PHOSPHORUS AND iROrN CONTENT, REDOX POTENTIAL ? - , AND PHOSPHORUS' FRACTiONATION OF LAKE _ .SEBiMENT - WATER INTERFACES Thesis for the Degree. of M. S. MICHIGAN STATE UNIVERSITY . :. D-ENiNis B. FEN-N , 1970 l ...... . u - \‘ \\\\\\\\|\\\\M\\\\\\\\\\NW\l _‘ Tm R A R y 5 8065 fl\\|\\\\\|\\\\||\ 3 1293 1 Michigan 5 rate Universi ,y PHOSPHORUS AND IRON CONTENT, REDOX POTENTIAL AND PHOSPHORUS FRACTIONATION OF LAKE SEDIMENT-WATER INTERFACES By 9 .2‘ r: l \x Dennis B)\Fenn AN ABSTRACT OF A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Crop and Soil Sciences 1970 ABSTRACT PHOSPHORUS AND IRON CONTENT, REDOX POTENTIAL AND PHOSPHORUS FRACTIONATION OF LAKE SEDIMENT-WATER INTERFACES By Dennis B. Fenn Six lakes of widely varying quality were sampled at 5 or 6 depths at 3 m intervals in June 1969, October 1969, and June 1970. The first objective was to study the effects of a reducing environ- ment upon the solubility of P. The second objective was to determine the forms in which P is stored when sorbed by the sediment and the degree to which it might be soluble under a reducing environment. The sedimentdwater interface was analyzed for Eh, P and Fe immediately after sampling and after 30 days of anaerobic incubation. It was found that anaerobic incubation increased the level of P in solution. Regres- = P at sion analysis yielded the equation P = 3.51 PS + .09 where P I S sampling time and PI = P after incubation, indicating that a 3.5 fold increase in P would be predicted during anaerobic incubation. In samples aerobically incubated for 30 days it was found that those originally low in P, below .445 ppm, also showed an increase in P during incubation. For those samples with more than .445 ppm P, aerobic incubation caused a decrease in P solubility. .L_n_ .— . both P view I the re lower Michi; tant of 59 miner accot too t 38111 from drai Cons in 5 S011 the Dennis B. Fenn Redox potentials were shown to have a qualitative influence upon both P and Fe in the lake waters, but the data would not support the view that Eh quantitatively controls the levels of P and Fe in solution. The P forms in the lake sediments were fractionated according to the revised method of Chang and Jackson. Al—P and Fe-P levels were much lower than in most mineral soils. Ca-P levels were similar to many Michigan soils but lower than expected in the marl lake bottoms. Reduc- tant soluble-P predominated in the sediments, reaching a maximum average of 591.9 ppm in Fenton Lake. This is much higher than is found in most mineral soils. This fraction is assumed to be reductant soluble Fe-P according to Chang and Jackson. The levels of total P extracted are much too high to have been contributed simply by the deposition of eroded agricultural from sources drain fields considerable in solution. soluble-P is the reducing soil alone. Considerable P must have been fed into the lakes such as municipal discharges, industrial effluents and septic of bordering homes. Sediments are capable of adsorbing quantities of P and can serve as a buffer upon the level of P The fact that the predominant storage form of P is reductant important because this form can become readily soluble under conditions produced at the bottoms of many lakes. PHOSPHORUS AND IRON CONTENT, REDOX POTENTIAL AND PHOSPHORUS FRACTIONATION OF LAKE SEDIMENT-WATER INTERFACES J By a A, “‘p Dennis BJ’ enn A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of CrOp and Soil Sciences 1970 To JACQUE This thesis is dedicated to my wonderful wife for without her support, understanding, compassion, inspiration, patience and active interest, this study would have been an impossibility. ii .—_._—._—~ The Dr. B. C study. inspirat Sir E. Erick APT Soil Sci growth. The manuSer ACKNOWLEDGEMENTS The author expresses sincere appreciation to his major professor, Dr. B. G. Ellis, for his interest, guidance and support during this study. His knowledge and interest in science has been a great inspiration. Sincere thanks is also expressed to Dr. A. R. Wolcott and Dr. A. E. Erickson for their generous offers of needed laboratory equipment. Appreciation is also expressed to other members of the Crop and Soil Sciences Department for the many opportunities for intellectual growth. Thanks is expressed to my wife for her efforts in typing this manuscript. iii Ill‘lllii INTROD' (LT-1.1.... SUMMA LITEP TABLE OF CONTENTS INTRODUCTION . . . . . . . . . LITERATURE REVIEW . . . . . . Redox Potentials . . . . Phosphorus and Reducing Conditions Sediments and Phosphorus . MATERIALS AND METHODS . . Sediment Sampling . . . . Redox Analysis . . . . . Phosphorus Analysis . . Iron Analysis . . . . . . Phosphorus Fractionation . Controlled Redox . . . . . RESULTS AND DISCUSSION . . . . Phosphorus and Redox Potential Phosphorus and Iron . . . Phosphorus Fractionation Controlled Redox . . . . SUMMARY AND CONCLUSIONS . . . . LITERATURE CITED . . . . . . iv Page 13 13 15 16 17 17 20 22 22 27 39 50 53 55 ll' TABLE 10. ll. 12. l3. l4. 10. ll. 12. 13. 14. LIST OF TABLES Fe, and redox potential of water interface Crooked Lake, July 1969 . Fe, and redox potential of water interface Crooked Lake, October 1969 Fe, pH and redox potential of water interface Crooked Lake, June 1970 . Fe and redox potential of water interface Chemung Lake, July 1969 . Fe, and redox potential of water interface Chemung Lake, October 1969 Fe, pH and redox potential of water interface Chemung Lake, June 1970 . Fe, and redox potential of water interface Ore Lake, July 1969 . . . Fe, and redox potential of water interface Ore Lake, October 1969 . Fe, pH and redox potential Ore Lake, June 1970 . . . Fe, and redox potential of Lobdell Lake, July 1969 . Fe, and redox potential of Lobdell Lake, October 1969 Fe, pH and redox potential Lobdell Lake, June 1970 . Fe, and redox potential of Fenton Lake, July 1969 . Fe, and redox potential of Fenton Lake, October 1969 of water interface water interface water interface of water interface water interface water interface Page 29 29 30 30 31 31 32 32 33 33 34 34 35 35 in. I. ESTOAIIIA LIST ' 15. 16. 17. 18. 19. 20. 21. 22. 23 24. 25. 26 III l I I I I I III I il III III LIST OF TABLES - Continued 15. l6. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. P, Fe, pH and redox potential of water interface of Fenton Lake, June 1970 . . . . . . . . . . . P, Fe, and redox potential of water interface of Ponemah Lake, July 1969 . . . . . . . . . . . . . . P, Fe, and redox potential of water interface of Ponemah Lake, October 1969 . . . . . . . . . . . . P, Fe, pH and redox potential of water interface of Ponemah Lake, June 1970 . . . . . . . . . . . . . . Fractionation of the forms of phosphorus in the sediments of Crooked Lake, July 1969, using the method of Chang and Jackson (1957), revised by Chang (1962) . . . . . . . . . . . . . . . . . . Fractionation of the forms of phosphorus in the sediments of Crooked Lake, October 1969, using the method of Chang and Jackson (1957), revised by Chang (1962) . . . . . . . . . . . . . . . . . . . Fractionation of the forms of phosphorus in the sediments of Chemung Lake, July 1969, using the method of Chang and Jackson (1957), revised by Chang (1962) . . . . . . . . . . . . . . . . . . . . . Fractionation of the forms of phosphorus in the sediments of Chemung Lake, October 1969, using the method of Chang and Jackson (1957), revised by Chang (1962) . . . . . . . . . . . . . . . . . . . . . Fractionation of the forms of phosphorus in the sediments of Ore Lake, July 1969, using the method of Chang and Jackson (1957), revised by Chang (1962) . . . . . . . . . . . . . . . . . . . . . . Fractionation of the forms of phosphorus in the sediments of Ore Lake, October 1969, using the method of Chang and Jackson (1957), revised by Chang (1962) . . . . . . . . . . . . . . . . . . . . . Fractionation of the forms of phosphorus in the sediments of Lobdell Lake, July 1969, using the method of Chang and Jackson (1957), revised by Chang (1962) . . . . . . . . . . . . . . . . . . . Fractionation of the forms of phosphorus in the sediments of Lobdell Lake, October 1969, using the method of Chang and Jackson (1957), revised by Chang (1962) . . . . . . . . . . . . . . . . vi Page 36 36 37 37 41 41 42 42 43 43 44 44 LIST 1 27 . 28 29 30 . 31 32 _.a..H-4TN1 N444 LIST OF TABLES - Continued Page 27. Fractionation of the forms of phosphorus in the sediments of Fenton Lake, July 1969, using the method of Chang and Jackson (1957), revised by Chang (1962) . . . . . . . . . . . . . . . . . . . . . . . 45 28. Fractionation of the forms of phosphorus in the sediments of Fenton Lake, October 1969, using the method of Chang and Jackson (1957), revised by Chang (1962) . . . . . . . . . . . . . . . . . . . . . . . 45 29. Fractionation of the forms of phosphorus in the sediments of Ponemah Lake, July 1969, using the method of Chang and Jackson (1957), revised by Chang (1962) . . . . . . . . . . . . . . . . . . . . . . . 46 30. Fractionation of the forms of phosphorus in the sediments of Ponemah Lake, October 1969, using the method of Chang and Jackson (1957), revised by Chang (1962) . . . . . . . . . . . . . . . . . . . . . . . 46 31. P and Fe analysis of solution after controlled redox incubation of a waterlogged Wisner soil . . . . . . . . . . 52 32. P and Fe analysis of solution after controlled redox incubation of a waterlogged Locke soil . . . . . . . . . . 52 vii 31195!" : 'u-n FIGURE 1. Schem. sampl- Georg 2. Incre Sampl 3. Incre Sampl 4- Deere Samp] 5- Relat Sampj 6' R61a1 July 7' Rela Octo LIST OF FIGURES FIGURE 1. Schematic drawing of sampler used to obtain an undisturbed sample of the sediment—water interface (developed by Dr. George H. Lauff, Gull Lake Research Center) . . . . . . . Increase in P during anaerobic incubation. Fenton Lake. Sampled July 1969 . . . . . . . . . . . . . . . . . . . Increase in P during aerobic incubation. Crooked Lake. sampled Ju1y 1969 I I I I I I I I I I I I I I I I I I I Decrease in P during aerobic incubation. Ponemah Lake. sampled Ju1y 1970 I I I I I I I I I I I I I I I I I I I Relationship between P and Redox Potential. Ore Lake. sampled OCtOber 1969 I I I I I I I I I I I I I I I I I I Relationship between P and Fe. Chemung Lake. Sampled Ju1y 1969 I I I I I I I I I I I I I I I I I I I I I I I Relationship between P and Fe. Fenton Lake. Sampled octOber 1969 I I I I I I I I I I I I I I I I I I I I I I viii Page 14 24 25 26 28 . 38 40 .——-_— .— .— —~ Th has com is shov this t1 has gr in the found P leve restr: uses. scum undes and 1 COnd. Amer nOt mor-E disc mine raw tre INTRODUCTION The pollution of our environment, especially our lakes and streams, has come to the point where practically every body of water in America is showing signs of eutrification. The awakening public concern about this tragic occurrence and its potential effect on the human population has greatly increased the need to determine the chief factors involved in the degradation of our aquatic environment. Phosphorus (P) has been found to be a very limiting nutrient in the eutrification process. When P levels in solution are low, algal blooms and higher plant growth is restricted—~the water remains pleasing for recreational and municipal uses. But as the input of P increases, water becomes choked with algae scum and weed growth and the fish population degrades to scavengers and undesirable species--the water is no longer of great value to man. The huge amounts of P fertilizer applied to our agricultural lands and the potential erosion of our fertile topsoil each year has brought condemnation upon agriculture as a chief cause of the pollution of America's waterways. The research conducted up to the present time has not totally relieved this condemnation, but it has shown that an even more important source of pollution is our industrial plants and their discharges. However, the chief source of P in our waters has been deter- mined to be the sewage effluents of our cities, many of whom still dump raw, untreated sewage into streams and rivers. Most present day sewage treatment plants do not have the capacity to adequately remove the levels inn-".— .— .— “—— .— -_ __ ‘- —_ .—-— _— -— _ .— of P t1 charges are frc P04, a1 Pl very 1] adds 1 runoff suffic have n be 105 condit by prc lower Winte large that the 2. of P that are in the sewage. The average person in this country dis- charges 3.5 lbs P/year. Of this amount, 1 lb is from waste and 2.5 lbs are from detergents. Detergents on the market today contain from O - 53% P04, and water softeners such as calgon contain up to 70% P04. Pollution by P from agriculture is still very real. Although P is very immobile in the soil and ground water seepage into lakes and streams adds little total P, areas of intense irrigation have large volumes of runoff which may contain P concentrations of up to 0.08 ppm P which is sufficient to support an extensive algal bloom. Studies over many years have noted that as much as 82% of the P added as fertilizer may ultimately be lost by erosion on some soils (Taylor, 1967), but this is an extreme condition. Most agricultural land today is well protected against erosion by proper conservation practices, and the losses of P by erosion are much lower than in Taylor's study. Losses from cattle feedlots, especially in winter when runoff is increased due to the frozen ground, can be very large and very damaging to the water environment. The objectives of this study were to: (1) study the effects of a reducing environment upon the solubility of P; (2) determine the forms that most P is stored in when sorbed by sediments; and, finally (3) study the influence of laboratory controlled redox potentials upon the solubility of P. Th a soil with, a conclug redox I J: and re. Yield Soil. Valua matior the Qt deVe1T LITERATURE REVIEW Redox Potentials The application of the concept of oxidation—reduction potentials to a soil system has been tried by numerous researchers; but they have met with, at best, only limited success. As will be seen below, contrasting conclusions have been reached by various workers as to usefulness of redox measurements as a valid analytical tool. Jeffery (1961) found that he could differentiate between oxidized and reduced soils with Eh measurements, but the error was too great to yield precise information about the state of reduction of the waterlogged soil. He also felt that the large experimental error made it of little value to correct for pH change. He suggested that a more accurate esti— mation of the state of reduction of a soil could be obtained by determining the concentrations of the two oxidation states of iron in the soil. He developed the equation Eh - 1.032 — 0.0601 log CFe - 0.180pH (30 C) as being applicable to waterlogged soils. In another paper, Jeffery (1961) expanded his equation and ran experi- ments in an effort to define oxidizing conditions, healthy reducing condi- tions, and extreme reducing conditions in quantitative terms. Using the above equation, he developed the term rh = Eh + 0.180pH and calculated the values rh:>l.34 volts for oxidizing conditions, rh = 1.27 - 1.21 volts for healthy reducing conditions and rh<;l.15 volts for extreme reducing conditions. J! l 3“ —" 5 the qu as a f of the (2) ur is ope transi datio: betwec of a‘ tion. 0.059 and F that the 1 SOiL meas was trOd it w long dqu e1e( the Ponnamperuma, Tianco, and Loy (1967) found, in a 17 week study, that the quantitative treatment of redox equilibria in a complex system such as a flooded soil was difficult because: (1) the highly dynamic nature of the flooded soil prevents the attainment of a stable, true equilibria; (2) uncertainty about which of a large number of possible redox systems is operating at a given time; (3) the lack of thermodynamic data on the transition substances present in the soils that undergo reversible oxi- dation-reduction; (4) complex formation which may alter redox equilibria between inorganic ions; and (5) the uncertainty about the true potential of a reduced soil--the soil potential or the potential of the soil solu- tion. However, they were able to determine that the equation Eh = 1.058 - 0.059 log Fe - 0.177pH held for most of the 17 weeks of submergence. They proposed that the participation of the metastable compounds Fe(OH)3 and Fe3(OH)8 in the equilibria was confirmed by their findings, indicating that the soil solution was the thermodynamically meaningful phase. Yamane and Sato (1968) conducted a series of experiments to determine the best way to get reproducible values of redox potentials of submerged soils. They found that the area of the electrode was related to the measured Eh values. The smaller the area of the electrode, the slower was the soil Eh value reached. Therefore, they felt that platinum elec— trodes should be at least 100 mm2 in area. Even with such large electrodes, it was necessary for the electrode to be in contact with the soil for longer than 6 hours and in most cases at least 24 hours to obtain repro- ducible results. Bohn (1968) attempted to relate the emf of gold, graphite and platinum electrodes in soil suspensions to electrode potentials of redox couples in the soil but was unsuccessful. He found that EAu and EC did not respond to aeration of the suspension or to the presence of sucrose as a reducing agent. Ept did respond to these conditions but did not correspond to the electrode potentials of the Mn+2 Mn oxide, 02 — H20, or H+ - H2 couples. He concluded that Ept is a mixed potential whose major com- ponents are the 02 - H20 couple in aerated suspensions, but whose value is not related to the Nernstian distribution of oxidized and reduced species. Although E is a qualitative measure of the oxidation, it has pt no quantitative meaning according to Bohn. He used a 1:10 soil-water suspension in his experiments; however, and at the low redox potentials measured, there was not enough total manganese in the soil to respond quantitatively to the potential as predicted by the Nernstian expression. This could also account for why no appreciable quantitative relationship was found. In a review of the topic of redox potentials, Bohn (1970) notes that the limit of oxidizing conditions in an aqueous system is the oxidation of water to molecular oxygen (02) and that the limit of reducing condi— tions in an aqueous system is the reduction of hydrogen ion to molecular hydrogen (H2). The redox potential or electron availability affects the oxidation states of H, C, N, O, 8, Mn, Fe, Co, and Cu in aqueous systems. But for the oxidation-reduction limits upon the stability of water, the list could be extended to include the entire periodic table. Bohn states that the reason redox measurements have been found want— ing in soil systems is because the soil redox potential is often a non- equilibrium potential rather than an equilibrium potential upon which the Nernst equation is based. Natural systems rarely reach oxidation-reduction equilibrium because of the continual addition of oxidizable organic matter. Being a mixed potential, redox potential measurements are quantitatively unrelated to the Nernstian distribution of ion oxidation states. In oxidized systems, the low concentration of redox couples decreases the stability, reproducibility and general usefulness of redox potential measurements. In reduced systems, the higher concentrations of redox couples increase the stability and utility of redox measurements. It is apparent that Bohn disagrees with the conclusions of Jeffery and Ponnamperuma that quantitative expressions can be applied to redox measurements in a soil system. They all agree, however, that certain ionic substances respond to oxidation and reduction potential changes. The debate remains in whether or not equilibrium is ever obtained between the redox couples and Eh. Phosphorus and Reducing Conditions Bartholomew (1931) found that flooding a soil caused a disappearance of soluble soil P and suggested that it may have been due to the reversion of soluble inorganic P to an organic form that is not available to plants. He noted an increase in water soluble organic P after three months of flooding which could have been produced by anaerobic bacterial action. The water used for irrigation was high in Ca(HC03)2. The amounts of soluble Ca, Fe, and Al added annually by the irrigation water were suf— ficient to have caused the reversion of large amounts of soluble P. Gasser (1956) noted that waterlogging soils caused an increase in acid soluble P. This increase could be attributed to the reduction of ferric phosphate to ferrous phosphate. Gasser proposed that in assessing the P status of rice soils, the ferric phosphate should be considered as an "available" form. Hayes and Phillips (1958) reported that the redox potential had little or no effect upon the level of P in solution. They felt that the biological - 7;; I'di‘W-w—uu—v L._.. system was was seconc and admin‘ 12 of the but, in 4 that thes system up Shap availabil flooded c in native both the to a mucl availabi] in Fe—P ; addition Mant °f P in . increase. Change 1 increase He State decOmpos soluble I deCIEase the effe. Manc phOSPhOrC system was the controlling factor and that the inorganic chemical system was secondary in importance. After waterlogging a series of soil samples and administering antibiotics to quench the biological system, 10 out of 12 of the soil samples showed an increase in P under reducing conditions; but, in 4 out of the 10, the difference was less than 5%. They concluded that these results indicated the minor influence of the inorganic reducing system upon P levels in solution. Shapiro (1958) showed that flooding caused an increase in soil P availability. He noted that applied P was utilized more efficiently under flooded conditions. These results held for soils high, moderate, and low in native P. His data showed that the increased P availability came from both the Fe—P and Al-P fractions but that the Fe-P fraction was affected to a much greater degree. He proposed that the small increase in Al-P availability was probably due to a chelation reaction. The large increase in Fe—P availability was due to the reducing conditions brought on by the addition of organic matter and flooding the soil. Mandal (1964) tested the effects of starch and lime on the availability of P in a waterlogged soil and found that waterlogging a soil only slightly increased acetic acid soluble P with a slight decrease in Fe-P and no change in the Al- or Ca-P. When starch was added, however, a considerable increase in acid soluble P and a decrease in the Ca-P fraction occurred. He stated that the release of large amounts of C02 formed during starch decomposition may have caused tricalcium phosphate to convert to more soluble di- and mono- calcium phosphates. The addition of lime caused a decrease in Fe— and Al-P, especially Fe-P. Ca-P was increased appreciably, the effect seeming to be a conversion from Fe-P to Ca-P upon liming. Mandal suggested that in acid soils having most of their inorganic Phosphorous in the ferric form, liming followed in a few days by the addition of orga conditi He fractic the A1: high eT increa: reduct reduct if it almost I Fe We] Fe pr. mostl extra -200 millj cate: the 1 Was COHd the of organic matter will result in an increase in P under waterlogged conditions. He further found that even after 105 days of waterlogging, the Al-P fraction remained unchanged. The Ca-P fraction showed results similar to the Al, except in very high organic matter soils where C02 evolution was high enough to be important. His data did not support the view that the increased availability of soil P upon waterlogging is largely due to the reduction of ferric phosphates. The available data suggested that the reduction of ferric phosphates did not occur to any great extent or that if it did proceed, a reversion reaction with Fe+3 in the soil occurred at almost an equal rate. Patrick (1964) found that extractible ferric and ferrous forms of Fe were very sensitive to changes in redox potential of the soil. Ferric Fe predominated at potentials above +200 millivolts, and extractible Fe was mostly ferrous below +200 millivolts. At the same time, he found that extractible P increased over threefold between redox potentials +200 and ~200 millivolts. The sharp break in the phosphate release curve at +200 millivolts, the same point at which ferric Fe began to be reduced, indi- cates that the conversion of P to an extractible form is dependent upon the reduction of ferric compounds in the soil. Broeshart, Haunold, and Fried (1965) noted that flooding significantly increased the availability of soil P in rice soils in which free CaCO3 is absent. The reduced availability of phosphates under upland conditions was not enough to account for the reduced growth of rice under upland conditions in their opinion, however. They also found that the availability of fertilizer P was similar under reduced or oxidized conditions but that ‘the efficiency of use was greater under flooded conditions. Williams and Simpson (1965), in conducting experiments on culti- vation and waterlogging, found that waterlogging for l - 2 days produced anaerobic conditions that caused a decrease in P availability and an increase in the sorption capacity. This reduced P availability applied both to soil P and applied P. They concluded that the decrease in P availability upon waterlogging was due to the effects of certain reactions during the waterlogging treatment. Reducible metals such as Mn could have interfered; the nature of the sorption sites may have been altered due to Fe reduction causing P to be more tightly bound; or, some of the P sorbed could have been occluded by reprecipitation of Fe upon restoration of aerobic conditions. They felt that the increases in soil P availability upon waterlogging as noted by several researchers, must be due to the presence of easily reducible ferric phosphate. In soils low in ferric phosphates, a decrease in P availability should be noted. Chiang (1968) noted that soils rich in organic matter showed rapid lowering of Eh and increased formation of organic acids and gases. Dis- integration of soil aggregates and lowered Eh values caused an increase in P solubility. These effects were more significant in Fe rich soils. H2, H28 and the organic acids, especially HOAc, increased P solubility in Fe and A1 rich soils. CO2 increased Ca-P solubility. Chakravarti and Ghoshal (1968) mixed two acid soils with 50% dried grass and each of the following treatments: (1) Fe—P; (2) Al-P; or (3) Fe- and Al-P. Then they waterlogged and incubated the samples at room temperature for 53 days. The amount of P released was found to be greatest in the Fe phosphate treatment followed by Fe and A1 phosphate treatment, Al phosphate, and the control in decreasing order. I‘ll“ 0f the 5' “191- 5.0 1Q Patr that the ditions 4 When pla greatly been red The mech tion of ; occluded Fe- and I (Phosphal that resc can Occu Fur weeks a sorrel; was at 8112a Al 3: COns aCcc ment: maximm true f0] 21 10 Patrick and Mahapatra (1968) and Mahapatra and Patrick (1969) showed that the greatest change in a P fraction between a soil under flooded con- ditions and under aerobic conditions occurs in the reductant soluble fraction. When placed under waterlogged conditions the reductant soluble fraction was greatly decreased, indicating that the ferric oxide coating on the P had been reduced to a soluble ferrous oxide and the P released to the solution. The mechanism of P release in a flooded soil may be explained by: (1) reduc- tion of ferric phosphate to soluble ferrous phosphate; (2) release of occluded P by reduction of ferric oxide coating; (3) displacement of P from Fe- and Al-P by organic anions; (4) hydrolysis of Fe- and Al-P; and (5) anion (phosphate) exchange between clay and organic anions. They warn, however, that resorption processes and refixation by unreduced ferric Fe processes can occur and waterlogging will not always increase available P. Furukawa and Kawaguchi (1969) submerged paddy soil samples for two weeks at 40 C and obtained up to a 21% decrease in organic P. This decrease correlated well with the increase of easily soluble P (Bray No. 2) and was attributed to the mineralization of organic P. This increased miner- alization was thought to be due to the enhanced solubility of the Fe or Al salts of inositol hexaphosphoric acid, the predominant organic P constituent, by reducing conditions or pH rise, and rapid hydrolysis accompanying submergence. Terman, Allen, and Engelstad (1970) conducted greenhouse pot experi- ments with flooded rice on a soil which was low in available P for upland crops. Marked yield responses by rice to applied P were obtained, but maximum yields were obtained at much lower rates of applied P than was true for most upland crops. Response to applied P decreased with liming of the soil and with increasing levels of acid soluble soil P. Granular, water soluble sources were most effective. The P in Fe-P was more 11 available than that in Al-P in the flooded soil. Both forms were more available in fines than as granules and in colloidal form rather than as fine crystals. Sediments and Phosphorus Hayes and Phillips (1958) proposed that there is a single pool of P belonging to lake water and solids which is distributed between them in a dynamic equilibrium or steady state. This dynamic equilibrium might be represented as: P in aqueous phase, a small.‘L___ P in solid phase, a large fraction of the whole -———77 fraction of the whole, with a constant value for each phase but a continuous exchange between them. The level of productivity, the state of oxidation and reduction, and the presence of green plants and bacteria all work together to produce the observed equilibrium. Hayes and Phillips concluded that bacteria are decisive and can, to a considerable degree, suppress the classical inorganic mechanism of oxidation-reduction that has been proposed by many as the controlling factor in P release and availability. Pomeroy, Smith, and Grant (1965) found that the exchange of P between water and sediment was a combination of a sorption reaction and a biologically controlled exchange. Their observations supported the view that sediments act to buffer the P content of water. They estimated the daily exchange across an undisturbed sediment boundary to be llqmole PO4/m2. For sedi- ments being stirred by wave action, the exchange is greater. In their opinion, this rate of release from the sediments is sufficient to support a continuous growth of plankton and that it is a continuous and rapid . ‘_(~‘ A as... release tha1 Barter to adsorb P water ratic the sedimeT the freshlj that the Ni loosely bo- sive water sorb much a few hour influxes o Frink highly cor increases, "atEI and fouud tha and had c The Ca-p 0.52. FT lake Sam; (Silt, c] This aCCC in Physic transport PhOSphate5 dePOSited 12 release that is not responsive to certain P limits within the water. Harter (1968) noted that lake sediment possessed a tremendous capacity to adsorb P from solution. He added P to a sediment sample (1:50 soil- water ratio) and allowed equilibrium to be reached. He then extracted the sediment with 0.5N NHAF to extract Fe-P and found that about half the freshly adsorbed P could be recovered in this way. He also discovered that the NHAF extraction contained a considerable amount of labile or loosely bound P and that this labile P could also be removed by succes- sive water extractions. His conclusions were that the sediment could sorb much of the P in this labile form and that it can be re-released in a few hours, days, or weeks for use by plants and algae. Hence, large influxes of P may be intercepted and stored for use over a period of time. Frink (1969) found that many of the sediment characteristics were highly correlated with depth of water. P increases in solution as depth increases. A correlation coefficient of 0.68 existed between depth of water and total P in solution. In his fractionation of the sediment, he found that Al-P, reductant soluble P, and Fe-P all increased with depth and had correlation coefficients of 0.80, 0.85, and 0.90 resPectively. The Ca-P fraction decreased with depth, however, and had a correlation of 0.52. Frink found that Ca—P correlated with the coarser sediments. The lake samples showed a particle size gradation of coarse (sandy) to fine (silt, clay, organic matter) from the bank to the center of the lake. This accounts for the decrease in Ca-P with depth. Most of the changes in physical characteristics are attributed to particle size sorting during transport and deposition within the lake. He noted a shift of stored phosphates from Al-P and Fe—P to Ca—P when sediments from acid soils were deposited in neutral lakes. MATERIALS AND METHODS Sediment Sampling The necessity of obtaining a relatively undisturbed sample of the sediment-water interface that would be suitable for analysis posed some interesting problems. It had to be accomplished without introducing air into the sample or the redox readings would be invalid. After reviewing the literature on sampling methods and personally contacting several agencies involved in water quality research, it was decided that a small sampler developed at the Gull Lake research station would best suit our purposes. A schematic drawing of the sampler is shown in Figure 1. The irrigation valve allows water to flow through as the sampler is dropped. This helps keep the sampler upright and also insures that the water picked up comes from the mud-water interface and not some other point nearer the surface. This valve seals when the sampler is lifted toward the surface. This prevents mixing and also forms a suction that holds the sample in the tube as the apparatus is hoisted to the surface. The sampler is small enough that the entire operation can be performed from a canoe. Once at the surface, the sampler is placed onto a number 13 stOpper which has been mounted on a bolt in the bottom of a galvanized bucket. This stopper will go inside the sampler head and lodge in poly— ethylene tube. Then the sampler is disassembled leaving only the sample in the polyethylene tube mounted on the stopper. The tube is carefully 13 14 ~—--——‘ U-bolt 3" nipple --— — one-way irrigation valve 1---- 4" nipple i {may polyethylene sample tube *---———»2" diameter galvanized steel pipe tooled head Figure 1. Schematic drawing of sampler used to obtain an undisturbed sample of the sedimentdwater interface (develOped by Dr. George H. Lauff, Gull Lake Research Center). 3“” 15 removed from the stopper and the sample transfered to a glass container. In this study, pint canning jars that had a rubber serum stopper tightly placed in the cap were used. The jars were filled with sediment and water in about equal volume and tightly capped. Within a few hours after sampling, the samples were returned to the laboratory for analysis. A 50 ml volume of solution was extracted by inserting a hyperdermic needle into the jar. Another needle connected to a nitrogen tank was inserted and nitrogen gas bubbled into the jar to fill volume left by extracted water. The redox potential, pH, phosphorous, and iron were then rapidly analyzed on the sample. Six lakes of diverse water quality characteristics were sampled in central—lower Michigan. It was suspected that conditions would get more reducing with depth, so each lake was sampled at 5 or 6 depths. Each lake was sampled in the spring or early summer and again in the fall to determine if seasonal variations were important. The fact that this sampler could meet the criteria of taking an undisturbed mud core and the water just above it and still be used from a canoe and lowered and raised by hand using a 1/4 inch nylon rope made it the ideal apparatus for this research. Redox Analysis The redox potential was measured using a Sargent model DR pH meter. .A.l cm2 bright platinum electrode was used along with a saturated calomel electrode to obtain the measurement. The polarity was determined by the Ilse of a standard Weston cell with a 1,083 millivolt emf. Each reading Dvas taken after a one minute equilibration period. In the final sampling, 16 June 1970, the redox potential and pH were measured in the field using a Beckman model G portable pH meter. Phosphorus Analysis A chlorostannous—reduced molybdophosphoric blue color method in a sulfuric acid system was utilized in this study, Jackson (1958). It is believed that P serves as a central coordinating atom in the formation of heteropoly complexes with molybdate ions. These complexes are slightly yellow but appear colorless in lower concentrations. The addition of a reducing agent will bring about a reduction in the phosphomolybdic complexes yielding a blue color that can be measured colorimetrically at 660 mg'. The concentration of P, molybdate ion, pH, and reductant must be carefully controlled to take advantage of a narrow region where only the heteropoly complex is reduced and not the excess molybdate or an appreciable amount of interfering ions such as arsenic (As). This optimum plateau is found to occur in this method when the phosphorous concentration is less than .4 ppm and the pH is about 0.5. This chlorostannous-reductant molybdophosphoric blue color method in a sulfuric acid system is the most sensitive analytical procedure for P and as such is well suited for analysis of infertile soils and low P waters. In systems where excessive Cl, As, Fe, and other complex forming ions are found, alternate analytical procedures should be employed. An aliquot of 1 - 25 ml of sample was taken and 2 ml of 2.5% sulfo- molybdic acid added, and the sample was diluted to 48 ml. Then .2 ml of .1N SnC12 was added to develop the color. The sample was diluted to 50 m1 and the absorbence determined on an Evelyn photoelectric colorimeter at (560 nu(. 17 Iron Analysis The orthophenanthroline analysis for iron was used in this work. 2 This procedure is based on the principle that Fe+ will form a stable, red complex with orthophenanthroline. The tri-(l,10) phenanthroline ferrous ion Fe(C H N )2+ 12 10 2 3 over two weeks. NHZOH'HCl is added first to reduce Fe+ forms almost immediately and is stable for 3 to Fe+2 and then the orthophenanthroline is added to develop the color. The reaction follows Beer's Law up to 6 ppm and the absorption peak is at 508 ml(' The color is stable within a pH range of 2 to 9, but for most soils the lower range is desired. This method is practically free from interference. Many constituents can be present in 200 - 500 times the concentration of Fe and not interfere with the color stability. An aliquot of 10 - 25 ml was taken, 2 ml 5N NH OAC and 1 ml 10% 4 NHZOH'HCl added, and the sample was shaken. Then 1 ml orthophenanthroline reagent was added along with 0.5 ml of 6N HCl. The sample was mixed, diluted to 50 ml and the adsorbence determined at a wavelength of 520 m;¢. Phosphorus Fractionation Following the completion of the analysis of the water, each sample was drained and air dried for a week, finely ground in a mortar, mixed, and placed back in the sample jar. A one gram sample of the air dried sedi- ment was placed in a 90 ml plastic centrifuge tube where the P forms were separately extracted and analyzed according to the method of Chang and .Jackson (1957) and Chang (1962) The exchange: added an: remove t? effluent 10 ml of were the solution 660 mg . The of satur. 0n the 5" Was cent' washed :1 The effl' remOVed ' CEntrifu ChlorOSt acid Sys Nex Nazszo4 water he a 100 ml pol'tiOns effluent A 5 ml a of diSti added. 18 The sample was shaken with 50 ml 1N NH4C1 for one hour to remove exchangeable Ca and water soluble P. Then 50 ml of .5N NH4F, pH 8.2 was added and the samples placed on a wrist action shaker for one hour to remove the Al—P fraction. The sample was centrifuged and 10 ml of the effluent was placed in a 50 ml volumetric flask; 15 ml of 0.8M boric acid, 10 ml of chloromolybdic acid solution, and 5 ml of chlorostannous reductant were then added, the sample being well mixed after each addition. The solution was brought up to volume and analyzed on the colorimeter at 660 m;(. The soil sample in the tube was then washed twice with 25 ml portions of saturated NaCl. Then 50 ml of 0.1N NaOH was added, and it was placed on the shaker for twelve hours to remove the Fe—P fraction. The sample was centrifuged and the supernatant removed. Again the soil sample was washed twice with 25 ml portions of saturated NaCl which were discarded. The effluent was usually darkly colored by organic matter. ,This was removed by adding 2 ml of 2M H2804, two drOps of concentrated H2804 and centrifuging. Then a 25 m1 portion of the effluent was analyzed by the chlorostannous—reduced molybdophosphoric blue color method in sulfuric acid system as previously discussed. Next, the soil sample in the tube was extracted with 1 gram of solid Na28204 and 40 m1 of 0.3M Na3C6H50 ‘ZHZO with constant stirring in a 90 C 7 water bath for 15 minutes and centrifuged. The Supernatant was placed in a 100 m1 volumetric flask and the soil was again washed twice with 25 ml portions of saturated NaCl. The NaCl washings were then added to the effluent in the flask and the contents brought up to volume and mixed. A 5 m1 aliquot of this solution was placed in a 250 m1 conical flask. 10 m1 of distilled water, a drop of 0.5M FeC13 and 10 ml of P—free 30% H20 were 2 added. The flask was placed over a bunsen burner and the oxidation allowed . - y ”a ‘I.’ $1" .wm ' placed back halOM: tate. The 250 ml {135. fuged and were broug Fe—P, was blue metho Final to the soij Analysis w. in sulfuri This some soils Yields re: Preper ex important the best. di‘calcit Phosphatt of NH4F 1 i°n inflt eXtract. NaOE iOn 0H7 r is eXclus 19 to go to completion. The sample was then taken to dryness on a steam bath. 10 m1 of 2N NaOH was added, the solution boiled for two minutes, and was placed back on the steam bath for five minutes. The sample was then placed in a 10 ml centrifuge tube and centrifuged to throw down the Fe precipi- tate. The supernatant liquid was placed in a 50 ml volumetric flask. The 250 ml flask was washed with 10 ml distilled water and the solution centri- fuged and added to sample. This process was repeated twice and the contents were brought up to volume and mixed. This fraction, the reductant soluble Fe-P, was then analyzed by the chlorostannous-reduced molybdophosphoric blue method in sulfuric acid system procedure. Finally, the Ca-P fraction was extracted by adding 50 ml of 0.5N H2804 to the soil sample in the tube, shaking for one hour and centrifuging. Analysis was by the chlorostannous-reduced molybdophosphoric blue method in sulfuric acid system. This fractionation procedure is not without its drawbacks, and for some soils it will not yield valid results. For most soils, however, it yields results quite close to actual fact. The order of extraction, the proper extraction, solid to solvent ratio, and extraction time are all important considerations. The 50:1 solvent to solid ratio has proven the best. The NH4C1 wash removes easily soluble phosphates such as di-calcium phosphate. NH4F dissolves aluminum phosphates readily, calcium phosphates hardly at all, and iron phosphate slightly. By raising the pH of NH4F to 8.2, the solubility of Fe-P is greatly reduced. The common ion influence of F‘ hinders the solubility of the apatite form in this extract. NaOH readily dissolves both Fe— and Al-P, but the high pH and common ion OH" restricts apatite solubility. By removing Al-P first, this step is exclusive for Fe—P. 20 The reductant soluble P is thought to be protected by a coating of Fe oxide which resists NHAF and NaOH solution. NaZSZO4 and Na3C6HSO7°2H20 reduce this coating of ferric oxide to a soluble ferrous oxide form and dissolve and extract the entrapped P forms. Sulfuric acid will dissolve apatite, Fe-P and Al-P. If the latter two forms have already been preferentially removed, this treatment will serve as a good extractant of the Ca-P fraction in the soil sample. Controlled Redox A system was designed to control the redox potential in a flooded soil sample by taking a 17 inch long piece of clear plastic tubing, 2 inches in diameter, and stoppering the ends. The t0p stopper was fitted with a platinum electrode, saturated calomel electrode, a gas inlet tube, and a gas outlet tube. A sample containing 600 grams of soil and 600 ml of water was placed in the tube with the gas inlet tube extending to the bottom of the column to insure complete mixing. The electrodes were hooked to a Coleman titrion automatic titrator and companion pH meter. The instrument was modified to deliver nitrogen (N2) or 02 gas instead of liquid titrant. To obtain reducing conditions, the titrator was hooked to a N2 tank. Bubbling N2 through the sample de-aerated the system and brought the desired reducing conditions. Bubbling 02 would bring oxi—