# One loop correction to $F^2$ in massless QED, question from Peskin & Schroeder

In Peskin & Schroeder chapter 19, about trace anomaly in massless QED, the trace of $\Theta^{\mu\nu}$ is given by $${\Theta^\mu} _\mu =-\frac{4-d}{4} (F_{\lambda\sigma})^2 + (1-d) \bar{\psi} i \displaystyle{\not}D \psi$$ The one-loop matrix element of ${\Theta^\mu} _\mu$ is given by the following three diagrams.

I do not understand these diagrams. Just as explaination in the P&S, the expectation of $(1-d) \bar{\psi} i \displaystyle{\not}D \psi$ is zero, the one-loop matrix element is from the first term which is can be written as $$-\frac{4-d}{2}A_\mu(-k) (k^2g^{\mu\nu}-k^\mu k^\nu)A_\nu(k)$$ This seems to be the usual photon propagator with somr projection operator inserted. Why not from the usual diagrams like the followings?

• Could you clarify what your exact question is? – ACuriousMind Nov 11 '14 at 21:37