I think it does and may be the loss of your new sample from a glass former and the excess from crystallization agent like vanadium oxide

Exactly, the glass forming ability strongly dependent on the compositions of the glass system.

I would suggest you consider the purity of your raw materials, the size of the batch, the accuracy of your weighing and any interactions or variations in interactions between the melt and the crucible / atmosphere. All of these can have an effect. Without knowing details of your glass composition it is indeed possible that a difference of 0.05mol% could cross the glass forming / crystalline boundary under the superheat / cooling rate conditions you used, but so many other factors could be in play that it would be wise to consider all of these as well. Unless your end-member compositions are exactly on the knife-edge of glass formation in the system and under your preparation conditions, a difference of 005mol% would not normally be expected to make such a large difference. However, of course, there are also cases where it can have such an effect. Basically, I'm advising that you consider all the parameters and potential sources of error and variability in this situation.

@ P.A. Bingham.

For the end compositions i have used alumina crucible but for middle composition i used silica crucible but the heating temp is between 800 - 900 deg celcius. so i dont think use of silica crucible would make a difference, as the M.P of silica is around 1600 deg celcius. However, the batch size of end compositions was 25gm and for middle comp. i used only 5gm batch. Does the less amount of batch or use of different crucibles effect the GFA?

I suspect this information may help you to extract the differences in preparation. This is a little bit of Sherlock Holmes. So - your "middle" glass crystallised. And this is one you melted in SiO2 and using a very small batch size whereas the others were melted in Al2O3 and using larger batches. Perhaps contamination from crucible corrosion and / or batch volatilisation and / or surface area to volume ratio and hence volatilisation / losses etc, could be the cause? Have you analysed your glasses compositionally to find out what they are?

@ P.A. Bingham

Yet I have to carry out XRD of that sample to confirm its amorphous/crystalline nature. Thank U sir for ur kind response. The appearance of the sample is light yellow shiny opaque.

From a theoretical point of view, the glass forming ability - both in covalent and oxide glasses - depends strongly on the average coordination and connectivity. First theory of glass using these topological aspects was created by J.C.Phillips and M.Thorpe, in early eighties. I have derived a nice formula for the glass transition temperature as a function of ratio between the coordination of glass-former and the modifier. Please see my book, "Models of Agglomeration and Glass Transition", Imperial College Press, 2007 - it contains a lot of material on glass transition and its theoretical description. See also my papers with R. Barrio, M.Micoulaut, D.M. dos Santos, and experimental results by P. Boolchand, K.Trachenko

@ Richard Kerner

I will go through the book and papers u have mentioned.

Thank u Sir.

Glass forming ability strongly depends on the glass compositions, i think your composition just out side of the glass forming region.

I would suggest XRF, ICP, or even SEM/EDS analysis of the chemical composition will tell you about any substantive compositional differences - using these should help you answer your question about composition. You can then trace this back to the glass formation behaviour.

We have investigated hundreds of ternry systems, thousands of glass compositions in halide, oxxyhalide and heavy metal oxide glass systems. Indeed, glass forming ability directly depends on glass composition. You must rely on experimental observations rather than theoretical approaches. Afterwards you may imagine somme rational explanation. However logical approach is often more convincing than accurate...

Certain glass compositions can be highly corrosive to Al2O3, SiO2 or similar crucibles, even at the temperatures you quote. For example PbO is a very good solvent of these materials. Hence Platinum is preferable.

You should also consider if your composition (including possible impurities from the container) is close to a line phase or a eutectic of the equilibrium phase diagram. Compositions close to eutectics will have better glass forming ability, and those close to line phases will have worse.

(....and this is one you melted in SiO2 and using a very small batch size !)

In addition to all reasonable suggestions of Paul, Marcel and others, I should emphasize that it is known, and we have observed in our lab, that melting SMALL glass batches often leads to INHOMOGENEOUS glasses , which in turn are prone to devitrification.

Paul has already mentioned that crucible corrosion and volatilization are significant in small glass batches...

I suggest that you look carefully at a polished glass sample for stria. I am almost sure you will find them!

Good luck!

Edgar Zanotto

Center for Research, Technology

and Education in Vitreous Materials

Department of Materials Engineering

Federal University of São Carlos

http://www.certev.ufscar.br