We have seen that the toughness of bulk (glassy) polymers, which craze under tension, begins to be understood through an original idea of H. Brown [18]. We shall now try to extend the Brown ideas to various systems of ‘weak junctions’. The junction may be a partly healed contact between two identical polymer blocks A/A, as in the experiments of the Lausanne group [19,20]. Alternatively, it could be a contact between two different polymers A and B.

In all our discussions, we shall assume these junctions to be perfect, with full contact between the two partners, and no gaps. Experimental arguments for the existence of these good contacts have been presented by Kausch and coworkers [19].

Our aim here is:

1. (a) to give a brief reminder of the theoretical description of the weak junctions;

2. (b) to show how some basic mechanical properties can be related to the structure.

One of the major conclusions, for the A/A case, is that chain ends play a crucial role. Thus, any attraction between a chain end and the free surface of one A block will react significantly on the A/A mechanical properties after welding.

This type of attraction was first suggested by systematic experiments on melts by D. Legrand and G. Gaines [21], showing that the surface tension γ of oligomers was often lower than the surface tension γ∞ of a high polymer, and that the correction has the form:

where N is the degree of polymerisation, and x an exponent of order 2/3. The fact that x < 1 shows that we are not dealing with a simple uniform dilution of chains ends (which would give a correction ∼N−1).