ABSTRACT EQUILIBRIUM CONCENTRATIONS OF INORGANIC PHOSPHORUS AND ADSORBEO PHOSPHORUS IN SOILS By Daulat Singh Investigations were undertaken on 29 Michigan soil profiles to determine (I) equilibrium conc. of inorganic P, (2) potential buffering capacity, and (3) P adsorption maximum as predicted by Langmuir adsorption isotherm. Above determinations were made with increasing conc of Ca(H2POu)2H20 in 0.0l M CaClz. Equilibration time of 2 hnL with a soil:solution ratio of lle was adopted. The podzoIic sandy profiles and B horizons in particular had exceedingly low conc of inorganic P at equilibriumuconc as low as 8.3 x lOaMP for Emmett loamy sand Bihr was obtained. Only few A horizons, McBride sandy loam, Miami loam, Miami sandy loam, Spinks loamy sand, Brookston clay loam and Muck had equilibrium P conc adequate enough to support an optimum plant growth, if tiose conc were maintained during the plant's life. These higher P conc were probably either a consequence of recent treatment with P fertilizers and/or low bonding energy of P on soil complexes. Daulat Singh The free energy of adsorption of P on soil adsorbing complexes estimated by slopelintercept from the Langmuir adsorption equation indicated higher values For B horizons with possibly Al and Fe adsorption complexes than those with A horizons suggesting lower energy of adsorption of P with organic matter. The potential buffering capacity and P adsorption maximum were highly correlated (r = 0.954). The higher correlation between potential buffering capacity and adsorption maximum alone, without the energy of adsorption term, i.e. K, was because of intencorrelation between adsorption maximum and K. Data strongly indicated that the Langmuir adsorption equation may be used to predict P conc in soil suspension or given a desired P conc in soil suspension for Optimum plant growth, the P needed to be added to soil may be calculated. Fine textured soils (high adsorption maximum and/or K) with no recent P additions, showed high P requirement to bring the P conc to any level compared to coarse textured soils. Data also suggested that P adsorption maximum, 3 unique propery of soil, should serve as indicator of soiIs ability to continue supply P to plants. EQUILIBRIUM CONCENTRATIONS OF INORGANIC PHOSPHORUS AND ADSORBED PHOSPHORUS IN SOILS By Daulat Singh A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Soil Science I969 " :2 L = / 4 Lie :3. [O (2 To My Wife Prabha ii ACKNOWLEDGMENT The author wishes to express his sincere appreciation and gratitude to: Dr. B. G. Ellis, Professor of Soil Chemistry Department of Soil Science for his expert advice and constructive criticism of the investigation. Drs.R. L. Cook, A. E. Erickson, M. Mortland, Department of Soil Science, Dr. K. Lawton, Director of Institute of International Agriculture and Dr. H. Eick, Department of Chemistry for acting as the members of the authors' guidance committee, and Michigan Water Resources Commission for financial assistance. TABLE OF CONTENTS INTRODUCTION . . . . . . . . . . . REVIEW OF LITERATURE . . . . . . . . . . Schofield Phosphate Potential . Phosphorus Sorption The Langmuir Adsorption Isotherm. MATERIALS AND METHODS. . . . Materials . . . . . . . . . . . Analytical Methods and Procedures RESULTS AND DISCUSSION Equilibrium PotentialSP and Equilibrium Conc of Inorganic P. P Adsorption Maximum. Potential Buffering Capacity (PBC). SUMMARY AND CONCLUSIONS. LITERATURE CITED . . . . . . . . . . Page I2 16 22 22 22 33 33 4l 62 76 BI LIST OF TABLES Table Page I Profile no., horizons, soil types and locations of soils. . . . . . . . . . . . . . . . 23 2 The potentialgg and TPeq in 29 Michigan soil profiles. . . . . . . . . . . . . . . . . . . . . 35 3 P adsorption maximum and K, the free energy of adsorption for 29 Michigan soil profiles . . . #2 A Extractable AI and free iron oxide content of the 29 Michigan soil profiles . . . . . . . . . . 48 5 Correlation coefficients between P adsorption maximum and Fe and Al in soils. . . . . . . . . . 53 6 Kg/ha 0-22 86cm of P needed to give l.OxlO'5MP in soil solution. . . . . . . . . . . . . . . . . 58 7 PBC of 29 Michigan soil profiles. . . . . . . . . 63 Fig. LIST OF FIGURES Relationship between potentialSP and 4/3 P/g SOII . . Langmuir adsorption plot for P adsorption data . . Relationship between percent saturation of adsorption maximum and equilibrium P conc in soil solution. Regression of adsorption maximum on PBC. vi Page 3h 1+6 56 68 LIST OF SYMBOLS AND ABBREVIATIONS p = mu = theta g = gram (5) mg = mg (5) Kg = killogram (5) ha hectare (s) no. = number (s) conc = concentration (5) M = mole (s) meq a milliequivalent pX = -log of activity of X potentialCP a chemical potential potentialSp a Schofield phosphate potential potentialzg a equilibrium phosphate potentil TPeq = total phosphorus at equilibrium vii INTRODUCTION As a limiting factor in world food production, P ranks next but to N (Black, I968) and perhaps to none in some parts. Conse- quently, much effort has been devoted to the chemistry of P in soil as related to plant growth. But the chemistry of P in itself and as it relates to plant growth is still not completely under- stood. Inorganic P in soil solution is the immediate source of P for plants. This quantity, however, is so small -- from below IO"7 M to about IO.6 M for P deficient soils and above IO'5 M in soils known to be well supplied (Russell, l96l) -- that a contin- uous replenishment of the soil solution is necessary to insure Optimum plant growth. As a measure of the inorganic P conc in soil suspension, or rather strictly the ease with which P may be removed, Schofield (I955) has introduced the concept of phosphate potential, l/2pCa+pH2POQ of monocalcium phosphate [Ca(H2POu)2H2D] or potentialSp (White and Beckett, I964). He further suggested that the soifs ability to maintain the appropriate chemical potential, that is replenishment of the soil solution against P withdrawal, was critical in a soiVs ability to supply P to plants. Beckett and White (l96h) have pr0posed that potential buffering capacity (PBC) of a soil reflects the soiVs ability to maintain the potentialSp and have devised methods to obtain PBC. Recent investigations (Murrmann and Peech, l969a,Ozanne and Shaw, I967, I968, and Rennie and Mckercher,l959) indicate that adsorbed P may be equally important in controlling the level of P in soil suspension and hence have an important influence on plant available P; notwithstanding the emphasis on the crystalline phOSphates as the soil-fertilizer reaction product of incorporated P in soils (Lindsay, Frazier and Stephenson, I962; Murrmann and Peech, I968; and Wright and Peech, I960). When P is placed in soil, most of it is traceable to Al-P and Fe-P (Bromfield, I967; Coleman, Thorup, and Jackson, I960; Dunbar and Baker, I965; Harter, I968; Hsu, I964; Huffman and Taylor, I963; and Mackenzie and Amer, l96h). AI-P fraction according to Chang and Jackson procedure (I957) contains P much of which is exchangeable with 32P (Dunbar and Baker, I965) indicating that the Al-P fraction must also include adsorbed or labile P in the soil. Recently Harter (I969) had concluded that P is adsorbed on organic matter through some type of anion exchange and significant linear correlation coefficients between Al and free iron oxides and P adsorbed were primarily due to inter-correlation between the independent variables. The present investigation, therefore, was undertaken to determine: l. Equilibrium phosphate potentialSp and hence equilibrium conc of inorganic P in soil solution of some Michigan soils. 2. Potential buffering capacity (PBC), and 3. The P adsorption maximum of the above soils and the interrelationships between the above parameters. REVIEW OF LITERATURE Schofield Phosphate Potential Schofield (I9A9) introduced the concept of chemical poten- tial of soil constituents and later (I955) suggested that - ”it seems reasonable to suppose that availability of soil phosphate is mainly determined by the appropriate chemical potential and by its rate of decrease with phosphate withdrawal.” He also suggested that the appropriate potential was the chemical potential of Ca(H2P04)2H20. Several workers have examined this proposition -- some treating theoretical aspects and techniques of measurement (Larsen and Court, I960; Barrow, I965; and I966) and others testing Schofield's hypothesis by measuring its ability to predict uptake of P by plants (Barrow, I967; Kudeyarova, I968; Mattingly, Russell, and Jephcott, I963; Moser, Sutherland, and Black, I959; Ramamoorthy and Subramanian, I960; White and Haydock, I967). The theory below is essentially as presented by Larsen and Court (I960). In any multiphase system at equilibrium the chemical potential or partial molar free energies of all diffusible chemical components are equal. The chemical potential of a component X of a chemical system is defined by the equation: uX = pOX + RT(InaX) Where uOX is the standard chemical potential referred to an arbitrary reference state, R is the universal gas constant, T is the absolute tempefiature, and aX is the activity of the compound X. The above equation can be rearranged to give: Ox - x 2. 0 = -log aX By the usual convention, -log X is written as pX: Ox - x _ 2.303 RT ’ pX In general, uOX is unknown and since the standard state is arbitrary,it is customary to define a modified chemical potential (uX) by the equation: Using this and the relationship pH + pOH = pKw for water, the potentials of a salt, an acid, and a base can be defined by: u salt = I/Z+ pM + l/Z' pA u acid = pH + l/Z' pA p base = pH - I/Z+ pM Where 2+ is the charge on the cation M and Z' is the charge on the anion A. For a system consisting of a solution phase in contact with a solid phase (e.g. soil system), the chemical potential at equilibrium of any component in solid phase is equal to that of the component in the solution phase and since the latter potential is easily determined from activity measurements, the potential of the solid phase can be found. Schofield's Ratio law (l9h7) requires that for a given soil with a given component of Ca and P ions the activity product (aCaé x aHzPOu) in the soil solution be independent of electrical potential differences and hence of other ions in the soil solution. Aslying (I950) and Nethsinghe (I958) have confirmed this postulate for a number of soils. The validity of ratio law depends on a number of assumptions which Schofield (I9h7) and Nethsinghe (I958) have shown do not always apply. Apart from this, RTln (aCa% x aHZPOQ) is a single valued pro- perty of the soil over the range of Ca and P conc normally found in soil solution, whereas RTln aHZPou is not. RTln (aCa%-x aHZPOQ) is a measure of the sum of the chemical potential of both Ca and P ions in whatever part of the soil controls their potential in the soil as a whole. The phosphate potentialsP was defined as the activity product aCa% x aHzPOu or rather strictly its logarithmic equivalent é.pCa + szPOu. Nevertheless, for soils of comparable Ca status RTIn- (aCa%.x aHzPOh) may also be used as a measure of the potentialSp of labile P in the solid phase of the soil. The use of phosphate potentialSP as a measure of P avail- ability depends on certain assumptions. The first is that the soils of which the potentialSp is measured confirm to the ratio law. The second assumption in the use of phosphate potentialSp to compare the availability of soil P to plants is that the potential of Ca in the soil solids is much less variable than the phosphate potential. The third assumption is that the act of measurement does not substantially alter the potentialCp of P and Ca in the solid phase, nor the potentialCR of any complementary ions such as Fe3+, Al3+ or OH' which may also influence the value of the activity product. The method of interpolation used in this work measures the potentialSp without altering the amount of P held by the soil. If the potentialSp also depends on the potentialsCR of other ions, it may be necessary to employ additional interpolations to obtain the equilibrium activities in solution of these other ions. The first experimental studies were made by Aslying (I950, I95h, and I964). After shaking samples for I-2 hours in both 0.0lM and 0.00lM CaCI2 solution, he found that the activity product was constant for a given ratio of soil to solution and independent of actual magnitude of individual activities. Cole and Olsen (I959) obtained similar results in calcareous soils where the partial pressure of C02 was controlled experimentally. Wild (I959) confirmed the dependence on the soil:solution ratio and also reported a variation with the time of shaking of the soil suspensions. Nethsinghe (I958), and Larsen and Court (I960) had also noted the dependence on the soil:solution ratio. As the conc of soil suspension increases these authors found a increase in the potentialép. It must, however, be noted that although the P conc decreases, the amount of P extracted per unit weight of soil increaseswith decreasing soil:solution ratio. Several explanations, not all independent, are possible. Larsen and Court (I960) suggested the following explanations assuming the presence of a one component P system. Similar postulates could be made for a multicomponent system. (I) (2) (3) (1+) (5) (6) Energy effects '7 as P is removed, the mean bonding energy of P that remains in soil may be higher than the initial mean bonding energy. Adsorption effects -- the equilibrium P conc may be dependent on the mechanism of adsorption and desorption which may be influenced by soil:solution ratio. Incongruent dissolution of P mineral -- mutually disagreeing solubility forms may come in the dissolution process and cause change in P conc with soil:solution ratios. The presence of a mineral containing a definite solubility product but for which the conc of one or more ionic compounds other than P also varies with soil:solution ratio so as to compen- sate for P changes. P of indefinite forms (adsorption complex) for which a constant solubility product would not be expected. Differences due to the uncertain form associated with soil organic matter -- Organic P compounds are not hydrolyzed in the analytical procedure for P and the solubilizing effect of organic matter will be negligible at the electrolyte conc of 0.0lM CaCI2 used. (7) Variation of cation balance with soil:solution ratio -- large changes however are not expected and small change will not affect the chemical potential since the compensating action of ionic strength means that at 0.0lM a l0% change of conc brings about only a I% change of potentiaIEP. Attention need be paid only to the first five explanations. Definite mechanisms are implied by (3), (h), and (5); whereas, explanation (2) is a quantitative description of relationships of postulate (5), and (I) is a quantitative thermodynamic description of all the others without regard to mechanism. These authors measured the P conc in solution after shaking a soil with a solution containing no P initially. The phosphate potentialSP determined this way depends on the amount of P released from the soil and thus on the soil:solution ratio. Ramamoorthy and Subramanian (I960) have rightly drawn attention to the discrepancies between potentialSp measured by a "null'l method (the equilibrium phosphate potentialsP) and the potentialSp obtained by equilibrating a soil with a solution containing no P initially. The “soil:solution ratio effect” is interpreted by White and Beckett (I96h) in terms of P desequilibrium -- a condition in soils in situ which can be due to either nonuniform removal of P or nonuniform distribution of soluble P recently added to soil. White and Beckett (I964) also found that upon allowing soil P to achieve equilibrium by prolonged storage under constant environmental conditions, the equilibrium potentialSp and sIOpe dQ/dl (the potential buffering capacity) remained constant and independent of soil:solution ratio. White and Beckett (I964) also noted change in phosphate potentialSp with drying, changes in temperature and aeration conditions. But with the exception of anaerobic effects, changes induced in the equilibrium potentialSp were considerably smaller than the range of potentialSp of the field soils examined. Nethsinghe (I958) had earlier reported that air drying of soils caused a change in chemical potentials of soil P which was not reversable upon rewetting. Following Schoefield (I949), Aslying (I954) used the chemical potential of Ca(0H)2, the “lime potential”, defined as (pH - l/2 pCa) and the phosphate potentialSP in an attempt to assess the presence and nature of Ca-P compounds in calcareous soils. Similar basic solubility relationships were used in Lindsay and Moreno's phase diagrams (I960). For the general case, Clark and Peech (I955) represented the formation of a hypothetical Ca P as: mCa(0H)2 + nCa(H2POL,)2‘:',Solid + ZHZO IO Where wiand n represent reacting moles of Ca(OH)2 and Ca(H2P04)2 forming one mole of solid + ZHZO. If the reactants are completely dissociated, the linear plot of I/2pCa + pHZPOh (potentiaISp) versus pH - I/2pCa (lime potential) should have a slope of m/n characteristic of the solid species, i.e. octacalcium phosphate Cah H(P0u)3 would have a slope of 5/3 = I.67. Similar solubility diagrams may be constructed for Al(0H)3 - AlPOh, and Fe(OH)3-FeP0u systems. Taylor and Gurney (I962) equilibrated an acid soil in dilute CaCl2 and plotted pH + pH2P04 as a function of pH - l/3 pAI. Points for the untreated soil fell on the variscite line. Since there are several reactions that may reduce the P conc in soil to a level whereby the points could fall on the variscite line, Taylor and Gurney (I962) concluded the finding of Al-P ion products similar to the solubility product of variscite is not a very satisfactory criterion for the existence of variscite in soil. To demonstrate that P status of a soil is governed by the precipitation or dissolution of a particular mineral the appropriate ion product must be proved independent of the base status and salt content of the soil. Chakravartiand Talibudeen (I962) examined P equilbria in 54 British and Indian soils by plotting Fe potentials (pH - l/3 Fe+3) and AI potentials (pH - I/3 AIBT) as a function of potential [1/3 p(Ai+3 or Fe+3)+ pHZPOuj. Equilibrium II conc were interpreted as related to soils grouped according to pH. A variscite type compound controlled P conc in temperate climate up to pH 4.7; above pH 4.7, nonstoichiometric P -- 0H adsorption complexes may be controlling. In the tropical soils (pH 4.3 - 5.8) an Al-P similar to variscite but more basic in character than in temperate climate soils coexists with hydroxides. Temperate soils seemed free from strengite whereas in trOpicaI soils it coexists with hydrated oxide over the entire pH range examined. Wada (I964) reported that P equilibria in some Ando soils, red-yellow podzolic soils and alluvial soils in Japan were governed by Fe and AI compounds from pH 4 to 7. Below pH 5.2 P activity was apparently controlled by variscite and gibbsite or possibly by strengite and Fe(0H)3. Stelly and Pierre (I943) compared the solubility versus pH curves of soils with those of known P minerals and found that a solubility pH curve similar to that of apatite was common with alkaline soils; whereas, acid soils usually displayed a solubility curve similar to that of Fe-P and/or AI-P. Similar results are also reported by Clark and Peech (I955); Lindsay, Peech, and\CIark (I959); Lindsay and Moreno (I960); Rinkenberger (I966); Weir and Soper (I962); Withee and Ellis (I965); and Ulrich and Khanna (I968). Wild (I964), on the basis of solubility product principle, calculated the P conc expected in the presence of variscite and various Ca-P compounds at different pH values and compared his I2 results with the P conc in several soil extracts determined by other workers. In the pH range 4.5 to 6.0 none of the phosphates considered could explain the level of P found in solution. Larsen and Court (I960) obtained scattered distribution of the points on the solubility diagram from solubility data of a great number of British soils. Although some of the points fell near the lines for pure compounds, the overall distribution suggested that either no definite forms of Ca-P compounds can be inferred to exist in soils or that the apparent solubility relationships are influenced by other factors. Solubility product concept and the chemical potential of P in soil suspensions, as evident from the above discussion, has been investigated by many workers for determining the nature of P compounds in soils. However, due to the complex nature of soil and the limitations of the solubility product principle itself in describing precipitation and dissolution phenomena, the results obtained may show the possibility of the existence of certain P compounds in soils, but do not show expli- 'citely\~hich form of the P will control the conc of P in the soil solution-solid system. Phosphorus Sorption Soils usually contain very low amounts of P, 0.04 to 0.Il%, among which the phosphates of Ca, Fe and AI are believed to be the predominate forms (Hemwall, I957). Ca phosphates are I3 abundant in alkaline or calcareous soils and AI and Fe phosphates are preponderant in acid soils (Hemwell, I957; Dahnke and Malcolm, I964; Kurtz, I953; Olsen, I953; and Smith, I965). The overall P problem is threefold: (I) a small total amount present in soils, (2) the unavailability of such native forms, and (3) a marked fixation of added soluble P. Since crop removal of P is relatively low and world P supplies are huge,‘ problem (I) of supplying total P is not serious. Increasing the availability of native soil P and retardation of fixation or reversion of added P are therefore of greatest importance. The literature on factors governing P fixation and avail- ability, (2) and (3), is enormous. Studies prior to I953 on compounds formed during soil genesis and as a result of P fertilization in both acid and calcareous soils were reviewed by Kurtz (I953) and Olsen (I953). Smith (I965) has presented a recent review of Al and Fe phosphates in soils. The nature of soil P after its sorption, even though the literature is flooded from contributions from all over the world, has been the subject of much controversy. Most of the controversy revolves around the question of whether soil phosphates are present as precipitates or as adsorbed anions. Kurtz, Deturk, and Bray (I946), Fried and Dean (I955), Baker (I960) and Murrnann and Peech (I969a)have shown that soil P acts like an adsorbed ion. Cole and Jackson (I950, l95l), Bradly l4 and Sieling (I953), and Metzger (I940) have concluded that P sorption is primarily a precipitation reaction. Kittrick and Jackson (I955) have even observed the crystalline precipitate under an electron microscope. Beaton, Charlton, and Speer (I963) have found dicalciumvphosphate dihydrate (CaHPOh; 2H20) and hydroxy apatite (caI0(OH)2(PO4)6) to be the soil fertilizer reaction product in a calcareous soil. Lindsay, et al. (I962) have identified about 30 crystalline P compounds as soil fertilizer reaction products. Several others have related soil phosphate potentialspto the potential of specific Ca-P compounds and solubility product principle to the presence of discrete P compounds in soils. Fried and Shapiro (I960) suggest that two approaches to P sorption, mineralogical precipitation and adsorption, are not necessarily incompatible. Since soil is a dynamic system, the possibility of both processes occuring is plausible (Smith, I965). Bache (I964) using pure clay and oxide mineral systems had suggested that P sorption takes place in three stages of reaction: (I) A high energy chemisorption of small amounts of P; (2) Precipitation of a separate P phase, and (3) A low energy sorption of P onto precipitates. Boischot, Coppenet and .Hebert (I950) and Olsen (I953) had reported that the mechanism of P sorption in calcareous soils was dependent on the amount of P in solution. The P was l5 adsorbed onto the CaCOB until a critical higher P conc was reached in solution, then precipitation occurred and equilibrium conc of P dropped below that found before precipitation began. The primary objective'of defining the nature of soil-P system is basically two fold: (I) creation of knowledge through basic research, and (2) prediction of the pattern of P supply to growing plants. Considerable effort has been devoted to the second objective. Total soil P failed to corre- late with plant uptake of P. Extraction of soils with various chemical mixtures designed to remove all or a portion of P postulated to be absorbed by plants have been tried. Although the theoretical basis for this approach is very doubtful, as a purely empirical procedure, it has proved useful among soils of the similar characteristics. Methods have been proposed to measure the P adsorbing capacity of acid soils (Bass and Sieling, I950; and Dean and Rubins, I947) and evidence presented to show how such measure- ments aid in the problem of estimating soil P availability. Recent studies have suggested that the adsorbed soil P (Fried and Shapiro, I956; Ozanne and Shaw, I967; and Rennie and Mckercher, I959) and percentage P saturation of the adsorption maximum is closely related with the equilibrium P conc in soil and hence have an important influence on plant available P. l6 The Langmuir adsorption isotherm has a sound theoretical de- rivation to obtain an adsorption maximum and is discussed in the following section. The Langmuir Adsorption Isotherm A model for the adsorption process and particularly for the chemisorption process was presented by Langmuir in I9l8 and led him to a simple but important theoretical derivation of an adsorption isotherm. The adsorption isotherm for an ideal case which is pertinent to the present study is discussed here. For such a case Langmuir made the following assumptions: (I) A unimolecular layer of adsorption-- only one layer of adsorbate adsorbed on the surface of the adsorbent. (2) Uniformity in adsorption sites -- each adsorbate has an equal probability of access, i.e. adsorption to each site and also the heat of adsorption is the same for all sites and does not depend on the fraction covered 9. (3) The solid surface contains a fixed number of adsorption sites. (4) Non-interacting adsorption -- occupancy of one site exerts no influence upon those adjacent (no adsorbate interaction). I7 The Langmuir theory suggests that the rate of evaporation is proportional to the fraction of the surface covered and can be written, therefore, as K19, where K] is some propor- tionality constant. This simple proportionality is an assumption that ignores the complications that often make the heat of adsorption dependent on the extent of coverage. The rate of condensation furthermore is taken to be prOportional both to the gas pressure, which according to the kinetic molecular theory determines the number of molecular collisions per unit area per unit time, and to the fraction of the surface not already covered by adsorbed molecules, i.e. to l-G. It is assumed that only collisions with this exposed surface can lead to the adsorption of a molecule to the surface. At equilibrium then, the rate of evaportation must equal the rate of condensation, i.e.: where K2 is another proportionality constant. Rearrangement gives KZP 9 = — (2) KI+K2P = KP (3) I+KP where K is the new constant K2, sometimes called adsorption KT coefficient. l8 Inspection of equation (3) shows that a chemisorption type isotherm is obtained from this theory. At small values of P, where KP in the denominator can be neglected compared with I, equation (3) reduces to a simple proportionality between 9 and P and this behavior is that corresponding to the initial steep rise of the chemisorption curve. For sufficiently large values of P, 0 approaches the constant maximum value of unity. For adsorption up to a monolayer, the amount of gas X/m, adsorbed at some pfessure P and the amount of gas b needed to form a monolayer are related to 9 according to: lém=e (4) and equation (3) becomes X/m = KbP (5) Em A more convenient form of the above equation is obtained by the arrangement P_I P 6 Ym'%+5 () A plot of P versus P will, if the experimental data are X m in accord with the Langmuir theory, yield a straight line with the intercept RE and the slope with the constant I . The b I9 adsorption maximum b, the amount of gas needed to form a mono- layer is simply the reciprocal of the slope of the regression line. The success of equation (6) in fitting experimental chemisorption type curves must not, of course, be taken as necessarily confirming the model and assumptions that have been used in the derivation. The same equation often applies to the adsorption of a solute from a solution onto a solid adsorbent (Graham I953), although the process is even more difficult to treat theoretically and hence the same rigorous theoretical basis is not as fully developed. When Langmuir adsorption isotherm is applied to liquids or ions, P (pressure of the gas) in equation (6) is simply replaced by C; conc of adsorbate in solution. Considering a function which represents adsorption equili- brium, Langmuir adsorption equation may also be derived for an ideal case as below: For the adsorption process: Free adsorbate molecules + vacant adsorptive sites = occupied sites; adsorption complex. The equation for the equilibrium constant is written - k = (activity of occupied sites) (activity of vacant sitesT (activity of free adsorbate molecules) 20 It is assumed that the activity coefficients of the occupied and unoccupied sites are the same and the equation becomes: K" II 9 _9 C (7) It is interesting to note that the equation for the equilibrium constant can be rewritten: Q = KC (8) I+KC which is the familiar Langmuir adsorption isotherm with kinetic terms replaced by constants readily determined from equilibrium data. The standard free energy of adsorption may be calculated for an ideal system from the equilibrium constant - - 13 F0 = R T an This quantity represents the decrease in free energy of the system and constitutes a measure of the strength of the adsorption bond. If K is known from equation (6), equation (7) may be used to predict the conc of adsorbate molecules in solution to any percentage saturation of adsorption maximum. P adsorption data have commonly been described by Freundlich Isotherm (Boischat, et al, I950; Russell and Low, I954; and Kurtz, Deturk and Bray, I946) an empirical equation of the form - 2I X/m = KCI/n where n is a constant greater than unity. This equation is not specific and generally applies to wide range of equilibrium P conc (not likely to be encountered in normal fertilizer applications of P) and cannot be used to calculate the adsorption maximum. Olsen and Watanabe (I957) and Friend and Shapiro (I956) have shown that constants calculated from Langmuir isotherm permit a sound theoretical approach to some of the problems of P sorption in soils. In the present investigation therefore, Langmuir adsorption equation was used to calculate the P adsorption maximum. MATERIALS AND METHODS Materials Soil profiles no. l-9 from the podzol region in the northern one-half of the lower peninsula of Michigan were collected by Drs. A. E. Erickson and B. G. Ellis in the summer of I967. Sixteen additional profile samples were obtained in the Fall of I967. Four additional soil profiles no. IO-l3, collected in I96l, were added to the above list. These samples were air dried, ground, mixed thoroughly, passed through a 2mm sieve, and stored for soil P to achieve equilibrium (White and Beckett, I964). The profiles and their locations are given in Table l. Analytical Methods And Procedures Time Adsorption Curves A preliminary study undertaken to determine the time of equilibration or adsorption indicated a soil type and time interaction altering the rate of initial adsorption reaction and second or precipitation phenomenon. Data indicated however, that a 2-hour shaking period would permit completion of the adsorption reaction and eliminate complicating secondary reactions. White and Beckett (I964) also had reported that 22 23 Table l. Profile rm» horizons, soil types and locations of soils Soil Profile no.‘ Horizon Soil Type Location I C Dune Sand Silver Lake (Oceana Co.) 2 A] Rubicon Sand Silver Lake (Oceana Co.) A2 .. " .. .. B u u n u C n n u u 3 A0 5 A] Rubicon Sand Lake of Woods (Antrim Co.) A2 " .. .. " B n n u u C n u u u 4 A Rubicon Sand Otsego Lake (Otsego Co.) B H II II n C n u u n 5 A2 Emmett Loamy West Branch of Sturgeon Sand (Otsego Co.) AB .. .. " .. Bi r " " " " 82 " " " " C u n u u 6 A] Emmett Loamy West Branch of Sturgeon Sand (Otsego Co.) A2 " " " " AB .. .. " .. Bihr ‘ " " " " Table I, cont'd 24 Soil Profile no. Horizon SOil Type Location 7 A] 8 A2 Grayling Sand East Branch of Au Sable (Crawford Co.) 8 II II II II Bihr II II II II C II II II II 8 A] 8 A2 Roseland Sand Au Sable River (Crawford Co.) 8 II II II II C II II II II 9 A] 8 A2 Rubicon Sand Au Sable River (Crawford Co.) below Grayling Bihr II II II II C II II II II C2 II II II II ID A] Kalamazoo Loam Richland (Kalamazoo Co.) 81 II II II II 82] II II II II 822 II II II II B3-C] II II II II D II II II II II A] Warsaw Loam Schoolcraft (Kalamazoo Co. 8] II II II II 82‘ II II II II 25 Table l cont'd Soil Profile no. Horizon Soil Type Location II 823 Warsaw Loam Schoolcraft (Kalamazoo Co.) D II II II II l2 A] Ontonagon Clay Ewen (Ontonagon Co.) A2 II II II II AB II II II II 82] II II II II 822 II II II II c II II II II l3 Ap Munising Sandy Larson Farm (Houghton Loam Co. Bi I. II II II II A2 II II II II B II II II II C II II II II l4 A Sims Clay Schean Farm (Saginaw Loam Co. B II II II II C II II II II IS A McBride Sandy Comden Farm (Montcalm Loam Co. 8 II II II II 26 Table I, cont'd Soil Profile Horizon Soil Type Location no. I6 A Conover Loam Sloan Creek (lngham Co.) B II II II II C II II II II I7 A ‘Parkhill Clay Davis Farm (Sanilac Loam Co.) B II II II II C II II II II l8 A Montcalm Hey Farm (Montcalm Loamy Sand Co.) B II II II II C II II II II l9 A Hillsdale Deercreek (lngham Co.) Sandy Loam 8 II II II II C II II II II 20 A Miami Loam Deercreek (lngham Co.) 8 II II II II C II II II II 2l A Spinks Loamy Sloan Creek (lngham Sand Co.) w 27 Table l, cont'd Soil Profile Horizon Soil Type Location no. 22 Spinks Loamy Water-shed (lngham 23 24 25 26 27 28 Sand Brookston Clay Loam Brookston Loam Sims Sandy Clay Loam Miami Sandy Loam Spinks Loamy Sand Co.) College Farm (lngham Co.) Sloan Creek (lngham Co.) Ferden Farm (Saginaw Co.) College Farm (lngham Co.) College Farm (lngham Co.) Water-shed (Corn) (lngham Co.) 28 Table l, cont'd Soil Profile Horizon Soil Type Location no. 28 C Spinks Loamy Water-shed (Corn) Sand (lngham Co.) 29 A Muck Muck Experimental Farm (Clinton Co.) 29 an equilibration period of I or 2 hours is sufficient for (a) most of the labile inorganic P to come in equilibrium with ambient solution and (b) for period greater than 2 hours, onset of microbial activity is pronounced causing a decrease in equilibrium P conc or increase in potentialSP measured. Typical time adsorption curves indicate an initial fast reaction followed by a secondary reaction which proceeds at a slow and almost constant rate. Studies on the effect of soil: solution ratio on the Freundlich isothorm had shown this variation to be small (Davis, I935; and Kurtz et al., I946). Since recent investigations by White and Beckett (I964) have confirmed that the potentialzg is independent of soil:solution ratio for soils under prolonged storage, this variable was not investigated further and a soil:solution ratio of l:lO with a 2 hour5* equilibriation time was adopted. PotentialSp Five gram soil samples were equilibrated for 2 hours with 50.0 ml of 0.0I M CaCIZ of varying P conc (Ca(H2POq)2 H20). Conc of P used varied in general between 0 and 50.0 x IO"5 M. For certain soils a level of P as high as l50.0 x IO‘5 M was used, for at lower levels equilibrium P conc in solution was too low to detect accurately. The suspensions were cleared by filtering and/or centrifuging at 3000 rpm for 5 minutes. Total inorgani 30 c P was determined by the Mo blue method in sulfuric acid system (Jackson, I958). The pH of the supernatant solution was measured with a Beckman Model G pH meter using glass electrode. Ca was determined on 303 Perkin Elmer Absorption Spectrophotometer. The latter did not change significantly from 0.0I M nor did the pH values vary over the range of P conc studied. The potentialsSp were calculated according to White and Beckett (I964) as outlined I/2 (Ca) Where: below: pCa = l/2 (-Iog(Ca)) f [Ca}n-conc in moles (Ca) = activity of Ca log f = -0.AZZ V? W 0:5 ”effective diameter” of the ion valency of the ion u = the ionic strength I/2ZM.Z? (Mi is molality I I I of i ion) 8 = constant for a given solvent (H2P04) = (H+)‘_ x total P where (HT) = H+ activity (from pH measurements) f1 activity f2 activity coefficient of HZPOQ' coefficient of HPOL,= 3I The ionic strength was taken to approximate that for CaClz because of the preponderance of Ca ions in the system. The potentialSp equals I/2 p Ca + pHZPOQ. Therefore, as the logarithm of a reciprocal, the potentialSp increases as the activity product aCaé x aHZPQ4 decreases. The difference between the initial P conc and equilibrium P conc of each solution gave the amount of P gained or lost by 5.09 of soil. Regression analysis by (bntrolled [Eta 3600 computer, was run between E10 P/g P soil and the corresponding potentialgp. The potentialzg was calculated from the regression equation whenzSP = O. This is the potentialSp of the solution with which the original soil undergoes no net exchange of P during the equilibration time. Potential Buffering Capacity (PBC) Beckett and White (I964) following Schofield's suggestion (I955) proposed PBC as the quantitative measure of ability of a soil to maintain the potentialSp against P withdrawl. The PBC for P was defined as (dQ/dl)] or (le/le)l_ll at a given potentialSp (I) or over a given range of potentialSp (IflI), respectively. The Q and I represent the quantity of P (extensive parameter) and potentialsP or P activity (intensive parameter), respectively. In the present study PBC of the soil was calculated by the regression equation between fifl P/g soil and potentialSp. 32 P Adsorption Maximum As discussed in the preceding chapter, Langmuir adsorption isotherms were run for the data used to calculate potentialzg. The zero levels of P however, were excluded from the run. The reciprocal of the slope-% of the Langmuir adsorption equation was obtained as the P adsorption maximum. Extractable Al and Free Iron Oxide Extractable Al was determined by ”aluminon” method as described by Mclean (I965) and free iron oxides by l'orthophenianthroline" method described by Olson (I965). Bulk Density Bulk density determinations were done in Soil Physics laboratory by ”Core Method” Blakes (I965). RESULTS AND DISCUSSION Equilibrium PotentiaISP and Equilibrium Conc of Inorganic P A typical plot of potentialSp vs t P/g soil is presented in Fig. I. The soils for the plot were selected to give a range in the slope and the X intercept. The f. P/g soil showed a fairly good linear regression on potentialSp, and a ”r” value on an average of 0.982 was obtained. The potentialég and total P at equilibrium (i.e. at J~P=0 herein after called TPeq) in Table 2. The TPeq was calculated by subtracting I/2 pCa from potentialgg calculated from regression equation are given and then applying the following equation: TPeq = _ + f l (lo’PH)f 2 where f I and f 2 are the activity coefficients of HZPOQ- and HPOA: respectively, and were discussed earlier. Data in Table 2, in general, reveal low amounts of TPeq. The TPeq however, ranged from 0.08 x IO'6M for Emmett loamy sand Bihr (6) to l7.37 x IO'6M for Spinks loamy sand A (22). 'The podzolic profiles no. l-9, and Munising sandy loam (l3) Iuad in general lower TPeq or higher potentialzg than found in (Jther profiles. In particular, the B horizons have exceedingly Icwv amounts of P at equilibrium. The A horizon in general had Iwigfmar TPeq within the group. This trend appears to be negatively ccn'related with the P adsorption maximum for the soil horizons discussed later. 33 34 30.0 f IOA y = 88.0 - I2.4x r = 0.978 l4B y = l05.I - I6.4x r = 0.985 4A y = 8.8 - I.2x r = 0.958 . 9A y = 23.9 - 3.3x r = 0.963 20.0f :: o o If) 0 3‘ '\ / IO / 04 x 10.0 _ 4.9 3’3 + C‘) '3 . :2 Q. Q 94 c. + . + K 0.0 l A I \ 5.00 6.00 . 7.0! Q Ti: _T PotentialSp '5.C 8 Fig. l. Relationship between potentialSp and/‘P/g soil. 35 Table 2. The potentialzg and TPeq in 29 Michigan soil profiles Soil type pH Potentialzg TPeq MxTO-b I Dune Sand 6.5 7.392 0.98 2 Rubicon Sand A] 7.2 7.9I2 0.62 A2 6.8 7.967 0.33 B 6.7 7.658 0.62 C 6.4 7.523 0.68 3 Rubicon Sand AO-A] 6.7 7.696 0.56 A2 5.5 7.283 0.94 B 4.9 7.422 0.66 C 5.7 7.069 I.58 4 Rubicon Sand A 6.4 7.400 0.90 B 6.5 7.455 0.84 C 6.6 7.677 0.54 5 Emmett Loamy Sand A2 7.0 7.48I I.28 A8 6.6 7.583 0.67 Bir' 6.9 7.657 0.75 82 6.8 7.758 0.54 C 7.I 7.833 0.65 6 Emmett Loamy Sand AI 7.I 7.562 I.2I A2 6.9 7.7II 0.67 AB 6.8 8.06I 0.27 Bihr 6.5 8.460 0.08 82 6.8 7.858 0.43 7 Grayling Sand AI 8 A2 5.6 7.3OI 0.9I B 6.5 7.499 0.76 Bihr 7.4 8.460 0.25 C 7.0 8.049 0.34 8 Roseland Sand A],A2 5.8 7.I48 I.33 B 6.8 8.l39 0.22 C 6.5 7.70l 0.48 D 6.3 8.037 0.20 36 Table 2, cont'd Soil type pH Potentialgg TPeq MxlO:6 9 Rubicon Sand Al 8 A2 4.I 7.529 0.5I Bihr 4.8 7.834 0.26 C 6.I 7.549 0.57 C2 6.2 7.72I 0.39 l0 Kalamazoo Loam Al 6.6 7.20 I.63 B] 6.7 7.380 I.l7 82] 6.0 7.590 0.50 B22 6.7 7.392 I.34 B3-C] 6.9 7.6I7 0.83 C 7.5 7.947 0.96 II Warsaw Loam A 6.6 7.I66 I.77 8I 5.0 7.346 0.79 BZI 4.9 7.346 0.79 822 4.7 7.323 0.83 823 5.3 7.504 0.56 C 5.9 7.79I 0.3I I2 Ontonagon Clay A] 5.4 7.32I 0.85 A2 5.4 7.253 1.00 AB 5.2 7.394 0.7I 821 6.5 7.375 1.02 822 6.9 7.5I9 I.04 C 7.2 7.7l7 0.98 I3 Munising Sandy Loam AP 5.6 7.I99 I.I5 Bir 5.4 7.573 0.48 A2 4.8 7.552 0.49 B 4.4 7.433 0.64 C 5.0 7.586 0.46 I4 Sims Clay Loam A 7.2 7.578 I.35 B 5.7 6.509 7.25 C 7.I 7.459 I.54 I5 McBride Sandy Loam A 6.I 6.59l 5.I8 B 6.3 7.I08 I.69 C 5.2 7.4OI 0.70 Table 2, cont'd -37 Soil type pH Potentialég TPeq IVIxIO'6 I6 Conover Loam A 6.I 7.000 2.0l B 6.6 7.I99 I.64 C 6.8 7.295 I.56 17 Parkhill Clay Loam A 6.9 7.I4I 2.49 B 6.I 7.009 l.97 C 6.9 7.660 0.75 l8 Montcalm Loamy Sand A 6.8 7.430 I.45 B 5.6 7.I5l I.29 C 5.7 7.246 I.05 I9 Hillsdale Sandy Loam A 6.2 7.I2I I.57 B 6.3 7.263 I.l8 C 6.5 7.I79 I.60 20 Miami Loam A 6.9 6.906 4.28 B 6.9 7.5l8 I.04 C 6.7 7.454 0.99 2l Spinks Loamy Sand A 6.3 7.I92 I.39 B 6.1 7.319 0.96 C 6.2 7.460 0.72 22 Spinks Loamy Sand A 6.3 6.I00 7.37 B 6.4 7.506 0.7] C 6.8 7.I55 2.I6 23 Brookston Clay A 6.5 6.476 8.l2 B 6.2 7.357 0.9I 24 Brookston Loam A 6.9 7.762 0.59 B 6.6 7.8I8 0.39 C 7.3 7.7I6 l.I5 '38 Table 2, cont'd Soil type pH Potentialgg TPeq MxIO‘S 25 Sims Sandy Clay A 6.7 6.949 3.I7 Loam B 6.5 7.726 0.45 C 6.8 7.960 0.34 26 Miami Sandy Loam A 6.I 6.766 3.45 B 7.2 7.449 I.82 C 6.8 7.824 0.46 27 Conover Loam A 6.I 7.4l3 0.77 B 5.8 7.558 0.52 c 6.7 7.477 0.94 28 Spinks Loamy Sand A 5.7 6.86 3.82 B 5.8 7.047 I.68 C 6.7 7.233 I.64 29 Muck A 6.0 6.I47 4.05 B 6.8 7.036 2.85 C 6.2 6.894 2.66 39 The Sims clay loam 8 (I4), McBride sandy loam A (l5), Miami loam A (20), Miami Sandy loam (26), Spinks loamy sand A (22,28), Brookston clay loam A (23), and Muck A (29) had higher TPeq conc than other soils. These P levels fall in the range needed for Optimum plant growth (Aslying, I954). Aslying (I954) found crops growing on soils with potentialsP greater than 8, corresponding to a P conc in the soil solution of I0'7M or less, usually responded very well to P fertilizer; whilst crops on soils with potentiaISP.of 6 or less, P conc of IO'SM or more, rarely responded. Recently Ozanne and Shaw (I968) found that pasture plants growing on soils not adsorbing P from equilibrium solution containing IO'SM P or more did not respond to applied P. Russell (I96I) also writes - I'Experiments to measure the minimum concentration of phosphate needed for good plant growth are technically difficult to carry out, but there seems no doubt that most crOps can make adequate and possibly optimum growth if phosophate concen- tration around their roots is kept at IO'SM, many crops may b2 able to make good growth if it is as low at IO' M, provided the conditions of growth allow good root deveIOpment, but that for at least some crops I0'7M is much too dilute; and these results are concordant with the observed field behaviour.” The higher TPeq in soil suspension in all the foresaid soil horizons are either because of high percentage saturation of the adsorption maximum, which probably is a consequence of recent treatment Wlth P fertilizers, and/or bonding energy of P on adsorbent. Recently Murrmann and Peech (I969a) observed simul- taneous increase in labile and soluble P in 3I soil samples 40 collected from old fertilizer experiments in Illinois, Kentucky, Ohio, and Pennsylvania. Spinks loamy sand A (22) which showed highest TPeq, l7.37xl0'6M has a low adsorption maximum (4.98 mg/IOOg, Table 3) and is also highly fertilized, (A.E. Erickson, I969, Personal communication, Michigan State University). The same is true for Spinks loamy sand A (28) except it has twice the adsorption maximum of Spinks loamy sand A (22) and it is probably for this reason the TPeq conc in the former is approximately I/5th of the latter. All horizons of Muck profile have higher TPeq with the A horizon as high as l4.05xl0’6M P. Apart from the degree of P saturation of the adsorption maximum, the low energy of adsorption of P on organic matter (Table 3) is probably another reason for the observed higher TPeq. Harter (I969) had attributed most of the P adsorption on organic matter to anion exchange and this will have a lower energy of adsorption than Fe and Al-P adsorption complex. Further the lack of Fe and AI in all the horizons of muck will decrease the chances of precipitation of adsorbed P. But the reason for an unusually high P content in Sims clay loam 8 (I4) is not apparent. The uptake of ID to 20 kg. of P/ha by an average crop requires hundreds of times as much P as is normally present in solution in a fertile soil at one time within the depth of rooting. Therefore replenishment of the soil solution, i.e. rate of release of P from soil to the solution, is of paramount 4I importance. The P adsorption maximum and potential buffering capacity is discussed in this context in the following sections. P Adsorption Maximum The adsorption maximum for the soils studied and K, the constant proportionaltofree energy of adsorption is given in Table 3. A typical Langmuir adsorption plot of the data is presented in Fig. 2. Surface P already present in soil is related to the equili- brium P conc according to the equation X/m = Kbc . Ideally I+Kc the adsorption would be determined on an adsorbate free surface, but because of nonfeasibility of this restriction with soils, a correction by adding the amount of surface P determined by a separate analysis to the X/m is made. Since this correction has negligible effect on %'(0Isen and Watanabe, I957) particularly on nonfertile soils, no attempt was made to correct the data for the amount of surface P present initially on the soils. The data (Fig. 2) show satisfactory agreement with the Langmuir isotherm as a straight line relationship and a r value, on an average of 0.985 was obtained. The soils studied show a great deal of variation in their ability to adsorb P (Table 3). This is also true for the horizons of most profiles, particularly the sandy podzolic group. The 42 Table 3. P adsorption maximum and K, the free energy of adsorption for 29 Michigan soil profiles Bulk 4 Type densityX le0 Adsorption Maximum 976m3 %_xT08 mg/TOOg Kg/ha 1 Dune 1.u9 4.51 1.88 85.4 2 Rubicon Sand Al I.24 0.77 4.24 l60.l A2 I.49 l.97 3.88 I76.0 B I.4I 4.00 I0.23 826.4 C I.50 5.8I l0.75 49I.6 3 Rubicon Sand AO-Al I.24 I.57 3.47 I3l.2 A2 1.49i 0.99 u.u0 200.1 B l.4l+ 5.93 ll.28 482.9 C I.50+ 2.70 II.49 525.5 4 Rubicon Sand A I.24+ 3.90 I.07 40.4 B l.4l+ 5.l4 27.78 Il93.8 C I.50+ 4.44 l2.50 57I.5 5 Emmett Loamy Sand A2 0.62 I.63 4.43 83.6 AB I.42+ 3.06 I0.20 44l.6 Bir I.42 3.43 9.7I 420.2 B2 I.64 3.22 9.7I 485.3 C I.64+ 3.I0 4.8l 240.3 6 Emmett Loamy Al ,I.I7+ I.8I 5.08 I8I.0 Sand A2 I.4l+ l.98 4.63 l99.0 AB I.42+ 4.52 I0.53 455.6 Bihr I.42+ 3.83 9.0l 389.9 82 I.64+ 4.09 I0.64 53l.8 7 Grayling Sand AI8A2 I.I7+ 2.32 8.77 3I2.8 B I.42+ 6.27 I4.49 627.3 Bihr l.46+ 7.43 I9.23 855.8 C I.60+ 4.70 l0.64 5l8.8 x - Data were provided by courtesy of Dr. A. E. Erickson, I969 Michigan State University, East Lansing. + - An estimate 43 Table 3, cont'd Bulk Type density le0“ Adsorption Maximum g/cm5 l_xl0“ m§71009 Kg7Ea M 8 Roseland Sand A08AI I.I7+ 2.8I 7.4I 264.I B I.42+ 8.80 22.73 983.7 C I.44 9.l7 I8.l8 83I.3 D I.60+ 6.44 9.7I 473.5 9 Rubicon Sand A08AI I.I7 36.70 2.73 97.2 B I.46 8.00 4I.67 I854.2 C I.50 7.55 I2.05 550.8 C2 I.55+ 2.64 6.89 325.8 IO Kalamazoo Loam AI I.34+ 2.24 9.7l 396.5 BI I.50+ I.69 20.4I 933.I BZI I.50+ II.I7 I4.93 682.4 B22 I.50+ 5.7I l2.50 57I.5 B3 l.60+ 6.00 7.94 387.I D I.60+ 3.59 9.62 468.9 II Warsaw Loam Al I.34+ 3.00 l8.52 756.4 BI I.50+ 4.I7 40.00 I828.8 BZI I.50+ 5.00 50.00 2286.0 822 I.50+ 5.00 22.22 IOI6.0 B3-C I.60+ 6.0I I0.99 535.9 0 I.60+ 5.45 4.60 223.7 I2 Ontonagon Clay AI I.34+ 2.22 50.00 2042.2 A2 I.34+ 3.75 33.33 I36I.4 AB I.50+ 5.67 29.4I I344.7 BZI I.50+ 4.50 l8.52 846.7 822 I.50+ 3.94 I4.93 682.4 C l.60+ 4.I2 I4.29 696.7 I3 Munising Sandy Ap I.34+ 5.I3 24.39 996.7 Loam Bir I.50+ I0.50 23.8I I088.6 A2 I.50+ 7.80 I2.82 586.2 8 I.50+ 4.80 20.83 952.5 C l.60+ 3.33 l0.00 480.8 I4 Sims Clay Loam A I.35 3.70 30.00 l234.4 B I.36 0.79 I2.20 505.5 C I.56 5.56 20.00 95I.0 44 Table 3, cont'd Bulk Type density le05 Adsorption Maximum g/cmg— ngIO2+ mg/IOOg Kg/ha l5 McBride Sandy A I.47 2.46 I0.42 466.7 Loam B I.49 4.II I2.82 582.3 C I.46 7.30 I3.70 609.6 l6 Conover Loam A I.58 3.37 I0.99 529.2 8 I.62 5.57 25.64 l266.l C I.83 4.4I l3.33 743.7 I7 Parkhill Clay A I.04 6.28 5.00 l58.5 Loam B I.47 2.3I 8.85 396.5 C l.60+ 6.67 l6.67 8l2.8 l8 Montcalm Loamy A I.80 4.4l 6.I4 336.6 Sand B I.58 5.I9 9.I7 44l.8 C I.50 3.38 5.32 245.8 l9 Hillsdale Sandy A I.49 3.I6 I2.66 574.9 Loam B I.50 5.00 II.II 508.0 C I.75 3.79 I0.99 586.2 20 Miami Loam A I.5I I.I5 l6.67 767.l B I.66 5.93 l2.05 609.6 C I.78 4.77 l6.l3 875.7 2I Spinks Loamy A I.26 2.68 I0.l0 387.9 Sand 8 I.40 6.33 l3.I6 56I.4 C I.I8 3.54 7.63 274.6 22 Spinks Loamy A I.39 I.50 4.98 2I0.8 Sand B 1.50 2.98 8.00 365.8 C I.45 I.I6 Il.49 508.0 23 Brookston Clay A I.32 0.56 27.03 l087.4 Loam B I.44 6.44 l7.24 756.7 24 Brookston Loam A I.39 6.00 l8.52 784.6 8 I.6l 6.89 l6.l3 79I.5 C I.77 2.96 l4.08 759.83 45 Table 3, cont'd Bulk Type density leOLI Adsorption Maximum ‘g/cmg' ‘A xlOII mg7IOOg Kg/ha 25 Sims Sandy Clay A .I.00 I.56 l2.50 38I.0 Loam B I.5I 2.95 I7.86 82I.9 C I.60 I0.20 I9.6I 956.2 26 Miami Sandy A I.34 0.53 I0.87 443.9 Loam B I.58 l.97 I4.08 678.3 C I.54 I0.80 l8.52 869.3 27 Conover Loam A I.45 2.26 I0.53 465.2 B I.6I II.I7 I4.92 732.4 C I.80 6.08 I3.70 75I.6 28 Spinks Loamy A I.28 I.48 Il.49 448.4 Sand B I.44 I.02 I7.54 770.0 C I.40 I.23 8.40 358.6 29 Muck A O.50+ 0.56 I9.50+ 297.2+ B O.55+ 0.60 28.57 478.0 C 0.60+ 0.60 2I.28 389.II 2A y = 0.75I 10 22 y I8A y = C/x/m 1 0.0 0.50 C. OIIO 46 .258x + 0.I3I r = 0.972 0.080x + 0.0I4 r = 0.990 .I63x + 0.037 r = 0.986 I I i I 1,0 . I.50 . 2.00 2.50 3.00 Equilibrium P conc moles x IO‘LI Fig. 2. Langmuir adsorption plot for P adsorption data 47 two extremes in P adsorption maximum were 40.4 and 2042.2 Kg/ha for Rubicon sand A (4) and Ontonagon clay A (I2); respectively. The Ontonagon Clay (l2) has a very high free iron oxide content (39.57 meq/IOOg) and has very high clay content; whereas, Rubicon sand A (4) represents minimal soil deveIOpment and contains low amounts of free iron oxide (5.I4) and extractable Al (0.067 meq/IOOg soil). A regression of P adsorption maximum on extractable Al and free iron oxide had 0.002 and 0.l63 partial correlation coefficients respectively, when all the soils were pooled together. However, when regression was run for the individual soils (soil with more than 3 horizons, since multiple correlation cannot be computed with only 3 points), significant correlation coefficients between extractable Al and free iron oxides on one hand,and P adsorption maximum on the other were obtained. This suggests that factors other than Al and iron oxide, such as the clay content (its quantity and quality) and others are important in P sorption and hence were masking the regression of P adsorption on Al and Fe when all soils were considered together. In general the A] and A2 horizons in sandy podzolic group have the lowest adsorption maximum and B horizons have compara- tively very high adsorption capacity within the group. This trend is correlated with extractable Al and free iron oxide content of the horizons (Table 4). For Rubicon sand (2) 48 Table 4. Extractable Al and free iron oxide content of the 29 Michigan soil profiles Type Extractable Al Free iron oxides l Dune Sand 2 Rubicon Sand A A2 3 Rubicon Sand A08AI 4 Rubicon Sand A 5 Emmett Loamy A2 Sand AB Bir B2 C 6 Emmett Loamy AI Sand A2 A8 Bir 82 7 Grayling Sand AI8A2 B Bihr C 8 Roseland Sand AI8A2 8 C 0 -'-‘OO --'—'00 O ONNO O—'-'O 00000 00000 .041. .027 .l56 .822 .I06 .059 .l28 .283 .l95 .067 .056 .278 .043 .293 .847 .267 .l6l .034 .l08 .30l .389 .I94 .256 .245 .828 .780 .687 .278 .278 .989 meq/TOOg soil ‘ d # WNNW Nd—‘CD \I-I-‘NkO-fl (13*me C‘C‘U‘l U'IOU‘IKD W-dNU‘I W ‘ :50 .35 .429 .929 .679 .072 .857 .643 .357 .00 .536 .I43 .429 .429 .I43 .943 .357 .572 .00 .286 .626 .50 .643 .I43 .572 .286 .57 .7l4 .429 49 Table 4, cont'd Type Extractable Al Free iron oxides meq/TOOg sOTT 9 Rubicon Sand AI8A2 0.220 3.07 Bihr 3.634 I6.429 C 0.894 4.286 C2 0.354 I.7l4 l0 Kalamazoo Loam Al 0.ll0 22.85 BI 0.289 25.00 82I 0.398 32.00 822 0.293 25.7l5 83CI 0.089 28.286 D 0.l33 26.429 II Warsaw Loam Al I.283 29.286 BI 2.222 33.572 82l 2.889 32.00 822 I.56l I9.786 823 0.709 I3.2l5 C-D 0.22I 5.7I4 I2 Ontonagon Clay Al 0.955 39.572 A2 0.689 40.572 AB 0.672 39.00 82l 0.394 39.00 822 0.36I 39.00 C 0.220 34.286 l3 Munising Sandy Ap 2.2II 25.00 Loam Bir 3.545 I8.286 A2 I.383 I6.00 8 I.828 22.286 C 0.5I3 I3.7I5 l4 Sims Clay Loam A 0.269 37.7I5 8 0.733 l2.l43 C 0.l53 I8.l43 I5 McBride Sandy A 0.366 l5.429 Loam B 0.83I I4.286 C 0.304 24.l43 50 Table 4, cont'd Type Extractable AI Free iron oxides meq7IOOg soil I6 Conover Loam A 0.ll6 20.7l5 8 0.522 22.I43 C 0.3II 35.7I5 I7 Parkhill Clay A 0.l33 8.572 Loam 8 0.256 I3.857 C 0.228 I0.00 l8 Montcalm Loamy A 0.209 34.286 Sand 8 0.9I7 I9.429 C 0.367 6.429 l9 Hillsdale A 0.598 2I.429 Sandy Loam 8 0.528 27.858 C 0.598 l8.572 20 Miami Loam A 0.l67 20.572 8 0.639 2I.429 C 0.4II 26.572 2l Spinks Loamy A 0.222 I5.00 Sand 8 0.950 I5.00 C 0.3ll Il.857 22 Spinks Loamy A 0.l33 20.286 Sand 8 0.222 20.286 23 Brookston A 0.089 22.I43 Clay Loam 8 0.197 30.00 24 Brookston A 0.l00 l2.l43 Loam 8 0.278 28.572 C 0.2ll 29.l43 25 Sims Sandy A 0.057 27.I43 Clay Loam B 0.l32 32.858 C 0.200 30.858 5l Table 4, cont'd Type Extractable AI Free iron oxides meq/TOOg soil 26 Miami Sandy A 0.079 21.429 Loam C 0.333 29.l43 27 Conover Loam A 0.098 2I.429 B 0.36l 35.l44 C 0.3II I9.429 28 Spinks Loamy A 0.247 26.429 Sand 8 0.339 35.I44 C 0.2II 16.857 29 Muck A 0.072 I0.00 8 0.076 7.857 C O.IO6 5.357 52 partial correlation coefficients of 0.999 and 0.979 for the regression of P adsorption maximum on extractable AI and free iron oxides respectively were obtained (Table 5). From the correlation coefficient data for other soils too, it appears that extractable Al is rather more important in P sorption for short equilibriation periods than the free iron oxide content of the soils. With the latter, negative correlation coefficients were even obtained (Table 5). However, not much reliance should be put on these correlations, because only one degree of freedom for error term was involved. 0f the fine textured soils, the Warsaw loam B], 82] and 822 (II), Ontonagon clay A], A2 and AB (l2) Munising sandy loam, Ap, Bir and 8 (I3), Sims clay loam A (I4), Conover loam 8 (I6), and Brookston clay loam A (23), possess exceptionally high adsorbing capacities. These horizons also have relatively higher extractable Al contents. The P adsorption maximum for the various horizons of the finer textured profiles were not as distinctly differentiated as for the sandy podzolic group. Correspondingly there was a lack of such a distinction in the AI and Fe content of the horizons. Parkhill clay loam (I7) has the lowest adsorption maximum for the A and B horizons among the fine textured soils. This is probably because of its calcareous nature and relatively lower content of Al and Fe. 53 Table 5. Correlation coefficients between P adsorption maximum and Fe and AI in soils Soil Type Correlation Coefficients Free iron oxides Extractable Al 2 Rubicon Sand 0,9798 0.999% 5 Emmett Loamy Sand 0.765ns 0.730nS 6 Emmett Loamy Sand 0.8llns 0.532ns 7 Grayling Sand -0.268ns 0.976* 8 Roseland Sand 0.3l9ns 0.907% 9 Rubicon Sand -0,996* 0.999** l0 Kalamazoo Loam -O.9.428ns 0.790ns ll Warsaw Loam -0.403ns 0.957A I2 Ontonagon Clay -0.9l2* 0,995** l3 Munising Sandy Loam 0,906 0.956A * significant at 5% level ** significant at I% level ns non significant 54 The bonding energy calculated from the Langmuir plot of the data (Table 3), (K = slope/intercept) has a large error term inherent in the method of calculation. However, there is a trend for higher energy of adsorption of P with B horizons or horizons with higher Al and Fe content. The surface horizons often have lower bonding energy. The value Of the bonding term calculated from the Langmuir equation provides an estimate of the average bonding energy of P on the major adsorbing surfaces. However, the P reacting initially at low surface coverage may be bound more strongly than that reacting subsequently if the bulk of the sorption sites are occupied. A linear fit of the data, however, with the Langmuir isotherm suggests near uniformity in bonding energy of P for the conc of P used. Investigators have shown two stages of P sorption -- a rapid initial reaction followed by a relatively slow reaction (Low and Black, I950). They indicated that this latter reaction would reach a definite end point with time. The rapid initial sorption involves P ions becoming attached to exchangeable Al+3, Fe+2+3, and Ca+2 ions of the clay and hydrous oxides or these same ions held in the outer edges of the lattice. This reaction is a type of chemical adsorption involving primary valence bonds rather than a physical adsorption. It seems plausible to suggest that the extent of this initial reaction is determined by the adsorption lnaximum from the Langmuir isotherm. 55 Cole, Olsen and Scott (I953) have presented evidence that in the range of equilibrium conc where the Langmuir equation applies, essentially all of the P adsorbed on CaC03 was exchangeable with 32P. Similar data implying a monolayer adsorption were found for P adsorbed on ferrated IRC 50 cation exchange resin (Fried and Dean, I955) over the range of conc which followed the Langmuir isotherm. With higher P levels, however, less of the adsorbed P was exchangeable with 32P and the isotherms deviated from a straight line. The deviation from the straight line was also noted in the present investigation at higher P conc for some soils, especially sandy podzolic A and A2 horizons and certain low adsorbing fine textured soils. The equation C = discussed earlier suggests that 9 TTT0)K a close relationship exists between percentage saturation of the adsorption maximum and TPeq in soil suspension. The theoretical relationship between these two variables for few typical A horizons representing the textural groups is illustrated in Fig. 3. The observed C values for the corres- ponding percent saturation of the adsorption maximum are also indicated. The close correlation between the percentage saturation of adsorption maximum and the corresponding TPeq, indicates that the amount of P in soil suspension closely reflects the degree of P saturation on the adsorbing surfaces c x IO'AMP 3.00 2.50‘ 2.0 I.50 I.0 0.50 56 O ..-a Rubicon Sand (2) 3>Ontonagon clay (I2) :+ Brookston lOam (24) O Hillsdale sandy loam (l9) 0» .00 20.0 40.0 '60.0 8010 ‘ J00.0 P adsorbed/adsorption maximum X100 Fig. 3. Relationship between percent saturation of adsorption maximum and equilibrium P conc in soil solution 57 of soil complex. This relationship, however, appears to deviate at higher saturation in some cases. Beyond this point the predicted and the observed C's for the corresponding percent saturation of the adsorption maximum fell wide apart. This probablyves because of precipitation at higher conc where data also deviate from the Langmuir isotherm. For normal soil fertilization practice however, the evidence presented herein strongly suggests that the adsorption maximum of a particular soil may be used to predict the TPeq; thus, given a required TPeq in soil suspension for optimum plant growth the degree of saturation of the adsorption maximum which is necessary to obtain the desired level may be calculated. The degree of saturation of the adsorption maximum required to give a l.0 x IO’SMP in soil suspension and P (Kg/ha 22.86 cm) required to bring the soil to that level is given in Table 6. The conc of l.0 x IO'SMP was chosen since if this conc is maintained at soil-root interface, in the absence of other limiting factors, it should support optimum plant growth for most crops (Russell I96l, Ozanne and Shaw I968). The fraction of the adsorption maximum already occupied was calculated using the equilibrium equation C = 9 and assuming that the TPeq obtained by White and Beckett procedure (I964) was governed by the above equation. Observed data do substantiate the assumption implied in calculation. Data in Table 6. K9/haO-22.86cm of P needed to give I.0xl0’5MP in soil solution 58 Soil type Initial 0 9 Needed P needed Kg7ha l Rubicon Sand 0.042 0.269 l7.2 2 Rubicon Sand 0.007 0.l20 l4.3 3 Rubicon Sand 0.00l 0.ll4 I2.9 4 Rubicon Sand 0.034 0.28] 7.5 5 Emmett Loamy Sand 0.020 0.l40 7.5 6 Emmett Loamy Sand 0.0l8 0.l59 20.2 7 Grayling Sand 0.02I 0.l89 39.4 8 Roseland Sand 0.036 0.220 43.5 9 Rubicon Sand 0.l63 0.786 42.6 l0 Kalamazoo Loam 0.035 0.l83 44.0 ll Warsaw Loam 0.050 0.23l l02.4 I2 Ontonagon Clay 0.0I9 0.l82 249.8 l3 Munising Sandy Loam 0.05I 0.339 2II.2 l4 Sims Clay Loam 0.048 0.270 27l.5 l5 McBride Sandy Loam 0.ll3 0.l98 29.6 l6 Conover Loam 0.063 0.253 74.9 l7 Parkhill Clay Loam 0.l35 0.386 29.8 I8 Montcalm Loamy Sand 0.048 0.306 65.l l9 Hillsdale Sandy Loam 0.047 0.240 83.l 20 Miami Loam 0.047 0.l04 32.4 59 Table 6, cont'd Soil Type Initial 9 0 Needed F’needed Kg7ha 2I Spinks Loamy Sand 0.036 0.2Il 5l.0 22 Spinks Loamy Sand 0.207 0.l30 -- 23 Brookston Clay Loam 0.044 0.053 7.7 24 Brookston Loam 0.034 0.375 200.5 25 Sims Sandy Clay Loam 0.045 0.l36 25.8 26 Miami Sandy Loam 0.0I8 0.050 l0.8 27 Conover Loam 0.0l7 0.l85 58.4 28 Spinks Loamy Sand 0.053 0.l29 25.3 29 Muck 0.072 0.056 -- 60 Table 5 demonstrate the differential capacity of soil to resist increase or decrease in P conc following addition or withdrawl of P bearing fertilizers to the soil. Ontonagon clay (l2), Sims clay loam (l4) and Brookston loam (24) have a very high requirement for P to yield desired P conc in solution; whereas,podzolic sands have a low requirement for P to bring the P conc to the same level. Moreover, at a given TPeq’in soil solution, the adsorbed P in the soil solids is less in soils of coarse texture than of fine texture. The close agreement with the theoretical and observed values of TPeq in soil suspension had another very important implication, that adsorbed P being in equilibrium with the P in solution should serve as the primary source from which P is released upon depletion by plants. Machold (I962) had reported high corre- lations between P that equilibrated with 32P and that absorbed by rye seedlings. In the light of the data obtained in the current investigation, Machold's (I962) findings may be inter- preted to mean that P less accessible to exchange with 32F, hence to equilibration, was also less available for uptake. Since total adsorbed P does not reflect the TPeq in a soil suspension (at any one final solution conc a range of aadsorbed P values were obtained depending on the characteristics (3f the soil) it would appear that surface P measurements alone vvill not be proportional to either available P or TPeq. 6I As the equation C = 0 suggests, K- the free energy (I-OIK of adsorption of P on the soil compex, besides 9; is important in producing a P buffering effect of soil against increase or decrease in P conc. The Ontonagon clay (I2) had only a slightly greater requirement of P to bring the P level to l.0 x IO'SM compared to Brookston loam (24). The former, however, had an adsorption maximum of 2042 Kg/ha compared to 784 Kg/ha for the Brookston loam (Table 3). The greater P buffering effect of Brookston loam is because of its higher energy of adsorption of P compared to Ontonagon clay (Table 3). The same is true for Sims clay loam (l4) which has an even higher P requirement than Ontonagon clay even though its adsorption maximum was only 2/3 of the latter. Since for surface horizons K is much less variable than the adsorption maximum,removal of a given quantity of P from soil should produce less decrease in P in solution for soils with higher adsorption maximum (fine textured, high Al, Fe and clay content) than those of low adsorption maximum because of greater capacity of the soil to supply P by dissolution. Olsen and Watanabe (I963) have suggested that the soils of fine texture would contain greater volume of water than would soils of coarse texture at the same inatri>