Improvement of the NH₂OH-HCl-HOAc method for extracting manganese and iron oxides in Japanese Andisols and other soil types in Japan

Abstract

We evaluated the validity of Tessier’s method as applied to the extraction of manganese (Mn) and iron (Fe) oxides in Japanese Andisols and other soil types in Japan. Using the original Tessier’s extractant mixture, 0.04 mol L⁻¹ hydroxylamine hydrochloride in 25% acetic acid (0.04 mol L^–1 NH₂OH-HCl in 25% HOAc), we found that substantial amounts of short-range-ordered Fe oxides were not extracted from allophanic Andisol samples and that considerable amounts of total Fe oxides were not extracted from all soil types. Relatively high extraction pH and large amounts of short-range-ordered Fe oxides in the Andisol samples might be responsible for incomplete extraction. Stoichiometric calculation indicated that the concentration of NH₂OH-HCl might be insufficient for complete extraction of Fe oxides. The extracted amounts of Mn and Fe increased with increasing concentration of NH₂OH-HCl in the extractant, and most of the Mn and Fe oxides in the soil samples, including samples with as much as 5.6% Fe, were extracted with 0.6 mol L^–1 NH₂OH–HCl in 25% HOAc. As judged from the simultaneous dissolution of aluminum (Al) and silicon (Si) minerals, extraction selectivity of Fe oxides with 0.6 mol L^–1 NH₂OH-HCl in 25% HOAc was comparable to that of the original Tessier’s method and better than that of a modified Community Bureau of Reference (BCR) sequential extraction procedure or a method using an extractant consisting of a mixture of oxalate and ascorbate, especially for Andisol samples.

Key words:

• Fe oxide
• heavy metals
• hydroxylamine hydrochloride
• Tessier

INTRODUCTION

Because of the importance of manganese (Mn) and iron (Fe) oxides to the dynamics of heavy metals in the environment, many methods for extracting Mn and Fe oxides and occluded heavy metals have been developed. A common method is extraction with acidified 0.5 mol L^–1 hydroxylamine hydrochloride (NH₂OH-HCl), the second-step fraction in the modified Community Bureau of Reference (BCR) sequential extraction procedure (SEP) (Sahuquillo et al. 1999). However, Suda et al. (2012) reported that this method was not suitable for Japanese Andisols, especially allophanic Andisols, which showed poor extraction selectivity and low rates of extraction of short-range-ordered Fe oxides such as ferrihydrite.

Dithionite-citrate has been used to extract both short-range ordered and well-crystalline Fe oxides in soil samples (McKeague and Day 1966). However, commercial dithionite could be contaminated with zinc; in addition, target metals would be precipitated as sulfide or sulfate (Gleyzes et al. 2002). Oxalate solutions and mixtures of oxalate solution and ascorbic acid have also been used for extracting heavy metals occluded in short-range ordered and total Fe oxides, respectively (Miller et al. 1986; Wilcke et al. 1998). One of the critical disadvantages of these oxalate-containing solutions is precipitation of some metals [cadmium (Cd) and lead (Pb)] as oxalate salts (Benitez and Dubois 1999).

Many researchers have used Tessier’s SEP (Tessier et al. 1979) for fractionation of heavy metals in soil samples. Tessier et al. (1979) used “sediment” samples, not “soil” samples, but most researchers have used Tessier’s procedure for the analysis of heavy metals in soil samples without any significant modifications. In the original Tessier’s method, sediment samples with heavy metals bound to Mn and Fe oxides are extracted with 0.04 mol L^–1 NH₂OH-HCl in 25% acetic acid (HOAc) in boiling water for 6 h with occasional agitation (HH₀₀₄-A extraction). Six hours was sufficient to extract all Fe oxides from the sediment samples in Tessier’s study. Therefore, the target phase of the Tessier’s method is generally considered to be heavy metals occluded in all Mn and Fe oxides, including short-range ordered to highly crystalline phases (e.g., Tack et al. 2006). However, on the basis of selective sequential extraction data and X-ray adsorption near-edge structure spectroscopy, La Force and Fendorf (2000) pointed out that part of the crystalline Fe oxide was not extracted from sediments contaminated with mine waste. This result indicates that the extractability of Fe oxides with HH₀₀₄-A extraction depends on the properties of the sediment or soil samples, especially their Fe oxide content and crystallinity.

Despite its widespread use, Tessier’s SEP has not been evaluated and sufficiently modified for various soil types. In Japanese agricultural fields, Andisols are among the most common soil types containing high amounts of Mn and Fe (Takeda et al. 2004). Mn oxides and short-range-ordered Fe oxides in Andisol samples are hardly extracted by 0.1 mol L^–1 NH₂OH-HCl (pH 2.0) and 0.5 mol L^–1 NH₂OH-HCl (pH 1.5), respectively (Suda et al. 2011, 2012). Inceptisols, classified as Brown Forest Soils in the Japanese soil classification system, are also widely distributed in Japan and generally contain abundant highly crystalline Fe oxides such as goethite. Fe oxides in these common Japanese soils might be incompletely extracted with HH₀₀₄-A extraction, and better extractants might be needed for these soils.

Here we evaluate the original Tessier’s HH₀₀₄-A extraction method as applied to samples of Japanese Andisols and other major soil types in Japan and improve the method to completely extract Mn and Fe oxides in these soils.

MATERIALS AND METHODS

Soil samples and their properties

Basic information about the three Entisols (E), three Inceptisols (I), five allophanic Andisols (AA), and three non-allophanic Andisols (NA) used in this study is compiled in Table 1. The soil samples were air-dried, passed through 2-mm mesh sieves, and stored at room temperature.

Table 1 Brief description of soil samples

	Soil classification			Site
Sample	Soil Taxonomy†	Japanese system‡	Horizon	(city, prefecture)	Land use
E-1	Aquents	Gray Fluvic soils	Ap	Suzaka, Nagano	Paddy Field
E-2	Aquents	Gray Fluvic soils	Ap	Tsukubamirai, Ibaraki	Paddy Field
E-3	Aquents	Gray Fluvic soils	Ap	Tochigi, Tochigi	Upland Field
I-1	Udepts	Haplic Brown Forest soils	Ap	Natori, Miyagi	Upland Field
I-2	Udepts	Haplic Brown Forest soils	Ap	Zentuji, Kagawa	Upland Field
I-3	Udepts	Haplic Brown Forest soils	Ap	Kanuma, Tochigi	Upland Field
AA-1	Udands	Allophanic Kuroboku soils	Ap	Tsukuba, Ibaraki	Upland Field
AA-2	Udands	Allophanic Kuroboku soils	Ap	Shimono, Tochigi	Upland Field
AA-3	Udands	Allophanic Kuroboku soils	Bw	Tateyama, Toyama	Forest
AA-4	Udands	Allophanic Kuroboku soils	Bw	Tsukuba, Ibaraki	Upland Field
AA-5	Udands	Allophanic Kuroboku soils	Bw	Chichibu, Saitama	Forest
NA-1	Udands	Non-allophanic Kuroboku soils	A	Oshaki, Miyagi	Forest
NA-2	Udands	Non-allophanic Kuroboku soils	A	Tateyama, Toyama	Forest
NA-3	Udands	Non-allophanic Kuroboku soils	O/A	Chichibu, Saitama	Forest

†According to USDA Soil Taxonomy (Soil Survey Staff 1998). ‡ According to Unified Soil Classification System of Japan 2nd Approximation (The Japanese Sociaty of Pedology 2003). E, Entisols; I, Inceptisols; AA, allophanic Andisols; NA, non-allophanic Andisols.

Soil pH was measured in H₂O or 1 mol L^–1 potassium chloride (KCl) with a glass electrode pH meter. The ratio of soil sample to H₂O or soil sample to 1 mol L^–1 KCl was 1:2.5 (weight/volume). Total carbon content of the soil sample was measured by a dry combustion method (Sumigraph NC-22F, Shimadzu, Kyoto, Japan). Dithionite-citrate extractable Fe (Fe_d) and Mn (Mn_d), and ammonium oxalate-extractable Al, Fe, Mn, and Si (Al_o, Fe_o, Mn_o, and Si_o) were extracted as described by Blakemore et al. (1987). Then the extract was centrifuged and filtrated by a 0.2-µm membrane filter (Merck Millipore, Burlington, MA, USA). Detailed extraction procedures for Fe_d, Fe_o, and Mn_o are described in subsequent paragraphs. Pyrophosphate-extractable Al and Fe (Al_p and Fe_p) were extracted and centrifuged as described by Blakemore et al. (1987), and the supernatant was filtered through a 0.025-µm filter (Merck Millipore) to minimize contamination of Fe from microcrystalline minerals (Schuppli et al. 1983). The concentrations of these elements in the filtrate were determined by inductively coupled plasma optical emission spectrometry (ICP-OES) (Vista-Pro, Varian, Palo Alto, CA, USA). Clay content was determined by a pipette method (Gee and Bauder 1986).

Experiment 1: Effects of increasing NH₂OH-HCl concentration in NH₂OH-HCl-HOAc extractant on the rates of extraction of Mn and Fe oxides

Extraction with 0.04, 0.2, 0.4, and 0.6 mol L^–1 NH₂OH-HCl in 25% HOAc solution

The extractants used in this experiment are NH₂OH-HCl in 25% HOAc (HH-A) solutions. To determine the optimal NH₂OH-HCl concentration for extraction of Mn and Fe oxides in soil samples, we evaluated the effect of increasing NH₂OH-HCl concentration in NH₂OH-HCl-HOAc extractant. For this purpose, we used NH₂OH-HCl concentrations of 0.04, 0.2, 0.4, and 0.6 mol L^–1 (HH₀₀₄-A, HH₀₂-A, HH₀₄-A, and HH₀₆-A).

An air-dried soil sample (1 g) was placed in a 50-mL polyethylene centrifuge tube, and 20 mL of HH-A solution was added. The mixture was then heated in boiling water for 6 h with occasional agitation. After extraction, the tube was cooled to approximately room temperature with running water for 10 min. After centrifugation at 3500 rpm (1700 g) for 10 min, the supernatant was filtered through a 0.2-μm mesh membrane filter, and the filtrate was diluted with 1% nitric acid. The pH of the residual supernatant was measured with a glass electrode pH meter. Fe, Mn, Al, and Si concentrations in the diluted solution were determined by ICP-OES. All extractions and measurements were carried out in triplicate. The extraction with original Tessier’s extractant of HH₀₀₄-A aims to extract both short-range-ordered and crystalline Mn and Fe oxides as well as metals occluded in these oxides.

Extraction of exchangeable and specifically adsorbed Mn and Fe

Exchangeable and specifically adsorbed Mn and Fe were extracted with a mixture of 1 mol L^–1 acetic acid and ammonium acetate (pH 5.0) (hereafter “acetate” solution). Air-dried soil (0.8 g) was placed in a 50-mL polyethylene centrifuge tube, 16 mL of acetate solution was added, and the mixture was mechanically shaken at room temperature for 2 h. After centrifugation at 3500 rpm (1700 g) for 10 min, the supernatant was filtered through a 0.2-μm mesh membrane filter, and the filtrate was diluted with 1% nitric acid. Mn and Fe concentrations in the diluted solution were determined by ICP-OES. All extractions and measurements were carried out in duplicate. Experimental conditions were based on those of Salbu et al. (1998).

Extraction of total Mn oxides

The amount of total Mn oxides was approximated using the value of oxalate-extractable Mn (Mn_o) according to Jarvis (1984). An air-dried soil sample (0.4 g) was placed in a 50-mL polyethylene centrifuge tube, and 40 mL of oxalate solution (mixture of 0.2 mol L^–1 oxalic acid and 0.2 mol L^–1 ammonium oxalate solution, pH 3.0) was added. The mixture was then mechanically shaken at 25°C for 4 h in the dark, a drop of polymeric flocculants (Superfloc, Kemira Co., Ltd., Helsinki, Finland) was added to an aliquot (about 25 mL) of the extract, and the resulting mixture was manually shaken. After centrifugation at 3500 rpm (1700 g) for 10 min, the supernatant was filtered through a 0.2-μm mesh membrane filter, and the filtrate was diluted with 1% nitric acid (HNO₃).

Oxalate is likely to extract Mn bound to organic matter in addition to oxide-forming Mn. Therefore, the amount of total Mn oxides should be estimated by the difference between the amount of oxalate-extractable Mn and the amount of organically bound Mn. Organically bound Mn is extracted with pyrophosphate (e.g., Iu et al. 1982) or hydrogen peroxide (e.g., Shuman 1979), but pyrophosphate extracts not only organically bound Mn but also oxide-forming Mn (Manh et al. 2010). Moreover, oxalate is produced by hydrogen peroxide treatment (Harada and Inoko 1977), resulting in dissolution of Mn oxides. Therefore, quantifying organically bound Mn is difficult. McBride (1982), however, showed that strong binding of Mn to organic matter is unlikely at pH < 6.0. Because the soil samples that are rich in organic matter in this study (i.e., the non-allophanic Andisols) are either moderately or strongly acidic (pH 4.43 to 5.67), the amount of organically bound Mn extracted is negligible. Thus, we approximated the amount of total Mn oxides as the Mn_o value.

Extraction of short-range-ordered Fe oxides

Oxalate extraction was used to approximate the amounts of short-range-ordered Fe oxides according to the method described by Blakemore et al. (1987). An air-dried soil sample (0.4 g) was placed in a 50-mL polyethylene centrifuge tube, and 40 mL of oxalate solution (0.2 mol L^–1 oxalic acid and ammonium oxalate solution, pH 3.0) was added. The mixture was mechanically shaken at 25°C for 4 h in the dark, a drop of the polymeric flocculants was added to an aliquot (about 25 mL) of the extract, and the resulting mixture was manually shaken. After centrifugation at 3500 rpm (1700 g) for 10 min, the supernatant was filtered through a 0.2-μm mesh membrane filter, and the filtrate was diluted with 1% HNO₃.

Extraction of total Fe oxides

Total Fe oxides were extracted with a dithionite-sodium citrate reagent. A ground, air-dried soil sample (0.4 g, < 212 μm) and dithionite powder (0.4 g) were placed in a 50-mL polyethylene centrifuge tube, 20 mL of 0.75 mol L^–1 sodium citrate solution was added, and the mixture was mechanically shaken at 25°C for 16 h. Ultrapure water (20 mL) and a drop of the polymeric flocculants were added, and the resulting mixture was manually shaken. After centrifugation at 3500 rpm (1700 g) for 10 min, the supernatant was filtered through a 0.2-μm mesh membrane filter, and the filtrate was diluted with 1% nitric acid. Fe concentration in the diluted solution was measured by ICP-OES. All extractions and measurements were carried out in triplicate. The experimental conditions were based on those of Blakemore et al. (1987). Both short-range-ordered and crystalline Fe oxides should be extracted under these conditions. Therefore, in this study, we consider dithionite-citrate extractable Fe as total Fe oxides.

Experiment 2: Comparison of extraction selectivity for Mn and Fe oxides for some major methods for extracting heavy metals occluded in Mn and Fe oxides

Extraction with 0.5 mol L^–1 NH₂OH-HCl (pH 1.5)

An air-dried soil sample (1 g) was placed in a 50-mL polyethylene centrifuge tube, 40 mL of 0.5 mol L^–1 NH₂OH-HCl (pH 1.5) solution was added, and the mixture was mechanically shaken at 25°C for 16 h. After centrifugation at 3500 rpm (1700 g) for 10 min, the supernatant was filtered through a 0.2-μm mesh membrane filter, and the filtrate was diluted with 1% nitric acid. The pH of the extract was measured with a glass electrode pH meter, and the Mn and Fe concentrations in the diluted solution were determined by ICP-OES. All extractions and measurements were carried out in triplicate. The experimental conditions were based on those used for the step 2 fraction (reducible fraction) in the modified BCR method (Sahuquillo et al. 1999). Hereafter, this NH₂OH-HCl extractant is designated HH_mBCR.

Extraction with a mixture of acid oxalate and ascorbic acid

The extraction reagent is acid oxalate (0.1 mol L^–1 oxalic acid and 0.175 mol L^–1 ammonium oxalate) plus 0.2 mol L^–1 ascorbic acid (hereafter Oxa-Asc). An air-dried soil sample (1 g) was placed in a 50-mL polyethylene centrifuge tube, and 30 mL of Oxa-Asc was added. The mixture was heated in boiling water for 1 h with occasional agitation and then cooled to approximately room temperature with running water for 10 min. After centrifugation at 3500 rpm (1700 g) for 10 min, the supernatant was filtered through a 0.2-μm mesh membrane filter, and the filtrate was diluted with 1% nitric acid. The Mn and Fe concentrations in the diluted solution were determined by ICP-OES. All extractions and measurements were carried out in triplicate. The experimental conditions were based on those reported by Shuman (1985), but slightly modified as described by Sadamoto et al. (1994). The modifications involved changes to the soil:solution ratio (from 1:50 to 1:30) and the extraction time (from 30 min to 1 h). The Oxa-Asc extraction method aims to extract both short-range-ordered and crystalline Mn and Fe oxides as well as the metals occluded in these oxides.

Calculation of the rates of extraction of Mn oxides and short-range ordered Fe oxides

We obtained the rates of extraction of Mn and Fe oxides for the acetate, HH-A, HH_mBCR, and Oxa-Asc extractants as follows:

where Mn_Ex and Fe_Ex are the amounts of Mn and Fe extracted by the HH-A, HH_mBCR and Oxa-Asc extractants. Mn_s is the amount of exchangeable and specifically adsorbed Mn extracted by the acetate solution, and Mn_o and Fe_o are the amounts of Mn and Fe extracted by oxalate. Fe_d is the amount of Fe extracted by dithionite-citrate. Exchangeable and specifically adsorbed metals do not form oxides, and Mn_s was likely extracted by the HH-A, HH_mBCR, and Oxa-Asc extractants. Therefore, Mn_s is subtracted from Mn extracted by the HH-A, HH_mBCR, and Oxa-Asc extractants. Because the amounts of exchangeable and specifically adsorbed Fe (Fe_s) were far less than the amounts of Fe extracted by the HH-A, HH_mBCR and Oxa-Asc extractants, as well as much less than Fe_o and Fe_d, we calculated the rate of extraction of Fe oxides by simply using the values of Fe extracted by the HH_mBCR and Oxa-Asc extractants.

RESULTS

Soil properties

Selected properties of the soil samples are listed in Table 2. The pH (H₂O) values of the soil samples ranged from acidic to neutral (pH 3.80 to 7.06). Total carbon (TC) ranged from 10.4 to 296 g kg^–1, and clay contents varied from 76.5 to 430 g kg^–1. The amounts of Fe_d ranged from 13.7 to 55.5 g kg^–1, and the amounts of Fe_o varied from 2.12 to 24.0 g kg^–1. Dithionite-citrate extractable Mn (Mn_d) was slightly higher than Mn_o for each sample. Al_p and Fe_p were abundant in the non-allophanic Andisol samples than other soil types.

Table 2 Selected properties of soil samples

Sample†	pH		TC	Fed‡	Alo	Feo	Mnd	Mno	Sio	Alp	Fep	Clay
	(H2O)	(KCl)					g kg^‒1
E-1	6.27	4.74	10.4	16.5	2.19	11.6	0.248	0.202	0.402	0.428	2.04	291
E-2	5.79	4.69	32.7	27.5	5.07	4.65	0.581	0.544	1.57	0.733	1.88	378
E-3	5.91	4.62	17.8	13.9	9.54	6.81	0.667	0.644	2.36	2.40	2.25	219
I-1	6.02	4.55	11.7	19.8	1.72	4.33	0.544	0.516	0.683	0.385	0.818	258
I-2	4.65	3.72	11.6	19.4	1.45	2.12	0.694	0.609	0.151	0.663	0.835	240
I-3	6.66	5.36	30.8	20.2	15.3	5.66	1.35	1.30	3.82	3.03	1.32	76.5
AA-1	7.06	5.75	42.0	55.5	52.1	18.6	1.20	0.918	24.5	3.12	0.479	279
AA-2	6.18	5.23	45.8	30.7	61.6	14.9	0.727	0.640	26.8	3.76	0.761	198
AA-3	4.71	4.22	23.5	23.9	23.0	15.6	0.644	0.581	6.18	4.49	1.60	87.0
AA-4	6.06	5.48	14.0	31.7	72.1	20.1	0.783	0.767	27.6	2.53	0.298	430
AA-5	5.28	4.89	30.0	32.4	42.4	24.0	0.551	0.535	15.2	3.12	0.355	162
NA-1	5.67	4.56	94.2	13.7	19.4	7.76	0.523	0.473	1.84	10.9	4.35	192
NA-2	4.86	3.86	183	17.5	16.3	9.82	0.170	0.141	1.45	8.67	5.85	289
NA-3	3.80	3.47	296	20.1	20.0	17.5	0.118	0.0865	2.83	11.7	9.90	219

† Some data are from Suda et al. (2012).

‡ Subscripts o, d, and p indicate oxalate-, dithionite-citrate-, and pyrophosphate-extractable elements, respectively. H₂O, water; KCl, potassium chloride; TC, total carbon; Fe, iron; Al, aluminum; Mn, manganese; Si, silicon; E, Entisols; I, Inceptisols; AA, allophanic Andisols; NA, non-allophanic Andisols.

Experiment 1: Effects of increasing NH₂OH-HCl concentration in NH₂OH-HCl-HOAc extractant on the extractability of Mn and Fe oxides

Extraction of exchangeable and specifically adsorbed Mn and Fe

The amount of exchangeable and specifically adsorbed Mn was 4.66 × 10^‒2 g kg^–1 on average, ranging from 2.73 × 10^‒3 to 0.243 g kg^–1. The amount of exchangeable and specifically adsorbed Fe ranged from 1.98 × 10^‒3 to 0.136 g kg^–1, with an average value of 3.53 × 10^‒2 g kg^–1.

Extracted amounts and rates of extraction of Mn and Fe for the HH₀₀₄-A, HH₀₂-A, HH₀₄-A, and HH₀₆-A extraction methods

The amounts of Mn extracted with HH-A (Mn_HHA) were comparable to or larger than those for oxalate extraction (Table 3). Mn_HHA and rates of extraction of Mn oxides increased slightly with increasing concentration of NH₂OH-HCl (Table 4). The rates of extraction of Mn oxides for the HH₀₀₄-A, HH₀₂-A, HH₀₄-A, and HH₀₆-A extraction methods from all samples averaged 117%, 130%, 136%, and 150%, respectively (Table 4).

Table 3 Amounts of manganese (Mn) extracted with hydroxylamine hydrochloride (NH₂OH-HCl) in acetic acid (HOAc) extractants

		Extracted amounts of Mn (g kg^‒1)
Soil type	Sample	Mno†	NH2OH-HCl concentration
			0.04 mol L^‒1	0.2 mol L^‒1	0.4 mol L^‒1	0.6 mol L^‒1
Entisols	E-1	0.202	0.203	0.249	0.285	0.322
	E-2	0.544	0.710	0.778	0.845	0.935
	E-3	0.644	0.732	0.788	0.769	0.877
	Average	0.463	0.548	0.605	0.633	0.711
	SD	0.232	0.300	0.309	0.303	0.339
Inceptisols	I-1	0.516	0.508	0.545	0.591	0.627
	I-2	0.609	0.774	0.787	0.824	0.831
	I-3	1.30	1.45	1.55	1.64	1.56
	Average	0.810	0.909	0.960	1.02	1.01
	SD	0.431	0.484	0.523	0.549	0.490
Allophanic Andisols	AA-1	0.918	0.985	1.06	1.07	1.17
	AA-2	0.640	0.650	0.722	0.746	0.820
	AA-3	0.581	0.591	0.656	0.685	0.742
	AA-4	0.767	1.01	1.17	1.19	1.27
	AA-5	0.535	0.601	0.682	0.715	0.801
	Average	0.688	0.767	0.859	0.882	0.962
	SD	0.155	0.211	0.240	0.233	0.244
Non-allophanic Andisols	NA-1	0.473	0.531	0.564	0.561	0.664
	NA-2	0.141	0.150	0.165	0.181	0.227
	NA-3	0.0865	0.0895	0.127	0.137	0.175
	Average	0.233	0.257	0.285	0.293	0.356
	SD	0.209	0.239	0.242	0.233	0.269

† Mn_o denotes oxalate-extractable Mn. SD, standard deviation.

Table 4 Rates of extraction of manganese (Mn) oxides with hydroxylamine hydrochloride (NH₂OH-HCl) in acetic acid (HOAc) extractants

		Extraction rate of Mn oxides (%)
		NH2OH-HCl concentration
Soil type	Sample	0.04 mol L^‒1	0.2 mol L^‒1	0.4 mol L^‒1	0.6 mol L^‒1
Entisols	E-1	100	134	160	187
	E-2	140	156	172	193
	E-3	114	123	120	138
	Average	118	138	151	173
	SD	19.9	16.6	27.1	30.4
Inceptisols	I-1	139	144	141	161
	I-2	96.9	111	127	140
	I-3	128	130	136	137
	Average	121	128	135	146
	SD	21.6	16.6	6.94	12.8
Allophanic Andisols	AA-1	113	122	129	114
	AA-2	107	116	117	128
	AA-3	102	113	117	129
	AA-4	102	113	118	128
	AA-5	132	153	155	167
	Average	111	123	127	133
	SD	12.5	16.8	16.6	19.6
Non-allophanic Andisols	NA-1	141	162	157	165
	NA-2	114	121	120	145
	NA-3	108	120	132	169
	Average	121	134	137	160
	SD	18.0	24.2	19.0	13.2
All	Average	117	130	136	150
	SD	15.9	17.3	18.4	23.8

SD, standard deviation.

The amounts of Fe extracted with the HH-A method (Fe_HHA) were highly dependent on the NH₂OH-HCl concentration (Table 5). The amounts of Fe extracted with the HH₀₀₄-A, HH₀₂-A, HH₀₄-A, and HH₀₆-A methods ranged from 4.69 to 20.9 g kg^–1, 9.21 to 31.3 g kg^–1, 11.9 to 37.0 g kg^–1 and 16.9 to 46.4 g kg^–1, respectively. For non-Andisol samples (except E-1 and E-3), the amounts of Fe extracted with HH₀₀₄-A extraction were generally greater than those for oxalate extraction but less than those for dithionite-citrate extraction. For Andisol samples, the amounts of Fe extracted with HH₀₀₄-A extraction were less than those with both oxalate extraction and dithionite-citrate extraction. Rates of extraction of short-range-ordered Fe oxides from the Entisols, Inceptisols, allophanic Andisols and non-allophanic Andisol samples with HH₀₀₄-A extraction averaged 133%, 158%, 74.4% and 80.0% (Table 6). Rates of extraction of total Fe oxides for the Entisols, Inceptisols, allophanic Andisols and non-allophanic Andisol samples averaged 48.2%, 28.9%, 42.6% and 52.1%. Rates of extraction of both short-range-ordered Fe oxides and total Fe oxides increased with increasing NH₂OH-HCl concentration. When the NH₂OH-HCl concentration was 0.6 mol L^–1, the rates of extraction of total Fe oxides from the Entisols, Inceptisols, allophanic Andisols and non-allophanic Andisol samples averaged 133%, 100%, 112% and 134%.

Table 5 Amounts of iron (Fe) extracted with hydroxylamine-acetic acid (NH₂OH-HOAc) extractants

		Extracted amounts of Fe (g kg^‒1)
		Feo†	Fed	FeHHA
Soil type	Sample			0.04 mol L^‒1	0.2 mol L^‒1	0.4 mol L^‒1	0.6 mol L^‒1
Entisols	E-1	11.6	16.5	11.0	16.8	20.4	25.9
	E-2	4.65	27.5	10.3	16.9	22.5	32.4
	E-3	6.81	13.9	5.62	9.21	12.3	17.4
	Average	7.67	19.3	8.97	14.3	18.4	25.2
	SD	3.54	7.25	2.92	4.42	5.35	7.54
Inceptisols	I-1	4.33	19.8	6.17	13.3	17.6	22.5
	I-2	2.12	19.4	4.69	10.7	15.0	18.7
	I-3	5.66	20.2	6.32	10.2	13.2	18.3
	Average	4.04	19.8	5.73	11.4	15.2	19.8
	SD	1.79	0.427	0.903	1.63	2.21	2.29
Allophanic Andisols	AA-1	18.6	55.5	14.9	24.3	31.6	42.8
	AA-2	14.9	30.7	9.44	15.7	19.6	26.1
	AA-3	15.6	23.9	9.35	15.6	20.0	26.8
	AA-4	20.1	31.7	16.4	28.1	35.7	45.5
	AA-5	24.0	32.4	20.9	31.3	37.0	46.4
	Average	18.7	34.8	14.2	23.0	28.8	21.2
	SD	3.70	12.0	4.91	7.14	8.45	10.2
Non-allophanic Andisols	NA-1	7.76	13.7	6.34	9.95	11.9	16.9
	NA-2	9.82	17.5	8.89	14.7	19.5	26.4
	NA-3	17.5	20.1	11.9	17.3	19.7	25.8
	Average	11.7	17.1	9.03	14.0	17.0	23.0
	SD	5.14	3.20	2.77	3.72	4.42	5.33

† Fe_o and Fe_d denote oxalate-extractable Fe and dithionite-citrate-extractable Fe, respectively. SD, standard deviation.

Table 6 Rates of extraction of iron (Fe) oxides with hydroxylamine hydrochloride (NH₂OH-HCl) in acetic acid (HOAc) extractants

		Extraction rate of short-range ordered Fe oxides (%)				Extraction rate of total Fe oxides (%)
		NH2OH-HCl concentration				NH2OH-HCl concentration
Soil type	Sample	0.04 mol L^‒1	0.2 mol L^‒1	0.4 mol L^‒1	0.6 mol L^‒1	0.04 mol L^‒1	0.2 mol L^‒1	0.4 mol L^‒1	0.6 mol L^‒1
Entisols	E-1	95.1	146	176	224	66.6	102	123	157
	E-2	222	364	484	698	37.4	61.5	81.7	118
	E-3	82.4	135	181	255	40.5	66.4	88.9	126
	Average	133	215	280	392	48.2	76.6	98.0	133
	SD	77.0	129	176	265	16.1	22.1	22.3	20.7
Inceptisols	I-1	143	306	405	519	31.1	66.9	88.6	113
	I-2	221	505	705	883	24.2	55.3	77.4	96.8
	I-3	112	181	233	323	31.3	50.7	65.1	90.6
	Average	158	331	448	575	28.9	57.6	77.0	100
	SD	56.3	163	239	284	4.03	8.37	11.7	11.8
Allophanic Andisols	AA-1	80.0	131	170	230	26.8	43.8	57.0	77.1
	AA-2	63.3	106	131	175	30.8	51.3	63.8	85.1
	AA-3	60.0	100	128	172	39.2	65.3	83.8	112
	AA-4	81.5	139	177	226	51.8	88.7	112.8	144
	AA-5	87.0	130	154	193	64.4	96.4	114.0	143
	Average	74.4	121	152	199	42.6	69.1	86.3	112
	SD	11.9	17.3	22.1	27.6	15.5	22.9	26.6	31.2
Non-allophanic Andisols	NA-1	81.7	128	154	218	46.2	72.4	86.9	123
	NA-2	90.5	149	198	269	50.9	83.9	112	151
	NA-3	67.7	98.7	112	147	59.1	86.1	98.0	128
	Average	80.0	125	155	211	52.1	80.8	98.8	134
	SD	11.5	25.4	43.0	61.2	6.52	7.33	12.3	15.0
All	Average	106	187	244	324	42.9	70.8	89.5	119
	SD	53.0	120	170	220	13.9	18.1	20.4	25.0

SD, standard deviation.

Thus, the amounts of Fe extracted with HH₀₆-A extraction were comparable to those with dithionite-citrate extraction. The pH values of the extracts decreased with increasing NH₂OH-HCl concentration (Fig. 1). Because pH values decreased during heated extraction (data not shown), the pH values after extraction were lower than before extraction in some cases. The average pH values after extraction were much higher for the allophanic Andisol samples than for the other soil types, and those for the non-allophanic Andisol samples were slightly higher than those for the Entisols and Inceptisols samples.

Figure 1 Relationship between hydroxylamine hydrochloride (NH₂OH-HCl) concentration and pH after extraction.

Experiment 2: Comparison of extraction selectivity for Mn and Fe oxides for some major methods for extracting occluded heavy metals

Amounts of Mn and Fe extracted with HH_mBCR and Oxa-Asc

The amounts of Mn extracted with HH_mBCR were comparable to those extracted with oxalate (Table 7). Oxa-Asc extracted smaller amounts of Mn than oxalate did from allophanic Andisol samples.

Table 7 Extracted amounts and rates of extraction of iron (Fe) and manganese (Mn) with 0.5 mol L^–1 hydroxylamine hydrochloride (NH₂OH-HCl) (pH 1.5) and with acid oxalate-ascorbic acid solution

				Extracted amounts (g kg^‒1)				Extraction rate of oxides (%)
	Mno†	Feo	Fed	Mn		Fe		Mn		Fe (SRO¶)		Fe (total)
Sample	g kg^‒1			HHmBCR‡	Oxa-Asc§	HHmBCR	Oxa-Asc	HHmBCR	Oxa-Asc	HHmBCR	Oxa-Asc	HHmBCR	Oxa-Asc
E-1	0.202	11.6	16.5	0.178	0.204	8.93	22.7	82.5	101	77.2	196	54.1	138
E-2	0.544	4.65	27.5	0.662	0.457	6.76	30.8	128	79.2	145	662	24.6	112
E-3	0.644	6.81	13.9	0.658	0.541	3.82	18.1	102	83.4	56.1	266	27.6	131
Average	0.463	7.67	19.3	0.499	0.401	6.50	23.9	104	87.9	92.9	375	35.4	127
SD	0.232	3.54	7.25	0.278	0.176	2.56	6.39	22.8	11.6	46.7	251	16.3	13.4
I-1	0.516	4.33	19.8	0.473	0.388	3.33	22.4	90.6	72.0	76.8	518	16.8	113
I-2	0.609	2.12	19.4	0.545	0.451	1.22	21.4	82.5	57.1	57.4	1008	6.29	111
I-3	1.30	5.66	20.2	1.37	1.03	3.10	20.0	105	78.5	54.7	354	15.3	99.2
Average	0.810	4.04	19.8	0.797	0.622	2.55	21.3	92.8	69.2	63.0	627	12.8	108
SD	0.431	1.79	0.427	0.500	0.352	1.16	1.20	11.6	11.0	12.1	340	5.68	7.44
AA-1	0.918	18.6	55.5	0.722	0.158	1.65	15.8	78.2	15.5	8.88	84.7	2.97	28.4
AA-2	0.640	14.9	30.7	0.470	0.430	1.38	18.1	73.1	66.8	9.27	121	4.50	58.9
AA-3	0.581	15.6	23.9	0.518	0.352	0.837	20.8	89.0	59.9	5.37	134	3.50	87.2
AA-4	0.767	20.1	31.7	0.818	0.397	2.71	14.9	107	51.6	13.5	74.1	8.56	47.1
AA-5	0.535	24.0	32.4	0.448	0.266	1.81	17.4	83.6	49.2	7.53	72.3	5.58	53.6
Average	0.688	18.7	34.8	0.595	0.321	1.68	17.4	86.1	48.6	8.90	97.2	5.02	55.0
SD	0.155	3.70	12.0	0.165	0.110	0.686	2.29	12.9	19.8	2.97	28.3	2.21	21.4
NA-1	0.473	7.76	13.7	0.437	0.523	1.42	18.0	92.1	111	18.3	232	10.3	131
NA-2	0.141	9.82	17.5	0.0730	0.156	1.73	26.3	31.1	115	17.6	267	9.91	150
NA-3	0.086	17.5	20.1	0.0701	0.121	2.08	22.4	76.7	149	11.8	128	10.3	111
Average	0.233	11.7	6.50	0.193	0.266	1.74	22.2	66.6	125	15.9	209	10.2	131
SD	0.209	5.14	2.56	0.211	0.222	0.328	4.15	31.7	20.9	3.54	72.5	0.246	19.5

† Subscripts o and d indicate oxalate- ad dithionite-citrate- extractable elements, respectively. ‡ HH_mBCR: 0.5 mol L⁻¹ NH₂OH-HCl (pH 1.5) extraction. § Oxa-Asc: acid oxalate-ascorbic acid extraction. ¶ SRO: Short-range ordered Fe oxides. Mn, manganese; Fe, iron; SD, standard deviation.

The amounts of Fe extracted from all samples (except E-2), especially Andisol samples, were smaller with HH_mBCR than with oxalate. Oxa-Asc extracted much greater amounts of Fe from non-Andisol samples than oxalate did; for some samples, the amount of Fe extracted was comparable to Fe_d. However, the amounts of Fe extracted from allophanic Andisol samples were comparable to Fe_o and much less than Fe_d. Oxa-Asc extracted smaller amounts of Fe from allophanic Andisol samples than from other soil samples, despite an abundance of Fe_o and Fe_d in the allophanic Andisol samples.

Amounts of Al and Si extracted with HH₀₀₄, HH₀₆, HH_mBCR, and Oxa-Asc

The amounts of Al extracted from the soil samples with the four extractants followed the order Oxa-Asc > HH₀₆-A > HH₀₀₄-A ≥ HH_mBCR (Fig. 2). The amounts of Al extracted from the Andisol samples were greater than those extracted from the non-Andisol samples. The order for Si was Oxa-Asc > HH₀₆-A > HH₀₀₄-A = HH_mBCR (Fig. 3). Greater amounts of Si were extracted from the allophanic Andisol samples than from the other soil samples.

Figure 2 Amount of aluminum (Al) extracted from soil samples with different extraction methods. HH_mBCR, HH₀₀₄-A, HH₀₆-A, and Oxa-Asc denote extraction with 0.5 mol L^-1 hydroxylamine hydrochloride (NH₂OH-HCl) (pH 1.5), 0.04 mol L^-1 NH₂OH-HCl in 25% acetic acid (HOAc), 0.6 mol L^-1 NH₂OH-HCl in 25% HOAc, and a mixture of acid oxalate and ascorbic acid, respectively. Bars represent means ± standard deviation (SD) (n = 3).


Figure 3 Amount of silicon (Si) extracted from soil samples with different extraction methods. HH_mBCR, HH₀₀₄-A, HH₀₆-A, and Oxa-Asc denote extraction with 0.5 mol L^-1 hydroxylamine hydrochloride (NH₂OH-HCl) (pH 1.5), 0.04 mol L^-1 NH₂OH-HCl in 25% acetic acid (HOAc), 0.6 mol L^-1 NH₂OH-HCl in 25% HOAc, and a mixture of acid oxalate and ascorbic acid, respectively. Bars represent means ± standard deviation (SD) (n = 3).


DISCUSSION

Relationships between extractability of Mn and Fe and NH₂OH-HCl concentration

The amounts of Mn extracted with HH-A was comparable to or greater than those extracted with oxalate and gradually increased with increasing NH₂OH-HCl concentration. Because Mn ions can be incorporated into crystalline Fe oxides (Ford et al. 1997), the small increase in the amounts of extracted Mn might be caused by the release of isomorphous-substituted Mn from the crystal lattice of these oxides. More Mn was extracted by HH₀₆-A than even Mn_d from most of the soil samples. This indicated that Mn was released from Fe oxides, such as coarse magnetite, which were relatively insoluble in dithionite-citrate. Most of the oxalate-extractable Mn is considered to form oxides. Therefore, we believe that the HH-A method extracted total Mn oxides from all soil samples used in this study.

For the non-Andisol samples, the amounts of Fe extracted with HH₀₀₄-A were the same as or greater than those extracted with oxalate. Most of the oxalate-extractable Fe originates from short-range ordered Fe oxides (Parfitt and Childs 1988). Therefore, HH-A completely extracted short-range-ordered Fe oxides from the non-Andisol samples, assuming that short-range ordered Fe oxides were extracted prior to most of crystalline Fe oxides.

The smaller amounts of Fe were extracted from the Andisol samples with HH₀₀₄-A than with oxalate (Tables 5 and 6). A substantial percentage of short-range ordered Fe oxide (allophanic samples: 13–40%, non-allophanic samples: 9–32%) was not extracted with HH₀₀₄-A. Because volcanic ashes contain magnetite, and oxalate extracts part of the oxide, the amounts of short-range ordered Fe oxides in samples of slightly weathered (i.e. low Fe_o) volcanic ash soils might be overestimated (Shoji et al. 1987). However, in a preliminary experiment, we found that more Fe was extracted from synthesized magnetite (Kanto Chemical Co., Inc., Tokyo, Japan) with HH₀₀₄-A than with oxalate. Therefore, we believe that, for our Andisol samples, some short-range ordered Fe oxides remained after extraction, with or without magnetite in the samples. La Force and Fendorf (2000) reported that HH₀₀₄-A extraction was insufficient for complete dissolution of crystalline Fe oxides in some polluted Fe-rich sediments. However, our study revealed that HH₀₀₄-A extraction would not extract a relatively large percentage of short-range ordered Fe oxides from non-Fe-polluted allophanic Andisol samples. For non-allophanic Andisol samples, Fe_o and Fe_d may include a measurable amount of Fe extracted from Fe–humus complex, as indicated by the high content of Fe_p (Table 2). This additional amount of Fe possibly led to underestimation of the rates of extraction of short-range ordered and total Fe oxides of the non-allophanic Andisol samples.

Extraction of Fe oxides by NH₂OH-HCl is highly dependent on extraction pH (Chao 1972). The pH values for the Andisol samples (allophanic samples: 2.43, non-allophanic samples: 2.15, on average) were higher than those for the non-Andisol samples (Entisols: 2.04, Inceptisols: 1.97, on average). Dissolution of abundant allophane-like minerals or amorphous Al oxides in allophanic Andisol samples might consume great amounts of protons during extraction. A decrease in proton concentration resulted in a decrease in the rates of both reduction and acid dissolution of Fe oxides. In addition, short-range ordered Fe oxides are more abundant in the Andisol samples than in the other soil samples. Therefore, larger amount of Fe oxides might result in a shortage of NH₂OH-HCl as a reducing agent during extraction. Stoichiometrically, when proton consumption by reduction of Fe³⁺ is compensated by hydrolysis of NH₂OH-HCl, 0.8 mmol of NH₂OH-HCl (i.e., 20 mL of 0.04 mol L^–1 solution) can reduce only 0.4 mmol (22.3 mg) of Fe³⁺. The same solution can reduce double the amounts (0.8 mmol, 44.7 mg) of Fe³⁺, assuming that all proton consumption is compensated by H⁺ from HOAc. The amount of NH₂OH-HCl in the extractant was stoichiometrically sufficient to reduce all short-range-ordered Fe oxides. However, because 4.22 to 19.1 mg of Fe was extracted by 20 mL of 0.04 mol L^–1 extractant, NH₂OH-HCl concentration during extraction should be remarkably decreased. Stoichiometric calculation also indicates that Tessier’s condition (20 mL of 0.04 mol L^‒1 NH₂OH-HCl for 1 g of sample) cannot reduce all Fe oxides from relatively Fe-rich, non-polluted soil samples (e.g., AA-1 contains greater than 44.7 mg of Fe_d per gram of air-dried soil sample). Moreover, since hydroxylamine (NH₂OH) itself reacts with O₂ and is oxidized on heating, the net decrease in NH₂OH-HCl during extraction would be greater than that calculated stoichiometrically. It might, however, be significantly important and not surprising that the original Tessier’s condition potentially has a severe shortage of NH₂OH-HCl to extract Fe oxides from soil samples in Japan.

On the basis of the foregoing discussion, the concentration of NH₂OH-HCl should be increased. Greater amounts of Fe were extracted with HH₀₂-A, HH₀₄-A, and HH₀₆-A than with HH₀₀₄-A (Table 5). Because NH₂OH was oxidized on heating and HCl remained in solution, extraction pH decreased with increasing NH₂OH-HCl concentration (Fig. 1). Therefore, this acceleration of Fe extraction was caused by the increase in both proton concentration and amount of reducing agent. On average, more Fe was extracted with HH₀₂-A and HH₀₄-A than with HH₀₀₄-A, and less Fe was extracted with HH₀₂-A and HH₀₄-A than with dithionite-citrate. However, the amounts of Fe extracted with HH₀₆-A were comparable to those extracted with dithionite-citrate. Therefore, most of the Fe oxides in the soil samples were extracted with HH₀₆-A. The amount of Fe extracted with HH₀₆-A was larger than that with dithionite-citrate. This indicates that HH₀₆-A extracts even from hardly-soluble Fe phases, such as coarse particles of crystalline Fe oxides. Among our soil samples, AA-1 contained the largest amount of Fe oxides (Table 2). A dataset of 151 Andisol samples (24 soil profiles) studied by Kurobokudo Co-operative Research Group (1986) indicates that about 85% of Andisol samples contained smaller amounts of Fe oxides than AA-1. Moreover, in general, Andisols (Andosols) contain much greater amounts of Fe than other soil types in Japan (Takeda et al. 2004). Hence, 0.6 mol L^–1 NH₂OH-HCl seems to be a sufficiently high concentration of NH₂OH-HCl to extract almost all Fe oxides in most Japanese soils.

Lower pH is also favorable for suppressing re-adsorption of metals during extraction (Gómez-Ariza et al. 1999). However, lower pH may accelerate the extraction of non-target phases such as Al and Si minerals. Therefore, we discuss the extraction selectivity of the HH₀₆-A method for Mn and Fe oxides in the next section.

Selectivities of HH₀₀₄, HH₀₆, HH_mBCR, and Oxa-Asc extractions for Fe oxides

The amounts of Al and Si extracted with the four extractants followed the order Oxa-Asc > HH₀₆-A > HH₀₀₄-A ≥ HH_mBCR for Al (Fig. 2) and Oxa-Asc > HH₀₆-A > HH₀₀₄-A = HH_mBCR for Si (Fig. 3). Oxa-Asc-extractable Al seems to include Al from the Al-humus complex, especially for the non-allophanic Andisol samples. For the other soil samples, much more Al was extracted with Oxa-Asc than with the other extractants, although Al_p values were small. Hence, we concluded that Oxa-Asc is a more destructive extractant for the Al and Si minerals than the other extractants. Less Al and Si were extracted with HH₀₆-A than with Oxa-Asc, and more Al and Si were extracted with HH₀₆-A than with HH₀₀₄ and HH_mBCR.

Heavy metal concentrations in Mn and Fe oxides are much higher than those in bulk soils (McKenzie 1989; Sipos et al. 2011). Therefore, the concentrations of heavy metals in Al and Si minerals are probably much lower than those in Mn and Fe oxides. However, Al and Si minerals more or less contain heavy metals in their lattices (Kitagawa 1973; Churakov and Dähn 2012), and the amounts of the metals derived from these non-target phases would increase as their dissolution proceeds during extraction. Hence, extraction selectivity for heavy metals in Fe oxides over Al and Si minerals is important if extracted metals are defined as a “reducible” or “Mn and Fe oxide occluded” fraction. However, this point has not been studied in detail. Because it is difficult to determine the amounts of heavy metals extracted from Fe oxides or from Al and Si minerals, we used Al/Fe and Si/Fe ratios as indicators for extraction selectivity. Because the amounts of Mn oxides were much less than those of Fe oxides, and most of the Mn oxides were extracted with each extraction method, Mn is ignored in the following discussion.

As judged by the selectivities indicated by the Al/Fe and Si/Fe ratios, HH₀₀₄-A and HH₀₆-A were better extractants than HH_mBCR and Oxa-Asc (Table 8). Because of its relatively strong acidity, HH_mBCR extracts large amounts of Al. The results for the non-allophanic Andisol samples indicate that substantial amounts of Al were extracted from the Al-humus complex, although this extraction has been carried out before organically bound heavy metals are extracted in the modified BCR SEP. In contrast, HH_mBCR extracted very small amounts of Fe from the Andisol samples, resulting in remarkably high Al/Fe ratios in these samples, being consistent with the results of Suda et al. (2012). Oxa-Asc extracted large amounts of Fe except for the allophanic Andisol samples. Unexpectedly, Oxa-Asc extracted much smaller amounts of Fe than Fe_d from the allophanic Andisol samples. This result may be due to a lack of oxalate; however, more detailed studies might be needed. The Al/Fe and Si/Fe ratios for HH₀₆-A extraction, as well as for HH₀₀₄-A extraction, were lower than those for HH_mBCR and Oxa-Asc extraction, especially for Andisol samples. Hence, we concluded that the selectivity of HH₀₆-A extraction was comparable to that of Tessier’s HH₀₀₄-A extraction and better than that of HH_mBCR and Oxa-Asc extractions.

Table 8 Comparison of rates of extraction of iron (Fe) and manganese (Mn) oxides, and aluminum (Al)/Fe and silicon (Si)/Fe mole ratios, for four extraction methods

Extraction methods	Extraction rate of oxides (%)			Mole ratio
Soils	Mn	Fe (SRO†)	Fe (total)	Al/Fe	Si/Fe
HHmBCR‡
Entisols	104	92.9	35.4	1.42	0.341
Inceptisols	92.8	63.0	12.8	2.98	0.766
Allophanic	86.1	8.90	50.2	17.7	4.86
Non-allophanic	66.6	15.9	10.2	12.5	0.967
Average	87.5	45.2	27.2	8.65	1.73
SD	15.8	39.9	19.1	7.77	2.10
Oxa-Asc§
Entisols	87.9	375	127	1.23	0.425
Inceptisols	69.2	627	108	1.35	0.450
Allophanic	48.6	93.3	62.6	5.31	1.43
Non-allophanic	125	209	131	2.34	0.392
Average	82.7	326	107	2.56	0.673
SD	32.5	231	31.3	1.90	0.503
HH004-A¶
Entisols	118	133	48.2	0.644	0.354
Inceptisols	121	158	28.9	1.06	0.619
Allophanic	111	74.4	42.6	3.00	0.493
Non-allophanic	121	80.0	52.1	2.18	0.294
Average	118	111	43.0	1.72	0.440
SD	4.69	40.8	10.1	1.07	0.146
HH06-A
Entisols	173	392	133	1.02	0.280
Inceptisols	146	575	100	1.16	0.339
Allophanic	133	199	112	2.27	0.268
Non-allophanic	160	211	134	2.14	0.308
Average	153	345	120	1.65	0.299
SD	17.1	177	16.7	0.647	0.0315

† SRO: Short-range ordered Fe oxides. ‡ HH_mBCR: 0.5 mol L^–1 hydroxylamine hydrochloride (NH₂OH-HCl) (pH 1.5) extraction. § Oxa-Asc: acid oxalate-ascorbic acid extraction. ¶ HH₀₀₄-A, HH₀₆-A: Extraction with 0.04 mol L^–1 and 0.6 mol L^–1 NH₂OH-HCl in 25% acetic acid (HOAc), respectively. SD, standard deviation.

CONCLUSION

Tessier’s HH₀₀₄-A extraction method might be unsuitable for allophanic Andisol samples because substantial amounts of short-range ordered Fe oxides are not extracted by this method. Higher extraction pH and relatively large amounts of short-range ordered Fe oxides might be responsible for incomplete extraction. For all soil samples in this study, a considerable amount of crystalline Fe oxides remained after HH₀₀₄-A extraction. However, increasing the concentration of NH₂OH-HCl increased the amounts of extracted Mn and Fe, and 0.6 mol L^–1 of NH₂OH-HCl was sufficient to extract almost all Mn and Fe oxides from the soil samples. The selectivity of the HH₀₆-A method for Fe oxides over Al and Si minerals was comparable to or better than that of the original Tessier’s method (HH₀₀₄-A). In contrast, the selectivities of the HH_mBCR and Ox-Asc methods were much lower than those for the HH-A extractions (Tessier’s extraction, HH₀₆-A extraction), especially when the HH_mBCR method was applied to both allophanic and non-allophanic Andisol samples and when the Oxa-Asc method was applied to allophanic Andisol samples. Therefore, the HH₀₆-A extraction method seems to be better than the HH₀₀₄-A, HH_mBCR, and Oxa-Asc extraction methods, especially for Japanese Andisol samples.

ACKNOWLEDGMENTS

We acknowledge Dr. S. Ono, Dr. T. Ito (Tohoku University), Dr. K. Kamewada, Dr. A. Aoki (Tochigi Prefecture Agricultural Experiment Station) and Dr. Y. Nagasaki (National Agricultural Research Center for the Western Region) for soil sampling. We also thank Dr. S. Wakabayashi (Agricultural Research Center) and Dr. J. Takahashi (University of Tsukuba) for providing some of the soil samples used in this study and for providing physicochemical data for some of the samples. We also thank Dr. Y. Maejima and Dr. I. Akahane (National Institute for Agro-Environmental Sciences) and Dr. K. Tamura (University of Tsukuba) for helpful comments. This work was supported by a Grant-in-Aid for Research Fellows of the Japan Society for the Promotion of Science (No. 23-444).