Njgonline.nl

Geologie en Mijnbouw / Netherlands Journal of Geosciences 79 (1): 59-71 (2000) Composition and genesis of rattlestones from Dutch soils
as shown by Mössbauer spectroscopy, INAA and XRD
J.J. van Loef 1
Interfacultair Reactor Instituut, TU Delft , Mekelweg 15, 2629 JB DELFT, The Netherlands; Manuscript received: 1999; accepted in revised form: 29-12-1999 Abstract
The chemical and mineralogical composition of rattlestones found near the main Dutch rivers has been studied by Möss-bauer spectroscopy, INAA and XRD. Rattlestones are concretions of iron, formed in an environment of lateral iron accumu-lation, under the influence of periodical oxidation, around a fine core of ferruginous sediments, mainly clay and sand. Thecore has shrunk and detached itself from the mantle around it. 57Fe Mössbauer spectroscopy was applied to identify the ironoxides, among which goethite is predominant. The goethite crystallinity was investigated by measuring its magnetic propertiesand its crystallinity, which is poorest at the outer side of the stone. The latter is confirmed by the broadening of the different X-ray reflections. In addition, illite and vermiculite were identified by XRD; these clay minerals were found mainly in the core.
The elemental composition was determined by INAA. The iron content in the mantle is about 50% by weight and gradual- ly decreases outwards, while the core contains 2-15% Fe by weight. Differences between rattlestones from the Middle Pleis-tocene East of the Meuse river and those from the Late Pleistocene North of it are the absence of lepidocrocite and a richermineralogy in the former.
It is concluded that the rattlestones are formed around a fine clayey core. Groundwater supplied the iron and other (trace) elements for the genesis. It is unlikely that rattlestones are the result of oxidation of siderite.
Keywords: crystallinity, goethite, iron accumulation, lepidocrocite, Mössbauer spectroscopy, rattlestones, siderite, trace elements Introduction
soluble layers have been removed by solution, leavinga central part detached from the outer part, such as a Rattlestones have been known for long. They were re- concretion of iron oxide filled with loose sand that ferred to as ‘aetites’ or ‘eagle stones’ in Roman times rattles on being shaken. Van der Burg (1969) suggest- (Adams, 1938; Bromehead, 1947). In spite of this, ed that rattlestones are formed by oxidation of side- surprisingly little information is available on the rite concretions which were deposited contemporane- chemical and mineralogical composition of these ously with the sediments in the beds where they are stones. Few previous publications dealt with rattle- stones (Van der Burg, 1969) and with the climatolo- The geochemical and mineralogical characteristics gical factors limiting their distribution in the Dutch of the rattlestones were investigated for the present Pleistocene (Van der Burg, 1971). The present contri- bution discusses the composition of several rattle-stones found near the main rivers in the Netherlands, Material
According to a recent definition (Jackson, 1997), a Eleven rattlestones have been investigated. They were rattlestone is a concretion composed of concentric collected in several sand and gravel pits near the main laminae of different composition, in which the more Dutch rivers, i.e. at Koningsbosch East of the Meuse Geologie en Mijnbouw / Netherlands Journal of Geosciences 79(1) 2000 river (Sterksel Formation, three specimens), and (Coney, 1988). The structural order of iron oxides en- North of the Meuse at Deest, Lathum and Wapenveld countered in natural environments ranges from rea- (Kreftenheye Formation, four specimens) and at sonably good to seemingly amorphous. This has a Leersum and Schaarsbergen (Drenthe Formation, great influence on the magnetic properties. Devia- four specimens). The codes used for these sites are tions of the magnetic properties of iron oxides of very Ko, De, La, Wa, Le and S, respectively; the Roman small particle size from those of coarse-grained coun- numbers I and II are used wherever necessary to dis- terparts lead to radical changes in their Mössbauer tinguish between stones from one single site. An alter- spectra (Van der Kraan & Van Loef, 1966; Murad, ated siderite concretion found in the Reuver clay at 1996). This is illustrated by the hyperfine field, which about 30 m depth (Kiezeloöliet Formation) was stud- has a single value at any temperature in a pure mag- netic crystal, but can vary greatly or even disappear at The specific density of the stones was obtained by room temperature in a soil iron oxide like goethite.
determining volume and weight. The latter varied be- This type of spectroscopy can be used to distinguish tween 30 and 500 g. The rattlestones were opened between the two valence states of iron, e.g. in goethite carefully in order to collect the loose, yellowish fine grains of the core, all of which weighed only a few Mössbauer spectroscopy is based on the recoil-free percent of the stone. The specific density of the outer resonance absorption of gamma rays in certain atom- part of the stone (called ‘mantle’), which was always ic nuclei, for example the stable iron isotope 57Fe (2% less than 10 mm thick, was obtained and the size of natural abundance). Gamma rays with an energy of 14.4 keV are emitted by radioactive 57Co with a half- Material from the core and several concentric lami- life of 270 days and the resonant absorption cross nae of the mantle of each rattlestone were powdered section of these (γ rays in 57Fe is very high at room separately. The lamination is irregular and the lami- temperature. The method allows determination of nu- nae with a width of 1-2 mm frequently merge into one clear energy levels to an extremely high accuracy, so another (Van der Burg, 1969). Differences in hardness that slight variations caused by different interactions and color are helpful in making a separation between between electrons and the nucleus become measur- laminae. The interior material of the mantle is hardest able. These interactions reflect differences in the elec- and often also the darkest; the exterior material of the tronic, magnetic, geometric or defect structure in rattlestone is mostly quite soft and easy to scrape off solids. The spectroscopy is based upon the principle the mantle. In total about 60 samples from the various that, by moving the radioactive source, a very small rattlestones were investigated. Core samples are coded energy shift can be attained in the emitted gamma 1 and samples from the interior and exterior mantle material have been coded 2 and 3; in case more layers The instrumental setup consists of the 57Co-source, in the mantle are distinguishable, the consecutive lay- detector and the iron-containing sample in between; ers have been numbered 2, 3 and 4. Material from dif- in such a transmission experiment, the Mössbauer ferent shells of the siderite concretion was sampled spectrum (MS) obtained consists of the intensity of and numbered 1 (central segment), 2 (first shell), 3 the (γ rays measured in the detector, plotted as a (second one) and so on. In practice about 200 mg ma- function of the source velocity. A minimum in MS terial per sample sufficed for detailed analysis. Mun- corresponds with resonance absorption in the sample sell hues were used to determine the color of each under investigation. A major value of Mössbauer sample (Cornell & Schwertmann, 1996).
spectroscopy as an analytical tool lies in the fact thatany iron contained in a solid must show up in MS, in- dependent of sample crystallinity or, of the form inwhich Fe is bound (Murad, 1988). Sample prepara- The methods used for the investigation are nuclear tion is usually very simple; the sample thickness is techniques: Mössbauer spectroscopy (Kuzmann et al.,1998) and instrumental neutron-activation analy- The distinction by Mössbauer spectroscopy of sis (INAA), which were complemented by X-ray dif- goethite, hematite, lepidocrocite and siderite is based on the magnetic hyperfine interaction. The first twooxides are magnetically ordered at room temperature and MS of both goethite and hematite consist of six-line spectra. The hyperfine field, B, in goethite at Mössbauer spectroscopy is a very useful technique for room temperature is known to be 38.5 T and that in investigating properties of Fe in soil iron oxides hematite 51.5 T; the latter is used to calibrate the Geologie en Mijnbouw / Netherlands Journal of Geosciences 79(1) 2000 source velocity in Mössbauer spectroscopy. TheMössbauer spectrum at 295 K (coded as MS[295K])of paramagnetic lepidocrocite consists of two lines ofequal intensity, the quadrupole splitting. The doubletis much larger in MS of siderite and readily distin-guished from that of lepidocrocite. Speciation in asample containing a few percent of Fe by weight isfeasible. Poorly crystalline iron oxides can still beidentified, in particular at lower temperatures. Spec-tral intensities of iron-containing species give infor-mation on their relative content in a sample. Furtherdetails on Mössbauer spectroscopy can be found inKuzmann (1988).
Instrumental neutron-activation analysis Fig.1. Cavity in the various rattlestones versus the density. The dashed line is drawn as a guide to the eye. The horizontal line indi- INAA is a quantitative method of high efficiency for cates the range of densities of the stones investigated by Van der the determination of a number of major and trace ele-ments and supplies geochemical information on thematerial that may be involved in the formation and INAA has been applied at the nuclear reactor of MS[295K] and MS[77K] of three samples from Ko IRI at Delft to determine major elements, in particu- II are shown in Figure 2. A central doublet dominates lar Fe, and many trace elements. The standard devia- the spectra at 295 K and has practically disappeared tion and detection limit for Fe are less than 2% and at 77 K. A six-line magnetic hyperfine splitting attrib- about 100 ppm, respectively (Parry, 1991).
uted to goethite dominates MS[77K]. The doublet inMS[295K] is mainly due to goethite and to a limited extent to clay minerals. This follows from their rela-tive intensities measured in MS[77K]. Remnants of a XRD was performed to identify clay minerals. XRD hyperfine splitting are still visible in the spectra of Ko patterns were obtained that enabled to determine II-1 and Ko II–2 at 295 K, but not in Ko II-3. The peak positions and line widths of the iron oxides.
doublet resembles that of a paramagnet, but it is also Diffractograms of about forty samples were made characteristic of superparamagnetic goethite, due to at the Laboratory of Soil Science and Geology of Wa- extremely small crystallites. This is most pronounced geningen University. The XRD was performed with a in Ko II-3. In general, the sextets at 77K consist of Philips PW 1710 diffractometer equipped with a asymmetrically broadened lines indicating a distribu- graphite monochromator using CoKα radiation.
tion of hyperfine fields due to a spread in goethite Peaks were recorded from 5o to 75o 2Θ. The clay min- crystallite size. The smaller the size, the lower the hy- erals were thus identified, and the peak position and perfine field (Van der Kraan, 1972). The distribution linewidth of the goethite reflections were determined.
is narrowest in MS[77K] of Ko II-1 and widest inthat of Ko II-3, while the hyperfine field is highest in the former and lowest in the latter (Table 1). Hence,the goethite crystallite size in the core is larger than in The cavities of the rattlestones vary widely and can take more than 50% by volume; this has a great effect In MS[77K] of Ko II-1 and Ko II–3, the resonant on the specific density of the rattlestone as a whole, as absorption in the centre is attributed to non-magnetic shown in Figure 1. The horizontal bar indicates the minerals. It can be assigned to illite (XRD). Besides, a density range of the numerous stones investigated by small signal is measured at a source velocity of about Van der Burg (1969), which only partly overlap with 3 mm/s that originates from an Fe2+ contribution; the accompanying line of the doublet is submerged in the The other characteristics, as shown by MS, INAA, central part of the spectrum. X-ray diffractograms XRD, and those regarding the hue of the specimens, confirm the assignments of goethite and illite. In the are dealt with in the following subsections.
XRD of Ko II-1, illite is more pronounced than in KoII-3 (see Table 2), which is contrary to the relative in- Geologie en Mijnbouw / Netherlands Journal of Geosciences 79(1) 2000 Fig. 2. Mössbauer spectra of samples Ko II-1 (core), Ko II-2 and Ko II–3 (both mantle) at 77 and 295 K. A central doublet dominates MS[295K] that has practically disappeared in MS[77K], where six-line spectra attributed to goethite are most pronounced. The positions of the lines that correspond to the hyperfine field, B’, are indicated. A central doublet in MS[77K] of Ko II-1 and Ko II–3 can be ascribed to il- lite, including the very small Fe2+ signal . The illite also contributes to the central doublet in MS[295K] but coincides with that of superpara- Table 1. Characteristics of the various samples as determined by XRD and Mössbauer spectra. The line width has not been corrected for in- strumental resolution of 0.1°. The goethite content in the core of Ko III and La was too low to measure the reflections.
these reflections have no interference with quartz.
the second-order reflection of illite may contribute to (020) because of its high concentration in Ko II-1.
the interference of the strong quartz reflections limited an accurate determination of the line width.
Geologie en Mijnbouw / Netherlands Journal of Geosciences 79(1) 2000 Table 2. Munsell hue and other characteristics of the various samples as obtained by INAA (total iron content), measured in Mössbauer spectra (Fe2+ fraction and iron compounds), and minerals identified by XRD. The diffraction lines of illite, vermiculite, goethite, hematite,lepidocrocite, siderite, dolomite and feldspar were measured and their relative intensities indicated.
Bold letters (g or i) mean: dominant in MS.
The intensity is referred to as: very much 4+; much 3+; moderate 2+; present +; questionable “; not proven –. Minerals analyzed: d = dolomite, f = feldspar, g = goethite, h = hematite, i = illite, l = lepidocrocite, v = vermiculite. n.d.= not determined.
tensities of the doublets in MS[77K]. It indicates nanocrystals of goethite with a mean needle width of that the degree of substitution of Fe by Al in illite is 9 nm, measured by Van der Kraan (1972). In summa- higher in Ko II-3 than in Ko II-1 probably because ry, the crystallite size of goethite in the Ko II samples the Fe/Al ratio in the former is higher by a factor of 3 is systematically smaller from the core outwards.
The total hyperfine fraction (h.f.f.) was determined MS between 77 K and 295 K were measured also by the area of the magnetically ordered portion of the during the slow warming up of the cryostat. In each spectra relative to the total spectrum. This fraction in spectrum, the hyperfine field of highest probability, each sample is also plotted versus temperature in Fig- B’, was determined by the positions of the steep outer ure 3. Both B’ and h.f.f. should show extrapolation to edge of the six individual hyperfine lines, as indicated zero at the same temperature, which is reasonably ful- in MS[77K] of Ko II-3 (Fig. 2). These hyperfine filled by the data for Ko II. The extrapolated tempera- fields are plotted versus temperature in Figure 3. The ture of the onset of magnetic ordering in goethite in B’ values for Ko II-1 are systematically higher than the samples is in the 300-320 K range, which is con- those for Ko II-2 and the latter are higher than those siderably below the Néel temperature of 393 K for for Ko II-3. In Figure 3, smooth B’-vs-T curves are pure goethite and also lower than that of the synthetic drawn through the data points of each sample, and nanocrystals, indicating an extremely small crystal they are extrapolated to zero field beyond 295 K. The size of goethite in Ko II. With decreasing tempera- shape of these curves is qualitatively similar to that of ture, h.f.f. increases in each spectrum to a maximum B in pure goethite (Cornwall, UK). A similarly at a temperature where magnetic ordering in goethite shaped curve, shown in Figure 3, is the temperature is complete. The ordered fraction is less than 100% dependence of the hyperfine field in synthetic pure because illite contributes to the non-magnetic part in Geologie en Mijnbouw / Netherlands Journal of Geosciences 79(1) 2000 MS. Figure 3 shows that Ko II-3 not only has thelowest temperature of onset of magnetic ordering ingoethite, but is also the least homogeneous with re-spect to goethite-crystallite size distribution, as thetemperature range in which magnetically ordered andsuperparamagnetic phases of goethite coexist iswidest. This corroborates the strongly broadened hy-perfine lines in MS[77K] of Ko II-3.
MS[295K] and MS[77K] of samples Le II are shown in Figure 4. The spectra have much in com-mon with those of Ko II, particularly with respect tothe magnetic properties of goethite. The hyperfinefield B’[77K] is highest in Le II-1 and lowest in Le II-3 (Table 1). The non-magnetic part in MS[77K] ofLe II-1 is different, however, as it consists of threesubspectra, a single Fe2+ doublet and two Fe3+ dou-blets. One of the latter can be attributed to lepi-docrocite that orders magnetically just below 77 K.
Fig. 3. The magnetic hyperfine field, B’, and the hyperfine fraction, This is verified by a measurement of MS at a temper- h.f.f., in the Mössbauer spectra of Ko II-1, Ko II-2 and Ko II–3 as ature of 4.2 K. Instead of a single sextet of goethite, a a function of temperature. The full curve refers to the hyperfine second one was found with a hyperfine field charac- field in a goethite crystal; the dashed curve refers to B’ in synthetic teristic for lepidocrocite at 4.2 K; XRD confirmed goethite nanocrystals with a mean width of 9 nm. Dot-dash lines this, and illite and vermiculite were identified in addi- through the experimental data points of the Ko samples are drawn tion (Table 2), contributing to the other two doublets Fig. 4. Mössbauer spectra of samples Le II-1(core), Le II-2 and Le II–3 (both mantle) at 77 and 295 K. A central doublet dominates MS[295K] that has practically disappeared in MS[77K], where six-line spectra attributed to goethite are most pronounced. The positions of the lines that correspond to the hyperfine field, B’, are indicated. The central contribution to MS[77K] of Le II-1 can be assigned to lepi- docrocite, illite and vermiculite. The FWHM of the central doublet in MS[295K] is given; note the increase of its line width in the presence Geologie en Mijnbouw / Netherlands Journal of Geosciences 79(1) 2000 Table 3. Major oxides and a number of trace elements in samples from various rattlestones and loam from Koningsbosch, as obtained by in MS[77K]. Temperature dependencies of both B’ and core of one of the central segments, Re-1 and Re- and h.f.f. in samples Le II (Fig. 5) show the same 1a, and from consecutive shells of the siderite concre- trend as found in Ko II. MS data on Le II suggest tions Re-2 through 4. Goethite is the dominant iron that the crystallite size of goethite is largest in the core oxide in all samples. Siderite was identified in Re-2 and smallest in the exterior part of the mantle.
only and hematite was present in Re-1 and Re-1a.
The results of Ko II and Le II are representative for The goethite crystallinity becomes worse in the shells the other rattlestones. Goethite is the predominant outwards, in the same way as in the mantle of rattle- iron oxide in all of them, in both core and mantle. The stones. Hematite in association with goethite has been goethite-crystallite size is in the nanometer range, found in several siderite concretions (Meyer, 1979; which follows from its superparamagnetic behaviour Senkayi et al., 1986; unpublished results from the au- at 295 K and the dramatic lowering of the magnetic thor); hematite has, however, not been identified in ordering temperature, and is also consistent with what is known from a comparison among common soilminerals (Schwertmann, 1988a). In general, the goethite-crystallite size is smallest and tends to bemost heterogeneous in the exterior part of the mantle.
In order to verify the Mössbauer results, X-ray dif- MS were obtained from samples of both surface fractograms were obtained from several samples.
Geologie en Mijnbouw / Netherlands Journal of Geosciences 79(1) 2000 Fig. 5. The magnetic hyperfine field, B’ , and the hyperfine fraction, h.f.f., in the ture. The full curve refers to the hyper- fine field in a goethite crystal. Dot-dash Table 1 lists peak positions and line widths of in several samples from rattlestones found at sites goethite reflections, and Table 2 lists the mineralogy.
north of the Meuse river (Table 2). Lepidocrocite The peak positions are similar to those of pure was mostly encountered in the mantle, except in Le goethite, which makes isomorphic substitution of Fe II-1. The line width of the (020) reflection is reason- by Al unlikely. The (020), (110) and (111) lines and ably narrow and hence lepidocrocite is more crys- the average of the (130), (021), (121) and (140) lines talline than goethite in these samples. Moreover, this of goethite were chosen to investigate line broaden- iron oxide is rarely evenly distributed over the whole ing. Although the first three lines in the core samples soil matrix (Schwertmann, 1988b). This might ex- were subject to interference with strong quartz lines, plain why lepidocrocite was found in some laminae the other four lines are not. A systematic increase in but is absent in adjacent ones. Lepidocrocite is not line broadening is observed in the samples Ko II and found in rattlestones from sites east of the Meuse, al- Le II from the core outwards, as shown in Table 1, though Riezebos et al. (1978) suggested that this iron indicating a decrease in crystallinity (Schwertmann, oxide is a constituent in middle terraces in south 1988a), consistent with MS data. XRD data on line broadening in goethite in the mantles of Ko III and XRD results on clay mineral identification are given La show a similar trend. The goethite content in the in Table 2, where they are also compared with MS re- core of these rattlestones was too low to determine sults. The overall correspondence is good. Besides the former. The weight percentage of goethite in each goethite, illite is found in most samples, not surpris- sample was derived from the absolute Fe content, ingly, as it is the main clay mineral in Dutch soils determined by INAA, in combination with the (Edelman & De Bruin, 1986). Vermiculite, identified goethite fraction measured in MS[77K]. In addition, by XRD in some core samples, can be associated the temperature of onset of magnetic ordering in with a relatively high Fe2+ signal in MS. Table 2 also goethite, T , is presented in Table 1. The lower this lists the XRD results on the siderite concretion. With temperature, the poorer the goethite crystallinity respect to the iron compounds they are consistent (Murad & Bowen, 1987) consistent with the stronger with MS. All Re samples contain dolomite, which is absent in the rattlestone samples. Clay minerals were In both MS and XRD, lepidocrocite was identified encountered only in the shells of the concretion.
Geologie en Mijnbouw / Netherlands Journal of Geosciences 79(1) 2000 trace elements show a similar trend: the amount ofRb and Cs, for instance, is positively correlated with Elemental composition of the samples has been ob- illite, which is consistent with previously reported re- tained by INAA. The Fe content in the core varies sults (Moura & Kroonenberg, 1990).
from 18% by weight in S II to 1.7% in De I; the latter Heavy metals can, as a rule, be associated with Fe has hardly any Fe accumulation at all. In contrast, the hydroxides – supplied by groundwater – that remain Fe content in the core segment of the siderite concre-tion is 52% by weight. The mantle of rattlestones hasa high iron accumulation (up to 50% by weight) thatvaries by less than a factor of 2 among the varioussamples (except for Wa-3). The Fe content relates tothat of Al as shown in Figure 6; in general, a high con-tent in the former corresponds to a low content in thelatter and vice versa. The amount of Fe gradually de-creases in each rattlestone in consecutive layers of themantle outwards, while the content of various otherelements just increases. A histogram (Fig. 7) illus-trates how the total Fe content in core and mantle offour different rattlestones is distributed amonggoethite, lepidocrocite and clay minerals, as deter-mined in MS. Goethite is the predominant iron ox-ide; lepidocrocite has been identified only in the rat-tlestones from north of the Meuse river (De I and LeII). The composition in terms of major oxides (inweight percentage) and a few trace elements (in ppm)is presented in Table 3.
The amount of silicon in the core is about twice that in the mantle, which is understandable as themuch higher Fe accumulation in the mantle lowersthe relative Si content in the latter. A quantitative esti-mate of illite has been made, assuming that K is fullyassociated with this clay mineral by using the chemi-cal formula, KAl Si O H . The illite content in the Fig. 7. A plot of the total Fe versus Al content, obtained by INAA, core in general exceeds that in the mantle by a sub- in core and mantle of the rattlestones (indicated by their codes).
stantial factor, and it tends to increase gradually out- Below the dotted line are the Fe and Al content of the core sam- wards in the latter. This could be a reason why several ples. Clayey loam from the Koningsbosch site is indicated by (o).
Geologie en Mijnbouw / Netherlands Journal of Geosciences 79(1) 2000 incorporated in the iron oxide after sedimentation.
in the Netherlands (Huisman et al., 1999). It appears The element Ba is also positively correlated with illite that the pattern in the mantle is similar to that in and the ratio Ba(ppm)/K O is similar to the value for groundwater. The relative depletion of LREE in the subsurface sediments in the southern Netherlands mantle is attributed to the fact that they form less eas- (Huisman et al., 1997). In the samples Ko II-2, Ko ily carbonate complexes than HREE, due to their II–3, Ko III-2 and Ko III–3, however, Ba is very larger ionic radii. It is known that bicarbonate ion strongly accumulated (Table 3). Such irregular accu- pairs can play an important role for the solution mulations of Ba without correlation with illite or Fe chemistry of HREE, which explains a significant rela- may indicate that it is present as a separate mineral tive fractionation and an enrichment in the latter phase or that it is built into the Fe-oxide structure in (Michard et al., 1987). The pattern in the core, on the varying ratios to iron (Huisman, 1998). Samples with other hand, is more shale-like without significant frac- an excessively high Ba content show a high Mn con- tionation, which is more common for river water tent, too. It follows also from Table 3 that the distribu- (McLennan, 1989). In summary, the mantle has been tion of Th and U in core and mantle is different; the formed from elements that are mainly supplied by mean ratio Th/U amounts to 4 in the core and is groundwater, whereas those in the core originate pri- about 1 in the mantle. These elements tend to frac- marily from the river water. In this connection it is tionate because of oxidation of U to soluble ions, and noted that loam from Koningsbosch also has a shale- because of selective adsorption of Th in clays and its retention in heavy resistant minerals (Wedepohl, Variations in the content of heavy metals within the mantle and between various rattlestones are probably The mean content of the rare earth elements in the caused by differences in groundwater composition mantle is systematically higher than in the core, ex- during the formation of these stones. Such diverse- cept for La, Ce and Nd, which elements show an op- ness could be the result of differences in aquifer mate- posite behaviour. The shale-normalised REE patterns rial, and by changes in pH and Eh of the ground- for both core and mantle are shown in Figure 8 and compared with that reported for shallow groundwater The Munsell hue of the samples, using the Munsellsoil color charts (1954) is also indicated in Table 2.
The color of the iron oxide may vary from sample tosample and often has sufficient consistency to be use-ful for the identification of these oxides in soils andsediments (Schwertmann, 1988a). The mineral-spe-cific Munsell hues of iron oxides in soils are in the fol-lowing ranges: hematite 5R-2.5YR, lepidocrocite 5YR-7.5YR, and goethite 7.5YR-2.5Y, sometimes ex-tending to 5YR.
Finely dispersed soil goethite and lepidocrocite usually become increasingly dark with poorer crys-tallinity; goethite crystals smaller than 50 nm, for in-stance, are brownish (Cornell & Schwertmann,1996). The first number behind the color hue, calledthe value, is a notation for darkness varying between 0(absolute black) and 8 (absolute white). Core sampleshave a higher value than those of the mantle, which isqualitatively consistent with the better crystallinity ofgoethite in the former.
Discussion
Fig. 8. REE patterns in core and mantle of rattlestones obtained by Goethite forms in all soil types and in early stages of INAA are compared with the patterns found in shallow groundwa- weathering it is the most commonly found iron oxide.
ter and in a sandpit. The concentrations are normalised using Dissolved ferrous iron, which is usually present in re- Geologie en Mijnbouw / Netherlands Journal of Geosciences 79(1) 2000 duced sediments, oxidizes to Fe(III) when it encoun- pects: (1) the iron content in the central segments of ters oxygen and converts to virtually insoluble Fe(III) the concretion is very high and clay minerals are ab- oxyhydrates (Cornell & Schwertmann, 1996). In al- sent in contrast to what is found in the core of rattle- ternating reducing and oxidizing conditions, a clay stones, (2) hematite is found in the central segments layer in a sandy matrix will accumulate these hydrates and absent in the rattlestone samples, (3) the latter through the following process. When the sediment may contain a strong and sometimes irregular accu- layer is reduced, Fe(II) is distributed evenly in the mulation of heavy metals such as Ba, which is unlikely aqueous phase. When this sediment dries out, iron in siderite as its structure cannot accommodate any will partially precipitate, being exposed to oxygen, accumulation of heavy metals (Huisman, 1998), and while the concentration in the remaining solution in- (4) the specific density of rattlestones can, due to a creases. Eventually the finer layers will still contain large cavity, be much smaller than the densities re- water and its dissolved iron. Upon complete dryout, ported by Van der Burg (1969), as shown in Figure 1.
this iron precipitates at the contact of the fine layer Septarian cracks have indeed been found in the alter- (which is still water-saturated) and its aerated coarse ated siderite concretion, which result in a relatively surroundings. This is a repetitive process. In addition small cavity only. These differences make it unlikely to accumulate iron, the clay layer may fracture upon that rattlestones are the result of oxidation of siderite.
drying out, whereby each of the fragments will be- It examplifies what Fitzpatrick (1988) has described come a nucleus of Fe accumulation. With time, the as iron compounds being indicators of pedogenic accumulated iron may form a rigid oxide coating around the clay core, and become a concretion. When Rattlestones are composed of clay minerals, mainly the concretion dries out further, the clay core will illite and vermiculite, and poorly crystalline goethite.
shrink and detach from the mantle, thus forming a The main non-clay mineral is quartz. This mixture cavity in the stone. Because the inner part of the rat- closely corresponds with what was called an assem- tlestone is not subject to periodical reduction and blage of young, moderately weathered soils in temper- precipitation any more, it can recrystallize to larger ate regions by Van Breemen & Buurman (1998). The size, while this is not the case for the active outer part official definition of a rattlestone given in the intro- (Van Breemen & Buurman, 1998). These processes duction (Jackson, 1997) will be more in accordance are governed throughout the year by groundwater so with the actual results when it is recognized that the that absolute accumulation of iron, conveyed from core material does not consist of loose sand only.
surrounding areas, may occur. As a result of fluctuat- The elemental analysis of several samples of a rat- ing groundwater levels, yellowish brown goethite con- tlestone gives additional geochemical information on centrates form laminae in the mantle of the stone the subsurface soil and sediments of the sampling with an Fe(III)-oxide content that is high relative to area. For example, the higher accumulation of various the surroundings. These oxides may cement the min- heavy metals in rattlestones from the older flood- eral grains giving rise to the hardness of the concre- plains east of the Meuse river compared to that in stones from north of the river indicates a richer min- The size of the cavity will depend on the amount eralogy in the former. On the other hand, the absence and mineralogy of the clay, and the original water of lepidocrocite in rattlestones from Koningsbosch content. Since illite is the dominant clay mineral, its might imply that this iron oxide has been transformed content in the core is positively correlated with the into goethite in the course of time though this process size of the cavity in the rattlestones as can be seen by is very slow under pedogenic conditions (Schwert- mann & Taylor, 1972). As lepidocrocite occurs pre- The former process of genesis of rattlestones is en- dominantly in younger postglacial soils following tirely different from the suggestion by Van der Burg Schwertmann & Taylor (1971), the reason for its ab- (1969) that they have been formed as a result of the sence is because these rattlestones are from the much oxidation of siderite concretions that had been de- earlier Sterksel Formation (Middle Pleistocene). Lepi- posited contemporaneously with the sediments in the docrocite was not found in the siderite concretion.
beds where they are found today. As siderite is particu-larly vulnerable to weathering under atmospheric Conclusions
conditions, it will oxidize preferentially to goethite(Pettijohn, 1975). Although MS and XRD results The chemical and mineralogical composition of rat- with respect to goethite crystallinity in the siderite tlestones found near the main Dutch rivers has been concretion are qualitatively similar to those obtained extensively studied by Mössbauer spectroscopy, with rattlestones, both systems differ in several as- INAA and XRD. Rattlestones are iron concretions Geologie en Mijnbouw / Netherlands Journal of Geosciences 79(1) 2000 formed in an environment of lateral accumulation of of the Mössbauer group in the Department of Radia- Fe under the influence of periodical oxidation, tion Physics (IRI) and to Ing. M.P. van Steenvoorden around a fine core of ferruginous sediments, mainly for giving me the opportunity to carry out the Möss- clay and sand. The core has shrunk and detached it- self from the mantle around it. The cavity formed canhave a considerable size, of up to 50% by volume.
References
57Fe Mössbauer spectroscopy was applied to identify Adams, F.D., 1938. The birth and development of the geological the iron oxides, among which goethite is predomi- sciences. Ballière, Tindall and Cox (London): 506 pp.
nant. The goethite crystallinity was investigated by Bromehead, C.N., 1947. Aetites or the eagle-stone. Antiquity 21: making use of its magnetic properties and showed to be poorest in material from the outer edge of the rat- Coey, J.M., 1988. Magnetic properties of iron in soil iron oxides and clay minerals. In: Stucki, J.W., Goodman, B.A. & Schwert- tlestone. The latter is confirmed by broadening of the mann, U. (eds.): Iron in soils and clay minerals. D. Riedel (Dor- various X-ray reflections. In addition, illite and vermi- culite were identified by XRD; they were found main- Cornell, R.M. & Schwertmann, U., 1996. The iron oxides: struc- ture, properties, reactions, occurrence and uses. Verlag Chemie The elemental composition was determined by Edelman, Th. & De Bruin, M., 1986. Background values of 32 ele- INAA. The iron content in the mantle is about 50% ments in Dutch topsoil, determined with non-destructive neu- by weight and gradually decreases outwards, while tron activation analysis. In: Assink, J.W. & Van den Brink, W.J.
the core contains 2-15% Fe by weight. Differences (eds.): Contaminated soil. Martinus Nijhoff Publishers (Dor- between rattlestones from a Middle Pleistocene site east of the Meuse river and those from Late Pleis- Fitzpatrick, R.W., 1988. Iron compounds as indicators of pedo- genic processes: examples from the southern hemisphere. In: tocene sites north of the river are the absence of lepi- Stucki, J.W., Goodman, B.A. & Schwertmann, U. (eds.): Iron in docrocite and a richer mineralogy in the former. As a soils and clay minerals. D. Riedel (Dordrecht): 351-395.
result of fluctuating groundwater levels, iron and oth- Huisman, D.J., 1998. Geochemical characterization of subsurface er (trace) elements were supplied and could lead to sediments in the Netherlands. Thesis Agricultural University Wa- concentrates, of iron hydroxides in particular, in the Huisman, D.J., Vermeulen, F.J.M., Baker, J., Veldkamp, A., Kroo- most porous parts of the clayey sediment. The de- nenberg, S.B. & Klaver, G.Th, 1997. A geological interpretation tailed mechanism describing the formation of rattle- of heavy metal concentrations in soils and sediments in the stones is consistent with experimental results. Its gen- southern Netherlands. Journal of Geochemical Exploration 59: esis is from inside outwards and is difficult to recon- cile with the concept of oxidation of a siderite concre- Huisman, D.J., Van Os, B.J.H., Klaver, G.Th. & Van Loef, J.J., 1999. Redistribution of REE in aquifers in the Netherlands. In: Ármannsson, H. (ed.): Geochemistry of earth’s surface. A.A.
Acknowledgements
Jackson, J.A., 1997. Glossary of geology (4th ed.). American Geo- logical Institute (Alexandria): 769 pp.
I like to express my high appreciation for the fruitful Kuzmann, E., Nagy, S., Vértes, A., Weiszburg, T.G. & Garg, V.K., 1998. Geological and mineralogical applications of Mössbauer discussion with and valuable support by Dr. P. Buur- spectroscopy. In: Vértes, A., Kuzmann, E., Nagy, S., Weiszburg, man and J.D.J. Van Doesburg for the XRD measure- T.G. & Garg, V.K. (eds.): Nuclear methods in mineralogy and ments in the Laboratory of Soil Science and Geology geology. Plenum Press (New York): 285-376.
at the Wageningen University. Discussions with Dr.
McLennan, S.M., 1989. Rare earth elements in sedimentary rocks: H. Huisman (NITG-TNO) on REE have been very influence of provenance and sedimentary processes. In: Lipin, B.R. & McKay, G.A. (eds.): Geochemistry and mineralogy of useful. I have profited from the generosity of readers rare earth elements. Reviews in Mineralogy 21: 169-200.
of the Dutch journal of the Nederlandse Geologische Meyer, B., 1979. Die Entcarbonatierungsrotung als Bodengeneti- Vereniging, Grondboor en Hamer, for providing me scher Teilprozess. Mitteilungen der deutschen Bodenkundlichen with a number of rattlestones. I am grateful to the late Ir. C. Maugenest of the Mineralogical and Geological Michard, A., Beaucaire, C. & Michard, G., 1987. Uranium and rare earth elements in CO -rich waters from Vals-les Bains Museum, Department of Applied Geoscience of (France). Geochimica et Cosmochimica Acta 51: 901-909.
Delft University of Technology, and to Dr. T. G. Nij- Moura, M.L. & Kroonenberg, S.B., 1990. Geochemistry of Qua- land, who helped me in an early stage of the research.
ternary fluvial and eolian sediments in the southeastern Nether- I appreciate the vivid interest of Drs. B.W. Zuurdeeg lands. Geologie en Mijnbouw 69: 359-373.
from Geochem Research (Utrecht) in this study. I Munsell Soil Color Charts, 1954. Munsell Company Inc. (Balti- thank Mrs. Drs. Th.G. van Meerten (Radiochemistry Murad, E. & Bowen, L.H., 1987. Magnetic ordering in Al-rich Department of IRI at Delft) for carrying out INAA. I goethites: influence of crystallinity. American Mineralogist 72: am very thankful to Dr. Ir. A.M.Van der Kraan, head Geologie en Mijnbouw / Netherlands Journal of Geosciences 79(1) 2000 Murad, E., 1988. Magnetic properties of iron in soil iron oxides tion of soil lepidocrocite to goethite. In: Pseudogley and gley.
and clay minerals. In: Stucki, J.W., Goodman, B.A. & Schwert- Transactions of Committees V and VI of the International Soci- mann, U. (eds.): Iron in solids and clay minerals. D. Riedel (Dor- ety of Soil Science (Stuttgart-Hohenheim): 45-54.
Schwertmann, U. & Taylor, R.M., 1972. The transformation of lep- Murad, E., 1996. Magnetic properties of microcrystalline iron (III) idocrocite to goethite. Clays and Clay Minerals 20: 151-158.
and related materials as reflected in their Mössbauer spectra.
Senkayi, A.L., Dixon, J.B. & Hossner, L.R., 1986. Todorokite, Physics and Chemistry of Minerals 23: 248-262.
goethite and hematite: alteration products of siderite in East Parry, S.J., 1991. Activation spectrometry in chemical analysis.
Texas lignite overburden. Soil Science 142: 36-42.
J.Wiley & Sons (NewYork) 243 pp.
Van Breemen, N. & Buurman, P., 1998. Soil formation. Kluwer Pettijohn, F.J., 1975. Sedimentary rocks. Harper & Row (New Academic Publishers (Dordrecht): 377 pp.
Van der Burg, W.J., 1969. The formation of rattle stones. Palaeo- Riezebos, P.A., Bisdom, E.B.A. & Boersma, O., 1978. Composite geography, Palaeoclimatology, Palaeoecology 6: 105-124.
grains in Maas sediments: a survey and a discussion of their Van der Burg, W.J., 1971. The climatological factors. Palaeogeogra- opaque components. Geologie en Mijnbouw 57: 417-431.
phy, Palaeoclimatology, Palaeoecology 7: 297-308.
Schwertmann, U., 1988a. Some properties of soil and synthetic Van der Kraan, A.M., 1972. Mössbauer effect studies of superpara- iron oxides. In: Stucki, J.W., Goodman, B.A. & Schwertmann, U.
magnetic α-FeOOH and α-Fe O . Thesis Delft University of (eds.): Iron in solids and clay minerals. D. Riedel (Dordrecht): Van der Kraan, A.M. & Van Loef, J.J., 1966 Superparamagnetism Schwertmann, U., 1988b. Occurrence and formation of iron oxides and submicroscopic α-FeOOH particles observed by the Möss- in various pedoenvironments. In: Stucki, J.W., Goodman, B.A. & bauer effect. Physics Letters 20: 614-616.
Schwertmann, U. (eds.): Iron in solids and clay minerals. D.
Wedepohl, K.H. (ed.), 1978. Handbook of geochemistry. Springer Schwertmann, U. & Taylor, R.M., 1971. The in vitro transforma- Geologie en Mijnbouw / Netherlands Journal of Geosciences 79(1) 2000

Source: http://www.njgonline.nl/publish/articles/000021/article.pdf