Last Updated: Sep 19, 2016 Views: 73
From Corning Museum of Glass research scientist emeritus, Dr. Robert Brill:
I believe the answer lies in the physical and structural changes that take place at a molecular level when a glass is heated to elevated temperatures. In order to conduct electricity, a material must have some charged species that are mobile and that can move through the material when a voltage is applied across it. (For example, electrons move rapidly and easily through a metal -- and cations and anions move easily through aqueous solutions.) At room temperature a glass can be viewed as consisting of a giant — and continuous — three-dimensional network of silicon and oxygen atoms connected by strong covalent bonds. Interspersed between them are positively-charged sodium and calcium cations that are held in place by ionic bonds. They are bonded to charged oxygen atoms hanging off the silicon-oxygen chains that compose the network. Everything is held firmly in place because all the chemical bonds are so strong. This is what makes a glass overall a rigid material.
The kinetic energy that all the atoms have is largely in the form of vibrations. If a glass is warmed, some of the weaker chemical bonds break (those which are somewhat distorted because the glass has not formed into a regular, symmetrical, crystalline lattice.) Thus, the glass begins to soften slightly. As the temperature is increased further, the next weakest bonds break, and the glass becomes softer still. At some point, the structure has collapsed to an extent that the positively charged cations (Na+, K+, Ca++, Fe+++, etc.) can break free and become mobile. This means that they can move through the softened glass. The motion of the charged species constitutes electrical conduction and redox reactions can then occur at the electrodes that are imposing the electric field. As the glass becomes hotter, it becomes more fluid (its viscosity decreases) and the ions can move more readily. Thus, in contrast to metallic conductors, the conductivity of the softened glass increases with temperature, instead of decreasing. I hope this helps. It is a nice way of illustrating that behavior we observe at a macro level can be visualized (I hope correctly) at what happens at a molecular level.