Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.issp.ac.ru/lhpp/PapersEfimchenko/9.pdf Äàòà èçìåíåíèÿ: Fri Mar 15 17:49:47 2013 Äàòà èíäåêñèðîâàíèÿ: Fri Feb 28 03:04:45 2014 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: molecular cloud

Article pubs.acs.org/JPCB

Hydrogen Solubility in Amorphous Silica at Pressures up to 75 kbar
Vadim S. Efimchenko,*, Vladimir K. Fedotov, Mikhail A. Kuzovnikov, Andrey S. Zhuravlev, and Boris M. Bulychev




Institute of Solid State Physics RAS, 142432 Chernogolovka, Moscow District, Russia Moscow State University, Vorob'evy gory, Moscow 119992, Russia ABSTRACT: The solubility of hydrogen in amorphous silica at a temperature of 250 °C and pressures up to 75 kbar is studied using a quenching technique. The molar ratio H2/SiO2 is found to linearly increase with pressure from X = 0.16 at P = 6 kbar to X = 0.53 at P = 75 kbar. An investigation of a sample with X = 0.47 by Raman spectroscopy demonstrated that hydrogen dissolves in silica in the form of H2 molecules, and these molecules occupy voids of, at least, two different types in the silica network. An X-ray diffraction study showed that the hydrogen molecules penetrating in the silica glass network prevented its irreversible densification occurring if the silica glass is compressed to the same pressure without hydrogen.

INTRODUCTION According to the most widely accepted model,1,2 all planets of the solar system were formed and evolved from a presolar nebula containing silica dust and molecular hydrogen. In the course of formation of the planetesimals, the hydrogen and other nebula gases could be trapped by the silica and further play a key role in forming the terrestrial and giant planets.1,2 In all cases, the SiO2-H2 phases, if formed, could only exist at high pressures because the hydrogen solubility in silica is very low under ambient conditions. Despite the importance of the problem for geosciences and planetary science, very little is yet known about the high-pressure H2 solubility in any of the crystalline and amorphous modifications of silica. One of the first investigations of the SiO2-H2 system at high pressure was performed by Hartwig in 1975.3 Using Raman scattering, he determined the pressure dependence of the hydrogen and deuterium solubility in the vitreous silica at a temperature of 90 °C and pressures up to 850 bar. In 1977, Shelby4 measured the hydrogen and deuterium solubility in different types of vitreous silica at pressures up to 850 bar and temperatures 0-800 °C using a volumetric technique. He also offered a model for calculating the solubility as a function of pressure and temperature. The experimental values of H2/SiO2 = 0.03 at T = 90 °C and the maximum pressure of P = 850 bar were the same in both papers.3,4 FTIR studies of the reaction of silica with water at high pressures and temperatures (P = 10-30 kbar, T = 1100-1250 °C) by Tomozawa et al.5 detected observable quantities of hydroxyl and hydride dissolved in the silica glass. The authors conjectured that the dissolved hydrogen was produced in the chemical reaction of water and graphite in the measuring cell and further diffused in the silica. Regretfully, the quantity of the dissolved hydrogen was not determined.
© 2012 American Chemical Society



Recently, a high helium solubility in silica glass was discovered by two research groups at pressures up to 90 kbar6 and 500 kbar.7 The solubility of helium at P = 90 kbar was estimated to be as high as He/SiO2 = 1 to 2.6 Taking into account that the effective diameters of the helium atom and hydrogen molecule do not differ much (their so-called kinetic diameters are 2.6 and 2.9 å,8 respectively), one could expect that the hydrogen solubility in amorphous silica at 75 kbar should also be high. In the present work, an isotherm of the hydrogen solubility in amorphous silica at a temperature of 250 °C and hydrogen pressures from 6 to 75 kbar is constructed using a quenching technique. Each samples of silica glass loaded with hydrogen at a given pressure was then rapidly cooled (quenched) to the N2 boiling temperature and only warmed above this temperature when its hydrogen content was measured by hot extraction in vacuum. The quenched samples were studied by X-ray diffraction and Raman spectroscopy at ambient pressure and N2 temperature.



EXPERIMENTAL METHODS Powder of amorphous silica with a grain size of 5-10 m was purchased from Sigma-Aldrich. The powder was preliminary annealed at 900 °C for 4 h to remove hydroxyls and water. Batches of this silica powder weighing about 100 mg each were exposed to an atmosphere of gaseous hydrogen at P = 6-75 kbar and T = 250 °C for 30 min. (Preliminary experiments at 75 kbar showed that an increase in the exposure time from 30 min to 1 and 2 h did not increase the hydrogen content of the
Received: October 9, 2012 Revised: December 5, 2012 Published: December 7, 2012
422
dx.doi.org/10.1021/jp309991x | J. Phys. Chem. B 2013, 117, 422-425


The Journal of Physical Chemistry B samples.) The hydrogen was produced by thermal decomposition of aminoborane (NH3BH3) used as an internal H2 source in the high-pressure cell. After the hydrogenation was complete, the batch was cooled to -196 °C to prevent the hydrogen losses in the course of further pressure release. When not in use, the samples thus prepared were stored in liquid nitrogen. The hydrogen content of the SiO2-H2 samples was determined by hot extraction into a pre-evacuated volume in the course of heating to 500 °C at a rate of 20 °C/min. Raman scattering studies of the initial silica and the SiO2-H2 sample with an H2/SiO2 ratio of X = 0.47 were performed with an EnSpectr Raman microscope. The microscope collects the spectra in a wide range of Raman shifts from 485 to 4770 cm-1 with a resolution of 6 cm-1, and therefore it is a convenient tool for detecting all known molecules, with a high vibration frequency, which can be dissolved in silica. The light source was a DPSS laser operated at 532 nm. The beam with a nominal power of 8 mW was polarized and directed normal to the face of the sample submerged in liquid nitrogen. The scattered light was collected with the same lens in backscattering geometry and resolved by an EnSpectr 532 spectrometer (EnSpectr microscope). The obtained SiO2-H2 samples were also studied by powder X-ray diffraction at ambient pressure and 80 K using a Siemens D500 diffractometer equipped with a home-designed nitrogen cryostat that permitted loading the powder samples without their intermediate warming.

Article

hydrogen concentration. Such an effect is typical of all metal hydrides and hydrogen clathrates.9,10 The open circles in Figure 2 show the hydrogen content X of the quenched SiO2-H2 samples vs the pressure of synthesis. As

Figure 2. Pressure dependence of the hydrogen and helium solubility in silica glass. The red solid line fits the experimental points of hydrogen solubility (open circles) at a temperature of 250 °C. The solid black circle indicates the hydrogen solubility at 90 °C determined in ref 3. The solid blue lines show the upper and low boundaries of the helium solubility in silica glass estimated in ref 6.

RESULTS AND DISCUSSION Hydrogen Solubility in Amorphous Silica. Figure 1 depicts thermal desorption curves of SiO2-H2 samples



Figure 1. Thermal desorption curves of SiO2-H2 samples quenched from different synthesis pressures. The curves a labeled by the pressures of the sample synthesis.

quenched from different synthesis pressures. The samples synthesized at minimal pressures (6 and 14 kbar) start evolving hydrogen at T = -174 °C. The desorption curves of these samples have a stepwise shape demonstrating the most intense hydrogen release at temperature from T = -116 to -83 °C to +25 °C. The hydrogen release from the samples synthesized at maximal pressures (56 and 75 kbar) begins on heating above T = -187 °C and also ceases at room temperature. In whole, the curves in Figure 1 demonstrate a gradual decrease in the thermal stability of the SiO2-H2 solid solutions with increasing
423

seen from Figure 2, the hydrogen content of the samples approximately linearly increases with pressure from X = 0.16 at P = 6 kbar to X = 0.53 at P = 75 kbar. To meet the requirement that X = 0 at P = 0, the red solid straight line fitting the experimental dependence at pressures from 6 to 75 kbar is smoothly turned downward at lower pressures. The solubility X = 0.05 at a pressure of 0.85 kbar read from this line is consistent with the maximal molar ratio X = 0.03 at P = 0.85 kbar and T = 90 °C obtained by Hartwig3 and shown by the solid circle. The solid blue curves in Figure 2 represent the pressure dependences of the upper and lower limits on the helium solubility in silica glass estimated in ref 6 (the solubility was not measured experimentally). As one can see, the He/SiO2 ratio can reach as much as 0.8 to 2.2 at a helium pressure of 75 kbar. The conventional model of void-size distribution11 predicts that in silica glass there are only about 0.1 voids per SiO2 formula unit accessible for He atoms with the kinetic diameter of 2.6 å .12 According to ref 6, a much larger number of He atoms can be "...continuously stacked in three-dimensionally connected voids, because voids in SiO2 glass may not be isolated...". The kinetic diameter of a H2 molecule equals 2.9 å .8 It is larger than the diameter of a He atom, and the hydrogen molecules can occupy a smaller number of about 0.05 voids per SiO2 unit (these are the voids inside the six-membered rings of silica).4 Similar to the He atoms, most H2 molecules dissolved in silica glass under high pressure should therefore occupy the "connected voids" proposed in ref 6. This suggests that the H2 molecules in our SiO2-H2 samples should have occupied the sites of, at least, two different types. Raman Spectroscopy. A Raman spectrum collected from the sample synthesized at P = 56 kbar and T = 250 °C and having a molar ratio of H2/SiO2 = 0.47 is shown in Figure 3. In the range of Raman shifts below 4000 cm-1, this spectrum virtually coincides with the spectrum of the initial silica glass: · The peak at 600 cm-1 labeled D2 is due to the breathing modes of the three-membered SiO2 rings;
dx.doi.org/10.1021/jp309991x | J. Phys. Chem. B 2013, 117, 422-425


The Journal of Physical Chemistry B

Article

Figure 3. Raman spectra of the SiO2-H2 sample synthesized at P = 56 kbar and T = 250 °C.

other vibrational bands of SiO2 are not seen because of the strong background luminescence arising at the low measuring temperature of -196 °C. · The strong peak at 2329 cm-1 is due to the stretching modes of N2 molecules in the 50 m layer of liquid nitrogen adjacent to the sample surface and located inside the focal spot of the laser beam. · The weak peak at 3442 cm-1 results from the OH stretching vibration modes. Such a peak was also observed in the Raman spectra of the initial SiO2 powder. The presence of this peak in the Raman spectra of the hydrogenated sample is therefore due to the incomplete removal of the hydroxyls and water from the initial silica glass by its preliminary annealing at 650 °Cin air. The peaks at 4168 and 4374 cm-1 are only intrinsic to the spectrum of the hydrogenated silica. The peak at 4168 cm-1 is close to the Q1(0) peak at 4161 cm-1 of the stretching modes of H2 molecules in a hydrogen gas.13 A peak at a slightly lower frequency of 4142 cm-1 was earlier observed in the Raman spectrum of hydrogenated amorphous silica with H2/SiO2 = 0.03, and it was ascribed to the vibrational modes of H2 molecules3 occupying voids in the silica glass network and bind only weakly to the silica. At the same time, as was noticed in ref 3, the weak interactions of the H2 molecules with the walls of the voids restricted the rotational freedom of the molecules, so that no hydrogen rotational modes were seen in the Raman spectra. No peaks that could be attributed to H2 rotational modes are seen in the spectrum of our SiO2-H2 sample with a much higher hydrogen content as well (Figure 3). The strongest peak in Figure 3 is the peak at 4374 cm-1. Such a peak was not present in the Raman spectra of the samples of hydrogenated silica glass with H2/SiO2 0.03 studied earlier,3 and its frequency much exceeds the highest observed stretching frequency 4280 cm-1 of the hydrogen molecule in solid hydrogen compressed to a pressure of 350 kbar at T = 5 K.14 On the other hand, substances with resonance vibrations at 4374 cm-1 are not known; therefore, we can only conjecture that the 4374 cm-1 peak is due to the stretching vibrations of the H2 molecules occupying the voids other than those occupied by the molecules vibrating at 4168 cm-1. Presumably, these are the "connected voids" proposed in ref 6 in order to explain the large solubility of He in amorphous silica. The locations and shapes of the "connected voids" are
424

not yet identified, but they should obviously be smaller, at least, in some directions, than the largest individual voids that can only give room to the H2 molecules according to the standard model.4 Consequently, the H2 molecules stacked in the "connected voids" should interact stronger with their walls that manifests itself in the increase in the H2 vibrational frequency. On the basis of the said above, it is reasonable to assume that the peak at 4168 cm-1 in Figure 3 corresponds to Xv 0.05 of the H2 molecules occupying the largest individual voids (mostly those inside the six-membered rings of silica),4 while the peak at 4374 cm-1 stems from the remaining Xcv molecules located in the "connected voids". From the intensity ratio I4374/I4168 8 of these peaks, one can roughly estimate that Xcv 8Xv 0.4. The value of Xv + Xcv 0.45 thus estimated agrees with the experimental value of the total hydrogen content X = 0.47(3) of the SiO2-H2 sample studied by Raman spectroscopy. As seen from Figure 3, the width of the 4168 and 4374 cm-1 peaks is much less than the distance between them. This suggests that most voids in amorphous silica incorporating H2 molecules under high pressure can roughly be divided into two different groups, and the scatter in the shapes, sizes, and other determinative properties of the voids within each group is much smaller than the difference in these properties of the voids from different groups. X-ray Diffraction. Figure 4 compares the structure factors S(Q) of the initial silica glass, the silica glass loaded with

Figure 4. Structure factor, S(Q), of the SiO2-H2 sample synthesized at P = 75 kbar and T = 250 °C (red solid line) in comparison with S(Q) of the sample of silica glass exposed to the same pressure and temperature without hydrogen (blue dashed line) and the initial silica glass (black dotted line). Cu K radiation, ambient pressure, T =85K.

hydrogen to X = 0.53 at a hydrogen pressure of 75 kbar and T = 250 °C, and also the silica glass exposed to the same pressure and temperature without hydrogen (in Teflon as the pressure transmitting medium) and quenched to the N2 temperature. As seen from Figure 4, the positions of the first sharp diffraction peak (FSDP) for the initial and hydrogenated samples coincide at Q = 1.55 å-1. The FSDP position of the sample exposed to P = 75 kbar and T = 250 °C without hydrogen is Q = 1.78 å-1. The observed shift of the FSDP by Q = 0.23 å-1 in the silica glass subjected to high pressure stems from the pressureinduced reduction of the ring sizes and a collapse of void spaces in the silica network.15 The absence of such a shift in the sample exposed to the same pressure and temperature in a hydrogen atmosphere suggests that the hydrogen penetrated in
dx.doi.org/10.1021/jp309991x | J. Phys. Chem. B 2013, 117, 422-425


The Journal of Physical Chemistry B the voids should have protected them from contracting. A similar effect of preventing the voids in amorphous silica from contracting at high helium pressures due to their filling with He atoms was earlier observed and thoroughly discussed in refs 6 and 7. The behavior of silica glass compressed in diamond anvil cells up to 190 kbar at room temperature was studied in ref 7 by in-situ X-ray diffraction and Raman spectroscopy using helium and hydrogen as the pressure-transmitting medium and also without such a medium. In agreement with earlier results,6 the FSDP position in the X-ray patterns of the silica glass in helium was nearly pressure-independent, thus demonstrating that the helium atoms were likely to enter the voids in the glass. Compressing silica glass in hydrogen resulted in approximately the same strong shift of FSDP to higher frequencies as in the case of using no pressure-transmitting medium. On the basis of this result, the authors of ref 7 concluded that, in contrast to helium, hydrogen cannot dissolve in silica glass in a large amount because of the larger kinetic diameter of H2 molecules (2.9 å8) compared to that of He atoms (2.6 å12). This conclusion actually predicts that concentrated hydrogen solutions in silica glass cannot be formed at any pressure and temperature and therefore contradicts our experimental results demonstrating that at a pressure of 75 kbar the H2 solubility in silica glass reaches a value of H2/SO2 0.5, which is similar to the estimated He solubility He/SO2 0.8-26 (see Figure 2). It can also be added that the He solubility in silica glass at high pressures has never been measured experimentally, so the difference between the He and H2 solubilities can even appear smaller. We think that the absence of any detectable H2 solubility in silica glass in the high-pressure experiments of ref 7 was a result of the low diffusion rate of hydrogen at room temperature. In fact, the diffusion coefficient for H2 molecules in silica glass at room temperature is D = 1 â 10-11 cm2/s,16 which is by 3 orders of magnitude lower than D = 2.4 â 10-8 cm2/s for He atoms.17 At the temperature of 250 °C used in the present work, the H2 diffusion coefficient in silica glass rises to a value of D = 1.9 â 10-8 cm2/s,16 close to that for the He coefficient at room temperature, and the concentrated SiO2- H2 solid solutions are easily formed.

Article

without hydrogen. A similar effect was earlier observed in situ at helium penetrations in the silica glass.6,7



AUTHOR INFORMATION

Corresponding Author

*Phone +7 906 495 24027; Fax +7(496) 522 8160; e-mail efimchen@issp.ac.ru.
Author Contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. All authors contributed equally.
Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The authors thank Dr. Vladimir S. Balitsky for his comments and useful discussions. The work was supported by Grant No. 12-02-00871 from the Russian Foundation for Basic Research and by the Program "Elementary Particle Physics, Fundamental Nuclear Physics and Nuclear Technologies" of the Russian Academy of Sciences. REFERENCES
(1) Montmerle, T.; Augereau, J. C.; Chaussidon, M.; et al. Earth, Moon, Planets 2006, 98, 39-95. (2) Safronov, V. S. Evolution of the Protoplanetary Cloud and Formation of the Earth and the Planets; Nauka Press: Moscow, USSR, 1969; Trans. NASA TTF 677, 1972. (3) Hartwig, C. M. J. Appl. Phys. 1976, 47, 956-959. (4) Shelby, J. E. J. Appl. Phys. 1977, 48, 3387-3394. (5) Li, C. Y.; Price, J.; Tomozawa, M.; Watson, E. B. J. Non-Cryst. Solids 2011, 357, 2081-2085. (6) Sato, T.; Funamori, N.; Yagi, T. Nat. Commun. 2011, 2, 345-350. (7) Shen, G.; Meia, Q.; Prakapenka, V. B.; et al. Proc. Natl. Acad. Sci. U. S. A. 2011, 108, 6004-6007. (8) Breck, D. W. Zeolite Molecular Sieves: Structure, Chemistry and Use; Wiley: NewYork, 1974. (9) Kirchheim, R. Acta Metall. 1982, 30, 1069-1078. (10) Efimchenko, V. S.; Kuzovnikov, M. A.; Fedotov, V. K.; et al. J. Alloys Compd. 2011, 509S, S860-S863. (11) Shackelford, J. F.; Masaryk, J. S. J. Non-Cryst. Solids 1978, 30, 127-139. (12) Shelby, J. E. J. Appl. Phys. 1976, 47, 135-139. (13) Herzberg, G. Infrared and Raman Spectra; Van Nostrand: Princeton, NJ, 1945. (14) Wijngaarden, R. J.; Lagendijk, A.; Silvera, I. F. Phys. Rev. B 1982, 26, 4957-4962. (15) Inamura, Y.; Katayama, Y.; Utsumi, W.; Funakoshi, K. Phys. Rev. Lett. 2004, 93, 015501-015504. (16) Shang, L.; Chou, I.; Lu, W.; Burruss, R. C.; Zhang, Y. Geochim. Cosmochim. Acta 2009, 73, 5435-5443. (17) Barrer, R. M.; Vaughan, D. E. W. Trans. Faraday Soc. 1967, 63, 2275-2290.

CONCLUSIONS The hydrogen solubility in amorphous silica at pressures up to 75 kbar and temperature 250 °C has been investigated for the first time. The composition study of the samples, hydrogenated under pressure and quenched to the liquid nitrogen temperature, revealed a linear dependence of the molar ratio H2/SiO2 versus synthesis pressure, from X = 0.16 at P = 6 kbar to X = 0.53 at P = 75 kbar. A Raman study of the sample with X = 0.47 showed that the hydrogen is dissolved in the form of H2 molecules, which occupy at least two different types of voids in the amorphous silica network. The H2 molecules occupying the voids of the first type (most likely, inside the six-membered rings of silica) and of the second type (not identified yet) have the vibration stretching bands at 4168 and 4374 cm-1, respectively. Judging by the ratio of integral intensities of these bands, the great majority (90%) of the dissolved hydrogen molecules are located in the voids of the second, yet unidentified type. Also, our X-ray diffraction studies showed that the hydrogen molecules penetrated in the silica glass network at a hydrogen pressure of 75 kbar prevented the sample from the significant irreversible densification occurring in the case of compressing the silica glass to the same pressure
425



dx.doi.org/10.1021/jp309991x | J. Phys. Chem. B 2013, 117, 422-425