Abstract: Removal of fertilizer phosphorous by adsorption and precipitation processes is an important factor for yield loss in low input soils of arid and semi-arid regions. The phosphate sorption characteristics of surface and subsurface horizons of four semi-arid region soils and their relationships with soil chemical and physical properties were investigated. Two sorption sites were identified. The adsorption capacity was greater for region 1 while bonding energy was higher for region 2 sites. The adsorption maxima (b1) of the low concentration range (region-I) was 625-1250 μg P/g and adsorption energy coefficient k1 ranged between 0.159-0.800 mL μg-1. The calculated b and k values of high concentration range (region-II) were not treated as real adsorption parameters, because as the concentration increased, the data weakly confirmed Langmuir isotherm. This revealed the fact that Langmuir isotherm can be applied to phosphorous adsorption in soils and soil materials to a limited extent. Adsorption maximum of soils for P were found to be greater for soils and horizons high in CaCO3, clay and CEC. Statistically significant positive relationships were found between adsorption maximum and CaCO3, clay and CEC whereas important negative relationships were found between adsorption capacity and organic matter, sand, Fe2O3 and Al2O3 content of soils.
INTRODUCTION
The reaction chemistry of phosphorous is very complex and it is very difficult to manage the behavior of P in arid and semi arid soils. Soil components such as clay, various pools of iron oxides (Castro and Torrent, 1998; Ryan et al., 1985; Li et al., 2007; Saavedra and Delgado, 2005), soil carbonates (Carreira et al., 2006; Hamad et al., 1992; Tunesi et al., 1999) and organic matter (Li et al., 2007) play very important role in the adsorption chemistry of P in calcareous soils. The mechanism of the adsorption of phosphate to CaCO3 has been previously studied (de Kanel and Morse, 1978; Millero et al., 2001; Stumm and Leckie, 1970). Earlier studies suggested that the initial uptake of phosphate on calcite occurs by chemisorption, followed by a slow transformation of amorphous calcium phosphate to crystalline apatite. Other studies aimed at defining the role of surface area and distinguishing between total and active carbonates have redefined the role of this mineral phase and have indicated that the low density of calcium atoms at the calcite surface reduces the importance of this mineral in P precipitation processes (Borrero et al., 1988).
Tunesi et al. (1999) reported that below an initial concentration of 0.5 mM the adsorption processes were predominant while above this level precipitation became predominant in calcareous soils. They also concluded that the surfaces of carbonate minerals were not necessary for induction of P precipitation. Ca-P precipitation can be governed by exchangeable-Ca ions at a convenient pH with increasing reaction time. On the other hand, surfaces of calcite are highly dynamic, even when it is exposed to air and able to incorporate adsorbed material into near-surface bulk (Stipp et al., 1996). This nature of calcite surfaces can continuously create new sorption sites that may have important implications in the behavior of adsorbed P in arid climate soils with high carbonate content.
The sorption of P in Vertisols and Inceptisols were highly correlated with CDB-extractable Fe and clay (Solis and Torrent, 1989). Similarly, Li et al. (2007) found that single point sorption index of purple soils significantly correlated with clay content. The P sorption maximum of pure CaCO3 was reported to be 9.2-25 mg P kg-1 (Hamad et al., 1992; Griffin and Jurinak, 1974) whereas limestones showed much higher P sorption maximum between 105-124 mg kg-1 due to presence of impurities, particularly Fe (Holford and Mattingly, 1975). Ming et al. (2002) pointed out that the P adsorption capacity of the organo-mineral colloidal complexes differed with the soil types, being higher for the acid and calcareous purple soils than for the neutral purple soils and partial removal of the organic matter increased the adsorption capacity of the colloidal complexes.
The aims of this study were to investigate sorption characteristics of surface and subsurface of four great groups developed on the alluvial and calcareous parent material over a very large concentration range of phosphate (0-300 μg mL-1), that enables both adsorption and precipitation and to define the relationships between Langmuir parameters and soil properties.
MATERIALS AND METHODS
Soils
The top two horizons of a Fluventic haploxeroll (A), a Typic xeropsamment
(B), a Typic calcixerept (C) and a Calcic haploxerept (D) from Ankara
(Turkey) were used in this study. The soils are derived from calcareous
and alluvial materials and occur at elevation about 800 m (Dengiz, 2002).
Climatic conditions in the area are semi-arid with 450 mm rainfall and
10-13°C average temperature. Mineralogical analysis showed that chlorite
was the dominant phyllosilicate in the Typic xeropsamment and Fluventic
haploxerepts whereas; calcite, quartz, feldspar and 2:1 type smectite
minerals were dominant in clay fraction of the Typic calcixerepts and
Calcic haploxerepts.
The samples were air-dried, gently crushed and passed through a 2 mm sieve. Particle size distribution was determined by hydrometer method (Gee and Bauder, 1986). Soil chemical properties were determined following standard procedures (Sparks, 1996), vis, pH in 1:2.5 soil: water suspension, electrical conductivity in saturation paste, organic matter by K2Cr2O7 oxidation, CaCO3 equivalent by a manometric method, available P by extraction with 0.5 M NaHCO3 at pH 8.5 (1 g: 20 mL), CEC by saturating the exchange sites with 1 M NaCH3COO at pH 8.2 then replacing the Na with M NH4CH3COO at pH 7.0. Total free Fe and Al oxides were extracted with citrate-dithionite buffer (Mehra and Jackson, 1960) and determined by AAS. The data are shown in Table 1.
Phosphate Sorption
P-sorption isotherms of soils were determined by equilibrating 1.000
g soil with 40 mL of 0, 10, 25, 50, 75, 100, 150, 200, 250 and 300 μg
P (as KH2PO4) mL-1 solution prepared
in 0.01 M CaCl2 for 24 h at 25°C. An equilibration time
of 24 h was chosen for experimental convenience. Any period ranging 6-48
h would ensure completion of the rapid reaction at the surface sorption
sites, without a significant contribution from the subsequent slow reactions
in which diffusion into the solid phase is likely to occur. At the end
of shaking process, the supernatant solutions were separated by
| Table 1: | Some physical and chemical properties of soils |
| 1: Typic xeropsamment (0-18 cm), 2: Typic xeropsammnet (18-47 cm), 3: Fluventic haploxerepts (0-25 cm), 4: Fluventic haploxerepts (25-65 cm), 5: Typic calcixerepts (0-13 cm), 6: Typic calcixerepts (25-60 cm), 7: Calcic haploxerepts (0-16 cm) and 8: Calcic haploxerepts (16-44 cm) | |
centrifugation and filtration through Whatman No. 42 filter paper. Then, the equilibrium P concentration was determined by spectrophotometry (Murphy and Riley, 1962). The amount of sorbed-P (x/m) was calculated from the difference between the amount of P added and that which remained in the equilibrium solution (C).
The data were fitted to the linear form of the Langmuir adsorption isotherm:
C/(x/m) = C/b+1/kb
Where:
b = The monolayer adsorption maximum (μg g-1)
k = The adsorption energy coefficient (mL μg-1)
Since very large initial concentration range (0-300 μg mL-1) was used, the sorption isotherms were split into two different regions by step wise regression analysis (Steel and Torrie, 1960). Adsorption was the dominant processes at the lower part whereas; precipitation was likely to occur at upper site of the isotherms as indicated by apparent decrease in the C/x ratio.
Correlation analysis between the Langmuir parameters and soil characteristics was performed by using SPSS.
RESULTS AND DISCUSSION
The soil samples from the four great soil groups with two horizon
almost completely adsorbed all of the applied phosphorous from the solutions
in the 0 to 150 μg mL-1 concentration range. As the phosphorus
concentration was increased, however, each great soil group developed
a different equilibrium concentration. The plots of C versus C/(x/m) for
great soil groups are shown in Fig. 1. There are two
main parts in the sorption isotherms where adsorption and precipitation
driven sorption mechanisms are displayed. Langmuir parameters of the soils
for different soil depth and isotherm regions with the regression coefficient
are shown in Table 2. Sorption maxima (b1) for region
1 ranged between 625 (Typic xeropsamment) to 1250 μg g-1
(Calcic haploxerepts) and for region 2 between 2702 (Fluventic haploxerepts)
to 25000 μg g-1 (Typic calcixerepts). Amrani et al.
(1999) reported a maximum adsorption range of 146 to 808 μg P g-1
for Moroccan calcareous soils. The data better conformed the theoretical
Langmuir isotherm in the low concentration range (region 1), that was
in agreement with findings of other researchers (Derici and Kaya, 1991;
Derici and Agca, 1991). On the other hand, P adsorption isotherm can be
better obtained at an initial concentration below 15 mg L-1
(Bertrand et al., 2003). But in present case due to common agricultural
practices which is rain-fed agriculture with minimal input management,
natural P content of the soils can be minimal (Table 1).
| Fig. 1: | Langmuir sorption isotherms of surface and subsurface horizon of different great soil groups. Arrow shows separation between adsorption and precipitation-driven reaction of phosphate | ||||
In fact, this consequence leads soils to adsorb more P than high input
soils. High input soils usually demonstrate monolayer adsorption maxima;
therefore, small additions of P make the soil attain maximum adsorption.
At high P concentration range (>150 mg P L-1) the precipitation
process was the dominant mechanism and the data in this range somehow
conformed to the Langmuir isotherm with smaller regression coefficient
than the first part. It is reported that precipitation driven reactions
can
| Table 2: | Parameters of Langmuir isotherms |
| Table 3: | The correlation coefficients between Langmuir parameters (b, k) and some soil properties |
| *: Significant at p<0.05 and **: at p<0.01 | |
be successfully described by the Langmuir isotherm (Veith and Sposito, 1977). There was a very strong deviation in the border line of the two process with an apparent increase before precipitation of any solid phase which was indication of the supersaturation of solution P against relevant solids or decrease after precipitation of a discrete solid phase. In general the mineralogy of the soils was likely an important factor in sorption maximum. Soils containing chlorite mineral (Typic xeropsamment and Fluventic haploxerepts) sorbed more P than montmorillonitic soils (Typic calcixerepts and Calcic haploxerepts). Mollisols and vertisols dominated by 2:1 clay minerals (e.g., montmorillonite) were reported to have relatively lower sorption capacities (Sims and Baker, 2003).
Correlation between soil properties such as clay content, oxides of Al and Fe, CaCO3 equivalent, organic matter and P sorption parameters of different theoretical models have been widely used to determine the main soil components in P behavior (Brennan et al., 1994; Bertrand et al., 2003). The relation between Langmuir parameters and some physical, chemical and mineralogical properties of experimental soils were investigated through correlation analyses and the results are given in Table 3. Correlation analysis revealed that adsorption maxima in region-1 were positively affected by CaCO3 (r2 = 0.863**) and clay (r2 = 0.649*) content whereas Fe2O3 (r2 = -0.648*), Al2O3 (r2 = -0.639*) and sand (r2 = -0.625*) content had a negative significant effect.
The contribution of exchangeable Ca-ions to Ca-phosphate precipitates was reported to be higher than that of CaCO3 (Akinremi and Cho, 1991a, b; Tunesi et al., 1999) but the role of CaCO3 in replenishing Ca in soil solution and exchange sites and calcification process during soil formation should not be discarded. On the other hand, above 150 mg L-1 initial concentrations (region-II) adsorption maxima was correlated with CaCO3 (p<0.01). This suggests that phosphorous was first precipitated with CaCO3 as thermodynamically less stable discrete solid phase (i.e., di/tri calcium phosphate) with subsequent formation of more stable phases such as apatite (Lindsay, 1979; Carreira et al., 2006). In addition, it was reported that majority of applied P can precipitate as poorly soluble Ca-phosphates in low P containing calcareous soils (Delgado and Torrent, 2000) such as the experimental soils.
Oxides of Fe and Al are important cementing agent in soils that reduce the total surface area thus sorption sites despite their very high adsorption ability in pure systems. The negative correlation between oxide content and sorption parameter may be attributed to alkaline pH that reduce the solubility of oxides while increasing the stability of surface coating on other soil particles and aggregating effect and oxides of Al and Fe approach to the point of zero charge, therefore the affinity of oxide surfaces decreases for P adsorption (McBride, 1994). On the contrary, a significant correlation between P retention capacity and a soil property does not necessarily imply a significant direct effect of the soil property on P retention (Ige et al., 2007). For example sand content, which have indirect effect, usually reduces the exchangeable Ca content of soils (Ige et al., 2007). Negative correlation between organic matter content and adsorption maxima could be related to competition between low molecular weight organic acids and phosphate for sorption sites that usually occurs in favor of organic acids and delays P adsorption (Geelhoed et al., 1999; Staunton and Leprince, 1996; Violante and Gianfreda, 1993). The coating of high affinity sorption sites by organic matter and organo-mineral complex formation could be more significant as organic matter content of soil increased and in especially fine textured clay soils. When different horizons are taken into consideration, horizons high in clay and CaCO3 and low in sand, adsorption maximum was significantly high. These findings are in agreement with the reason given above.
CONCLUSIONS
Phosphorus added to calcareous soils was removed from soil solution by the combination of adsorption and precipitation processes. At lower end of the sorption isotherm, the adsorption was the dominant process whereas higher concentration addition promoted precipitation. Initial P concentration above 150 μg mL-1 precipitation-driven reaction was dominant. In general, P sorption characteristics of arid region soils are likely related to the amount of clay and reactive oxides or carbonates present in the soil and are influenced by the physical and chemical characteristics of the solid phase (e.g., mineralogy and crystallinity). Soils containing chlorite clay mineral sorbed more P than montmorillonitic soils. Oxides minerals and organic matter are negatively correlated with the sorption parameters. It can be concluded that P sorption characteristics along with the soil properties should be considered for economically and environmentally friendly fertilization.