• Register
PhysicsOverflow is a next-generation academic platform for physicists and astronomers, including a community peer review system and a postgraduate-level discussion forum analogous to MathOverflow.

Welcome to PhysicsOverflow! PhysicsOverflow is an open platform for community peer review and graduate-level Physics discussion.

Please help promote PhysicsOverflow ads elsewhere if you like it.


New printer friendly PO pages!

Migration to Bielefeld University was successful!

Please vote for this year's PhysicsOverflow ads!

Please do help out in categorising submissions. Submit a paper to PhysicsOverflow!

... see more

Tools for paper authors

Submit paper
Claim Paper Authorship

Tools for SE users

Search User
Reclaim SE Account
Request Account Merger
Nativise imported posts
Claim post (deleted users)
Import SE post

Users whose questions have been imported from Physics Stack Exchange, Theoretical Physics Stack Exchange, or any other Stack Exchange site are kindly requested to reclaim their account and not to register as a new user.

Public \(\beta\) tools

Report a bug with a feature
Request a new functionality
404 page design
Send feedback


(propose a free ad)

Site Statistics

203 submissions , 161 unreviewed
5,007 questions , 2,163 unanswered
5,342 answers , 22,657 comments
1,470 users with positive rep
815 active unimported users
More ...

  Why isn't the renormalization of composite operators determined by the renormalization of elementary fields?

+ 7 like - 0 dislike

In typical expositions of renormalization of composite operators, one needs such renormalization to tame the divergences of some matrix elements such as (say in an interacting scalar theory)

\[\langle \alpha|\phi^2(x)|\beta\rangle\]

One renormalizes $\phi^2$ by introducing the renormalized composite operator


My question is, doesn't the above relation imply $Z_{(\phi^2)}=Z^2_{\phi}$, where $Z_{\phi}$ is the field strength renormalization of $\phi$? That is, once the elementary field is renormalized, how come one can still have the freedom to renormalize the composite operator?

Useful scholarpedia introduction on the subject(still scratching my head reading it): http://www.scholarpedia.org/article/Local_operator#Definition_of_normal_products_and_operator-mixing

asked Mar 8, 2015 in Theoretical Physics by Jia Yiyang (2,640 points) [ revision history ]
edited Apr 26, 2015 by Jia Yiyang

In the book ''The theory of quark and gluon interactions'' by F.J. Yndurain (4th edition), there is on p.76 beginning with (3.4.2) an explicit example for the 1-loop renormalization of a composite operator in QCD. 

@ArnoldNeumaier, thanks for the new reference!

2 Answers

+ 5 like - 0 dislike

One has no extra freedom in renormalizing the composite fields (i.e., no new parameters in the renormalization prescription).

However, the normalization factors are not multiplicative since the operator product expansion is singular, and the limits do not commute as it would be needed to show multiplicativity. Therefore the renormalization of $\phi^2$ is quite different from the square of the renormalization of $\phi$ - which is not even defined, since all local fields are distributions only that cannot be multiplied. (In conformal field theories, not even the scaling dimensions add, as one would expect from a naive argument.)

answered Mar 8, 2015 by Arnold Neumaier (15,777 points) [ revision history ]
edited Mar 9, 2015 by Arnold Neumaier
Most voted comments show all comments

 I must admit there must be something very elementary that I don't understand: In the renormalized perturbation theory, when we write down the renormalized Lagrangian, the $\phi^2$ terms simply become $Z^2\phi^2$, where $Z$ is the field strength renormalization factor. However in composite operator renormalization this does not look so. What's the difference in the two cases?

@JiaYiyang: This is because a composite field is not just a product, but a renormalized version of the deformed normally ordered product.

Even in the free case, the product $\phi^2$ is ill-defined, and only $:\phi^2:$ makes sense. In the interacting case, everything is deformed, the vacuum, the field, and the recipe for normal ordering, by infinitely many terms to be constructed perturbatively.

If you use your expression, hence $:(Z\phi)^2:$ in the renormalized Lagrangian, you renormalized $\phi$ but not $\phi^2$, since the normal ordering is still that of the free procedure.

In a Lagrangian setting, what determines the renormalization is the success in getting a meaningful asymptotic perturbation theory. So one just follows the practitioners in the field, who had worked out what works, and does the same to verify that it indeed works.

I know that this is somewhat unsatisfactory, but at present nobody has better rules. A somewhat better understanding of field products is given by the operator product expansion. The cleanest procedure is perhaps to take this as the starting point, as Hollands and Olbermann do; then only physical constants arise.


 Weinberg used a very arbitrary (to my eyes) condition in equation (18.1.11)(Vol2 page 117), which basically requires ⟨α|ϕ^2|β⟩ to be normalized to 1 at zero momentum, what does this mean?

Since $Z_{\phi^2}$ is infinite in the limit of infinite cutoff, any finite rescaling of it has the same property. hence the scale of the renormalized field is determined only after fixing some expectation of it. This can be arbitrary, as different ways or choosing the scale just select different multiples of the same field and call this multiple the renormalized field.

The thing is you don't have such arbitrariness in the renormalization of field strength, mass and other coupling constants. Due to LSZ, S-matrix becomes a very solid anchoring point for us to write down the renormalization conditions, and seemingly different renormalization schemes either result from different regularization schemes, or different schemes of relating correlation function to S-matrix, so the freedom of such renormalizations is very limited. What would be the "anchoring point" for the renormalization of local products of operators?

 I gradually realize it is really a question about the meaning of local products, maybe I'll start a new thread if necessary.

Stefan Hollands has several papers on perturbative OPE construction (for scalar fields), which ''proves'' the existence of all local products (i.e., demonstrates their existence at the level of rigor of theoretical physics).

Most recent comments show all comments

@JiaYiyang: As the whole procedure and how to talk about it is ill-defined, it is difficult for me to identify in full detail how to make it well-defined. My way of processing the literature on ill-defined objects is simply to take for granted the properties and tricks that proved to work, and trust an authour unless another one contradicts. So on that level I can say little more than Weinberg, apart from providing my own ill-defined intuition.

Independent of that I am looking for a coherent way to organize things in a more consistent way. In the case of operator products, the rigorous interpretation (though nowhere rigorously completed) is along the lines of Hollands: For free fields, the well-defined variations of the products are the normally ordered products with respect to a vacuum state (or more generally a coherent state - these typically provide inequivalent representations of the CCR). Thus $\phi^2$ should be understood as $:\phi^2:$, and $\phi^4\sim(\phi^2)^2$ as $:\phi^4:=:(:\phi^2:)^2:$. In the case with interactions, everything is deformed according to finite rules that were derived by Hollands in perturbation theory. This deformation defines the renormalized interaction, and $\phi$ and $\phi^2$ are defomed differently. I believe that there lies the future, and some version of this might become fully rigorous one day.

How these two pictures (the traditional one and the near-rigorous one) relate to each other is again given only by some ill-defined intuition that I tried to convey. Don't attach too much meaning to it.

@ArnoldNeumaier, thanks, but I'm hardly convinced, let me think about it.

+ 3 like - 0 dislike

This answer relies pretty heavily on http://www.scholarpedia.org/article/Local_operator. Some material can also be found in Peskin and Schröder and Weinberg, but the above article is quite self contained and gives a good overview.

I think there is a misconception regarding different types of divergences in QFT's. Consider for example a renormalizable QFT of a scalar field \(\phi_0\), where this denotes the bare field. The standard procedure then is that you rescale/shift the fields, masses and coupling constants so that your Lagrangian is expressed in terms of renormalized (physical) fields, masses, coupling constants and counterterms. The counterterms are fixed by appropriate renormalization conditions. Then all Green's-functions of the renormalized fields \(\phi_R= Z^{-\frac{1}{2}} \phi_0\) at different spacetime positions are well defined. However that does not mean that for example \(\langle \phi_R(x) \phi_R(x) \rangle_0\) is finite (0 denotes the interacting vacuum here). As an example consider  a non interacting theory of a scalar field (where no field renormalization occurs at all). \( \phi(x) \phi(x) \) is ill defined since for example the vacuum expectation value (VEV) diverges. A regular operator mimicking the \(\phi^2\) operator in the free case is the normal ordered version \(: \phi^2 :\) which can be written (schematically) as \(: \phi^2 : = \phi^2 -\mathbb{1} \Delta(0) \), where the propagator and the identity operator appear. 

This procedure of normal ordering inspires the proper definition of local regular operators in any interacting QFT. For any operator \(O(x)\) that is a product of elementary (and already renormalized) field operators and their derivatives at the same spacetime point we can define an operator \(N_{\delta}[O(x)]\), which can be expressed as a finite sum \(\Sigma_l Z_l O_l(x)\) where \(Z_l\) are (usually) infinite​ constants, and \(O_l(x)\) are other local field operators built from elementary fields and their derivatives. In some special cases \(N_{\delta}[O(x)]= Z O(x)\), but in the general case other operators appear on the righthandside. This is then called operator mixing. \(\delta\) is a parameter related to the mass dimension of the operator. 

The prescription on how to construct these renormalized versions of local operators requires additional conditions, called the BPHZ normalization conditions. All of the above is a highly abbreviated and simplified version of http://www.scholarpedia.org/article/Local_operator. This article also mentions why operators like the electromagnetic current \(J(x)\) which consists of 2 fermion fields does not require this procedure. Due to it being conserved/ gauge invariant we have \(N_{\delta}[J(x)]=J(x)\). I hope this helps a bit

answered Oct 29, 2019 by jm [ no revision ]

Your answer

Please use answers only to (at least partly) answer questions. To comment, discuss, or ask for clarification, leave a comment instead.
To mask links under text, please type your text, highlight it, and click the "link" button. You can then enter your link URL.
Please consult the FAQ for as to how to format your post.
This is the answer box; if you want to write a comment instead, please use the 'add comment' button.
Live preview (may slow down editor)   Preview
Your name to display (optional):
Privacy: Your email address will only be used for sending these notifications.
Anti-spam verification:
If you are a human please identify the position of the character covered by the symbol $\varnothing$ in the following word:
Then drag the red bullet below over the corresponding character of our banner. When you drop it there, the bullet changes to green (on slow internet connections after a few seconds).
To avoid this verification in future, please log in or register.

user contributions licensed under cc by-sa 3.0 with attribution required

Your rights