Problem 30 - Calculus On Manifolds: A Modern Approach To Classical Theorems Of Advanced Calculus

Question

If \(\omega\) is a 1 -form on \(\mathbf{R}^{2}-0\) such that \(d \omega=0\), prove that $$ \omega=\lambda d \theta+d g $$ for some \(\lambda \in \mathbf{R}\) and \(g: \mathbf{R}^{2}-0 \rightarrow \mathbf{R}\). Hint: If $$ c_{R .1} *(\omega)=\lambda_{R} d x+d\left(g_{R}\right) $$ show that all numbers \(\lambda_{R}\) have the same value \(\lambda\).

Step-by-Step Solution

Answer

\(\omega = \lambda d \theta + d g\) for some \(\lambda \in \mathbf{R}\) and function \(g: \mathbf{R}^2 - 0 \rightarrow \mathbf{R}\).

Step 1: Verify the Closed Form

Given that \(d\omega = 0\), \(\omega\) is closed. By Poincaré's lemma, in a star-shaped domain, every closed form is also exact. \( \mathbf{R}^{2} - 0 \) is not simply connected, but for local considerations, let’s proceed.

Step 2: Utilize Polar Coordinates

Transform to polar coordinates \( (r, \theta) \). Here, \( d\theta \) is a well-defined 1-form except at the origin.

Step 3: Express \( \omega \) in Polar Coordinates

Assume \( \omega = f(r, \theta) dr + g(r, \theta) d\theta \). Since \( d\omega = 0 \), equate the components of \ (d\omega) \ to zero. This gives \( \frac{\partial f}{\partial \theta} = \frac{\partial g}{\partial r} \).

Step 4: Integrate the Condition

From the above relation \( \frac{\partial g}{\partial r} = \frac{\partial f}{\partial \theta} \), integrate w.r.t. \(r \) to get \( g(r, \theta) = \int f_r(r, \theta) dr + h(\theta) \). Then \( \omega = f(r, \theta) dr + \left( \int f_r(r, \theta) dr + h(\theta) \right) d\theta \).

Step 5: Consider the Form on a Circle

Define the integral curve \ c_{R, 1} = (R, \theta) \ for fixed \ R \ and variable \ \theta \ within boundary (0, 2π). At \ r = R \, \( \omega = \lambda_{R} d\theta + d(g_{R}) \) simplifies due to periodic \ \theta \ dependency.

Step 6: Show Constants Across Different Radii

For different \ R \, constants \(\lambda_{R}\) must be equal by consistency under integral evaluation, proving that \ \lambda \ is unique where \ \lambda = \lambda_R \ for all \ R.

Step 7: Conclusion

Combine information to state \( \omega = \lambda d \theta + d g \) where \ \lambda \ is a constant and \( g: \mathbf{R}^2 - 0 \rightarrow \mathbf{R} \) is some smooth function.

Key Concepts

Poincaré's lemma

Poincaré's lemma is a fundamental result in differential geometry. It states that if \(\omega\) is a closed p-form (meaning \(d\omega = 0\)) on a star-shaped domain, then \(\omega\) is also exact. This means there exists some (p-1)-form \(\alpha\) such that \(\omega = d\alpha\). A domain is star-shaped if there is a point such that every other point can be connected to it by a straight line entirely within the domain. For our problem, we consider \(\mathbf{R}^{2} - 0\), which is not star-shaped because excluding the origin makes the domain non-simply connected. However, locally, Poincaré's lemma gives insight that closed forms resemble exact forms in small neighborhoods.

exact forms

An exact form is a differential form that can be expressed as the exterior derivative of another form. For example, in the case of 1-forms, \(\omega\) is exact if there exists a 0-form (or function) \(f \) such that \(\omega = df\). Given \(d\omega = 0\), by Poincaré's lemma in appropriate domains, we can express \(\omega\) as \(df\). This is useful because if we can show that \(\omega\) is exact, many integrals become straightforward through the evaluation of functions rather than complex forms.

polar coordinates

Polar coordinates (\r, \(\theta \)) express points on the plane with a radius and angle relative to the origin rather than x and y coordinates. This is particularly useful for problems with circular symmetry. The conversion from Cartesian coordinates (x, y) to polar coordinates is given by:

• \(r = \sqrt{x^2 + y^2}\)
• \(\theta = \arctan(y/x)\)

In polar coordinates, the 1-form \(d\theta\) plays a vital role, as it measures angular change and is defined except at \(r = 0\). For a 1-form \(\omega\), expressing it in polar coordinates often simplifies the problem by focusing on radial and angular components separately.

closed forms

A differential form \(\omega\) is called closed if \(d\omega = 0\). This condition implies that \(\omega\) has no 'curl', similar to how a conservative vector field has zero curl in vector calculus. Closed forms are significant because they provide essential information about the topology of the underlying space. In the given exercise, knowing \(d\omega = 0\) allows us to deduce properties about \(\omega\), such as expressing it in terms of other differential forms. In essence, while closed forms are not necessarily exact in complex domains, in simple connected regions or locally, they hold the properties of exact forms.