Ionescu C., Hoeck, V., Gruian C., Simon V. (2014): Insights into the EPR characteristics of heated carbonate-rich illitic clay. Applied Clay Science, 97-98: 138-145.

Applied Clay Science 97–98 (2014) 138–145 Contents lists available at ScienceDirect Applied Clay Science journal homepage: www.elsevier.com/locate/clay Research paper Insights into the EPR characteristics of heated carbonate-rich illitic clay Corina Ionescu a,⁎, Volker Hoeck a,b, Cristina Gruian c, Viorica Simon c a b c Department of Geology, Babeş-Bolyai University, 1 Kogălniceanu Str., 400084 Cluj-Napoca, Romania Division Geography and Geology, Paris Lodron University, 34 Hellbrunner Str., A-5020 Salzburg, Austria Faculty of Physics and Institute of Interdisciplinary Research on Bio-Nano-Sciences, Babeş-Bolyai University, Cluj-Napoca, 400084, Romania a r t i c l e i n f o Article history: Received 18 October 2013 Received in revised form 23 May 2014 Accepted 26 May 2014 Available online 13 June 2014 Keywords: Carbonate-rich illitic clay Thermal treatment EPR Clay-based ceramics Archaeometry a b s t r a c t The response of carbonate-rich illitic clay heated up to 1200 °C was investigated by means of electron paramagnetic resonance (EPR) in order to define the factors influencing the shape of the resonance signals and to establish whether this method can be used for evaluation of firing temperature for clay-based ceramic objects. The results show that the dominating hyperfine sextet, at g ≅ 2, due to Mn2+, is replaced over 700 °C by a large signal, mainly due to Fe3+. Oxidation of Mn2+ (EPR active) to Mn3+ (EPR silent) or Mn4+, and Fe2+ (EPR silent) to Fe3+ (EPR active) respectively, combined with changes in their environment produce the resonance signals. The destruction of the carriers such as Fe-oxihydroxides, clinochlore, calcite, dolomite, altered biotite, illite and muscovite, as well as the formation of new minerals and glass are the main mineralogical processes influencing the width of the resonance signals. The results of this study can be used in conjunction with mineralogical and microstructural data for the investigation of technological conditions such as firing temperature and atmosphere related to archaeological ceramic objects. Data gathered from other methods may also help to constrain the EPR signal shape. © 2014 Elsevier B.V. All rights reserved. 1. Introduction Clays and ceramic objects play a very important role in the development of human society. Since the first ceramic statuettes from Central Europe dated to the Late Palaeolithic (~ 26,000 B.P.; Vandiver et al., 1989) and the invention of pottery in South China around 20,000 B.P. (Wu et al., 2012), clays have been regarded as raw materials. They are widely distributed and in many cases available close at hand. More recently, clays have also been involved in the study, restoration and preservation of cultural heritage, especially archaeological ceramic objects. Mudstone – here referred as ‘clay’ – involved in obtaining so-called ‘traditional’ or ‘clay-based’ ceramics, consists in most cases of clay minerals (illite, kaolinite and smectites) associated with quartz, feldspars, micas, carbonates, sulfate, Fe- and Mn-oxihydroxides, heavy minerals, as well as rock fragments and organic material. Their compositional, structural and textural changes recorded at various temperatures are used for estimating the technological conditions of firing (e.g., Cultrone et al., 2001; Maggetti, 1982; Riccardi et al., 1999). The most used investigation methods are polarized light optical microscopy (OM), electron microprobe analysis (EMPA), scanning electron microscopy (SEM), X-ray powder diffraction (XRPD) and thermoanalysis. However, inferring the conditions, in particular the temperature at which clay-based objects were produced, even with a large margin of ⁎ Corresponding author. E-mail addresses: corina.ionescu@ubbcluj.ro (C. Ionescu), volker.hoeck@sbg.ac.at (V. Hoeck), cristinagruian@yahoo.com (C. Gruian), viosimon@phys.ubbcluj.ro (V. Simon). http://dx.doi.org/10.1016/j.clay.2014.05.023 0169-1317/© 2014 Elsevier B.V. All rights reserved. error, is a complex issue. This is due to the nonstoichiometric mineral reactions, the occurrence of phases which are not in equilibrium, and last but not least, the wide range of temperatures recorded within a single kiln load and even within the same pot (Gosselain, 1992; Maggetti et al., 2011). In search of a more precise evaluation of the firing temperature for ancient ceramics, several spectroscopic methods are used: Fourier transform infrared, Raman, and electron paramagnetic resonance (EPR). The latter, also called electron spin resonance (ESR), ever since it was invented by Zavoisky (1945), has been widely applied in various fields, such as crystal chemistry (e.g., coordination, distortion, oxidation state etc.), behaviour of transition metals compounds, as well as effects of natural or artificial radiation, and dating. Of particular interest in archaeometry are the EPR studies on clay minerals (e.g., Allard et al., 2012; Balan et al., 2000; Elsass and Olivier, 1978; Manhães et al., 2002; Mestdagh et al., 1980; Morichon et al., 2008) and fired clays and ceramics (Bensimon et al., 1999; Cano et al., 2013; Dobosz and Krzyminiewski, 2007; Felicissimo et al., 2010; Gualtieri and Del Monaco, 1996; Ionescu et al., 2010; Matsuoka and Ikeya, 1995; Mota et al., 2009; Presciutti et al., 2005). Clay minerals and clays contain paramagnetic ions, paramagnetic defect centres and organic free radicals (Lück et al., 1993), which produce resonance signals, therefore clays are suitable for EPR investigation. “A key point in the characterisation of pottery is to assess firing conditions” (Nodari et al., 2004) and this can be achieved by various methods, including OM, XRD, SEM and EMPA. Some of these methods need a relatively large amount of material, others reveal thermal changes only in isolated C. Ionescu et al. / Applied Clay Science 97–98 (2014) 138–145 points, and thus the results might be misleading. On the other hand, the behaviour of paramagnetic species gives an overall insight and very little sample is needed. A clay with a known composition was selected in order to investigate its EPR behaviour upon thermal treatment. The aims of this study were: a) to define the relation between the resonance signal characteristics on one side, and temperature and composition of thermally treated clay on the other side, and b) to demonstrate the relevance of EPR studies in archaeometry, in particular for ancient ceramics technology. This paper will not deal with the physical principles behind the EPR spectra, but rather with the relation between the shape of the resonance signal on one side, and composition and firing temperature, on the other side. 2. Experimental: samples and methods This study was carried out on a Miocene mudstone (named here ‘raw clay’ and ‘clay’) extracted from a quarry in the eastern part of the Transylvanian Basin (Romania). It is used, after light grinding and 139 being passed through a 1 mm sieve, without any additional tempering material, to produce a well-known type of pottery, either black or red in colour, depending on firing atmosphere. The workshop is located in the village of Marginea, in NE Romania. Twenty six briquettes ~ 50 × 50 × 10 mm in size, hand-modelled from the raw clay, were dried at room temperature for 72 h and subsequently heated with a rate of 10 °C/min in an electric kiln (Nabertherm furnace) at 1 atmosphere of pressure and in oxidizing conditions (Fig. 1). The briquettes were kept at 400 °C, 500 °C, 600 °C, 700 °C, 800 °C, 900 °C, 1000 °C and 1100 °C for 2, 4 and 8 h soaking time, respectively. The heating at 1200 °C was interrupted after 10 min as the sample melted and started to boil. The recorded data show some differences for various soaking times. Only the data for 2 h heating will be presented here because it can be regarded as having produced effects similar to that of fast bonfire heating (Shepard 1976) or slow heating in a (primitive) kiln (e.g., Maniatis, 2009). The influence of soaking time upon resonance signal will be presented elsewhere. Fig. 1. Change in colour from the raw clay (a) to the clay briquettes heated for 2 h at 400 °C (b), 600 °C (c), 700 °C (d), 800 °C, 900 °C (f), 1000 °C (g), 1100 °C (h) and 1200 °C (i). Scale bar = 1 cm. 140 C. Ionescu et al. / Applied Clay Science 97–98 (2014) 138–145 From each heated briquette a slice was cut in order to prepare thin sections for optical microscopy. Additionally, to identify the Mn- and Fe-bearing main phases, four thin sections and four polished thin sections were obtained from sherd of red ceramic vessels fired for one day, at ~ 850 °C, at the Marginea workshop. About 12 g of raw clay and fired briquettes (bulk samples) were handmilled in an agate mortar. The mineral composition of the samples was determined by XRPD using a Shimatzu 6000 diffractometer with CuKα radiation (λ = 1.5418 Å) and Ni filter. The samples were measured from 5 to 60°2θ, with a scan speed of 2°/min. The chemistry of the major oxides in the bulk raw clay (~ 10 g of powdered sample) was obtained by inductively-coupled emission spectroscopy (ICP-ES) at Acme Labs, Vancouver (Canada). For analytical details see Hoeck et al. (2009). The mineral chemistry of the red ceramic sherds was obtained by electron microprobe from the polished thin sections coated with carbon. The JXA Superprobe 8600 (University of Salzburg) equipped with four wave-dispersive and one Si(Li) energy-dispersive spectrometer operated at 15 kV accelerating voltage, 40 nA beam current and ≤5 μm electron-beam diameter. The final composition of minerals and matrix was calculated following the ZAF procedure. The detection limits (2 σ) were 0.02 mass% for MgO, CaO, Na2O and K2O, 0.03 mass% for Al2O3 and MnO, 0.04 mass% for TiO2, 0.05 wt.% for SiO2, 0.06 mass% for FeO and 0.07 mass% for P2O5. The counting time was 20 s for the peak and 10 s for the background. For more details on analytical conditions see Ionescu et al. (2011a) and Ionescu and Hoeck (2011). A few grams of powdered samples, including the red ceramic sherds, were analysed by EPR in order to investigate the behaviour of paramagnetic ions such as Fe3+ and Mn2+ when a magnetic field is applied. An ADANI spectrometer operating in the X-band frequency (9.4 GHz), with 4 Gauss (G) amplitude modulation, was employed for the EPR investigation. The value of g factor was calculated based on the DPPH standard of g = 2.0036. The spectra were collected in 100 s of scanning time, at room temperature. The investigation conditions were similar for both the raw clay and the heated clay briquettes. The recorded spectra show ‘resonance lines (bands)’ or ‘resonance signals’ in the form of soft and sharp curves. For a more precise view of the position and shape of the resonance lines, the EPR spectrum is displayed as the first derivative of the original signal. A signal is characterised by four main parameters: the g factor, the line width (ΔB), the amplitude and the asymmetry. The g value derives from the resonance condition related to the external magnetic field B, measured in gauss (Chiesa and Giamello, 2000). The line width ΔB is the peak-to-peak distance in respect to the external magnetic field. The amplitude, which is subject of gain conditions, is the peak-to-peak distance along the vertical axis, measured in arbitrary units. The asymmetry is the deviation from an ideal spectrum (Langer et al., 2001; Watanabe et al., 2009; Weiss et al., 2004). A Shimadzu DTG-60H derivatograph with 5 °C/min heating rate and alumina open crucibles was employed for differential thermal analysis (DTA) and thermogravimetric analysis (TGA) in order to identify organic material in the raw clay. 3. Results The optical microscopy and XRPD (Figs. 2 and 3) show that raw clay consists predominantly of illite and muscovite as well as quartz and calcite. Subordinately, clinochlore, K-feldspar, plagioclase, altered biotite and dolomite occur. In common clays, small amounts of Fe- and Mn-oxihydroxides might be present as nanoscale grains (Post, 1999) but in sample studied they were not detected by XRPD. Chemically, the raw clay has a Ca- and Fe-rich composition, with: 51.57 mass% SiO2, 14.91 mass% Al2O3, 5.90 mass% Fe2O3TOT, 2.33 mass% MgO, 7.31 mass% CaO, 0.92 mass% Na2O, 2.78 mass% K2O, 0.71 mass% TiO2, 0.11 mass% P2O5, 0.11 mass% MnO, 12.9 mass% loss on ignition and 2.33 mass% TOTC. The thermal analyses show mainly the loss of H2O and OH−. Some traces of organic matter might be also present as indicated by the weak exothermic event at 330 °C (Tămăşan et al., 2009). By heating, the light brown grey colour of the raw clay (Fig. 1a) changes to reddish hues (Fig. 1b–h). The reddish yellow acquired at 400 °C remains till 600 °C. With 700 °C and 800 °C the colour turns to light reddish brown, whereas at 900 °C and 1000 °C it becomes red. At 1100 °C the samples become again a reddish brown. The melt formed after 10 min heating at 1200 °C turned upon cooling into a reddish black glassy mass (Fig. 1i) similar to the ‘ceramic slag’ which results from over-firing ceramics (Hoeck et al., 2012). Table 1 shows the Munsell (1994) classification of the sample colours. Optical microscopy and XRPD (Ionescu et al., 2011b; Tămăşan et al., 2009) reveal the progressive diminishing of illite, muscovite, clinochlore, calcite and dolomite lines and the formation of new minerals (hematite, gehlenite, clinopyroxene and maghemite) as well as amorphous phase (glass), with increasing temperature (Fig. 3). The birefringent and inhomogeneous light brown matrix at lower temperatures (up to 700 °C), gradually transforms into a reddish glassy mass at 1200 °C. At this temperature, the resulting material is almost entirely isotropic and opaque, and includes only scarce grains of maghemite (γ-Fe2O3) and some relics comprised mostly of quartz. Fig. 2. X-ray powder diffractogram of the raw clay. Mineral abbreviations according to Whitney and Evans (2010): Ilt for illite, Qz for quartz, Ms for muscovite, Clc for clinochlore, Bt for altered biotite, Fsp for feldspar, Cal for calcite, and Dol for dolomite. C. Ionescu et al. / Applied Clay Science 97–98 (2014) 138–145 141 Fig. 3. Mineral composition of raw and heated clay as inferred from OM and XRPD. Abbreviations: Ilt for illite, and Ms for muscovite. Here only the manganese and iron-bearing minerals identified by EMPA in the red ceramic sherds are considered. The most important carriers of MnO (Mntot expressed as MnO) are: clinochlore (0.02–0.6 mass%), Fe-oxihydroxides (0.1–5.5 mass%) and calcite and dolomite (up to 0.2 mass%). The same holds for FeO (Fetot expressed as FeO), which reaches up to ~ 67 mass% in Fe-oxihydroxides, ~34–35 mass% in chlorite, and up to ~2 mass% in carbonates. Additionally, Fe is found in muscovite (3–4 mass%), altered biotite (~23 mass%) and illite (up to ~9 mass%). As newly-formed phases, small and isolated Fe-rich minerals (hematite), gehlenite and clinopyroxene were identified by EMPA. The EPR spectrum recorded for the raw clay (Fig. 4) displays several resonance lines. The first is a weak signal positioned at 1607 G, with g = 4.18. The second is a sextet of narrow and sharp lines (hyperfine signals) equally spaced, occurring between ~ 3200 G and ~ 3750 G magnetic field, at g ≅ 2 (values from 1.98 to 2.03). Accompanying the hyperfine sextet and covered by it, there are two weak and large satellite lines. In between the sextet lines there are doublets with lower amplitude (‘forbidden transitions’) with a similar symmetrical shape, except the middle one at g ≅ 2, which is asymmetrical. The amplitude of the hyperfine lines decreases at higher values of the magnetic field. Compared to the raw clay (Fig. 4), the heated clay produces significantly different resonance signals (Figs. 4 and 5). With increasing temperature, the small resonance line at g = 4.18 shifts to slightly higher g (4.23). The apparent disappearance of g ~4.2 is due to the change in gain during the measurement. In the 400 °C to 600 °C temperature interval, the hyperfine sextet still persists, marked by only a slight and gradual diminishing of the amplitude. The asymmetric line at g ≅ 2 becomes weaker at 400 °C and is not seen any more over 600 °C. At 700 °C only remnants of the hyperfine structure overprinting a broader signal are preserved. Above 700 °C the spectra show a shape marked by narrowing and broadening. Fig. 6 illustrates the non-linear trend of evolution of the resonance line width at g ≅ 2, with increasing temperature. Between 700 °C and 1000 °C the width continuously decreases, from 780 G to 450 G. Above 1000 °C the width increases again, reaching the maximum of 870 G at 1200 °C. 4. Discussion 4.1. Interpretation of the EPR spectra The resonance spectra of clays arouse problems in interpretation due to the mineral phase complexity and the spectral overlap of various EPR active species. It becomes even more problematic if the clay is heated to high temperatures, over 700 °C as for example in the case of most ancient ceramics. For archaeometric purposes, the question regarding to what extent the shape of the EPR spectra at a given mineral composition is a function of temperature, is important. Table 1 Munsell colours (Munsell, 1994) of raw and heated clay. Raw clay Heated clay (temperature) 400 °C 500 °C 600 °C 700 °C 800 °C 900 °C 1000 °C 1100 °C 1200 °C Munsell colour index Munsell colour 2.5 Y 6/2 7.5 YR 6/6 7.5 YR 6/6 7.5 YR 6/6 5 YR 6/4 5 YR 6/4 2.5 YR 5/6 2.5 YR 5/6 2.5 YR 5/4 10 R 2.5/1 Light brown grey Reddish yellow Reddish yellow Reddish yellow Light reddish brown Light reddish brown Red Red Reddish brown Reddish black Fig. 4. The first derivative of the EPR spectra recorded from the raw clay and the thermally treated clay from 400 °C to 700 °C, for 2 h. The 500 °C spectrum is not shown here, as it is identical with that from 400 °C. 142 C. Ionescu et al. / Applied Clay Science 97–98 (2014) 138–145 Fig. 5. Resonance spectra obtained for the raw clay and clay heated between 700 °C and 1200 °C. It is known that the characteristics of the EPR line depend on several physical parameters, e.g., type, behaviour, amount and distribution of paramagnetic ions (Chiesa and Giamello, 2000; Guskos et al., 2002; Mohan et al., 2008) as well as the grain size (Sholom et al., 1998). In general, a narrow EPR line is assigned to a paramagnetic centre disposed in an ordered (crystalline) environment (Chiesa and Giamello, 2000). A broad line is thought to reflect a more disordered system around the resonant centre such as an amorphous or glassy one. The interaction between different paramagnetic ions also has an influence on the signal width. Isolated paramagnetic centres will produce narrow lines whereas their clustering (agglomeration) will enlarge the signal (Park et al., 2001; Rabenstein and Shin, 1995). The relation between the type of the host mineral and the EPR signal is presented in detail by Mestdagh et al. (1980). In the following, firstly the resonance spectrum recorded from the raw clay and after that the spectra of the heated clay briquettes will be presented. 4.1.1. Raw clay The comparison of the EPR spectrum (Fig. 4) with the mineralogy, mineral chemistry and chemical composition of the raw clay suggests that the recorded signals are due most likely to Mn2+ and Fe3 + ions Fig. 6. Non-linear variation of the resonance line width ΔB at g ≅ 2, between 700 °C and 1200 °C. The vertical bars represent the accuracy of measurements. and to structural defects in Mn and Fe-bearing phases. As shown above, the most important Mn-bearing minerals identified by EMPA in the red ceramics are Fe-oxihydroxides, clinochlore, and carbonates, where Mn2+ may replace Fe2+, Ca2+ and Mg2+. Such manganese carriers were also described by Wakabayashi (1963), Lupei et al. (1972), Prissok and Lehmann (1986), Lloyd et al. (1993), Götze et al. (2002), Franco et al. (2003), Polikreti and Maniatis (2004), Garribba and Micera (2006). In the studied clay, Fe2 + and Fe3 + are found in the same Mn2 + carriers, and additionally in biotite, illite and muscovite (see also Balan et al., 2000; Cesare et al., 2005; Murad and Wagner, 1996). The weak resonance line at g ≅ 4.2 can be assigned to Fe3+ located in orthorhombic positions (Polikreti and Maniatis, 2002) and the substitution of Fe3+ for Al3+ in the structure of clay minerals (Komusiński et al., 1981; Lück et al., 1993; Meads and Malden, 1975; Murad and Wagner, 1996). The sextet of hyperfine signals centred at g ≅ 2 might be due to several causes, among which Mn2+ ions located in an environment with octahedral symmetry is widely accepted (Crook et al., 2002; Elsass and Olivier, 1978; Franco et al., 2003; Gehring et al., 1993; McBride, 1995; Mohan et al., 2008; Muller and Calas, 1993). The sextet accompanied by external satellites and especially the doublets of the ‘forbidden transitions’ in the interspace of the hyperfine lines are typical for Mn2+ in calcite and dolomite (Garribba and Micera, 2006; Lupei et al., 1972). The two satellite signals outside the sextet range may represent also overprints from Fe3 + ions in Fe-oxihydroxides (Gehring et al., 1990; Guskos et al., 2002). Structural defects in silicates, e.g., vacancy-oxygen pairs, Si\O− and Al\O−-Al centres (Allard et al., 1994; Delineau et al., 1994; Markevich et al., 1998; Tokuda and Seki, 2000) are responsible for the narrow and asymmetric signal at g ≅ 2 seen in the middle of the hyperfine sextet (Fig. 4). O-centred and C-centred free radicals (Malhotra and Graham, 1984), radiation induced defects (Allard et al., 2012) and broken C bonds originating in some organic material may also give rise to this signal. On the other hand, resonance signals at g ≅ 2 can also be assigned to Fe3 + ions (Berger et al., 1995; Clerjaud, 1975; Kurkjian and Sigety, 1968), in particular to Fe3 + located in octahedral sites (Albon et al., 2008). 4.1.2. Heated clay The signal at g ≅ 4.2 is preserved over the whole range of temperatures, with only small amplitude variations. This situation argues for both the Fe3+ which remained associated to few orthorhombic positions and to a small amount to Fe3+ replacing Al3+ in the structure of clay minerals and released upon their gradual collapse (see also Murad and Wagner, 1996; Nodari et al., 2007). The most obvious changes occur at g ≅ 2. The sextet of sharp and narrow signals from the raw clay (Fig. 4) persists up to 600 °C. Very small traces of the sextet can be still seen at 700 °C. At temperatures over 700 °C, the hyperfine sextet is replaced by a single broad and intense signal. The disappearance of the hyperfine sextet is generally related to the oxidation of Mn2 + (EPR active) to Mn3 +, which is EPR silent (Mohan et al., 2008; Ravi Kumar et al., 2012; Reddy et al., 2006). In the samples studied, this process relates to Mn-bearing Feoxihydroxides and clinochlore. In calcite and dolomite, a part of the oxidized Mn3+ could be accommodated by lattice defects or might occur alternatively as nanoscale impurities with a grain size well below the resolution of the microprobe. The occurrence of Mn4+ ions due to oxidation during heating can also contribute to the broad component in the g ≅ 2 region (see also Ivanova et al., 2011). Along with increasing amount of Mn3+, the decrease of Mn2+ in octahedral positions supports the decrease of the signal amplitude (Mohan et al., 2008). However, to which extent Mn2+ is oxidized to Mn3+ or Mn4+ during heating cannot be solved with EMPA. C. Ionescu et al. / Applied Clay Science 97–98 (2014) 138–145 A continuously narrowing resonance signal is recorded up to 1000 °C, in agreement with the mineralogical observations pointing to the main changes in this temperature range. The change of the width of EPR signal g ≅ 2 over 700 °C reflects both the oxidation of Fe2+ into Fe3+ (Mangueira et al., 2011; Presciutti et al., 2005) and the gradual entering of Fe3+ in new crystalline phases (see also Molera et al., 1998). The oxidation to Fe3 + starts around 400 °C and ends after 700 °C (Maritan et al., 2006; Shimada et al., 2003). Most of hematite nucleates and grows before 700–750 °C (Maniatis et al., 1981; Nodari et al., 2007), on the expenses of Fe-oxihydroxides (Gualtieri and Venturelli, 1999) or clinochlore (Bai et al., 1993; Barlow et al., 1997; Jones, 1981). Over 900 °C, decomposition of biotite (Cesare et al., 2005), muscovite and illite will also release iron, which will be trapped in the structure of newly-formed phases (see also Kreimeyer, 1987; Molera et al., 1998) such as gehlenite and clinopyroxene. At 1200 °C, maghemite forms (Ionescu et al., 2011b; Tămăşan et al., 2009). Over 1000 °C, the destruction of the crystalline structure of most of the mineral phases and the formation of an increased amount of amorphous phase influence the resonance signal. The Fe3+ ions will record a structurally disordered environment and consequently the width of the resonance line at g ≅ 2 will increase. The slight asymmetry of the signal is due to overlapping of two or more resonance signals (Lyons, 1996) and points to the coexistence of amorphous/vitreous phase and relic crystalline phases. 4.2. Relation of the EPR spectra with heating temperature So far, previous attempts to link the resonance spectra to the firing temperature of a multi-mineral material, e.g., ancient ceramics, have failed to give precise constraints which might be useful for any calibration. This is due mostly to the manifold causes of resonance signals and in particular to spectral overprint of different paramagnetic ions. Nevertheless, several authors have noticed a strong connection between the resonance signal and temperature (Bartoll et al., 1996; Chiesa and Giamello, 2000; Dobosz and Krzyminiewski, 2007; Gehring et al., 1990, 1993; Lück et al., 1993; Mangueira et al., 2011; Morichon et al., 2008; Mota et al., 2009). Attempts based mainly on refiring experiments, were made in order to define the firing temperature or at least the temperature range for ancient ceramics (Felicissimo et al., 2010; Matsuoka and Ikeya, 1995; Presciutti et al., 2005; Warashina et al., 1981). The disappearance of the weak resonance signals at g ≅ 2 was observed for clays fired at up to 800 °C (Bensimon et al., 1999). Felicissimo et al. (2010) described the narrowing of the hyperfine sextet lines assigned to Mn2+ from calcite, if heated up to 700 °C. Similarly, change in the resonance signal assigned to Mn2+ between 600 °C and 700 °C, when dehydroxylation of clay minerals takes place is presented by Gehring et al. (1993). According to these authors, the broad signal at g = 2 occurring at 800 °C is due to formation of Fe-phases such as hematite, after the collapse of the clay mineral structure. A continuous narrowing of the signal at g ≅ 2 up to ~ 900–1000 °C, followed by the widening of the line at higher temperatures is presented also by Mota et al. (2009) and Mangueira et al. (2011) for heated clay. The relation between temperature and width has a non-linear trend (Fig. 6). Up to around 1000 °C, the width of the resonance line is governed by solid-state reactions. Heating above 1000 °C will provoke the enlargement of the signals, most likely due to a higher amount of amorphous phase. Thus, two domains can be envisaged, one which is generated to a large extent by crystalline phases up to a temperature range of 900–1000 °C and another one, which is above this temperature threshold and is characterised by a vitreous structure. This assumption is confirmed by mineralogical data which show birefringent (crystalline) phases at up to 900 °C. Above this temperature, there is a mixture of birefringent and isotropic (amorphous or vitreous) phases. The latter occurs in increasing amount mainly over 1000 °C and becomes predominant at 1200 °C. The illitic clay relatively rich in CaO (7.31 mass%) and Fe2O3TOT (5.90 mass%), heated between 700 °C and 1200 °C, gives resonance 143 spectra with different width, intensity and asymmetry, due to: a) the oxidation of Fe2+ and Mn2+ from the raw clay into Fe3+ and Mn3+ respectively, b) the variable environment provided for the resonance sources (Fe3 +, Mn2 + and defect centres), in particular formation of new phases (e.g., hematite, gehlenite, clinopyroxene), including glass. All spectra are dominated by resonance signals at g ≅ 2. The gradual narrowing of the resonance signal continues only up to 900–1000 °C. Heating above this temperature causes the enlargement and flattening of the EPR line due to melting, confirming previous observations by Mota et al. (2009). However, the overlap of resonance signals collected from a wide range of sources, the different responses of mineral phases to heating, and the formation of different environments have to be taken into account when assigning a certain resonance signal to a certain temperature. 4.3. Heating atmosphere One of the best evidences of the firing atmosphere is the colour of the ceramics. Between 700 °C and 1100 °C, the clay briquettes (Fig. 1b–h) show a colour specifically gained by carbonate-rich clay with relatively high Fe content, when heated in an oxidizing atmosphere (Kreimeyer, 1987; Maritan, 2004; Molera et al., 1998; Murad and Wagner, 1998; Nodari et al., 2007). The colour change is due not so much to variations in the hematite content but most likely to incorporation of Fe into newly formed phases, e.g., pyroxene (see also Kreimeyer, 1987; Molera et al., 1998). The behaviour of the resonance signals also reflects a predominantly oxidizing environment, i.e. transformation of Mn2 + to Mn3 + and Fe2 + to Fe3 + respectively. The formation of some maghemite (Ionescu et al., 2011b) reflects a sluggish diffusion of oxygen within the melt formed in the kiln at 1200 °C, rather than decomposition of organic matter (e.g. in Maritan et al., 2006). 5. Conclusions The EPR signals recorded from raw and heated carbonate-rich illitic clay show systematic change as a result of a complex interaction of several factors: a) temperature, b) primary chemical and mineralogical composition of the material, c) heating atmosphere, d) specific phase transformations at various temperatures, e) distribution of paramagnetic species within various phases and f) the local environment for each paramagnetic ion. As most of the carbonate-rich clays used for traditional ceramics contain small amounts of Mn and Fe, an EPR study can be useful in the interpretation of firing temperature range. It has to be kept in mind that the heating of clay, as long as it is in crystalline state, has a trend in resonance signal width (narrowing), opposite to the signal evolution at temperatures where amorphous/vitreous structure prevails (broadening). Ancient ceramics were very rarely fired above 1000 °C therefore narrow resonance signals at g ~ 2 would point rather to temperatures close but below this threshold (i.e. 800–950 °C). A broader signal with remnants of the hyperfine sextet will indicate a temperature around 700 °C, whereas the lack of these remnants in a broad signal would suggest temperatures over 1000 °C. Nevertheless, the EPR data should be supported by other observations such as optical microscopy, XRPD, SEM or EMPA, to obtain a more realistic view of the technological constraints for ancient pottery. These data may help to constrain the EPR signal shape. Even more, in cases where the raw material of ancient ceramics is known and the ceramic paste is not contaminated much by (Mn and Fe-bearing) temper material, a comparison between re-fired ceramics at a certain temperature range and the EPR data obtained from firing experiments of the raw clay could lead to reasonable results. Additional combination with other thermometers will improve the potential of EPR as a useful tool in archaeometry. 144 C. Ionescu et al. / Applied Clay Science 97–98 (2014) 138–145 Acknowledgments This study was supported by the Romanian National Authority for Scientific Research, CNCS-UEFISCDI, through PN-II-ID-PCE-2011-30881 project. The comments and observations of anonymous reviewers significantly contributed to the improvement of the paper. Thanks are due also to Dr. Otis N. Crandell for the careful checking of the English language and to Mrs. Monica Mereu of Babeş-Bolyai in University ClujNapoca for the computer-assisted graphics, respectively. References Albon, C., Mureşan, D., Vandenberghe, R.E., Simon, S., 2008. Iron environment in calciumsoda-phosphate glasses and vitroceramics. J. Non-Cryst. Solids 354, 4603–4608. Allard, T., Muller, J.P., Dran, J.C., Menager, M.T., 1994. Radiation-induced paramagnetic defects in natural kaolinites: alpha dosimetry with ion beam irradiation. Phys. Chem. Miner. 21, 85–96. Allard, T., Balan, E., Calas, G., Fourdrin, C., Morichon, E., Sorieul, S., 2012. Radiation-induced defects in clay minerals: a review. Nucl. Inst. Methods Phys. Res. B 277, 112–120. Bai, T.B., Guggenheim, S., Wang, S.J., Rancourt, D.G., Koster van Groos, A.F., 1993. Metastable phase relations in the chlorite-H2O system. Am. Mineral. 78, 1208–1216. Balan, E., Allard, T., Boizot, B., Morin, G., Muller, J.P., 2000. Quantitative measurement of paramagnetic Fe3+ in kaolinite. Clay Clay Miner. 48, 439–445. Barlow, S.G., Manning, D.A.C., Hill, P.I., 1997. Influence of time and temperature on reactions and transformations of clinochlore as a ceramic clay mineral. Br. Ceram. Trans. 96 (5), 195–198. Bartoll, J., Tani, A., Ikeya, M., Inada, T., 1996. ERS investigations of burnt soil. Appl. Magn. Reson. 11, 577–596. Bensimon, Y., Deroide, B., Zanchetta, J.V., 1999. Comparison between the electron paramagnetic resonance spectra obtained in X- and W-bands on a fired clay: a preliminary study. J. Phys. Chem. Solids 60, 813–818. Berger, R., Kliava, J., Yahiaoui, E.M., Bissey, J.C., Zinsou, P.K., Béziade, P., 1995. Diluted and non-diluted ferric ions in borate glasses studied by electron paramagnetic resonance. J. Non-Cryst. Solids 180, 151–163. Cano, N.F., Munita, C.S., Watanabe, S., Barbosa, R.F., Chubaci, J.F.D., Tatumi, S.H., Neves, E.G., 2013. OSL and EPR dating of pottery from the archaeological sites in Amazon Valley, Brazil. Quat. Int. http://dx.doi.org/10.1016/j.quaint.2013.05.042 (in print). Cesare, B., Meli, S., Nodari, L., Russo, U., 2005. Fe3+ reduction during biotite melting in graphitic metapelites: another origin of CO2 in granulites. Contrib. Mineral. Petrol. 149, 129–140. Chiesa, M., Giamello, E., 2000. Electron spin resonance spectroscopy. In: Meyers, R.A. (Ed.), Encyclopedia of analytical chemistry. Applications, theory and instrumentation, vol. 13. John Wiley & Sons Ltd., Chichester, New York, Brisbane, Singapore, Toronto, pp. 12040–12070. Clerjaud, B., 1975. Intensities of the forbidden EPR transitions AM = 2 for different orientations of the RF field related to the static field. Phys. Status Solidi B 72, K33–K37. Crook, N.P., Hoon, S.R., Taylor, K.G., Perry, C.T., 2002. Electron spin resonance as a high sensitivity technique for environmental magnetism: determination of contamination in carbonate sediments. Geophys. J. Int. 149, 328–337. Cultrone, G., Rodriguez-Navarro, C., Sebastian, E., Cazalla, O., De la Torre, M.J., 2001. Carbonate and silicate phase reactions during ceramic firing. Eur. J. Mineral. 13, 621–634. Delineau, T., Allard, T., Muller, J.-P., Barres, O., Yvon, J., Cases, J.-M., 1994. FTIR reflectance vs. EPR studies of structural iron in kaolinites. Clay Clay Miner. 42 (3), 308–320. Dobosz, B., Krzyminiewski, R., 2007. Characteristic of paramagnetic centres in burnt clay and pottery by the EPR method. Radiat. Meas. 42 (2), 213–219. Elsass, F., Olivier, D., 1978. Infrared and electron spin resonance studies of clays representative of the sedimentary evolution of the basin of Autun. Clay Miner. 13, 299–308. Felicissimo, M.P., Peixoto, J.L., Bittencourt, C., Tomasi, R., Houssiau, L., Pireaux, J.J., Rodrigues-Filho, U.P., 2010. SEM, EPR and ToF-SIMS analyses applied to unravel the technology employed for pottery-making by pre-colonial Indian tribes from Pantanal, Brazil. J. Archaeol. Sci. 37, 2179–2187. Franco, R.W.A., Pelegrini, F., Rossi, A.M., 2003. Identification and valuation of paramagnetic radicals in natural dolomites as an indicator of geological events. Phys. Chem. Miner. 30, 39–43. Garribba, E., Micera, G., 2006. Determination of the hyperfine coupling constant and zerofield splitting in the ESR spectrum of Mn2+ in calcite. Magn. Reson. Chem. 44, 11–19. Gehring, A.U., Karthein, R., Reller, A., 1990. Activated state in the lepidocrocite structure during thermal treatment. Naturwissenschaften 77 (4), 177–179. Gehring, A.U., Fry, I.V., Sposito, G., 1993. Residual manganese(II) entrapped in singlelayer-hydrate montmorillonite interlayers. Clay Clay Miner. 41 (5), 565–569. Gosselain, O.P., 1992. Bonfire of the enquiries. Pottery firing temperatures in archaeology: what for? J. Archaeol. Sci. 19, 243–259. Götze, J., Plötze, M., Götte, T., Neuser, R.D., Richter, D.K., 2002. Cathodoluminiscence (CL) and electron paramagnetic resonance (EPR) studies of clay minerals. Mineral. Petrol. 76, 195–212. Gualtieri, S., Del Monaco, S., 1996. An attractive way to recover information from ancient ceramics using EPR spectroscopy. Rev. Chem. Intermed. 22, 435–438. Gualtieri, A.F., Venturelli, P., 1999. In situ study of the goethite-hematite phase transformation by real time synchrotron powder diffraction. Am. Mineral. 84, 895–904. Guskos, N., Papdopoulos, G., Likodimos, V., Patapis, S., Yarmis, D., Przepiera, A., Przepiera, K., Majszczyk, J., Typek, J., Wabia, M., Aidinis, K., Drazek, Z., 2002. Photoacustic, EPR and electrical conductivity investigations of three synthetic mineral pigments: hematite, goethite and magnetite. Mater. Res. Bull. 37, 1051–1061. Hoeck, V., Ionescu, C., Ghergari, L., Precup, C., 2009. Towards mineralogical and geochemical reference groups for some Bronze Age ceramics from Transylvania (Romania). Stud. Univ. Babeş-Bolyai Geol. 54 (2), 41–51. Hoeck, V., Ionescu, C., Metzner-Nebelsick, C., Nebelsick, L.D., 2012. Mineralogy of the ceramic slags from the Bronze Age funerary site in Lăpuş, NW Romania. Geol. Q. 56 (4), 649–664. Ionescu, C., Hoeck, V., 2011. Firing-induced transformations in Copper Age ceramics from NE Romania. Eur. J. Mineral. 23 (6), 937–958. Ionescu, C., Simon, V., Ghergari, L., 2010. EPR spectroscopy applied to the study of ancient ceramics: preliminary results for Bronze Age pottery from Ilişua. Stud. Univ. BabeşBolyai Phys. LV, 79–83. Ionescu, C., Hoeck, V., Ghergari, L., 2011a. Electron microprobe analysis of ancient ceramics: a case study from Romania. Appl. Clay Sci. 53, 466–475. Ionescu, C., Hoeck, V., Simon, V., 2011b. Effect of the temperature and the heating time on the composition of an illite-rich clay: an XRPD study. Stud. Univ. Babeş-Bolyai Phys. LVI, 69–78. Ivanova, S., Zhecheva, E., Nihtianova, D., Stoyanova, R., 2011. Nano-domain structure of Li4Mn5O12 spinel. J. Mater. Sci. 46, 7098–7105. Jones, A.A., 1981. Charges on the surfaces of two chlorites. Clay Miner. 16, 347–359. Komusiński, J., Stoch, L., Dubiel, S.M., 1981. Application of electron paramagnetic resonance and mossbauer spectroscopy in the investigation of kaolinite-group minerals. Clay Clay Miner. 29 (1), 23–30. Kreimeyer, R., 1987. Some notes on the firing colour of clay bricks. Appl. Clay Sci. 2 (2), 175–183. Kurkjian, C.R., Sigety, E.A., 1968. Coordination of Fe3+ in glass. Phys. Chem. Glasses 9, 73–83. Langer, J.J., Krzyminiewski, R., Kruczynski, Z., Gibinski, T., Czajkowski, I., Framski, G., 2001. EPR and electrical conductivity in microporous polyaniline. Synth. Met. 122, 359–362. Lloyd, R.V., Morrison, J.W., Lumsden, D.N., 1993. The influence of ferrous and ferric iron on the Mn+2 partitioning ratio and ESR signal of synthetic dolomite. Geochim. Cosmochim. Acta 57 (5), 1071–1078. Lück, R., Stösser, R., Gyepesová, D., Slosiariková, H., Kolditz, L., 1993. EPR study of montmorillonites from several deposits. Chem. Pap. 47 (2), 79–84. Lupei, V., Lupei, A., Ursu, I., 1972. Further EPR studies of forbidden hyperfine transitions of Mn2+ in calcite. Phys. Rev. B 6 (11), 4125–4131. Lyons, R.G., 1996. Back to basics: qualitative spectral analysis as an investigatory tool, using calcite as a case study. Appl. Radiat. Isot. 47 (11–12), 1385–1391. Maggetti, M., 1982. Phase analysis and its significance for technology and origin. In: Olin, J.S., Franklin, A.D. (Eds.), Archaeological ceramics. Smithsonian Institute Press, Washington D.C., pp. 121–133. Maggetti, M., Neururer, C., Ramseyer, D., 2011. Temperature evolution inside a pot during experimental surface (bonfire) firing. Appl. Clay Sci. 53, 500–508. Malhotra, V.M., Graham, W.R.M., 1984. Study of mineral fines in tar sand bitumen and their acid sensitivity using EPR and FTIR techniques. Symposium on characterization and chemistry of oil shales presented before the divisions of fuel chemistry and petroleum chemistry, Inc. American Chemical Society, St. Louis Meeting, April 8–13 1984, pp. 261–268. Mangueira, G.M., Toledo, R., Teixeira, S., Franco, R.W.A., 2011. A study of the firing temperature of archeological pottery by X-ray diffraction and electron paramagnetic resonance. J. Phys. Chem. Solids 72, 90–96. Manhães, R.S.T., Auler, L.T., Sthel, M.S., Alexandre, J., Massunaga, M.S.O., Carrió, J.G., dos Santos, D.R., da Silva, E.C., Garcia-Quiroz, A., Vargas, H., 2002. Soil characterisation using X-ray diffraction, photoacoustic spectroscopy and electron paramagnetic resonance. Appl. Clay Sci. 21, 303–311. Maniatis, Y., 2009. The emergence of ceramic technology and its evolution as revealed with the use of scientific techniques. In: Shortland, A.J., Freestone, I.C., Rehren, T. (Eds.), From mine to microscope: Advances in the study of ancient technology. Oxbow Books, Oxford, pp. 11–28. Maniatis, Y., Simopoulos, A., Kostikas, A., 1981. Mössbauer study of the effect of calcium content on iron oxide transformations in fired clays. J. Am. Ceram. Soc. 64, 263–269. Maritan, L., 2004. Archaeometric study of Etruscan-Padan type pottery from the Veneto region: petrographic, mineralogical and geochemical-physical characterization. Eur. J. Mineral. 16, 297–307. Maritan, L., Nodari, L., Mazzoli, C., Milano, A., Russo, U., 2006. Influence of firing conditions on ceramic products: Experimental study on clay rich in organic matter. Appl. Clay Sci. 31, 1–15. Markevich, V.P., Mchedlidze, T., Suezawa, M., Murin, L.I., 1998. EPR study of hydrogenrelated radiation-induced shallow donors in silicon. Phys. Status Solidi B 210, 545–549. Matsuoka, Y., Ikeya, M., 1995. Electron paramagnetic resonance and Mössbauer spectra of iron ions in Bizen pottery. Jpn. J. Appl. Phys. 34, 6068–6072. McBride, M.B., 1995. On the natural Mn(II) EPR signal of SWy-1 montmorillonite. Clay Clay Miner. 43 (3), 383–384. Meads, R.E., Malden, P.J., 1975. Electron spin resonance in natural kaolinites containing Fe3+ and other transition metal ions. Clay Miner. 10, 313–345. Mestdagh, M.M., Vielvoye, L., Herbillon, A.J., 1980. Iron in kaolinite: II. The relationship between kaolinite crystallinity and iron content. Clay Miner. 15, 1–13. Mohan, N.K., Reddy, M.R., Jayasankar, C.K., Veeraiah, N., 2008. Spectroscopic and dielectric studies on MnO doped PbO-Nb2O5-P2O5 glass system. J. Alloys Compd. 458 (1–2), 66–76. Molera, J., Pradell, T., Vendrell-Saz, M., 1998. The colours of Ca-rich ceramic pastes: origin and characterization. Appl. Clay Sci. 13, 187–202. Morichon, E., Allard, T., Beaufort, D., Patrier, P., 2008. Evidence of native radiation-induced paramagnetic defects in natural illites from unconformity-type uranium deposits. Phys. Chem. Miner. 35, 339–346. C. Ionescu et al. / Applied Clay Science 97–98 (2014) 138–145 Mota, L., Toledo, R., Faria Jr., R.T., da Silva, E.C., Vargas, H., Delgadillo-Holtfort, I., 2009. Thermally treated soil clays as ceramic raw materials: characterization by X-ray diffraction, photoacoustic spectroscopy and electron spin resonance. Appl. Clay Sci. 43, 243–247. Muller, J.P., Calas, G., 1993. Mn2+-bearing kaolinites from lateritic weathering profiles: geochemical significance. Geochim. Cosmochim. Acta 57 (5), 1029–1037. Munsell, A., 1994. Munsell soil colour charts. Munsell Colour, New Windsor NY, (10 pp.). Murad, E., Wagner, U., 1996. The thermal behavior of an Fe-rich illite. Clay Miner. 31, 45–52. Murad, E., Wagner, U., 1998. Clays and clay minerals: the firing process. Hyperfine Interact. 117, 337–356. Nodari, L., Maritan, L., Mazzoli, C., Russo, U., 2004. Sandwich structures in the EtruscanPadan type pottery. Appl. Clay Sci. 27, 119–128. Nodari, L., Marcuz, E., Maritan, L., Mazzoli, C., Russo, U., 2007. Hematite nucleation and growth in the firing of carbonate-rich clay for pottery production. J. Eur. Ceram. Soc. 27, 4665–4673. Park, K., Novotny, M.A., Dalal, N.S., Hill, S., Rikvold, P.A., 2001. Effects of D-strain, g-strain, and dipolar interactions on EPR linewidths of the molecular magnets Fe8 and Mn12. Phys. Rev. B 65 (1), 1–5 (014426). Polikreti, K., Maniatis, Y., 2002. A new methodology for the provenance of marble based on EPR spectroscopy. Archaeometry 44 (1), 1–21. Polikreti, K., Maniatis, Y., 2004. Distribution changes of Mn2+ and Fe3+ on weathered marble surfaces measured by EPR spectroscopy. Atmos. Environ. 38, 3617–3624. Post, J.E., 1999. Manganese oxide minerals: Crystal structures and economic and environmental significance. PNAS 96 (7), 3447–3454. Presciutti, F., Capitani, D., Sgamellotti, A., Brunetti, B.G., Constantino, F., Viel, S., Segre, A., 2005. Electron paramagnetic resonance, scanning electron microscopy with energy dispersion X-ray spectrometry, X-ray powder diffraction, and NMR characterization of iron-rich fired clay. J. Phys. Chem. B 109, 22147–22158. Prissok, J., Lehmann, G., 1986. An EPR study of Mn2+ and Fe3+ in dolomites. Phys. Chem. Miner. 13, 331–336. Rabenstein, M.D., Shin, Y.K., 1995. Determination of the distance between two spin labels attached to a macromolecule. Proc. Natl. Acad. Sci. U. S. A. 92, 8239–8243. Ravi Kumar, A.V., Srinivasa Rao, C., Narasimha Rao, N., Ravi Kumar, V., Kityk, I.V., Veeraiah, N., 2012. Influence of valence and coordination of manganese ions on spectral and dielectric features of Na2SO4-B2O3-P2O5 glasses. J. Non-Cryst. Solids 358 (10), 1278–1286. 145 Reddy, M.S., Krishna, G.M., Veeraiah, N., 2006. Spectroscopic and magnetic studies of manganese ions in ZnO-Sb2O3-B2O3 glass system. J. Phys. Chem. Solids 67 (4), 789–795. Riccardi, M.P., Messiga, B., Duminuco, P., 1999. An approach to the dynamics of clay firing. Appl. Clay Sci. 15, 393–409. Shepard, A.O., 1976. Ceramics for the archaeologist, 9th ed. Carnegie Inst., Washington, (414 pp.). Shimada, I., Häusler, W., Hutzelmann, T., Wagner, U., 2003. Early pottery making in Northern Coastal Peru. Part I: Mössbauer study of clays. Hyperfine Interact. 150 (1–4), 73–89. Sholom, S.V., Haskell, E.H., Hayes, R.B., Chumak, V.V., Kenner, G.H., 1998. Influence of crushing and additive irradiation procedures on EPR dosimetry of tooth enamel. Radiat. Meas. 29 (1), 105–111. Tămăşan, M., Mocuţa, H., Ionescu, C., Simon, V., 2009. Structural changes induced in mineral clays by high temperature heat treatments. Stud. Univ. Babeş-Bolyai Phys. LIV (2), 3–8. Tokuda, Y., Seki, T., 2000. Interaction of hydrogen with the vacancy-oxygen pair produced in n-type silicon by electron irradiation. Semicond. Sci. Technol. 15, 126–129. Vandiver, P.B., Soffer, O., Klima, B., Svoboda, J., 1989. The origins of ceramic technology at Dolni Věstonice, Czechoslovakia. Science 246, 1002–1008. Wakabayashi, J., 1963. Paramagnetic resonance spectrum of Fe3+ in calcite. J. Chem. Phys. 38, 1910–1912. Warashina, T., Higashimura, T., Maeda, Y., 1981. Determination of the firing temperature of ancient pottery by means of ESR spectrometry. Occas. Pap. Br. Mus. 19, 117–128. Watanabe, S., Akutagawa, S., Sawada, K., Iwasa, T., Shimoyama, Y., 2009. A ferromagnetic resonance study of iron complexes as biologically synthesized in magnetic bacteria. Mater. Trans. 50, 2187–2191. Weiss, B.P., Sam Kim, S., Kirschvink, J.L., Kopp, R.E., Sankaran, M., Kobayashi, A., Komeili, A., 2004. Ferromagnetic resonance and low-temperature magnetic tests for biogenic magnetite. Earth Planet. Sci. Lett. 224 (1), 73–89. Whitney, D.L., Evans, B.W., 2010. Abbreviations for names of rock-forming minerals. Am. Mineral. 95, 185–187. Wu, X., Zhang, C., Goldberg, P., Cohen, P., Pan, Y., Arpin, T., Bar-Yosef, O., 2012. Early pottery at 20,000 years ago in Xianrendong Cave, China. Science 336, 1696–1700. Zavoisky, E., 1945. Spin-magnetic resonance in paramagnetics. J. Phys. 9, 245.