Title: Extreme Stability in Organic Glasses: Challenges to the Kauzmann Paradox and the Idea of an Ideal Glass Transition

Abstract: A major premise in many theories of glass formation in complex fluids is the concept of an ideal, thermodynamic glass transition taking place at or near the Kauzmann temperature T K where the system entropy extrapolates to a zero value. This thermodynamically based T K is associated with the Vogel, Fulcher, Tammann temperature T VFT at which the relaxation time or the viscosity extrapolate to infinite values above zero Kelvin. At the same time, it is known that many non-crystallizable materials, especially polymers, form glasses. This suggests that the importance of the Kauzmann paradox to glass formation may be problematic. Here we show results from two non-crystallizable ultra-stable glasses (evidenced by greatly reduced
fictive temperatures T F ) that give new insights into the conundrum of the Kauzmann paradox. In the first instance we use vapor deposition to make an ultra-stable amorphous perfluoropolymer (CYTOP TM ) that shows a fictive temperature T F that is more than 60 K below the glass transition temperature T g and approximately 11 K below T VFT . The results also show (and validate) the potential of using the vapor deposition process to create extremely stable glasses even when the material to be deposited is a polymer. Examples of a similar amorphous perfluoropolymer (Teflon AF TM ) studied previously had suggested the possibility extreme stability in a vapor deposited polymer. The second instance uses length change dilatometry measurements on a 50-million-year-old amber from Fushun, China. In this case direct measurements of the fictive temperature show that the material is not only ultra-stable, but also that the TF is as much as 196 K below the T g , i.e., over 100 K below T VFT . In addition to the dilatometry, the experiments were conducted in such a fashion that the material dynamics were obtained for glasses having different fictive temperatures as a function of temperature, fictive temperature and density. We find that the case in which the value of the test temperature T > T F , confirms prior results that suggest non-diverging relaxation times even though this condition should give upper bounds to the relaxation times at any given temperature. In addition, we find that the relaxation times for the condition when T=T F (which should be a state for which the volume is equal to the equilibrium liquid volume at the temperature of interest) are exponential in temperature and do not diverge, though the times measured are in excess of yotta seconds (10 24 s). This is a striking difference from prior findings that suggest a change in dynamics from super-Arrhenius to Arrhenius behavior as one goes below the glass temperature T g . The implications of these findings will be discussed.

Host: Laura Clarke