Improvement of nitrate-leaching control technology using an anion exchange compound on agriculture 1: synthesis of a Mg-Fe system layered double hydroxide and its anion exchange characteristics

ABSTRACT

A hydrotalcite-like compound for reducing the leaching of nitrogen from farmland was prepared via a new synthesis procedure, and its chemical structure and anion exchange characteristics were described. The advantage of this new synthesis method is that it is a simple, inexpensive process that involves the mixing of only two types of materials, followed by a hydrothermal reaction that does not require pH or gas control. The materials for synthesis are MgO and FeCl₃. Chemical analysis, X-ray diffraction, and scanning electron microscopy were used to confirm an Mg-Fe system hydrotalcite-like compound. The anion exchange capacity was 166 cmol_c kg⁻¹, which was determined using NO₃⁻ ion as an exchangeable anion. That capacity is equivalent to 23 g N kg⁻¹. The nitrate adsorption isotherm of this compound approximately matched to the isotherm that was predicted by the Langmuir formula. The anion exchange selectivity coefficients of the compound were in the following order: NO₃⁻ > Cl⁻ > SO₄²⁻ > HPO₄²⁻. The developed compound is a prospective material for reducing leaching of NO₃−N from farmland.

KEY WORDS:

• Anion exchange
• groundwater
• hydrotalcite
• leaching
• nitrate

1. Introduction

Nitrate nitrogen (NO₃−N) has been listed in the Environmental Quality Standards (EQSs) for groundwater since 1999 in Japan, and 10 mg dm⁻³ is the critical concentration from the standpoint of human health. Leaching of NO₃−N is known to be more common than the leaching of ammonium nitrogen because both NO₃−N and soil are negatively charged. Reducing the leaching of NO₃−N from soil is an important factor for environmental protection. One of the main origins of NO₃−N in groundwater might be applied fertilizers (Nakanishi et al. 2001; Nonaka and Kamura 1995; Nonaka et al. 1996). To reduce losses of NO₃−N, researchers have investigated controlled-release fertilizer, partial fertilization (Togashi and Yamazaki 2000), and organic materials (Ito and Togashi 2002; Ito and Togashi 2010). Results have shown that the recovery rate of fertilizer N by crops increased and the losses of applied NO₃−N were reduced. However, according to the results of the FY 2015 Water Quality Survey of Ground-water, 2.9% of the surveyed wells exceed the EQS for NO₃−N or NO₂−N (Ministry of Environment 2016). The results of this survey suggest that more effective leaching control technology is required to reduce groundwater pollution on sandy soil, particularly soil with high water percolation rate. Shimada and Takahashi (1979) reported that mixing of an ion exchange resin with soil resulted in lower leaching of NO₃−N than in untreated soil. However, ion exchange resins are too expensive for use in agricultural production. A low-cost material is needed.

NO₃−N adsorption by biochar (BC) produced from wood-based biomass at 800°C was significantly higher than in BC produced from non-wood-based biomass at 800°C and can be adequate as a soil amendment material (Kameyama et al. 2016). However, the adsorption capacity was not high, even for BC with the highest adsorption ability. Mori and Ono (1995) reported that charcoal (wood-based BC) could not adsorb NO₃−N, but charcoal treated with ferric chloride (FeCl₃) solution adsorbed NO₃−N due to the positive charge of the ferric ion. Ito and Yamazaki (2008) reported the effect of defatted rice bran carbide, which was treated with FeCl₃ solution to reduce leaching of N from the sandy field. In their test group, the material was supplied with 3 Mg ha⁻¹, and leaching of N was reduced to 2 kg ha⁻¹. The amount of reduced N reaching was not satisfactory, instead of the large amount of application of the soil amendment material of 3 Mg ha⁻¹. Therefore, the author tried to develop a new soil amendment having much larger adsorption capacity in the present study. The material, which has a larger adsorption capacity, was required to reduce the leaching of N.

Hydrotalcite is a clay mineral with a high anion exchange capacity (AEC). A reduction in NO₃−N leaching is expected when this clay mineral is applied to soil. Hydrotalcite is a layered double hydroxide (LDH), and its formula is represented as M²⁺_xM³⁺_y(OH)_2x+3y−nz (Aⁿ⁻)_z mH₂O, where M²⁺ and M³⁺ are divalent and trivalent cations, respectively, and Aⁿ⁻ represents n-valent anions (Miyata and Kumura 1973). Many hydrotalcite-like compounds have been used in industrial or medical applications. The trivalent metal of LDH is typically aluminum (Miyata 1983), but aluminum compounds are not desirable for agricultural use in Japan because the crystal structure of hydrotalcite-like compounds is stable under neutral- to high-pH conditions but will dissolve under strongly acidic conditions (Hibino 2006); additionally, most upland soils in Japan are acidic (Takahashi et al. 2013). Hydrotalcite has well-known isomorphous and polytypic compounds. The Fe³⁺ derivative of hydrotalcite is known as pyroaurite or iowaite. Pyroaurite has been reported to exhibit the same NO₃−N adsorption ability as hydrotalcite with Al³⁺ as the trivalent metal (Wuxian et al. 1995; Wuxian et al. 1996). Yamamoto (2006) studied the utility of the Mg–Fe LDH system and reported that this LDH exhibits a high NO₃–N adsorption ability. However, the influence of coexisting anions (i.e., phosphate and sulfate anions), which have been extensively applied as a soil conditioner or fertilizer on the adsorption of NO₃−N, was not considered. In addition, the application of large amounts of LDH has been observed to increase soil pH and inhibit crop growth (Yamamoto 2006).

Most types of LDH are prepared using co-precipitation and hydrothermal methods. For example, LDH was prepared from a mixed solution of divalent and trivalent metals by adjusting the pH from 9 to 11 and treating the solution in an autoclave at 150°C for 3 d; the autoclave was filled with N₂ gas beforehand. The resulting precipitate was subsequently washed with deionized and decarbonated water, dried, and ground to a powder (Wuxian et al. 1995). These procedures required pH adjustment of the solution, more than three different materials were needed, and the pH of the wastewater was high. Thus, this procedure for preparing LDH is expensive and complex, giving rise to pollution problems. Therefore, an easier procedure for producing LDH for agriculture is needed. In response, I developed a new procedure to prepare hydrotalcite-like LDH under easier, less expensive synthesis conditions than the conventional method. I characterized the new product, including its NO₃−N adsorption ability and selective adsorbing capacity among NO₃⁻, PO₄³⁻, SO₄²⁻, and Cl⁻.

2. Materials and methods

2.1. Sample preparation

LDH was prepared by mixing an MgO suspension and a FeCl₃ solution followed by a hydrothermal method; alkaline materials (e.g., sodium hydroxide solution or aqueous ammonia) were not required. The samples were synthesized via reaction of 0.1 dm³ of 0.5, 1.0, 1.5, or 2.0 mol dm⁻³ FeCl₃ (guaranteed reagent; Wako Pure Chemical Industries, Ltd., Osaka, Japan) solution with a 0.1 dm³ of 8 mol dm⁻³ MgO (heavy; Wako Pure Chemical Industries, Ltd, Osaka, Japan) suspension. Samples are referred as PA_0.5, PA_1.0, PA_1.5, or PA_2.0, respectively. The FeCl₃ solution was added to a stirring suspension of MgO, and the pH of the solution was not adjusted. The precipitate settled via stirring for 30 s of the reactant mixture. Next, 0.2 dm³ of distilled water was added to the reactant to form a suspension. The reactant in the autoclave (SP-51; Yamato Scientific Co., Ltd., Tokyo Japan) was heated at 120°C for 2 h under autogenous pressure. The precipitate was collected by decantation, dried at 150°C for 1 d and subsequently washed with distilled water. The products were dried at 200°C for 1 d and subsequently ground into a powder.

2.2. Nitrate adsorption capacity of prepared samples

Five hundred milligrams of samples (i.e., PA_0.5−PA_2.0) were placed in plastic centrifuge tubes, and 25 × 10^–3 dm³ of a 16 mmol dm⁻³ Ca(NO₃)₂ solution was added. The tubes were shaken for 1 h on a reciprocal shaker (SA-31; Yamato Scientific Co., Ltd., Osaka, Japan) and centrifuged at 1670 g for 3 min. Supernatants were diluted to 50-fold and filtered through a disk filter (PP Syringe Filter 0.45-μm, Membrane Solutions Co., Ltd., TX., USA). The amount of NO₃⁻ in the solution was determined by ion chromatography (SCL-10Asp; Shimadzu Co., Kyoto, Japan). The amount of NO₃⁻ adsorbed with samples was calculated according to the following equation:

(1)

where QNO₃ is the adsorbed NO₃ (cmol kg⁻¹), C₀ and C_eq are the initial and equilibrium concentration of NO₃, respectively (mmol dm⁻³), V₁ is the volume of extraction (dm³), and W₁ is the mass of PA (kg).

2.3. Analysis of prepared samples

The Mg and Fe contents of PA_1.5 were determined by atomic absorption spectrometry (SpectrAA 220FS; Varian Australia Pty Ltd., Australia) and Cl⁻ content was determined by ion chromatography after the PA_1.5 was dissolved in 1 mol dm⁻³ HNO₃. The structure of PA_1.5 was analyzed by X-ray powder diffraction (XRD), and the pattern was recorded on a diffractometer using CuKα radiation (λ = 1.54056 Å) with a Bragg angle of 2–70° (2θ) and steps of 0.02°counting 0.6 s/step (MiniFlex; Rigaku Co., Tokyo, Japan). PA_1.5 was coated with a thin layer of evaporated carbon, and images were obtained using a HITACHI SU8000 scanning electron microscope (SEM) at a 15-kV accelerating voltage.

2.3.1. Nitrate adsorption isotherms for PA_1.5

Solutions with the initial NO₃⁻ concentrations of 0, 1.6, 3.2, 8.0, or 16 mmol dm⁻³ were prepared using Ca(NO₃)₂. One gram of PA_1.5 was suspended with 25 × 10⁻³ dm³ of each Ca(NO₃)₂ solution in plastic centrifuge tubes; after being shaken for 1 h, each tube was allowed to stand for 12 h. Supernatant was diluted 50-fold and filtered through a disk filter. The amount of NO₃⁻ and Cl⁻ in solution was determined by the aforementioned ion chromatography. The amounts of NO₃⁻ adsorbed with PA_1.5 were calculated using Equation (1). The NO₃⁻ exchange isotherm of PA_1.5 was adapted to the Langmuir formula using the least squares method.

The amount of desorbed Cl⁻ was calculated according to the following equation:

(2)

where Q_Cl is the concentration of Cl⁻ (mmol kg⁻¹), C_eq is the equilibrium concentration of Cl⁻ (mmol dm⁻³), C₀ is the concentration of Cl⁻ (mmol dm⁻³) in the equilibrium solution not including NO₃⁻, V₂ is the volume of extraction (dm³), and W₂ is the mass of PA_1.5 (kg).

2.4. Determination of the AEC for PA_1.5

AEC was determined according to the method of Nakahara and Wada (1993). Although Nakahara and Wada measured AEC with Cl⁻ as an exchangeable anion, in this study, AEC was measured using NO₃⁻ as an exchangeable anion because the adsorptive target ion is NO₃⁻, and the procedure is as follows. The PA_1.5 (0.5 × 10⁻³ kg) was suspended as a 25 × 10⁻³ dm³ of 1 mol dm⁻³ KNO₃ solution in a plastic centrifuge tube that was weighed in advance. The suspension was shaken for an hour in the abovementioned shaker and left to stand overnight with occasional stirring. The supernatant was disposed after centrifugation by 1670 g, 3 min. Next, the PA_1.5 that remained in the tube was washed five times with 0.02 mol dm⁻³ KNO₃ solution. The content of NO₃⁻ in the equilibrium solution was measured with supernatant after the fifth washing. Next, the PA_1.5 was washed five times with a 1 mol dm⁻³ KCl solution, and supernatants were collected to measure the NO₃⁻ content. AEC was calculated by the following equation:

(3)

where AEC is the anion exchange capacity (cmol_c kg⁻¹), n_ex is the amount of NO₃⁻ in the extraction (mmol), m_eq is the mass molar concentration of the equilibrium solution (mmol kg⁻¹), W_residual is the mass of the equilibrium solution that was residual in the tube (kg), and W₃ is the mass of PA_1.5 (kg).

2.5. Evaluation of anion exchange selectivity of PA_1.5

The following anions in the soil due to fertilization were used as target anions: nitrate, phosphate, and sulfate ions. The anion solutions were prepared using NaNO₃ (extra pure reagent: Kanto Chemical Co. Inc., Tokyo, Japan), NaH₂PO₄ (guaranteed reagent; Wako Pure Chemical Industries, Ltd., Osaka, Japan), and Na₂SO₄ (guaranteed reagent; Wako Pure Chemical Industries, Ltd., Osaka, Japan). Then, 0.1 × 10⁻³ kg of PA_1.5 was suspended in 25 × 10⁻³ dm³ of anion solution. Total negative charge of anion contained in the test solution was prepared so as not to exceed the amount that 0.1 × 10⁻³ kg of PA_1.5 could adsorb. The initial concentrations of each anion and the pH of each solution were adjusted to 3 mmol dm⁻³ and 7.0, respectively. However, a Cl⁻ solution was not prepared because PA_1.5 already contains Cl⁻ intercalated into the layered structure. The concentration and pH for the three anions mixed solution was adjusted to the same of single anion solution (i.e., each anion concentration and pH was 3 mmol dm⁻³ and 7.0, respectively). For this reason, the total negative charge of anions in the three anions mixed solution was 0.2 meq above the amount that 0.1 × 10⁻³ kg of PA_1.5 could adsorb.

Because it is assumed PA_1.5 will be used for agricultural purposes, a competitive adsorption experiment on imitative soil solution was conducted. An imitative soil solution was prepared including Cl⁻, NO₃^−, PO₄³⁻, and SO₄²⁻ to 2 mmol dm⁻³, 10 mmol dm⁻³, 1 mmol dm⁻³, and 2 mmol dm^−3, respectively, based on the author’s experience in the chemical analysis of dune fields. The amount of PA_1.5 application was assumed to be 1 Mg ha⁻¹, and the amount of soil solution included in 1 ha of dune field was estimated at 250 Mg (assuming cultivated soil depth is 20 cm, bulk density is 1.3 g cm⁻³, and water content is 9–10%). Therefore, 0.1 × 10⁻³ kg of PA_1.5 was suspended in 25 × 10⁻³ dm³ of imitative soil solution. The negative charge of ions contained in the imitative soil solution was 0.28 meq above the amount that 0.1 × 10⁻³ kg of PA_1.5 could adsorb. The anion exchange selectivity of the aforementioned anions was determined as follows. PA_1.5 was suspended in solutions containing the aforementioned anions and shaken for 1 h. The supernatant was collected by decanting after centrifugation (1670 g, 3 min) to determine the anion content in the equilibrium solution. The concentration of each anion in the equilibrium solution was measured using ion chromatography. The adsorption of each anion was calculated using the following equation:

(4)

where Q is the adsorbed anion (cmol kg⁻¹), C₀ and C_eq are the initial and equilibrium concentration of anion, respectively (mmol dm⁻³), V₄ is the volume of extraction (dm³), and W₄ is the mass of PA_1.5 (kg).

The anion exchange selectivity coefficients for each anion were related to Cl⁻ (K_GT) (Gaines and Thomas 1953) and the distribution coefficients (K_d dm³ kg⁻¹) were calculated using the following equation:

(5)

(6)

Fundamentally, the ion exchange selectivity coefficient is calculated using the ion activity in solution. The activity can be computed by multiplying the molar concentration by activity coefficient. In this study, the selectivity coefficient was computed using the molar concentration of solute ions instead of activities, because the activity coefficients will approach 1 in dilute solutions.

The selectivity coefficient (K_GT) defined using the charge fraction is known as Gains–Thomas selectivity coefficient (Wada 1987). E_A and E_Cl are charge fractions of adsorbed anion (mol kg⁻¹) and chloride (mol kg⁻¹), and [A^Z−] and [Cl⁻] are concentrations in the solution phase (mol dm⁻³). The anion adsorption selectivity was evaluated using log K_GT and K_d. The experiments were performed at ambient temperature (25°C).

3. Results and discussion

3.1. Anion adsorption capacity of prepared samples

The relationship between the Fe/Mg molar ratio of reactants and NO₃⁻ adsorption ability of the samples is shown in Fig. 1. PA_1.0 and PA_1.5 exhibited the highest NO₃⁻ adsorption ability among the investigated samples. The molar ratio of Fe/Mg in reactants for the synthesis of samples with high nitrate ion adsorption ability was thus surmised to be 0.13–0.19.

Figure 1. NO₃^– adsorption properties of prepared samples.

3.2. Determination of the structure of PA_1.5

The chemical composition of PA_1.5 is shown in Table 1. The molar ratio of cations (Fe/Mg) of the product was 0.23, and chloride ion was included in the compound. However, the mole of Cl⁻ in the unit mass of PA_1.5 was smaller than for substituted Fe which is the cause of the positive charge on PA_1.5. It was deduced that HCO₃⁻ and CO₃²⁻ exist at the exchange site, because PA_1.5 was synthesized using distilled water, and the procedure was performed in an air atmosphere with an alkaline reactant.

Table 1. Chemical composition of PA_1.5.

element	cmol kg^−1
Mg	1178
Fe	271
Cl	170

The XRD data for PA_1.5 is shown in Table 2. The data in this work are very similar to that for iowaite (a mineral of the hydrotalcite group) in previous reports (Kohls and Rodda 1967; Braithwaite et al. 1994). The lattice constants of PA_1.5 were determined to be approximately a = 311 pm and c = 2407 pm and are close to those for natural iowaite, a = 311.9 pm and c = 2425 pm (Cavani et al. 1991). The basal spacing of the present compound was determined to be d(003) = 802 pm. The results of XRD indicated that PA_1.5 is a hydrotalcite-like compound similar to iowaite. The SEM micrograph in Fig. 2 shows the morphology of the PA_1.5 surface, which consists of an agglomerate of flaky crystals.

Table 2. X-ray powder diffraction data.

Kohls and Rodda (1967) ^a		Braithwaite et al. (1994) ^a		Present work
dobs (pm)	R.I.meas^b	dobs (pm)	R.I.meas^b	dobs (pm)	R.I.meas^b	h	k	l
810.9	100	804.3	10	802.2	100	0	0	3
404.7	40	404.0	5	405.5	61	0	0	6
269.7	<1					0	0	9
263.9	17	263.7	4	262.7	52	0	1	2
236.3	27	235.3	4	234.5	51	0	1	5
201.9	23	201.4	4	201.1	30	0	1	8
180.5	4	179.1	<1			1	0	10
171.0	5	170.5	1			0	1	11
156.0	8	156.6	1	155.5	81	1	1	0
153.0	13	153.6	2	153.0	49	1	1	3

^a Data of Iowaites.

^b Relative intensity.

Figure 2. SEM image of PA_1.5 surface.

3.3. Adsorption characteristics of PA_1.5

3.3.1. Adsorption of NO₃−N and AEC

The ion exchange isotherm of NO₃⁻ at 25°C is shown in Fig. 3. The amount of adsorbed NO₃⁻ increased with increasing equilibrium concentration of NO₃⁻ in the aqueous solution. The positive charge of the soil (e.g., andosol) is the variable charge reduced by proton adsorption at low-pH condition (Nakahara and Wada 1993; Wada 1988). The positive charge of the PA_1.5 was caused by Fe³⁺ which was substituted for Mg²⁺ in the brucite sheet, and the NO₃⁻ adsorptive capacity of PA_1.5 was observed under a high-pH condition (Table 3). The hydrotalcite-like compound has positive charge and electrostatic balance is maintained by anion intercalation between the layers (Hibino 2006). The relationship between the amount of NO₃⁻ adsorption and the amount of Cl⁻ desorbed into equilibrium solutions is shown in Fig. 4. A high positive correlation was observed between the amount of NO₃⁻ adsorption and the amount of Cl⁻ desorption. It was surmised that the adsorption of NO₃⁻ occurs via ionic exchange with intercalated or adsorbed Cl⁻.

Table 3. The pH value at equilibrium and AEC of PA_1.5.

	AEC
pH	(cmolc kg^−1)
9.8	166

AEC: Anion Exchange Capacity.

Figure 3. NO₃^– exchange isotherm of the PA_1.5. The approximated curve was drawn by the formula of Langmuir.

Figure 4. Relationship between NO₃^– adsorption and the Cl^–desorption. Simple regression analysis was done by Excel software.

The AEC of PA_1.5 was 166 cmol_c kg⁻¹ (Table 3). This value of AEC is equivalent to approximately 23 g of N per kilogram of PA_1.5. This compound exhibits NO₃⁻ exchange capacity approximately equal to that of a pyroaurite-like compound synthesized using a conventional method (i.e., 170–190 cmol_c kg⁻¹) (Wuxian et al. 1995; Wuxian et al. 1996).

3.3.2. Adsorption selectivity of anions

Table 4 shows the adsorbed anion amounts, distribution coefficient (K_d), and anion exchange selectivity coefficient for PA_1.5 for anions in the solution containing one anion. The pH of the solution suggests that phosphate ions in the solution were divalent (HPO₄²⁻). The order of adsorbed anion amounts was NO₃⁻ > SO₄²⁻ > HPO₄²⁻. The distribution coefficients (K_d) for PA_1.5 for anions was also in the order NO₃⁻ > SO₄²⁻ > HPO₄²⁻. The logarithm of the anion exchange selectivity coefficient (K_GT) of NO₃⁻ related to Cl⁻ was above 0. However, coefficients of HPO₄²⁻ and SO₄²⁻ were below 0, and the coefficient for HPO₄²⁻ was lower than for SO₄²⁻. Therefore, it was inferred that the order of anion exchange selectivity on PA_1.5 was as follows: NO₃⁻ > Cl⁻ > SO₄²⁻ > HPO₄²⁻.

Table 4. The anion adsorption, selectivity coefficient (K_GT), and the distribution coefficient (K_d) for each anion on the equilibrium solution containing one kind of anion.

Anion				Cl^–
		Solution phase	Solid phase	Solution phase	Solid phase		Kd
Anions	pH	(mmol dm^−3)	(mmol kg^−1)	(mmol dm^−3)	(mmol kg^−1)	log KGT	(dm^3 kg^−1)
NO3^−	8.8	0.9	530	2.9	980	0.2	604
PO4^3−(HPO4^2−)	8.9	1.6	344	2.0	1210	−2.9	212
SO4^2−	9.8	1.4	393	3.3	872	−1.9	276

K_GT, Gaines–Thomas selectivity coefficient.

In the case of the solution with three anions, the amount of adsorbed NO₃⁻ was also greater than for SO₄²⁻ and HPO₄²⁻ (Table 5). Ion exchange selectivity usually increases when valence is high. These conflicting results between the valence of anions and the anion exchange selectivity may be explained by the interlayer space of PA_1.5 and the ionic radii of anions. The distance of the interlayer of PA_1.5 was 325.3 pm, which was calculated by subtracting the thickness of the brucite layer (476.9 pm: NipponKagakukai 1966) from the thickness (802.2 pm) of the unit layer. PA_1.5 was prepared as an LDH with intercalated Cl⁻ (ionic radius is 167 pm: Saito 2006), and the interlayer space of dried PA_1.5 was approximately twice the ionic radius of the intercalated anion (Hibino 2006). The ionic radius of NO₃⁻ is 165 pm (Saito 2006) with a planar structure that comprised oxygen bound to nitrogen. In contrast, sulfate SO₄²⁻ exhibits a tetrahedral ion structure, and its ionic radius is 230 pm (Saito 2006). These ionic radii indicate that NO₃⁻ can prioritize exchange with Cl⁻ over SO₄²⁻.

Table 5. The anion adsorption and distribution coefficient (K_d) for each anion on the equilibrium solution containing three kinds of anion.

		Solution phase	Solid phase	Kd
Anions	pH	(mmol dm^−3)	(mmol kg^−1)	(dm^3 kg^−1)
NO3^−		1.4	413	306
PO4^3− (HPO4^2−)	9.1	1.7	333	200
SO4^2−		2.1	230	111

Table 5 shows that the amounts of adsorbed NO₃⁻ and SO₄²⁻ and K_d were lower than for the solution containing a single anion species. It was suggested that there was competition for adsorption among coexisting ions. However, the adsorbed amount and K_d for HPO₄²⁻ did not differ from that in the solution with a single anion (Tables 4 and 5). In addition, the amount of Cl⁻ transferred from the solid phase to solution phase in HPO₄²⁻ solution was significantly smaller than for other solutions (Table 4). It was surmised that HPO₄²⁻ was adsorbed via other processes in addition to ionic exchange. Elution of Mg from Mg-Fe-Al-Cl hydrotalcite was reported at the initial pH of 7.0 or below, and phosphate ion uptake by coagulative precipitation was reported (Kuwabara et al. 2007). The present experiment was conducted with a solution adjusted to pH 7.0 from initial solution, and therefore it was surmised that phosphate uptake occurred via coagulative precipitation. The pH suggests that the precipitate was MgHPO₄. Moreover, the surface complexation of phosphate with FeOH or MgOH group at the edge of PA_1.5 can also be a phosphate removal process. Even if phosphate was taken up by coagulative precipitation or surface complexation by surface hydroxyl groups, it cannot be concluded that phosphate uptake by plant would be hindered. Because MgHPO₄ dissolves in a weak acid, coagulative precipitated phosphate is regarded as an available form. Moreover, despite complexation by surface hydroxyl groups, some plants can take up Fe- or Al-bound phosphorus by organic acids in root exudates (Otani and Ae 1996).

The competitive adsorption of anions by PA_1.5 on an imitative soil solution is shown in Table 6; similar to the aforementioned competitive adsorption experiment, more NO₃⁻ was adsorbed compared with SO₄²⁻. The most HPO₄²⁻ was adsorbed, and Cl⁻ was not adsorbed.

Table 6. The competitive adsorption for each anion in the imitative soil solution.

Anions	Equilibrium		Adsorption ratio (%)	Kd (dm^3 kg^−1)
	Solution phase (mmol dm^−3)	Solid phase (mmol kg^−1)
Cl^−	8.3	–	–	–
NO3^−	6.4	900	36	141
PO4^3− (HPO4^2−)	0.0	250	100	–
SO4^2−	1.7	68	14	39

The anion adsorption characteristics of PA_1.5 suggest that leaching of NO₃₋N will be reduced when PA_1.5 is applied to an agricultural field.

4. Conclusions

An Mg-Fe system hydrotalcite-like compound (PA_1.5) was synthesized using a new method, and its structure and anion adsorption characteristics were determined. The results were as follows.

1. PA_1.5 was synthesized using an easy, cost-effective method that simply involves mixing of a 8 mol dm⁻³ MgO suspension and a 1.5 mol dm⁻³ FeCl₃ solution, followed by a hydrothermal reaction without pH or gas control.

2. PA_1.5 is a hydrotalcite similar to iowaite; the layered structure was confirmed by XRD.

3. PA_1.5 has AEC with an AEC of 166 cmol_c kg⁻¹, equivalent to 23 g for N per kilogram of PA_1.5.

4. The adsorption of NO₃⁻ occurs via ionic exchange with intercalated or adsorbed Cl⁻.

5. The anion exchange selectivity coefficient (K_GT) of PA_1.5 was in the order NO₃⁻ > Cl⁻ > SO₄²⁻ >HPO₄²⁻.

6. PA_1.5 adsorbed HPO₄²₋ via other processes in addition to ionic exchange.

7. PA_1.5 is expected to control the leaching of NO₃₋N from farmland.

Acknowledgments

The author is grateful to Dr Masami Nanzyo for analysis by powder XRD, SEM at Tohoku University Graduate School of Agricultural Science, and Ho Ando for his valuable advice on this paper.