Surface Integrals
In this lesson we will study surface integrals, which are a type of integral in which we integrate a function over an arbitrary surface. Just as for line integrals, the integrand can be either a scalar or vector field.
Surface Integrals of Scalar Fields
Let’s begin by discussing surface integrals used to integrate scalar fields over a surface. Formally, the definition can be stated as follows.
Let $(\Sigma, r)$ be a regular parametrized surface and let $f : \Sigma \to \mathbb{R}$ be a continuous scalar field. Suppose that $r: D \to \Sigma$ is a parametrization that covers the surface once, except possibly along its boundary. The surface integral of the scalar field $f$ over $\Sigma$ is defined as: $$ \iint_{\Sigma} f \, d\sigma = \iint_D f (r(u,v)) \cdot \Vert r_u \times r_v \Vert \, du \, dv $$
Let’s see a practical example.
Note: the following exercises is taken from Marco Bramanti’s Esercitazioni di Analisi Matematica 2 (2012) (page 542, exercise 6.38).
Consider the torus defined by:
$$ \Sigma = \begin{cases} x = (R + r \cos \varphi) \cos \psi\\ y = (R + r \cos \varphi) \sin \psi\\ z = r \sin \varphi \end{cases}, \qquad \varphi, \psi \in [0, 2\pi] $$
- Find the area of $\Sigma$;
- Find the moment of inertia of $\Sigma$, assuming it is a homogeneous material surface of mass $M$ about the $z$-axis.
Let’s begin with the first question. We assume that $R > r > 0$, where $R$ and $r$ denote the major and the minor radii, respectively. The parameter $\psi$ describes the rotation around the $z$-axis while $\varphi$ describes the circular cross-section of the torus.
Thus, we can define the following parametrization: $$ r(\varphi, \psi) = \begin{pmatrix} (R + r \cos \varphi) \cos \psi\\ (R + r \cos \varphi) \sin \psi\\ r \sin \varphi \end{pmatrix} $$
and calculate the following partial derivatives:
$$ r_{\varphi} = \begin{pmatrix} -r \sin \varphi \cos \psi\\ -r \sin \varphi \sin \psi\\ r \cos \varphi \end{pmatrix}, \quad r_{\psi} = \begin{pmatrix} -(R + r \cos \varphi) \sin \psi\\ (R + r \cos \varphi) \cos \psi\\ 0 \end{pmatrix} $$
Then, we can determine the vector product:
$$ r_{\varphi} \times r_{\psi} = \begin{pmatrix} -r (R + r \cos \varphi) \cos \varphi \cos \psi\\ -r (R + r \cos \varphi) \cos \varphi \sin \psi\\ -r(R + r \cos \varphi) \sin \varphi \end{pmatrix} $$
and its magnitude:
$$ \begin{aligned} \Vert r_{\varphi} \times r_{\psi} \Vert &= r(R + r \cos \varphi)\\ &\quad\cdot \sqrt{\cos^2 \varphi \cos^2 \psi + \cos^2 \varphi \sin^2 \psi + \sin^2 \varphi} \end{aligned} $$
By recalling that $\cos^2 \psi + \sin^2 \psi = 1$ and $\cos^2 \varphi + \sin^2 \varphi = 1$, we obtain:
$$ \Vert r_{\varphi} \times r_{\psi} \Vert = r(R + r \cos \varphi) $$
The infinitesimal area element becomes:
$$ d\sigma = r(R + r \cos \varphi) \, d\varphi \, d\psi $$
Hence:
$$ \begin{aligned} A(\Sigma) &= \int_0^{2\pi} \int_0^{2\pi} r(R + r \cos \varphi) \, d\varphi \, d\psi\\ &= r \int_0^{2\pi} \left( \int_0^{2\pi} (R + r \cos \varphi) \, d\varphi \, \right) d\psi\\ &= r \int_0^{2\pi} \left[ R \varphi + r \sin \varphi \right]_0^{2\pi} \, d\psi\\ &= r \int_0^{2\pi} 2 \pi R \, d\psi\\ &= \boxed{4 \pi^2 R r} \end{aligned} $$
The second question, instead, asks us to determine the moment of inertia, which is a scalar measure of an object’s resistance to changes in its rotational motion about a specific axis and depends on how the object’s mass is distributed relative to that axis. The formula to calculate it is:
$$ I_z = \iint_{\Sigma} d_z^2 \, dm $$
where $d_z$ represents the distance of each point of the surface from the $z$-axis. In our case, this axis corresponds to the intersection of the planes $x=0$ and $y=0$, while the distance of a given point $(x,y,z)$ from the $z$-axis is:
$$ d_z = \sqrt{x^2 + y^2} \implies d^2_z = x^2 + y^2 $$
Therefore:
$$ I_z = \iint_{\Sigma} (x^2 + y^2) \, dm $$
Furthermore, since the surface is homogeneous, its surface mass density is constant, that is:
$$ \delta = \frac{M}{A(\Sigma)} \iff \delta = \frac{M}{4 \pi^2 Rr} $$
This value represents the mass of each area element. Therefore, the infinitesimal mass $dm$ associated with the surface element $d\sigma$ is:
$$ dm = \delta \, d\sigma $$
Thus:
$$ I_z = \iint_{\Sigma} (x^2 + y^2) \delta \, d\sigma = \delta \iint_{\Sigma} (x^2 + y^2) \, d\sigma $$
Now let’s evaluate the scalar field along the parametrization:
$$ \begin{aligned} x^2 + y^2 &= (R + r \cos \varphi)^2 \cos^2 \psi + (R + r \cos \varphi)^2 \sin^2 \psi\\ &= (R + r \cos \varphi)^2 (\underbrace{\cos^2 \psi + \sin^2 \psi}_1)\\ &= (R + r \cos \varphi)^2 \end{aligned} $$
We then arrive at the following result:
$$ \begin{aligned} I_z &= \delta \iint_D (R + r \cos \varphi)^2 r (R + r \cos \varphi) \, d\varphi \, d\psi\\ &= \delta r \int_0^{2\pi} \int_0^{2\pi} (R + r \cos \varphi)^3 \, d\varphi \, d\psi \end{aligned} $$
Since the integrand does not depend by $\psi$, we can split the two integrals:
$$ I_z = \delta r \left( \int_0^{2\pi} d \psi \right) \left( \int_0^{2\pi} (R + r \cos \varphi)^3 \, d\varphi \right) $$
The first one is immediate:
$$ \int_0^{2\pi} d \psi = 2 \pi $$
For the second one, let’s expand the following cube:
$$ (R + r \cos \varphi)^3 = R^3 + 3R^2 r \cos \varphi + 3Rr^2 \cos^2 \varphi + r^3 \cos^3 \varphi $$
Then, by recalling the following immediate integrals:
$$ \int_0^{2\pi} \cos \varphi \, d\varphi = 0, \quad \int_0^{2\pi} \cos^2 \varphi \, d\varphi = \pi, \quad \int_0^{2\pi} \cos^3 \varphi \, d\varphi = 0 $$
we conclude that:
$$ \begin{aligned} \int_0^{2\pi} (R + r \cos \varphi)^3 \, d\varphi &= 2 \pi R^3 + 3\pi R r^2\\ &= \pi R (2R^2 + 3r^2) \end{aligned} $$
By combining the two contributions, we obtain:
$$ \begin{aligned} I_z &= \delta r \cdot 2\pi \cdot \pi R (2R^2 + 3r^2)\\ &= 2\pi^2 \delta R r(2R^2 + 3r^2) \end{aligned} $$
Let’s now substitute the constant surface density $\delta$:
$$ \begin{aligned} I_z &= 2\pi^2 \frac{M}{4\pi^2 R r} Rr (2R^2 + 3r^2)\\ &= \frac{M}{2} (2R^2 + 3r^2)\\ &= \boxed{M \left( R^2 + \frac{3}{2} r^2 \right)} \end{aligned} $$
Surface Integrals of Vector Fields
In the second part of this lesson, we will study surface integrals of vector fields over surfaces. Unlike the previous type of surface integral, this one depends on the orientation of the surface.
Consider the following vector field:
$$ \mathbf{F}(x,y,z) = (P(x,y,z), Q(x,y,z), R(x,y,z)) $$
and suppose we want to measure the net amount of $\mathbf{F}$ passing through an arbitrary oriented surface $\Sigma$. This scalar quantity is called the flux of the vector field and we can distinguish the following cases at each point of the surface:
- If $\mathbf{F}$ is tangent to $\Sigma$, then the vector field does not cross the surface at that point;
- If $\mathbf{F}$ is perpendicular to $\Sigma$ and points in the same direction as the chosen normal, then the flux is positive;
- If $\mathbf{F}$ is perpendicular to $\Sigma$ but points in the opposite direction to the chosen normal, then the flux is negative.
The component of the vector field that effectively passes through the surface is its projection onto the unit normal vector, that is, $\langle \mathbf{F}, \mathbf{n} \rangle$.
Hence the surface integral of a vector field over an oriented surface is:
$$ \iint_{\Sigma} \langle \mathbf{F}, \mathbf{n} \rangle \, d\sigma $$
By recalling that:
$$ d\sigma = \Vert r_u \times r_v \Vert \, du \, dv $$
We obtain:
$$ \begin{aligned} \iint_{\Sigma} \langle \mathbf{F}, \mathbf{n} \rangle \, d\sigma &= \iint_D \mathbf{F}(r(u,v)) \cdot \frac{r_u \times r_v}{\cancel{\Vert r_u \times r_v \Vert}} \cdot \cancel{\Vert r_u \times r_v \Vert} \, du \, dv\\ &= \iint_D \mathbf{F}(r(u,v)) \cdot (r_u \times r_v) \, du \, dv \end{aligned} $$
Notice that reversing the orientation of the surface changes the sign of the flux.
Let’s now look at a final example.
Note: the following exercises is taken from Marco Bramanti’s Esercitazioni di Analisi Matematica 2 (2012) (page 560, exercise 6.52).
Find the outward flux of the vector field:
$$ \mathbf{E}(x,y,z) = \frac{(x,y,z)}{(x^2 + y^2 + z^2)^2} $$
across the boundary of the three-dimensional region defined by:
$$ \Omega = \left\{ (x,y,z) \in \mathbb{R}^3 : R_1 \leq \sqrt{x^2 + y^2 + z^2} \leq R_2 \right\} $$
with $0 < R_1 < R_2$.
We immediately notice that $\Omega$ denotes a spherical shell, that is, the region included between two concentric spheres of radii $R_1$ and $R_2$, respectively. Moreover, its border is formed by two surfaces:
$$ \partial \Omega = \Sigma_1 \cup \Sigma_2 $$
where $\Sigma_1$ represents the internal sphere of radius $R_1$ while $\Sigma_2$ represent the external sphere of radius $R_2$.
Since we want to determine the outward flux of $\Omega$, the normal on $\Sigma_2$ points outwards, away from the origin, while on the internal sphere $\Sigma_1$, the normal points inwards, towards the origin.
Let’s begin by determining a parametrization of the two spheres:
$$ r(\theta, \varphi) = \begin{pmatrix} R \sin \theta \cos \varphi\\ R \sin \theta \sin \varphi\\ R \cos \theta \end{pmatrix}, \quad \begin{matrix} &\theta \in [0, \pi]\\ &\varphi \in [0, 2\pi] \end{matrix} $$
where for $\Sigma_1$ we set $R = R_1$, while for $\Sigma_2$ we set $R = R_2$.
Let’s now calculate the normal vector:
$$ r_{\theta} = \begin{pmatrix} R \cos \theta \cos \varphi\\ R \cos \theta \sin \varphi\\ -R \sin \theta \end{pmatrix}, \quad r_{\varphi} = \begin{pmatrix} -R \sin \theta \sin \varphi\\ R \sin \theta \cos \varphi\\ 0 \end{pmatrix} $$
then their vector product:
$$ r_{\theta} \times r_{\varphi} = R^2 \sin \theta \begin{pmatrix} \sin \theta \cos \varphi\\ \sin \theta \sin \varphi\\ \cos \theta \end{pmatrix} $$
and finally its magnitude:
$$ \Vert r_{\theta} \times r_{\varphi} \Vert = R^2 \sin \theta $$
Therefore, we have:
$$ d\sigma = R^2 \sin \theta \, d\theta \, d\varphi $$
while the outward-facing unit normal of the external sphere $\Sigma_2$ is:
$$ \mathbf{n} = \begin{pmatrix} \sin \theta \cos \varphi\\ \sin \theta \sin \varphi\\ \cos \theta \end{pmatrix} $$
Conversely, for the internal sphere $\Sigma_1$, we will use $-\mathbf{n}$.
Let’s now calculate the flux of the external sphere $\Sigma_2$:
$$ \Phi_2 = \oiint_{\Sigma_2} \langle \mathbf{E}, \mathbf{n} \rangle \, d\sigma $$
On the sphere, we have:
$$ x^2 + y^2 + z^2 = R_2^2 $$
Therefore, the scalar field evaluated on the parametrization becomes:
$$ \begin{aligned} \mathbf{E}(r(\theta, \varphi)) &= \frac{1}{R_2^4} \begin{pmatrix} R_2 \sin \theta \cos \varphi\\ R_2 \sin \theta \sin \varphi\\ R_2 \cos \theta \end{pmatrix}\\ &= \frac{1}{R^3_2} \begin{pmatrix} \sin \theta \cos \varphi\\ \sin \theta \sin \varphi\\ \cos \theta \end{pmatrix}\\ \end{aligned} $$
We notice that the field is radially outward. Let’s now compute the dot product:
$$ \begin{aligned} \langle \mathbf{E}, \mathbf{n} \rangle &= \frac{1}{R_2^3} (\sin^2 \theta \cos^2 \varphi + \sin^2 \theta \sin^2 \varphi + \cos^2 \theta)\\ &= \frac{1}{R_2^3} \end{aligned} $$
We obtain:
$$ \begin{aligned} \Phi_2 &= \int_0^{2\pi} \int_0^{\pi} \frac{1}{R_2^3} R_2^2 \sin \theta \, d\theta \, d\varphi\\ &= \frac{1}{R_2} \int_0^{2\pi} \int_0^{\pi} \sin \theta \, d\theta \, d\varphi\\ &= \frac{1}{R_2} \left(\int_0^{\pi} \sin \theta \, d\theta \right) \left(\int_0^{2\pi} d\varphi \right)\\ &= \frac{1}{R_2} \cdot 2 \cdot 2\pi\\ &= \boxed{\frac{4\pi}{R_2}} \end{aligned} $$
Let’s now do the same for the flux of the internal sphere $\Sigma_1$. Since the outward normal from $\Omega$ points towards the origin, we have:
$$ \mathbf{n}_{\mathrm{int}} = - \mathbf{n} $$
The vector field, instead, continues to be radially outward. Hence, the dot product changes its sign:
$$ \langle \mathbf{E}, \mathbf{n}_{\text{int}} \rangle = - \frac{1}{R_1^3} $$
The integral becomes:
$$ \begin{aligned} \Phi_1 = \oiint_{\Sigma_1} \langle \mathbf{E}, -\mathbf{n} \rangle \, d\sigma &= \int_0^{2\pi} \int_0^{\pi} - \frac{1}{R_1^3} R_1^2 \sin \theta \, d\theta \, d\varphi\\ &= - \frac{1}{R_1} \int_0^{2\pi} \int_0^{\pi} \sin \theta \, d\theta \, d\varphi\\ &= \boxed{- \frac{4\pi}{R_1}} \end{aligned} $$
Therefore, since the boundary is formed by the two surfaces, the total flux is given by:
$$ \Phi = \Phi_1 + \Phi_2 $$
That is:
$$ \begin{aligned} \Phi &= - \frac{4\pi}{R_1} + \frac{4\pi}{R_2}\\ &= \boxed{4 \pi \left( \frac{1}{R_2} - \frac{1}{R_1} \right)} \end{aligned} $$
Since $R_1 < R_2$, we conclude that the magnitude of the inward flux through the internal sphere is greater than the outward flux through the external sphere. Consequently, the total outward flux is negative.
Finally, we can verify this result by applying Gauss’s theorem, which states the following:
$$ \iint_{\partial \Omega} \langle \mathbf{E}, \mathbf{n} \rangle \, d\sigma = \iiint_{\Omega} \nabla \cdot \mathbf{E} \, dV $$
We can begin by determining the divergence of the vector field:
$$ \mathbf{E} = \left( \frac{x}{(x^2+y^2+z^2)^2}, \, \frac{y}{(x^2+y^2+z^2)^2}, \, \frac{z}{(x^2+y^2+z^2)^2} \right) $$
Let’s now choose:
$$ \rho^2 = x^2 + y^2 + z^2 $$
Thus:
$$ \begin{aligned} \nabla \cdot \mathbf{E} &= \frac{\partial}{\partial x} \left(\frac{x}{\rho^4} \right) + \frac{\partial}{\partial y} \left( \frac{y}{\rho^4} \right) + \frac{\partial}{\partial z} \left( \frac{z}{\rho^4} \right)\\[1em] &= \frac{3}{\rho^4} - \frac{4(x^2 + y^2 + z^2)}{\rho^6}\\ &= \frac{3}{\rho^4} - \frac{4\rho^2}{\rho^6}\\ &= \frac{3}{\rho^4} - \frac{4}{\rho^4}\\ &= - \frac{1}{\rho^4} \end{aligned} $$
We can now perform a change of variables to the spherical coordinate system, where the volume element $dV$ is determined by:
$$ dV = \rho^2 \sin \theta \, d\rho \, d\theta \, d\varphi $$
The flux is therefore:
$$ \begin{aligned} \Phi &= \int_{R_1}^{R_2} \int_0^{\pi} \int_0^{2\pi} - \frac{1}{\rho^4} \rho^2 \sin \theta \, d\varphi \, d\theta \, d\rho\\ &= -4\pi \int_{R_1}^{R_2} \frac{1}{\rho^2} \, d\rho\\ &= -4\pi \left[ -\frac{1}{\rho} \right]_{R_1}^{R_2}\\ &= -4\pi \left(-\frac{1}{R_2} + \frac{1}{R_1} \right)\\ &= \frac{4\pi}{R_2} - \frac{4\pi}{R_1}\\ &= \boxed{4\pi \left(\frac{1}{R_2} - \frac{1}{R_1} \right)} \end{aligned} $$