The disordering effects on the crystalline structure of feldspars using both static (diamond anvil cells) and kinetic (shock recovery) high pressure experiments have been studied using dominantly infrared absorbance and Raman spectroscopy combined with x-ray diffraction [e.g., Velde and Boyer, 1985; Velde et al., 1987, 1989; Heymann and Hörz, 1990; Daniel et al., 1995, 1997; Reynard et al., 1999]. Disordering of feldspar begins generally at pressures above 15–20 GPa and progressively increases until a dominantly amorphous state is reached by about 30 GPa [e.g., Stöffler and Hornemann, 1972; Arndt et al., 1982; Ostertag, 1983; Langenhorst, 1989]. Within this pressure region the characteristic, fourfold (tetrahedral) strong coordination bonds of silicon and aluminum in feldspars distort to weaker, less polymerized bonds that approach sixfold (octahedral) coordination. The increased structural disorder and density results in the mutual existence of crystalline phases and diaplectic glasses throughout this pressure region which provide characteristic vibrational frequencies in the thermal infrared. Absorptions in the 900–1200 cm−1 region are due to Si-O antisymmetric stretch motions of the silica tetrahedral units in the structure. SiO6 octahedral stretching vibrations occur between 750–850 cm−1 whereas Si-O-Si octahedral bending vibrations cause several smaller absorptions between about 700–450 cm−1. Between 400–550 cm−1 bending vibrations in the Si-Al-O planar ring structures in tectosilicates and diaplectic glasses occur [Bunch et al., 1967; Iiishi et al., 1971; Stöffler and Hornemann, 1972; Arndt et al., 1982; Velde et al., 1987; Williams and Jeanloz, 1988, 1989; Daniel et al., 1995, 1997; Williams, 1998].
 In shocked samples crystalline and amorphous phases likely coexist as intimate mixtures with the proportion of diaplectic glass increasing with shock pressure [Ostertag, 1983; Heymann and Hörz, 1990; Yamaguchi and Sekine, 2000]. The spectral features of anorthosite that change with increasing peak shock pressure are summarized in Table 2. The disappearance of the small bands between 500–650 cm−1 by 25.5 GPa is due to the depolymerization of the silica tetrahedra [e.g., Williams, 1998]. This behavior is similar to that observed in transmission spectra of diaplectic laboradorite glass from Manicouagan crater by Arndt et al.  and in absorption spectra of shocked feldspars by Ostertag , as is the appearance of the band near 450 cm−1 with increasing pressure (Figure 1). For the stronger Si-O stretching bands, the higher-wave number band (1115 cm−1) disappears by 38.2 GPa, and the lower-wave number band (940 cm−1) shifts to higher wave numbers with pressure. These effects are similar to the feldspar observations of Arndt et al.  and Stöffler and Hornemann . The shift in the Christiansen feature near 1250 cm−1 to lower wave numbers with increasing pressure in the chip samples (Figure 4) may be a side-effect of the loss of the 1115 cm−1 stretching band. The change in CF position is not observed in the powder sample spectra (Figure 6), consistent with results of Nash and Salisbury  for biconical reflectance spectra of crystalline and fused powders of the same plagioclase composition.
Table 2. Summary of Observed Changes in Unshocked Spectral Feature Positions With Increasing Shock Pressure
|Sample||Feature Position, cm−1||Change With Increasing Pressure|
|Anorthosite (S2-104)||538, 590, 630, 1115||band depth decreases|
|450||band depth increases|
|940||band shifts to higher cm−1|
|1250||position shifts to lower cm−1|
|830 (powder)||band depth decreases|
|Orthopyroxenite (S-77)||976, 567||band depth decreases|
 The transparency feature near 830 cm−1 in powdered samples (Figure 6) results from their fine grain size, which reduces the spectral contrast of the reststrahlen bands and allows volume scattering to dominate [e.g., Salisbury et al., 1991a, 1991b]. The transparency feature disappears at high pressures, similar to the fused glass spectra of anorthite presented by Nash and Salisbury . This likely occurs because of the structural disorder in the highly shocked samples, which prevents significant volume scattering.
 Although transparency features can be difficult to detect on airless planetary surfaces due to severe temperature gradients [e.g., Salisbury et al., 1991a, 1997], Henderson and Jakosky  suggested that the atmospheric pressure on Mars is sufficient to minimize these effects. Indeed, although TES spectra of high albedo regions are dominantly blackbody-like (particularly in the reststrahlen region) [e.g., Christensen et al., 2000c], ongoing analyses of some spectra reveal an absorption near 825 cm−1, which is interpreted as a transparency feature associated with feldspar [e.g., Ruff and Christensen, 2002]. If this feature is absent in some high albedo regions, one explanation could be the presence of highly shocked, fine-grained feldspars. In support of this concept, we note that the similarity of impact melt compositions at Meteor Crater with Martian surface materials suggested to Hörz et al.  that shock-produced glasses may be suitable analogs to some Martian surface materials. Schultz and Mustard  came to similar conclusions upon comparison of visible/near-infrared spectra from Mars Pathfinder to spectra of glasses produced from impacts into terrestrial loess.
 Feldspar minerals with compositions other than the bytownite investigated here likely experience similar effects in their thermal infrared spectra at high shock pressures. Although differences in spectral detail of unshocked feldspars exist and are well documented [e.g., Iiishi et al., 1971; Nash and Salisbury, 1991], Ostertag  showed that changes in the thermal infrared transmission spectra of shocked feldspar crystals are very similar no matter the spectral shape of the unshocked samples. While the precise peak shock pressures at which structural disorder and melting occur may vary among feldspar compositions [e.g., Williams, 1998], the stretching and bending modes associated with Si-O, Si-O-Si, and Si-Al-O bonds generally weaken with increasing pressure and cause shifts in band positions similar to those observed here for bytownite. Emission and reflectance spectra of an experimentally shocked albite-rich rock are being acquired to test these observations and will be reported in a subsequent paper.
 In contrast to the multiple features in anorthosite spectra that provide sensitive barometers to shock pressure, the orthopyroxenite studied here shows little change in spectral properties with increasing pressure (Figures 8 and 9). The two minor bands at 976 cm−1 and 567 cm−1 degrade with increasing shock pressure, but the stronger bands maintain their relative depths and positions at high shock pressures. This is consistent with previous observations of shocked pyroxenes that demonstrated their resiliency and incompressibility to high pressures [e.g., Ahrens and Gaffney, 1971; Dundon and Hafner, 1971; Stöffler et al., 1991; Leroux et al., 1994; Schmitt and Deutsch, 1995].
 The disappearance of the small bands observed in the spectra here may result from minor amounts of melting similar to that observed in petrologic studies of samples shocked to pressures greater than about 45 GPa. Our spectra show that such melting apparently is insufficient to alter significantly the major stretching and bending vibration bands, at least to the maximum pressure of 62.5 GPa in the orthopyroxenite studied here. Indeed, Estep et al.  required peak shock pressures of 100 GPa to observe a reduction in spectral detail in their transmission spectra.
 Although the only pyroxene composition we examined was bronzite, other pyroxenes probably also exhibit limited spectral changes over the range of shock pressures investigated here, as evidenced by numerous studies of both orthopyroxenes and clinopyroxenes which suggest that mechanical defects and fractures dominantly occur at these pressures [e.g., Lambert, 1982; Sazonova et al., 1996; Kotelnikov and Feldman, 1998].