Transformation operators for the canonical Dirac system (Q2269324)

Let \(E\) be the \(2\times 2\) identity matrix, and let \[ \sigma_1={0\quad i\choose -i\;0},\quad \sigma_2={1\quad 0\choose 0 \;-1},\quad \sigma_3={0\quad 1\choose 1\quad 0} \] be the well-known Pauli matrices, which have the following properties: \(\sigma^2_k=E\), \(\sigma^*_k=\sigma_k\) (selfadjointness), \(k=1,2,3\), and \(\sigma_k,\sigma_j=-\sigma_j\sigma_k\), \(k\neq j\) (anticommutativity). Consider the canonical Dirac system \[ \ell y\equiv By'+\Omega(x)y=\lambda y,\tag{1} \] where \(B=\frac1i\,\sigma_1\) and the matrix \(\Omega(x)=\sigma_2p(x)+\sigma_3q(x)\) is usually referred to as a potential. For locally integrable coefficients, we prove the existence of transformation operators and estimate the kernels of these operators. We also give estimates for these kernels for the case in which the coefficients belong to the space \(L^2_{\text{loc}}\). We establish a relationship between the kernel of the transformation operators and the potential matrix.