Boletín de la Sociedad Chilena de Química
versión impresa ISSN 0366-1644
Bol. Soc. Chil. Quím. v.45 n.2 Concepción jun. 2000
SOME CAUTIONS ON THE INTERPRETATION OF MÖSSBAUER
SPECTRA IN MINERALOGICAL STUDIES
OF VOLCANIC SOILS
(1)Universidad de Santiago de Chile, Av. B. O'Higgins 3363, C40-33, Santiago, Chile.
(2)Centro Nacional de Investigaciones Científicas, La Habana, Cuba.
(3)Universidade Federal de Minas Gerais, 31270-901 Belo Horizonte, MG, Brazil.
(Received: January 27, 2000 - Accepted: April 12, 2000)
In memoriam of Doctor Guido S. Canessa C.
Room temperature Mössbauer spectra of bulk Chilean soil samples derived from volcanic materials (Ultisols and Andisols) are very complex and least-squares fitting not seldom leads to many different interpretations. In most cases the presence of commonly occurring iron oxides in those soils, namely maghemite (ideal formula, g Fe2O3), hematite ( gFe2O3) and goethite (aFeOOH) in Andisols, and maghemite, magnetite (Fe3O4), ferrihydrate (Fe5HO8.4H2O) and/or goethite in the Ultisol can only be speculated. In the present paper we present X-ray diffraction and Mössbauer analyses of particle size fractions (soil, sand, silt and clay), which were submitted to magnetic extraction (sand fractions) and subsequently to NaOH selective chemical treatment (silt and clay fractions), in laboratory procedures which were conveniently adapted to such volcanic soil-samples. When chemically treated soil separates of Ultisol are analyzed, the presence of partially oxidized magnetite is confirmed, and the existence of hematite is also inferred. The magnetic fraction separation and chemical treatments are to be essential steps in sample preparation in the mineralogical assessment of Chilean soils derived from volcanic materials.
KEY WORDS: Mössbauer spectroscopy, iron oxides, volcanic soils, soil mineralogy
Los espectros Mössbauer de suelos Chilenos derivados de materiales volcánicos (Ultisoles y Andisoles), tomados a temperatura ambiente, son muy complejos y los ajustes realizados a través del método de cuadrados mínimos, llevan a varias interpretaciones diferentes. En la mayoría de los casos, sólo es posible postular la presencia de óxidos de hierro comunmente presentes como maghemita (fórmula ideal , g Fe2O3), hematita (aFe2O3) y goethita (aFeOOH) en Andisoles, y maghemita, magnetita (Fe3O4), ferrihidrita (Fe5HO8.4H2O) y/o goethita en el Ultisol. En este artículo, se presentan estudios realizados empleando difracción de rayos X y espectroscopía Mössbauer de fracciones de diferente tamaño de partícula provenientes de suelos Chilenos (suelo completo, arena, limo y arcilla), las cuales han sido sometidas a extracciones magnéticas (arena) y tratamiento químico disolutivo selectivo con NaOH (limo y arcilla), adaptados a suelos volcánicos. Cuando se analizan los resultados obtenidos con las fracciones tratadas, sólo se confirma la presencia de magnetita en el Ultisol y se infiere la existencia de hematita. La separación de la fracción magnética y los tratamientos químicos resultan ser etapas esenciales en la preparación de muestras para estudios mineralógicos de suelos Chilenos derivados de materiales volcánicos.
PALABRAS CLAVES: Espectroscopía Mössbauer, óxidos de hierro, suelos volcánicos, mineralogía de suelos.
Most agricultural activities in Chile are carried out in soils derived from volcanic materials, that encompasses between 50% to 60% of the Chilean cultivated land. Andisols and Ultisols, commonly found in southern Chile, are the most important, in terms of their potential in agricultural production. Andisols are characterized by a mineralogy that is dominated by a poorly organized aluminosilicate (allophane, nAl2O3.n'SiO 2.n''H2O), along with significant amounts of iron oxides and high organic matter contents. Ultisols are older than Andisols. It has a mineralogy that is characterized by more crystalline compounds with higher iron oxide, and lower organic matter contents than Andisols1.
Mineralogical analysis of soils is an elaborate process and standard physical and chemical methods currently used give only partial or ambiguous information about the intrinsic nature of soil minerals. The mineralogy of volcanic soils is even more difficult to deduce because of the compositional and structural variability of their minerals, particularly of the iron oxide group. For example, powder X-ray diffraction patterns of Chilean soils show only a few broad peaks2,3 that does not allow an unequivocal interpretation. Despite of their widespread occurrence, the main characteristics of the iron oxides in Chilean volcanic soils are primarily inferred from indirect data2,4,5. Moreover, iron oxides play an important role in defining the surface charge of variable charge soils6. For all these reasons, their mineralogy merits more thorough investigation.
57Fe Mössbauer spectroscopy is a valuable tool on the characterization of soil-iron oxides, but its application to soils requires that much special attention must be paid on sample preparation (particularly in physical and chemical dissolution treatments of the bulk samples to concentrate mineralogical phases of interest), measurements both at room and low temperatures, and use of externally applied magnetic field.
In general, Mössbauer analysis of soil samples may not give a decisive qualitative information about iron-bearing compounds mainly because of:
1.- A mixture of paramagnetic and magnetic subspectra is mostly obtained; the presence of Fe3+ in octahedral sites of different silicates results in doublets with similar hyperfine parameters.
2.- The poor crystallization of some iron oxides and oxyhydroxides in volcanic soils, or the superparamagnetic behavior of crystalline minerals as a consequence of their occurrence in finely divided particles (this latter situation may be circumvented through qualitative analysis of Mössbauer spectra obtained with the sample at very low temperature).
3.- The isomorphic replacement of iron by another element such as aluminum, titanium, magnesium, etc., which may modify the hyperfine parameters of the bulk soil sample Mössbauer spectrum.
Recently, Saragovi et al.7 pointed to the experimental difficulties and rose some warnings on the semiquantitative Mössbauer analysis of a low-iron-content Mollisol from Argentine.
In this paper, Mössbauer spectra at room temperature complemented with powder X-ray diffraction analysis of relatively iron-rich soil-samples, and of their particle size fractions (sand, silt, and clay), which were submitted to physical and chemical concentration treatments are compared, to demonstrate (i) the ambiguous interpretation of iron oxides mineralogy, if only Mössbauer spectra at room temperature of volcanic soil are used, and (ii) the improvements in such interpretation when chemical dissolution treatments and magnetic separation are used to prepare the samples.
The B horizon (15-30 cm) of three soils located in southern Chile were selected, one of them is classified as Ultisol (Collipulli), one as Andisol (Osorno) and one as Andisol with iron oxide pan and constrained seasonal drainage (Frutillar). Soil classification and characterization are summarized in Table 1.
|Table I: Identification and characterization of soil samples.|
|Soil sample||Classification||Soil order||Organic Carbon||pH||Fe2O3|
Soil-samples were screened in the field to pass a sieve with 2 mm openings, and stored in a cold room at field moisture content until use.
The particle size fractionation was carried out by dispersing an aliquot of the soil-sample in 0.6 M NH4OH. The suspension was then sieved, to separate the sand fraction (f = 2 mm - 0.05 mm). The silt (f = 0.05 mm - 0.002 mm) and the clay (f < 0.002 mm) fractions were separated by centrifuging the remainder at 600 rpm for 5 min8. The sand fractions were further separated into magnetic and non-magnetic sub-fractions by suspending the sample in water, and picking up magnetic particles with a hand-held magnet.
Silicates and gibbsite in silt and clay fractions were removed by three sequential extractions that consisted of adding 75 mL of 5 M NaOH to 5 g of the soil material and heated to approximately 90 C for 1 hour9. The solid residue so obtained was then washed with diluted hydrochloric acid, to neutralize the sodium hydroxide excess and to remove byproducts of the reaction, such as sodalite.
Mössbauer analysis was performed with a conventional constant acceleration transmission spectrometer, set up with nuclear instrument modules, mainly from ORTEC, and a 50 mCi 57Co/Rh source, at room temperature (RT). Data were collected with an ORTEC MCS card, coupled into an IBM PC. The powder X-ray patterns were obtained with a Rigaku Geigerflex diffractometer equipped with a graphite diffracted beam monochromator, using Cu(Ka) radiation.
Chemical analysis for cations was carried out by atomic absorption spectroscopy after sample dissolution in a teflon bomb. Organic matter content was determined by the Walkley-Black method10. Samples pH were measured in double distilled water soil suspension at 1:2.5 weight to volume ratio.
RESULTS AND DISCUSSION
All soils investigated were acidic with pH ranging from 4.1 to 5.8. The chemical properties of the Andisol with iron oxide pan and seasonal water logging conditions (Frutillar) were not different from the other Andisol except for the significantly high organic matter content of Frutillar.
The X-ray diffraction patterns of the soils provided limited information on the aluminosilicates in Ultisols2,3. No substantive peak was observed in Andisols. Mössbauer spectra of the soils showed, in all cases, a well defined symmetric or asymmetric doublet and a less prominent multiplet (Figure 1). The central doublet in soil samples are either due to the presence of Fe3+ for Collipulli, Osorno and Frutillar soils and Fe2+ for the Frutillar sample in paramagnetic/superparamagnetic mineral phases. In all cases the sextuplet are fitted with different sets of hyperfine parameters. As an example, in Figure 2 and in Table II the different fitting models and Mössbauer parameters for the Collipulli soil are exhibited. The fittings were obtained considering three sextuplet and one doublet (Fittings 1, 2, and 3), and two sextuplet and one doublet (Fitting 4). In the Table II the variation of values of magnetic field from one fitting to other is observed.
Fig 1. Room temperature Mössbauer spectra of B horizon for the Collipulli, Osorno, and Frutillar soils.
Fig.2 Fitting models of the room temperature Mössbauer spectrum of the Collipulli soil-sample.
|Fig.3 Room temperature Mössbauer spectra of the Collipulli whole soil and corresponding particle-size fractions (sand, silt, and clay).|
The highest value of magnetic field for spectrum fitting 1 (sextuplet 1, Table II) may be assigned to hematite. For fittings 2 and 3 the magnetic field value for the same sextuplet corresponds to maghemite. The sextuplets 2 and 3 with lower values of magnetic field correspond to magnetite, considering fitting 1 and 2, however, for fitting 3, the fitted Bhf = 38.17 T is assignable to goethite. For fitting 4, the two sextuplets have similar values of magnetic field, which do not allow a clear identification of any iron compounds, and may be explained by a distribution of magnetic field, as it can be drawn from the by a broad-line spectrum.
In the least-squares fitting algorithm, the normalized
|, where v = n-r is the|
degree of freedom, n = number of experimental points, r = number of fitting parameters, s2 = variance (in a Poison distribution, the expectation value of s2 = zi, the number of detected counts) and f(x) is the calculated number of counts of point i, is iteratively minimized. The expected value of the normalized c2 is the unity. This is the used criteria for the goodness fitting (c2 closer to 1 corresponds to the best fitting). In present case, the best c2-value corresponds to fitting 4 (Table II), but the fitted hyperfine parameters do not allow a clear separation of the iron sites, due to the poor resolution of the experimental spectrum (Figure 2). Conversely, the best resolved spectra have the lowest c2-values (Fittings 1 and 2).
|Table II.- Fitting models and Mössbauer hyperfine parameters for Collipulli soil (d=isomer shift relative to aFe; D=quadrupole splitting; 2eQ=quadrupole shift; Bhf=hyperfine field; A=relative subspectral area; c2=goodness of fitting criterion)|
|Fitting||d/mm s-1||2eQ /mms-1||Bhf/tesla||A(%)||c2|
Thus, considering the room temperature Mössbauer spectra of bulk soils only, the presence of maghemite, hematite and goethite in the Andisols and maghemite, magnetite and ferrihydrate and/or goethite in the Ultisol may be only tentatively assigned, and obviously further studies are needed to arrive at more definitive conclusions.
The spectra of soil separates did not shown significant differences with the respective Mössbauer spectra of the soils, which is exemplified with the Collipulli soil and fractions (Figure 3). However, the spectra of the magnetic separates of sand fractions provided some additional information. Mössbauer spectra of magnetic sand fractions for the two Andisols (Osorno and Frutillar) exhibited a complex mixture of iron oxide spinel (confirmed by the X-ray diffractograms) and a paramagnetic subspectrum (the central doublet). For the Collipulli soil the X-ray diffractogram of magnetic separate sand fraction (Figure 4) shows a high intensity 311 reflection corresponding to spinel structure, and the fitting of the Mössbauer spectrum with a model-independent hyperfine field distribution leads to two distribution blocks meaning a probability curve for Bhf, P(Bhf) (figure not shown). One of them, with averaged isomer shift relative to aFe and a quadrupole shift which is nearly the characteristic value for the mixed valence Fe3+/2+ in the octahedral site of magnetite, and a second distribution due to Fe3+ in tetrahedral sites. The rest of the spectrum was explained by one small paramagnetic subspectrum as a central doublet.
|Fig.4 Powder X-ray diffraction patterns of the magnetic separate of Collipulli sand fraction.|
Silt and clay fractions of soils were treated with NaOH to dissolve amorphous Fe oxides, Fe-organic matter complexes, silicates and gibbsite, and to concentrate magnetic and crystalline Fe compounds. In all soils the chemical dissolution treatment significantly improve the signal quality of the obtained Mössbauer spectra. Not all spectra could be accurately interpreted. Collipulli soil was used as an example because it produced the best spectrum (Figure 5). The Mössbauer spectrum of this soil revealed the characteristic features of partially oxidized magnetite. The best fitting leads to two P(Bhf) profiles: a mixed valence Fe3+/2+ in octahedral sites, and a contribution from Fe3+ in tetrahedral sites.
|Fig.5 Room temperature Mössbauer spectrum of the Collipulli soil-silt fraction after three NaOH treatments.|
These two iron sites exhibited Mössbauer parameters comparable with those found for the magnetic separates of corresponding sand fraction (Table III), and assigned to magnetite. A minor partial spectra could be assigned to a magnetic ordered phase, hematite.
|Table III.- Mössbauer parameters of magnetic separate of sand fraction and NaOH treated silt of Collipulli soil (d=isomer shift relative to aFe; 2eQ=quadrupole shift; Bhf=hyperfine field).|
|Room temperature||Magnetic separate of the sand||NaOH treated silt|
|Sextuplet 1||Sextuplet 2||Sextuplet 1||Sextuplet 2|
The general Mössbauer spectra revealed many possible Fe minerals in the soils from different fittings of the same spectra. The magnetic fraction separation and chemical treatments would be essential steps in sample preparation in the mineralogical assessment of volcanic materials derived soils. After separation and NaOH chemical treatment the presence of maghemite, hematite and goethite in Andisols, and maghemite, magnetite and goethite in Ultisol may be tentatively assigned. The NaOH-treatment was effective in concentrating the iron oxide minerals but new modifications in chemical extraction could further increase the iron oxide concentration in the soil fractions and to arrive to more definitive conclusions. Ammonium oxalate, dithionite-citrate-bicarbonate, and pyrophosphate extraction procedures might be helpful to better define the crystalline and amorphous Fe oxides (as ferrihydrite) and Fe oxides associated to inorganic compounds.
Work supported by FAPEMIG, CNPq and FINEP (Brazil) and CONICYT, FONDECYT (1980031 and 7980043) and RELAQ/UNESCO (Chile).
1.- Besoaín, E. Mineralogía de los Suelos Volcánicos del Centro-Sur de Chile. In: J. Tosso (Ed.) Suelos Volcánicos de Chile, INIA, Santiago, Chile, pp. 107-302 (1985). [ Links ]
2.- Escudey, M., and Galindo, G. J. Colloid Interface Sci. 93, 78 (1983). [ Links ]
3.- Escudey, M., and Galindo, G. Bol. Soc. Chil. Quím. 39, 63 (1994). [ Links ]
4.- Escudey, M., Galindo, G., and Ervin, J. Clays and Clay Miner. 34, 108 (1986). [ Links ]
5.- Escudey, M., and Galindo, G. Geoderma 41, 275 (1988). [ Links ]
6.- Cornell, R. M., and Schwertmann, U. Crystal Structure. In: The Iron Oxides, VHC Publishers: New York, pp. 7-34 (1996). [ Links ]
7.- Saragovi, C., and Mijovilovich A. Clays Clay Miner. 45, 480 (1997). [ Links ]
8.- Jackson, M. L. Soil Chemical Analysis: Advanced Course, 3rd ed. (Madison, Wisconsin, published by the author) pp. 894 (1969). [ Links ]
9.- Norrish, K. and Taylor, R.M. J. Soil Sci. 12, 294 (1961). [ Links ]
10.- Allison, L. E. Organic Carbon. In : C. A. Black, D. D. Evans, J. L. White, L. E. Ensminger, and F. E. Clark (Eds.), Methods of Soil Analysis. Part 2. Agronomy 9. American Society of Agronomy, Madison, WI. pp 1367-96 (1965). [ Links ]