|
|
| home |
|
|
| people |
| publications |
| research |
| software |
| resources |
| facilities |
| course pages |
| positions |
| contact us |
|
|
| |
Controlled Pore Glasses
About porous glasses
Preparation of experimental glasses
Spinodal Decomposition
Models
What's the Point?
Publications
Visualizations
Links
About Porous Glasses
Controlled pore glass is widely used as a stationary phase in chromatography. Controlled pore glasses (CPG's) and the related Vycor glasses have excellent mechanical properties, and can be prepared with a wide range of porosities and average pore sizes. They can be modified to include a variety of functional groups, and the adsorption strength of the glasses can be adjusted over a wide range of values. Although controlled pore glasses were developed for use in size-exclusion chromatography, derivatized glasses can show a high chemical affinity for certain biomolecules, and can even be used as catalytic agents and bioreactors.Because of these properties, porous glasses are often used as a substrate in studies of the thermodynamics of confined systems, especially liquid-vapor (capillary) equilibria and liquid-liquid equilibria. Our own interest is in modeling these phenomena. Previous attempts at modeling confined systems have relied on idealized models of the pore systems themselves, which often compare poorly with experimental data.
Preparation of Experimental Glasses
Vycor glass is prepared from a quaternary glass mixture, of typical composition 62.7% SiO2, 26.9% B2O3, 6.6% Na2O, and 3.5% Al2O3. This glass is melted and formed into the desired shape, and then held at a temperature above the annealing point but below that which would cause deformation. The material phase-separates (on a microscopic scale) into two continuous phases, one rich in silica, the other in borosilicate and alkali. It is then treated with a hot dilute acid solution, which dissolves away the borate, leaving some (small) colloidal silica particles inside the pores of the other phase. The finished glass is 96% silica.Vycors have a porosity of 28%, an average internal pore diameter somewhere between 4 and 6 nanometers, and a surface area of between around 90 and 200 m2/g, calculated from BET analysis of nitrogen adsorption isotherms.
CPG's are prepared like Vycor glasses. The starting material is 50-75% SiO2 1--10% Na2O, and the remainder B2O3. The molten glass is phase separated by cooling to between 500 and 750 C. The time taken for this treatment determines the extent of phase separation and the resulting average pore size. The borate phase is leached out by acid solutions at high temperatures. The remaining glass contains also colloidal silica particles, which are removed by a treatment with NaOH followed by washing with water. The final glass has a porosity between 50% and 75%, and an average pore size between 4.5 nm and 400 nm. CPG has a surface area somewhere between 10 and 350 m2/g, depending on the pore size.
Spinodal Decomposition
The fundamental process behind the preparation of both types of porous glass is near-critical spinodal decomposition . When a mixture of two liquids near to their critical composition is quenched, a process of dynamic phase separation occurs. The two phases form complex, highly connected network structures, which grow in time according to a power law. A movie of this process is available on our visualizations page.The morphology of the network structures appears to scale in time, so that this process can be used to produce a series of materials with similar topological properties but different average pore sizes. Furthermore, the properties of the network appear to be insensitive to the exact liquids used, so that many partially miscible systems exhibit similar behavior and pattern formation upon quenching.
Models
Our approach is to simulate the production of real glasses using molecular dynamics methods, in order to generate morphologically correct glass structures "ab initio". Because the phase separation process is similar for many different mixtures, we can use a much simpler mixture in this study.We have used a symmetric binary Lennard-Jones mixture for this, which shows extremely fast phase separation, and is very simple to work with. Using a realistic model of the silicate/borosilicate mixture would be prohibitively expensive; the model would be very complex, and the necessary length of simulation would be much greater. (In addition, current state-of-the-art two-body potentials proposed for these mixtures do not phase separate correctly; some potential development work would also be necessary.)By using a simple model we greatly reduce the cost of the computation. In addition, we can simulate for longer times, which allows access to a larger range of pore sizes. The trade-off is that we lose detailed information about the chemistry of the glass surface.
It may be possible, in future work, to regenerate this detail by use of a templating process, or even by the development of a cheaper model of the real system which exhibits the correct tetrahedral liquid structure but nonetheless has no long-range forces to calculate.
What's the Point?
Computer simulation is in many ways an "ideal" laboratory. All the parameters in a simulation (pressure, temperature, etc.) are exactly controllable, there are no impurities, and essentially any quantity which can be microscopically defined can be measured. Measurements only suffer from statistical error, so that very high precision data can be obtained by using sufficiently long runs. The downside, of course, is that you study a model system rather than a real one. As a result, computer simulations of this type are ideal for testing the effectiveness of general theories and predictions. For instance, we have used these glass models and Monte Carlo simulations of nitrogen adsorption to evaluate the efficacy of the BET method of surface area determination for this class of materials. Because the method is not dependent on any particular properties of nitrogen, or any particular surface chemistry, using a simplified model to test it is acceptable. This study found systematic deviations in the BET surface area due to strongly curved surfaces in small pores, and by looking at microscopic quantities difficult to obtain experimentally, we could explain these deviations at a semi-quantitative level.Publications
The first paper on this work appeared in the April 14, 1998 issue of Langmuir .
We even got the cover photograph
(warning: 100Kb JPG !)
Downloadable copies of this paper and others that we've written about these models are available on the publications page.
Visualizations
A whole page full of visualizations of these model materials is here.Links
Here are some links for CPG and Vycor glass sources:- Vycor glasses are available from Corning (Note that Vycors comes in two varieties; porous and non-porous, with the non-porous glass being produced from the porous one.)
- Prime Synthesis produces both bulk CPG and other reagents and materials for DNA synthesis.