Theory of sparse random matrices and vibrational spectra of amorphous solids

The random matrix theory has been used for analyzing vibrational spectra of amorphous solids. The random dynamical matrix M = AAT with nonnegative eigenvalues ɛ = ω2 has been investigated. The matrix A is an arbitrary square (N-by-N) real sparse random matrix with n nonzero elements in each row, mean values 〈Aij〉 = 0, and finite variance 〈Aij2〉 = V2. It has been demonstrated that the density of vibrational states g(ω) of this matrix at N, n ≫ 1 is described by the Wigner quarter-circle law with the radius independent of N. For n ≪ N, this representation of the dynamical matrix M = AAT makes it possible in a number of cases to adequately describe the interaction of atoms in amorphous solids. The statistics of levels (eigenfrequencies) of the matrix M is adequately described by the Wigner surmise formula and indicates the repulsion of vibrational terms. The participation ratio of the vibrational modes is approximately equal to 0.2–0.3 almost over the entire range of frequencies. The conclusions are in qualitative and, frequently, quantitative agreement with the results of numerical calculations performed by molecular dynamics methods for real amorphous systems.