Exponential Map: Connecting Lie Algebras to Lie Groups

Welcome back. In our previous lesson, we finalized the abstract definition of a Lie algebra by introducing the Jacobi identity, the crucial constraint that arises from the associativity of the underlying Lie group. We now possess a complete picture of the local structure of a Lie group near its identity element.

This lesson addresses the natural next question: how do we get back from this local, linear structure of the Lie algebra to the global, curved structure of the Lie group? We need a bridge that takes us from the infinitesimal (the algebra) to the finite (the group). This bridge is the exponential map. We will see that this map allows us to reconstruct finite group transformations by "integrating" the infinitesimal transformations represented by the algebra's generators.

1. From Tangent Vectors to Group Paths: One-Parameter Subgroups

Imagine the Lie algebra $g$ as the tangent space to the Lie group $G$ at the identity element $e$ . An element $X \in g$ is like a velocity vector, pointing in a specific direction from the identity. If we "flow" along this direction for a certain "time" $t$ , we trace out a path on the group manifold. This path, $γ (t)$ , starts at the identity, $γ (0) = e$ , and its initial velocity is the vector $X$ we started with.

This path has a special property inherited from the group structure: moving for time $s$ and then for time $t$ is the same as moving for time $s + t$ . In the language of group theory, this is $γ (s) γ (t) = γ (s + t)$ . This makes the map $γ : (R, +) \to G$ a Lie group homomorphism. Such a path is called a one-parameter subgroup of $G$ .

For every element $X$ in the Lie algebra, there exists a unique one-parameter subgroup $γ (t)$ such that $\frac{d γ}{d t}_{t = 0} = X$ . This establishes a direct link between an algebra element and a curve in the group.

For the matrix Lie groups that are central to physics, this relationship becomes wonderfully concrete. The differential equation for the path is $γ ˙ (t) = X γ (t)$ with the initial condition $γ (0) = I$ . As you know from your studies of differential equations, the unique solution to this is the matrix exponential:

γ (t) = e^{tX} = k = 0 \sum \infty \frac{( tX ) ^{k}}{k !}

2. The Exponential Map

This connection allows us to define the map that takes any algebra element $X$ and maps it to a specific group element. We simply follow the one-parameter subgroup generated by $X$ for a "time" of $t = 1$ .

Definition: The Exponential Map
The exponential map, $exp : g \to G$ , is defined by

exp (X) = γ_{X} (1)

where $γ_{X} (t)$ is the unique one-parameter subgroup in $G$ whose tangent vector at the identity is $X$ . For matrix Lie groups, this definition simplifies to the familiar matrix exponential:

exp (X) = e^{X}

The following reading provides a concise but rigorous definition of the matrix exponential and its most important properties, then generalizes to the formal definition of the exponential map via one-parameter subgroups.

Lie Groups and Lie Algebras for Physicists

This document from the University of Vienna provides an excellent introduction to the exponential map.

First, read the section on The Matrix Exponential. Focus on the definition of the series and the list of properties in Proposition 11. Pay special attention to property 4, which is the heart of non-commutativity. Next, read the section on One-parameter subgroups and the exponential map. This formalizes the intuition we just discussed. Read from the beginning of the section up to the formal definition of the map, Definition 13. Note how this connects back to the matrix exponential.

The key properties from the reading are worth summarizing:

$exp (0) = I$
$exp (X)$ is always invertible, with $(exp X)^{- 1} = exp (- X)$ .
$det (exp X) = e^{Tr (X)}$ . This is a vital identity. It immediately shows why the Lie algebra for $S L (n)$ (special linear group, $det = 1$ ) consists of traceless matrices. If $Tr (X) = 0$ , then $det (e^{X}) = e^{0} = 1$ .
If $[X, Y] = 0$ , then $exp (X + Y) = exp (X) exp (Y)$ . The failure of this identity in the general non-commuting case is precisely what makes Lie group theory so rich.

3. Geometric Intuition

Geometrically, the exponential map takes the flat vector space $g$ (the tangent plane at the identity) and wraps it onto the curved manifold of the group $G$ . A straight line through the origin in $g$ , given by $tX$ , is mapped to a geodesic on $G$ , the one-parameter subgroup $γ_{X} (t)$ .

A depiction of the exponential map. A tangent vector $v$ in the tangent plane $T_x\mathcal{M}$ is mapped to a point $q$ on the manifold $\mathcal{M}$ by following a geodesic (the red curve) for a unit parameter distance. The exponential map provides a local chart that relates the linear algebra of the tangent space to the non-linear geometry of the manifold.

4. Application: From Group Elements to Generators

The exponential map is not just an abstract concept; it is a practical tool for finding the generators of a group. The process involves reversing the map: we start with a known group element parameterized by an angle, say $R (θ)$ , and find the matrix $X$ such that $R (θ) = e^{θ X}$ . The matrix $X$ is the generator of that transformation.

The following video provides an excellent walkthrough of this process for the rotation group SO(3).

Particle Physics Lecture 7: Lie Groups, Lie Algebras and an SO(3) Case Study

In his "Particle Physics Lecture 7", Alex Flournoy gives a clear, step-by-step derivation of a generator for SO(3).

Please watch the segment from the start of the example. The lecturer starts with the familiar matrix for a rotation about the y-z plane, expands it as a Taylor series in the angle \theta, and compares it to the series expansion of e^{i\theta g} to identify the generator g.

Let's recap the method shown in the video, using the physics convention where generators are often made Hermitian by pulling out a factor of $i$ . A finite rotation by an angle $ϕ$ around the x-axis is given by:

R_{x} (ϕ) = 100 0 cos ϕ sin ϕ 0 - sin ϕ cos ϕ

We postulate that this can be written as $R_{x} (ϕ) = exp (i ϕ J_{x})$ for some generator $J_{x}$ . Expanding both sides for small $ϕ$ :

R_{x} (ϕ) \approx 100 01 ϕ 0 - ϕ 1 = I + ϕ 000001 0 - 1 0 + O (ϕ^{2})

exp (i ϕ J_{x}) \approx I + i ϕ J_{x} + O (ϕ^{2})

Comparing the terms linear in $ϕ$ , we can identify $i J_{x}$ . This gives us the generator:

J_{x} = - i 000001 0 - 1 0 = 000 00 - i 0 i 0

This is one of the familiar angular momentum matrices from quantum mechanics, up to a factor of $ℏ$ . This procedure connects the abstract algebra directly to the physical operators you've worked with.

You can verify that exponentiating this generator $J_{x}$ recovers the full rotation matrix. The powers of $J_{x}$ have a cyclic structure that collapses the infinite exponential series into sines and cosines, exactly as shown for a similar generator in the Steinacker PDF (equation 25).

5. Reconstructing the Group Law: The Baker-Campbell-Hausdorff Formula

We now know how to go from an algebra element $X$ to a group element $exp (X)$ . But how does the algebra structure, specifically the Lie bracket, determine the group's multiplication law? If we multiply two group elements, $exp (X)$ and $exp (Y)$ , the result must be another group element, which we can write as $exp (Z)$ . What is $Z$ ?

If $X$ and $Y$ commute, the answer is simple: $Z = X + Y$ . But in the non-abelian case, the answer is given by the Baker-Campbell-Hausdorff (BCH) formula:

Z = lo g (e^{X} e^{Y}) = X + Y + \frac{1}{2} [X, Y] + \frac{1}{12} ([X, [X, Y]] - [Y, [X, Y]]) + \dots

The full formula is an infinite series of nested Lie brackets.

This formula is profound. It demonstrates that the entire group multiplication law in a neighborhood of the identity is completely determined by the Lie algebra and its bracket operation. All the information about how to compose finite transformations is encoded in the commutation relations of the generators.

Mathematical Structure of Quantum Theory Lecture 4: Lie Groups and Lie Algebras

The video "Mathematical Structure of Quantum Theory Lecture 4" introduces the exponential map and the Baker-Campbell-Hausdorff formula.

Watch the segment from the exponential map definition. The lecturer defines the multiplication of two exponentials and introduces the BCH formula as the solution for Z. The subsequent part of the video delves into a detailed proof, which you are not required to follow, but it's available should your curiosity lead you there.

The BCH formula is the final piece of the puzzle connecting the local algebraic structure to the local group structure. It guarantees that if you have a valid Lie algebra, you can (at least locally) reconstruct the corresponding Lie group's multiplication.

Conclusion

In this lesson, we have constructed the essential bridge from the Lie algebra back to the Lie group. The exponential map provides a concrete way to generate finite transformations from the algebra's infinitesimal generators.

Key Takeaways:

The exponential map, $exp : g \to G$ , connects the Lie algebra to the Lie group. For matrix groups, it is the matrix exponential, $e^{X}$ .
An algebra element $X$ generates a one-parameter subgroup $γ (t) = e^{tX}$ , which is a curve of finite transformations.
By postulating $G (θ) = e^{θ X}$ and comparing Taylor series expansions, we can derive the generator $X$ for a given parameterized group element $G (θ)$ .
The Baker-Campbell-Hausdorff (BCH) formula shows how the Lie bracket determines the group multiplication law near the identity, expressing the product $e^{X} e^{Y}$ as the exponential of a series of nested commutators.

We now have a powerful toolkit for analyzing continuous symmetries. In the next lesson, we will apply these tools in a detailed case study of the groups SO(3) and SU(2). We will formally derive their Lie algebras, establish the deep connection between them, and see why this relationship is at the very foundation of the quantum theory of spin.

Recommended resources