Sorption capacity of phosphate in mineral soils II Dependence of sorption capacity on soil properties

The dependenceof the indicator of phosphate sorption capacity on extractable Al and Fe and other soil properties was studied in a material consisting of 102 mineral soil samples. The sum of P adsorbed on soil during two days from a solution containing P 5 mmol/1 and P extracted by 0.02 M EDTA (pH 5.3) as an estimate of the initial P content in the soil was used as the indicator of P sorption capacity. In clay and silt soils (n =51), the Al and Fe extracted by 0.05 M oxalate (pH 2.9) together with the organic C content explained 85 Vo, the Al and Fe extracted by 0.05 M K 4 P2 07 (pH 10) together with the clay content 87 Vo, the Al and Fe extracted by 0.02 M EDTA (pH 5.3) 91 Vo, and the Al extracted by 1 M CH 3COONH4 (pH 4.8) together with the organic C and clay contents 78 Vo of the variation of the indicator of phosphate sorption capacity. In coarse soils (n =51), the variation of the indicator was explained well only by oxalate-extractable metals, which together with soil pH and clay content explained 80 Vo of the variation. Extractable Al was generally the most important explainer of variation. The results suggest that forms of extractable Al and Fe explaining the variation of the indicator of P sorption capacity in clay and silt soils are partially different from those in coarse soils. Index words: acetate-extractable AI, EDTA-extractable Al, Fe and P, oxalate-extractable A 1 and Fe, pyrophosphateextractable A 1 and Fe, soil pH, clay content, organic carbon content


Introduction
In the previous paper (Niskanen 1990), the phosphate sorption capacity of some mineral soils was estimated by means of sorption isotherms.When sorption properties of large soil material are to be studied, it is, however, laborious to determine complete sorption curves.Bache and Williams (1971) proposed the use of a single value sorption index to characterize the phosphate sorption properties of soil.Sorption determined from only one concentration may, however, be unsatis- factory if the initial phosphate content in soil is high (Barrow 1978).In such cases it may be valid to correct the sorption results by ad- ding an estimate of the initial P content.The aim of this paper was to study the dependence of the corrected indicator of P sorption ca- pacity, determined by the one-point method, on extractable Al and Fe and other soil properties.

Material and methods
The material consisted of 51 clay or silt soils and 51 coarse soil samples, which have been presented more thoroughly in a previous paper (Niskanen 1989) (Table 1).The sam- ples were air-dried at room temperature and ground to pass through a 2-mm sieve.The pH of the soil was measured in a soil-0.01M CaCl 2 suspension (1:2.5) (Ryti 1965).The particle-size distribution of inorganic matter was determined by the pipette method (Elonen 1971), the organic carbon content by the Alten wet combustion method (Graham   1948).Soil aluminium and iron were extracted by the following methods (Niskanen 1989): 0.05 M oxalate (pH 2.9, ratio 1:20 w/v, shak- ing time 2 h), 0.05 M K 4P 2 0 7 (pH 10, 1:100 w/v, 3 h) and 0.02 M Na 2 -EDTA (pH 5.3, 1:50 w/v, 3 h).Aluminium was additionally extracted by 1 M ammonium acetate (pH 4.8, 1:10 w/v, 2 h) (McLean et al. 1958).The metals contained in filtrated extracts were de- termined by atomic absorption spectrophotometry.
To determine the phosphate sorption, 5 g of soil was treated at + 20°C for two days with 100 ml of solution containing KH 2 PG 4 5 mmol/1.The ionic strength of the solution was 0.01, adjusted with KCI 5 mmol/1.To in- hibit microbial activity, the solution contained 0.01 % NaNj.The suspensions were daily shaken for eight hours.At the beginning and the end of the experiment, the pH of the sus- pensions was measured.The phosphorus con- centration in filtrates was determined by a modified molybdenum blue method (Kaila 1955).The amount of retained phosphate was calculated as the difference between the phos- phate quantity present initially and that re- maining in the supernatant.The experiment was carried out in triplicate.
Phosphorus extracted by 0.02 M Na 2 -EDTA and determined by the molybdenum blue method was used as the estimate of the initial phosphorus content in soil.The sum of re- tained and initial phosphate content was used as the indicator of the P sorption capacity of the soil.

Results
The phosphate concentration (5 mmol/1) used to determine the sorption of phosphate on experimental soils was the same as Bache  and Williams (1971) used in determination of the P sorption index.The initial pH of the soil suspensions was 4.0-6.6 and after the sorption, lasting two days, the pH was slightly higher, the increase being no more than 0.5 pH units.The experimental soils sorbed, on average, a little more than 20 % of the added P (100 mmol/kg soil) (Table 2).The sorption was higher in clay and silt soils than in coarser soils.The material consisted of 23 samples, mainly clay and silt soils, which ad- sorbed more than 30 % of the added P. In all, 20 samples adsorbed less than 10 % of added P; these were mainly coarser soils.There were two coarser soil samples which released P rather than adsorbed it.The phosphorus extracted by EDTA was used as an estimate of the initial content of the adsorbed P in the soil (Table 2).EDTA extracted, on average, more P from surface soils than from subsoils, and more P from clay and silt soils than from coarser ones.In sur- face soils, the EDTA-extractable P seemed to increase with increasing pH (r = o.6B***,n = 53).In surface layers of clay and silt soils, the correlation coefficient between extractable P and soil pH was o.Bo*** (n = 27).In coarse surface soils, the extractability of P increased with an increasing clay content (r = o.69***,n = 26).
EDTA-extractable P is thought to be con- nected with inorganic P fractions in soil which are bound to aluminium and iron (Alexan- der and Robertson 1972) and also to calcium (Ahmed and Islam 1975, Sahrawat 1977,  Hartikainen 1979).Several studies (Alexander and Robertson 1972, Ahmed and Islam   1975, Olsen 1975, Sahrawat 1977, Onken et   al. 1980) verified that acidic EDTA solution extracts P which is available to plants.
The indicator of P sorption capacity (Table 2), which includes sorbed and EDTAextractable P, amounted, on average, to about 25 mmol/kg soil.The mean value of the indi- cator was higher in clay and silt soils than in coarser soils.
The dependence of the indicator of P sorp- tion capacity on soil properties was studied by means of linearregression analysis.The vari- ables were as follows: X, = indicator of P sorption capacity (mmol/kg soil) X 2 = extractable aluminium (mmol/kg soil) X 3 = extractable iron (mmol/kg soil) X 4 = organic carbon content (%) X 5 = soil pH X 6 = clay content (%) In clay and silt soils, oxalate-extractable aluminium and iron together with the organic carbon content explained (P = 0.001) 85 % of the variation of the indicator of P sorption ca- pacity.The regression equation and partial correlation and P coefficients were as follows; X, = -8.13+ 0.293X 2 + 0.117X 3 + 2.01X 4 r 12.34 (3,2.34 = 0-64 r ,3.24 = 0-56*** (3,3.24= 0.27 r 14.23 = o.46***(314.23 = 0.31 The content of oxalate-extractable Al ex- plained 54 °7o, the content of iron 31 % and the content of organic carbon 21 % of the var- iation in sorption when the effect of the oth- er two independent variables was eliminated. Oxalate-extractable Al and Fe, soil pH and clay content explained (P = 0.001) 80 % of the variation in the indicator of P sorption capacity in coarse soils, the regression equation being: X,= -3.33 + 0.185X 2 + 0.094X 3 -3.18X 5 + 0. The content of oxalate-extractable Al explained 76 °7o, the content of Fe 19 %, the soil pH 19 % and the clay content 11 % of the variation when the effect of the other independent variables was eliminated.In both soil groups, oxalate-extractable Al was the most important explainer of the variation in the in- dicator of P sorption capacity.
Acetate-extractable Al and the organic car- bon and clay contents explained (P = 0.001) 78 % of the variation in the indicator of P sorption capacity in clay and silt soils accord- ing to the regression equation: X, = 5.50 + 0.859X 2 + 0.272X 4 + 0. Al explained 65 %, the organic carbon con- tent 30 % and the clay content 12 % of the variation when the effect of the other two independent variables was eliminated.In coarse soils, acetate-extractable Al explained (P = 0.001) 36 % of the variation in the indi- cator according to the equation: X[ = 8.75 + 1.112X 2 .

Discussion
In the present soil material, AI and Fe ex- plained the variation of the indicator of P sorption capacity.Aluminium was generally a more important explainer than iron.In numer- ous papers concerning the dependence of the P sorption on soil properties (e.g.Williams et al. 1958, Kaila 1959, 1963, Bromfield  1964, 1965, Saini and MacLEAN 1965, Ahen-  korah 1968, Lopez-Hernandez and Burnham  1973, 1974, Hartikainen 1979, Wada and  Gunjioake 1979), extractable aluminium has been more important than iron in explaining the retention of P. In sorption studies, the Freundlich constant k, the anion exchange ca- pacity (Piper 1944) and the P sorption index (Bache and Williams 1971) are used as indi- cators of P sorption capacity, and metals are extracted by oxalate, dithionite, HCI, ammo- nium acetate and acetic acid solutions.In the study of Kaila (1963), oxalate-extractable Al and Fe explained nearly 80 % of the variation of the Freundlich constant k in clay soils and about 60 % in coarse soils.The partial corre- lation coefficient for the relationship between k and Al was o.7B*** (n= 151) in clay soils and o.63*** (n= 109) in coarse soils, and for the relationship between k and Fe o.46*** in clay soils and o.3o*** in coarse soils.Kaila's results show the same tendency as those ob- tained in the present investigation.
In some cases iron has been found to be a more important explainer of P sorption than aluminium (Kaila 1963, Bromfield   1965, Ahenkorah 1968, Lopez-Hernandez  and Burnham 1973, 1974).In these studies, however, Al is often extracted by a different method than Fe.According to Lopez-Her-  nandez and Burnham (1973), extractable Al and Fe explain the sorption of P the better the more efficient the extraction method of metals is.
The greater significance of soil aluminium than soil iron in the adsorption of phosphate may be related partially to the different way these metals occur in soil.Iron oxides and hydroxides are generally present as discrete mineral particles, even when they are present on surfaces of clay minerals (Deshpande et  al. 1964, Greenland et al. 1968).Aluminium hydroxides tend to form films over clay par- ticles (El Swaify and Emerson 1975), offer- ing a large surface area for phosphate sorp- tion.In the present study the significance of a larger sorption surface was manifested by the higher sorption of phosphate in clay and silt soils than in coarse soils.
In this study, the organic carbon content ex- plained the variation of the indicator of P sorption capacity in clay and silt soils.In many earlier papers, P sorption has been found to depend on the content of organic matter in soil (Williams et al. 1958, Saini  and MacLEAN 1965, Ahenkorah 1968, Lopez-  Hernandez and Burnham 1974).This indi- cates that active Al and Fe are closely con- nected with soil organic matter, which retards the crystallization of oxides and thus enhances their activity in sorption (Williams et al.   1958, Schwertmann et al. 1968).In coarse soils of the present material, pH, clay content and oxalate-extractable Fe were about equally important explainers of the variation in the indicator of P sorption capacity.In most of the previously mentioned studies, the dependence of the indicator of P sorption on soil pH and clay content is weak or insignificant.
The present study showed that the forms of extractable Al and Fe explaining the variation of the indicator of P sorption capacity in clay and silt soils are partially different from those of coarse soils.In clay and silt soils, the vari- ation of the indicator was rather well explained by the metals extracted by all of the methods studied, whereas in coarse soils, the variation in the indicator of P sorption capa- city was explained well only by oxalate-ex- tractable metals.In clay and silt soils, the in- dicator of P sorption capacity, in particular, seemed to be related to the EDTA-extractable metals, whereas in coarse soils this fraction of metals explained the variation in P retention only weakly.It has been mentioned before that particularly in clay soils, Al hydroxides occur as films on clay particles, thus providing large surface for retention of P. It may be that EDTA extracted such hydroxide films, whereas EDTA was able to extract only poorly the more crystalline forms, which may con- stitute a considerable proportion of the P retaining material in coarse soils.

Table I .
Soil characteristics.