Two-network structures

One might imagine a situation where two amorphous sublattices could be defined topologically, as sketched in Fig. 1, but it seems unlikely that this ever arises in practice.
There remains, however, the possibility of distinguishing two magnetic subnetworks in amorphous solids on a chemical basis. Generally, these are the sublattices composed of 3d and 4f atoms, The d–d exchange is strongly ferromagnetic, defining a ferromagnetic 3d subnetwork, and the 3d–4f interactions then tend to align the 4f subnetwork spins antiparallel to the 3d subnetwork. Hence the subnetwork moments are aligned parallel for heavy rare-earths, but antiparallel for the light rare-earths. We can therefore have an amorphous ferrimagnet where the A and B subsystems are defined chemically. An example is a-Gd₂₅Co₇₅. There may be a compensation point at the temperature where M_4f >M_3d , just as in crystalline ferrimagnets.
For rare-earths with strong uniaxial anisotropy and weak exchange coupling to the 3d subnetwork, their local easy axes are defined by the local crystal-field

Figure 1: Types of disorder on two-dimensional monatomic and diatomic networks.

interaction D_iJ ²_zi . These local easy axes are random, leading to the sperimagnetic structures illustrated in Fig. 2.

Figure 2: Possible two-subnetwork magnetic structures in amorphous binary alloys.