Professional Documents
Culture Documents
Donu Arapura
February 5, 2016
The calculus of differential forms give an alternative to vector calculus which
is ultimately simpler and more flexible. Unfortunately it is rarely encountered
at the undergraduate level. However, the last few times I taught undergraduate
advanced calculus I decided I would do it this way. So I wrote up this brief
supplement which explains how to work with them, and what they are good
for. By the time I got to this topic, I had covered a certain amount of standard
material, which is briefly summarized at the end of these notes.
My thanks to Jo
ao Carvalho, John Crow, Mat
us Goljer and Josh Hill for
catching some typos.
Contents
1 1-forms
1.1 1-forms . . . . . . . . . . . . .
1.2 Exactness in R2 . . . . . . . . .
1.3 Parametric curves . . . . . . .
1.4 Line integrals . . . . . . . . . .
1.5 Work . . . . . . . . . . . . . .
1.6 Greens theorem for a rectangle
1.7 Exercise Set 1 . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2
2
3
3
4
7
8
8
2 2-forms
2.1 2-forms . . . . . . . . . . . . . . . .
2.2 Exactness in R3 and conservation of
2.3 d of a 2-form and divergence . . .
2.4 Poincares lemma for 2-forms . . . .
2.5 Exercise Set 2 . . . . . . . . . . . .
. . . .
energy
. . . .
. . . .
. . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
10
10
12
13
14
15
3 Surface integrals
3.1 Parameterized Surfaces . . . .
3.2 Surface Integrals . . . . . . .
3.3 Surface Integrals (continued)
3.4 Length and Area . . . . . . .
3.5 Exercise Set 3 . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
17
17
19
21
23
25
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Stokes Theorem
4.1 Green and Stokes . . . .
4.2 Proof of Stokes theorem
4.3 Cauchys theorem* . . .
4.4 Exercise Set 4 . . . . . .
5 Gauss theorem
5.1 Triple integrals . . .
5.2 Gauss theorem . . .
5.3 Proof for the cube .
5.4 Gravitational Flux .
5.5 Laplaces equation* .
5.6 Exercise Set 5 . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
28
28
29
31
34
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
36
36
36
37
38
40
41
6 Beyond 3 dimensions*
43
6.1 Beyond 3D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
6.2 Maxwells equations in R4 . . . . . . . . . . . . . . . . . . . . . . 46
7 Further reading
48
1
1.1
1-forms
1-forms
where i, j, k are the standard unit vectors along the x, y, z axes. Under this set
up, the gradient f corresponds to df . Thus it might seem that all we are doing
is writing the previous concepts in a funny notation. However, the notation is
very suggestive and ultimately quite powerful. Suppose that that x, y, z depend
on some parameter t, and f depends on x, y, z, then the chain rule says
f dx f dy f dz
df
=
+
+
dt
x dt
y dt
z dt
Thus the formula for df can be obtained by canceling dt.
1.2
Exactness in R2
1.3
Parametric curves
Before discussing line integrals, we have to say a few words about parametric
curves. A parametric curve in the plane is vector valued function C : [a, b] R2 .
3
dt
du
6= 0.
x = f (h(u)), y = g(h(u))
dt
It the derivative du
is everywhere positive, we want to view the oriented curves
C and C 0 as the equivalent. If this derivative is everywhere negative, then C
and C 0 are equivalent. For example, the curves
C : x = cos , y = sin , 0 2
C 0 : x = sin t, y = cos t, 0 t 2
represent going once around the unit circle counterclockwise and clockwise respectively. So C 0 should be equivalent to C. We can see this rigorously by
making a change of variable = /2 t.
It will be convient to allow piecewise C 1 curves. We can treat these as unions
of C 1 curves, where one starts where the previous one ends. We can talk about
parametrized curves in R3 in pretty much the same way.
1.4
Line integrals
Now comes the real question. Given a differential F dx + Gdy, when is it exact?
Or equivalently, how can we tell whether a force is conservative or not? Checking
that its closed is easy, and as weve seen, if a differential is not closed, then
it cant be exact. The amazing thing is that the converse statement is often
(although not always) true:
THEOREM 1.4.1 If F (x, y)dx + G(x, y)dy is a closed form on all of R2 with
C 1 coefficients, then it is exact.
dy
dx
+ G(x(t), y(t))
dt
F (x(t), y(t))
dt
dt
y
x
dx + 2
dy = sin (cos )0 d + cos (sin )0 d = d
x2 + y 2
x + y2
and therefore
Z
y
x
2
dx + 2
dy =
2
x
+
y
x
+
y2
C
d = 2
0
Z
F dx + Gdy =
F dx + Gdy
C
F dx + Gdy
C0
While were at it, we can also define a line integral in R3 . Suppose that
F dx + Gdy + Hdz is a differential form with C 1 coeffients. Let C : [a, b] R3
be a piecewise C 1 parametric curve, then
DEFINITION 1.4.4
Z
F dx + Gdy + Hdz =
C
Z
a
dy
dz
dx
F (x(t), y(t), z(t))
+ G(x(t), y(t), z(t))
+ H(x(t), y(t), z(t))
dt
dt
dt
dt
C2
Its quite easy to see why this works. If = df and C1 : [a, b] R3 then
Z
Z b
df
dt
df =
C1
a dt
by the chain rule. Now the fundamental theorem of calculus shows that the
last integral equals f (C1 (b)) f (C1 (a)), which is to say the value of f at the
endpoint minus its value at the starting point. A similar calculation shows that
the integral over C2 gives same answer. If the C is closed, which means that
the starting point is the endpoint, then this argument gives
R
COROLLARY 1.4.6 If is exact and C is closed, then C = 0.
Now we can prove theorem 1.4.1. If F dx + Gdy is a closed form on R2 , set
Z
f (x, y) =
F dx + Gdy
C
(x,y)
(x,0)
(0,0)
f
=
G(x, t)dt
y
y 0
= G(x, y)
6
Slightly trickier is
f
x
Z y
F (x, 0)dt +
G(x, t)dt
x 0
0
Z y
G(x, t)
= F (x, 0) +
dt
x
Z0 y
F (x, t)
= F (x, 0) +
dt
t
0
= F (x, 0) + F (x, y) F (x, 0)
=
= F (x, y)
The same proof works if if we replace R2 by an open rectangle. However, it
will fail for more general open sets. For example,
x2
x
y
dx + 2
dy
2
+y
x + y2
is C 1 1-form on the open set {(x, y) | (x, y) 6= (0, 0)} which is closed. But it is
not exact (see exercise 6). In more advanced treatments, this failure of closed
forms to be exact can be measured by something called the de Rham cohomology
of the set.
1.5
Work
Line integrals have many important uses. One very direct application in physics
comes from the idea of work. If you pick up a rock off the ground, or perhaps
roll it up a ramp, it takes energy. The energy expended is called work. If youre
moving the rock in straight line for a short distance, then the displacement can
be represented by a vector d = (x, y, z) and the force of gravity by a vector
F = (F1 , F2 , F3 ). Then the work done is simply
F d = (F1 x + F2 y + F3 z).
On the other hand, if you decide to shoot a rocket up into space, then you would
have to take into account that the trajectory c may not be straight nor can the
force F be assumed to be constant (its a vector field). However as the notation
suggests, for the work we would now need to calculate the integral
Z
F1 dx + F2 dy + F3 dz
c
F ds
c
1.6
Let R be the rectangle in the xy-plane with vertices (0, 0), (a, 0), (a, b), (0, b).
Let C be the boundary curve of the rectangle oriented counter clockwise. Given
C 1 functions P (x, y), Q(x, y) on R, the fundamental theorem of calculus yields
Z b
Z
Q
dxdy =
[Q(a, y)) Q(0, y)]dy =
Q(x, y)dy
0
D x
C
Z a
Z
ZZ
P
dydx =
[P (x, b) P (x, 0)]dx =
P (x, y)dx
0
C
D y
ZZ
ZZ
P dx + Qdy =
Q P
x
y
dxdy
x
y
is the integrand of the right in Greens theorem. In general, one can discover
the formula for the other components of (P i + Qj + Rk) by expressing the
integrals of P i + Qj + Rk around the boundaries of rectangles in the xz and yz
planes and rewriting them as double integrals. To make a long story short,
(P i + Qj + Rk) = (Ry Qz )i + (Qx Py )k + (Pz Rx )j
(In practice, this is often written as a determinant
i
j
(P i + Qj + Rk) = x
y
P
Q
k
z .
R
But this should really be treated as a memory aid and nothing more.)
1.7
Exercise Set 1
2
2.1
2-forms
2-forms
dxdy
x
y
in a more direct way than was done above. But first we need to understand
how to work with expressions of the form F dxdy. In fact, for reasons that will
be clear later, we wish to introduce symbols dx dy, . . . which carries slightly
more information; namely a sense of direction.
The cross product of vectors uv is a very useful operation in 3 dimensional
geometry. Its length gives the area of the parallelogram spanned by u, v, and it
determines the plane containing this parallelogram. There is no direct analogue
of the cross product in higher dimensions. However, there are certain features
which do generalize. The essential idea is to define new objects called 2-vectors
which are to planes what ordinary vectors are to lines. A vector in the usual
sense is a directed line segment. It is determined by its length, the line it lies
on, and a choice of one of two possible directions along the line. The basic
building block of a 2-vector is an oriented parallelogram (we usually omit the
modifier oriented). It has three attributes: its area, the plane on which it
lies, and the orientation which is a choice of direction for walking around its
perimeter. Two oriented parallelograms are considered equal if the areas are
equal, the planes are parallel, and the orientations match. A parallelogram is
equal to zero precisely when its area is, in which case the other attributes can
be arbitrary. Given a parallelogram P and a number c, we define cP to be a
parallelogram with the same plane, and area given by |c| area(P ) and the same
orientation if c > 0 and the opposite orientation if c < 0. Given two vectors
u, v we define the parallelogram u v to be given as follows ( is pronounced
wedge).
u
Evidently
v u = u v
uu=0
and
c(u v) = (cu) v = u (cv)
10
The sum of vectors can be defined geometrically using the parallelogram law.
There is a limited version of this for 2-vectors: if two parallelograms share one
side, then the remaining sides can be added using the parallelogram law. In
terms of algebra, this is just the distributive law:
u v + u w = u (v + w)
Unfortunately there is no simple geometric rule for adding more complicated
pairs of parallelograms, so we just add them formally without worrying about
the meaning. A 2-vector is a finite sum of oriented parallelograms. The set
of 2-vectors with these operations form a vector space. In other words, all the
expected rules for + and apply. We want to emphasize that the formalism of
2-vectors works in any dimension. In R3 , we can identify 2-vectors with vectors
via u v u v. However, we will usually refrain from doing this, because
it would lead to more confusion in the long run. It is best to think of 2-vectors
and vectors as separate objects. For this reason, is often called the exterior
product.
A 2-form is like a 2-vector but built using forms. On R3 , this would be an
expression:
F (x, y, z)dx dy + G(x, y, z)dy dz + H(x, y, z)dz dx
where F, G and H are functions defined on an open subset of R3 . Any wedge
product of two 1-forms can be put in this format. For example, using the above
rules, we can see that
(3dx + dy) (dx + 2dy) = 6dx dy + dy dx = 5dx dy
The real significance of 2-forms will come later when we do surface integrals.
A 2-form will be an expression that can be integrated over a surface in the same
way that a 1-form can be integrated over a curve.
In order to make the comparison with traditional vector calculus, we note
that we can convert vector fields to 2-forms and back
F1 i + F2 j + F3 k F1 dy dz + F2 dz dx + F3 dx dy,
Earlier, we learned how to convert a vector field to a 1-form:
F1 i + F2 j + F3 k F1 dx + F2 dy + F3 dz
To complete the triangle, we can interchange 1-forms and 2-forms using the so
called Hodge star operator.
(F1 dx + F2 dy + F3 dz) = F1 dy dz + F2 dz dx + F3 dx dy
(F1 dy dz + F2 dz dx + F3 dx dy) = F1 dx + F2 dy + F3 dz
Given a 1-form F (x, y, z)dx + G(x, y, z)dy + H(x, y, z)dz. We want to define
its derivative d which will be a 2-form. The rules we use to evaluate it are:
11
d( + ) = d + d
d(f ) = (df ) + f d
d(dx) = d(dy) = d(dz) = 0
where and are 1-forms and f is a function. Recall that
df = fx dx + fy dy + fz dz
where fx =
f
x
2.2
12
To appreciate the importance of this concept, recall from physics that the
kinetic energy of a particle of constant mass m and velocity
dx dy dz
v=
, ,
dt dt dt
is
K=
1
1
m||v||2 = mv v.
2
2
dv
=F
dt
dv
+ v P = 0
dt
(1)
(2)
2.3
Earlier we introduced 2-vectors which correspond to sums of oriented parallelograms. We also have 3-vectors which correspond to oriented parallelopipeds.
Given three vectors u, v, w R3 , we think of the 3-vector u v w as the
oriented parallelopiped with u, v, w as the first, second and third sides. The
only attribute that will distinguish one from another is the oriented volume,
which is the usual volume if u, v, w is right-handed, otherwise it is minus the
usual volume. With these rules, we see that
u v w = v u w = v w u = . . .
A 3-form is simply an expression
f (x, y, z)dx dy dz = f (x, y, z)dy dx dz = f (x, y, z)dy dz dx = . . .
These are things that will eventually get integrated over solid regions. The
important thing for the present is an operation which takes 2-forms to 3-forms
once again denoted by d.
d(F dy dz + Gdz dx + Hdx dy) = (Fx + Gy + Hz )dx dy dz
13
Its probably easier to understand the pattern after converting the above 2form to the vector field V = F i + Gj + Hk. Then the coeffiecient of dx dy dz
is the divergence
V = Fx + Gy + Hz
So far weve applied d to functions to obtain 1-forms, and then to 1-forms to
get 2-forms, and finally to 2-forms. The real power of this notation is contained
in the following simple-looking formula
PROPOSITION 2.3.1 d2 = 0
What this means is that given a C 2 real valued function defined on an open
subset of R3 , then d(df ) = 0, and given a 1-form = F dx + Gdy + Hdz with
C 2 coefficents defined on an open subset of R3 , d(d) = 0. Both of these are
quite easy to check:
d(df ) = (fyx fxy )dx dy + (fzy fyz )dy dz + (fxz fzx )dz dx = 0
d(d) = [Gxz Fyz + Hyx Gzx + Fzy Hxy ]dx dy dz = 0
In terms of standard vector notation this is equivalent to
(f ) = 0
( V) = 0
2.4
Poincar
es lemma for 2-forms
(3)
14
2.5
Exercise Set 2
Let = xdx + ydy + zdz, = zdx + xdy + ydz and = xydz in the following
problems.
1. Calculate
(a)
(b)
(c)
(d) ( + ) ( + )
2. Calculate
15
(a) d
(b) d
(c) d( + )
(d) d(x)
3. Given = f dx + gdy + hdz such that dz = 0, what can we conclude
about f, g and h?
4. (a) Let = F dx + Gdy + Hdz and let C be the straight line connecting
(0, 0, 0) to (x0 , y0 , z0 ) show that
Z
Z
=
16
Surface integrals
3.1
Parameterized Surfaces
x=u
y=
v
z = 1 u2 v 2
2
u + v2 1
2) The upper half sphere can be parameterized using spherical coordinates
x = sin() cos()
y = sin() sin()
z = cos()
0 /2, 0 < 2
(Since there is more than one convention, we should probably be clear about
this. For us, is the polar angle, and is the angle measured from the positve
z-axis.)
3) The upper half sphere can be parameterized using cylindrical coordinates
x = r cos()
y=
r sin()
z = 1 r2
0 r 1, 0 < 2
An orientation on a curve is a choice of a direction for the curve. For a
surface an orientation is a choice of up or down. The easist way to make
this precise is to view an orientation as a choice of (an upward, downward,
outward or inward pointing) unit normal vector field n on S. A parameterized
surface S
x = f (u, v)
y = g(u, v)
z = h(u, v)
(u, v) D
is called smooth provided that f, g, h are C 1 , the function that they define from
D R3 is one to one, and the tangent vector fields
x y z
Tu =
,
,
u u u
x y z
Tv =
,
,
v v v
17
are linearly independent. In this case, once we pick an ordering of the variables
(say u first, v second) an orientation is determined by the normal
n=
Tu Tv
||Tu Tv ||
n
Tv
S
u=constant
v=constant
FIGURE 1
x = f (u, v)
y = g(u, v)
z = h(u, v)
(u, v) D R2
then the image of C in R3 will be the boundary of S. For example, the boundary
18
x = sin() cos()
y = sin() sin()
z = cos()
0 /2, 0 < 2
is the circle C given by
x = cos(), y = sin(), z = 0, 0 2
In what follows, we will need to match up the orientation of S and its boundary
curve. This will be done by the right hand rule: if the fingers of the right hand
point in the direction of C, then the direction of the thumb should be up.
n
FIGURE 2
3.2
Surface Integrals
x = f (u, v)
y = g(u, v)
z = h(u, v)
(u, v) D
with orientation corresponding to the ordering u, v. The symbols dx etc. can
be converted to the new coordinates as follows
dx =
x
x
du +
dv
u
v
19
dy =
y
y
du +
dv
u
v
dx dy
x
x
y
y
=
du +
dv
du +
dv
u
v
u
v
x y x y
=
du dv
u v
v u
(x, y)
=
du dv
(u, v)
(x, y)
d d = cos() sin()d d
(, )
So
ZZ
/2
dx dy =
S
cos() sin()dd =
0
On the other hand if use the same surface parameterized using cylindrical
coordinates
(x, y)
dx dy =
dr d = rdr d
(r, )
Then
ZZ
dx dy =
S
rdrd =
0
which leads to the same answer as one would hope. The general result is:
THEOREM 3.2.2 Suppose that an oriented surface S has two different smooth
C 1 parameterizations, then for any 2-form , the expression for the integrals of
calculated with respect to both parameterizations agree.
(This theorem needs to be applied to the half sphere with the point (0, 0, 1)
removed in the above examples.) Let us give the proof. Suppose that u, v
is a parameterization as above, and u = p(s, t), v = q(s, t) for one to one C 1 functions p, q on a domain E in pq-plane. We also want to assume that (u,v)
(s,t) > 0
on E. This will ensure that s, t gives a new parameterization of S with the
20
correct orientation.
RR Let us assume for simplicity that we have = F dx dy.
Then computing S in two ways gives either
ZZ
F
(x, y)
dudv
(u, v)
(x, y)
dsdt
(s, t)
or
ZZ
E
We have to prove that these integrals are equal. The chain rule can be expressed
as a matrix equation
x x x x u u
s
t
v
s
t
= u
y
y
y
y
v
v
s
(4)
3.3
21
boundary
We will insist that if any two triangles touch, they either meet only at a
vertex, or they share an entire edge. In addition, any edge lies on at most two
triangles. So the picture below is not a surface from our point of view. We define
the boundary of a surface to be the union of all edges which are not shared. The
surface is called closed if the boundary is empty.
Given a surface which has been divided up into patches, we can integrate a
2-form on it by summing up the integrals over each patch. However, we require
that the orientations match up, which is possible if the surface has two sides.
Below is a picture of a one sided, or nonorientable, surface called the Mobius
strip.
22
Once we have picked an orientation of S, we get one for the boundary using
the right hand rule.
In many situations arising in physics, one needs to integrate a vector field
F = F1 i + F2 j + F3 k over a surface. The resulting quantity is often
R Rcalled a
flux.
We
will
simply
define
this
integral,
which
is
usually
written
as
F dS
S
RR
or S F n dS, to mean
ZZ
F1 dy dz + F2 dz dx + F3 dx dy
S
3.4
It is important to realize some line and surface integrals are not expressible as
integrals of differential forms in general. Two notable examples are the arclength
and area integrals.
23
whereas for 1-forms the integral would change sign. Nevertheless, ds (or more
accurately its square) is a sort of generalization of a differential form called a
tensor. To get a sense what this means, let us calculate the arclength of a curve
lying on a surface. Suppose that S is a parameterized surface given by
x = f (u, v)
y = g(u, v)
z = h(u, v)
(u, v) D
and suppose that C lies on S. This means that there are functions k, ` : [a, b]
R such that x = f (k(t), `(t)), . . . determines C. We can calculate the arclength
of C by applying the chain to the above integral all at once. Instead, we want
to break this down into a series of steps.
dx =
x
x
du +
dv
u
v
y
y
du +
dv
u
v
z
z
dz =
du +
dv
u
v
For the next step, we introduce a new product (indicated by juxtaposition)
which is distributive and unlike the wedge product is commutative. We square
the previous formulas and add them up. (The objects that we are getting are
tensors.)
2
2
x
x x
x
dx2 =
du2 + 2
dudv +
dv 2
u
u v
v
2
2
y
y y
y
dy 2 =
du2 + 2
dudv +
dv 2
u
u v
v
2
2
z
z z
z
dz 2 =
du2 + 2
dudv +
dv 2
u
u v
v
dy =
(5)
in the notation of section 3.1. The expression in (5) is called the metric tensor
of the surface. We can easily deduce a formula for arclength in terms of it:
s
Z
Z b
du 2
du dv
dv 2
E
+ 2F
+G
dt
ds =
dt
dt dt
dt
C
a
The area of S can also be expressed in terms of the metric tensor. First
recall that
DEFINITION 3.4.2 The area of S is given by
ZZ
ZZ
dS =
||Tu Tv ||dudv
S
area(S) =
sin dd = 4
0
as expected.
3.5
Exercise Set 3
Let = xdx + ydy + zdz, = zdx + xdy + ydz and = xydz in the following
problems.
1. Let D be the square 0 x 1, 0 y 1, z = 1 oriented with the upward
normal. Calculate
RR
(a) D
RR
(b) D
RR
(c) D
25
y = (a + b cos ) sin
z = b sin
RR
6. With T as in the last problem, compute T zdx dy.
27
Stokes Theorem
4.1
Stokes theorem is really the fundamental theorem of calculus for surface integrals. We assume that the surfaces can be triangulated.
THEOREM 4.1.1 (Stokes theorem) Let S be an oriented smooth surface
with smooth boundary curve C. If C is oriented using the right hand rule, then
for any C 1 1-form on an open set of R3 containing S,
ZZ
Z
d =
If the surface lies in the plane, it is possible make this very explicit:
THEOREM 4.1.2 (Greens theorem) Let C be a closed C 1 curve in R2
oriented counterclockwise and D be the interior of C. If P (x, y) and Q(x, y) are
both C 1 functions then
Z
ZZ
Q P
dxdy
P dx + Qdy =
x
y
C
D
As an application, Greens theorem shows that the area of D can be computed as a line integral on the boundary
ZZ
Z
dxdy =
ydx
D
=0
S
S1
S2
B
t
where t is time. If we integrate both sides over S, apply Stokes theorem and
simplify, we obtain Faradays law of induction:
Z
ZZ
E ds =
B n dS
t
C
S
To get a sense of what this says, imagine that C is wire loop and that we are
dragging a magnet through it. This action will induce an electric current; the
left hand integral is precisely the induced voltage and the right side is related to
the strength of the magnet and the rate at which it is being dragged through.
Stokes theorem works even if the boundary has several components. However, the inner an outer components would have opposite directions.
2
S
THEOREM 4.1.3 (Stokes theorem II) Let S be an oriented smooth surface with smooth boundary curves C1 , C2 . . .. If Ci is oriented using the right
hand rule, then for any C 1 1-form on an open set of R3 containing S,
ZZ
Z
Z
d =
+
+ ...
S
4.2
C1
C2
Suppose that
= f dx + gdy + hdz
3
x
y
z
x
y
z
+g
+ h )du + (f
+g
+ h )dv
u
u
u
v
v
v
Z
P (u, 0)du
=
0
Z
P (u, 1 u)du
Q(1 v, v)dv
Q(0, v)dv +
0
30
1v
1u
Pv dvdu
0
Z
Qu dudv
Si
Ci1
Ci3
Ci2
where Ci1 , Ci2 , Ci3 are the edges of Si . We separate these into exterior curves,
which are contained in the boundary C, and interior curves. A curve Ci1 is an
interior curve if it also occurs as an edge, say Cj1 of an adjacent triangle Sj . If
we take into account the orientations, then we see that Cj1 = Ci1 .
C i1
C j1
Therefore the sum of integrals along the interior curves cancel. The conclusion is that
ZZ
Z
Z
X
d =
=
4.3
exterior Cik
Cik
Cauchys theorem*
Recall
that a complex number is an expression z = a + bi where a, b R and
i = 1, so that i2 = 1. The components a and b are called the real and
imaginary parts of z. We can identify the set of complex numbers C with the
plane R2 = {(a, b) | a, b R}. Addition and subtraction of complex numbers
correspond to the usual vector operations:
(a + bi) (c + di) = (a c) + (b d)i
31
h0
f (z + h) f (z)
h
Notice that h is a complex number. For the limit to exist, we should get the
same value no matter how it approaches 0. If h = x approaches along the
x-axis, we get
f 0 (z)
=
=
y0
lim
y0
dx dy = 0
d(f (z)dz) =
y
x
x y
Therefore Stokes theorem implies what may be thought of as the fundamental
theorem of complex analysis:
THEOREM 4.3.1 (Cauchys theorem) If f (z) is analytic on a region with
boundary C then
Z
f (z)dz = 0
C
Cr
Here Cr is a circle of radius r around 0. We can parameterize this with the help
of Eulers formula by
z = r cos() + ri sin() = rei , 0 2
Then dz = riei d, so that
Z
Z
g(z)dz = ri
Cr
f (rei ) i
e d = i
rei
f (rei )d
4.4
Exercise Set 4
35
5
5.1
Gauss theorem
Triple integrals
Orientations are also implicit here. To be consistent with earlier rules, we have
to take
ZZZ
ZZZ
f (x, y, z)dy dx dz =
f (x, y, z)dxdydz
V
for example. The rationale may be clearer if we distinguish left and right handed
coordinate systems. The system (x, y, z) is right handed because the thumb of
the right hand points in the direction of the third coordinate axis, when the
fingers are curled from first to second. If we reorder, say to (y, x, z) then we get
a left handed system. For more general coordinate systems, such as spherical,
it may be harder to visualize what right handed means. So let us give a more
precise mathematical definition. Suppose that u, v, w is a new ordered system
of coordinates, we will say that it is right handed if
(x, y, z)
>0
(u, v, w)
holds everywhere; left handed means that this strictly negative. Now given ,
we can convert it to an expression g(u, v, w)du dv dw. Let W be the region
corresponding to V in uvw-space. Then
ZZZ
ZZZ
=
g(u, v, w)dudvdw
V
where we choose a plus sign if these are right handed coordinates, and a minus
sign if these are left handed.
5.2
Gauss theorem
This is also called the divergence theorem. The reason for this name becomes
clear if we express this in standard vector notation, where it reads
ZZZ
ZZ
F dV =
F n dS
V
36
dV
(v)dV =
t
V
V
which yields
ZZZ
(v) +
dV = 0.
t
V
The only way this can hold for all possible regions V is that the integrand
(v) +
=0
t
(7)
5.3
S1
S2
We will just give the proof for the easiest case, where we can choose right handed
coordinates u, v, w, so that V corresponds to the the unit cube W = {(u, v, w) |
0 u, v, w 1}. The boundary consist of 6 pieces, the top
T = {(1, v, w) | 0 v, w 1}
the bottom
B = {(0, v, w) | 0 v, w 1}
and 4 sides that we wont label.
The proof is similar to the proof of Stokes theorem. Let denote by the
conversion of to uvw-coordinates.
THEOREM 5.3.1 d = d
37
We skip the proof which would involve a messy computation. Now let us
write
= P dv dw + Qdw du + Rdu dv
Then
ZZ
ZZ
ZZ
P dv dw
=
S
P dv dw + . . .
P (1, u, v)dudv
=
0
P (0, u, v)dudv + . . .
(8)
ZZZ
d =
(Pu + Qv + Rw )dudvdw
P (0, u, v)dudv
0
and similar expression for the integrals of Qu and Rw . Adding these up gives
us (8).
5.4
Gravitational Flux
mr
r3
p
where r = ||r|| = x2 + y 2 + z 2 . This has singularity at 0, so it is a vector field
on R3 {0}. The corresponding 2-form is given by
= m
(9)
Let BR be the ball of radius R around 0, and let SR be its boundary. Since the
outward unit normal to SR is just n = r/r. One might expect that the flux
ZZ
ZZ
m
F ndS = 2
dS
R
SR
SR
=
m
m
area(SR ) = 2 (4R2 )
R2
R
38
However, this is really just a proof by notation at this point. To justify it, we
work in spherical coordinates. SR is given by
x = sin() cos()
y = sin() sin()
z = cos()
=R
0 , 0 < 2
Rewriting in these coordinates and simplifying:
= m sin()d d
Therefore,
ZZ
ZZ
SR
F ndS =
=
SR
sin()dd = 4m
0
if the interior V does not contain 0. On the other hand, if V contains 0, let BR
be a small ball contained in V , and let V BR denote part of V lying outside
of BR . We use the second form of Gauss theorem
ZZ
ZZ
ZZZ
=
d = 0
S
V BR
SR
We are subtracting the second surface integral, since we are supposed to use the
inner normal for SR . Thus
ZZ
= 4m
S
From here, we can easily extract an expression for the flux for several particles or even a continuous distribution of matter. If F is the force of gravity
associated to some mass distribution, for any closed surface S oriented by the
outer normal, then the flux
ZZ
F ndS
S
is 4 times theRRR
mass inside S. For a continuous distribution with density ,
this is given by
dV . Applying Gauss theorem again, in this case, yields
ZZZ
( F + 4)dV = 0
V
(10)
5.5
Laplaces equation*
f =
r
x
y
r
dx +
dy = dx + dy
x
y
r
r
Similarly
d =
So that
dr =
y
x
dx + 2 dy
2
r
r
x
y
dy dx = rd
r
r
40
x
1
y
dy 2 dx = dr
2
r
r
r
1
1
(dr d) = ( dx dy) =
r
r
d =
Thus
f
f
f = d
dr +
d
r
f
1 f
= d r d
dr
r
r
1 f
1 2f
2f
+ 2 + 2 2
r r
r
r
If f is radially symmetric, then it depends only on r so we obtain
1 d
d2 f
df
1 df
+ 2 =
r
=0
r dr
dr
r dr
dr
=
D
r
5.6
Exercise Set 5
RRR
6. Show that for any fixed value of m 6= 0, given in (9) is closed i.e. d = 0
but not exact. Therefore theorem 2.4.1 fails for R3 {0}.
7. Prove Greens identities
RR
RRR
(a) S f dg =
(df dg + f d dg)
V
(b) df dg = (fx gx + fy gy + fz gz )dx dy dz = dg df
RR
RRR
(c) S (f dg g df ) =
(f d dg gd df )
V
RR
Conclude that S (f dg g df ) = 0 if both f and g are harmonic.
42
6
6.1
Beyond 3 dimensions*
Beyond 3D
x1 = f1 (u1 , . . . uk )
x2 = f2 (u1 , . . . uk )
...
xn = fn (u1 , . . . uk )
(u1 , . . . uk ) D Rk open
x1
n
, . . . x
such that the map D Rn is one to one and the tangent vectors ( u
u1 ), . . .
1
x1
xn
( uk , . . . uk ) are linearly independent for all values of the coordinates (u1 , . . . uk ).
Given a k-form on Rn , we can express it as a linear combination of k-fold
wedges of dx1 . . . dxn , and then rewrite it as g(u1 , . . . uk )du1 . . . duk . The
surface integral is defined as
Z
Z
Z
= ...
g(u1 . . . uk )du1 . . . duk
(11)
M
In order to get a feeling for how this works, lets calculate the volume V
of the 4-dimensional ball B = {(x, y, z, t) | x2 + y 2 + z 2 + t2 R} of radius
R in R4 in two ways. This is a 4-manifold with boundary S = {(x, y, z, t) |
x2 + y 2 + z 2 + t2 = R}.
V can be expressed as the integral
ZZZZ
Z
V =
dxdydzdt =
dx dy dz dt
B
44
y
z
x
In these coordinate B is described as
0
0 2
0R
To simply computations, we note that form will get multiplied by the Jacobian
when we change coordinates:
dx dy dz dt =
(x, y, z, t)
d d d d
(, , , )
= 3 sin2 sin d d d d
Note that the Jacobian is positive, and this what it means to say the coordinate
system , , , is right handed or positively oriented. The volume is now easily
computed
Z R
Z
Z 2 Z
1
3 d
sin2 d
d
sin d = 2 R4
2
0
0
0
0
Alternatively, we can use Stokes theorem, to see that
Z
Z
V =
dx dy dz dt = tdx dy dz
B
The parameter x, y, z gives a left hand coordinate system on the upper hemisphere U = S {t > 0}. It is left handed because t, x, y, z is left handed on R4 .
45
For similar reasons, x, y, z gives a right handed system on the lower hemisphere
L where t < 0. Therefore
Z
Z
V =
tdx dy dz
tdx dy dz
L
Z ZUZ
p
= 2
R2 x2 y 2 z 2 dxdydz
x2 +y 2 +z 2 R
R
Z
=
R2 2 d
sin d
0
d
0
/2
Z
=
R4 sin2 cos2 d
8
0
6.2
1 2 4
R
2
Maxwells equations in R4
+
dx dy dt + . . .
x
y
z
x
y
t
Two of Maxwells equations for electromagnetism
B = 0,
E=
B
t
E = A4
46
A2
A3
A1
i+
j+
k
t
t
t
B=
E
+ 4J
t
where is the electric charge density, and J is the electric current. The first law
is really an analog of (10) for the electric field. After applying the divergence
theorem, it implies that the electric flux through a closed surface equals (4)
times the electric charge inside it. These last two Maxwell equations can also
be replaced by the single equation dF = 4J of 3-forms. Here
F = E3 dx dy + E1 dy dz + E2 dz dx B1 dx dt B2 dy dt B3 dz dt
and
J = dx dy dz J3 dx dy dt J1 dy dz dt J2 dz dx dt
(We have been relying on explicit formulas to avoid technicalities about the
definition of the -operator. In principle however, it involves a metric, and in
this case we use the so called Lorenz metric.)
Lets see how the calculus of differential forms can be used to extract a
physically meaningful consequence of these laws. Proposition 2.3.1 (in extended
1 2
d F = 0. Expanding this out yields
form) implies that dJ = 4
J3
dt dx dy dz
dz dx dy dt + . . . =
t
z
+ J dx dy dz dt = 0
t
Thus the expression in brackets is zero. This really an analog of the equation
(7). To appreciate the meaning integrate
t over a solid region V with boundary
S. Then this equals
ZZZ
ZZ
JdV =
J ndS
V
In other words, the rate of change of the electric charge in V equals minus the
flux of the current accross the surface. This is the law of conservation of electric
charge.
47
Further reading
For more information about differential forms, see the books [Fl, S, W]. Most
the physics background which is only used here for illustration, in any case
can be found in [Fe]. A standard reference for complex analysis is [A]. The
material of the appendix can be found in any book on advanced calculus. For
a rigorous treatment, see [R, S].
References
[A] L. Ahlfors, Complex Analysis
[Fe] R. Feynman et. al., Lectures on Physics
[Fl] H. Flanders, Differential forms and applications to the physical sciences
[R] W. Rudin, Principles of mathematical analysis
[S] M. Spivak, Calculus on manifolds
[W] S. Weintraub, Differential forms: theory and practice
48
A
A.1
To simplify the review, well stick to two variables, but the corresponding statements hold more generally. Let f (x, y) be a real valued function defined on open
subset of R2 . Recall that the limit
lim
f (x, y) = L
(x,y)(a,b)
means that f (x, y) is approximately L whenever (x, y) is close to (a, b). The
precise meaning is as follows. If we specified > 0 (say = 0.0005), then
we could pick a tolerance > 0 which would guarantee that |f (x, y) L| <
(i.e. f (x, y) agrees with L up to the first 3 digits for = 0.0005) whenever the
distance between (x, y) and (a, b) is less than . A function f (x, y) is continuous
at (a, b) if
lim
f (x, y)
(x,y)(a,b)
exists and equals f (a, b). It is continuous if it is so at each point of its domain.
We say that f is differentiable, if near any point p = (x0 , y0 , f (x0 , y0 )) the
graph z = f (x, y) can approximated by a plane passing through p. In other
words, there exists quantities A, B such that we may write
f (x, y) = f (x0 , y0 ) + A(x x0 ) + B(y y0 ) + remainder
with
|remainder|
=0
(x,y)(x0 ,y0 ) |(x, y) (x0 , y0 )|
lim
The last condition says that as the distance |(x, y) (x0 , y0 )| goes to zero, the
remainder goes to zero at an even faster rate. We can see that the coefficients
are nothing but the partial derivatives
A=
f
f (x0 + h, y0 ) f (x0 , y0 )
(x0 , y0 ) = lim
h0
x
h
f
f (x0 , y0 + h) f (x0 , y0 )
(x0 , y0 ) = lim
h0
y
h
There is a stronger condition which is generally easier to check. f is called
continuously differentiable or C 1 if it and its partial derivatives exist and are
continuous. Consider the following example
( 3
x
if (x, y) 6= (0, 0)
2
2
f (x, y) = x +y
0
if (x, y) = (0, 0)
B=
2
2
x
x +y
(x2 + y 2 )
49
has no limit as (x, y) (0, 0). To see this, note that along the x-axis y = 0, we
have f
x = 1. So the limit would have to be 1 if it existed. On the other hand,
along the y-axis x = 0, f
x = 0, which shows that there is no limit. So f (x, y)
is not C 1 .
Partial derivatives can be used to determine maxima and minima.
THEOREM A.1.1 If (a, b) is local maximum or minimum of a C 1 function
f (x, y) , then (a, b) is a critical point, i.e.
f
f
(a, b) =
(a, b) = 0
x
y
THEOREM A.1.2 (Chain Rule) If f, g, h : R2 R are C 1 functions, then
f (g(u, v), h(u, v)) is also C 1 and if z = f (x, y), x = g(u, v), y = h(u, v) then
z x z y
z
=
+
u
x u y u
z
z x z y
=
+
v
x v
y v
A function f (x, y) is twice continously differentiable or C 2 if is C 1 and if its
partial derivatives are also C 1 . We have the following basic fact:
THEOREM A.1.3 If f (x, y) is C 2 then the mixed partials
f
2f
=
yx
y x
2f
f
=
xy
x y
are equal.
If f is C 2 , then we have a Taylor approximation
f
1 2f
f
(a, b) (x a) +
(a, b) (y b) +
(a,
b)
(x a)2
f (x, y) f (a, b) +
x
y
2 x2
2
f
1 2f
+
(a, b) (x a)(y b) +
(a,
b)
(y b)2
yx
2 y 2
More precisely, the remainder, which is the difference of left and right, should
go to zero faster than |(x, y) (a, b)|2 goes to zero. Since it is relatively easy to
determine when quadratic polynomials have maxima or minima, this leads to
the second derivative test.
THEOREM A.1.4 A critical point (a, b) of a C 2 function f (x, y) is a local
minimum (respectively maximum) precisely when the matrix
!
2f
2f
(a,
b)
(a,
b)
2
x
yx
2f
yx (a, b)
2f
y 2 (a, b)
The above conditions are often formulated in a more elementary but ad hoc
way in calculus books. Positive definiteness is equivalent to requiring
2f
(a, b) > 0
x2
2
2
2
2
f
f
f
(a, b)
(a, b)
(a, b) > 0
x2
y 2
yx
A.2
Integral Calculus
Integrals can be defined using Riemanns method. This has some limitations
but its the easiest to explain. Given a rectangle R = [a, b] [c, d] R2 , choose
dc
integers m, n > 0 and let x = ba
m y = n . Choose a set of sample points
P = {(x1 , y1 ), . . . (xm , yn )} R with
(xi , yj ) Rij = [a + (i 1)x, a + ix] [b + (j 1)y, b + jy]
The Riemann sum
S(m, n, P ) =
f (xi , yj )xy
i,j
lim S(m, n, P )
m,n
This definition is not really that precise because we need to choose P for each
pair m, n. For the integral to exist, we really have to require that the limit
exists for any choice of P , and that any two choices lead to the same answer.
The usual way to resolve the above issues is to make the two extreme choices.
Define upper and lower sums
X
U (m, n) =
Mij xy
i,j
L(m, n) =
mij xy
i,j
where
Mij = max{f (x, y) | (x, y) Rij }
mij = min{f (x, y) | (x, y) Rij }
In the event that the maxima or minima dont exist, we should use the greatest
lower bound and least upper bound instead. As m, n the numbers L(m, n)
tend to increase. So their limit can be understood as the least upper bound, i.e.
the smallest number L L(m, n). Likewise we define the limit U as the largest
number U U (m, n). If these limits coincide, the common value is taken to be
ZZ
f (x, y)dxdy = L = U
R
THEOREM A.2.1
RR
R
The integral of
(
1 if (x, y) has rational coordinates
f (x, y) =
0 otherwise
would be undefined from the present point of view, because L = 0 and U = 1.
Although, in fact the integral can be defined using the more powerful Lebesgue
theory [R]; in this example the Lebesgue integral is 0.
For more complicated regions D R, set
ZZ
ZZ
f (x, y)D (x, y)dxdy
f (x, y)dxdy =
R
h(x)
f (x, y)dxdy =
D
f (x, y)dy dx
a
g(x)
=
=
x y x y
u v
v u
x x
u
v
.
det
y
y
u
52