The distribution of tetravalent network glasses

Abstract

A model tetravalent network fluid, crystal and many glasses are studied by molecular dynamics simulation. 171 glasses were made by compressing the fluid with five orders of magnitude variation in the quench rate. The pressure and entropy of each glass are expressed as functions of a single variable, the quench rate dependent limiting density, z ₀ of the rigidly jammed state where the pressure diverges. The number of possible glasses with limiting density z₀ is approximated by a Gaussian distribution N _g(z ₀)dz₀ = exp{1·2N}exp(-123N(Δz ₀)²}dz₀, where Δz ₀ = z ₀ −0·766 and N is the number of molecules. That distribution implies that In {N _g(z ₀)}/N → 0 as z ₀ → 0·864, which suggests that an ideal glass transition would occur to a glass with z ₀ = 0·864 with slow compression rates, if the fluid did not freeze. We show that the free energy and pressure of the dense fluid can be simply expressed in terms of the properties of the glasses.