Building Affine Lie Algebras: Loop Algebras to Central Extensions

Hello! Welcome to the sixth module of our course.

In our previous lessons on 2D CFT, we focused extensively on the consequences of conformal symmetry, which is generated by the stress-energy tensor $T (z)$ . This led us to the Virasoro algebra, its representations (Verma modules, characters), and the powerful constraints of modular invariance.

However, many interesting CFTs possess additional symmetries beyond the conformal group. These symmetries, described by a Lie group $G$ , give rise to conserved currents $j^{a} (z)$ via Noether's theorem. In 2D, these currents are holomorphic fields of conformal weight (1,0) and generate a so-called current algebra. Today, we will explore the structure of this algebra. You will learn that the quantum nature of the theory introduces a crucial modification, a "central extension," to the classical algebra of symmetries.

Your learning objective is to: Construct affine Lie algebras (Kac-Moody algebras) as central extensions of loop algebras.

This lesson marks a shift from the purely conformal structure to the richer algebraic framework that underpins theories like the Wess-Zumino-Witten (WZW) models, which are central to string theory and condensed matter physics.

1. From Symmetries to Current Algebras in CFT

Let's begin from a familiar place in quantum field theory. A continuous global symmetry with Lie group $G$ implies the existence of a set of conserved Noether currents $j_{μ}^{a} (x)$ , where the index $a$ runs over the dimension of the Lie algebra $g$ . In a 2D CFT, it's natural to work in complex coordinates $z = x^{0} + i x^{1}$ and $z ˉ = x^{0} - i x^{1}$ . The conservation equation $\partial^{μ} j_{μ}^{a} = 0$ splits into two independent equations for the holomorphic and anti-holomorphic components:

\partial ˉ j^{a} (z) = 0 and \partial j ˉ^{a} (z ˉ) = 0

This means we have a set of holomorphic fields $j^{a} (z)$ , which have conformal dimension $(h, h ˉ) = (1, 0)$ . The core idea is that the algebraic relations between these currents are captured by their Operator Product Expansion (OPE).

The general OPE of two such currents in a CFT takes the form:

j^{a} (z) j^{b} (w) \sim \frac{k δ ^{ab}}{( z - w ) ^{2}} + \frac{i f ^{ab c} j ^{c} ( w )}{z - w} + regular terms

Let's analyze the two singular terms:

The Simple Pole: The term with the $1/ (z - w)$ singularity contains the structure constants $f^{ab c}$ of the Lie algebra $g$ . This term reflects the classical symmetry algebra. If this were the only term, the currents would form a representation of $g$ at each point $z$ .
The Double Pole: The term with the $1/ (z - w)^{2}$ singularity is a pure quantum effect. It is a c-number (i.e., not an operator) and its coefficient $k$ , called the level, is a crucial parameter of the theory. Its presence will fundamentally alter the algebra of the symmetry generators.

To find the algebra, we expand the current $j^{a} (z)$ in modes, just as we did for the stress-energy tensor:

j^{a} (z) = n \in Z \sum J_{n}^{a} z^{- n - 1}

By using the standard CFT technique of taking contour integrals of the OPE, we can derive the commutation relations for these modes. The OPE above translates to:

[J_{n}^{a}, J_{m}^{b}] = i f^{ab c} J_{n + m}^{c} + k n δ^{ab} δ_{n + m, 0}

This is the affine Lie algebra, or more precisely, the untwisted affine Kac-Moody algebra $g^$ . Let's see how the physics literature arrives at this result.

TOPICS IN CONFORMAL FIELD THEORY ' * '

The following reading from Alvarez-Gaumé, Sierra, and Gomez's 'Topics in Conformal Field Theory' demonstrates how this algebra arises directly from the operator formalism of CFT.

Please read section 4.1, 'Generalities', starting from the introduction to current algebras. Follow the development from the OPE of currents (eq. 4.8) to the commutation relations for their modes (eq. 4.10 and its normalization in eq. 4.15). The key takeaway is the statement following eq. 4.16, which identifies this structure as a 'central extension of the loop algebra of G'. You can stop reading after this statement on page 51.

As the text highlights, the resulting algebra has two distinct parts. The term $i f^{ab c} J_{n + m}^{c}$ looks like the original Lie algebra $g$ , but with an extra index from the mode expansion. The second term, $k n δ^{ab} δ_{n + m, 0}$ , is something new. To understand the structure of this algebra, we need to formally define what a "loop algebra" and a "central extension" are.

2. The Mathematical Structure: Loop Algebras and Central Extensions

Your appreciation for mathematical structure will be useful here as we deconstruct the affine algebra into its constituent parts. We will see that the result derived from the physics of OPEs corresponds precisely to a well-defined mathematical construction.

a) The Loop Algebra $L (g)$

First, let's ignore the double pole term in the OPE for a moment. The algebra would then be:

[J_{n}^{a}, J_{m}^{b}] = i f^{ab c} J_{n + m}^{c}

This structure is known as a loop algebra. Why?
An element can be thought of as a pair $(T^{a}, n)$ , where $T^{a}$ is a generator of the Lie algebra $g$ and $n \in Z$ . We can represent this formally as $T^{a} \otimes t^{n}$ , where $t$ is a formal variable. The commutation relation becomes:

[T^{a} \otimes t^{n}, T^{b} \otimes t^{m}] = [T^{a}, T^{b}] \otimes t^{n + m} = i f^{ab c} T^{c} \otimes t^{n + m}

This algebra is isomorphic to the algebra of Laurent polynomials with values in $g$ , i.e., $L (g) = g \otimes C [t, t^{- 1}]$ . If we think of $t$ as $e^{i θ}$ , this is the algebra of maps from a circle $S^{1}$ (a loop) into the Lie algebra $g$ , with the bracket defined pointwise.

Affine Lie Algebras

The following resource provides the formal definition of a loop algebra. This will solidify the mathematical picture of the 'classical' part of our current algebra.

Please read Section 3, 'The Loop Algebra'. Focus on the definition of the loop algebra as the vector space of maps from a circle to a Lie algebra, and its equivalent formulation as \mathbb{C}[t, t^{-1}] \otimes_{\mathbb{C}} L.

b) Central Extensions

Now, let's reintroduce the second term from our commutator: $k n δ^{ab} δ_{n + m, 0}$ . This term is proportional to the identity matrix in the space of operators, and it commutes with all generators $J_{n}^{a}$ . Such an element is called a central element.

A central extension $L ~$ of a Lie algebra $L$ is constructed by adding a central element $K$ and modifying the bracket. We define the new space as $L ~ = L \oplus C K$ and the new bracket as:

[x + λ K, y + μ K]_{L ~} := [x, y]_{L} + ψ (x, y) K

For this new bracket to satisfy the Jacobi identity, the map $ψ : L \times L \to C$ must be a 2-cocycle. This construction should remind you of how the Virasoro algebra is built from the Witt algebra; the Virasoro algebra is the unique non-trivial central extension of the Witt algebra.

Affine Lie Algebras

Let's look at the formal definition of a central extension.

Please read Section 2, 'Central Extensions of a Lie Algebra'. Understand the definition of the extended bracket in equation (2.1) and the conditions on the function \psi(x,y) that make it a 2-cocycle.

c) The Affine Lie Algebra

We can now put everything together. The affine Lie algebra is the central extension of the loop algebra.

Start with the loop algebra $L (g) = g \otimes C [t, t^{- 1}]$ .
Extend it by a central element $K$ . The resulting space is $g^= (g \otimes C [t, t^{- 1}]) \oplus C K$ .
Define the new bracket with a specific 2-cocycle. It can be shown that for a semi-simple Lie algebra $g$ , there is essentially one unique way to do this. The 2-cocycle can be defined using the invariant Killing form $κ (,)$ on $g$ .

The resulting bracket is:

[X \otimes t^{n}, Y \otimes t^{m}] = [X, Y] \otimes t^{n + m} + κ (X, Y) n δ_{n + m, 0} K

If we choose an orthonormal basis ${T^{a}}$ for $g$ such that $κ (T^{a}, T^{b}) = δ^{ab}$ and identify the central element $K$ with the level $k$ , we recover exactly the algebra we found from the OPE in CFT:

[J_{n}^{a}, J_{m}^{b}] = i f^{ab c} J_{n + m}^{c} + k n δ^{ab} δ_{n + m, 0}

where $J_{n}^{a}$ now represents $T^{a} \otimes t^{n}$ and the central term is written as $k n δ^{ab} δ_{n + m, 0}$ (where the right-hand side is implicitly proportional to the central element $K$ ).

Test your understanding!

Consider the affine algebra $[J_{n}^{a}, J_{m}^{b}] = i f^{ab c} J_{n + m}^{c} + k n δ^{ab} δ_{n + m, 0}$ .
The zero-modes ${J_{0}^{a}}$ form a subalgebra. What is this subalgebra? How does this demonstrate the relationship between the affine algebra $g^$ and the original Lie algebra $g$ ?

Show answer

To find the subalgebra of the zero-modes, we set $n = m = 0$ in the commutation relation:

[J_{0}^{a}, J_{0}^{b}] = i f^{ab c} J_{0 + 0}^{c} + k \cdot 0 \cdot δ^{ab} δ_{0 + 0, 0}

The central extension term vanishes because $n = 0$ . The equation simplifies to:

[J_{0}^{a}, J_{0}^{b}] = i f^{ab c} J_{0}^{c}

This is precisely the commutation relation for the finite-dimensional Lie algebra $g$ . This shows that the original "global" symmetry algebra $g$ is faithfully embedded within the infinite-dimensional affine algebra as the subalgebra of its zero-modes.

Conclusion

In this lesson, we have constructed the affine Kac-Moody algebra, a cornerstone of modern theoretical physics. We took a two-pronged approach, demonstrating that this structure arises both from the physical operator product expansion of conserved currents in 2D CFT and from the rigorous mathematical construction of a central extension of a loop algebra.

Key Takeaways:

Continuous symmetries in a 2D CFT are generated by holomorphic currents $j^{a} (z)$ of spin 1.
The OPE of these currents contains two key pieces: a simple pole encoding the classical Lie algebra $g$ , and a double pole, a quantum effect characterized by the level $k$ .
The algebra of the current modes $J_{n}^{a}$ is an affine Lie algebra, which is mathematically defined as the unique non-trivial central extension of the loop algebra $L (g)$ .
The loop algebra $L (g)$ is the algebra of maps from a circle into $g$ .
The central extension introduces a new generator $K$ (the "center") that modifies the bracket with a 2-cocycle, which is physically realized as the term $k n δ^{ab} δ_{n + m, 0}$ .

Having established the fundamental algebraic structure, our next step is to study its representations. Just as we built representations of the Virasoro algebra from highest-weight states, we will now do the same for affine Lie algebras, where the concept of the level $k$ will play a critical role in classifying the possible representations.

Recommended resources