Effect of helium on structure and compression behavior of SiO₂ glass

Abstract

The behavior of volatiles is crucial for understanding the evolution of the Earth’s interior, hydrosphere, and atmosphere. Noble gases as neutral species can serve as probes and be used for examining gas solubility in silicate melts and structural responses to any gas inclusion. Here, we report experimental results that reveal a strong effect of helium on the intermediate range structural order of SiO₂ glass and an unusually rigid behavior of the glass. The structure factor data show that the first sharp diffraction peak position of SiO₂ glass in helium medium remains essentially the same under pressures up to 18.6 GPa, suggesting that helium may have entered in the voids in SiO₂ glass under pressure. The dissolved helium makes the SiO₂ glass much less compressible at high pressures. GeO₂ glass and SiO₂ glass with H₂ as pressure medium do not display this effect. These observations suggest that the effect of helium on the structure and compression of SiO₂ glass is unique.

SiO₂ glass is a prototypical network forming glass whose structure can be understood in terms of a continuous network of corner shared SiO₄ tetrahedra, with a high degree of intermediate range order (IRO). Numerous studies on structure and compressibility have been performed on SiO₂ glass under high pressure due to its importance as a model geological component, and an increased interest on polyamorphism of glasses and liquids (1–7). The large increase in density under pressure has been attributed to significant modification of the IRO structure, which was manifested by a drastic change in the first sharp diffraction peak (FSDP) in the structure factor S(Q) (3, 8, 9). For example, the FSDP appears at 1.55 Å^-1 for SiO₂ glass at ambient condition and shifts to 2.12 Å^-1 at 19.2 GPa (3). Several studies have suggested that the IRO structure evolution with pressure occurs through a reduction in the ring sizes and a collapse of void spaces in the network (2, 10–12). It has also been demonstrated that the modification in IRO structure is irreversible if the sample is recovered from pressures above a threshold value of 12–14 GPa, based on the observations of the permanent densification and the FSDP position not fully recovered to its original position (5, 9, 13–15).

Results and Discussion

When a SiO₂ glass is pressurized in helium pressure medium in a diamond anvil cell (DAC), however, we observed unique behavior of both structure and density evolution with pressure. Fig. 1 shows the high-pressure structure factors for SiO₂ glass, measured at the High Pressure Collaborative Access Team (HPCAT) at the Advanced Photon Source (APS). The clear difference between the data collected with and without a helium medium correlates with the FSDP. With no pressure medium (3), the FSDP displays a large shift from 1.55 to 2.12 Å^-1 as pressure increases from ambient condition to 19.2 GPa. In our data with helium loading, the FSDP shifts only slightly from 1.61 Å^-1 to 1.63 Å^-1 as pressure increases from 2.4 to 18.6 GPa (Fig. 1). This preservation of the FSDP may reflect an effect of dissolved He on the structure of SiO₂ glass and, therefore, compression mechanisms. Because SiO₂ glass contains mainly 4-, 5-, 6-, 7-, and 8-membered rings (16), which form a network of interconnected cages, or interstitial solubility sites (ISS) (17), it is possible for helium atoms to diffuse into the relatively larger ISS in SiO₂ glass. Assuming that the ISS size in 6-membered rings is similar to that of 6-membered rings in β-cristobalite silica, based on their similarities in pair distribution function up to 7.5 Å (18), the ISS size of 6-membered rings is estimated to be 0.30 nm in diameter (19), comparing to the kinetic diameter of helium atom of 0.26 nm (20). The size consideration gives a feasible mechanism for helium entering in 6- and larger ( 7- and 8-) membered rings in the glass.

Fig. 1.

The structure factors of SiO₂ glass with helium loading at (A) 2.4 GPa, (B) 7.2 GPa, (C) 9.4 GPa, (D) 11.2 GPa, (E) 16.7 GPa, and (F) 18.6 GPa (solid line), in comparison with the data from ref. 3 without a pressure medium at (A) 2.3 GPa, (B) 7.5 GPa, (C) 9.9 GPa, (D) 11.9 GPa, (E) 17.2 GPa, and (F) 19.2 GPa (dotted line). The sharp peaks are Bragg diffraction from the rhenium gasket.

As shown in Fig. 2, the dissolved helium makes the glass much less compressible than that without helium (1), indicating that the helium inclusion may have prevented the voids from contracting in the pressure range of this study (up to 48.9 GPa). It can be seen in Fig. 2 that without helium the molar volume of SiO₂ glass (1) decreases greatly with pressure, exceeding those of the fourfold coordinated crystalline polymorphs of SiO₂, quartz (21), and coesite (22), at around 20 GPa and approaching to that of stishovite (23) at approximately 40 GPa. In contrast, the volume of SiO₂ glass examined here does not exceed those of both fourfold and sixfold crystalline polymorphs of SiO₂ (21–23). Such a big difference in compressibility can only be explained by the inclusion of a large amount of helium entering in the void spaces of SiO₂ glass.

Fig. 2.

Pressure dependence of the molar volume of SiO₂ glass with helium loading in this work and without helium from ref. 1. The thick black line represents the equation of states of crystalline polymorphs of SiO₂ (quartz—ref. 21, coesite—ref. 22, stishovite—ref. 23).

Incorporation of helium is verified by the in situ Raman spectroscopy for SiO₂ glass at high pressure (Fig. 3, see also SI Text for band assignments). From the work (2) without helium, the spectral width of the Raman band at approximately 440 cm^-1 decreases significantly with increasing pressure up to 8.0 GPa (2), which implies a narrowing of the intertetrahedral angle distribution. With helium medium in this study, the reduction in band width is less pronounced.

Fig. 3.

The Raman spectroscopy of SiO₂ glass with helium loading at (A) 0.1 MPa, (B) 3.1 GPa, (C) 8.5 GPa, (D) 16.8 GPa, and (E) 0.1 MPa released from 19.4 GPa (solid line), in comparison with data from refs. 2 and 5 at (A) 0.1 MPa, (B) 3.6 GPa, (C) 8.0 GPa, (D) 16.1 GPa, (E) 0.1 MPa released from 30 GPa (dotted line) and (E) 0.1 MPa released from 18 GPa (dash dotted line).

The experiments on recovered samples further support the helium inclusion in SiO₂ glass. SiO₂ glass displays elastic compression at pressures below 12–14 GPa (5, 9, 14, 15). Above that, the modification in IRO structure becomes irreversible, resulting in a permanently densified structure with a large shift in the FSDP position accompanied with peak broadening on the recovered samples (5). In contrast, our results (Fig. 4) of a glass retrieved from 19.4 GPa with helium loading, show that the FSDP position shifts back to its original position with a similar peak width, suggesting that the IRO structure may be largely recovered from a pressure of 19.4 GPa. This recovery may be because of the dissolved helium in the network. Similar effect can be seen from the Raman data (Fig. 3) on a recovered sample from 19.4 GPa. The Raman band at 440 cm^-1 associated with the symmetric bending motion of the Si-O-Si linkage shifts to approximately 460 cm^-1 and its shape is largely recovered, together with the presence of the D1 “defect” peak at 495 cm^-1 (24, 25). The Raman data (Fig. 3) of the recovered sample do not show a complete recovery to its original shape, which may be interpreted by local angular distortions because the Raman band at approximately 440 cm^-1 is more sensitive to the Si-O-Si bond angle distribution (26).

Fig. 4.

The X-ray S(Q) at ambient pressure of the SiO₂ glass released from 19.4 GPa with helium loading, in comparison with that of the starting material SiO₂ glass at ambient condition.

The strong effect of dissolved helium may have implications in Earth’s evolution models and is also important in interpreting the high-pressure experiments in general because helium is widely used as pressure medium. To find out whether the helium diffusion in SiO₂ glass is a general phenomenon, we have tested two related systems: helium in GeO₂ glass and hydrogen in SiO₂ glass. A recent report (27) of GeO₂ glass in a helium medium showed that helium does not have clear effect on the structure and volume evolution with pressure. Fig. 5A plots the pressure dependence of the FSDP position of GeO₂ glass with helium loading (5), in comparison with the results obtained without pressure medium (28, 29). It can be seen that the use of helium medium has negligible effect on the FSDP positions, implying minimal effect on IRO structure change in GeO₂ glass. This may be explained by the relatively small sizes of ISS in GeO₂ glass. Tournour et al. (30) studied inert gas solubility in binary GeO₂-SiO₂ glasses and concluded that the solubility of helium in GeO₂ is relatively low because GeO₂ contains only a small concentration of appropriately sized interstices available for helium atoms. We note that the ISS size in 6-membered rings in α-quartz germania is 0.25 nm (31), too small for helium (0.26 nm) to fill in. It is likely that the helium can only occupy the ISS in 7- and 8-membered rings in the GeO₂ network where the concentration of the available ISS is then much reduced. Therefore, the amount of atoms diffused in the GeO₂ network may not be significant enough to interfere with the IRO structure changes under pressure.

Fig. 5.

(A). The pressure dependence of the FSDP position in the S(Q) for GeO₂ glass with helium loading, in comparison with the results obtained by Guthrie et al. (28) and Hong et al. (29) without pressure medium. (B). The pressure dependence of the FSDP position in the S(Q) for SiO₂ glass with helium loading, hydrogen loading, in comparison with the results obtained by Benmore et al. without pressure medium (7). The solid symbols represent data on compression, whereas the corresponding open symbols are taken on recovered samples after pressure release.

Fig. 5B shows the FSDP position as a function of pressure for SiO₂ glass with hydrogen and helium loadings, in comparison with those for SiO₂ glass without pressure medium (7). The FSDP position for SiO₂ glass with hydrogen loading behaves in the identical manner as that of SiO₂ glass without pressure medium within experimental errors. It has been reported that the concentration of ISS for hydrogen in silica is lower than that for helium [H₂ = 1.07 × 10²¹/cm³ (32), He = 2.3 × 10²¹/cm³ (33)], which is consistent with the larger kinetic diameter of H₂ (0.29 nm) (20) compared to that of helium atom (0.26 nm).

Our findings imply that any elements or molecules with kinetic diameters larger than that of H₂ (0.29 nm) can not enter the voids of SiO₂ glass in a large amount. Therefore, the observed helium effect on the IRO structure change and the rigid compression behavior of SiO₂ glass is unique, because of the combination of controlling factors including the small kinetic diameter of helium, the high degree of IRO with large rings in SiO₂ glass, and the large fraction of ISS occupied by helium. It should be noted, however, that the reported evidence is not a direct measure of helium inclusion in SiO₂ glass. Whereas helium diffusion may be detected at ambient pressure (32, 34), experimental techniques for samples under high pressures are still not available. Our observed helium effect on structure and compression behavior provides strong indirect evidence indicating helium inclusion in SiO₂ glass under high pressure. It remains to be checked whether there is strong helium diffusion effect in other silicate glasses and liquids.

It has been reported that there could be discontinues drops of noble gas solubility in the silicate melt at certain threshold pressures (35–37). Our results from the glasses may have implications for the melt. Helium with the smallest size among noble gases might have unique solubility in the silicate melt at high pressures. The dissolved helium may affect its structural evolution with pressure and, therefore, compression mechanisms.

Materials and Methods

A fused silica glass with 99.995 SiO₂ wt.% hydroxyl (OH-) content less than 5 ppm was used for the experiment (Chemglass, Inc). The in situ high-pressure X-ray diffraction measurements were conducted at the 16ID-B beamline at HPCAT at the APS (38), using a double-crystal monochromator with an incident beam focused to a size of 5 × 7 μm² at energy of 29.353 keV. The Raman measurements were performed with a micro-Raman spectrometer and a Kaiser spectrograph with high numerical aperture optics at the Uppsala University. Pressure was applied to the sample using diamond anvil cells with c-BN seats (39), allowing reliable diffraction patterns to be taken up to 8 Å^-1. The SiO₂ glass sample along with two ruby balls was loaded in a 165 μm diameter, 40 μm thick sample chamber in a rhenium gasket. Helium was loaded using the high-pressure gas-loading system at GeoSoilEnviro Center for Advanced Radiation Sources at the APS, whereas hydrogen gas was loaded in separate runs using the gas-loading system at Geophysical Laboratory at the Carnegie Institution of Washington. The volume of SiO₂ glass was measured as a function of pressure up to 48.9 GPa (27) (see SI Text for a description).