Vector Calculus and Differential

Forms with Applications to

Electromagnetism

Sean Roberson

May 7, 2015
P R E FA C E

This paper is written as a final project for a course in vector analysis,

taught at Texas A&M University - San Antonio in the spring of 2015
as an independent study course.
Students in mathematics, physics, engineering, and the sciences
usually go through a sequence of three calculus courses before go-
ing on to differential equations, real analysis, and linear algebra. In
the third course, traditionally reserved for multivariable calculus, stu-
dents usually learn how to differentiate functions of several variable
and integrate over general domains in space. Very rarely, as was my
case, will professors have time to cover the important integral theo-
rems using vector functions: Green’s Theorem, Stokes’ Theorem, etc.
In some universities, such as UCSD and Cornell, honors students are
able to take an accelerated calculus sequence using the text Vector Cal-
culus, Linear Algebra, and Differential Forms by John Hamal Hubbard
and Barbara Burke Hubbard. Here, students learn multivariable cal-
culus using linear algebra and real analysis, and then they generalize
familiar integral theorems using the language of differential forms.
This paper was written over the course of one semester, where the
majority of the book was covered.
Some details, such as orientation of manifolds, topology, and the
foundation of the integral were skipped to save length. The paper
should still be readable by a student with at least three semesters of
calculus, one course in linear algebra, and one course in real analysis
- all at the undergraduate level. Many of the fundamental theorems
do not have their proofs, for some would be lengthy or require many
details from real analysis.
The idea of this paper is to introduce differential forms - an object
that allows Maxwell’s equations of electromagnetism to be concisely
represented in two lines as opposed to four. The development of this
idea takes time, but it is worth it in the long run.

-Sean Zachary Roberson

2
ACKNOWLEDGEMENTS

Many individuals supported me in writing this paper. First, I would

like to acknowledge my professor Donald F. Myers for teaching this
class in a way that tapped into my interests and offered me a chal-
lenge. He has also provided me with numerous resources that ex-
tended my knowledge of mathematics and my methods of problem
solving. I would also like to thank Dr. John G. Romo for being able
to have this course offered. It served as an excellent extension of my
knowledge from real analysis taken the semester before, as well as
linear algebra and differential equations.
While writing this paper, I have received numerous amounts of sup-
port from my coworkers, each of whom are writing tutors. I would
like to extend my thanks to Brandi Anderson, Stephanie Gil, Carlos
Lopez, Coral Rosario, Teresa Ruiz, Lisa Sanna, and Sheridan Santens
for their encouragement in motivating me to finish this project, all
while still tending to my responsibilities at work. I also would like
to thank Chris Tingwald and Dr. Katherine T. Bridgman for their
support as well.
Most importantly, I could not have been motivated to finish this
paper without the undying support of my family and friends, espe-
cially my mother, Denise Ginther, for giving me the motivation to
finish college and pursue a degree.
Without all your help, I would not have finished this paper, nor
would it have been a quality product.

3
 1
F U N D A M E N TA L D I F F E R E N T I A L O P E R AT O R S

In vector analysis, it is of interest to examine the rate of flow of var-

ious quantities such as heat, electricity, and fluids. The concept of
derivative can be extended to vectors in three ways, each using some
sort of partial derivative. Each differential operator uses the symbol
∇ to operate on a function of several variables. This symbol is treated
as the following vector:
 ∂

∂x1
 ∂ 
 ∂x2 
 ∂ 
∇ :=  ∂x 
 3
 .. 
 . 
∂
∂xn

1.1 gradient

The first differential operator is the gradient, denoted by ∇ f or grad f .

The gradient operates on a C 1 function f : Rn → R and returns a
vector of first derivatives. For example, if f ( x, y, z) = 2x + 3xy − 4yz2 ,
then
 
2 + 3y
∇ f = 3x − 4z2 
−8yz
There are scalar product and quotient rules for gradients, with

grad( f g) = f grad g + g grad f

and

g grad f − f grad g
 
f
grad = ,
g g2
similar to the product and quotient rules familiar in single variable
calculus.

4
 1.2 divergence

1.2 divergence

The next operator is the divergence, and it returns a sum of partial

derivatives of a vector function F : Rn → Rn . Using the nabla nota-
tion, divergence can be denoted by ∇ · F, or as div F. By definition,

∂F j
div F := ∑ ∂x j
.
1≤ j ≤ n

Since div is a differential operator, there are product and quotient

rules for differentiation. The product rule states that for a scalar func-
tion f and a vector function G,

div( f G) = G · grad f + f div G.

Similarly, the quotient rule states

f div G − G · grad f
 
G
div = .
f f2
Another differential operator can be built from the divergence op-
erator. The Laplacian is a differential operator that returns a sum of
second derivatives. It it defined as

∇2 : = ∇ · ∇,
or, when acting on a function f ,

∇2 f = div grad f
= ∇ · ∇f
∂2 f
= ∑ 2
.
1≤ j≤n ∂x j

The Laplacian appears in many partial differential equations, such

as the wave equation

∂2 u
= c2 ∇2 u,
∂t2
the diffusion equation

∂u
= k∇2 u,
∂t
and the Laplace equation ∇2 u = 0. It is important to note that the
Laplacian operator acts on spatial variables such as x, y, z and not the
time variable t (Strauss, 14).

5
 1.3 curl

1.3 curl

The last fundamental differential operator exists only in R3 . The curl

returns a vector of derivatives according to the rule ∇ × F. That is,
 ∂F ∂F2

∂y
3
− ∂z
 ∂F1 ∂F3 
curl F :=  ∂z − ∂x 
.
∂F2 ∂F1
∂x − ∂y

The curl of a product of a scalar and vector function is given by

curl( f G = f curl G + grad f × G,

and the curl of the quotient of a vector function with a scalar func-
tion is

f curl G − G × (grad f )
 
G
curl = .
f f2
We also have two important results regarding the curl.

Theorem 1. Suppose f is differentiable. Then curl grad f = 0.

Proof. By definition,

∂2 f ∂2 f
 
−
 ∂y∂z ∂z∂y
 ∂2 f ∂2 f 

curl grad f =  ∂x∂z − ∂z∂x 
∂2 f ∂2 f
∂x∂y − ∂y∂x

=0

since mixed partials commute.

Theorem 2. Suppose F is a differentiable vector function on R3 . Then

div curl F = 0.

Proof. By definition,

 ∂F ∂F2

 T ∂y
3
− ∂z
∂ ∂ ∂  ∂F3 ∂F1 
div curl F = ∂x ∂y ∂z  ∂x − ∂z 
∂F2 ∂F1
∂x − ∂y
∂2 F3 ∂2 F2 ∂2 F2 2 ∂2 F1
∂ F1 ∂2 F3
= − + − + −
∂x∂y ∂x∂z ∂y∂z ∂y∂x ∂z∂x ∂z∂y
= 0,

since, as above, mixed partials commute.

6
 1.3 curl

There are many other product rules for vector derivatives. Griffiths
gives the following three product rules utilizing both the inner and
cross products:

1. grad(F · G) = (F × curl G) + (G × curl F) + (F · ∇)G + (G · ∇)F

2. div(F × G) = G · curl F − F · curl G

3. curl(F × G) = (G · ∇)F − (F · ∇)G + F(div G) − G(div F)

7
 2
C L A S S I C A L T H E O R E M S A N D R E S U LT S

As with any branch of mathematics, vector calculus has its own set of
classic results. The theorems here are presented in the order in which
they appear in James Stewart’s Calculus: Early Transcendentals.

2.1 fundamental theorem for line integrals

The Fundamental Theorem for Line Integrals, also called the Gradi-
ent Theorem by David Griffiths gives a way to evaluate special line
integrals (p. 29). In order to understand the theorem, we must first
define a conservative vector field and a potential function.

Definition 1 (Conservative Vector Field, Potential Function). A

vector field is conservative if it is the gradient of some differentiable
function. That is, F is conservative iff F = grad f . Such a function f
is called a potential function for F.

Finding these potential functions amounts to solving an exact dif-

ferential equation, defined below, along with the definition of an exact
differential.

Definition 2 (Exact Differential, Exact Differential Equation). Let

z = f ( x, y) be a differentiable function. The symbol dz denotes the
∂f ∂f
differential of f , given as ∂x dx + ∂y dy. The expression M dx + N dy
∂f ∂f
is an exact differential if M = ∂x and N = ∂y - that is, it corresponds
to the differential of a function. The associated differential equation
M dx + N dy = 0 is exact if the left side is an exact differential.

Solving these exact differential equations uses the equality of mixed

derivatives. This establishes the following exactness condition; a simi-
lar expression will be used in the computations for Green’s Theorem.

Theorem 3 (Condition for Exactness). A necessary and sufficient

condition for M dx + N dy to an exact differential on some rectangle R
is

8
 2.1 fundamental theorem for line integrals

∂M ∂N
= ,
∂y ∂x
provided that M and N are differentiable on R.

The proof for the necessity is given in Zill, while the proof of the
sufficiency outlines the method of solution to an exact differential
equation (p. 64). To solve such an equation, one of the functions M
or N is selected and then integrated with respect to their “attached”
variable - that is, integrate M with respect to x, and N with respect
to y. Afterwards, differentiate with respect to the remaining variable
and then solve for f ( x, y). This procedure is similar to finding a po-
tential function for a conservative vector field.

Theorem 4 (Conservative Criterion). If F = P( x, y)i + Q( x, y)j is

conservative, then

∂P ∂Q
= .
∂y ∂x
The converse holds on an open, simply-connected region - a region
whose boundary curve does not intersect itself and is convex.

Thus, to find a potential function for the vector field F described

above, it suffices to solve the exact differential equation

P dx + Q dy = 0.
We are now ready to state the Fundamental Theorem of Line Inte-
grals.

Theorem 5 (Fundamental Theorem of Line Integrals, Gradient

Theorem). Let C be a smooth curve parameterized by r(t), a ≤ t ≤ b.
Suppose f is differentiable with a continuous gradient on C. Then
Z
grad f · dr = f (r(b)) − f (r( a)).
C

From the Gradient Theorem follows a property of integrals taken

over a closed loop.

R
Theorem 6 (Path Independence Theorem). The integral F · dr is
R C
path independent on a domain D iff F · dr = 0 for each closed path
C
C ⊂ D.

9
 2.2 green’s theorem

2.2 green’s theorem

The following result is due to George Green. It relates the line integral
of a function over a closed curved to an associated double integral
taken over the interior of the curve.

Theorem 7 (Green’s Theorem). Suppose C is a curve with positive

orientation (that is, oriented in the counterclockwise direction). Suppose
also that C is smooth and closed. If P and Q are of class C 1 on some
open set contained in C̊, then
Z ZZ
∂Q ∂P
P dx + Q dy = − dA.
C ∂x ∂y
C̊

It immediately follows from the theorem that if the integrand on

the left is an exact differential, the integral is zero.
We may also write Green’s Theorem in two vector forms: one with
the curl and the other with the divergence.

Theorem 8 (Vector Forms of Green’s Theorem). Suppose F = Pi +

Qj is a vector field in R3 whose component functions are of class C 1 ,
and suppose C is a simple, smooth, closed curve. Then
I ZZ
F · dr = (curl F) · k dA.
C
C̊

Equivalently, if n is an outward-oriented normal to the curve C, then

I ZZ
F · n ds = ∇ · F dA.
C
C̊

From the vector forms of Green’s theorem, two identities follows.

The first can be viewed as a generalized form of integration by parts.

Corollary 1 (Green’s First Identity). Suppose f is of class C 1 and

g is of class C 2 , and that the compact set D and its boundary satisfy
Green’s Theorem. Then
ZZ I ZZ
2
f ∇ g dA = f (∇ g) · n ds − ∇ f · ∇ g dA.
∂D
D D

In order to prove the identity, the following property about diver-

gence is needed.

10
 2.2 green’s theorem

Lemma 1. Suppose f is C 1 and g is C 2 . Then ∇ · ( f ∇ g) = f ∇2 g +

∇ f · ∇ g.

Proof of Lemma. We have

∂ f gx ∂ f gy ∂ f gz
∇ · ( f ∇ g) = + +
∂x ∂y ∂z
= f gxx + f x gx + f gyy + f y gy + f gzz + f z gz
= f ∇2 g + ∇ f · ∇ g.

We are now ready to prove Green’s First Identity.

Proof of Green’s First Identity. Begin with the integral of f ∇ g · n taken

over the boundary of D. By the lemma and the second vector form of
Green’s Theorem,

I ZZ
f ∇ g · n ds = ∇ · ( f ∇ g) dA
∂D
D
ZZ
= f ∇2 g + ∇ f · ∇ g dA.
D

Rearranging yields the desired.

Corollary 2 (Green’s Second Identity). Suppose f and g satisfy the

hypotheses of the First Identity. Then
ZZ I
2 2
f ∇ g − g∇ f dA = ( f ∇ g − g∇ f ) · n ds.
∂D
D

Proof. Apply the First Identity to the integral on the left, separating
the integrand. Then the only term that remains is the line integral.

Using Green’s identities, we can examine what happens to func-

tions that are harmonic on a set. That is, if a function f satisfies
Laplace’s equation:

∇2 f = 0,
then the following theorem holds.

11
 2.3 stokes’ theorem

Corollary 3. Suppose D isH a region that satisfies Green’s theorem and

g is harmonic on D. Then ∂D ∇ g · n ds = 0.

Proof. Let f ( x, y, z) = 1. By the First Identity,

ZZ I ZZ
f ∇2 g dA = ∇ g · n ds − ∇ f · ∇ g dA
∂D
D D
I ZZ
= ∇ g · n ds − 0 · ∇ g dA
∂D
D
I
= ∇ g · n ds
∂D

Now, since g is harmonic on D, the first integral is zero. Hence the

result holds.

Corollary 4. Suppose f is harmonic

RR on a 2region D satisfying Green’s
Theorem and f ≡ 0 on ∂D. Then |∇ f | dA = 0.
D

Proof. Set f = g in the First Identity. Then

ZZ I ZZ
2
f ∇ f dA = f ∇ f · n ds − |∇ f |2 dA
∂D
D D
ZZ
2
=− |∇ f | dA.
D

But ∇2 f = 0 on D, so the integral on the left is zero. Also, note that

the norm functions is non-negative, and the only way |∇ f |2 can be
zero is if ∇ f = 0. Hence, the entire integral is zero (and, consequently,
f is identically zero on the entire domain D).

2.3 stokes’ theorem

The next theorem is due to George Stokes. As another integral theo-

rem, it relates a line integral to a surface integral under appropriate
conditions. The generalization of Stokes’ Theorem will be explored
once differential forms and exterior derivatives are introduced.

Theorem 9 (Stokes’ Theorem). Let S be a surface whose boundary

curve is simple, closed, and has positive orientation. Suppose that the
vector field F : R3 → R3 has continuous partial derivatives on an open
set containing S. Then
I ZZ
F · dr = curl F · n dS,
∂S
S

12
 2.4 gauss’ divergence theorem

Where the integral on the right is a surface integral.

As with Green’s Theorem, Stokes’ Theorem changes difficult inte-

grals into easier problems.

2.4 gauss’ divergence theorem

In mathematics, it is common for the same result to be discovered in-

dependently by different mathematicians. For example, the Cauchy-
Schwarz inequality for sums was discovered by Augustin-Louis Cauchy,
while the related inequality for integrals was given by Viktor Bun-
yakovsky. The next theorem is credited to Carl Gauss, while Mikhail
Ostrogradsky published the result in 1826 (Stewart, p.1129).

Theorem 10 (Divergence Theorem). Suppose E is a solid and S is the

surface (or union of surfaces), with outward orientation, that encloses
E. Let F be a C 1 vector field on an open set containing E. Then
ZZ ZZZ
F · dS = ∇ · F dV.
S E

Corollary 5 (Volume Formula). Suppose E is a solid satisfying the

hypothesis
RR of the Divergence Theorem. Then the volume of E is given
1
by 3 ( xi + yj + zk) · dS.
S

Proof. Note that ∇ · ( xi + yj + zk) = 3. The result follows.

13
 3
DIFFERENTIAL FORMS

It is common in mathematics to find multiple representations for the

same object or quantity. Such methods are useful in vector calcu-
lus. In the typical Euclidean vector spaces R2 and R3 , results in
multivariable calculus involving areas and volumes are easy to write
with integrals, but generalized results in Rn may be cumbersome to
write down. The use of differential forms (shortened to forms) al-
lows these results to be streamlined using a generalized differential
operator while maintaining traditional methods of integration.

3.1 forms

The differential form is a function that acts on a given set of vectors

and returns a number. There are three defining properties of forms:

1. Forms are antisymmetric. That is, exchanging any two vectors

changes the sign of the result.

2. Forms are normalizing. That is, when a form acts on the stan-
dard basis vectors, it returns 1.

3. Forms are multilinear. That is, a linear combination of vectors

can be separated in the form.

Notice that the properties of forms are similar to those for determi-
nants. Thus the determinant is an example of a form on Rn .
A form is classified by the number of vectors it acts on and the
vector space it lives in. For example, a 2-form on R3 can look like
2x + y − z dx ∧ dz, and a 3-form on R6 can look like x1 x2 dx2 ∧ dx5 ∧
dx6 . The wedges determine how to act on the vectors provided (the
example forms given above do not have specific vectors). For now,
we can focus on the wedge product and not the function at the front
of the product. To see this, suppose ϕ is a k-form on Rn . In particular,
let ϕ = dxi1 ∧ dxi2 ∧ . . . ∧ dxik , and let v1 , . . . , vk be a set of vectors in
Rn . Build the matrix A whose columns are each v j . Then ϕ takes, in
succession, each row i1 , i2 , . . . ik from A, builds a new matrix B, then
takes the determinant of B. Notice that if im = in , then ϕ returns the
value zero.

14
 3.2 forms and integrals

3.2 forms and integrals

Students in traditional multivariable calculus courses see differential

forms, but do not use them in the context of vector analysis and exte-
rior algebra. The exposure of differential forms is usually restricted to
line integrals
R in these courses. For example, what is the value of the
integral C x dy, where C is the quarter-circle joining (2, 0) and (0, 2)?
In a traditional class, the path C is given a parameterization γ(t), then
the derivative of γ is taken. From there, the students substitute the
appropriate expressions and evaluate an integral in t. Using the stan-
dard parameterization γ(t) = (2 cos t, 2 sin t), the computation is as
follows:

Z Z
x dy = (2 cos t) d(2 sin t)
C [0, π2 ]
Z 2
= 4 cos2 t dt
0
=π

The computation is the same using differential forms, except we

must now consider the 1-form x dy as a function that acts on the vec-
tor (−2 sin t, 2 cos t). We then replace x with 2 cos t, as in the parame-
terization γ, then replace dy with 2 cos t dt. The second replacement
occurs from the y in the differential dy that selects the second entry
in the vector. The computation of the integral remains the same.
We can extend this process to ”larger” forms. These integrals of
forms have connections to standard integrals in vector calculus. For
instance, what is the integral of the 2-form x dy ∧ dz, over the surface
parametrized by γ(u, v) = (u2 , u + v, v3 ), where (u, v) ∈ S = [−1, 1]2 ?
To start, we must first examine the derivative of the parametrization
γ. These derivatives can be placed in a 3 × 2 matrix, where each
column represents a partial derivative with respect to each parameter
and each row represents the gradient of the function in a particular
entry. The derivative is then

2u 0
[Dγ] =  1 1 .
0 3v2
Z
We use this derivative to evaluate x dy ∧ dz. First, we examine
[γ(S)]
the parametrization and replace x with u2 . Now, the wedge product
dy ∧ dz says to take the second and third rows of the derivative matrix
and take the determinant. Call this matrix B, so
 
1 1
B= ,
0 3v2

15
 3.3 the exterior derivative

and hence det B = 3v2 . The last step is to write the integral as an
ordinary double integral, making appropriate choices for the limits
of integration (the limits are given in the constraint for S). So,

Z Z 1 Z 1
x dy ∧ dz = 3u2 v2 du dv
[γ(S)] −1 −1
Z 1
2
2
=3 w dw
−1
2
= 3·
3
= 2.

It is important to note that integrals of 2-forms are analogous to

surface integrals in traditional multivariable calculus courses. In the
traditional manner, we may express the integral as
ZZ
x dσ,
S
where σ is a surface area element of S.

3.3 the exterior derivative

The derivative has many interpretations in calculus. It usually repre-

sents slopes of curves, but can be generalized to represent a rate of
change of a certain quantity. For functions of several variables, the
partial derivatives examine change in one variable while others are
held fixed. Directional derivatives give change in the direction of a
particular unit vector, usually combining partial derivatives. The gra-
dient is a special differential operator that returns a vector of partial
derivatives. This vector points in the direction of greatest change.
In vector analysis (and in some cases, tensor analysis), a derivative
is needed for integrals (such as in Stokes’ Theorem) and relating var-
ious forms. To achieve this goal, the exterior derivative is needed.
Hubbard and Hubbard define the exterior derivative as follows:

Definition 3. Let U be an open subset of Rn and ϕ a k+1-form on U.

The exterior derivative dϕ is given by:

1
Z
dϕ( Px (v1 , . . . , vk+1 )) = lim ϕ,
h →0 h k +1 ∂Px (hv1 ,...,hvk+1 )

where Px (·) represents a parallelogram in k dimensions spanned by

the vectors inside the parentheses and with a vertex at x.

This definition is motivated by replacing the ordinary difference

quotient with an integral. For functions of one variable, the derivative
of f is given by

16
 3.3 the exterior derivative

f ( x + h) − f ( x )
f 0 ( x ) = lim .
h →0 h

R The expression f ( x + h) − f ( x ) can be replaced by the integral

∂Px (h)
f , which says to evaluate f at x + h, then subtract the value of f
at x. Notice how there is no mention of antiderivatives. When using
this formulation for the exterior derivative of differential forms, one
is typically not interested in antiderivatives, since integrals of forms
can be converted to standard multiple integrals.
As with the definition of the derivative in one variable, using the
definition to compute an exterior derivative is cumbersome. There
are rules used to find exterior derivatives.

1. The operator d is linear.

2. If ϕ is a constant form, then dϕ = 0.

3. The exterior derivative of a 0-form (a function) f is its total

differential:

∂f
df = ∑ ∂x j
dx j
1≤ j ≤ n
!
= df ∧
V V
4. If f is a function, then d f dxi j dxi j .
1≤ j ≤ k 1≤ j ≤ k

To see how the exterior derivative works, first consider the function
of two variables f ( x, y) = 3 ln xy − 2 sinh x. Then by the third rule,
d f = d f , that is,

∂ ∂
df = (3 ln xy − 2 sinh x ) dx + (3 ln xy − 2 sinh x ) dy
∂x ∂y
3 3
= − 2 cosh x dx + dy.
x y

As another example, let ϕ = 2xz2 dx ∧ dy − 3x3 dy ∧ dz. Then

dϕ = d(2xz2 ) ∧ dx ∧ dy − d(3x3 ) ∧ dy ∧ dz
= (2z2 dx + 4xz dz) ∧ dx ∧ dy − 9x2 dx ∧ dy ∧ dz
= 2z2 dx ∧ dx ∧ dy + 4xz dz ∧ dx ∧ dy − 9x2 dx ∧ dy ∧ dz
= 4xz dx ∧ dy ∧ dz − 9x2 dx ∧ dy ∧ dz
= (4xz − 9x2 ) dx ∧ dy ∧ dz.

Perhaps an interesting property of the exterior derivative is that

second exterior derivatives are always zero.

17
 3.4 some basic forms: work, flux, and mass

Theorem 11. Suppose ϕ is a k-form of class C 2 . Then d(dϕ) = 0.

Proof. If ϕ is a 0-form (namely a function f ), then

!
∂f
d(dϕ) = d ∑ ∂x j dx j
1≤ j ≤ n
 
∂f
= ∑ d dx j
1≤ j ≤ n
∂x j
 
∂f
= ∑ d ∧ dx j
1≤ j ≤ n
∂x j
 2 
∂ f
= ∑ dx j ∧ dxi
1≤i,j≤n
∂xi ∂x j
= 0.

Otherwise, if ϕ is a k-form, use the preceding to obtain the result.

There is also a product rule for forms.

Theorem 12 (Wedge Product Rule). Suppose ϕ is a k-form and ψ is

an l-form. Then d( ϕ ∧ ψ) = dϕ ∧ ψ + (−1)k ϕ ∧ dψ.

Note that the product rule does not depend on l.

3.4 some basic forms: work, flux, and mass

There are three kinds of forms that will characterize the integral the-
orems. They come from their interpretations in physics and engineer-
ing. Also covered in Hubbard and Hubbard is a 0-form field, which
is simply a function that returns a real number.

3.4.1 Work

In physics, the work done by a force is the product of that force’s

magnitude and the distance the object moves. With this in mind, we
define the following form.

Definition 4 (Work Form). Let F be a vector field in Rn . The work

form WF of F is defined as

WF := ∑ Fj dx j .
1≤ j ≤ n

18
 3.4 some basic forms: work, flux, and mass

For example, a work form on R2 looks like 2x dx + 3y dy and one

on R4 looks like sin t dt + cos 2xt dx − 3yz dy − 2tz2 dz. Work forms do
not need every variable present; for example, the force of gravity has
the work form − gm dz when associated to the vector field F( x, y, z) =
− gme3 (where e3 is the standard basis vector associated with the z
direction in R3 ).
If we wish to compute the work done by a vector field over a path,
we can use a line integral.

Definition 5 (Work Integral). The work done by a vector field F over

an oriented curve C is the line integral
Z
WF .
C

Work is measured in units of energy, such as the Joule or the erg.

The vector field associated to a work form gives a measure of energy
per unit length.

3.4.2 Flux

Flux is the measure of flow through an object. The term is commonly

used in electromagnetism to describe the flow of electric field lines
through a surface (Serway, p. 673). Gauss’ law gives a way to com-
pute flux. Here we define a differential form used in computing flux
for vector fields in R3 .

Definition 6 (Flux Form). A flux form ΦF is the 2-form field

F1 dx ∧ dy − F2 dx ∧ dz + F3 dy ∧ dz.

The minus sign on the F2 term is necessary since it arises when

computing the determinant det[F, v, w] where v and w are the vectors
(usually derivatives) that the flux form acts on.
Computing the flux through a surface is similar to computing work.

Definition 7 (Flux Integral). The flux of a vector field F through the

surface S is the surface integral
Z
ΦF .
S

Flux is measured in mass per unit area.

19
 3.5 integral theorems in the language of forms

3.4.3 Mass

The last form that will be used in the integral theorems is the mass
form.

Definition 8. The mass form of a function f defined on a set U ⊂ R3

is defined by

M f := f dx ∧ dy ∧ dz.

This comes from the expression f (x) det[v1 , v2 , v3 ], where x, v1 , v2 , v3 ∈

R3 and v1 , v2 , v3 are the vectors the form acts on, usually derivatives.
We can compute masses of objects given their densities.

Definition 9 (Mass Integral). The mass of an object with density

function f is
Z
Mf .
[γ(U )]

The mass form is measured in mass per unit volume, or in the case
of electrostatics, charge per unit volume.

3.4.4 Derivatives

To obtain a different version of Maxwell’s equations using forms later,

we need the following facts for functions and vector fields on R3 and
the derivatives of the associated forms.

1. d f = W∇F

2. dWF = Φ∇×F

3. dΦF = M∇·F

So, the exterior derivative takes functions to work forms of gradi-

ents, work forms to flux forms of curls, and flux forms to mass forms
of divergences.

3.5 integral theorems in the language of forms

All the classical results in vector calculus can be cast in the language
of forms. In particular, the integral theorems can be expressed us-
ing forms, usually in a more concise manner. The results presented
here are given in the order in which they appear in Hubbard’s Vector
Calculus, Linear Algebra, and Differential Forms.

20
 3.5 integral theorems in the language of forms

3.5.1 Stokes’ Theorem, Generalized

All of the integral theorems in calculus can be summarized with the

most general form of Stokes’ Theorem. The general statement re-
quires knowledge of manifolds - a generalization of curves and sur-
faces to higher dimensions. Part of the general theorem mentions
a “compact piece-with-boundary.” A piece-with-boundary of a mani-
fold M is a compact set X ⊂ M - a set that is both closed and bounded
- such that the set of non-smooth points of the boundary of X in M
has volume zero, and the smooth boundary has finite volume.

Theorem 13 (Generalized Stokes’ Theorem). Let X be a compact

piece-with-boundary of an oriented manifold M. Give ∂X the boundary
orientation - either “clockwise” (direct) or “counterclockwise” (indirect).
Let ϕ be a (k − 1)-form defined on an open set containing X. Then
Z Z
ϕ= dϕ.
∂X X

Stokes’ Theorem says that the integral of a form taken over the
boundary of a manifold is equal to the integral of the exterior deriva-
tive of the form taken over the manifold. Compare this to the one-
variable Fundamental Theorem of Calculus:
Z b
f ( x ) dx = F (b) − F ( a)
a
where the right side is an “integral” over the boundary of the inter-
val [ a, b]: a 1-dimensional manifold in R. A general proof of Stokes’
Theorem is difficult to write: Hubbard and Hubbard give an infor-
mal proof that relies on estimating the integral using parallelograms
in n dimensions; a much more generalized proof is outlined in the
appendix of their text. Their generalized proof requires a solid foun-
dation in real analysis, linear algebra, and the theory of calculus on
manifolds.

3.5.2 The Fundamental Theorem for Line Integrals

We now express the Gradient Theorem without using the word “gra-
dient.”

Theorem 14 (Fundamental Theorem for Line Integrals: The Forms

Version). Let C be an oriented curve γ(t) (a ≤ t ≤ b) in Rn with an
oriented boundary that designates b as the final point and a as the initial
point. Let f be a function defined on a neighborhood of C then

21
 3.5 integral theorems in the language of forms

Z
d f = f (γ(b)) − f (γ( a)).
C

Here, the exterior derivative is the gradient. We are also requiring

that d f be exact so the potential function f can be constructed. Hence,
if constructed appropriately, d can be viewed as a conservative vector
field.

3.5.3 Green’s Theorem

We now express Green’s theorem using work forms.

Theorem 15 (Green’s Theorem: Work Forms Version). Suppose

that S is a bounded region in the plane whose boundary consists of
some number of curves Cj . Give the union of these curves the boundary
orientation (for simplicity, the counterclockwise diretion). Let F be a
vector field defined on a neighborhood of S. Then
Z Z
dWF = ∑ WF .
S 1≤ j ≤ n C
j

This still amounts to writing P dx + Q dy in the right integrand

∂Q
∂y − ∂x in the left integrand. We still also have the fact that if the
and ∂P
work form is an exact differential, the integral is zero.

3.5.4 Stokes’ Theorem for Three Dimensions

A forms version of the traditional Stokes’ theorem uses flux forms.

Theorem 16 (Traditional Stokes’ Theorem: The Forms Version).

Let S be an oriented surface in R3 (usually, by the outward normal)
bounded by a curve C given the boundary orientation (again, usually
counterclockwise). Let ϕ be a 1-form field defined on a neighborhood of
S. Then
Z Z
dϕ = ϕ.
S C

Since ϕ is a 1-form, it represents a work form. The exterior deriva-

tive of such a form is a flux form. Flux integrals are surface integrals,
and the function on the right is actually the curl of the vector field
represented by ϕ.

22
 3.5 integral theorems in the language of forms

3.5.5 Divergence Theorem

The divergence theorem, when written using forms, looks similar to

the forms version of Green’s theorem.

Theorem 17 (Divergence Theorem: The Forms Version). Let X be

a bounded domain in R3 with the standard orientation of space (this is
given by the determinant of a change of basis matrix; usually this does
not affect the integral). Let ∂X be a union of surfacees S j each oriented
by an outward normal. If ϕ is a 2-form field defined on a neighborhood
of X, then
Z Z
dϕ = ∑ ϕ.
X 1≤ j ≤ n S
j

This theorem still states to integrate the surface’s boundary using

the divergence of a vector field in the integrand. This vector field
comes from the flux form on the right side, and the divergence ap-
pears in the mass form on the left.

23
 4
M A X W E L L’ S E Q U AT I O N S

In electromagnetism, the behavior of electric fields, magnetic fields,

and closed surfaces can be characterized by a set of four equations.
These equations, developed by James Clerk Maxwell (for whom the
equations are named for), describe both the flux of magnetic and elec-
tric fields and the imaginary rotation of these fields. Appropriately,
the divergence and curl operators are used to describe the respective
property.
There are two ways Maxwell’s equations are traditionally written.
One formulation uses partial differential equations that arise from
the vector operators, while another expresses the same quantities as
integral equations. In order to write these equations, we must first
define three vector fields that exist in space-time.
The first field is E, the electric field. It is defined as the force a test
charge experiences at a point divided by the charge itself. In symbols,

F
E= ,
q0
where F is the force the test charge experiences in Newtons and q0 is
the charge in coulombs (Serway, 688).
The second vector field is the magnetic field B. To define this field,
it is useful to examine the force exerted by this field. Suppose a
charge q moves with velocity v in a magnetic field B. Then the force
in this magnetic field can be described by the Lorentz force law:

F = qv × B

The force then acts in a direction normal to the plane containing

the velocity vector and the magentic field at a point. The magnitude
of the magnetic field at a point can be expressed using the formula for
|F|
the magnitude of the cross product; this gives |B| = , where
q|v| sin θ
θ is the angle between the vectors B and v. Since the force is normal
to the magnetic field and the velocity, the work done by B is therefore
zero since no component of force is in the direction of motion.
Note that in the presence of an electric field, the Lorentz law can
be written as F = q(E + v × B) (Griffiths, 204).

24
 4.1 the maxwell equations - traditional forms

The last field needed to generate Maxwell’s equations is the current

density J. This density is measured in statcolumbs per cubic centime-
ter in the centimeter-gram-second system. This vector field appears
in the Maxwell-Ampere law of Maxwell’s equations.

4.1 the maxwell equations - traditional forms

There are four equations that characterize electricity and magnetism.

There are two tradional ways that the Maxwell equations can be writ-
ten.

4.1.1 Partial Differential Equations

It is common to see the Maxwell equations written as partial differ-

ential equations via vector operators. In this fashion, the equations
are as follows, with c representing the speed of light and ρ the charge
density:

1 ∂B
− = curl E
c ∂t
∇·B =0
1 ∂E 4πJ
= curl B −
c ∂t c
∇·E = 4πρ

The first equation is Faraday’s law; the second is Gauss’s law in

magnetic fields; the third, Ampere’s law; the fourth, Gauss’s law in
electric fields. This representation is given in the Hubbard text.

4.1.2 Integral Equations

Some texts, such as Serway’s Physics for Scientists and Engineers also
give integral representations for Maxwell’s equations. The integrals
can be checked for validity using both Stokes’ and Gauss’ theorems
in the appropriate contexts.
It is easy to obtain an integral representation for the second equa-
tion. Taking a triple integral on both sides and applying the diver-
gence theorem backwards gives a surface integral of the magnetic
field equal to zero.
For the fourth equation, the divergence theorem produces a surface
integral of the electric field on the left side and a triple integral of the
charge density on the right. But the integral of a density function
gives the total amount of the measurement - in this case, charge. So,
ZZ
E · dS = 4πq,
σ

25
 4.2 differential forms and maxwell’s equations

q
or in SI units, e0 .
The curl equations are slightly more difficult to obtain an integral
representation for. First, consider the integral
I
E · dr
C

taken over a closed curve C. Assuming the conditions for Stokes’

theorem hold for E and C,

I ZZ
E · dr = curl E · dS
C σ
1 ∂B
ZZ
= − · dS
c ∂t
σ
1 d
ZZ
=− B · dS
c dt
σ

where the last equality follows from differentiation under the integral
sign. As a result, we have that the work done by the electric field
along a closed curve C is proportional to the time derivative of the
flux of B through the surface S bounded by C.
To obtain an integral form for the remaining equation, begin with
the line integral of B over some closed curve C. Again, under appro-
priate conditions, the following holds:

I ZZ
B · dr = curl B · dS
C σ
1 ∂E 4πJ
ZZ
= + · dS
c ∂t c
σ
1 d 16π 2 I
ZZ
=− B · dS +
c2 dt c2
σ

where I represents the total current. Again, the derivative of flux

appears in this integral representation.

4.2 differential forms and maxwell’s equations

We now come to a large application of differential forms. Maxwell’s

equations can be written concisely in two lines using three differential
forms. Define the following forms:

1. The Faraday 2-form F = WE ∧ c dt + ΦB

2. The Maxwell 2-form M = WB ∧ c dt − ΦE

3. The current 3-form J = 1c ΦJ ∧ c dt − Mρ

26
 4.2 differential forms and maxwell’s equations

We will use the exterior derivative and the differential equations

version of Maxwell’s equations to derivative the differential forms
version.

4.2.1 The Faraday Form

Begin with the Faraday form F and take its exterior derivative.

dF = Φ∇×E ∧ c dt + dΦB
= Φ∇×E ∧ c dt + M∇·B + Φ 1 ∂B ∧ c dt
c ∂t
 
= Φ∇×E+ 1 ∂B ∧ c dt + Mdiv B
c ∂t

=0

Each equality used either the derivatives of the basic forms or the
partial differential equations from Maxwell. Thus, for two equations,
we have the compact formulation dF = 0.

4.2.2 The Maxwell Form

We use the same process as above to find the exterior derivative of

the Maxwell form.

dM = Φ∇×B ∧ c dt − M∇·E − Φ 1 ∂E ∧ c dt
c ∂t
 
= 4π Φ J ∧ c dt − Mρ
c

= 4πJ

The last line relied on converting to the current form J to condense

the equation. Thus the remaining two of Maxwell’s equations can be
written as dM = 4πJ.
Differential forms allow for many concepts in electromagnetism to
be simplified by separating spatial variables and the time variable.
The use of differential forms in physics can be extended to quantum
mechanics, as seen in Garrity.

27
 4.2 differential forms and maxwell’s equations

references
Fleisch, D. (2008). A Student’s Guide to Maxwell’s Equations.
Cambridge University Press.

Garrity, T. A. (2015). Electricity and Magnetism for Mathematicians.

Cambridge University Press.

Griffiths, D. J. (1999). Introduction to Electrodynamics (3rd ed.).

Prentice Hall.

Hubbard, J. H., & Hubbard, B. B. (2009). Vector Calculus, Linear

Algebra, and Differential Forms (4th ed.). Matrix Editions.

Serway, R. A., & John W. Jewett, J. (2008). Physics for Scientists and
Engineers with Modern Physics (7th ed.). Brooks/Cole.

Stewart, J. (2012). Calculus: Early Transcendentals (7th ed.).

Brooks/Cole.

Strauss, W. A. (2008). Partial Differential Equations: An Introduction

(2nd ed.).

Zill, D. G. (2009). A First Course in Differential Equations with

Modeling Applications (9th ed.). Brooks/Cole.

28