Simple random matrix model for the vibrational spectrum of structural glasses

To better understand the surprising low-frequency vibrational modes in structural glasses, where the density of states D(ω) deviates from mean field predictions, we study the spectra of a large ensemble of sparse random matrices where disorder is controlled by the distribution of bond weights and network coordination. We find D(ω) has three regimes: a very low-frequency regime that can be predicted analytically using extremal statistics, an intermediate regime with quasilocalized modes, and a plateau in D(ω). When there is a finite probability of bond weights approaching zero strength, the intermediate regime displays a scaling consistent with D(ω)∼ω4, independent of network coordination and system size, just as in simulated structural glasses.