2
$\begingroup$

I am trying to follow Matthew Schwartz's renormalization group lectures (pdf or see Chapter 23 of QFT and the SM by Matthew Schwartz), but I am having trouble with Eq. (book 23.31/pdf 29).

I understand that the $\beta$-function comes from the renormalization group equation (RGE), $$ \beta(\alpha)\equiv \mu\frac{\partial\alpha}{\partial\mu} \tag{book 23.30/pdf 28}, $$ but then the notes introduce $$ \beta(\alpha) = -2\alpha\biggl[\frac{\epsilon}{2} + \biggl(\frac{\alpha}{4\pi}\biggr)\beta_0 + \biggl(\frac{\alpha}{4\pi}\biggr)^2\beta_1 + \biggl(\frac{\alpha}{4\pi}\biggr)^3\beta_2 + \cdots\biggr] \tag{23.31/29}, $$ where I've corrected a typo in the pdf. I totally did not understand this expansion. Then it finally goes to $$ \alpha(\mu)=\frac{2\pi}{\beta_0}\frac{1}{\ln(\frac{\mu}{\Lambda_\text{QED}})}\tag{23.32/30} $$ where $\Lambda_\text{QED}$ is the Landau pole and the fundamental scale of QED. Where does it come from? I'd appreciate it if anyone could help me to derive Eq. 23.31/29 and 23.32/30.

$\endgroup$
1
$\begingroup$

He is writing the $\beta$ function as a perturbation series in the small parameter $\alpha$, $$\beta(\alpha) = \sum_{n=1}^\infty c_n \alpha^n.$$ By convention, it's common to write this expansion in the form he has given. The first coefficient is always $c_1 = - 4\pi\epsilon$ (where $\epsilon$ is the dimensional regularization parameter). The remaining terms define the coefficients $\beta_n$, which one computes in perturbation theory. For example, he has computed the series up to one loop for QED, which he writes in eqn. 27: $$\mu\frac{\mathrm{d}e}{\mathrm{d}\mu} = -\frac{\epsilon}{2} e+ \frac{e^3}{12\pi^2}+\cdots$$ Or, in terms of $\alpha :=e^2/4\pi$, $$\mu \frac{\mathrm{d}\alpha}{\mathrm{d}\mu} = \frac{e}{2\pi} \mu \frac{\mathrm{d}e}{\mathrm{d}\mu} = -\frac{\epsilon}{4\pi}e^2+\frac{e^4}{24\pi^3} +\cdots = -\epsilon\alpha+\frac{2}{3\pi}\alpha^2+\cdots,$$ i.e. $\beta_0 = -4/3$. He sends the regulator $\epsilon$ to 0 and solves the differential equation $$\int_{\alpha(\mu_0)}^{\alpha(\mu)}\frac{\mathrm{d}\alpha}{\alpha^2} = \frac{2}{3\pi} \int_{\mu_0}^\mu\frac{\mathrm{d}\mu}{\mu},$$ which gives $$\alpha(\mu) = -\frac{3\pi}{2}\frac{1}{\log(\mu/\mu_0)}.$$ $\mu_0$ is an integration constant which he has chosen to be $\Lambda_{\text{QED}}$.

$\endgroup$
3
$\begingroup$

Schwartz is simply noting that the $\beta$-function has a generic expansion in QED of the form (29) where $\beta_{0,1,2}$ are some numbers that can be computed by explicitly calculating the various Feynman diagrams.

For instance the leading $\epsilon/2$ is the tree-level result in $d=4-\epsilon$ dimensions. This can be easily seen as follows. In $d=4-\epsilon$, we have $[e]=\epsilon/2$ (do you know how to see this??). Then, we find $[\alpha] = [e^2] = \epsilon$. Thus, we can define the dimensionless renormalized coupling $$ \alpha(\mu) = \mu^{-\epsilon} \alpha $$ This is of course a tree-level result as generically $\alpha$ obtains an anomalous dimension. From this equation, we can determine $$ \beta(\alpha) = \mu \frac{d\alpha}{d\mu} = - \epsilon \alpha $$ This is precisely the first term that he has in the $\beta$-function expansion. If I then take into consideration quantum renormalization effects the $\beta$-function takes the more general form (29).

Good! Now, we understand (29). From (29), we can derive (30). Keeping terms to leading order in $\alpha$ in $d=4\implies \epsilon=0$, we find $$ \beta (\alpha) = \mu \frac{d\alpha}{d\mu} = - \frac{\alpha^2}{2\pi} \beta_0 $$ This differential equation can be solved. Doing so introduces an integration constant $\Lambda_{QED}$ and the solution is of the form $$ \alpha(\mu) = \frac{2\pi}{\beta_0} \frac{1}{\log \frac{\mu}{\Lambda_{QED}}} $$ We can now interpret $\Lambda_{QED}$ as the energy scale where $\alpha(\mu) \to \infty$, i.e. it is the scale where the UV description of the theory breaks down. It is known as the Landau pole.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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