Impact of hydrothermal treatment of FEBEX and MX80 bentonites in water, HNO3 and Lu(NO3)3 media: Implications for radioactive waste control

1  Impact of hydrothermal treatment of FEBEX and MX80 1 bentonites in water, HNO3 and Lu(NO3)3 media: Implications 2 for radioactive waste control 3 Francisco J. Osuna1, Pablo Chain2, Agustín Cota3, 4 Esperanza Pavón1,4, María D. Alba1* 5 6 1Instituto Ciencia de los Materiales de Sevilla- Departamento de Química Inorgánica, CSIC-7 Universidad de Sevilla. Avda. Américo Vespucio, 49. 41092, Sevilla. Spain 8 2Consejería de Educación. Junta de Andalucía. 41092, Sevilla. Spain 9 3 Laboratorio de Rayos-X. Centro de Investigación, Tecnología e Innovación de la Universidad de 10 Sevilla (CITIUS), Universidad de Sevilla. Avda. Reina Mercedes, 4, 41012 Sevilla, Spain. 11 4Center for the Development of Nanoscience and Nanotechnology, CEDENNA, 9170124 Santiago, 12 Chile 13 14 ABSTRACT 15 Engineered barriers of deep geological repositories (DGR) are commonly 16 constructed with bentonite. FEBEX and MX80 bentonites have been selected by different 17 countries as reference materials for the sealing of repositories; however, their chemical 18 reactivity with high-level long-lived radioactive wastes (HLRW) under subcritical 19 conditions had not been explored before. The hydrothermal stability in neutral and acid 20 media and chemical reactivity in contact with an actinide analogous compound were both 21 studied. The long-range and short-range structural changes were analyzed by X-ray 22 diffraction, nuclear magnetic resonance and scanning electron microscopy. Both bentonites 23  *Corresponding author.Tel.: +34 954489546 E-mail address: [email protected]  2  have exhibited a good stability in neutral and acid media and have generated a new phase 24 immobilizing the actinide analogous compound. The extent of the chemical reaction is 25 higher in MX80 bentonite than in FEBEX bentonite. 26 27 Key Words—Bentonite, Engineered barrier, rare-earth, hydrothermal treatment, subcritical 28 conditions, radioactive waste. 29 30 3  1. Introduction 31 32 The complexity of radioactive waste management is one of the most significant 33 challenges faced by all countries involved in nuclear power generation. It is also a concern 34 in many other countries using radioactive materials for medical, industrial, or research 35 purposes. Radioactive wastes have to be disposed in ways that minimize their negative 36 impact. In particular, the most hazardous and long-lived waste, such as spent nuclear fuel or 37 reprocessing waste arising from the operation and dismantling of nuclear reactors, must be 38 contained and isolated in order to ensure the long-term radiological protection of humans 39 and the environment. Disposal of such waste in deep underground-geological repositories 40 has become an effective and universally preferred option (Allègre, 1999; McCombie et al., 41 2000; Savage et al. 2010). 42 A deep geological repository (DGR), designed for the long-term management of 43 high-level long-lived radioactive waste (HLRW), is a multi-barrier system consisting of 44 engineered and backfill barriers surrounded by an additional geosphere barrier (Astudillo, 45 2001; Chapman, 2006; McCombie et al., 2000). Engineered barriers are commonly 46 constructed with bentonite due to its low hydraulic conductivity and diffusivity and its high 47 cation sorption capacity and buffering properties (Landais, 2006; Meunier et al., 1998; 48 Pusch, 2006). In addition, bentonite retards the diffusion of radionuclides through sorption 49 processes, and it is able to react with radioactive leachates (Bradbury and Baeyens, 2005; 50 Coppin et al., 2002). The barrier, made of compacted bentonite (a highly swelling material), 51 will get saturated with groundwater while it is subjected to high temperatures due to the 52 radioactive decay of the wastes (Villar and Gómez-Espina 2008). 53 4  Under the conditions expected in HLRW repositories, alteration of a smectite (the 54 main mineral component of bentonites) is expected to increase its hydraulic conductivity 55 (permeability) and porosity and decrease its swelling pressure and cation exchange capacity 56 (CEC) (Bildstein et al., 2006; Carlson et al., 2007; Gaudin et al., 2009; Savage et al., 2010). 57 An alternative and effective immobilization mechanism that would overcome these 58 drawbacks could involve the chemical retention of actinides by the bentonite. Lutetium 59 disilicate (Lu2Si2O7) is formed when a nitric acid solution of lutetium reacts with pure 60 smectite under subcritical pressures and temperatures as it is expected to occur in Deep 61 Geological Repositories (DGR), even after weathering of the smectite (Alba and Chain, 62 2005; Alba et al., 2009, Perdigón, 2002). Moreover, the persistence of Lu2Si2O7 under 63 hydrothermal conditions covering a wide range of pH and different salt solutions was 64 demonstrated (Alba et al., 2009). Disilicate-phase synthesis was, therefore, proposed to be a 65 possible long-term retention solution for trivalent actinide cations. In this sense, Chain et al. 66 (2013) analyzed the structural properties of smectites that enhance their reactivity with Lu3+ 67 ions and lead to formation of the disilicate phase. However, the international HLRW 68 programs do not consider using purified smectite but raw bentonite as buffer in DGR for 69 both logistic and economic reasons. In this sense, FEBEX and MX80 bentonites were 70 selected by different countries as reference materials for the sealing of HLRW repositories 71 (Tripathy et al., 2004; Villar and Gómez-Espina, 2008), but their capacity for 72 immobilization of actinides by chemical reaction under subcritical condition need further 73 investigations. 74 Both the chemical reaction and interaction of bentonites with groundwater can, 75 however, modify their physical and chemical properties, thus decreasing their retention 76 capacity (Pusch et al., 2007). The hydrothermal stability of clay minerals in water, basic 77 5  and acidic media is of great importance, particularly where radioactive waste confinement 78 is concerned. Thus, a precise knowledge of the properties of clay minerals such as the long-79 term integrity of the bentonite when it is in contact with either acidic tailings pore water (in 80 the case of slurry wall filter cake) or acidic stored mine water (in the case of acid mine 81 drainage collection ponds), is, therefore, of crucial importance for the safety of radioactive 82 waste disposal concepts. 83 In a previous study, Villa-Alfageme et al. (2014) studied the reactivity of bentonite 84 with Eu3+. They observed that Eu3+ is immobilized in a higher amount than expected based 85 on the CEC of the bentonite and concluded that an additional chemical mechanism may 86 therefore participate to the immobilization. Since short-range study by NMR was not 87 carried out, the extension and nature of the involved chemical reaction(s) were not 88 evaluated. The main aim of this study is, therefore, to characterize the formation of 89 Lu2Si2O7 in the bentonites FEBEX and MX80 upon treatment with a Lu3+ solution under 90 subcritical conditions, with the aim to study the structural changes occurred on the 91 bentonite and quantify the disilicate phase formation, and the hydrothermal stability of the 92 two bentonites in two media in which the reaction may take place, namely water and strong 93 acid. Although Lu3+ is not the best rare earth element (REE) analogous compound for 94 trivalent actinides most commonly found in high activity radioactive waste (Np, Am, and 95 Cm) (Astudillo, 2001), it was chosen because the formation of rare-earth disilicates is 96 common to all REE, with lutetium being the most reactive (Alba et al., 2009). 97 98 2. Materials and method 99 100 2.1.Materials 101 6  102 The two selected bentonites are the Spanish bentonite FEBEX and the American 103 bentonite MX80. Those bentonites were provided by ENRESA Company (the Spanish 104 Company in charge of radioactive wastes management) and CIEMAT, Spain. 105 Bentonite FEBEX is a Ca, Mg, Na-bentonite (predominantly Ca) extracted from the 106 Cortijo de Archidona deposit (Serrata de Níjar, Almería, Spain). It contains more than 90 % 107 w/w of montmorillonite and variable quantities of quartz, plagioclase, K-feldspar, calcite 108 and cristobalite (Caballero et al., 1983, 2005; Lloret, 2004; Villar 2002). The structural 109 formula of the montmorillonite component is (Ca0.5Na0.08K0.11)(Si7.78Al0.22) 110 (Al2.78Fe3+0.33Fe2+0.02Mg0.81)O20(OH)4 and the bulk cation exchange capacity (CEC) of the 111 bentonite is 158.2 meq/100 g (Galunin et al., 2011). 112 Bentonite MX80 is a Na-bentonite from Wyoming, USA. This bentonite consists 113 mainly of montmorillonite (70–85 % w/w), but also contains quartz, feldspars and smaller 114 quantities of cristobalite and calcite (Monts-H et al., 2005). The structural formula of the 115 montmorillonite component is (Ca0.2Na0.36)(Si7.96Al0.04)(Al3.1Fe3+0.18Fe2+0.16Mg0.56)O20(OH)4 116 and the bulk CEC of the bentonite is 102.1 meq/100 g (Galunin et al., 2011). 117 118 2.2. Hydrothermal treatments 119 120 300 mg of powdered bentonite and 50 mL of 7.3·10-2 M Lu(NO3)3·3.6H2O solution, 121 distilled water or 0.01 M HNO3 solution were transferred into a stainless steel T316SS 122 hydrothermal reactor and heated to 300º C for 48 hours under autogenous pressure. 123 Although geochemical waste degradation and waste/rock interaction processes in 124 hydrothermal environments remain predictable up to temperatures of about 200 ºC, many 125 7  studies focused on simulating deep geological disposal conditions use temperatures up to 126 350 ºC to increase the reaction rates (Alba and Chaín, 2007; Allen et al., 1988; Mather et 127 al., 1982; Savage and Chapman, 1982). The reactors were then cooled to room temperature 128 and the solids separated by filtration, washed repeatedly with distilled water, and allowed to 129 dry in air at 25 ºC. 130 131 2.3. Analytical methods 132 133 X-ray powder diffraction patterns (XRD) were obtained using a Bruker D8I 134 instrument. This diffractometer is operated with a Ni-filtered Cu K radiation, at 40 kV and 135 40 mA. Diffractograms were obtained from 3 to 70° 2 at a scan step of 0.05° 2 and with 136 a counting time of 3 s. 137 Scanning electron microscopy (SEM) observations of morphology and Energy 138 Dispersive X-Ray Spectroscopy (EDX) analyses of chemical composition were performed 139 at 20 kV with a JEOL JSM 5400 microscope , equipped with a LINK Pantafet probe 140 (Oxford Link ISIS) allowing chemical analyses using a Si/Li detector with Be window. 141 29Si and 27Al solid state nuclear magnetic spectra (MAS NMR) were recorded with 142 a magnetic field of 9.36 T using a Bruker DRX400 spectrometer equipped with a 143 multinuclear probe. Powdered samples were packed in 4 mm zirconia rotors and spun at 12 144 kHz. 29Si MAS NMR spectra were obtained at a frequency of 79.49 MHz using a π/6 pulse 145 width of 2.7 μs and a delay time of 60 s. 27Al MAS NMR spectra were recorded at 104.26 146 MHz with a π/20 pulse width of 1.1 μs and delay time of 3 s. Chemical shifts are reported 147 in ppm with respect to tetramethylsilane for 29Si and a 0.1 M solution of AlCl3 for 27Al. All 148 8  spectra were deconvoluted into single Gaussian-Lorentzian bands using a modified version 149 of the Bruker Winfit program (Massiot et al., 2002). The fitted parameters were intensity, 150 position, band-width, and Gaussian/Lorentzian ratio. 151 152 3. Results and discussion 153 154 3.1. Hydrothermal stability of bentonites 155 156 Structural changes in the long-range order of both bentonites after hydrothermal 157 treatments with water and 0.01M HNO3 were analyzed by XRD (Fig. 1). The XRD patterns 158 of the raw bentonites (Fig. 1a and 1d) are consistent with the general (hk) and basal (00l) 159 reflections of a smectite, with minor reflections corresponding to impurities of quartz (PDF 160 04-007-0522), cristobalite (PDF 04-008-7742), and feldspar (PDF 04-017-1349). The 060 161 reflections are observed at a distance between the reflection planes of 0.15 nm, typical of 162 dioctahedral 2:1 phyllosilicates (Alba et al. 2001a). The 001 basal reflections corresponds 163 to a d001 value of 1.47 nm for bentonite FEBEX and 1.23 nm for bentonite MX80 and are 164 due to a water bilayer and monolayer, respectively, around the interlayer cations, mainly 165 Ca2+ in bentonite FEBEX and Na+ in bentonite MX80 (Alba et al. 2001a). 166 The positions of the hk reflections of both bentonites do not change upon 167 hydrothermal treatments (Fig. 1b, 1c, 1e and 1f), thus indicating that the basic structure of 168 the silicates remained intact and that no other crystalline phase was formed. However, their 169 widths and intensities changed after the hydrothermal treatments, especially in acidic 170 media, indicating the occurrence of structural disorder, also evidenced by the more 171 pronounced pattern background. 172 9  The position and intensity of the 00l reflection of bentonite FEBEX does not 173 change after the hydrothermal treatments in water and acidic media. The hydrothermal 174 treatments did however cause the swelling of the bentonite MX80 layers up to a d001 175 spacing of 1.42 nm in water medium and 1.52 nm in acidic medium. This swelling may be 176 due to the presence of cations with a higher hydration state. Similar results were observed 177 by Alba et al. (2001a, 2001b) after hydrothermal treatment where a leaching of divalent and 178 trivalent cations from the octahedral sheet with the replacement of Na+ in the interlayer 179 space was observed. 180 The SEM micrographies (Fig. 2a-2c and Fig. 3a-3c) evidence the remaining 181 lamellar structure of the both bentonite after the hydrothermal treatments. The untreated 182 bentonites show EDX spectra (Fig. 2d and Fig. 3d) with the emission lines characteristic of 183 the elements found in their chemical composition. The hydrothermal treatment in water 184 (Fig. 2e and Fig. 3e) does not cause any change in their chemical composition, however, 185 the acid treatment (Fig. 2f and Fig. 3f) results in a decrease of the K, Ca and Na Kα1 lines 186 due to the replacement of the initial interlayer cations by those leached from the 187 framework. Finally, this treatment provokes a relative decrease of the intensity of the Si Kα1 188 line, in agreement with the observed decrease of the XRD reflections intensities of the 189 impurities. 190 Therefore, both bentonites show a good stability under hydrothermal treatment in 191 water and acidic media and the observed structural changes are more pronounced in the 192 acidic medium. Short-range structural effect of the acid treatment was therefore analyzed 193 by MAS NMR spectroscopy. 194 29Si MAS NMR spectra of the untreated bentonites (Fig. 4a and 4c) show two set of 195 signals, the main one is in the range of -100 to -80 ppm and is due to Q3(mAl), typical 196 16  energy relationships and estimates of surface binding constants for some selected 337 heavy metals and actinides. Geochimica. Cosm. Acta 69, 875–892. 338 Caballero, E., Fernandes Porto, M.J., Linares, J., Reyes, E., 1983. Mineralogy, 339 geochemistry and mineral genesis of bentonites from Serrata de Nijar (Almeria, 340 Spain), Estudios Geológicos 39, 121–140. 341 Caballero, E., Jiménez de Cisneros, C., Huertas, F.J., Huertas, F., Pozzuoli, A., Linares, J., 342 2005. Bentonite from Cabo de Gata, Almería, Spain: a mineralogical and 343 geochemical overview. Clay Min. 40, 463–480. 344 Carlson, L., Karnland, O., Oversby, V.M., Rance, A.P., Smart, N.R., Snellma, M., 345 Vähänen, M., Werme, L.O., 2007. Experimental studies of the interactions between 346 anaerobically corroding iron and bentonite. Phys. Chem. Earth 32, 334–345. 347 Chain, P., Cota, A., El Mrabet, S., Pavón, E., Pazos, M.C., Alba, M.D., 2013. Evaluation 348 of rare earth on layered silicates under subcritical conditions: Effect of the 349 framework and interlayer space composition. Chem. Geol. 347, 208-216. 350 Chapman, N., 2006. Geological disposal of radioactive waste — concept, status and trends. 351 J. Iber. Geol. 32, 7–14.352 Coppin, F., Berger, G., Bauer, A., Castet, S., Loubet, M., 2002. Sorption of lanthanides on 353 smectite and kaolinite. Chem. Geol. 182, 57–68. 354 Engelhardt, G., Michel, D., 1987. High-resolution Solid-State NMR of Silicates and 355 Zeolites. John Wiley & Sons, New York. 356 Galunin, E., Alba, M.D., Santos, M.J., Abrao, T., Vidal, M., 2011. Examination of 357 competitive lanthanide sorption onto smectites and its significance in the 358 magnagement of radiactive waste. J. Hazar. Mater. 186, 1930-1941.359 17  Gaudin, A., Gaboreau, S., Tinseau, E., Bartier, D., Petit, S., Grauby, O., Foct, F., 360 Beaufort, D., 2009. Mineralogical reactions in the Tournemire argillite after in-situ 361 interaction with steels. App. Clay Sci. 43, 196–207. 362 Landais, P., 2006. Advances in geochemical research for the underground disposal of high-363 level, long-lived radioactive waste in a clay formation. J. Geochem. Expl. 88, 32–364 36. 365 Lloret, A., Romero, E., Villar, M.V., 2004.FEBEX II Project. Final report on thermo-366 hydro-mechanical laboratory tests. Publicación Técnica ENRESA 10/04. Madrid, pp 367 180. 368 Massiot, D., Fayon, F., Capron, M., King, I., Le Calvé, S., Alonso, B., Durand, J.O., Bujoli, 369 B., Gan, Z., Hoatson, G., 2002. Modelling one- and two-dimensional solid-state 370 NMR spectra. Mag. Reson. Chem. 40, 70-76. 371 Mather, J.D., Chapman, N.A., Black, J.H., Lintern, B.C., 1982. The geological disposal of 372 high-level radioactive waste-a review of the Institute of Geological Sciences 373 Research Programme. Nucl. Ener. 21, 167-173. 374 McCombie, C., Pentz, D.L., Kurzeme, M., Miller, I., 2000. Deep geological repositories: a 375 safe and secure solution to disposal of nuclear wastes. GeoEng2000 — An 376 International Conference on Geotechnical & Geological Engineering, Melbourne, 377 Australia. Technomic, Lancaster. 378 Meunier, A., Velde, B., Griffault, L., 1998. The reactivity of bentonites: a review. An 379 application to clay barrier stability for nuclear waste storage. Clay Clay Miner. 33, 380 187–196. 381 18  Montes-H, G., Marty, N., Fritz, B., Clement, A., Michau, N., 2005. Modelling of long-term 382 diffusion-reaction in a bentonite barrier for radiactive waste confinement. Appl. 383 Clay Sci. 30, 181-198. 384 Perdigón, A.C., 2002. Estudio del sistema saponita/Lu(NO3)3/H2O en condiciones 385 hidrotérmicas, Ph.D. thesis, University of Seville, Seville, Spain. 386 Pusch, R., 2006. Engineered barriers. In: Popov, V., Pusch, R. (Eds.), Disposal of 387 Hazardous Waste in Underground Mines. Wessex Institute of Technology, UK , pp. 388 35–40. 389 Pusch, R., Kasbohm, J., Pakovsky, J., Cechova, Z., 2007. Are all smectite clays suitable as 390 “buffers”. Phys. Chem. Earth 32, 116–122. 391 Savage, D., Chapman, N.A., 1982. Hydrothermal behaviour of simulated waste glass and 392 waste-rock interaction under repository conditions. Chem. Geol. 36, 59–86. 393 Savage, D., Watson, C., Benbow, S., Wilson, J., 2010. Modelling iron–bentonite 394 interaction. App. Clay Sci., 47, 91–98. 395 Tripathy, S., Sriharan, A., Schanz, T., 2004. Swelling pressure of compacted bentonites 396 from diffuse double layer theory. Can. Geot. J., 41, 437–450. 397 Villa-Alfageme, M., Hurtado, S., Castro, M.A., El Mrabet, S., Orta, M.M., Pazos, M.C., 398 Alba, M.D., 2014. Quantification and comparison of the reaction properties of 399 FEBEX and MX-80 clays with saponite: Europium immobiliser under subcritical 400 conditions. Appl. Clay Sci., 101, 10-15. 401 Villar, M.V., 2002. Thermo-hydro-mechanical characterisation of a bentonite from Cabo de 402 Gata. A study applied to the use of bentonite as sealing material in high level 403 radioactive waste repositories. Publicación Técnica ENRESA 01/2002, pp 258. 404 Madrid. 405 19  Villar, M.V., Gómez-Espina, R., 2008. Effect of temperature on the water retention 406 capacity of FEBEX and MX-80 bentonites. Unsaturated Solids: Advances in Geo-407 engineering, 257-262.408 Weiss, C.A., Altaner, S.P., Kirpatrick, R.J., 1987. High-resolution 29Si NMR spectroscopy 409 of 2:1 layer silicates: Correlation among chemical shift, structural distortions, and 410 chemical variations. Am. Miner. 72, 935-942. 411 412 . 413 20  414 415 416 Table 1. 29Si chemical shift, linewidth (FWHM) and area under the curve of the different contribution obtained from the fitting of the 29Si MAS NMR spectra of bentonite FEBEX and MX80: starting materials, and, after hydrothermal treatment at 300 ºC for 48 h in 0.01 M HNO3 solution. sample δ (ppm) FWWH (KHz) % Phase Bentonite FEBEX raw -88.05 0.819 18.4 Q3(1Al) -93.33 0.325 76.3 Q3(0Al) -110.34 0.661 5.3 Q4(0Al) treated in acid -93.54 0.437 100 Q3(0Al) Bentonite MX80 raw -93.80 0.758 79.7 Q3(0Al) -107.85 0.077 2.4 Q4(0Al) -109.52 0.296 17.9 Q4(0Al) treated in acid -93.41 0.536 89.0 Q3(0Al) -107.87 0.118 11.0 Q4(0Al)  21  FIGURE CAPTION 417 418 Fig. 1. XRD patterns of bentonite FEBEX (upper) and MX80 (down): (a, d) starting materials, and, 419 after hydrothermal treatment at 300 ºC for 48 h in two different chemical media (b, e) water and (c, 420 f) 0.01 M HNO3 solution. q=quartz (PDF 04-007-0522), c=cristobalite (PDF 04-008-7742), and, 421 f=feldspar (PDF 04-017-1349). 422 Fig. 2. SEM images and EDX spectra of bentonite FEBEX: (a,d) starting materials, and, after 423 hydrothermal treatment at 300 ºC for 48 h in two different chemical media (b,e) water and (c,f) 0.01 424 M HNO3 solution. 425 Fig. 3. SEM images and EDX spectra of bentonite MX80: (a,d) starting materials, and, after 426 hydrothermal treatment at 300 ºC for 48 h in two different chemical media (b,e) water and (c,f) 0.01 427 M HNO3 solution. 428 Fig. 4. 29Si and 27Al MAS NMR spectra of bentonite FEBEX (upper) and MX80 (down): (a, c) 429 starting materials, and, (b, d) after hydrothermal treatment at 300 ºC for 48 h in 0.01 M HNO3 430 solution. 431 Fig. 5. XRD patterns of bentonite FEBEX (upper) and MX80 (down): (a, c) starting materials, and, 432 (b, d) after hydrothermal treatment at 300 ºC for 48 h in a 7.3·10-2 M Lu(NO3)3 solution. q=quartz 433 (PDF 04-007-0522), c=cristobalite (PDF 04-008-7742), f=feldspar (PDF 04-017-1349), 434 cr=chrysotile (PDF 00-025-0645), k=kaolinite (PDF 01-075-0938), and, *=Lu2Si2O7 (PDF 00-035-435 0326). 436 Fig. 6. 29Si and 27Al MAS NMR spectra of bentonite FEBEX (upper) and MX80 (down): (a, c) 437 starting materials, and, (b, d) after hydrothermal treatment at 300 ºC for 48 h in a 7.3·10-2 M 438 Lu(NO3)3 solution. 439 Fig. 7. a) SEM images of bentonite FEBEX after hydrothermal treatment at 300 ºC for 48 h in a 440 7.3·10-2 M Lu(NO3)3 solution. EDX spectra of bentonite FEBEX: (b) starting materials, and, (c,d) 441 after hydrothermal treatment at 300 ºC for 48 h in a 7.3·10-2 M Lu(NO3)3 solution. 442 22  Fig. 8. (a, b) SEM images of bentonite MX80 after hydrothermal treatment at 300 ºC for 48 h in a 443 7.3·10-2 M Lu(NO3)3 solution. EDX spectra of bentonite FEBEX: (c) starting materials, (d) lamellar 444 particles of Fig. 8.a., and, (e) the ball particles of Fig. 8.b. 445 446 23  447 Fig.1448 449 450 5 10152025303540455055606570 c b [60] 2 [001] a q q q c c MX80 5 10152025303540455055606570 f e [60] f 2 f [001] d q q q q FEBEX 24  Fig.2451 00452 1234567 Energy (keV) Si Al Mg O KCa Fe Na d e f a b c 25  Fig.3453 454 1234567 Na Energy (keV) Si Al Mg O Fe d e f a b c