On the Hartree-Fock equations of the electron-positron field (Q2568929)

Let \(D_0:=-i\alpha\cdot\nabla+m\beta\), and \(F\): the set of all self-adjoint operators \(\delta(x,y):=\sum_n\lambda_n\xi_n(x)\overline {\xi_n(y)}\) on \({\mathcal H}:=L^2({\mathbb R}^3)\otimes\mathbb{C}^4\) with finite \(\text{tr} (|D_0 |^{1/2}|\delta||D_0|^{1/2})\). \(\rho_\delta({\mathfrak x}):=\sum_{s=1\sim 4} \sum_n\lambda_n| \xi_n({\mathfrak x},s)|^2\), \(\varphi^{(\delta)}({\mathfrak x})= \int_{{\mathbb R}^3}\{\rho_\delta({\mathfrak y})/|{\mathfrak x}-{\mathfrak y}|\}\,d{\mathfrak y}\), \(X^{(\delta)}:\psi\to\int\{\delta(x,y)\psi(y)/|{\mathfrak x}-{\mathfrak y}|\| \,dy\). The authors study the D-C operator \(D_z:=D_0-\alpha Z/|{\mathfrak x}|\) and the D-F operator \(D^{(\delta)}:=D_z+(\varphi^{(\delta)}-X^{(\delta)})\). Let \(\Lambda_+^{(\delta)}\) be the projection onto the positive spectral subspace of \(D^{(\delta)}\),\break \(\Lambda_-^{(\delta)}=1-\Lambda_+^{(\delta)}\), \(D[\rho,\sigma]:= 2^{-1}\int_{{\mathbb R}^3} d{\mathfrak x}\int_{{\mathbb R}^3}d{\mathfrak y}\rho({\mathfrak x}) \sigma ({ \mathfrak y})/|{\mathfrak x}-{\mathfrak y}|\), and \[ E[\gamma,\gamma']:=2^{-1} \int dx \int dy\gamma(x,y)\gamma'(x,y)/|{\mathfrak x}-{\mathfrak y}|\quad\text{for }\gamma,\gamma'\in F. \] Let \(S^{(\delta)}:=\{\gamma\in F;-\Lambda_-^{(\delta)}\leq\gamma\leq \Lambda_+^{( \delta)}\), \(\Lambda_-^{(\delta)}\gamma\Lambda_+^{(\delta)}=0\}\), \(S_{\partial q}^{(\delta)}:=\{\gamma\in S^{(\delta)};\text{tr}\, \gamma=q>0\}\). The functional \({\mathcal E}:\gamma\in S^{(\delta)}\to \text{tr}(D_z\gamma)+\alpha Q[\gamma,\gamma] \in{\mathbb R}\) gives the energy, where \(Q:(\gamma,\gamma')\to D[\rho_\gamma, \rho_{\tau'}]-E[\gamma,\gamma']\). When \(q\in\mathbb{N}\) \(q\leq Z\), \({\mathcal E}\) has a minimizer in \(S_{\partial q}^{(\delta)}\). Assume that \(\gamma\) is an orthogonal projection minimizing \({\mathcal E}\) in \(S_{\partial q}^{( \delta) }\). Then \(\gamma=\sum_{i=1}^q|\psi_i\rangle\langle\psi_i|\), \(\Lambda_+^{(\delta)} D^{(\gamma)}\Lambda_+^{(\delta)} \psi_i = \varepsilon_i\psi_i\), \(\varepsilon_i\in [0,1]\). Finally they study the density matrix \(\Gamma:=\Lambda_-^{(\delta)}+ \gamma\), and obtain the Euler equation \([D^{(\gamma)},\Gamma]=[D^{(\gamma)}, \Lambda]=0\).