Download as pdf or txt
Download as pdf or txt
You are on page 1of 7

A Summary on Vector Calculus

The study of electric and magnetic fields requires you to be familiar with scalars and vectors, and
various coordinate systems that exploit symmetries of the problems at hand.
The handling of scalar and vector fields, coordinate systems, and how these change is known as vector
calculus and is an immensely powerful tool used in many different fields. Due to the vast number of
applications, there is (unfortunately) an equally vast number of different conventions and notations in
the literature.
In this course, we will try and adhere to the notation used by the prescribed book Purcell. A summary
of different conventions that you might encounter are listed in section 2.

1 Scalars and Vectors


You should have, hopefully, already encountered and be somewhat familiar with scalars and vectors,
though we will try and give you a physical intuition behind these. First, lets recap the basic differences
between scalars and vectors:
scalar A scalar is a quantity which has only a single value (a magnitude), which no dependence on
direction. Examples include electric charge and temperature. They are not denoted in any special
way mathematically.
vector A vector is a quantity which has both a magnitude and a dependence on direction. Examples
include the electric and magnetic fields, and forces. They are denoted usually using a bold font in
typed notes, F , and usually with an error in handwritten notes, F~ . Occasionally, you will see it
written as Fi or F i .
The magnitude of the vector is denoted by kF k or sometimes with a non-bold symbol: F .
Tensor Although we will not be dealing with this much, a tensor can be seen as a generalization of a
vector, but with a dependence on multiple directions.
Unless analyzing only the simplest of problems, all these quantities will have a dependence on space
and time and is denoted by T (x, t) and F (x, t) for scalar and vector quantities. Writing out the
dependence on space and time can become cumbersome, so it is often implicit that T and F vary.

1.1 Inner Products


The inner product is an operation on two vectors which returns a scalar, and geometrically takes into
account how much of one vector goes in the direction of another vector,
a b := kakkbk cos , (1)
where is the angle between a and b.
Notice that there are two special cases. The dot product of a vector with itself is equal to the squared
magnitude of the vector:
a a kak2 . (2)

1
Also, if two vector are perpendicular, then the inner product is zero:

a b = 0 a b. (3)

Notice that if b is chosen such that kbk 1, which we will denote as b, then

a b = kak cos . (4)

This is the projection of a along the direction of b.

1.2 Coordinate Systems


Physically, a vector is a quantity which has both magnitude and direction; it is not just a series of numbers.
In other words, is a vector, but (0, 2, 1) is not. Mathematically though, it is hard to manipulate ,
so we decompose a vector in terms of a basis.
For 3 dimensions, we need 3 elements in a basis and well call them (e1 , e2 , e3 ). The basis need not
be orthogonal nor normalized, but they do need to be linearly independent. We will also assume that
they are normalized such that each kei k 1 (we dont have to do this, but it makes the following
maths a lot easier).
Now that we have this basis, we can then fully describe the vector v = by looking at how much
of is in each ei :
v (v e1 )e1 + (v e2 )e2 + (v e3 )e3 (5)
If we are using the same (e1 , e2 , e3 ) every time, it becomes cumbersome to always specify them so they
are often implicitly defined and then all we care about are the projections of v onto each basis. This is
where the notation of a vector being a list of numbers comes about:

v = v e1 , v e2 , v e3 = (v1 , v2 , v3 ) (6)

But always remember that the above notation is really an abbreviation for eq. (5), and without specifying
each ei , it is completely meaningless.
A particular choice of (e1 , e2 , e3 ) corresponds to a coordinate system. Some common coordinate
systems are listed in section 3. One of the most common coordinate system is the Cartesian coordinate
system in which (e1 , e2 , e3 ) = (x, y, z). In this coordinate system, the vector is written as:

v = (v x, v y, v z) = (vx , vy , vz ). (7)

1.2.1 Inner Products with Coordinate Systems


Now that we have introduced coordinate systems, lets look at what happens when we take two vectors
and try and get an intuitive understanding of the origin of the usual definition of the dot product you
have seen. Lets define

a = (a1 , a2 , a3 ), b = (b1 , b2 , b3 ), (8)

2
and calculate the inner product:

(a1 e1 + a2 e2 + a3 e3 ) (b1 e1 + b2 e2 + b3 e3 )
= a1 b1 e1 e1 + a1 b2 e1 e2 + a1 b3 e1 e3
+ a2 b1 e2 e1 + a1 b2 e1 e2 + a2 b3 e2 e3
+ a3 b1 e1 e1 + a3 b2 e1 e2 + a1 b3 e3 e3
T
a1 e1 e1 e1 e2 e1 e3 b1
= a2 e2 e1 e2 e2 e2 e3 b2 .
a3 e3 e1 e3 e2 e3 e3 b3

The matrix in the last line above is known as the metric tensor. Note that in this course, you will not
need to know much about metric tensors with the exception of the Euclidean metric and Minkowski
metric (both defined below). They are introduced here to give you a physical idea of where the dot
product comes from, and also to satisfy your curiosity.
In the vast majority of cases, we will also be dealing with an orthonormal basis. In this case, the
various inner products between the basis elements simplify greatly as ei ej ij , and the metric
tensor simplifies to the identity. In this way, we get our usual definition of the dot product:
T
a1 1 0 0 b1
a b = a2 0 1 0 b2 = a1 b1 + a2 b2 + a3 b3 . (9)
a3 0 0 1 b3

This particular metric is known as the Euclidean metric.


When looking at spacetime vector a = (at , a), you can write the inner product in a similar way:
T
at 1 0 0 0 bt
a1 0 1 0 0 b1
a b =

a2 0 0 1 0 b2 = at bt (a b) (10)

a3 0 0 0 1 b3

This particular metric tensor is known as the Minkowski metric.

1.2.2 Rates of Change


In the usual Cartesian coordinates, calculating rates of change is greatly simplified as the rate of change
in the x direction is completely independent of the rates of change in the y and z direction. When
dealing with other coordinate systems, this is not always the case.
If we consider 2d polar coordinates (r, ), you should be able to convince yourself that the rate of
change in is dependent on r. Consider some infinitesimal change in of d, the corresponding length
that this change represents on the 2d plane is r d.
This factor of r above is known as a scale factor and appear in expressions for the gradient, divergence,
curl and the Laplacian. There exists general expressions for scale factors in arbitrary coordinates, though
we will be primarily concerned with Cartesian, cylindrical and spherical coordinates (see section 3).

3
1.3 Cross Product
The cross product is a bit of a special operator which only works in 3d.1 We will try and give one physical
reasoning behind the cross product, though it will involve concepts which are a bit beyond the scope of
this course. If you dont understand all the arguments that following, just remember the final result.
Lets begin by considering a familiar example with circular motion, and notice what happens when
we take a mirror image:

p p

r r

On the left, the particle is going counter-clockwise around the original; however, on the right the particle
is moving clockwise. This will become crucial later on. Notice also that if we interchange the vectors p
and r, the motion changes from being counter-clockwise to clockwise (and vice-versa).
Given that we have two (linearly independent) vectors, they describe a 2d plane of 3d space, and the
magnitude of angular momentum must be proportional to both krk and kpk. This if we are to define
angular momentum, it should resemble something like:
?
L := krkkpkPlane(r, p). (11)

Since we are in 3d though, the plane spanned by r and p can be uniquely specified by taking a vector
that is perpendicular to both r and p. Well define f (a, b) to be the function that returns a unit vector
perpendicular to both a and b. We can now write angular momentum as

L := krkkpkf (r, p). (12)

So far, all we have decided is that f must be perpendicular to both r and p; however as noted above,
the motion changes direction if we interchange r and p thus f (r, p) = f (p, r). Lastly as we
noted above, if take a mirror image by taking (x, y, z) 7 (x, y, z) (and more generally, if we take
(x, y, z) 7 (x, y, z)), then f 7 f .
With this, all the necessary properties of f have been defined and all remains to find what it looks
like. In the case of 3d vectors, this is:

x y z
r p krkkpkf (r, p) = det rx ry rz (13)
px py pz

If you want to read more about this, look up what a pseudovector is.
1Strictly speaking, it is possible to define a cross product in 7d as well.

4
2 Conventions
The list below will explain some of the more common conventions you might encounter. Remember
though to use common sense if you find a form you havent encountered before. If were calculating
the work done by a field, it should be a line integral even if they write their infinitesimal line segment as
dS for some reason.
Also remember that although mathematics is much more precise than English, it still relies context in
order to be fully understood. In particular, it is often
R obvious from context whether a one, two or RRthree
dimensional
RRR integral is being performed and thus is usually used in all cases. The only uses of and
occur when they might otherwise be ambiguous, when the bounds are being explicitly stated, or
when simply being pedantic. Similarly, the domain of integration is usually omitted and taken from
context.

2.1 Line Integrals


The line integral over some curve C will usually be denoted by
Z
F ds. (14)
C
H R
A closed line integral can be specified with instead of . Since the boundary of a surface A forms a
closed curve, then it is also possible to write S instead of C. The infinitesimal line element might also
be represented by d or dx.

2.2 Surface Integrals


The surface integral over some surface S will usually be denoted by
Z
F da. (15)
S
H R
A closed surface integral can be specified with instead of . Since the boundary of a volume V forms
a closed surface, then it is also possible to write V instead of S. The infinitesimal area element might
also be represented by dS, dA or d2 x.

2.3 Volume Integrals


The volume integral over some volume V will usually be denoted by
Z
f dv. (16)
V

The infinitesimal area element might also be represented by dV , d3 x, d . Note that instead of the
above expression, you might also find Z
df. (17)
V
One concrete
R example of this is to expression the total charge as the integral of all infinitesimal charges:
Q = V dq.

5
3 Coordinate Systems
3.1 Cartesian
Cartesian coordinates are the typical (x, y, z) coordinates. The unit vector are usually x, y and z,
though it is not uncommon to see , and k.

f f f
f = x + + z
x y z
Fx Fy Fz
F = x + y + z
x y z
     
Fz Fy Fx Fz Fy Fx
F = x + y + z
y z z x x y
2 Fx 2 Fy 2 Fz
2 f = + +
x2 y 2 z 2

3.2 Cylindrical Coordinates


Cylindrical coordinates exploit the symmetry of a cylinder. They consist of 2d polar coordinates (, )
along with the z Cartesian coordinate. Note that the radial distance is very often expressed as r instead
of . It is also common to see or instead of .
p
x = cos = x2 + y 2
y = sin = arctan(y/x)
z=z z=z

ds = d + d + dzz dv = d d dz

f 1 f f
f = + + z
z
1 (F ) 1 F Fz
F = + +
z
     
1 Fz F F Fz 1 (F ) F
F = + + z
z z
1 2f 2f
 
2 1 f
f= + 2 +
2 z2

6
3.3 Spherical Coordinates
Spherical coordinates exploit the symmetry of a sphere. They consist of a radial distance r, a polar
angle and an azimuthal angle : (r, , ). Note that in the maths literature, the polar and azimuthal
components are interchanged. Just as with cylindrical coordinates, the radial component is often
denoted as .
p
x = r cos sin r = x2 + y 2 + z 2
y = r sin sin = arccos(z/r)
z = r cos = arctan(y/x)

ds = drr + r d + r sin d dv = r2 sin d d d

f 1 f 1 f
f = r + +
r 2 r sin
1 (r2 Fr ) 1 (sin F ) 1 F
F = 2 + +
r r r sin r sin
 
1 (F sin ) F
F = r
r sin
 
1 1 Fr (rF )
+
r sin r
 
1 (rF ) Fr
+
r r
2f
   
1 f 1 f 1
2 f = 2 r2 + 2 sin + 2 2
r r r r sin r sin 2
|  {z  }
2
= + r2 r

f
r 2

You might also like