0
$\begingroup$

I have two questions about photon momentum in QED, which I think are related.

1) In Peskin and Schroeder, p. 186, they deduce that the vertex correction must be of the form

$$\Gamma^\mu = \gamma^\mu\cdot A + (p'{}^{\mu}+p^\mu)\cdot B + (p'{}^{\mu}-p^\mu)\cdot C,\tag{6.31}$$

where $p^\mu$ is the momentum of the incoming electron and $p'{}^\mu$ is the momentum of the outgoing electron. Now, my question is why the second term vanishes when dotted with $q_\mu$, the momentum of the scattering photon, as does the first when sandwiched between $\overline{u}(p')$ and $u(p)$?

2) On page 191 below eq. (6.44), they say that $q^2 < 0$ for a scattering process. But if $q$ is the momentum of a photon, shouldn't its square always be zero?

$\endgroup$
1

1 Answer 1

2
$\begingroup$

$$\newcommand{\slashed}[1]{#1\!\!\!/}$$

Since $\Gamma_\mu$ has a Lorentz index $\mu$, it must involve $\gamma_\mu$, $p_\mu, p'_\mu$ (or equivalently, the linear combinations $p_\mu\pm p'_\mu$) such that: $$\Gamma^\mu=\gamma^\mu A(p,p')+(p^\mu+p'^\mu)B(p,p')+(p^\mu-p'^\mu)C(p,p')\tag{1}$$ where the coefficients $A, B$ and $C$ could involve Dirac matrices dotted into vectors i.e., $\slashed p,\slashed p'$. Notice that, $\Gamma_\mu$ is sandwiched between $\bar u(p')$ and $u(p)$. Since $$\slashed{p} u(p) = mu(p) ,~ \hspace{1cm} \bar u(p')\slashed p' = \bar u(p')m\tag{2}$$ we can write the coefficients in terms of ordinary numbers without loss of generality.

Next we can use the Ward Identity $q_\mu\Gamma^\mu=0$ which would further restrict the form of $\Gamma_\mu$. From the diagram $q=p-p'$ and hence the third term $$q\cdot (p'-p)=p'^2-p^2=m^2-m^2=0\tag{3}$$ because the incoming (and outgoing) particles are on shell but the photon is off-shell i.e. $q^2\neq 0$. The first term also vanishes when evaluated between $\bar u(p')$ and $u(p)$ we get, $$\bar u(p^\prime)\slashed q u(p)=\bar u(p^\prime)(\slashed p'-\slashed p) u(p)=\bar u(p')(\slashed p'-\slashed p) u(p)=0\tag{4}$$ Therefore for the Ward identity to hold $C=0$. Hence, $$\Gamma^\mu=\gamma^\mu A(p,p')+(p^\mu+p'^\mu)B(p,p')\tag{5}$$


$q^2$ need not be zero for a virtual photon i.e, need not be on-shell. Note that, $q^2=p-p^\prime$ or $$q^2=p^2+p^{\prime2}-2p\cdot p^\prime=2(m^2-p\cdot p^\prime)\tag{6}.$$ Since both sides are Lorentz invariant, their values are identical in all inertial frames. In particular, if we use the rest frame of the initial electron i.e., $p^\mu=(m,\textbf{0})$ and $p^{\prime}_\mu=(E^\prime,-\textbf{p}^\prime)$, $$q^2=2m(m-E^\prime)<0\tag{7}$$ since $E^\prime>m$.

$\endgroup$
2
  • $\begingroup$ Ok, but why is $q^2$ strictly negative for a scattering process? $\endgroup$
    – Rodrigo
    Commented Feb 14, 2020 at 4:26
  • 1
    $\begingroup$ I have updated the answer. Does it help? @Rodrigo $\endgroup$
    – SRS
    Commented Feb 14, 2020 at 12:35

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.