FORMS OF NEWLY RETAINED PHOSPHORUS IN MINERAL SOILS Armi Kaila University of Helsinki, Department of Agricultural Chemistry Received January 11, 1964 In most soils a large part of the added water-soluble phosphate will be rapidly removed from solution by soil constituents. The mechanism involved may be ad- sorbtion or chemical precipitation, and the products formed are likely to represent a large variety of compounds with aluminium, iron, or calcium, and their oxides and hydroxides. Usually, more than 80 per cent of the added soluble phosphate is fixed as aluminium and iron compounds, and only in soils with a high calcium carbonate content retention as calcium compounds may be marked (8, 12, 14). In the present paper an attempt is made to study the distribution of applied water-soluble phosphate in the various fractions of inorganic phosphorus in Finnish mineral soils, over a relatively short period of contact. The fractionation procedure by Chang and Jackson (3) is employed. The possible connection is tried to be found between the net increase in the various phosphorus fractions and soil proper- ties such as the contents of active aluminium and iron, pH, the content of organic carbon, the phosphate fixing capacity, and the degree of phosphorus saturation. Methods Soil pH was measured in a 1 to 2.5 suspension in 0.01 M CaCl2 by the glass electrode. The content of organic carbon was determined by the Walkley method using the iodometric titration. Aluminium and iron were extracted by Tamm’s acid ammonium oxalate solution in the ratio of soil to solution of 1 to 20, the period of extraction was two hours. Aluminium was determined by the Aluminon method, and iron by the sulfosalicylic acid procedure, after the organic matter in the extract was destructed by ignition. The indicator of the phosphate sorption capacity, k, was estimated by a somewhat modified method of Teräsvuori (6) which is based on the Freundlich adsoption isotherm. The degree of satura- tion y O/k, is, accordning to Teräsvuori, the ratio between the »exchangeable» P, or P extracted by alkali, and the indicator of the sorption capacity. The soil samples were treated with water-soluble phosphate in the following way: I g of air-dried and ground soil was weighed in a centrifuge tube, and 50 ml of KH 2P0 4 solution containing 6 mg P/l was added. The tube was shaken by machine for 4 hours, let stand over night, and shaken for 2 hours https://www.c-info.fi/en/info/?token=B1vkXkBmUTyQag9W.Ly6ygJH0BFdtnYmBGFPyig.tEFq_FZQfUbyRmuNbbfWIMQ4gxhqSvFOVa7WLE9TW76T6OHpXX-glpe3A6UgWQsejgTwtjsTuSsXmq4EQMVEEZSwuQZGCOloay51cFiLynoremi9hxXWUCGSp5hTzsmlVPg25pIonomA4kOxxmnQmR_ncCcKDBK8 66 before centrifuging. The supernatant solution was decanted, and its content of P determined. The moist soil in the centrifuge tube was let stand for 3 days at room temperature, washed with saturated NaCl- solution, and successively extracted with alkaline 0.5 N NH,F, 0.1 N NaOH, and 0.5 N H 2 S04 according to the procedure by Chang and Jackson (3). Untreated samples were fractionated in the same way, and the difference between the phosphorus contents of the treated and untreated samples was taken to represent the increase due to the added P. Soil samples 180 samples of mineral soils were collected from various parts of the country both from the plough layer, or the corresponding layer in the virgin soils, and from the deeper layers between the depths of 20 cm and 70 cm. Since only 21 samples were from the virgin lands, they were not treated separately. According to the results of the mechanical analysis, the samples were divided to the groups of sand and fine sand soils with less than 30 per cent clay and more than 50 per cent fine sand, loam and silt soils with less than 30 per cent clay and less than 50 per cent fine sand, and the group of clay soils with more than 30 per cent clay. The number of samples in the various groups and the means of the pH- values, the contents of organic carbon, and oxalate soluble aluminium and iron are listed in Table 1. The low acidity of some glacial clay soils of the deeper layers increases the mean pH of the groups of subsoil samples and clay soil samples. In the other groups the mean pH value is rather low. Table 1. Soil samples Number Org. C Oxalate soluble Al* Fe* mmol/kg mmol/kg of pH* samples %* Topsoil Sand and fine sand 45 5.2 ± 0.2 3.7 ± 0.7 95 ± 15 67 ± 9 Loam and silt 38 5.2 ± 0.2 3.8 ± 0.6 101 ±l7 81 ± 838 5.2 ± 0.2 3.8 ± 0.6 101 ±l7 81 ± 8 Clay 47 5.2 ± 0.2 3.8 ± 0.4 123 ± 15 121 ± 18 All 130 5.2 ± 0.1 3.7 ± 0.4 107 ± 9 90 ± 8 Subsoil Sand and fine sand 17 5.2 ± 0.2 0.8 ± 0.3 138 ± 52 67 ± 20 Loam and silt 10 5.3 ± 0.4 0.7 ± 0.3 80 ± 42 70 ± 2210 5.3 ± 0.4 0.7 ± 0.3 80 ± 42 70 ± 22 Clay 23 5.6 ± 0.3 0.4 ±O.l 78 ± 8 56 ± 8 All 50 5.6 ± 0.2 0.6 ± 0.1 99 ± 20 63 ± 8 All Sand and fine sand 62 5.2 ± 0.2 Loam and silt 48 5.2 ± 0.2 107 ± 18 67 ± 24 96 ± 15 78 ± 7 108 ±ll 99 ± 14 48 5.2 ± 0.2 Clay 70 5.5 ± 0.2 * Means with the confidence limits at the 95 per cent level 67 In the topsoil samples, the content of oxalate soluble aluminium and iron tend to increase from sand and fine sand to clay soils, but in the subsoil samples the sand and fine sand group has a higher mean content of aluminium than the clay soils have. The phosphorus condition of the soil groups is characterized by the data in Table 2. In the topsoils, the phosphate sorption capacity tends to be highest in the clay soils and lowest in the loam and silt soils. In the subsoil samples, some of the sand and fine sand soils have a markedly high capacity. The average degree of saturation with phosphate is almost equal in all the groups of topsoil samples, on one hand, and in all the groups of subsoil samples, on the other hand, with a statisti- cally insignificant tendency for the highest value in the loam and silt soils. In the sand and fine sand soils the ammonium fluoride soluble inorganic phos- phorus (»Al-P» in Table 2) tends to be higher than in the other soils, and not markedly lower than the alkali-soluble phosphorus (»Fe-P»). This fraction is highest in the clay soils from the surface layer. The acid-soluble phosphorus (»Ca-P») tends to be lower Table 2. Indicator of phosphate sorption capacity, k, degree of saturation, yO/k, and fractions of inorganic P (ppm) in the soil groups. (Mean values with the confidence limits at the 95 per cent level.) k Yo/k »Al-P» »Fe-P# »Ca-P» Topsoil Sand and fine sand 249 ± 38 0.76 ± 0.13 98 ± 29 139 ± 34 150 ± 30 Loam and silt 199 ± 27 0.78 ± 0.11 62 ± 21 142 ± 25 188 ± 40199 ± 27 0.78 ±O.ll 62 ± 21 142 ± 25 188 ± 40 Clay 337 ± 55 0.74 ± 0.06 72 ± 15 218 ± 30 194 ± 27 All 267 ± 26 0.76 ± 0.06 78 ± 13 169 ± 18 177 ± 18 Subsoil Sand and fine sand 319 ± 134 0.43 ± 0.10 58 ± 20 76 ± 25 136 ± 65 Loam and silt 176 ± 116 0.50 ± 0,15 22 ± 16 102 ± 84 325 ± 14;176 ± 116 0.50 ± 0.15 22 ± 16 102 ± 84 325 ± 143 Clay 223 ± 34 0.35 ± 0.05 15 ± 3 98 ± 15 315 ± 52 All 246 ± 52 0.41 ± 0,05 31 ± 9 88 ± 18 256 ± 47 All Sand and fine sand 268 ± 45 0.67 ± 0.10 85 ± 22 119 ± 26 146 ± 27 Loam and silt 194 ± 30 0.72 ± 0.10 54 ± 17 134 ± 25 216 ± 43194 ± 30 0.72 ± 0.10 54 ± 17 134 ± 25 216 ± 43 Clay 300 ± 40 0.61 ± 0.06 53 ± 12 175 ± 25 234 ± 28 in the sand and fine sand soils than in the other groups. The typical differences between the forms of phosphorus in the topsoils and subsoils is evident: the topsoils contain more fluoride-soluble and alkali-soluble phosphorus than the subsoils, but less acid-soluble phosphorus. 2 68 Fractions of sorbed phosphate The amount of phosphorus retained by the soil samples during the treatment with KH 2P0 4-solution was calculated on the basis of the phosphorus concentration in the solution removed by centrifuging (Table 3.). In the group of the topsoil samples this amount was highest in the clay soils, but in the subsoil samples a fairly high retention occurred also in the group of the sand and fine sand soils. On the average, about one half of the 250 mg of phosphorus added per kilogram of soil was retained during the period of contact of 24 hours. Very little phosphorus was removed by the washing with saturated NaCl which means that after the further three days of contact only a small part of the added phosphate remained in easily soluble forms. Only in the groups of the clay soils, the total amount of phosphorus recovered by the fractionation is somewhat lower than the amount calculated on the basis of the phosphate concentration in the solution removed. Table 3. Phosphorus sorbed by the samples (Means with the confidence limits at the 95 per cent level) Sorbed P Sorbcd P ppm % extracted by of added P NH.-F NaOH H.,5() 4 Topsoil Sand and fine sand 41 ± 6 62 ± 11 40 ± 10 3 ± 3 Loam and silt 44 ± 5 64 ± 12 38 ± 6 5 ± 3 Clay 56 ± 5 71 ± 9 61 ± 11 3 ± 3 All 47 ±3 67 ±6 47 ±6 4±2 Subsoil Sand and fine sand 55 ± 14 88 ± 24 43 ± 13 5 ± 5 Loam and silt 43 ± 16 58 ± 23 48 ± 22 3 ± 7 Clay 57 ±6 61 ±5 57 ±8 5±5 All 54 ± 14 70 ± 10 51 ± 7 5 ± 3 Sand and fine sand 45 ± 6 69 ± 10 41 ± 8 4 ± 2 Loam and silt 44 ± 5 63 ± 10 40 ± 6 5±3 Clay 57 ± 4 68 ± 6 60 ± 8 4 ± 2 In all the soil groups the largest part of sorbed phosphorus is extracted by ammonium fluoride, but in the groups of clay soils and in the subsoil samples of loam and silt Soils the alkali-soluble fraction is not statistically significantly lower than the fluoride soluble fraction. The amount of phosphorus left in the acid-soluble fraction is low. In about two thirds of the samples no increase in the phosphorus content of this fraction could be detected as a result of the treatment with the phosphate solution. 69 No statistically significant difference occurs between the various kinds of soils, or between the samples of topsoil and subsoil, in the amount of sorbed phos- phorus extracted by fluoride. The amount of sorbed phosphorus in the alkali extract is highest in the clay soils, although their superiority in this respect is not statisti- cally significant in the subsoil samples. Thus it seems that the various soils accumulate about the same amounts of applied phosphorus in the fraction supposed to be bound by aluminium, and the more effective sorption of the clay soils is due to a larger accumulation of iron bound phosphorus. In all the samples, both topsoil and subsoil, on the average, about 56 per cent of the sorbed phosphorus is in the fluoride soluble fraction and about 40 per cent in the alkali-soluble forms. In the various soil groups this percentile distribution is the following: topsoil samples fluoride-soluble P alkali-soluble P sand and fine sand 59 % 38 % loam and silt 60 % 36 % clay 53 % 45 % subsoil samples sand and fine sand 65 % 32 % loam and silt 53 % 44 % clay 50 % 46 % Also these figures show the relative great importance of iron in the retention of phosphorus by clay soils as compared especially with the retention by the sand and fine sand soils. Cable i Ratio between acid oxalate loluble aluminium and iron and between the amount« of phos- phorus extracted by NH,F and NaOH (Means with the confidence limits at the 95 per cent level) »Al-P»/»Fe-P» Al/Fe initial sorbed total Topsoil Sand and fine sand 1.5 ± 0.2 0.7 ± 0.1 2.5 ± 0.7 1.1 ± 0.2 Loam and silt 1.2 ± 0.1 0.4 ± 0.06 1.8 ± 0.3 0.8 ± 0.1 Clay 1.1 ±O.l 0.3 ± 0.08 1.5 ± 0.4 0.6 ± 0.1 All 1.3 ± 0.1 0.5 ± 0.06 1.9 ± 0.3 0.8 ± 0.1 Subsoil Sand and fine sand 2.0 ± 0.6 0.7 ± 0.2 21 ± 0.8 1.2 ± 0.3 Loam and silt 1.1 ± 0.4 0.2 ± 0.16 1.4 ± 0.3 0.7 ± 0.2 Clay 1.5 ± 0.2 0.2 ± 0.02 1.2 ± 0.2 0.6 ± 0.2 All 1.6 ±0.2 0.4 ±O.l 1.5 ± 0.3 0.8 ± 0.1 All Sand and fine sand 1.7 ± 0.2 0.7 ± 0.1 2.5 ± 0.5 1.2 ± 0.2 Loam and silt 1.2 ± 0.1 0.4 ± 0.06 1.7 ± 0.2 0.7 ± 0.1 Clay 1.2 ±O.l 0.3 ± 0.06 1.4 ± 0.3 0.6 ± 0.1 70 It is of interest to study in more detail the distribution of the added phosphorus between the fractions which are supposed to be bound by aluminium or iron. The ratios between these fractions in the original samples and the samples treated with phosphorus were calculated and are listed in Table 4, which also contains the ratios between the aluminium and iron (as mols) extracted by acid ammonium oxalate. In accordance with some previous results (7), the ratio of fluoride soluble phosphorus to alkali soluble phosphorus is in the untreated soils rather low, and in the sand and fine sand soils significantly higher than in the soils of the finer texture. The same order between the soil groups may be found also in respect to the ratios of these fractions in the samples treated with phosphorus and the ratios of the net increase in these fractions due to the treatment. The ratio of the sorbed »Al-P» to the sorbed »Fe-P» is in all soil groups markedly higher than the correspond- ing ratio in the untreated samples, thus indicating that the distribution of the applied phosphorus in these fractions is essentially different from the distribution of the inorganic phosphorus which has been in contact with the soil for a longer period. The ratio of the mols of the oxalate soluble aluminium to iron is highest in the sand and fine sand soils which offers an explanation for the corresponding order in the ratios of aluminium bound and iron bound phosphorus in the soil groups. Relation between some soil properties and fractions of sorbed phosphorus The amount of applied soluble phospate retained in the fluoride-soluble forms is likely to depend on the content of active aluminium and its compounds in the soil. The total linear correlation coefficients between the sorbed »Al-P» and acid oxalate soluble A 1 (Table 5) prove that there is this kind of relation also in the present material, although the correlation is close only in the subsoil samples. The corresponding connection between the sorbed »Fe-P» and oxalate soluble Fe Table 5. Total linear correlation coefficients between sorbed »Al-P» or »Fe-P» and contents of oxalate soluble sesquioxides, organic carbon and pH Correlation coefficient Group of samples Sand and r between Topsoil Subsoil fine sand Loam and silt Clay sorbed »Al-P» and Al o.s7*** o.B7*** o.69*** o.73*** o.sl*** pH -0.25** -0.21 -0.14 -0.35* -o.4o*** C % o.49*** 0.23 0.14 0.38* o.46*** sorbed »Fe-P» and Fe o.72*** 0.33* o.49*** o.s9*** o.66*** pH -o.46*** -0.08 -0.28* -o.ss*** -o.sl*** C % 0.21* 0.07 0.19 0.13 0.09 * Significant at 5 per cent level. ** Significant at 1 per cent level. *** Significant at 0.1 per cent level. 71 Tiililc 8. Coefficients of determination, r 2, and multiple determination,R2 , for the relationship of sorbed »Al-P» and »Fe-P» with other soil properties Group of soil Samples Sand and Loam and Topsoil Subsoil fine sand silt Clay r 2 A i_i. A| 0.33 0.76 0.48 0.53 0.26 R 2 Al.,> A] ~,, 0.34 0.77 0.49 0.56 0.29 R« A 1.,, A 1 _H ,• 0.38 0.77 0.50 0.56 0.30 r 2 Ke. P Ke 0.51 0.11 0.24 0.35 0.43 R 2 Fe-P. Fe nil 0.61 ° l2 °-30 °-B1 °- 47 R 2 l-c-l'. IV pll C °- 61 ° l2 °- 50 or>6 o,il is very weak in the subsoils, but fairly distinct in the topsoils. In some groups a weak negative correlation exists between the amounts of sorbed phosphorus and soil pH, but the low correlation coefficients between the sorbed phosphorus and soil organic carbon content are positive. According to the coefficients of determination and multiple determination, the part of the variation in the amounts of sorbed »Al-P» or »Fe-P» which may be explained on the basis of the variation in these three soil properties differs markedly in the various soil groups. The figures in Table 6 show that while in the subsoil samples as much as 77 per cent of the variation in the amount of sorbed »Al-P» may be explained by the content of A 1 and pH, taking into account also the content of organic C determines only 38 per cent of it in the topsoil samples, and in the clay soils this part remains as low as 30 per cent. It seems that the variations in pH and the content of organic C are of more importance in explaining the variation in the sorbed »Fe-P» than in the sorbed »Al-P». Yet. the highest part of the variation in the former which is explainable by the variation in Fe, pH and organic C is only 61 per cent, and in the subsoils this part is almost negligible. To find some of the factors on which the distribution of applied phosphate in the aluminium bound and iron bound fractions may depend, the relation between the ratio of sorbed »Al-P» to sorbed »Fe-P» and the ratio of acid oxalate soluble A 1 to Fe, pH, the content of organic carbon, the indicator of phosphate sorption capacity, k, and the degree of saturation, y„/k, was studied. The total linear cor- relation coefficients computed indicated that there is a positive correlation between the ratios »Al-P»/»Fe-P» and Al/Fe, the coefficients of correlation ranging from 0.43** in the group of loam and silt soils to o.66*** in the group of topsoil samples. No total linear correlation could be found between the ratio »Al-P»/»Fe-P» and the other characteristics studied. Further statistical treatment showed, however, that in some of the soil groups, the phosphate sorption capacity is of certain importance in determining the distribution of the sorbed P between the fluoride-soluble and alkali-soluble fractions. The coefficients of determination and multiple determina- tion in Table 7 indicate that in some soil groups taking in account the value k, in 72 Table 7. Coefficients of determination, r 2, and multiple determination, R 2, for the relation between the rations of sorbed »Al-P»/*Fe-P» and total »Al-P»/»Fe-P» with other soil properties Group of soil samples Sand an