Introduction

Various remediation technologies have been applied to cadmium (Cd)-contaminated arable soils to make the crops grown in these soils suitable for human consumption [1,2,3,4,5,6]. Many studies on Cd chemical immobilization in arable soils using phosphate (P) fertilizer have been conducted. It is well known that P materials are effective in reducing the bioavailability of Cd in soils [7,8,9,10,11].

The immobilization of Cd with P compounds in soils is caused by various mechanism: (i) the formation of Cd minerals such as Cd3(PO4)2; (ii) Cd2+ adsorption induced by P; (iii) cation exchange; and (iv) surface complexation. Indeed, several studies observed that Cd was immobilized with P through forming a compound mineral with varying concentrations of these two elements. However, others reported that the precipitation of a Cd-P mineral did not occur under their experimental conditions [12, 13]. Some researchers reported that the addition of P immobilized Cd by increasing its adsorption onto the surface of soil colloids [4, 14, 15], but did not elucidate the exact pathway of P-induced Cd2+ adsorption. In a previous study [4], we suggested that Cd might not be precipitated as Cd-P minerals under 1,600 mg P kg−1 of P addition in soil containing low Cd concentration (total Cd: 5.57 mg kg−1). In such soil conditions, an increase in the pH-induced negative charge of soil could play an important role in the immobilization of Cd. However, we did not determine the soil condition under which Cd could precipitate with P addition and did not investigate the exact role of a pH-induced negative charge in Cd immobilization. So far, the exact mechanism of Cd immobilization with P in arable soils containing low Cd concentration remains poorly understood.

Therefore, this study was designed to investigate the mechanism of Cd immobilization by P in the arable soil containing low Cd concentration, by estimating (1) the P addition rate under which Cd could precipitate as a Cd-P mineral and identify the main species of Cd minerals formed by the reaction of Cd and P; (2) the contribution of negative charge induced by P to Cd2+ adsorption in soil.

Materials and methods

Incubation test for the precipitation of cadmium and phosphate

An arable soil located near the Bongsan gold mine area in Hapcheon, South Korea (128°01′N and 34°37′E) was collected for the precipitation test. The total Cd concentration in this soil was 5.57 mg kg−1, which exceeded the warning criteria for Cd in the arable soil established by the Korean Soil Environmental Conservation Act [16]. The chemical and physical properties of the collected arable soil are shown in Table 1.

Table 1 Selected chemical and physical properties of soil for precipitation and adsorption test
Full size table

To determine the suitable P addition rate to induce Cd precipitation as Cd-P minerals, the effects of P addition rate on the solubility of Cd minerals were investigated in this study. A 500-g soil sample was mixed with 0, 200, 400, 800, 1600, 3200, 8000, and 16,000 mg P kg−1 of K2HPO4 in a plastic beaker and placed in the dark at 25 °C for 8 weeks. Distilled water was added to fill up 70% of the soil pore volume. Soil water content was maintained by periodically weighing the beakers and adjusting the weight by adding distilled water. Triplicates were arranged in a completely randomized design in an incubator.

Phosphate adsorption

For the P and Cd adsorption test, uncontaminated arable soil was collected from the experimental farm of Pusan National University, Miryang, South Korea (128°43′ N and 35°30′ E). The physical and chemical properties of the studied soil are also given in Table 1.

To adsorb P onto the uncontaminated soil, a 1-g soil sample was reacted with 20 ml of 0–10,000 mg P L−1 (K2HPO4) solution without controlling solution pH and shaken at 25 °C for 48 h. Solution pH condition is a significant factor affecting adsorptions of P and Cd. We assumed that results of the test could be derived from the wrong thing, if adsorption test was conducted in controlled pH condition. For this reason, solution pH was not controlled in the test. The amount of P adsorbed was calculated by subtracting its final concentration in the equilibrium solution from its initial concentration. The adsorption test was carried out in six replicates. After shaking for 48 h, the soil residues were rinsed with 20 ml of 95% ethyl alcohol; three replicates were used to analyze for the negative charge of the soil and the other three were used to measure the amount of adsorbed Cd.

Cadmium adsorption

The soil residues mentioned earlier were reacted with 20 ml of 500 mg Cd L−1 (CdCl2) solution and shaken at 25 °C for 48 h to determine the amount of Cd adsorbed by P fixed to the soil. The Cd adsorption capacity was calculated by subtracting its final concentration in the equilibrium solution from its initial concentration.

Determination of pH-induced negative charge and phosphate-induced negative charge

The amount of Cd adsorbed to P added to the soil could be dependent on the change in soil negative charge brought about by P adsorption. An increase in the negative charge of the soil caused by P adsorption could be attributed to corresponding increases in (1) the pH-dependent charge by specific adsorption of P (pH-induced negative charge) and (2) negative charge from adsorbed orthophosphate (P-induced negative charge). Therefore, we determined the contribution of each of the pH-induced negative charge and P-induced negative charge to the total increase in negative charge of the soil elicited by P adsorption.

The soil used for the P and Cd adsorption test was used to determine the contribution of pH-induced negative charge in the soil. 25 ml of distilled water was added to 5 g of the soil and shaken in a polyethylene centrifuge tube at 25 °C until the pH was adjusted to pH between 3 and 10 using either 0.1 M NaOH or HCl. After centrifuging for 10 min at 5000 rpm, the supernatant solutions were decanted and 20 ml of 95% ethyl alcohol was added to rinse the soils before analyzing the pH-induced negative charge of the soil.

The phosphate-induced negative charge was calculated by subtracting pH-induced negative charge obtained at equilibrial P adsorption from the total negative charge attained at P adsorption equilibrium.

Cd-mineral speciation

The Visual MINTEQ (ver 2.23) was used to predict the precipitation or dissolution of Cd minerals such as Cd3(PO4)2, CdCO3, and Cd(OH)2. The input parameters were the respective concentrations of cations and anions in the soil solution. The concentration of dissolved organic carbon (DOC), ionic strength, pH, and temperature (25 °C) were also inputted. In this study, we assumed a constant CO2 concentration of 0.003 atm.

Chemical analysis

The chemical properties of the collected soils (< 2 mm) were analyzed using the following methods: pH (1:5 with H2O), organic matter content [17], and concentrations of exchangeable Ca2+, Mg2+, and K+ (1 M NH4-acetate pH 7.0, AA, Shimazu 660). The available P content was determined using the Lancaster method [18]. The negative charge of the soil was measured using 0.1 M NaCl after performing Schofield’s ion retention method [19]. The concentrations of cations in the soil solution were measured using an inductively coupled plasma-optical emission spectrophotometer (ICP-OES, Perkin Elmer Model OPTIMA 4300 DV, Shelton USA). The anions in the soil solution were quantified using an ion chromatograph (IC, Dionex Corporation Model ICS-2000 IC, USA). DOC was quantified using a TOC analyzer (Total Organic Carbon Analyzer Shimadzu Model TOC-VCPN, Japan).

After the P adsorption test, the pH of the reacted solution was measured using a pH meter. The concentration of P in the equilibrium solution was measured spectrophotometrically at 880 nm [20]. After the Cd adsorption test, Cd concentration in the equilibrium solution was measured using ICP-OES.

Results and discussion

Precipitation of cadmium with phosphate

The pH of the soil solution increased with increasing K2HPO4 concentrations (Table 2). That was because the hydroxyl (OH−) group was released into the soil solution when the specific adsorption of P to the edge of the crystals of clay minerals occurred. Log activity of CO32− increased significantly as more K2HPO4 was added, corresponding to an increase in soil pH. The dissociation of carbonic acid in the soil solution is represented in the following equation:

$$ \text{H}_{2} {\text{C}}\text{O}_{3} \leftrightarrow \text{H}^{{ + }} \text{ + }{\text{H}}{\text{C}}\text{O}_{3}^{{ - }} \leftrightarrow \text{H}^{{ + }} { + }{\text{C}}\text{O}_{3}^{{{2 - }}} $$
(1)
Table 2 Calculated activities of Cd2+, H2PO4−, and CO32−, as well as the ionic strength and pH of soil solutions amended with different rates of K2HPO4 after eight weeks of incubation
Full size table

where pKa1 and pKa2 in carbonic acid are 6.36 and 10.33, respectively. Very few cadmium carbonate (CdCO3) molecules can form in acidic soil as the activity of CO32− is low, but this increases in alkaline soil, allowing more CdCO3 molecules to form [21]. The log activity of H2PO4− increased up to the peak concentration of 1600 mg P kg−1, but decreased afterwards as more K2HPO4 was added (Table 2). This change is related to an increase in soil pH as more K2HPO4 is added. The dissociation of phosphoric acid in soil solution is represented in the following equation:

$$ \text{H}_{3} \text{P}\text{O}_{4} \leftrightarrow \text{H}^{{ + }} \text{ + } \text{H}_{2} \text{P}\text{O}_{4}^{{ - }} \leftrightarrow \text{H}^{{ + }} \text{ + }\text{H}\text{P}\text{O}_{4}^{{\text{2 - }}} \leftrightarrow \text{H}^{{ + }} \text{ +} \text{P}\text{O}_{4}^{{\text{3 - }}} $$
(2)

where pKa1, pKa2, and pKa3 of phosphoric acid are 2.15, 7.20, and 12.35, respectively. Therefore, H2PO4− is the predominant chemical species of phosphoric acid in the pH range of 2.15–7.20. However, HPO42− dominates above pH 7.20 [21]. The log activity of H2PO4− increased up to 1,600 mg P kg−1 as more K2HPO4 was added, as the solution pH was below 7.20 (and hence H2PO4− was predominant) (Table 2). However, it decreased when we added more than 1,600 mg P kg−1 of K2HPO4, as soil pH increased above 7.20, in which HPO42− is predominant.

The results from the calculated Cd solubility diagram (Fig. 1) suggest that CdCO3 might control the solubility of Cd at 8000 mg P kg−1 of K2HPO4 addition. When pH exceeds 7.5; depending on CO2 (g), Cd2+ activity is limited by CdCO3 concentration [21]. Soil pH was 8.29 with 8000 mg P kg−1 of K2HPO4 addition (Table 2) and the soil solution was supersaturated with respect to only CdCO3. However, the soil solution was supersaturated with respect to both CdCO3 and Cd3(PO4)2 under 16,000 mg P kg−1 of P addition (Fig. 1). In general, ions in a given solution can precipitate if the IAP of ion-induced minerals is higher than the mineral solubility product (Ksp) [22]. Log IAP values of Cd3(PO4)2 were −30.7 with 16,000 mg P kg−1 of K2HPO4 addition and higher than Ksp values (−32.6) of Cd3(PO4)2. This implies that Cd might precipitate as Cd3(PO4)2 under 16,000 mg P kg−1 conditions. Some studies reported formation of Cd3(PO4)2 in Cd-contaminated soils amended with a suitable concentration of P [23, 24]. The application of large amounts of P fertilizer can release P from the arable land and subsequently, cause eutrophication of nearby water bodies. For instance, a concentration of 33.5 kg P ha−1 of P fertilizer is recommended for radish cultivation in Korea [25], using a product concentration of ca. 18.6 mg P kg−1. In this study, the soil solution was undersaturated compared to Cd3(PO4)2 and CdCO3 below 400 mg P kg−1 (low P addition) and Cd might not precipitate as Cd3(PO4)2 or CdCO3 (Fig. 1). As a result, Cd solubility might be controlled by other mechanisms rather than precipitation under low P application systems.

Fig. 1
figure 1

Cadmium solubility diagram with soil solution speciation by increasing K2HPO4 addition rate after 8 weeks of incubation at 25 °C. The CdCO3 line assumes CO2 (g) = 10–2.5 M

Full size image

Phosphate adsorption

Soil phosphate adsorption increased as the P equilibrium concentration increased (Fig. 2). In general, the P adsorption occurred with an L-curve type on the soil surface in the controlled pH condition. However, a C-curve type of P adsorption was observed in this study, probably due to changes in the soil solution pH depending on the concentration of K2HPO4. In this study, the P adsorption experiment was conducted without controlling solution pH to understand P dynamics under the natural soil condition in which P is specifically adsorbed or precipitated with cations. The final pH of equilibrium solution increased with increasing adsorbed P up to 689 mmolc P kg−1 of adsorbed P, but there was no further increase in final pH with additional P adsorption (Fig. 3). The final pH of equilibrium solution was 8.71 at 689 mmolc P kg−1 of adsorbed P. The pH of the soil solution increased as more OH− was released into the soil solution through ligand exchange when specific P adsorption occured. We assumed that the maximum amount ofshared charge by sharing the P adsorption in the studied soil was 689 mmolc P kg−1, as the final pH was highest under this specific concentration of adsorbed P. Phosphate removal from the above equilibrium solution might be attributed to P precipitation with calcium in alkaline pH conditions, instead of specific adsorption of P (Fig. 2).

Fig. 2
figure 2

Phosphate adsorption characteristics of the selected clay loam soil at 25 °C

Full size image
Fig. 3
figure 3

Final pH of equilibrium solution after phosphate adsorption

Full size image

The trend in the soil negative charge recorded after reacting with P was similar to that of the final pH of equilibrium solution (Fig. 4). The soil negative charge was 120 mmolc kg−1 in the control, increased up to 164 mmolc kg−1 at 689 mmolc P kg−1 of adsorbed P. This implies that total 44 mmolc kg−1 of soil negative charge was increased by P adsorption. This increase in soil negative charge by P adsorption could be attributed to two reasons, first, the increase in pH-induced negative charge and second, the increase in P-induced negative charge. Phosphate adsorption could increase the pH-induced soil negative charge as OH− was released from the edge of the crystals of clay minerals when specific adsorption of P occurred. Furthermore, the negative charge of the oxide group contained in P itself could contribute to the increase in soil negative charge. Each oxygen (O−2) in the orthophosphate (PO43−) has − 0.75 of shared charge by sharing the charge with the P+5 atom. This “shared charge” is determined by dividing the valence of P by the number of bonded O atoms [26].

Fig. 4
figure 4

Changes in soil negative charge due to increased phosphate adsorption to clay loam soil at 25 °C

Full size image

Cadmium adsorption

The pattern of Cd adsorption to P adsorbed soil was similar to that of the soil negative charge (Fig. 5). Cadmium adsorption was 139 mmolc kg−1 in the control and increased up to 161 mmolc kg−1 at 689 mmolc P kg−1 of adsorbed P, implying that a total of 22 mmolc kg−1 of Cd was adsorbed onto the negative site in the soil. The Cd maximal adsorption corresponded to maximal P adsorption (689 mmolc P kg−1), possibly implying that Cd possibly occupied all adsorption sites made by P adsorption, and that the maximum concentration of P adsorbed soil did not provide more Cd adsorption sites. However, the negative charge at P maximal adsorption (689 mmolc kg−1) was 164 mmolc kg−1 (Fig. 4) and was not the same as the Cd maximal adsorption (161 mmolc kg−1) (Fig. 5). Moreover, Cd adsorption did not have a significant correlation with P adsorption but showed a significantly positive correlation with soil negative charge (r = 0.841**) (Table 3). Based on the above results, we assumed that Cd was not adsorbed to all adsorption sites created by adsorbed P.

Fig. 5
figure 5

Cd-adsorption to phosphate adsorbed soil after reacting with 500 mg Cd L−1 of CdCl2 solution at 25 °C

Full size image
Table 3 Correlation coefficients between each parameter
Full size table

Contribution of pH-induced negative charge to Cd adsorption

The soil negative charge with P adsorption was found to increase as both pH- and P-induced negative charges increased. The soil negative charge had a more significant positive correlation with the final pH (r = 0.977***) than with P adsorption (r = 0.830**) (Table 3), implying that changes in soil pH caused by P adsorption could explain changes in negative charge better than that of P adsorption itself. The soil negative charge used for the adsorption test increased significantly with increasing soil pH (in the range of pH 3.1–10.0; Fig. 6). The soil negative charge was 120 mmolc kg−1 at pH 6.97 which was the initial pH of soil used for the adsorption test. This increased to 161 mmolc kg−1 at pH 8.70, which was the soil pH at 689 mmolc kg−1 of P adsorption. Therefore, 41 mmolc kg−1 of the total increase in soil negative charge (44 mmolc kg−1) with 689 mmolc kg−1 of P adsorption was caused by an increase in pH-induced negative charge and 3 mmolc kg−1 of that was originated from an increase in P-induced negative charge. Therefore, contributions of the increase in pH-induced negative charge and P-induced negative charge to the total increase in soil negative charge with 689 mmolc kg−1 of P adsorption were 93.2 and 6.8%, respectively.

Fig. 6
figure 6

Changes in negative charge after adjusting pH of clay loam soil

Full size image

In conclusion, Cd might precipitate as Cd3(PO4)2 under extremely high P addition (16,000 mg Pkg−1). The formation of Cd minerals might not occur with P under the recommended field application rate of P fertilizer. Cd solubility might be controlled by another mechanism rather than precipitation under low P application systems. Phosphate adsorption increased the soil negative charge and Cd adsorption. Contributions of the increase in pH- and P-induced negative charges to the total increase in soil negative charge were 93.2 and 6.8%, respectively. The concurrent increase in Cd adsorption with P adsorption was mainly attributed to the increase in pH-induced negative charge.