SEARCH

SEARCH BY CITATION

Keywords:

  • diamond anvil cell;
  • plagioclase;
  • pressure-induced amorphization;
  • Raman spectroscopy;
  • transmission electron microscopy;
  • shock metamorphism

Abstract

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Experimental Techniques
  5. 3. Results
  6. 4. Discussion
  7. Acknowledgments
  8. References

[1] High-pressure and high-temperature experiments of albitic plagioclase up to 41 GPa and 270 °C were carried out using an externally heated diamond anvil cell. Raman spectroscopy and transmission electron microscopy of the recovered samples revealed that the amorphization of albite was complete at ∼37 GPa and room temperature. The amorphization pressure at 170 °C was nearly the same as that at room temperature. In contrast, the pressure largely decreased to ∼31 GPa at 270 °C. In comparison with the amorphization pressure of albite in laboratory shock experiments, that in the present static compression experiments is significantly lower (>10 GPa) even at room temperature. This suggests that shorter pressure duration results in a lower degree of amorphization of plagioclase. The formation of maskelynite in shocked meteorites does not necessarily require the very high shock pressure (30–90 GPa) that was previously estimated on the basis of shock recovery experiments.

1. Introduction

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Experimental Techniques
  5. 3. Results
  6. 4. Discussion
  7. Acknowledgments
  8. References

[2] In heavily shocked ordinary chondrites and Martian meteorites, plagioclase in the host rock has partially or totally transformed into high-density glass (maskelynite), preserving morphological features of the previous crystalline state [Stöffler, 1972]. Maskelynite has been thought to have formed in a solid-state reaction, although there is some debate on whether the formation of maskelynite involves a solid-state reaction and/or quenching from dense plagioclase melt at high pressure, because some of the maskelynite grains have petrographic evidence for shock-induced melting, such as flow texture [Chen and El Goresy, 2000]. Maskelynite is considered to be an important indicator of the shock stages of ordinary chondrites proposed by Stöffler et al. [1991]. Existence of maskelynite is a prime indicator of the strongly shocked S5 stage. According to the shock-stage classification, chondrites in shock stages S5 and S6 containing maskelynite are considered to have experienced the shock pressure in the range of 30–90 GPa.

[3] The formation pressure of maskelynite has been estimated only on the basis of shock recovery experiments [Ostertag, 1983; Stöffler et al., 1986]. However, the shock pressure duration should be significantly different between laboratory shock and natural shock because the pressure duration in shock compression is proportional to the size of target materials. In the case of laboratory shock experiments, the target size and pressure duration are 10−3 m and 10−6 seconds, respectively. The large parent bodies of chondrites result in much longer timescales for shock pressures resulting from natural impact compared with laboratory shock experiments. In fact, several researchers claimed that the peak pressure of natural impact persists from tens of milliseconds to several seconds in terms of the ultrahigh-pressure mineralogy and cation-diffusion profiles in shock-induced veins [Ohtani et al., 2004; Beck et al., 2005; Xie and Sharp, 2007]. Therefore, if the kinetics of phase transformations are sluggish enough relative to the timescale of shock experiments, shock pressures of heavily shocked meteorites could be overestimated.

[4] Previously, static compression experiments of plagioclase were conducted using a diamond anvil cell (DAC) that can generate longer pressure duration than laboratory shock experiments, but those experiments were limited in terms of considering an anorthitic composition at room temperature [Williams and Jeanloz, 1989; Daniel et al., 1997]. Recently, Kubo et al. [2010] first conducted an in situ X-ray diffraction study of albite-rich (Ab98) and intermediate composition (Ab45) plagioclase at high pressures and temperatures using a Kawai-type multianvil apparatus up to 29 GPa and 1300 °C. They pointed out that the pressures of complete amorphization of both plagioclases decreases with increasing temperature significantly, although the data for the amorphization of albitic plagioclase are limited to the temperature range of 900 °C–1100 °C.

[5] In the present study, we present the results of static compression experiments of albitic plagioclase under moderately high temperature using an externally heated DAC with pressure duration of tens of minutes. We discuss the temperature and time dependence of the amorphization pressure of albitic plagioclase and its mechanism of amorphization on the basis of the Raman spectroscopy and transmission electron microscopy (TEM) of recovered specimens of DAC experiments.

2. Experimental Techniques

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Experimental Techniques
  5. 3. Results
  6. 4. Discussion
  7. Acknowledgments
  8. References

[6] In this study, natural albitic plagioclase crystals (Ab98.8An0.1 Or1.1) were crushed into fine-grains ∼5–30 μm in size. The plagioclase powder was directly put into a 100 μm hole of a rhenium gasket and uni-axially compressed to desired pressures. Applied pressures were measured from a peak shift of the R1 fluorescent line of ruby chips placed in the gasket hole using the pressure scale of Mao et al. [1978]. The pressure gradient in the sample was less than ∼0.05 GPa/μm.

[7] After compression, moderately high temperatures up to 270 °C were applied by a nichrome wire heater surrounding the outer cylinder of the DAC. The samples were kept at elevated pressures and temperatures for around 30 minutes. The temperature dependence of the shift of the R1 line was considered in pressure measurements at high temperature [Ragan et al., 1992]. Pressures varied during heating/cooling and pressures after cooling to room temperature were the highest in the temperature paths in all the runs. Therefore, pressures measured after cooling are regarded as the run pressures in the present study. The pressure ranges were at 20–39 GPa, 21–41 GPa and 24–31 GPa at room temperature, 170 °C and 270 °C, respectively. After heating, the samples were cooled to room temperature and then decompressed to ambient pressure.

[8] Recovered samples were initially examined using a laser Raman spectrometer equipped with an Argon laser with wavelength of 488 nm at the Institute for Study of the Earth's Interior. The portion near the center of the gasket hole was measured to minimize the effect of the pressure gradient as much as possible. The samples compressed at 27, 33 and 37 GPa at room temperature were examined under a JEOL JEM-2010 transmission electron microscope equipped with a Noran System SIX energy dispersive spectrometer system (EDS) at Kobe University. To avoid vitrification during the TEM-sample preparation, we adopted the crushing method rather than the conventional Ar-ion-milling technique. The samples were removed from the gasket and gently crushed with a tungsten needle. The isolated grains were then mounted on Cu grids, each of which was covered with a microgrid and a formvar/carbon film. Qualitative EDS analyses were carried out to differentiate contaminant particles from the samples. Mineral phases of the sample particles were identified on the basis of selected area electron diffraction.

3. Results

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Experimental Techniques
  5. 3. Results
  6. 4. Discussion
  7. Acknowledgments
  8. References

[9] In the sample compressed at 20 GPa and room temperature, sharp Raman peaks of plagioclase mainly at 291, 329, 455, 479, 508, 763, 814, 977 and 1098 cm−1 were observed (Figure 1a). These peaks weakened with increasing pressure. At 32 GPa, broad Raman peaks appeared at ∼490, ∼590, ∼780 and ∼1080 cm−1 in addition to those of plagioclase. The profile of broad peaks is similar to that of maskelynite in shocked meteorites [Chen and El Goresy, 2000; Fritz et al., 2005] and synthetic plagioclase glass produced in shock experiments [Velde et al., 1989]. At 39 GPa, Raman peaks of plagioclase completely disappeared. The broad peaks of plagioclase glass overlapping peaks of crystalline plagioclase make phase identification difficult. Since the Raman peak at 291 cm−1 does not overlap the glass peaks and is not broadened by comparison with increasing pressure, the disappearance of this peak is used to determine when the samples have undergone complete amorphization. In this paper, the pressure for complete amorphization is termed “amorphization pressure”.

image

Figure 1. (a) Raman spectra of the recovered samples of albitic (Ab99) plagioclase compressed by a diamond anvil cell at room temperature. Broad peaks indicated by filled circles are of plagioclase glass (maskelynite). Asterisks and crosses denote unassigned peaks and grating ghosts from the fluorescent lines of the ruby chip, respectively. Arrows indicate diagnostic peaks for the presence of plagioclase. (b) Raman spectra of the recovered samples of albitic (Ab99) plagioclase compressed by a diamond anvil cell at 170 °C. Note that the sample pressures were measured after cooling, and the Raman spectra were taken at room temperature and ambient pressure (see text).

Download figure to PowerPoint

[10] In the TEM observation of the sample compressed at 27 GPa, electron diffraction indicates that all of the observed grains are crystalline (Figure 2a). At 33 GPa, some of grains showed powder diffraction rings consistent with extremely fine grains, although it is difficult to recognize respective grains of several tens of nanometers in size owing to the large thicknesses of the samples produced by the crushing method (Figure 2b). However, most grains in this sample showed diffused scattering in electron diffraction that suggests an amorphous nature (Figure 2c). In the sample compressed at 37 GPa, electron diffraction patterns are consistent with all grains being amorphous (Figure 2d). Planar amorphous lamellae, which were reported in the single crystal of experimentally shocked anorthite [Kitamura et al., 1977], were not observed in any sample.

image

Figure 2. Transmission electron micrographs of the albite samples compressed at room temperature. (a) Sample compressed at 27 GPa showing a diffraction pattern of feldspar, (b) sample compressed at 33 GPa showing powder diffraction rings of polycrystalline plagioclase, (c) sample compressed at 33 GPa showing a diffused diffraction pattern, (d) sample compressed at 37 GPa showing a diffused diffraction pattern. Electron diffraction patterns from the circled areas are shown in the insets.

Download figure to PowerPoint

[11] While compressing plagioclase at 170 °C, broad Raman peaks appeared in the sample compressed at 27 GPa and amorphization was complete at 38 GPa (Figure 1b). In contrast, at 270 °C, all the plagioclase peaks disappeared in the sample at 31 GPa. High-pressure phases of plagioclase, such as the calcium ferrite-phase (NaAlSiO4) + stishovite (SiO2) pair and the jadeite (NaAlSi2O6) + stishovite pair and lingunite (NaAlSi3O8), were not observed in Raman or TEM measurements of any of the recovered samples.

4. Discussion

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Experimental Techniques
  5. 3. Results
  6. 4. Discussion
  7. Acknowledgments
  8. References

[12] The compression experiments of albitic plagioclase in the present study clarified that the amorphization of Ab99 plagioclase starts between 26 and 32 GPa and is complete at ∼37 GPa and room temperature (Figure 3). For compression with heating at 170 °C, there is no obvious difference in amorphization pressure. However, the amorphization pressure at 270 °C dropped significantly to ∼31 GPa. Combining the data of the in situ X-ray diffraction study of Ab98 plagioclase [Kubo et al., 2010] and those of the present study, we find a clear trend where the amorphization pressure decreases with increasing temperature (Figure 3). This is consistent with the results of the in situ X-ray diffraction study of Ab45 plagioclase by Kubo et al. [2010]. Meanwhile, in shock recovery experiments on preheated plagioclase (Ab∼83), Huffman et al. [1993] clarified that the amorphization pressure decreased by ∼5 GPa at a preheating temperature of 800 °C. These results suggest that the amorphization of plagioclase is temperature dependent.

image

Figure 3. Amorphization pressures of albitic plagioclase obtained in static high-pressure experiments as functions of temperature. Open squares, circles, and diamonds represent Raman measurements of plagioclase, plagioclase + maskelynite and maskelynite, respectively. Black and gray symbols are for the same phases indentified by electron diffraction patterns in the present study and those indentified in the in situ X-ray diffraction study of Kubo et al. [2010], respectively.

Download figure to PowerPoint

[13] Under static compression, amorphous phases in oxides and silicates have been regarded to be metastable phases at sufficiently low temperatures where the crystallization of their high-pressure phases is inhibited kinetically. Pressure-induced amorphization was previously reported in a variety of materials such as quartz [Hemley et al., 1988], olivine [Andrault et al., 1995], germanium oxide [Wolf et al., 1992; Yamanaka et al., 1992] and magnesium germanate [Nagai et al., 1994] in DAC experiments. It was pointed out that a metastable extension of the melting curve is possibly identical to the negative slope of the amorphization boundary [Hemley et al., 1988; Kubo et al., 2010]. However, if maskelynite forms in this way, plagioclase and maskelynite should show a clear phase boundary. Nevertheless, plagioclase vitrified gradually with increasing pressure in an in situ X-ray study [Kubo et al., 2010], shock experiments [Velde et al., 1989; Ostertag, 1983] and DAC studies [Williams and Jeanloz, 1989; Daniel et al., 1997; this study]. This suggests that amorphization would be activated when plagioclase is pressurized above the metastable melting curve, but the amorphization would be kinetically too sluggish to be completed near the phase boundary on the experimental timescale. The metastable melting curve may correspond to a lower pressure bound for the kinetic phase boundary. Alternatively, the amorphous plagioclase may be formed as a stable phase at low temperatures as being suggested by the positive melting slope in SnI4 which refutes the pseudo-melting hypothesis [Grocholski et al., 2010].

[14] The amorphization pressure of albitic (Ab99) plagioclase was earlier investigated in dynamic shock compression experiments and the Raman spectroscopy of recovered samples [Velde et al., 1989]. After shock compression to 50 GPa, albite crystal partly, but not completely, transforms into maskelynite. Therefore, it was assumed that the pressure of complete amorphization exceeds 50 GPa. Since shock compression is a specific adiabatic process associated with increase of entropy, the sample temperature increases with increasing pressure. The estimated shock temperature and post-shock temperature at 50 GPa are higher than 1800 °C and 400 °C respectively on the basis of the Hugoniot calculation of plagioclase [Ahrens et al., 1969]. Such elevated temperatures in shock experiments are likely to activate pressure-induce transformations. Nevertheless, the amorphization pressure in the shock experiments is significantly higher (>10 GPa) than those in the present DAC experiments, in spite of the shock temperature being higher than the temperature in the present experiments. Therefore, the kinetic effect on the amorphization of plagioclase is certainly not negligible and is one of the main factors of the different amorphization pressures in the two experiments.

[15] Previously, Kitamura et al. [1977] demonstrated shock recovery experiments for anorthitic plagioclase. They reported lamellar diaplectic glass in the (001), (010), (−1–21) and (−111) planes of single-crystal anorthite shocked at 30 GPa. However, albitic plagioclase in the present study did not show such lamellar texture but showed the aggregates of randomly oriented fine grains (Figure 2b). The different mode of amorphization may be due to the difference in the initial grains sizes in the two experiments. The amorphization of relatively small grains possibly proceeds owing to a size reduction of the crystalline domains as reported in the amorphization of GeO2 in a DAC [Wolf et al., 1992] rather than by planer deformation associated with lamellar amorphization in millimeter-scale quartz crystals in shock recovery experiments [Tattevin et al., 1990; Langenhorst, 1994], although further investigation on the effect of grain size on the amorphization mechanism is necessary.

[16] The results of the present study suggest that shorter pressure duration results in a lower degree of amorphization in plagioclase. The formation of maskelynite in shocked meteorites does not necessarily require the very high shock pressure (30–90 GPa) that was previously estimated in shock recovery experiments [Stöffler et al., 1991]. Further detailed kinetic studies would enable the estimation of amorphization pressures of natural maskelynite through the interpolation of pressure versus timescale data between static high-pressure experiments (100–103 s) and natural impacts (10−2–100 s) also taking temperature and compositional effects into account. This is of particular importance as mineralogical investigations of shocked chondrites, classified as shock stages S5 and S6, contain the assemblages of ultrahigh-pressure minerals including wadsleyite, ringwoodite, akimotoite, majorite, lingunite and MgSiO3-perovskite, suggesting shock pressures of ∼13–26 GPa [e.g., Xie et al., 2006; Ozawa et al., 2009] on the basis of their stability conditions. The shock pressure estimation considering ultra-high pressure minerals is more plausible than that from shock experiments. Hence, the shock pressure calibration, especially for high shock stages, needs further reevaluation.

Acknowledgments

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Experimental Techniques
  5. 3. Results
  6. 4. Discussion
  7. Acknowledgments
  8. References

[17] We thank T. Yamanaka, T. Narita, and Y. Seto for their technical assistance on the DAC experiments. M. Kanzaki generously assisted with the laser-Raman spectroscopy. We also thank A. El Goresy for his valuable discussion and anonymous reviewers for their careful and critical reviews. This work was supported by a grant-in-aid for scientific research (20540467) to N. Tomioka.

References

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Experimental Techniques
  5. 3. Results
  6. 4. Discussion
  7. Acknowledgments
  8. References