Unit 4: Parametric Surfaces & The First Fundamental Form
Foundations of two-dimensional surface theory in Euclidean 3-space: smooth coordinate patches, regularity conditions, tangent planes and surface normals, the First Fundamental Form (surface metric) $I = E du^2 + 2F dudv + G dv^2$, positive definiteness, arc-length of surface curves, angles between tangent vectors, orthogonal coordinates, and the intrinsic surface area element.
ยง4.1 Parametric Surfaces, Coordinate Patches & Regular Points
1. Vector Parametrization of Surfaces in $\mathbb{R}^3$
Just as a curve is described by a single scalar parameter, a surface in $\mathbb{R}^3$ is locally parametrized by two independent real parameters $(u, v)$.
Definition 4.1 (Parametrized Surface Patch): Let $U \subseteq \mathbb{R}^2$ be an open connected domain in the $uv$-plane. A parametrized surface (or local coordinate patch) is a smooth vector-valued function:
The set of points $S = \mathbf{r}(U) \subset \mathbb{R}^3$ is the trace or image of the surface patch.
2. Partial Derivatives and Coordinate Curves
For a fixed $v = v_0$, the mapping $u \mapsto \mathbf{r}(u, v_0)$ traces a curve on $S$ called the $u$-coordinate curve (or $u$-parameter curve). Similarly, for a fixed $u = u_0$, $v \mapsto \mathbf{r}(u_0, v)$ traces a $v$-coordinate curve.
The partial derivative vectors are tangent to these coordinate curves:
3. Regularity and the Tangent Plane
Definition 4.2 (Regular Point of a Surface): A point $p = \mathbf{r}(u_0, v_0)$ is called a regular point of the surface patch if the partial derivative vectors $\mathbf{r}_u$ and $\mathbf{r}_v$ are linearly independent at $(u_0, v_0)$:
Equivalently, the Jacobian matrix $J = \begin{pmatrix} x_u & x_v \\ y_u & y_v \\ z_u & z_v \end{pmatrix}$ has maximal rank $2$. A surface patch is regular if every point in $U$ is regular.
Definition 4.3 (Tangent Plane $T_p S$ and Unit Normal $\mathbf{n}$): At any regular point $p \in S$, the vectors $\mathbf{r}_u$ and $\mathbf{r}_v$ span a two-dimensional vector subspace of $\mathbb{R}^3$ called the tangent plane to $S$ at $p$, denoted $T_p S$:
The unit normal vector $\mathbf{n}(u, v)$ to the surface at $p$ is defined by:
By construction, $\mathbf{n} \cdot \mathbf{r}_u = 0$, $\mathbf{n} \cdot \mathbf{r}_v = 0$, and $\|\mathbf{n}\| = 1$.
The Cartesian equation of the tangent plane passing through $\mathbf{r}(u_0, v_0)$ is:
where $\mathbf{R} = (X, Y, Z)^T$ is an arbitrary point on the plane.
ยง4.2 The First Fundamental Form: Metric Coefficients E, F, G & Positive Definiteness
1. Differential of the Position Vector
Let $\mathbf{r}: U \to \mathbb{R}^3$ be a regular surface. Consider an infinitesimal displacement on the parameter domain $d\mathbf{u} = (du, dv)^T$. The corresponding infinitesimal displacement vector on the surface in $\mathbb{R}^3$ is given by the differential:
The square of the infinitesimal Euclidean distance between $\mathbf{r}(u, v)$ and $\mathbf{r}(u + du, v + dv)$ is:
Expanding the dot product bilinearly:
2. Definition of the First Fundamental Form
Definition 4.4 (First Fundamental Form): The First Fundamental Form of a surface $S$, denoted by $I$ or $I_p$, is the quadratic form on the tangent space $T_p S$ induced by the Euclidean metric of $\mathbb{R}^3$:
where the metric coefficients (Gauss coefficients) are:
In matrix notation, for a tangent vector $\mathbf{w} = \lambda \mathbf{r}_u + \mu \mathbf{r}_v \in T_p S$, represented in coordinates by $\mathbf{\xi} = \begin{pmatrix} \lambda \\ \mu \end{pmatrix}$:
where $g = \begin{pmatrix} E & F \\ F & G \end{pmatrix}$ is the metric tensor matrix.
3. Positive Definiteness and the Metric Determinant
Theorem 4.1 (Positive Definiteness of the First Fundamental Form): At every regular point of a surface, the First Fundamental Form is strictly positive definite:
Moreover, the determinant of the metric tensor satisfies:
Proof: By Lagrange's vector identity for any two vectors $\mathbf{a}, \mathbf{b} \in \mathbb{R}^3$:
Setting $\mathbf{a} = \mathbf{r}_u$ and $\mathbf{b} = \mathbf{r}_v$:
Because the point is regular, $\mathbf{r}_u \times \mathbf{r}_v \ne \mathbf{0}$, which implies:
Next, consider $I(du, dv) = E du^2 + 2F dudv + G dv^2$. Since $\mathbf{r}_u \ne \mathbf{0}$, $E = \|\mathbf{r}_u\|^2 > 0$. We complete the square:
Both coefficients $E > 0$ and $\frac{EG - F^2}{E} > 0$ are strictly positive. Thus $I(du, dv) \ge 0$, with equality holding if and only if $dv = 0$ and $du + \frac{F}{E} dv = 0 \implies du = 0$. Hence, $I$ is strictly positive definite. $\blacksquare$
ยง4.3 Arc-Length of Surface Curves, Isometries & Conformal Mappings
1. Arc-Length of Curves Lying on a Surface
Let $C$ be a smooth curve lying entirely on the surface $S$, defined parametrically by $u = u(t), v = v(t)$ for $t \in [a, b]$. The position vector of the curve in $\mathbb{R}^3$ is:
Applying the chain rule, the velocity vector is:
The speed of the curve is the norm of the velocity vector:
Theorem 4.2 (Arc-Length Formula on Surfaces): The arc-length $L$ of the curve $C$ from $t = a$ to $t = b$ is given intrinsically by:
This fundamental result demonstrates that distance along curves on a surface can be computed solely from the metric coefficients $E(u, v), F(u, v), G(u, v)$ without knowing how the surface is embedded in 3D space!
2. Isometries and Intrinsic Geometry
Definition 4.5 (Local Isometry): A diffeomorphism $\phi: S_1 \to S_2$ between two surfaces is called a local isometry if it preserves the lengths of all curves. Equivalently, for any point $p \in S_1$ and tangent vectors $\mathbf{w}_1, \mathbf{w}_2 \in T_p S_1$:
If $S_1$ and $S_2$ are parametrized by coordinates $(u, v)$ such that their metric coefficients satisfy:
then the mapping is an isometry.
Surfaces related by an isometry share all intrinsic geometric properties (such as arc-length, angles, and Gaussian curvature), even though their spatial embeddings may look completely different (e.g., a plane sheet rolling into a cylinder).
3. Conformal Mappings and Orthogonal Coordinates
Definition 4.6 (Conformal Mapping): A diffeomorphism $\phi: S_1 \to S_2$ is conformal (angle-preserving) if there exists a smooth positive function $\lambda(u, v) > 0$ (the conformal factor) such that:
Definition 4.7 (Orthogonal Coordinate Systems): A coordinate patch $(u, v)$ is called orthogonal if the coordinate curves intersect at right angles everywhere on $U$:
When $F \equiv 0$, the First Fundamental Form simplifies to:
In addition, if $E = G = \lambda^2(u, v)$ and $F = 0$, the coordinates are called isothermal (or conformal):
ยง4.4 Angles Between Curves on Surfaces & Direction Fields
1. Angle Between Tangent Vectors on a Surface
Let $p \in S$ be a regular point, and let $\mathbf{w}_1, \mathbf{w}_2 \in T_p S$ be two non-zero tangent vectors. In local coordinates:
The inner product of these tangent vectors is:
Expanding:
Theorem 4.3 (Angle Between Directions on a Surface): The angle $\theta \in [0, \pi]$ between the directions $d\mathbf{r} = (du, dv)$ and $\delta\mathbf{r} = (\delta u, \delta v)$ is given by:
In particular, the two directions are orthogonal ($\theta = \pi/2$) if and only if:
2. Angle of a Curve with Coordinate Curves
For the $u$-coordinate curve ($dv = 0, du > 0$), the tangent vector is $\mathbf{r}_u$. The angle $\alpha$ between an arbitrary curve direction $(du, dv)$ and the $u$-coordinate curve satisfies:
For an orthogonal coordinate system ($F = 0$):
3. Orthogonal Trajectories of a Family of Curves
Suppose a family of curves on $S$ is defined by a differential equation:
To find the family of orthogonal trajectories $(\delta u, \delta v)$, we apply the orthogonality condition:
Dividing by $du \, \delta u$:
Substituting $\frac{dv}{du} = -\frac{P}{Q}$:
Multiplying by $Q$:
This first-order ODE governs the orthogonal trajectories across the surface patch.
ยง4.5 Surface Area Element, Jacobians & Integrals on Surfaces
1. Infinitesimal Area Element on a Surface
Consider an infinitesimal curvilinear parallelogram on the surface bounded by the vectors $\mathbf{r}_u \, du$ and $\mathbf{r}_v \, dv$. The area $dA$ of this infinitesimal parallelogram in $\mathbb{R}^3$ is the magnitude of their cross product:
Using Lagrange's identity from Section 4.2:
Definition 4.8 (Surface Area Element): The intrinsic area element (or Riemannian volume element $d\sigma$) on a regular surface patch is:
The positive quantity $W = \sqrt{EG - F^2} > 0$ is the Gram determinant factor.
2. Surface Integral and Total Area
Definition 4.9 (Surface Area): Let $S = \mathbf{r}(U)$ be a regular surface patch where $U \subset \mathbb{R}^2$ is bounded. The surface area of $S$ is defined by the double integral:
For a scalar function $f: S \to \mathbb{R}$, the surface integral of $f$ over $S$ is:
3. Invariance Under Reparametrization
Theorem 4.4 (Invariance of Surface Area): The surface area $\operatorname{Area}(S)$ is invariant under orientation-preserving or orientation-reversing smooth reparametrizations.
Proof: Let $(\bar{u}, \bar{v})$ be an alternative coordinate system related to $(u, v)$ by a diffeomorphism $\Phi: \bar{U} \to U$, $(u, v) = \Phi(\bar{u}, \bar{v})$. By the multi-variable chain rule:
Taking the cross product:
Since $\mathbf{r}_u \times \mathbf{r}_u = \mathbf{0}$ and $\mathbf{r}_v \times \mathbf{r}_v = \mathbf{0}$, this simplifies to:
Taking norms on both sides:
By the multivariable change-of-variables theorem for double integrals:
Thus the surface area is completely independent of the choice of coordinate patch! $\blacksquare$
Step-by-step rigorous derivations with unskipped proofs, categorized into Foundational Concepts, Advanced Structural Analysis, and Honors / Proof Challenge tiers.
Consider the 2-sphere of radius $R > 0$ parametrized by spherical angles $(\theta, \phi)$:
- Compute the partial derivatives $\mathbf{r}_\theta$ and $\mathbf{r}_\phi$.
- Calculate the metric coefficients $E, F, G$ and write down the First Fundamental Form $I$.
- Compute the area element $dA = \sqrt{EG - F^2} \, d\theta \, d\phi$ and evaluate the total surface area of the sphere.
1. Partial Derivatives
Differentiating $\mathbf{r}(\theta, \phi)$ with respect to $\theta$:
Differentiating with respect to $\phi$:
2. Metric Coefficients and First Fundamental Form
Computing $E = \mathbf{r}_\theta \cdot \mathbf{r}_\theta$:
Computing $F = \mathbf{r}_\theta \cdot \mathbf{r}_\phi$:
Since $F = 0$, the spherical coordinate lines are orthogonal everywhere!
Computing $G = \mathbf{r}_\phi \cdot \mathbf{r}_\phi$:
Thus, the First Fundamental Form of the sphere is:
3. Surface Area Element and Total Area
The metric determinant is:
Since $\theta \in (0, \pi)$, $\sin\theta > 0$, so:
The area element is:
Evaluating the total surface area:
The Catenoid $S_{\text{cat}}$ and the Helicoid $S_{\text{hel}}$ are parametrized by:
- Compute the metric coefficients $E, F, G$ of the Catenoid.
- Compute the metric coefficients $\bar{E}, \bar{F}, \bar{G}$ of the Helicoid.
- Show that under the coordinate transformation $\bar{u} = \sinh u$ and $\bar{v} = v$, the two First Fundamental Forms coincide, proving that the Catenoid and Helicoid are locally isometric.
1. Metric Coefficients of the Catenoid
For $\mathbf{r}_{\text{cat}}(u, v) = (\cosh u \cos v, \cosh u \sin v, u)^T$:
Now compute the dot products:
Thus, the First Fundamental Form of the catenoid is:
2. Metric Coefficients of the Helicoid
For $\mathbf{r}_{\text{hel}}(\bar{u}, \bar{v}) = (\bar{u} \cos \bar{v}, \bar{u} \sin \bar{v}, \bar{v})^T$:
Now compute the metric coefficients:
Thus, the First Fundamental Form of the helicoid is:
3. Coordinate Transformation and Local Isometry
Let $\bar{u} = \sinh u$ and $\bar{v} = v$. Then:
Substitute these into $I_{\text{hel}}$:
Therefore:
Since the First Fundamental Forms match identically under this smooth bijection, the Catenoid and Helicoid are locally isometric! $\blacksquare$
A regular coordinate system $(u, v)$ on a surface $S$ is called isothermal (or conformal) if the First Fundamental Form takes the form:
where $\lambda(u, v) > 0$ is a smooth non-vanishing function.
- Let $\mathbf{w}_1 = \mathbf{r}_u \, du_1 + \mathbf{r}_v \, dv_1$ and $\mathbf{w}_2 = \mathbf{r}_u \, du_2 + \mathbf{rv} \, dv_2$ be two tangent vectors at $p = \mathbf{r}(u, v)$. Prove that the geometric angle $\theta \in [0, \pi]$ between $\mathbf{w}_1$ and $\mathbf{w}_2$ in $\mathbb{R}^3$ equals the Euclidean angle $\alpha$ between the parameter displacement vectors $\mathbf{v}_1 = (du_1, dv_1)^T$ and $\mathbf{v}_2 = (du_2, dv_2)^T$ in the parameter plane $\mathbb{R}^2$.
- Conclude that the parameter mapping $\mathbf{r}: U \subset \mathbb{R}^2 \to S \subset \mathbb{R}^3$ is a conformal map, preserving all angles and shapes of infinitesimal figures.
1. Inner Product and Norms Under Isothermal Coordinates
Assume the metric coefficients satisfy:
with $\lambda(u, v) > 0$. The inner product of the tangent vectors $\mathbf{w}_1, \mathbf{w}_2 \in T_p S$ is given by the bilinear form:
Substituting $E = G = \lambda^2$ and $F = 0$:
Notice that:
where $\mathbf{v}_1 = (du_1, dv_1)^T$ and $\mathbf{v}_2 = (du_2, dv_2)^T$ are vectors in the flat parameter plane $\mathbb{R}^2$. Thus:
Next, compute the norms of $\mathbf{w}_1$ and $\mathbf{w}_2$:
2. Angle Equivalence
The cosine of the 3D angle $\theta$ between $\mathbf{w}_1$ and $\mathbf{w}_2$ on the surface is:
Substituting the expressions derived above:
The right-hand side is precisely the definition of $\cos \alpha$, where $\alpha$ is the standard Euclidean angle between $\mathbf{v}_1$ and $\mathbf{v}_2$ in $\mathbb{R}^2$:
Since both $\theta, \alpha \in [0, \pi]$, we have:
3. Conclusion: Conformal Mapping
Because $\theta = \alpha$ for any two arbitrary non-zero tangent directions at every point $p \in S$:
- Every angle between curves on the surface is identical to the angle between their preimage curves in the $uv$-plane.
- The coordinate mapping $\mathbf{r}: U \to S$ is a conformal map (angle-preserving).
- Infinitesimal circles in the parameter plane are mapped to infinitesimal circles on the surface, scaled uniformly in all directions by the factor $\lambda(u, v)$ without angular shearing. $\blacksquare$