Conformal Transformations: Types and Classification

Hello! Welcome back to our course on Conformal Field Theory.

In our previous lesson, we investigated the consequences of requiring angle-preserving symmetry on an infinitesimal level. By solving the conformal Killing equation, we discovered that for dimensions $d > 2$ , the symmetry is generated by four types of transformations: translations ( $P_{μ}$ ), rotations/boosts ( $M_{μν}$ ), dilatations ( $D$ ), and special conformal transformations ( $K_{μ}$ ). We also established the commutation relations that define the conformal algebra.

Today, we will build directly on that foundation. Our goal is to determine the finite form of these transformations. We will "integrate" or "exponentiate" the infinitesimal transformations to see their global effect on spacetime coordinates. This will allow us to classify the four fundamental types of finite conformal transformations, which form the building blocks of the conformal group.

From Infinitesimal to Finite Transformations

As you know from your work with Lie groups in QFT, finite transformations are generated by exponentiating the elements of the corresponding Lie algebra. A finite conformal transformation is a composition of the four basic types we are about to classify. While translations, rotations, and dilatations have familiar finite forms, the Special Conformal Transformation (SCT) is more intricate and reveals a deeper structure.

To begin, please read the following brief excerpt which sets out our task.

EPFL Lectures on Conformal Field Theory in D ⩾ 3 ...

This excerpt from the 'EPFL Lectures on Conformal Field Theory' by Slava Rychkov states our goal: to find the finite form of the conformal transformations by integrating their infinitesimal versions.

Read the paragraph starting 'Let us integrate the infinitesimal forms...' until the end of the paragraph. This will frame our classification task for this lesson.

Now, let's systematically derive and classify these transformations. We will primarily follow the clear and structured exposition in the notes "Conformal field theory for particle physicists" by Marco Meineri.

Classifying Finite Conformal Transformations

The resource below details how each infinitesimal transformation exponentiates into its finite group element.

Conformal field theory for particle physicists

The following reading explains how to derive the finite form of each of the four transformation types. The derivations for translations, rotations, and dilatations are straightforward. Pay special attention to the clever method used for the Special Conformal Transformation, which involves introducing the inversion transformation.

Please read Section 2.3, 'Finite transformations' (pages 9-10). Focus on understanding how each transformation type is exponentiated. For the SCT, grasp the definition of the inversion transformation I and how the SCT can be constructed as a composition I ∘ T ∘ I.

Let's summarize the key results from the reading. A general finite conformal transformation is a composition of the following four types:

1. Translations

These are generated by $P_{μ} = i \partial_{μ}$ . Exponentiating a constant infinitesimal shift $x^{μ} \to x^{μ} + ϵ^{μ}$ gives the familiar finite translation:
$x^{' μ} = x^{μ} + a^{μ}$
where $a^{μ}$ is a constant vector.

2. Rotations and Lorentz Transformations

These are generated by $M_{μν}$ . Exponentiation gives a finite rotation (in Euclidean space) or a Lorentz transformation (in Minkowski space):
$x^{' μ} = Λ^{μ}_{ν} x^{ν}$
where $Λ$ is an element of the group $S O (d)$ or $S O (1, d - 1)$ , respectively.

3. Dilatations (Scale Transformations)

Generated by $D = i x^{μ} \partial_{μ}$ , these transformations rescale the coordinates. The finite form is:
$x^{' μ} = λ x^{μ}$
where $λ$ is a positive real number.

4. Special Conformal Transformations (SCTs)

This is the most interesting case. As the reading explains, a finite SCT cannot be found by simple exponentiation. The key insight is to introduce the inversion transformation:
$I (x^{μ}) = \frac{x ^{μ}}{x ^{2}}$
where $x^{2} = x^{μ} x_{μ}$ . An inversion is itself a conformal transformation (though a discrete one, not connected to the identity), which maps the origin to infinity and vice-versa.

A finite SCT with parameter vector $b^{μ}$ can be constructed as a sequence of an inversion, a translation by $- b^{μ}$ , and another inversion:
$K_{b} (x) = I \circ T_{- b} \circ I (x)$
Working through this composition, as you saw in the reading, yields the finite form:
$x^{' μ} = \frac{x ^{μ} - b ^{μ} x ^{2}}{1 - 2 ( b \cdot x ) + b ^{2} x ^{2}}$

This transformation has some peculiar properties. It keeps the origin fixed ( $x = 0 ⟹ x^{'} = 0$ ), but it can map finite points to infinity (any $x$ for which the denominator is zero) and maps the point at infinity to a finite location ( $x \to \infty ⟹ x^{'} \to - b^{μ} / b^{2}$ ).

This structure strongly suggests that the natural setting for the conformal group is not $R^{d}$ itself, but a compactified version of it where the "point at infinity" is treated on an equal footing with all other points.

Special Conformal Transformation Illustration — This image provides a 2D visualization of a special conformal transformation, $z' = z / (2 - z)$, where $z = x_1 + i x_2$. Notice how the grid of straight lines on the left is mapped to a grid of circles and circular arcs on the right. This illustrates a key geometric property of conformal transformations: they map lines and circles to other lines and circles, while preserving the angles of intersection.

Test your understanding!

Using the formula for a finite Special Conformal Transformation,
$x^{' μ} = \frac{x ^{μ} - b ^{μ} x ^{2}}{1 - 2 ( b \cdot x ) + b ^{2} x ^{2}}$
show that a point at $x^{μ} = b^{μ} / b^{2}$ is mapped to infinity.

Show answer

We simply need to evaluate the denominator at $x^{μ} = b^{μ} / b^{2}$ .
First, let's find the relevant scalar products:

$x^{2} = x^{μ} x_{μ} = \frac{b ^{μ}}{b ^{2}} \frac{b _{μ}}{b ^{2}} = \frac{b ^{2}}{( b ^{2} ) ^{2}} = \frac{1}{b ^{2}}$
$b \cdot x = b^{μ} x_{μ} = b^{μ} \frac{b _{μ}}{b ^{2}} = \frac{b ^{2}}{b ^{2}} = 1$

Now, substitute these into the denominator:
$1 - 2 (b \cdot x) + b^{2} x^{2} = 1 - 2 (1) + b^{2} (\frac{1}{b ^{2}}) = 1 - 2 + 1 = 0$
Since the denominator is zero, the transformation is singular at this point, and $x^{' μ}$ is undefined, which we interpret as the point being mapped to infinity.

Conclusion

In this lesson, we have successfully classified the four types of finite conformal transformations that, together, constitute the conformal group in $d > 2$ dimensions.

Key Takeaways:

The conformal group is generated by four types of finite transformations: translations, rotations/Lorentz transformations, dilatations, and special conformal transformations (SCTs).
The first three have simple, familiar forms.
SCTs have a more complex structure and are best understood as a composition: an inversion, followed by a translation, followed by another inversion ( $I \circ T \circ I$ ).
The finite form of an SCT is $x^{' μ} = (x^{μ} - b^{μ} x^{2}) / (1 - 2 (b \cdot x) + b^{2} x^{2})$ .
The singular nature of SCTs, which interchange finite and infinite points, suggests that the conformal group acts most naturally on a compactified spacetime.

This last point provides the perfect motivation for our next lesson. We will see how this idea of including the "point at infinity" can be made mathematically precise. We will construct the conformal group as the special orthogonal group $S O (d + 1, 1)$ , which acts linearly on a higher-dimensional space. This will unify all four transformation types into a single, elegant geometric picture.

Recommended resources