ITPO 2020 Solutions: Problem 2
ITPO 2020 Solutions: Problem 2
Problem 2
Consider a free, massless scalar field φ in two-dimensional Minkowski spacetime. Show that
the action is invariant under the transformation rule
¯
φ 7→ φ + δφ ≡ φ − ε(z)∂φ − ε̄(z̄)∂φ (1)
Where z and z̄ are the so-called “lightcone coordinates” in two dimensions, ∂ and ∂¯ are
their corresponding derivatives, and ε(z) and ε̄(z̄) are small but arbitrary functions of the
indicated coordinates. Next, for a small parameter α consider the following Lagrangian
density, as an extension of the free scalar:
¯ + α(∂φ∂φ)
L = ∂φ∂φ ¯ 2 (2)
Is it possible to add O(α) contributions to δφ which are first-order in φ-derivatives and which
preserve this invariance?
Rubric:
Solution:
The first part of the problem is a textbook exercise. Let us replace φ 7→ φ + δφ in L and
consider only the terms which are first order in ε. Without loss of generality we will discard
ε̄ terms, because the analysis will be identical (due to the Z2 symmetry of the action under
∂ ↔ ∂,¯ ε ↔ ε̄). We find:
¯ − ∂φ∂(ε∂φ)
δL = −∂(ε∂φ)∂φ ¯ (3)
¯ − ε∂∂φ∂φ
= −∂ε∂φ∂φ ¯ − ε∂φ∂∂φ
¯
Now, since we are in flat space, we can discard boundary terms – integrating the first term
by parts gives:
¯ + ε∂φ∂ ∂φ
δL = ε∂∂φ∂φ ¯ − ε∂∂φ∂φ
¯ − ε∂φ∂ ∂φ
¯ =0 (4)
1
Which establishes invariance of the original action under these transformations, which are
called “Weyl transformations.”
In order for δ (1) φ to be well-defined for all ε and ε̄, we need to equate (7) with (8), which in
particular requires us to integrate (7) by parts to find a term proportional to ∂ ∂φ. ¯
There are a number of ways to show that you cannot find such a term, and therefore one
cannot define δ (1) φ. Intuitively, it this happens because integrating by parts any given
derivative in (7) will yield terms involving both ∂ ∂¯ and either of ∂∂ or ∂¯∂.
¯ Exhausting all
the possibilities, one formally proves that such additions to the transformation rules cannot
be formulated.
This order-α addition to the Lagrangian density is for various reasons called the T T̄ defor-
mation of the massless, free scalar field. Such deformations are known not to preserve Weyl
invariance, which is what we have basically shown here.