Amorphous silica and the intergranular structure of nanocrystalline silica

ER van Hoek, R Winter; Phys Chem Glass 43C (2002) 80
Abstract. In the present study, the local structure of nanocrystalline SiO2 with varying particle size, obtained by high-energy ball-milling, is compared with coarse-grained quartz and commercial amorphous fused-silica nano-particles. The average grain size of the nanocrystalline samples is determined by x-ray diffraction, and decreases with milling time down to about 5nm. Both x-ray diffraction (XRD) and 29Si magic-angle spinning (MAS) nuclear magnetic resonance (NMR) signals split into narrow and broad components after several hours of ball milling. On the basis of line shape analysis of diffractograms and spectra, the progressive effect of ball milling on the structure is described in terms of a two-stage process involving the formation of wide-angle grain boundaries and subsequent pressure-amorphisation. An in-situ annealing experiment tracked by XRD reveals that the effect of the first stage is reversible at moderate temperature, while the second stage, which transforms the powder into a nanocrystalline material with grain and intergranular components, is not reversible at temperatures up to 1025oC.
Fig.1: X-ray diffractograms of ball-milled silica Nanocrystalline quartz was prepared by ball milling for up to 128 hours. Polycrystalline quartz and commercial amorphous nanoparticles were used at reference materials in the analysis of NMR spectra and (the former) diffractograms.
Fig.1: X-ray diffractograms of silica particles. Top: Raw material, ball-milled material after 16 and 32 hours. Bottom: Evolution of the two strongest lines and collapse of the structure after 128 hours of grinding. (In both plots, the angle scale corresponds to the bottom diffractogram only; all other diffractograms are progressively shifted by 2deg each.)
The upper part of the figure shows the increase in line width of all diffraction lines with milling time. After 32 hours, only the two most prominent lines, (100) and (101), remain visible above the background. The increase in line width is due to the reduction in particle size. The fewer lattice planes contribute to the Bragg condition, the less sharp is the reflection.
The bottom part of the figure focuses on these lines. After 128 hours, the individual lines can no longer be distinguished. The diffractogram consists only of a very broad amorphous background signal with a small local maximum at the former position of the (101) line. This indicates that the structure is nearly entirely collapsed at this stage.
Fig.2: NMR spectra of ball-milled silica Fig.2: 29Si MAS NMR spectra of silica particles after (bottom to top) 0, 1, 8, 16, 32, 64 hours of ball milling. (The chemical shift scale corresponds to bottom spectrum only; all other spectra are shifted by 2.5ppm each.)
NMR spectra of the nanocrystalline samples show a progressive line broadening also. In this case the line broadening is indicative of the increasing bond-angle disorder around some of the probe nuclei. After 8 hours of grinding, the spectral lines lose their simple Gaussian shape, and a broad line (interfaces) develops underneath the narrow component of the crystalline cores. After longer milling times, the fraction of the broad component continues to increase, and it shifts to less negative chemical shift values indicating a change in the electron density of the interface component.
Fig.3: Comparison and deconvolution of NMR spectra of amorphous, nano-, and polycrystalline silica Fig.3: 29Si MAS NMR spectra of ball-milled (32 hours) silica particles. The broken lines indicate the two components of a double-Gaussian fit (thin line). For comparison, spectra of amorphous silica (top, with single-Gaussian fit) and of coarse-grained silica (not ground, bottom) are shown.
The line width of the interface component of the nanocrystalline sample is comparable to that of amorphous silica. Hence, the extent of bond-angle disorder is similar. The remains of the crystalline line are broadened also, indicating that even the remaining crystalline regions have a rather distorted structure.
The position of the shifted interface line corresponds to one found in pressure amorphised forms of silica glass such as that found in some meteorites or prepared by shock amorphisation (e.g. bullet impact), while the chemical shift of the line of silica glass obtained at ambient pressure corresponds to that of crystalline quartz.
Fig.4: In-situ annealing diffractograms of 128h ball-milled silica If the nanoscale structure of the material is dominated by wide-angle grain boundaries created by severe mechanical impact, the structural integrity can be restored by annealing far below the melting point since only relatively minor atomic rearrangements are required. Amorphised silica on the other hand , being a good glass former, will not recrystallise under such moderate conditions.
Fig.4: Diffractograms of ball-milled (128 hours) silica after heat treatment at a rate of 15K/min. Bottom to top: before annealing, at ~240oC, at ~450oC, and at 1025oC. The grey curve displayed in the background is the diffractogram of coarse-grained quartz with its (100) and (101) reflections, for comparison.
In the temperature range from about 240oC to 450 oC, annealing takes place. This leads to the narrowing of the main structural peak and to the growth of a small (101) peak from the shoulder observed prior to the annealing experiment. The XRD pattern does not markedly change any further up to 1025oC, even on prolonged treatment for several hours at that temperature. This indicates that most of the structure is thoroughly amorphised, and only a small fraction of remaining crystalline material, signified by the reappearance of the (101) line, is retained.
Ball milling of quartz involves a two-step mechanical effect: At first, wide-angle grain boundaries are created, causing all XRD lines to broaden. Atoms in or near the grain boundaries are in a rather disordered environment with a density similar to that of the bulk crystal (broadening of NMR line; no significant change of chemical shift). The second step causes the structure to collapse, with chemical shifts typical of pressure-amorphised silica variants and an NMR line width (bond-angle disorder) typical of a glassy silicate. While the effects of the first step are reversible by annealing, the second step is irreversible.
Acknowledgements. We would like to thank Ray Jones, Materials Support Laboratory, CLRC Daresbury Laboratory, for the opportunity to use their diffractometers. Financial support from the Engineering and Physical Sciences Research Council and the University of Wales Research Fund is also acknowledged.
Physics Home
RW Home
Other nano-ceramics results
Reprint requests
content last modified 030108, RW