2.5.3 Structure and the glass transition temperature

There is a relation between the ease of chain rotation (controlling conformation) and the locked-in configuration of polymer backbone chains. It is most easily appreciated by examining the effect of different backbone configurations on the glass-transition temperature or T_g. As already noted above, the T_g is the temperature when a rigid amorphous thermoplastic becomes elastomeric, and its stiffness drops steeply. How can this transition temperature be interpreted at a molecular level?

The simplest way at looking at the problem is in terms of the chain rotational model of Section 2.3, where the effect of raising temperature on the conformation of polyethylene was considered. It was of course implicitly assumed that the chain was flexible, and that rotation about carbon-carbon bonds created chain flexibility. And that is true at ambient temperatures of 25 °C, say, provided the level of crystallinity is low or absent. But what happens if the temperature is lowered? As the temperature decreases, there must come a point when all rotation about chain bonds ceases entirely; in other words, the energy available locally in the form of thermal vibration is insufficient to cause neighbouring atoms to twist around one another. The energy needed to achieve rotation is actually shown by the potential energy banners in Figure 22(b). They lie at about 13 kJ mole⁻¹ and 16 kJ mole₋₁ above the energy minimum for the trans conformer in n-butane.

The temperature at which chain molecular rotation ceases is the glass transition temperature, because the chains can no longer respond to external strain by uncoiling and lengthening by chain rotation. In other words, the polymer becomes glassy and rigid. For polyethylene, the T_g is very low and occurs at about −90 °C. However, it is important to mention that the T_g is not necessarily a sharp transition, like the melting point, for example. It can be very broad indeed, with a progressive stiffening effect as temperature is lowered. Indeed with PE, substantial stiffening is already present by −20 °C.

So how do changes in chain structure affect T_g, if the transition is largely controlled by rotation about chain bonds? Introduction of atoms like oxygen where there are no hydrogens to create steric hindrance in the chain would be expected to lower T_g, and this is found to be in general the case (Table 5). POM or acetal resin and PEO also have low T_gS, and silicone polymer is exceptional in having one of the lowest T_gS of any material, at −125 °C. This is why the (crosslinked) rubber is widely used in gaskets and fuel hose for aircraft, where low temperatures will be encountered when flying at height.

It would also be expected that chains having double bonds, such as BR and NR might have low T_gs since there is little steric hindrance adjacent to this bond. Again, this is found to be the case (Table 5).

On the other hand, if ways of hindering chain rotation are used, for example by increasing the physical size of pendant groups, then the T_g would be expected to increase. Thus polypropylene with a large methyl group on every alternate chain atom has a T_g of about 5 °C. a value that implies PP milk bottle crates could crack on a frosty morning! This problem was overcome by using a copolymer grade with ethylene to lower the T_g to below 0 °C. Increasing the size of the side group in vinyl polymers shows a reasonably regular increase in T_g, with PVA having a value of about 30 °C, PVC a value of about 80 °C and polystyrene with a very large benzene pendant ring hindering rotation has a T_g of 97 °C.

Larger side groups than this are uncommon, so it is worth returning to the structure of the main chain. Benzene rings trapped within the backbone should increase steric hindrance, and hence T_g. The effect is shown in many polymers and was indeed a strategy used for developing polymers stable to high-temperatures. PET for example, is an aromatic polyester and contains such a ring (C₆H₄) in its repeat unit together with a short aliphatic portion (Table 5). It has a T_g of about 65 °C, but a polymer like polycarbonate is much more hindered by its repeat unit, since the short chain is absent. Here, the large group which forms the bulk of the unit is a bisphenol A group consisting of two benzene rings connected by a carbon atom with two pendant methyl groups, so there is considerable resistance to rotation. Its T_g is about 149 °C. Finally, there are materials like aramids and polyimides where the chain is either completed prevented from rotation at all (PI), or degrades by chain breakage before rotation is possible (aramids). The concept of T_g in these cases becomes redundant.