Retention of phosphorus on calcite and dolomite: speciation and modeling

Abstract

The intensive application of phosphate fertilizers in agriculture has created an important source of diffuse phosphorus pollution. The interaction of phosphorus with carbonate minerals plays a role in the fate and transport of phosphorus in soil. The object of the present study was to investigate the speciation of phosphorus on two common carbonate minerals, calcite and dolomite, using a combination of batch experiments, ATR-FTIR spectroscopy, XANES analysis, and diffuse layer modeling. Within the pH range 6.0–7.0, the retention of phosphorus by calcite and dolomite is mainly attributed to the formation of amorphous calcium phosphate (Ca₃(PO₄)₂, ACP), dibasic calcium phosphate (CaHPO₄·2H₂O, DCP), and hydroxyapatite (Ca₅(PO₄)₃OH, HAP). At pH ≥ 8.0 the immobilized phosphorus takes the form of complexes =CaPO₄Ca⁰/=sCaPO₄Ca⁰ on the surface of calcite, followed by the formation of Ca–P phases, including ACP, DCP, and HAP, with increasing phosphorus levels (>2 mg L⁻¹). However, the dolomite surface is initially dominated by the adsorption complex =MgHPO₄Ca⁺ at =Mg sites, and at higher phosphorus levels it then grows due to Ca–P phases and the formation of newberyite (MgHPO₄). It is interesting to note that the Mg content in dolomite favors the rapid growth of DCP at phosphorus levels >200 mg L⁻¹. As a result, at pH ≥ 8.0, dolomite shows a stronger capacity for immobilizing phosphorus than does calcite. Dolomite therefore serves as a better phosphorus sink than calcite in calcareous soil environments.

---

1. Introduction

The intensive application of chemical fertilizer in soils has been reported throughout the world.¹ In particular the excessive use of phosphate fertilizers in agriculture has increased the risk of phosphorus in water runoff, which can result in an excess of nutrients (eutrophication) in the water body.² Meanwhile, a number of researchers have focused on the chemical speciation of phosphorus in soils, determining its bioavailability.³ The reactivity of high phosphorus levels with soil minerals is one of most important factors leading to phosphorus solubility and speciation in surface or subsurface soils. Fe, Al and Ca minerals are predominant in soils.

Many studies have investigated phosphorus sorption on Fe oxides,⁴ Al oxides⁵ and in soils themselves.^6,7 It has been found that phosphorus is predominantly adsorbed in calcareous soils.⁸ In particular, calcite and dolomite are prevalent in calcareous soils, and this may influence the behavior of phosphorus compounds.⁹ Most studies on calcite,^10,11 aragonite,¹² and dolomite^13,14 have been based on very low phosphorus concentrations, and it has been reported that the immobilization of phosphate is not only affected by soil pH but also by the initial phosphate concentration.¹⁵ It is therefore necessary to study the immobilization and speciation of high concentrations of phosphorus by different carbonate minerals in phosphate-fertilized soils.

Quantification of the interaction between dissolved phosphorus species and soil minerals can be derived from the investigation of sorption envelopes and from surface sorption modeling.¹⁶ The available models for phosphorus sorption on calcite have been based on very low total phosphorus concentrations (<3.88 × 10⁻³ mg L⁻¹), without allowing for the surface area of the calcite.^12,17,18 On the other hand, surface complexation models (SCMs) include much more information regarding ion adsorption complexation and the surface areas of solids, and also the density of adsorption sites on their surface, but unfortunately few researches have studied the retention of phosphorus on carbonate minerals using SCMs. The constant capacitance model has been employed as one of the SCMs in the PHREEQC program,¹¹ a combination of surface complexation^19,20 that describes well the low sorption of phosphorus (<1.55 mg L⁻¹) by calcite. To the best of our knowledge, the use of the diffuse layer model (DLM) to quantify the surface reactions between high levels of phosphorus and calcite or dolomite remains limited.

An understanding of the underlying mechanism at the molecular level can significantly improve the quantification of the capacity of carbonate minerals to retain soluble forms of phosphorus. Elemental composition (particularly Ca and Mg levels) and a slightly alkaline pH are dominant in determining the phosphorus speciation of the solid phases.²¹ X-ray absorption spectroscopy is a valuable tool for identifying solid phase speciation of elements such as phosphorus in a number of environments. A study using high-resolution solid state ³¹P NMR and X-ray absorption near-edge structure (XANES) analysis has indicated that tricalcium phosphate (TCP) has been found as a precipitate on the surface of CaCO₃, and dibasic calcium phosphate (CaHPO₄, DCP) has also been detected.²² In addition, newberyite (MgHPO₄) is the probable magnesium orthophosphate phase present in dairy manure.²³ Nevertheless, there is little information available regarding the possible phosphorus speciation on the surface of calcite or dolomite at the molecular level.

The objective of the present study was to establish the mechanism of the high phosphorus retention by calcite and dolomite under various soil pH conditions. Specifically, the study would:

• compare the sorption isotherms on calcite and dolomite at high phosphorus concentrations;

• predict the phosphorus retention by fitting the DLM;

• study the effect of pH and phosphorus concentration on surface speciation;

• assess P speciation formed on the surface of calcite and dolomite using FTIR and XANES.

2. Materials and methods

2.1 Materials

The minerals calcite and dolomite were purchased from Zhejiang Changxing South, Inc. (China). The surface areas of calcite and dolomite were determined as 4.17 and 3.02 m² g⁻¹, respectively, by a nitrogen multi-point BET isotherm method. Scanning electron microscopy (SEM) JSM–7001F analysis indicated a crystalline aggregate structure for both calcite and dolomite (Fig. A1 in ESI†). X-ray diffractometry (XRD) confirmed the purity of the calcite (PDF# 01–085–1108) and dolomite (PDF# 01–086–2335) used in the study.

2.2 Batch sorption

All chemicals used in the study were analytical grade. Sorption isotherms were investigated by adding calcite and dolomite (5 g L⁻¹) to NaH₂PO₄·2H₂O solutions (0–250 mg L⁻¹ phosphorus). Batch sorption experiments were performed at room temperature in 0.01 M NaCl background electrolyte after 18 h equilibration. After equilibration the suspensions were centrifuged and passed through a 0.45 μm pore filter. Finally, the phosphorus concentration of the filtrate was determined colorimetrically using the molybdenum blue method.

The extent of sorption was established by the difference between the initial and final phosphate concentration in the solution. In addition, sorption envelopes for calcite and dolomite suspensions containing phosphate were obtained by adjusting the pH in the range 6.0–13.0 using dilute HCl or NaOH solution. The equilibrated pH value was recorded and plotted against the fraction of phosphate sorbed.

2.3 ATR-FTIR and XANES analysis

All suspensions were prepared in a similar manner to batch samples. After equilibration, the residues deposited were collected and air dried and the solids were then characterized for any change in the functional groups of the two minerals using a Nicolet iS10 Fourier transform infrared spectroscope (ATR-FTIR) provided with a Smart iTR accessory. To identify the speciation of phosphorus sorbed on calcite and dolomite, the solids were subjected to phosphorus X-ray absorption near-edge structure (XANES) spectroscopy. K-edge XANES spectra were collected at Beamline 4B7A of the Beijing Synchrotron Radiation Facility. A variety of phosphorus compounds were purchased from Sigma Aldrich Inc., including TCP, ACP, HAP, DCP, octacalcium phosphate (Ca₈H₂(PO₄)₆·5H₂O, OCP), dibasic calcium phosphate dehydrate (CaHPO₄·2H₂O, DCPD), magnesium phosphate hydrate (Mg₃(PO₄)₂·H₂O), and magnesium hydrogen phosphate trihydrate (MgHPO₄·3H₂O), and were analyzed for reference.

The spectra of the reference standards were determined in total electron yield mode. The electron storage ring was operated at 2.5 GeV, with an electron current in the range 150–250 mA. The chamber pressure was maintained at 10⁻⁶ Torr during the measurements. The spectra of the adsorption samples were determined as ground powder by the silicon drifted detector (PGT, USA) in partial fluorescence yield mode. A reference energy (E₀) value of 2152 eV was assigned to the maximum peak of the spectrum, scanning from 2120 to 2200 eV in steps of 0.2 eV. Multiple scans of each sample were averaged in order to improve the signal to noise ratio. All the XANES spectra were performed and normalized using ATHENA Version 0.8.056.²⁴

2.4 Modeling

The surface of calcite and dolomite was characterized by surface complexation models.^17,19,25 The DLM was applied using the 2-pK diffuse-layer model in Visual MINTEQ 3.0 to describe chemical reactions on the surface of solids.²⁶ It may be noted that the ion pair CaHPO₄⁰ is the species adsorbed on the surface of calcite.²⁷ H₂PO₄⁻, HPO₄²⁻ and CaHPO₄⁰ are the dominant species in the supersaturation system at pH above 7.0.⁸ In addition to the above species, CaPO₄⁻ is also included as an adsorbed species in our model in order to obtain better agreement with the experimental data.¹¹

At the carbonate–solution interface the surface sites are presumed to be =Me (Me = Ca or Mg) and =CO₃, with 1 : 1 stoichiometry. The DLM model thus adopts the =Ca surface for calcite, and =Ca and =Mg for dolomite. Subsequent hydration and sorption of the constituent ions from solution lead to the formation of the surface species =MeOH₂⁺, =MeO⁻, =MeHCO₃⁰, =MeCO₃⁻, =CO₃⁻, =CO₃H⁰, and =CO₃Me⁺. The two types of =sMe and =Me sites are introduced as “strong sites”, with a low density of 0.91 μmol m⁻² and “weak sites” with a higher density of 7.31 μmol m⁻², respectively.^11,20 The values of intrinsic stability constants (K^o_int) in the model for P surface complexation reactions at 25 °C are shown in Table 1.^11,19

Table 1 Surface complexation reactions for phosphorus sorption at the solid–solution interface

Solid–solution surface complexation reactions	log K^oint
=CaOH → =CaO^− + H^+	−12
=CaOH + H^+ → =CaOH2^+	11.5
=MgOH → =MgO^− + H^+	−12
=MgOH + H^+ → =MgOH2^+	10.6
=MeOH + CO3^2− + H^+ → =MeCO3^− + H2O	17.1
=MeOH + CO3^2− + 2H^+ → =MeHCO3^o + H2O	23.5
=MeCO3^− + CaHPO4(aq)^0 → =MeHPO4Ca^+ + CO3^2−	−1.75
=MeCO3^− + CaPO4^− → =MePO4Ca^0 + CO3^2−	−0.79
=sMeCO3^− + CaHPO4(aq)^0 → =sMeHPO4Ca^+ + CO3^2−	0.9
=sMeCO3^− + CaPO4^− → =sMePO4Ca^0 + CO3^2−	2.21

---

3. Results and discussion

3.1 Sorption isotherms of phosphorus

Fig. 1 illustrates the sorption isotherms of phosphate (5–250 mg L⁻¹) on calcite and dolomite at pH 6.0, 7.0, and 8.0, respectively. In the case of calcite (Fig. 1(a)), the phosphorus uptake grew steeply with increasing initial concentration of phosphate at pH 6.0 and 7.0, as shown by the linear sorption isotherms. This contributed to the precipitation of calcium phosphate in supersaturated acidic systems.²⁸ Meanwhile, similar linear isotherms were also observed for dolomite at the corresponding pH (Fig. 1(b)), and suggest the precipitation of calcium phosphate in this system also. This was possibly the result of the dissolution of dolomite (Fig. A1 in ESI†); for example, 30% of calcite or dolomite dissolved at pH 7.0 led to a large quantity of soluble Ca(II) ions in solution. In addition, it has been reported that a high concentration of background electrolytes increases the dissolution of calcite.²⁹ However, calcite and dolomite are difficult to dissolve at pH ≥ 8.0, as shown in Fig. A1 in ESI.† Accordingly, sorption isotherms for calcite and dolomite at pH 8.0 were different from those at pH 6.0 and 7.0 (Fig. 1), which may suggest different retention processes.

Fig. 1 Sorption isotherms and modeling of phosphorus (0–250 mg L⁻¹) by (a) calcite and (b) dolomite at different pH.

The reaction of phosphorus with CaCO₃ begins with sorption, and is followed by the mineral surface-induced nucleation of Ca–P phases (such as ACP, OCP, brushite, apatite, and HAP) with increasing phosphorus concentration at the various ionic strengths relevant to soil solution.^30–32 In particular, Ca–P phases have adsorptive components.²² Furthermore, at pH 8.0 sorption isotherms fit the Langmuir model well, with R² of 0.98 and 0.96 for calcite and dolomite, respectively. The sorption capacity of calcite and dolomite are 31.07 mg g⁻¹ and 38.26 mg g⁻¹, respectively. It is known that capacity for sorption to phosphate depends on the specific surface area of CaCO₃.³³ The higher retention ability of phosphate to dolomite than to calcite could therefore reasonably be attributed to the Mg in its structure, which provides different reaction sites in the crystal lattice. Previous studies have concluded that aragonite is significantly more effective as a phosphorus adsorbent than calcite over a wide range of environmental conditions.¹¹

3.2 ATR-FTIR analysis

Fig. 2 shows the ATR-FTIR spectra of calcite and dolomite loaded with phosphate at various pH values. CO₃²⁻ is associated with characteristic bands at 872 and 1440 cm⁻¹. No apparent vibration bands of sorption species were observed for calcite and dolomite at 5 mg L⁻¹ phosphorus levels, probably resulting from a low or undetectable resolution signal. When the concentration of phosphorus increased to 50 mg L⁻¹ on calcite at pH 7.0 and 8.0, the HPO₄²⁻ species had ν₁ band activation at 850 cm⁻¹. Characteristic broad absorption bands are expected at 3400 cm⁻¹ (H₂O), 1640 cm⁻¹ (H₂O), 1090 cm⁻¹ (PO₄ ν₃), and 872 cm⁻¹ (HPO₄ ν₅), indicating the formation of ACP.³⁴ Adsorption arising from the HPO₄ groups of DCP was also observed at 1120, 1032, and 872 cm⁻¹. The CO₃ ν₂ band was superimposed on the HPO₄²⁻ band at 872 cm⁻¹. It is noticeable that the PO₄ ν₁ band became sharper and was shifted from 950 cm⁻¹ to 962 cm⁻¹, characterized as HAP.³⁵ With an increase in phosphate from 50 to 200 mg L⁻¹, the characteristic bands of DCP were significantly increased, with a minor shift from 1032 cm⁻¹ to 1028 cm⁻¹. When the phosphorus concentration reached 500 mg L⁻¹, the IR band of DCP shifted further, to 1024 cm⁻¹. Meanwhile, the ν₁ band of the HPO₄²⁻ species at 850 cm⁻¹ gradually disappeared, suggesting the absence of free phosphate ions.

Fig. 2 FTIR spectra of calcite (a–c) and dolomite (a′–c′) after phosphate retention at different pH.

A number of studies have confirmed that HAP is present in all soils, regardless of pH.^6,23 A similar phenomenon has also been found in the present system. The IR adsorption characteristics of ACP, HAP and DCP are observed in dolomite at both pH 7.0 and 8.0. Careful inspection of the spectra reveals that the appearance of the bands characteristic of ACP, HAP and DCP is also seen in calcite and dolomite at pH 6.0. Although the band positions differed slightly at different pH, the numbers of bands related to symmetry were consistent. For example, the position of the adsorption band arising from DCP at pH 6.0 was shifted from 1028 cm⁻¹ to 1024 cm⁻¹, and then to 1020 cm⁻¹, corresponding to an initial phosphorus concentration of 50, 200, and 500 mg L⁻¹, respectively.

These results do not however exclude the possible existence of Mg–P phase crystallites which are too small to be detected by FTIR. In order to obtain direct information regarding the contribution of =Mg sites in the dolomite structure, the IR spectra of the synthesized MgCO₃ minerals with 200 mg L⁻¹ P retention were investigated. The MgCO₃ mineral synthesized had a typical intense band at 1420 cm⁻¹, the characteristic absorption peak of CO₃²⁻, shown in Fig. A3 in ESI.†³² Magnesium, at 865 cm⁻¹, caused increased structural disorder, resulting in an additional weak band for CO₃²⁻ at 1085 cm⁻¹ (ν₁).^4,36 The band at 1050 cm⁻¹ is attributed to the HCO₃⁻ bicarbonate symmetrical stretching mode.^37,38 The bands at ∼3510 and 3450 cm⁻¹ are assigned to water of crystallization. After loading by phosphate, the FTIR spectrum of MgCO₃ is as presented in Fig. A4 in the ESI,† showing that the typical peak of CO₃²⁻ has weakened, with a shift from 1420 cm⁻¹ to 1455 cm⁻¹.³⁹ The peak at 1650 cm⁻¹ is associated with adsorbed water.

Typical IR bands for PO₄³⁻ at 1075 and 590 cm⁻¹ were observed for the ν₃ and ν₄ bands of sorbed phosphate. In comparison with calcite or dolomite in Fig. A4,† the disappearance of resolution between peaks at 1090 and 1050 cm⁻¹ is observed, or between peaks at 605 and 565 cm⁻¹ for phosphate retention on MgCO₃ mineral. There is therefore a clear possibility of Mg–P speciation formed on dolomite, which shows a negligible IR signal masked by the strong typical peak of Ca–P speciation. It was therefore necessary to employ XANES to collect further information regarding possible phosphorus speciation at the molecular level.

3.3 XANES analysis

According to the spectra of all the references examined in Fig. 3, some spectral characteristics are evident. Apart from line (a), the first and most important feature is the presence of a shoulder past the edge at 2152–2160 eV, regardless of structure or crystallinity. As expected, the peaks at 2155 eV (b) are sharper in crystalline HAP than in the amorphous speciation, ACP. Careful inspection of the spectra reveals a loss of the slight shoulder due to the Mg–P species, Mg₃(PO₄)₂ and MgHPO₄, past the edge at 2155 eV and a shift in position at 2158 eV (c). In addition, it is interesting that a second distinctive peak is commonly seen in all Ca–P and Mg–P species at 2170 eV (e), interpreted as oxygen oscillation.⁴⁰ Furthermore, an additional peak occurred at 2162 eV (d) in all Ca–P species, although a slight shift in this peak was noticed in DCPD and β-TCP. These spectral features are consistent with the results of other studies.^6,41 However, the absence of a peak at 2162 eV was noticeable in Mg–P speciation.

Fig. 3 Normalized phosphorus absorption near-edge structure (XANES) spectra for different inorganic phosphorus reference compounds. The dashed lines indicate different spectral features of importance and are labeled as follows: (a) absorption edge for P; (b) unique spectral features of hydroxyapatite-type calcium phosphates; (c) unique spectral feature of magnesium phosphate type phases; (d) unique spectral feature of dibasic calcium phosphate-type phases; (e) oxygen oscillation.

All the standard phosphorus compounds used in the present study were employed to identify the XANES data of phosphorus speciation on calcite and dolomite at pH 6.0, 7.0 and 8.0, respectively. The differences in phosphorus spectra between calcite and dolomite were quite subtle, as shown in Fig. 4. Sharp peaks at 2155 eV were present in all samples, suggesting that the formation of HAP-type species was probably occurring at various pH values. Furthermore, a loss of the slight shoulder at 2155 eV is seen with increasing phosphorus concentration. Note that one notable peak appeared in all phosphorus-sorbed samples of calcite and dolomite at 2162 eV, attributable to the formation of DCP.⁴⁰

Fig. 4 Normalized phosphorus absorption near-edge structure (XANES) spectra for phosphate sorption on calcite (A–C) and dolomite (A′–C′) at pH 6.0, 7.0, and 8.0, respectively. The dashed lines indicate different spectral features of importance and are labeled as follows: (a) unique spectral features of hydroxyapatite-type calcium phosphates; (b) unique spectral feature of dibasic calcium phosphate-type phases; (c) oxygen oscillation.

The proportions of dominant species in the spectra are reasonably estimated using goodness of fit criteria (R factor < 0.0001) from the statistical variation through linear combination (LC) fitting. The LC fitting result in Table 2 suggests that surface Ca–P phases, including ACP, DCP, and HAP, are formed on calcite and dolomite. A recent study has reported that the water-dispersible phosphorus that predominates in inorganic form in soils is distributed as DCP, Fe–P and HAP.⁴² Our results agree with the conclusion obtained by Sato et al. that soils with short-term manure treatment contain Fe–P, DCP and ACP, whereas soils with long-term manure application show the dominance of the more stable TCP.⁴⁰ As expected, newberyite (MgHPO₄) is detectable in dolomite, although in small amounts, and this probably accounts for the greater sorption capability of dolomite than calcite (Fig. 1). However, other phosphorus species such as β-TCP, OCP and Mg₃(PO₄)₂ were not observed in all samples. Previous studies with high-resolution solid state ³¹P NMR and XANES showed β-TCP, DCP, DCPD and HAP present on the surface of CaCO₃ and soils.^6,21,40 In all cases, HAP was the main Ca–P speciation on the solid surfaces after phosphorus retention (≤200 mg L⁻¹), independently of pH. In particular, both ACP and HAP were the main speciation of calcite. However, an increase in phosphorus concentration from 200 to 500 mg L⁻¹ significantly improved the growth of DCP on dolomite.

Table 2 Proportion of phosphorus species sorbed in calcite and dolomite from linear combination fitting (standard error of fit in parenthesis)

			Goodness of fit		Linear combination fit
pH	P conc. (mg L^−1)	Sample	R factor	x^2	DCP (%)	ACP (%)	HAP (%)	MgHPO4 (%)
6.0	50	Calcite	0.000145	0.070	6.4 ± 0.3	27.6 ± 1.1	64.9 ± 1.1	—
		Dolomite	0.000198	0.096	—	15.0 ± 0.4	70.6 ± 0.4	12.3 ± 1.0
	200	Calcite	0.000419	0.191	8.3 ± 0.4	37.2 ± 1.8	52.5 ± 0.0	—
		Dolomite	0.000848	0.305	57.4 ± 1.3	20.1 ± 0.5	4.7 ± 0.0	15.1 ± 0.9
	500	Calcite	0.001077	0.414	40.8 ± 1.2	16.7 ± 0.7	39.9 ± 1.2	—
		Dolomite	0.000939	0.383	71.0 ± 1.2	10.2 ± 2.1	10.4 ± 0.0	7.7 ± 1.3
7.0	50	Calcite	0.000231	0.119	16.7 ± 0.0	5.4 ± 0.0	77.7 ± 0.0	—
		Dolomite	0.000227	0.122	15.8 ± 3.5	5.2 ± 3.6	77.5 ± 0.0	3.6 ± 0.6
	200	Calcite	0.000489	0.218	7.8 ± 0.4	36.3 ± 2.0	52.5 ± 0.0	—
		Dolomite	0.000111	0.051	35.5 ± 1.8	13.9 ± 0.5	44.8 ± 0.0	5.6 ± 0.3
	500	Calcite	0.000852	0.360	42.6 ± 2.6	13.8 ± 0.6	40.9 ± 0.6	—
		Dolomite	0.000594	0.250	67.6 ± 3.5	15.4 ± 0.9	17.4 ± 0.0	8.7 ± 0.8
8.0	50	Calcite	0.000583	0.310	15.2 ± 0.0	4.1 ± 0.0	81.3 ± 0.0	—
		Dolomite	0.000494	0.272	—	—	92.1 ± 0.6	8.4 ± 0.6
	200	Calcite	0.000177	0.083	5.8 ± 0.3	29.6 ± 1.2	62.0 ± 0.0	—
		Dolomite	0.000652	0.290	11.0 ± 0.7	—	64.7 ± 0.0	19.0 ± 1.3
	500	Calcite	0.000800	0.333	40.6 ± 2.5	13.0 ± 0.6	42.6 ± 0.0	—
		Dolomite	0.000493	0.214	77.9 ± 0.6	—	11.3 ± 0.0	12.2 ± 0.3

---

The proposed model suggested that the increase in phosphate concentration favored the formation of HAP on calcite in saturated systems.⁹ Our XANES data indicate that at high phosphate concentrations (>200 mg L⁻¹), as intermediate Ca–P phases, DCP may be stabilized on dolomite for certain periods, despite the fact that the formation of HAP is the most stable phase. This could be attributed to the Mg in the dolomite structure providing =Mg sites, favoring the generation of Mg–P phases. Our previous study showed that the adsorbency of Mg-containing calcium carbonate favored the formation of DCP on the surface with increasing Mg species in the adsorbent.⁴³ Salimi et al. showed that Mg²⁺ has a marked inhibiting effect on HAP growth, but hardly any influence on the growth of DCPD.⁴⁴ Furthermore, the presence of Mg²⁺ inhibited the conversion from ACP to HAP, resulting from competition for structural sites by the chemically similar but larger Ca²⁺.⁴⁵ Accordingly, dolomite is able to immobilize more phosphorus than calcite in the soil environment.

3.4 Modeling phosphorus retention

As is already recognized, ACP is not stable thermodynamically and transforms to the more stable HAP, possibly via the intermediates DCPD and OCP.^46–48 As a result, the presence of various Ca–P phases on the surface of calcite or dolomite cannot be excluded, although there are difficulties in characterizing the small amounts of sorbed phosphorus using spectroscopy. Furthermore, the phosphorus sorbed in the form of HAP or ACP is present at greater than 1.15 μg PO₄ m⁻² on calcite.³² It is has been reported that SCMs can successfully be used to describe the low (<1.55 mg L⁻¹) sorption of phosphorus on calcite.^19,20 Accordingly, the DLM approach may provide significant qualitative information on the low amounts of phosphorus at the interface between solid and water. To obtain better modeling, the sorption and formation of mineral surface-induced Ca–P phases (e.g. ACP and HAP) were therefore included in the DLM model at two phosphate concentration levels, 2 and 10 mg L⁻¹.

The DLM model provides a satisfactory fit for the experimental data on phosphorus retention by calcite in the alkaline region (pH ≥ 8), as illustrated in Fig. 5(a). However, the modeling seems to overestimate phosphate retention between pH 6.0 and 8.0. Again, this difference could be attributed to calcite dissolution (see Fig. A1(a) in ESI†), which leads to lower uptake of phosphorus by calcite than the modeling prediction.²⁰ The modeling results presented in Fig. 5(b) show that the monodentate complexes =CaPO₄Ca⁰ and =sCaPO₄Ca⁰ are dominant in the surface solids at a phosphorus level of 2 mg L⁻¹ within the pH range 5.5–10.5, but that the formation of Ca–P phases becomes predominant at pH > 10.5. As shown in Fig. 5(c), when the phosphorus concentration increases to 10 mg L⁻¹, Ca–P phases are formed above pH 6.0 over a wide range, while the sorbed monodentate complexes of =CaPO₄Ca⁰ and =sCaPO₄Ca⁰ (i.e. about 20% of the phosphorus) are insignificant at these pH values.

Fig. 5 (a) Modeling of phosphate retention on calcite; and species distribution of phosphate at (b) 2 mg L⁻¹ and (c) 10 mg L⁻¹ on calcite as a function of pH (electrolyte = 0.01 M NaCl).

Based on a similar modeling approach, phosphate retention by dolomite is suitably modeled in Fig. 6(a) using an analogous combination of Ca–P phases in complexes of =MgHPO₄Ca⁺ and =CaPO₄Ca⁰. However, the degree of retention is considerably overestimated in the pH range 6.0–8.0. Again, this is possibly due to dissolution of dolomite, as shown in Fig. A1(b) in the ESI.† The modeling results for phosphate (P = 2 mg L⁻¹) in Fig. 6(b) show that 80% of the sorption species at pH > 7 is the protonated monodentate complex =MgHPO₄Ca⁺, while 20% of phosphorus is formed as Ca–P phases on the surface of dolomite. However, at pH > 6.5 the dolomite surface is dominated by Ca–P phases and by another complex of =MgHPO₄Ca⁺ in the pH range 5.5–6.5 at the 10 mg L⁻¹ phosphorus level (Fig. 6(c)). The modeling results are consistent with FTIR and XANES spectroscopic analysis, showing that the growth of Ca–P nucleation phases on calcite or dolomite takes place mainly >10 mg L⁻¹ phosphorus.

Fig. 6 (a) Modeling of phosphate retention on dolomite; and species distribution of phosphate at (b) 2 mg L⁻¹ and (c) 10 mg L⁻¹ on dolomite as a function of pH (electrolyte = 0.01 M NaCl).

At pH ≥ 8.0, the retention processes may perhaps best be described in terms of simplified reactions at the solid–water interface, as in Scheme 1. The dissolution of calcite or dolomite and phosphate adsorption is the earliest reaction, and is then followed by the formation of Ca–P phases and nucleation growth when the solution system becomes supersaturated.

Scheme 1 Schematic diagrams of possible retention modes of phosphate on metal carbonates; the upper scheme is for calcite and the lower is for dolomite.

The supersaturation indices (SI) of various Ca–P phases are included in Fig. A5 in ESI.† They show an increase in SI with increasing pH, and that the calcite systems HAP, Ca₃(PO₄)₂, and Ca₄H(PO₄)₃·3H₂O are mainly saturated; and in particular that the dolomite systems Mg₃(PO₄)₂, HAP, and Ca₄H(PO₄)₃·3H₂O are also predominantly saturated. Accordingly, similar processes are taking place, except that the surface of dolomite is initially dominated by adsorption complexes of =MgHPO₄Ca⁺, and then by a variety of Ca–P and Mg–P nucleation. In addition, the intermediate DCP can be stabilized to a certain extent at phosphorus concentrations above 200 mg L⁻¹.

4. Conclusions

Phosphate retention on calcite and dolomite was investigated at different pH and concentrations of phosphate, both as a combination of sorption isotherms and by FTIR and XANES spectroscopy. The experimental results suggest that the processes of retention of high phosphorus levels in calcareous soils are different at pH 6.0, 7.0, and 8.0. The retention of high phosphate on calcite and dolomite at pH 6.0–7.0 is mainly attributable to the precipitation of calcium phosphate. A DLM model fitting the experimental data shows that the sorption complexes =CaPO₄Ca⁰/=sCaPO₄Ca⁰ and =MgHPO₄Ca⁺ are predominantly formed for phosphate retention (P ≤ 2 mg L⁻¹) on the surface of calcite and dolomite at pH ≥ 8.0. When the phosphate loading is above 10 mg L⁻¹, the nucleation growth of Ca–P phases is largely maintained on calcite and dolomite. FTIR and XANES analysis suggest that Ca–P phases of HAP are formed on the surface of calcite and dolomite, regardless of pH and phosphorus concentration. It is observed that that DCP is preferentially formed on the surface of dolomite when the phosphorus concentration is greater than 200 mg L⁻¹, the probable mechanism being that the magnesium atom in the dolomite structure provides a =Mg site for nucleation growth of the Mg–P phases. Accordingly, dolomite offers a stronger sorption capacity than calcite at pH ≥ 8.0 at high concentrations of phosphate.

Acknowledgements

The authors wish to acknowledge financial support provided by the National Natural Science Foundation of China (grants no. 21107077, 21377090, and 21277094), the Natural Science Foundation of Jiangsu Province (grant no. BK20131152), Key Innovative Projects in Colleges of Jiangsu Province (201310332018Z), Qing Lan Project, Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD), and Postgraduate Research and Innovative Projects of Suzhou University of Science and Technology (SKCX13S-058). The authors are also grateful to the Public User Program provided by the Medium-energy Beam-line Station of the Beijing Synchrotron Radiation Facility at the Institute of High Energy Physics, Chinese Academy of Sciences.

References

1. FAO Resource STAT-Fertilizer, Food and agriculture organization of the United Nations, http://faostat.fao.org/site/575/Desktop%20Default.aspx?, PageID=575#ancor, 12.03.2009, accessed 2009.
2. A. N. Sharpley, R. W. McDowell and P. J. A. Kleinman, Phosphorus loss from land to water: Integrating agricultural and environmental management, Plant Soil, 2001, 237, 287–307.
3. S. J. Traina and V. Laperche, Contaminant bioavailability in soils, sediments, and aquatic environments, Proc. Natl. Acad. Sci. U. S. A, 1999, 96, 3365–3371.
4. E. J. Elzinga and D. L. Sparks, Phosphate adsorption onto hematite: an in situ ATR-FTIR investigation of the effects of pH and loading level on the mode of phosphate surface complexation, J. Colloid Interface Sci., 2007, 308, 53–70.
5. B. B. Johnson, A. V. Ivanov, O. N. Antzutkin and W. Forsling, ³¹P nuclear magnetic resonance study of the adsorption of phosphate and phenyl phosphates on γ-Al₂O₃, Langmuir, 2002, 18, 1104–1111.
6. S. Beauchemin, D. Hesterberg, J. Chou, M. Beauchemin, R. R. Simard and D. E. Sayers, Speciation of phosphorus in phosphorus-enriched agricultural soils using X-ray absorption near-edge structure spectroscopy and chemical fractionation, J. Environ. Qual., 2003, 32, 1809–1819.
7. Y. B. Lee and P. J. Kim, Reduction of phosphate sorption by ion competition with silicate in Soil, Korean J. Environ. Agric., 2007, 26, 286–293.
8. B. Ajiboye, O. O. Akinremi, Y. Hu and A. Jürgensen, XANES speciation of phosphorus in organically amended and fertilized Vertisol and Mollisol, Soil Sci. Soc. Am. J., 2008, 72, 1256–1262.
9. L. Wang, E. Ruiz-Agudo, C. V. Putnis, M. Menneken and A. Putnis, Kinetics of calcium phosphate nucleation and growth on calcite: implications for predicting the fate of dissolved phosphate species in alkaline soils, Environ. Sci. Technol., 2011, 46, 834–842.
10. K. Karageorgiou, M. Paschalis and G. N. Anastassakis, Removal of phosphate species from solution by adsorption onto calcite used as natural adsorbent, J. Hazard. Mater., 2007, 139, 447–452.
11. H. U. Sø, D. Postma, R. Jakobsen and F. Larsen, Sorption of phosphate onto calcite; results from batch experiments and surface complexation modeling, Geochim. Cosmochim. Acta, 2011, 75, 2911–2923.
12. F. Millero, F. Huang, X. Zhu, X. Liu and J. Z. Zhang, Adsorption and desorption of phosphate on calcite and aragonite in seawater, Aquat. Geochem., 2001, 7, 33–56.
13. S. Karaca, A. Gürses, M. Ejder and M. Açıkyıldız, Kinetic modeling of liquid-phase adsorption of phosphate on dolomite, J. Colloid Interface Sci., 2004, 277, 257–263.
14. S. Karaca, A. Gürses, M. Ejder and M. Açıkyıldız, Adsorptive removal of phosphate from aqueous solutions using raw and calcinated dolomite, J. Hazard. Mater., 2006, 128, 273–279.
15. T. Suzuki, S. Inomata and K. Sawada, Adsorption of phosphate on calcite, J. Chem. Soc., Faraday Trans., 1986, 82, 1733–1743.
16. N. Devau, E. L. Cadre, P. Hinsinger, B. Jaillard and F. Gérard, Soil pH controls the environmental availability of phosphorus: experimental and mechanistic modeling approaches, Appl. Geochem., 2009, 24, 2163–2174.
17. C. E. Cowan, J. M. Zachara and C. T. Resch, Solution ion effects on the surface exchange of selenite on calcite, Geochim. Cosmochim. Acta, 1990, 54, 2223–2234.
18. R. D. Van der Weijden, J. Meima and R. N. J. Comans, Sorption and sorption reversibility of cadmium on calcite in the presence of phosphate and sulfate, Mar. Chem., 1997, 57, 119–132.
19. O. S. Pokrovsky, J. A. Mielczarski, O. Barres and J. Schott, Surface speciation models of calcite and dolomite/aqueous solution interfaces and their spectroscopic evaluation, Langmuir, 2000, 16, 2677–2688.
20. O. S. Pokrovsky and J. Schott, Surface chemistry and dissolution kinetics of divalent metal carbonates, Environ. Sci. Technol., 2002, 36, 426–432.
21. K. Güngör, A. Jürgensen and K. G. Karthikeyan, Determination of phosphorus speciation in dairy manure using XRD and XANES spectroscopy, J. Environ. Qual., 2007, 36, 1856–1863.
22. S. Hunger, H. Cho, J. T. Sims and D. L. Sparks, Direct speciation of phosphorus in alum-amended poultry litter: Solid-state ³¹P NMR investigation, Environ. Sci. Technol., 2004, 38, 674–681.
23. K. Güngör and K. G. Karthikeyan, Probable phosphorus solid phases and their stability in anaerobically digested dairy manure, Trans. ASAE, 2005, 48, 1509–1520.
24. Á. Ravel and M. Newville, ATHENA, ARTEMIS, HEPHAESTUS: data analysis for X-ray absorption spectroscopy using IFEFFIT, J. Synchrotron Radiat., 2005, 12, 537–541.
25. O. S. Pokrovsky, J. Schott and F. Thomas, Processes at the magnesium-bearing carbonates/solution interface. I. A surface speciation model for magnesite, Geochim. Cosmochim. Acta, 1999, 63, 863–880.
26. J. P. Gustafsson, Visual MINTEQ version 3.0, Stockholm, Sweden, 2011. http://www2.lwr.kth.se/English/OurSoftware/vminteq.
27. Y. P. Lin and P. C. Singer, Inhibition of calcite precipitation by orthophosphate: speciation and thermodynamic considerations, Geochim. Cosmochim. Acta, 2006, 70, 2530–2539.
28. N. Xu, H. Yin, Z. Chen, S. Liu, M. Chen and J. Zhang, Mechanisms of phosphate retention by calcite: effects of magnesium and pH, J. Soils Sediments, 2014, 14, 495–503.
29. E. Ruiz-Agudo, M. Kowacz, C. V. Putnis and A. Putnis, The role of background electrolytes on the kinetics and mechanism of calcite dissolution, Geochim. Cosmochim. Acta, 2010, 74, 1256–1267.
30. J. S. Freeman and D. L. Rowell, The adsorption and precipitation of phosphate onto calcite, Soil Sci., 1981, 32, 75–84.
31. Z. R. Hinedi, S. Goldberg, A. C. Chang and J. P. Yesinowski, A ³¹P and ¹H MAS NMR study of phosphate sorption onto calcium carbonate, J. Colloid Interface Sci., 1992, 152, 141–160.
32. W. Stumm, in Chemistry of solid–water interface. Wiley. New York, 1992, p. 428.
33. E. A. Burton and L. M. Walter, The role of pH in phosphate inhibition of calcite and aragonite precipitation rates in seawater, Geochim. Cosmochim. Acta, 1990, 54, 797–808.
34. R. Alvarez, L. A. Evans, P. J. Milham and M. A. Wilson, Effects of humic material on the precipitation of calcium phosphate, Geoderma, 2004, 118, 245–260.
35. X. Cao and W. Harris, Carbonate and Magnesium Interactive Effect on Calcium Phosphate Precipitation, Environ. Sci. Technol., 2008, 42, 436–442.
36. X. Long, Y. Ma and L. Qi, In vitro synthesis of high Mg calcite under ambient conditions and its implication for biomineralization process, Cryst. Growth Des., 2011, 11, 2866–2873.
37. A. Botha and C. A. Strydom, DTA and FT-IR analysis of the rehydration of basic magnesium carbonate, J. Therm. Anal. Calorim., 2003, 71, 987–995.
38. R. L. Frost and S. J. Palmer, Infrared and infrared emission spectroscopy of nesquehonite Mg(OH)(HCO₃)·2H₂O-implications for the formula of nesquehonite, Spectrochim. Acta, Part A, 2011, 78, 1255–1260.
39. J. Kapolos and P. G. Koutsoukos, Formation of calcium phosphates in aqueous solutions in the presence of carbonate ions, Langmuir, 1999, 15, 6557–6565.
40. S. Sato, D. Solomon, C. Hyland, Q. M. Ketterings and J. Lehmann, Phosphorus speciation in manure and manure-amended soils using XANES spectroscopy, Environ. Sci. Technol., 2005, 39, 7485–7491.
41. D. Peak, J. T. Sims and D. L. Sparks, Solid-sate speciation of natural and alum-amended poultry little using XANES spectroscopy, Environ. Sci. Technol., 2002, 36, 4253–4261.
42. J. Liu, J. Yang, B. J. Cade-Menun, X. Liang, Y. Hu and C. W. Liu, et al. Complementary phosphorus speciation in agricultural soils by sequential fractionation, solution 31P NMR, and P K-edge XANES spectroscopy, J. Environ. Qual., 2013, 42, 1763–1770.
43. N. Xu, Y. Li, L. Zheng, Y. Gao, H. Yin, J. Zhao, Z. Chen, J. Chen and M. Chen, Synthesis and application of magnesium amorphous calcium carbonate for removal of high concentration of phosphate, Chem. Eng. J., 2014, 251, 102–110.
44. M. H. Salimi, J. C. Heughebaert and G. H. Nancollas, Crystal growth of calcium phosphates in the presence of magnesium ions, Langmuir, 1985, 1, 119–122.
45. W. J. Suchanek, K. Byrappa, P. Shuk, R. E. Riman, V. F. Janas and K. S. Tenhuisen, Mechanochemical-hydrothermal synthesis of calcium phosphate powders with coupled magnesium and carbonate substitution, J. Solid State Chem., 2004, 17, 793–799.
46. G. A. O'Connor, K. L. Knudtsen and G. A. Connel, Phosphorus solubility in sludge-amended calcareous soils, J. Environ. Qual., 1986, 15, 308–312.
47. T. Moutin, J. Y. Gal, H. El Halouani, B. Picot and J. Bontoux, Decrease of phosphate concentration in a high rate pond by precipitation of calcium phosphate: Theoretical and experimental results, Water Res., 1992, 11(26), 1445–1450.
48. C. S. Dunets and Y. Zheng, Removal of phosphate from greenhouse wastewater using hydrated lime, Env. Technol., 2014 DOI:10.1080/09593330.2014.924567 in press.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra05461j

---