Let A∈M2(Z2) be an expanding matrix and let D∈Z2 be a digit set as follows: D={(0,0)*, (1,0)*, (0,1)*}, where v* means the compose of the vector v. Then there exists a unique probability measure μA,D satisfying that which is called a Sierpinski measure.In this thesis we study the spectrality of Sierpinski measures, that is, to inves-tigate the existence of a countable set Λ such that E(Λ)={e-2πi(λ,x):λ∈Λ} is an orthonormal basis (Fourier basis) for L2((μA,D). We prove that the Sierpinski measure μA,D is a spectral one if and only if (A, D) is admissible. This answer the open problem raised by Li (Sci. China Math.,56 (2013),1619-1628). Combining the series works of Li on this issue, we obtain the characteristic conditions for the spectrality of Sierpinski measures. In the same time, we give a positive answer for the following open problem:Under what conditions on an integral matrix A and digit set C(?)Z2, the integral self-affine measure μA,C is a spectral one if and only if (A, C) is admissible. That is, we show that the open problem is true for Sierpinski measures. |