A Ca-Al phosphate or A1 phosphate more soluble than variscite might control the system. The calcareous Wisner soil is a dominant Ca—P soil. Equilibrium points for the untreated soil fell into a group half way between the octacalcium phosphate and hydroxyapa— tite isotherms (Figure 6). No consistent changes in the treated or in the untreated soil equilibrium points were noted when 1ncubated for two months. The addition of 200 ppm P as CaHPO caused a vertical shift toward the octa- 4 calcium phosphate isotherm. Essentially no change in pH was accompanied by an increase in P activity. This was also noted at 500 ppm P as the points shifted closer to the octa— calcium phosphate isotherm. Dicalcium phosphate raised the P activity in solution with little change in pH.. The for- mation of significant amounts of apatite is improbable as (3" No P [j-— 200 ppm P as CaHPO A“ 500 ppm P as CaHPO 4 4 3 OF— Sample No. Time (days) 7 14 28 56 28 (acidified) 56 (acidified) mmbwwr—I ‘3» V "’c. \ g/ \\\\ .9 \A A ‘8 Ek\°‘ ‘90 E \\\@ Figure 6. Solubility—phase diagram, pH2PO4 vs. pH for Wisner silty clay loam. .Each point is an average of duplicate samples. 55 this mineral forms slowly. Further, the phase diagram seems to indicate that a Ca—P similar to octacalcium phosphate still remains after two months incubation as no large shift at the various sampling dates is noted. A Ca—P compound more soluble than octacalcium phosphate could be present in the treated soil, and not by revealed or a solubility dia— gram. This is probable as a four day shaking period in dilute CaCl2 may result in its shift to more insoluble com— pounds such as octacalcium phosphate. Acidifying the system at the end of two months incubation caused the untreated points to move parallel to the CaHPO4 line. This is evi- dence of a Ca-P controlled P activity in solution. A plot of lime potential (pH—%pCa) versus phosphate potential (%pCa + pH2P04) for the treated and untreated Wisner soil indicated similar results as compared to the phase diagram (Figure 7). Aluminum potential (pH—1/3 pAl) versus phosphate po- tential (pH2PO4 + 1/3 pAl) are presented on a single solu— bility diagram (Figure 8). Two variscite isotherms are represented as szp(V) = 22.52 (88), and szp(V) = 30.50 (58). Results of analysis of the Warsaw and Iron River soils after two months incubation indicated similar results to the phase diagram. The untreated Iron River soil "equi— librium" points are on the szp(V) = 22.52 variscite line. Points on the szp(V) = 30.5 variscite isotherm occurred with 200 ppm P in the Iron River soil. The 500 ppm P level in O %p Ca + szP04 0 56 0 CPNo P . — U-ZOO ppm P as CaHPO4 "' A 500 ppm P as CaHPO4 I Sample No. Time (days) 1 7 2 14 3 28 4 56 5 28 (acidified) 6 56 (acidified) (Q9 ‘6 Y . =40 '92 a \ “’6 \\ . \\ \ ‘8 \\\\° \\ V\ El \ \ )8 \ \ \ 1E \ \§\\\ \\® ‘98) I l l 1 1 1 1 | pH_%p Ca6.0 7.0 Figure 7. Lime vs. phosphate potential for Wisner silty clay loam. Each point is an average of duplicate samples. 57 <3- No P D— 200 ppm P as CaHPO4 A— 500 ppm P as CaHPO4 Iron River Iron River, acidified - Warsaw Warsaw, acidified {v szp = 30.5 #60er pH2P04 + 1/3 pAl l 2 pH - 1/3 pAl 3 Figure 8. Aluminum vs. phosphate potential for Iron River silt loam and warsaw loam° Each point is an average of duplicate samples. 58 Iron River seemed to be supersaturated with respect to variscite after two months. Upon acidification both treated and untreated points moved parallel to the isotherms indi— cating variscite like compounds probably control the P activity. Again, compounds more soluble than those present in the Iron River soil seemed to control P activity in the Warsaw soil. The untreated soil equilibrium points are located just above the szp(V) = 30.5 isotherm. The 200 and 500 ppm P treated equilibrium points were located vertically above the untreated "equilibrium" points by about one phos— phate potential unit. Upon acidification the treated and un— treated warsaw points moved parallel to the variscite iso- therms indicating variscite—like, possibly amorphous AlPO4 compounds controlling P activity in solution. Inorganic phosphorus fractionation results and the two types of solubility diagrams will be compared. These two procedures measure different phases of soil P. The fractionation procedure measures the amount of P in the solid phases. Solubility product diagrams are used to de- termine the dominant forms of mineral P which control P activity in solution. Fractionation of untreated iron river indicated the dominance of Fe-P, and some Al-P and Ca—P. Solubility dia— grams indicated that quite insoluble compounds like variscite or strengite control P activity in the soil solution. When 59 treated with 200 or 500 ppm P as CaHPO4 an increase in the Al—P fraction is immediately noted. Solubility diagrams dis— play an increased P activity in solution as amorphous AlPO4 is probably formed. After two months incubation the Fe~P fraction increased somewhat at the expense of the Al—P fraction. An indication that CaHPO does not persist in the 4 acid Iron River soil is evidenced by the trace of P in the water soluble fraction. Thus, both methods suggest that Al-P‘s control P activity over a short period of time; how— ever, P fractions suggest that in time the newly formed Al-P may revert to Fe-P. Fractionation of untreated Warsaw soil indicated Al-P dominance. When treated with CaHPO the greatest in- 4 crease was noted in the Al—P fraction. A smaller increase in the water soluble fraction was noted. Solubility dia— grams show increased P activity in the P treated soils, with an amorphous AlPO4 compound more soluble than that present in the Iron River soil, probably controlling P activity. Both methods suggest Al-P's controls P activity in the Warsaw soil. Dicalcium phosphate does not appear to be trans- formed as rapidly in the acid Warsaw as in the Iron River soil. Successive NH Cl extractions of warsaw soil treated 4 with 200 ppm P as CaHPO showed that only about one—third 4 could be removed with one 50 ml extraction, and that only about 70 percent with five extractions. Whether the seventy 60 percent P removed was CaHPO4 or was some other form of P as a result of a transformation from CaHPO4 during the ex- traction period is not known. Generally, only the most soluble P compounds dissolve or precipitate fast enough in soils to govern the P activity in solution (58); thus, the possibility of a shift during extraction is evident. Fractionation of the treated Wisner soil indicated the greatest increases in the Al-P fraction, and a con_ siderable increase in the water soluble fraction. The con— siderable increase in Al—P was indicated to probably be CaHPO not removed by one NH Cl extraction, Since successive 4 4 NH4C1 extraction removed sufficient P that no gain was shown in the Al—P fraction. Solubility diagrams indicate that Ca-P's control P activity in both treated and untreated soil. Fractionation data agrees with this only if care is exerted to remove all water-soluble P prior to fractionation. SUMMARY AND CONCLUSIONS Chang and Jackson's (l4) procedure for fractionating soil inorganic P into water soluble, Al-P, Fe—P, and Ca-P fractions has shown large increases in the Al—P fraction when P was added to both acid and calcareous soils. These results are not expected in calcareous soils. It was thought that considerable CaHPO4 was not extracted with NH4C1 and was therefore reported as Al-P. In an attempt to verify prOper separation into these fractions, two acid soils and one calcareous soil were treated with CaHPO4 and incubated for two months. Fractionation data and phase equilibrium data from samples throughout the incubation period were com— pared. Also, successive NH Cl extractions were carried out 4 to remove CaHPO4 before other fractions were extracted. One extraction of NH Cl was inadequate in CaHPO4 re- 4 moval from the treated calcareous Wisner soil. The second increment of NH4C1 removed considerably more P than did the first, with the third, fourth, and fifth increments removing decreasing amounts of P. Possibly high Ca activity in the first increment resulting in a common ion effect kept much of the CaHPO4 out of solution. One NH4C1 extraction seemed to be adequate when employed on the untreated Wisner soil. The acid Iron River and Warsaw soils both treated and untreated 61 62 seemed to be adequately extracted with one increment of NH Cl, 4 thus little or no carry over of CaHPO into the Al—P fraction 4 would occur. Fractionation of Wisner soil treated with 200 and 500 ppm P as CaHPO indicated that most of the added P was 4 present in the Al-P and water soluble (NH Cl) fractions. 4 Both the P potential diagrams and the P phase solubility diagram showed Ca—P's to be controlling P activity in so- lution. Equilibrium points for untreated soil were located between the octacalcium phosphate and.hydroxyapatite iso- therms. Treated soil points were closer to the octacalcium phosphate isotherm. Upon acidification of the dilute CaCl 2 systems, equilibrium points shifted parallel to the CaHPO4 isotherm. A large increase in the Al-P fraction as reported by fractionation is more likely to be a fairly soluble Ca-P 4, that was not removed with one NH4C1 extraction and was then extracted with NH F and reported as Al-P. Al-P's such as CaHPO 4 are not likely to form in large quantities in this calcareous soil. Fractionation is in agreement with solubility data when care is taken to remove all water-soluble (NH4C1) P prior to fractionation. Thus, Ca—P's are indicated to con- trol P activity in the Wisner soil. Solubility diagrams indicated that in the Iron River soil quite insoluble compounds like variscite or strengite control P activity in solution. Fractionation indicated only traces of P in the water soluble fraction, 8 small Ca-P 63 fraction, and a Fe—P dominance with considerable Al—P fraction in the untreated soil. Upon the addition of CaHPO4 the greatest increase was noted in the Al-P fraction with some increase in the Fe-P fraction. The water soluble and Ca—P fractions were essentially unaffected. After two months incubation the Al-P fraction decreased somewhat with an accompanied increase in the Fe—P fraction. Both methods suggest that Al-P‘s control P activity over a short period of time; however, P fractions suggest that in time the newly formed Al-P may revert to Fe—P. Fractionation of Warsaw untreated soil revealed Al—P dominance. Solubility diagrams indicated that amorphous Al-P probably controls P activity in solution. Treated soils contained most of the added P in the Al—P fraction, with some in the water soluble fraction. The phase diagram indi- cated increased P activity in the treated soil. But upon acidification little change in P activity was noted. Phos— phate potential plotted against Al potential showed movement somewhat parallel to the variscite isotherm when treated and untreated systems were acidified. More soluble compounds were indicated to control P activity in the warsaw soil in comparison to the Iron River soil. Both methods suggest Al-P's control P activity in Warsaw $011. Interpretation of fractionation and solubility pro_ duct data is approximate, and gives only an indication as to 64 quantities of different P solid phases present in soils. These phases control P activity in the soil solution. Much is still unknown about forms and properties of P minerals in soils. Especially complex is the problem of understanding their reactions in a dynamic soil system when P is added to the soil. BIBLIOGRAPHY 10. BIBLIOGRAPHY Aslying, H. C. (1954). The lime and phosphate po- tentials of soils. 1954 Yrb. Royal Vet. and Agric. College, Copenhagen Denmark. Avnimelech, Y., and J. Hagin. (1965). Phosphorus fixation as a transfer controlled phenomenon. Soil Sci. Soc. of Amer. Proc. 29:394-396. Bache, B. W. (1963). Aluminum and iron phosphate studies relating to soils. I Solution and hydro- lysis of variscite and strengite. Jor. Soil Sci. 14:113. Bache, B. W. (1964). Aluminum and iron phosphate studies relating to soils. II Reactions between phosphate and hydrous oxides. Jor. Soil. Sci. 15:110—116. Baker, D. E. (1960). Phosphorus equilibrium and availability in soils. Diss. Abstr. 21:1317-18. Boischott, P., M. COppenet, and J. Hebert. (1950). The fixation of phosphoric acid on calcium carbon- ate in soils. Plant and Soils 2:311—322. Bouldin, D. R., J. R. Lehr, and E. Co Sample. (1960). The effect of associated salts on transformations of MCPM at site of application. Soil Sci. Soc. Amer. Proc. 24:464. Brown, W. E., and J. R. Lehr. (1959). Application of phase rule to behavior of MCP in soils. Soil Sci. Soc. Amer. Proc. 23:7—10. Chai, M. C., and A. C. Caldwell. (1959). Forms of phosphorus and fixation in soils. Soil Sci. Soc. Amer. Proc. 23:458-460. Chakravarti, S. N., and O. Talibudeen. (1962). Phos— phate equilibria in acid soils. Jorn. Soil Sci. 13:231-240. 66 ll. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 67 Chang, S. C. (1961). Modified procedure for fraction- ation of soil phosphorus. Agri. Chem. Bull., National Taiwan University, 1961, pp. 28-31. Chang, S. C., W. K. Chu. (1961). The fate of soluble phosphate applied to soils. Jorn. Soil Sci. 12: 286-293. Chang, S. C., and M. L. Jackson. (1957). Solubility product of iron phosphate. Soil Sci. Soc. Amer. Proc. 21:265-269. Chang, S. C., and M. L. Jackson. (1957). Fraction- ation of soil phosphorus. Soil Sci. 84:133-144. Chang, S. C., and M. L. Jackson. (1958). Soil phos- phorus fractions in some representative soils. Jorn. Soil Sci. 9:109-119. Chang, S. C., and A. S. R. Juo. (1963). Available phosphorus in relation to forms of phosphates in soils. Soil Sci. 95:91-96. Chang, S. C., and F. H. Liaw. (1962). Separation of aluminum phosphate from iron phosphate in soils. Science 136:386. Chu, C. R., W. W. Moschler, and G. Thomas. (1962). Rock phosphate transformations in acid soils. Soil Sci. Soc. Amer. Proc. 26:476—478. Clark, J. S., and M. Peech. (1955). Solubility cri- teria for the existence of calcium and aluminum phosphates in soils. Soil Sci. Soc. Amer. Proc. 19:171-174. Clark, J. S., and R. C. Turner. (1955). Reactions between solid calcium carbonate and orthophosphate solutions. Canad. J. Chem. 33:665-671. Cole, C. V., and M. L. Jackson. (1951). Solubility equilibrium constant of dihydroxy aluminum di- hydrogen phosphate relating to a mechanism of phosphate fixation in soils. Soil Sci. Soc. Amer. Proc. 15:84-89. Coleman, R. (1944). Phosphorus fixation by the coarse and fine clay fractions of kaolinitic and montmorillonitic clays. Soil Sci. 58:71. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 68 Coleman, N. T., J. T. Thorup, and W. A. Jackson. (1960). Phosphate-sorption reactions that involve exchangeable aluminum. Soil Sci. 90:1-7. Dean, L. A. (1949). Fixation of Soil Phosphorus. Advances in Agronomy 1:391-411. Egan, E. P., Jr., Z. T. Wakefield, and B. B. Luff. (1961). Low temperature heat capacity, entropy, and heat of formation of crystalline and colloidal ferric phosphate dihydrate. Jorn. Physical Chem. 65:1265—1270. Eisenberger, S., A. Lehrman, and W. D. Turner. (1940). The basic calcium phosphates and related systems. Some theoretical and practical aspects. Chem. Revs. 26:257—296. Ellis, R., Jr., and E. Truog. (1955). Phosphate fixation by montmorillonite. Soil Sci. Soc. Amer. Proc. 19:441-454. Fife, C. V. (1959). An evaluation of ammonium fluoride as a selective extractant for aluminum— bound soil phosphate: I Preliminary studies on non—soil systems. Soil Sci. 87:13—21. Fife, C. V. (1949). Ibid. II Preliminary studies on soils. Soil Sci. 87:83—88. Fife, C. V. (1962). Ibid. III Detailed studies on selected soils. Soil Sci. 93:113-123. Fife, C. V. (1963). Ibid. IV Detailed studies on soils. Soil Sci. 96:112-120. Fried, M., and R. E. Shapiro. (1960). SoilvPlant re- lations in phosphorus uptake. Soil Sci. 90:67—76. Frink, C. R., and M. Peech. (1962). The solubility of gibbsite in aqueous solutions and soil extracts. Soil Sci. Soc. Amer. Proc. 26:346. Glenn, R. C. (1959). Flow sheet for soil phosphate fractionation. Agronomy abstracts, Nov. 16-20, 1959, Madison, Wisconsin. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 69 Haseman, J. F., J. R. Lehr, and J. P. Smith. (1951). Mineralogical character of some iron and aluminum phosphates containing potassium and ammonium. Soil Sci. Soc. Amer. Proc. 15:76—84. Hemwall, J. B. (1957). The fixation of phosphorus by soils. Advances in Agronomy 9:95-112. Hemwall, J. B. (1957). The role of soil clay minerals in phosphorus fixation. Soil Sci. 83: 101-107. Huffman, E. W., W. E. Cate, M. E. Deming, and K. L. Elmore. (1960). Rates and mechanisms of disso- lution of some iron and aluminum phosphates. 7th Intern. Congress of Soil Science, Madison, Wisconsin, U.S.A. 1960, Vol. II:404—412. Huffman, E. 0., and A. W. Taylor. (1963). The be— havior of water—soluble phosphate in soils. Jorn. Agric. Food Chem. 11:182—187. Hsu, Pa Ho, and D. A. Rennie. (1962). Reactions of phosphate in aluminum systems: II Precipitation of phosphate by exchangeable aluminum on a cation ex— change resin. Canad. Jor. Soil Sci. 42:210-221. Jackson, M. L. Soil Chemical Analysis. (1958). Prentice-Hall, Inc., New Jersey. Kittrick, J. A., and M. L. Jackson. (1955). Rate of phosphate reaction with soil minerals and electron microscope observation of the reaction mechanism. Soil Sci. Soc. Amer. Proc. 19:415—421. Khin, A., and G. W. Leeper. (1960). Modification in Chang and Jackson's procedure for soil phosphorus fractionation. Agrochinica 4:246—254. Lamb, A. B., and A. G. Jacques. (1938). The slow hydrolysis of ferric chloride in dilute solution: II The change in hydrogen ion concentration. Jor. Amer. Chem. Soc. 60:1215-1228. Larsen, S. (1965). The influence of calcium chloride concentration on the determination of lime and phosphate potentials of soil. Jor. Soil Sci. 16: 275-278. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 7O Larsen, S. (1966). The solubility of phosphate in a calcareous soil. Jor. Soil Sci. 17:121-126. Larsen, S., and M. N. Court. (1960). Soil phosphate solubility. Nature London 189:164-165. Larsen, S., and M. N. Court. (1960). The Chemical potentials of phosphate ions in soil solutions. 7th Inter. Congress of Soil Sci. Madison, Wisconsin. Larsen, S., D. Gumay, and J. R. Devine. Stability of granular dicalcium phosphate dihydrate in soils. Nature London 204:114. Larsen, J. E., G. F. Warren, and R. Langston. (1959). Effect of iron, aluminum, and humic acid on phos— phorus fixation by organic soils. Soil Sci. Soc. Amer. Proc. 23:438—440. Larsen, S., and A. E. Widdowson. (1964). Effect of soil to solution ratio on determining the chemical potentials of phosphate ions in soil solutions. Nature 203:942. Laverty, J. C., F. O. McLean. (1961). Factors af— fecting yields and uptake of phosphorus by crOps: III Kinds of phosphates. Soil Sci. 91:166-171. Leaver, J. P., and E. W. Russell. (1957). The re- action between phosphate and phosphate fixing soils. Jor. Soil Sci. 82113—126. Lehr, J. R., and W. E. Brown. (1958). Calcium phos- phate fertilizers: II. A petrographic study of their alterations in soils. Soil Sci. Soc. Amer. Proc. 22:29-32. Lewin, S. (1960). The solubility product principle. Pitman, 1960. Lindsay, W. L. (1959). Solubility criteria for the existence of variscite in soils. Soil Sci. Soc. Amer. Proc. 23:357—360. Lindsay, W. L., M. Peech, and J. S. Clark. (1959). Determination of aluminum ion activity in soil ex— tracts. Soil Sci. Soc. Amer. Proc. 23:266-269. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71 Lindsay, W. L., and E. C. Moreno. (1960). Phosphate phase equilibria in soils. Soil Sci. Soc. Amer. Proc. 24:177—182. Lindsay, W. L., and H. F. Stephenson. (1959). Re— actions of monocalcium phosphate in soils: I. Solutions that react with the soil. Soil Sci. Soc. Amer. Proc. 23:12. Ibid. II. Dissolution and precipitation reactions in- volving iron, aluminum, manganese, and calcium. Soil Sci. Soc. Amer. Proc. 23:18. Ibid. III. Studies with metastable-triple—point— solution. Soil Sci. Soc. Amer. Proc. 23:342-345. Ibid. IV. Repeated reactions with metastable triple- point solution. Soil Sci. Soc. Amer. Proc. 23: 440-445. MacKenzie, A. F. (1962). Inorganic soil phosphorus fractions of some Ontario soils as studied using isotopic exchange and solubility criteria. Canad. Jorn. Soil Sci. 42:150-156. MacKenzie, A. F., and S. A. Amer. (1964). Reactions of iron, aluminum, and calcium phosphates in six Ontario soils. Plant and Soils 21:17—25. Mandal, L. N. (1964). Effect of time, starch, and lime on the transformation of inorganic phosphorus in a water-logged rice soil. Soil Sci. 97:127-132. Manning, P. B., and M. Salomon. (1965). Forms of phosphorus in a soil after long continued fertili- zation. Soil Sci. Soc. Amer. Proc. 29:421-423. Moore, W. J. (1963). Physical Chemistry, 3rd Edition. Prentice—Hall. Moreno, E. C., W. E. Brown, and G. Osborn. (1960). Solubility of dicalcium phosphate dihydrate in aqueous systems. Soil Sci. Soc. Amer. Proc. 24: 94—98. Ibid. Stability of dicalcium phosphate dihydrate in aqueous solutions and solubility of octacalcium phosphate. Soil Sci. Soc. Amer. Proc. 24299-102. Moreno, E. C., W. L. Lindsay, and G. Osborn. (1960). Fractions of dicalcium phosphate dihydrate in soils. Soil Sci. 90:58—68. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 72 Nethsinghe, D. A. (1958). Chemical potential of ex- changeable phosphate in soils. D. Phil. Thesis, Oxford, 1958. Olsen, S. R. (1953). Soil and fertilizer phosphorus in crOp nutrition. Agronomy, a series of mono— graphs Vol. IV. Academic Press. N. Y., 1953. Perkins, A. T., R. D. Dragsdorf, and R. D. Bhangoo. (1957). Reactions between phosphates and kao— lonite decomposition products. Soil Sci. Soc. Amer. Proc. 21:154-157. Pratt, P. F., and D. W. Thorne. (1948). Solubility and Physiological availability of phosphate in sodium and calcium systems. Soil Sci. Soc. Amer. Proc. 13:213-217. Rathje, W. (1960). The dependence of the solubility of phosphorus and potassium in soil on the po— tential hydrogen-ion concentration. Plant and Soil 13:159-165. Raupach, M. (1963). Solubility of simple aluminum compounds expected in soils. III. Aluminum ions in soil solution and aluminum phosphates in soils. Australian Jorn. Soil Res. 1:46. Rennie, D. A., and R. B. McKercher. (1959). Adsorp- tion of phosphorus by four Saskatchewan soils. Can. Journ."Soil Sci. 39:64-75. Russell, E. W. (1961). Soil Conditions and Plant Growth. 9th ed. Longmans, Green and Co. Ltd., London. Saeki, H. (1965). Dicalcium phosphate reactions with soils during 2.N ammonium sulfate extraction. Soil Sci. and Plant Nutrition 11:38—47. Saeki, H., and M. Okamoto. (1960). Fractionation of phosphate enriched soil. Soil and Plant Food 6:96. Salmon, R. C. (1965). Changes in phosphate potential on rewetting air-dry soil. Nature 205:316. Sanchez, S. (1965). Evaluation of certain chemical transformations at soluble fertilizer phosphorus applied to three Michigan soils. Ph.D. Thesis. Mich. State University, 1965. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 73 Schofield, R. K. (1949). Thermodynamic potentials of soil constituents. Annual Rep. Rothamsted Exp. Sta. p. 29. Smith, A. N. (1965). Aluminum and iron phosphates in soils. Jorn. Aust. Inst. Agric. Sci. 31:110— 126. Stelly, M., and W. H. Pierre. (1942). Forms of in— organic phosphorus in the "C" horizon of some Iowa soils. Soil Sci. Soc. Amer. Proc. 7:139-147. Swenson, R. M., C. V. Cole, and D. H. Sieling. (1949). Fixation of phosphate by iron and aluminum and re- placement by organic and inorganic ions. Soil Sci. 67:3—21. Taylor, A. W., and E. L. Gurney. (1962). Phosphate equilibria in an acid soil. Jorn. Soil Sci. 13: 187—197. Taylor, A. W., and E. L. Gurney. (1963). Solubility of variscite. Soil Sci. 98:9-13. Taylor, A. W., and E. L. Gurney. (1965). The effect of lime on the phosphate potential and resin ex- tractable phosphate in five acid soils. Soil Sci. Soc. Amer. Proc. 29:482-483. Taylor, A. W., E. L. Gurney, and E. C. Moreno. (1964). Precipitation of phosphate from calcium phosphate solutions by iron oxide and aluminum hydroxide. Soil Sci. Soc. Amer. Proc. 28:49-52. Thompson, L. M., and C. A. Black. (1947). The effect of temperature on the mineralization of soil organic phosphorus. Soil Sci. Soc. Amer. Proc. 12:323. Toth, S. J. (1937). Anion adsorption by soil col- loids in relation to changes in free iron oxides. Soil Sci. 44:299-314. Thurlow, D. L. (1965). Interaction of rate and method of fertilizer application and soil phosphorus with yield and chemical composition of sugar beets. Ph.D. Thesis. Mich. State University, 1965. Volk, V. V., and E. 0. McLean. (1963). The fate of applied phosphorus in four Ohio soils. Soil Sci. Soc. Amer. Proc. 27:53-58. 95. 96. 97. 98. 99. 100. 101. 102. 103. 74 Wada, K. (1964). Phosphate equilibria in arable soils different in soil type and management (1963). Soil Sci. and Plant Nutrition 10:191—198. Webber, M. D., and G. E. G. Mattingly. (1965). Rep. Rothamsted Exp. Station 67, 1964. Weir, C. C., and R. J. Soper. (1962). Adsorption and exchange studies of phosphorus in some Manitoba Soils. Canad. Jorn. soil Sci. 42:31—42. Weir, C. C., and R. J. SOper. (1963). Solubility studies of phosphorus in some calcareous Manitoba soils. Jorn. Soil Sci. 14:256-261. White, R. E., and P. H. T. Beckett. (1964). Studies on the phosphate potential of soils: I. The measurement of phosphate potential. Plant and SOil 20:1-16. White, R. E. (1964). Studies on the phosphate po— tential of soils: II. Microbial effects. Plant and Soil 20:184-193° Wild, A. (1950). The retention of phosphate by soil, ‘ A review. Jorn. Soil Sci. 1:221—238. Withee, L. V., and R. Ellis, Jr. (1965). Change of phosphate potentials of calcareous soils on adding phosphorus. Soil Sci. Soc. Amer. Proc. 29:511- 513. Wright, B. C., and M. Peech. (1960). Characterizar— ation of phosphate potential reaction products in acid soils by the application of solubility cri- teria. Soil Sci. 90:32-43. Il1l1111l|1|1|1l11Illfllllllllufllfllml 31293 01748 4233