Introduction

Tea is a very important crop in tropical and subtropical regions. Worldwide, the total area of tea plantations is about 2.4 million ha, and China is the single largest producer with about 1.0 million ha (Wang et al. 1997). Tea plants grow well in soils of low pH and high aluminum (Al) concentrations. Recently, Ma et al. (2000) observed that soils of tea plantations had very low pH of <4.0 in at least 44% of a total of 212 soil samples from the three major tea producing provinces of China during 1998–1999. Several studies have shown similarly low pHs produced under tea plantations, while studying the tea farming system (Ding and Huang 1991; Ma et al. 2000; Ruan et al. 2000; Fung et al. 2003; Xu et al. 2003). Acidification is determined by the age of the tea plantation, the origin of soil parent material and the use of fertilizers (Obukhov et al. 1982; Abe et al. 2006; Oh et al. 2006; Wu et al. 2006).

Tea plants are known to be typical Al accumulators and can take up large amounts of Al, most of which is accumulated in the leaves. The uptake of Al by tea plants was determined with forms of labile Al in the acidic soils (Dong et al. 2001). Biogeochemical cycling of Al in tea litter may also cause soil acidification in tea plantations (Ding and Huang 1991); however, most studies have attributed soil acidification in tea plantations to heavy application of chemical fertilizer (Abe et al. 2006; Oh et al. 2006; Ruan et al. 2006). Application of ammonium (NH4 +)–N fertilizers to increase yields of tea can greatly accelerate soil acidification through nitrification of NH4 + (Ruan et al. 2000, 2004; Oh et al. 2006).

Although adapted to acidic soils, the growth and product quality of tea plants can be negatively affected by the acidification of soil and other soil processes under tea plantations (Fung et al. 2008). The long-term exploitation of soils in tea plantations can lead to different types of soil degradations: (1) chemical, e.g. decreasing organic matter contents, lower exchangeable base cations, higher levels of toxic Al, leaching of nutrients, and accumulation of toxins (polyphenols) in tea leaves (Wang et al. 1997; Senapati et al. 1999; Abe et al. 2006; Oh et al. 2006); (2) physical, e.g. reduced water-holding capacity, compaction of the soil surface and soil erosion (Wang et al. 1997; Senapati et al. 1999); and (3) biological, e.g. loss of up to 70% of important soil biota (Han et al. 2007). Heavy fertilization of tea fields results in chemical pollution in local environments by leaching, runoff and/or erosion of applied nutrients (Abe et al. 2006; Oh et al. 2006). Tea plantations and soil acidification of tea plantations also induces activation of soil heavy metals such as Pb, Cu, Cd, Cr and Zn and increases accumulation of these metals in tea leaves (Zhang et al. 2006; Zhang and Fang 2007), which not only affects tea quality but also threatens human health.

The effects of soil acidification on silicate mineral weathering have been investigated extensively. Soil acidification induced by agricultural stress and acid deposition accelerates the chemical weathering of soil minerals and leads to leaching of major cations and transformation of minerals in acidic soils (Starr and Lindroos 2006; Herre et al. 2007; Nakao et al. 2009; Pierson-Wickmann et al. 2009a). Although there have been investigations of soil acidification in tea plantations, little information is available on the effect of soil acidification induced by tea cultivation on soil mineralogy and magnetic properties. Our study investigated changes in the following properties of Alfisols in eastern China under a tea plantation aged 17 years: (1) basic soil properties, (2) chemical composition, (3) mineralogical composition, and (4) magnetic properties.

Materials and methods

Nanjing is in a transitional region between subtropical and temperate regions of eastern China, where an Alfisol (Yellow Brown Earth) developed from loess is the characteristic zonal soil. Soil samples from two profiles, under a 17-year-old tea plantation and unused land (control), were collected for investigation. The standard irrigations and applications of chemical fertilizers were used for the entire experimental period.

Soil samples were collected in October 2005, air-dried and ground to pass a 60-mesh sieve. Soil pH was determined using a glass electrode with the ratio of soil:water 1:2.5. Cation exchange capacity (CEC) was determined using the ammonium acetate method, and soil organic matter by the dichromate method. Soil exchangeable H+ and Al3+ were extracted by 1.0 mol L−1 KCl and then determined by base titration. Soil exchangeable base cations were extracted with 1.0 mol L−1 ammonium acetate; the Ca2+ and Mg2+ ions in solution were determined by atomic absorption spectroscopy, and the Na+ and K+ ions by flame photometry.

Free Fe and Al oxides were extracted with the Na-dithionite-citrate-bicarbonate system (Fed and Ald). The poorly crystalline oxides were extracted with NH4-oxalate (Feo and Alo). Fe and Al concentrations in the extractants were measured using ICP-AES.

A SPECTROSCAN MAKC-GV desktop XRF crystal diffraction scanning-spectrometer was used to determine elemental concentrations. The results were optimized using the computer simulation program, and were preliminarily calibrated using standard soil samples of known elemental concentrations.

Mineralogical composition of soil clay fractions was analyzed by X-ray diffraction (XRD, CuKα radiation) method. Clay fractions (<2 μm) were separated by sedimentation and centrifugation. Oriented specimens were prepared by sedimentation on glass slides and step-scanned in steps 0.1 °2θ and 10 s/step counting time. Mg-saturated clays were examined at room temperature, after ethylene glycol solvation and also after heating at 300 and 550°C. K-saturated clays were examined at room temperature and after heating at 100, 300 and 500°C.

Mineralogy of Fe oxides was examined by magnetic methods. Environmental magnetism is a relatively new direction in science and involves numerous interdisciplinary studies including soils. Detailed overviews of the state and progress in such investigations were presented in several monographs (Thompson and Oldfield 1986; Maher and Thompson 1999; Evans and Heller 2003; Maher 2008). The magnetic characteristics of sediments and soils reflect the amount and quality of ferruginous minerals they contain and were first found to be related to their content, mineralogy and grain size. These magnetic techniques are sensitive even at trace amounts of magnetic minerals. The soil samples were examined using a range of room-temperature magnetic measurements, including low- and high-frequency magnetic susceptibility (MS), anhysteretic remanence (ARM), incremental acquisition of magnetic remanence (IRM), demagnetization, and a high-field IRM (HIRM). Each sample was dried and packed into 10-cm3 plastic cylinders. MS (χ) reflects the total concentration of ferrimagnetic or total concentration of paramagnetic minerals and antiferromagnetic by low content of ferromagnetic. MS was measured at low (0.46 kHz) and high (4.6 kHz) frequencies, using a Bartington Instruments MS2 susceptibility meter.

ARMs were imparted in an alternating field of 80 mT with a biasing field of 0.08 mT (Molspin AF demagnetizer, with ARM attachment). The susceptibility of ARM (χARM) was calculated by normalizing the ARM by the intensity of the applied bias field. IRMs were imparted in pulsed fields of 10, 20, 50, 100 and 300 mT (Molspin pulse magnetizer) and a field of 1,000 mT (Newport 4” Electromagnet). All magnetic remanences were measured using a fluxgate magnetometer (Molspin Ltd, sensitivity ~10−7 A m2). The magnetic hardness of the high-field IRM (HIRM, i.e. acquired beyond 100 mT) was also examined, first by applying a magnetic field of 1 T to representative samples, to produce an SIRM, which was then demagnetized in a field of 100 mT. The remanence remained following this demagnetization (the HIRM100 mT af), reflected the concentration of stable, high-coercivity minerals, such as haematite (Liu et al. 2002; Maher et al. 2003).

Results and discussion

Soil properties

The granulometry of the parent loess and soil samples is a sandy clay loam, where fine sand (0.25–0.05 mm) and clay predominate. Both profiles were characterized by the eluvial–illuvial redistribution of clay, and upper portions of soils were enriched with fine sand (Table 1).

Table 1 Granulometry of Alfisol-control profile and profile under tea plantation
Full size table

Control soil had pH 6.0–7.3 (Table 2), and the tea plantation soil had pH 4.5–5.7. The acidification under tea occurred within a depth of 70 cm; however, the maximum difference in pH (∆pH = 2.80) was in the upper 17 cm. With the decreased pH there was a decline in organic matter content and a change in the exchangeable complex of soil under the tea plantation. No significant concentration of exchangeable H+ was accompanied as by the appearance of exchangeable Al3+. Control soil had CEC and exchangeable Ca2+ of 1.5 and six times higher, respectively, compared to tea plantation soil.

Table 2 Chemical properties of a natural soil and soil from a tea garden
Full size table

Compared with other soil samples of the tea plantation, the rhizosphere soil was characterized by higher organic matter content, CEC, and exchangeable K+ and Ca2+; however, concentrations of exchangeable Al3+ and Na+ were lower.

In both soil profiles free Fe and Al oxides were predominant: Feo/Fed < 0.2 and Alo/Ald < 0.6 (Table 3). These ratios were larger for the control than the tea plantation soil. Tea plantation soil was characterized by higher Ald and Fed and lower Feo. The maximum differences were for subsoil (depth 35–70 cm).

Table 3 Soil free Fe/Al oxides and amorphous Fe/Al oxides
Full size table

Chemical composition

Total concentrations of the most important elements in these studied soils along with their clay fractions are given in Tables 4 and 5, respectively. The bulk mass of soil contained >70% silica (SiO2) followed by Al2O3 (≈10%) (Table 4). The profile that developed under the tea plantation was characterized by leaching of Fe and Al and accumulation of Si (Table 5). In many soils, the most important hydrolysis reactions are the incongruent dissolution of minerals such as feldspars to yield clays and alkali and alkaline earth cations in solution. A useful proxy for the progress of this reaction in soils is the molar ratio of Al (representing clay) to the sum of Ca, Mg, Na and K (representing major cationic nutrients lost into soil solution). A large database of North American soils (Sheldon et al. 2002; Sheldon and Tabor 2009) showed that a similar ratio was usually <2 for fertile soils, but >2 in less fertile soils. Another molar ratio (Rb/Sr) reflects leaching of Rb and Sr, and indicates the weathering effect based on the differential partitioning of the two elements: Rb is associated with K feldspar, while Sr is associated with carbonate and other easily weatherable minerals (Gallet et al. 1996; Chen et al. 1999). The values of this ratio were consistently higher for the profile under tea compared to controls, confirming a significantly higher degree of weathering intensity in this profile. Rhizosphere samples showed the maximum concentrations of Al2O3 and MnO, and high contents of Fe2O3 and Rb/Sr, under tea.

Table 4 Total elemental concentrations in the tea garden soil and control soil profile (bulk samples)
Full size table
Table 5 Total elemental concentrations in the tea garden soil and control soil profile (clay fraction)
Full size table

Mineralogical composition of clay fraction

The parent loess contains four main minerals: vermiculite, mica, kaolinite and quartz, along with traces of gibbsite and goethite. Some decrease of the intensity of the 14-Å peak after ethylene glycol solvation could be connected with the presence of swelling (smectite type) layers within the vermiculite structure (Fig. 1). Heating of specimens showed vermiculite hydroxy-interlayering or vermiculite chloritization. Heating to 300 and 500°C resulted in the non-complete collapse of the 14-Å peak and development of a visible asymmetry of the 10-Å peak with low angles. The stability of interlayers depends on the degree of interlayer filling and their composition. The XRD spectra of the samples heated at 300°C indicate lower stability of the interlayer. Barnhisel and Bertsch (1989) found a XRD examination of heat-treated K-saturated samples was the most useful procedure to determine the degree of interlayer filling. Spectra of K-saturated specimens heated at 25, 100, 300 and 500°C are shown in Fig. 2. The temperature required to collapse the 14-Å peak toward 10 Å can be used to estimate the relative degree of filling, as the higher the temperature, the larger the degree of filling. In this case, room-temperature K-saturated specimens gave almost complete collapse, which can be explained by the high-charge vermiculite phase (originated from mica). The pattern from the 500°C-heated sample indicated that no chlorite was present (i.e. no 14-Å peak); however, the presence of the 12-Å peak was due to the hydroxy-interlayer form of vermiculite. The degree of filling and their thermal stability was low. This mineral is thought to originate from weathering of mica following the sequence: mica → vermiculite → chloritized vermiculite → kaolinite + oxides (Barnhisel and Bertsch 1989). Chloritized vermiculite originates from the deposition of hydroxy-Al polymeric components within the interlayer space of vermiculite. It was suggested that moderately acidic conditions, low organic matter content, oxidizing conditions and frequent wetting–drying cycles are optimal for the hydroxy-Al interlayer formation (Rich 1968). The control profile was non-differentiated and mineralogy of all studied layers was identical.

Fig. 1
figure 1

X-ray diffraction (XRD) patterns of the <2-μm-fraction for the control Alfisol profile. a Mg-saturated. b Ethylene glycol solvated. c Heated to 300°C. d Heated to 500°C (values in Å)

Full size image
Fig. 2
figure 2

X-ray diffraction (XRD) patterns of the <2-μm-fraction Mg-saturated, K-saturated and heated to 100, 300 and 500°C for the control Alfisol profile. a 0–20 cm. b 50–100 cm (values in Å)

Full size image

The profile under the tea plantation showed a visible decrease of intensity of the 14-Å peak on the Mg-saturated spectra of all three horizons, with the maximal change in the topsoil (Fig. 3). The further decrease of the intensity of this peak after ethylene glycol solvation could be due to further dissolution of vermiculite. Specimens heated at 300 and 500°C showed a nearly complete collapse of the 14-Å peak, which was confirmed by XRD examination of K-saturated specimens that were heated at 100, 300 and 500°C. Both subsoil and surface horizons showed only small asymmetry of the 10-Å peak on the spectra heated at 500°C (Fig. 4). Thus we concluded that the acidification induced by the tea plantation led to the dissolution of hydroxy-Al interlayers within the structure of vermiculite and the dissolution of vermiculite. The maximum intensity of these processes was within the upper 35 cm, where the decrease of pH was maximum (ΔpH = 2.20–2.80).

Fig. 3
figure 3

X-ray diffraction (XRD) patterns of the <2-μm-fraction for the Alfisol profile under the tea plantation. a Mg-saturated. b Ethylene glycol solvated. c Heated to 300°C. d Heated to 500°C (values in Å)

Full size image
Fig. 4
figure 4

X-ray diffraction (XRD) patterns of the <2-μm-fraction Mg-saturated, K-saturated and heated to 100, 300 and 500°C for the Alfisol profile under the tea plantation. a 0–17 cm. b 35–70 cm (values in Å)

Full size image

Obtained data showed good correlations with results in Tables 2 and 3, where the soil profile under the tea plantation, the increase of exchangeable Al, and free Al oxides (Ald) were shown. The source of Al could be both hydroxy-Al interlayers and vermiculite. It is known that in the acid-soil environments vermiculite and smectite are stabilized by hydroxyl-Al polymeric interlayers, and inhibit gibbsite formation (the so-called antigibbsite effect). Gibbsite does not form until the expandable 2:1 minerals are decomposed (Barnhisel and Bertsch 1989; Robert 1995).

Magnetic properties

The magnetic data obtained for the soils is summarized in Table 6. MS values ranged from ≈40 × 10−8 to 160 × 10−8 m3/kg in the bulk samples and ≈60 × 10−8 to 360 × 10−8 m3/kg in the clay fraction (Fig. 5). MS represented the total contribution of Fe-bearing minerals in the mineral assemblage and was usually controlled by the total ferrimagnetic concentration (Thompson and Oldfield 1986). The parent material susceptibility values (≈150 × 10−8 m3/kg) indicated a magnetite (maghemite) concentration of ≈0.4%. Compared to the control, the tea plantation soil profile showed lower MS, which was correlated with the leaching of Fe. The major content of Fe oxides was associated with the clay fraction.

Table 6 Magnetic properties of bulk soils and clay fraction
Full size table
Fig. 5
figure 5

Magnetic susceptibility data for control and tea plantation soil profiles. a Bulk soils. b Clay fractions

Full size image

Values of frequency-dependent susceptibility, which reflect the concentration of superparamagnetic (SP), ultrafine-grained (< ≈30 nm in magnetite and maghematite) particles (Maher 1988; Dearing et al. 1996), had a range of 8–14%. This clearly demonstrated the presence of all ferrimagnetics in the super-dispersed form (Table 6). ARM (expressed here as the susceptibility of χARM) was particularly sensitive to the presence of sub-micrometer ferrimagnets, ≈30 nm (Özdemir and Banerjee 1982; Maher 1988). ARM and MS showed a strong direct correlation, both for the tea plantation and control profiles, while χARM decreased twofold under the tea plantation (Fig. 6). Maximum IRM values for the C-horizons were ≈100 × 10−4 Am2/kg, with ≈80% of the IRM acquired within 10–100 mT, and only ≈20% acquired at high applied fields of 300–1,000 mT. The majority of ferrimagnetics were fully saturated at fields of 100 mT; thus IRM100 allowed the evaluation of ferrimagnet contents (magnetite and maghemite).

Fig. 6
figure 6

Incremental acquisition of magnetic remanence (IRM) data for control and tea plantation soil profiles. a Bulk soils. b Clay fractions (1- IRM100, 10−5 Am2/kg; 2- HIRM100, 10−5 Am2/kg)

Full size image

As with the MS data, the IRM (SIRM) values indicate varying ferrimagnetic concentrations in soils. ARM and MS showed a strong direct correlation. In comparison with control, tea plantation soil had a clear change in magnetic mineralogy, the decreased content of ferrimagnetic minerals (magnetite and maghemite), which were not stable in acid conditions.

General discussion

It is well known that the cultivation of tea plants can acidify soil. However, the effect of soil acidification induced by tea plantations on magnetic characteristics and mineralogical properties of soils has remained unknown. The data of our study show that acidification occurred within 70-cm depth of the soil profile under a tea plantation for 17 years. The decline in soil pH was up to 2.80 units for the upper 17 cm, compared with samples from the unused control soil. Soil acidification resulted in decreases in soil exchangeable Ca2+ and K+, soil CEC and soil total Ca, and the increase in soil exchangeable Al. Soil acidification under tea accelerated the chemical weathering of soil minerals and thus led to CaO depletion, and accumulation of SiO2 and free Fe and Al oxides. This is similar to observations in soils influenced by acid deposition and agricultural practices (Guicharnaud and Paton 2006; Starr and Lindroos 2006; Pierson-Wickmann et al. 2009b).

XRD spectra revealed that soil acidification induced by the tea plantation led to the dissolution of hydroxy-Al interlayers within the structure of vermiculite, the dissolution of vermiculite, and the transformation of soil minerals following the sequence: vermiculite → chloritized vermiculite → kaolinite + oxides. The decomposition of 2:1 minerals and the formation of 1:1 minerals were the main reasons for decreased soil CEC. Traditionally, the decreased soil CEC due to acidification has been ascribed to the blockage effect of hydroxyl-Al polymers formed during soil acidification on negatively charged sites of soil (Ulrich 1991). However, the decline of negatively charged sites resulting from the transformation of hydroxyl-interlayered vermiculite–vermiculite–smectite was recently observed in upper layers of podzolic soils under very acidic conditions (Nakao et al. 2009). This is similar to the observations of our study.

The data of MS and IRM (SIRM) indicated that ARM and MS were strongly correlated with each other. Compared with control soil, the content of ferrimagnetic minerals (magnetite and maghemite) in tea plantation soil was lower due to serious acidification from tea cultivation. This is consistent with the content of total Fe oxides in the two soil profiles (Table 4). In soil under the tea plantation, the content of Fe2O3 in the 35–70-cm layer was much greater than in the 0–35-cm layer and also than in the entire control soil profile. These results, incorporated with the data of MS and IRM, suggest that dissolution of ferrimagnetic minerals and movement of Fe from upper to deeper layers occurred because these minerals are not stable under acid conditions. Therefore, soil acidification under the tea plantation accelerated not only the transformation of silicate minerals, but also the changes in ferrimagnetic minerals in the soil.

Conclusions

Comparison of chemical and mineralogical compositions and magnetic properties of Alfisols from eastern China under a tea plantation with samples from unused land showed that the tea plantation accelerated soil acidification, with the maximum decline in soil pH in the topsoil. Soil acidification decreased total Ca and exchangeable Ca and K, and increased exchangeable Al and accumulation of SiO2. Acidification induced by the tea plantation led to the destruction of vermiculite followed by the dissolution of hydroxy-Al interlayers within its structure. Soil weathering was enhanced by acidification caused by tea cultivation. The acidification of soil under the tea plantation accelerated not only the transformation of silicate minerals, but also the changes in ferrimagnetic minerals in the soils. The transformation of minerals led to decreased soil CEC. Therefore, acidification induced by the tea plantation reduced soil fertility and increased Al toxicity to plants including tea. This information is essential to understanding the geochemical behavior of soils in tea plantations, and their relationship to the accumulation and transport of nutrients and pollutants in the environment, thus meriting attention for understanding and regulating environmental health.