1.

Vectors
1.1 Definitions and properties
1

S
V
P
D
p
A
D
E

C
D
M
C
N

© Stuart Dalziel (Michaelmas, 2021) 12

. Vectors Definitions and properties

In the typeset notes we will identify a vector variable using bold

Roman font, e.g. u or q . This notation is widely used although it
may be set using bold italics e.g. u . During the lectures, vectors
will be indicated using a ~ beneath the corresponding variable
name, e.g. u or q , or sometimes using an underline, e.g. u or q .
Another commonly used notation puts an arrow above the variable
name, e.g. u or q . This is particularly common in North America.
We shall frequently use this notation when describing the vector
between two points. For example, AB between points A and B ,
or a position vector OA or OB .

1.1.1 Adding vectors

The easiest way to think about adding the two vectors a and b is to
think geometrically about adding displacements.
b b

a a

a
a+b b+a
b

Figure 1. Vector addition: adding the displacement is commutative.

© Stuart Dalziel (Michaelmas, 2021) 13

. Vectors Definitions and properties

S
c

F
I
t
T

© Stuart Dalziel (Michaelmas, 2021) 14

. Vectors Definitions and properties

As well as being commutative, vector addition is also associative,

that is the order in which the vectors are added does not matter;
(a b) c a (b c) a b c . (2)
B B
b b
C C
a a
(b + c)
(a + b) c c

A A
(a + b) + c D a + (b + c) D

Figure 3: Vector addition is associative. Both (a b) c and a (b c) start at

point A and finish at point D .

We can also describe the situation shown in figure 3 using the over-
arrow position vector notation. If

a AB b BC c CD

then

a b AB BC AC
b c BC CD BD

so

(a b) c AC CD AD
a (b c) AB BD AD .
a b c AB BC CD AD

© Stuart Dalziel (Michaelmas, 2021) 15

. Vectors Definitions and properties

If vectors not co-planar (later), then if vectors are edges of a

parallelepiped sum is the diagonal

Figure 4: The sum of three vectors is the diagonal of a parallelepiped

a b c diagonal .

1.1.2 Multiplying a vector by a scalar

A vector can be multiplied by a scalar, to give another vector. For
instance, the vector a is simply the vector a but with its length
scaled by . Multiplication by a scalar is distributive, which means
that
(a b) a b

a a.

a (a + b) = a+ b

a+b

Figure 5: Multiplication of a vector by a scalar is distributive.

Multiplication by a negative real scalar reverses the direction of the

vector.

© Stuart Dalziel (Michaelmas, 2021) 16

. Vectors Definitions and properties

1.1.3 Coordinate axes

For vector components, we need to know the axes
3

A
N
A
A

© Stuart Dalziel (Michaelmas, 2021) 17

. Vectors Definitions and properties

Cartesian coordinates
4

M
E
H
W
B

For Cartesian coordinates the x, y, z axes are mutually

perpendicular and emanate from a fixed origin O . We will always
use right-handed systems, i.e. the directions Ox , Oy and Oz must
line up with the thumb, first finger and second finger of ones right
hand, in that order.

Figure: 6: Right-handed coordinate system. Using your right hand: thumb for x,
index finger for y and middle finger for z.

While any Cartesian coordinate satisfying the right-hand rule is

‘right-handed’, shifting the origin or changing the orientation of the
axes changes the coordinate system and the coordinates of an
arbitrary point P.

© Stuart Dalziel (Michaelmas, 2021) 18

. Vectors Definitions and properties

Cartesian coordinates of point P

The coordinates of a point P , relative to these axes, are denoted

x y z and correspond to the lengths of the projections of the vector
OP onto the three axes. The vector OP can then be written in
coordinate form ( x y z ) or as a column vector

x
OP r y .
z

We introduce vectors xˆ yˆ zˆ which are parallel to the three axes and

which each have unit length. In fact, the hat notation x̂ means a
vector of unit length, a unit vector. Commonly, xˆ yˆ zˆ are also
called ˆi ˆj kˆ or sometimes e x , e y , e z . It follows that

ˆi 1,0,0 , ˆj 0,1,0 , kˆ 0,0,1 ,

and then that

OP r ( x y z ) xˆi yˆj zkˆ

1 0 0 x
x 0 y 1 z 0 y
0 0 1 z

The vectors ˆi ˆj kˆ are said to form a basis of three-dimensional

space, because these three vectors are sufficient that any point P
can be defined by its coordinates relative to these vectors in only
one way. The basis is said to span the space is provides a way of
accessing every point. Of course, this is not the only basis of three-
dimensional space; we could obtain an infinite number of other
bases by simply rotating ˆi ˆj kˆ by an arbitrary angle about the
origin, or applying a translation to the origin. We will come back to
the idea of bases in section 1.10.

© Stuart Dalziel (Michaelmas, 2021) 19

. Vectors Definitions and properties

1.1.4 Basic arithmetic with Cartesian coordinates

5

© Stuart Dalziel (Michaelmas, 2021) 20

. Vectors Definitions and properties

1.1.5 Kinematics
6

K
P

© Stuart Dalziel (Michaelmas, 2021) 21

. Vectors Definitions and properties

Example 1: Aeroplane flight path

An aeroplane wishes to fly from city A to city B, which is at a
distance of 1000km north-west of A. There is a wind of speed
50km hr 1 from the south-west, and the average speed of the plane
relative to the air is 800km hr 1 . What direction would the pilot fly
in, and how long does the journey take?

800 kph

50 kph
A

1 50
sin 3.58
800
=> a compass bearing of 315 – 3.58 = 311.42 .
Speed made good towards B u 8002 502 798.44 kph
1000 1000
To travel the 1000 km takes a time t 1.252 hr
u 798.44
Note that because the wind speed is much smaller than the speed
of the aeroplane, the distance travelled is only slightly greater: only
around 7 seconds!

© Stuart Dalziel (Michaelmas, 2021) 22

. Vectors Definitions and properties

Example 2: Crystalline lattice

A crystalline lattice consists of cubic unit cells. Each unit cell
consists of an atom of type A at each corner of the cube, and a
single atom of type B at the centre of the cube. The sides of the
cube have length a . Find the distance from an atom B to its nearest
neighbouring atom of type (i) A and (ii) B . What distances are the
second nearest neighbours from B ?

Take the coordinate system aligned with the lattice and the origin
be at the location of one of the A atoms.
There are type A atoms at (0,0,0), (a,0,0), (0,a,0), (0,0,a), (a,a,a),
etc., and type B atoms at (a/2,a/2,a/2), etc.
rA a,0,0 and rB 2 a, 2 a, 2 a be the positions of the next A
1 1 1

atom to the right and the closest B atom to the origin, respectively.

(a) The distance between a B atom and its nearest A atom is

a 1
2 a 1
2 a
rBA rA rB 0 2a
1
2a
1

0 1
2 a 1
2 a
2 2 2 3
1
2 a 1
2 a 1
2 a 3
4 a2 a
2
rB

© Stuart Dalziel (Michaelmas, 2021) 23

. Vectors Definitions and properties

(b) Between B and the next B (B , say):

B
B

B
O

3
2 a 1
2 a a
rBB rB rB 1
2 a 1
2 a 0 a.
1
2 a 1
2 a 0
Note that we could have expressed rB as rB rA so that
rB rB rB rA rB rA .
The second nearest A are
1
2 a a 3
2 a
rBA rBA rAA 1
2 a 0 1
2 a
1
2 a 0 1
2 a
2 2 2
3
2 a 1
2 a 1
2 a 11
4 a 1
2 11 a
and second nearest B
a 0 a
rBB rB rB rB rB rB rB 0 a a
0 0 0
2a

© Stuart Dalziel (Michaelmas, 2021) 24

. Vectors Definitions and properties

Example 3: Angle between lines

Use the cosine rule to calculate the angle between the lines OA and
OB , where O is the origin, a (1 2 3) and b (2 3 2) .
7

© Stuart Dalziel (Michaelmas, 2021) 25

. Vectors The equation of a line

1.2 The equation of a line

Vector form
Consider a straight line through the points A and B , whose position
vectors relative to the origin O are a and b , respectively. We want
to write down an equation for the general point P on the line with
position vector r .
8

b
A
r

Figure 7: Vector definition of a line through the points A and B.

Note first that the line between A and P is necessarily parallel to

the line between A and B . This means that AP must be a scalar
multiple of AB , or in other words
r a (b a) ,

for some scalar , and rearranging gives

r a b a . (5)

Equation (5) is the vector equation of a straight line, since any

point r on the line can be written in this form for some . The idea
is that the scalar runs through all real values, thereby describing
the whole line – the original points A and B on the line correspond
to the special cases 0 and 1 respectively.

© Stuart Dalziel (Michaelmas, 2021) 26

. Vectors The equation of a line

Instead of knowing two points on the line, we are sometimes given

just one point (say a ) together with the direction of the line (call it
the unit vector t̂ ). In that case, equation (5) can easily be modified
to give the alternate general form
b a ˆtˆ ,
r a b a a (6)
b a

where tˆ (b a) b a is a unit vector (i.e. tˆ 1 ).

Component form
Equation (5) can be recast in a form that does not involve at all
by using components. Write a (ax a y az ) , etc., then see that the
component form of (5) becomes

r ( x y z ) ( ax (bx ax ) a y (by a y ) az (bz az ))

and equating the components on the left and right gives us three
equations, which can be rearranged to give

x ax y ay z az
bx ax by ay bz az

Of course is the same in each of these three equations, so

therefore the right hand sides must all be equal to each other. This
gives

x ax y ay z az
, (7)
bx ax by ay bz az

which is an alternative form of the equation of a straight line which

does not involve the scalar .
What happens if, for example, ax bx ?
What if ax bx , a y by and az bz ?

© Stuart Dalziel (Michaelmas, 2021) 27

. Vectors The equation of a line

Example 4: Straight line through two points

Find the equation of the straight line through the points a (1 2 3) ,
b (4 0 6) . Does this line pass through the point (4 0 0) ? For what
value of does the point (10 4 ) lie on the line?
The line is given by
r a (b a)
1 4 1 1 3
2 0 2 2 2
3 6 3 3 3
Test if there is a value of giving r 4,0,0 .
For x component, need 1 3 4 1
r 1 3 ,2 2 ,3 3 4,0,6
the line does not pass through 4,0,0

When does r (10 4 ) ?

Need 1 3 10 3
r 1 3 ,2 2 ,3 3 10, 4,12
12

Example 5: Centroid of a triangle

Consider the triangle ABC , with the position vectors of the points
relative to the origin O being a , b , c respectively. Show that the
medians (the lines from each vertex to the middle of the opposite
side) meet in a single point, the so-called centroid (cf. question 9 on
Example Sheet 1).

© Stuart Dalziel (Michaelmas, 2021) 28

. Vectors Scalar product

C +O
A
The middle of the side BC is given by a 1
2 b c and so the
median through point A is given by
r a 1
2 b c a
for some parameter . Similarly, the median through the point B
(and the side CA) for parameter is
s b 1
2 a c b
These lines intersect when r s
a 1
2 b c a b 1
2 a c b
a1 1
2 b1 1
2 c 12 0
1 3
2 0 2
3

and r a 2
3
1
2 b c a r s 1
3 a b c
Similarly, for the median passing through point C.

1.3 Scalar product

We will consider two different ways of multiplying vectors
together. The first is the scalar product, in which two vectors are
multiplied together to give a scalar.
This is called the ‘scalar product’ because the operation produces a
scalar value from two vector values. The same operation is also
known as the ‘dot product’ because of the common notation
adopted.
© Stuart Dalziel (Michaelmas, 2021) 29
. Vectors Scalar product

The scalar product is a special case of an ‘inner product’.

Note that the dot is generally typeset as elevated (a \cdot in
LaTeX) to distinguish from a full stop or decimal point.

Scalar product definition and properties

9

P
C

© Stuart Dalziel (Michaelmas, 2021) 30

. Vectors Scalar product

We can now work out the scalar product of two vectors that are
given in (Cartesian) component form. We have

a b (ax ˆi a y ˆj az kˆ ) (bx ˆi by ˆj bz kˆ )

We use the distributive property (10) to expand the brackets, and

then the commutative property (9) to see that ˆi ˆj ˆj ˆi etc. This
leaves us with

a b axbx ˆi ˆi a y by ˆj ˆj az bz kˆ kˆ
axby a y bx ˆi ˆj a y bz az by ˆj kˆ az bx axbz kˆ ˆi

Orthonormality ( ˆi ˆi 1, ˆi ˆj 0, ) yields

a b axbx a y by az bz (12)

© Stuart Dalziel (Michaelmas, 2021) 31

. Vectors Scalar product

We often need to work out the component of a vector a in a given

direction, say in the direction of the unit vector n̂ . By resolving,
this component is a cos , where is the angle between a and
n̂ . We can now use the scalar product to work this out: the
component of the vector a in the direction of the unit vector n̂ is
a nˆ .

Example 6: Angle from dot product

Find the acute angle between the vectors (5 0 3) and ( 1 3 4) .
We compute a b (5 0 3) 1,3,4 5 0 12 7
But we also have a b a b cos
a b 7 7
cos
ab 52 32 12 32 42 52 32 12 32 42
7 7
34 26 2 17 13
7
2 221
1 7
cos 76.38
2 221

Example 7: Vector component in a direction

Find the component of the vector (2 1 7) in the direction (1 1 1) .
1
The unit vector in the direction (1 1 1) is (1 1 1) , so the
3
component of (2 1 7) in this direction is
1 1 8
(2 1 7) 1,1,1 2 1 7 .
3 3 3

© Stuart Dalziel (Michaelmas, 2021) 32

. Vectors The equation of a plane

1.4 The equation of a plane

In section 1.2 we found the equation for a straight line in several
different forms. We will now do the same for a plane.

Orientation + direction
The orientation of a plane is given by the direction of the normal
vector to the plane n . If we know one point on the plane, say A
given by a , the plane is then uniquely specified.
10

U
d
b
T

© Stuart Dalziel (Michaelmas, 2021) 33

 . Vectors The equation of a plane

n̂
n̂
r a P
A

a r
p

O
Figure 8: A plane may be defined by a unit normal n̂ and the position vector a of
one point on the plane. The closest point to the origin is on the normal vector
passing through the origin.

Orientation + distance
One important property of the plane is its perpendicular distance to
the origin, p . In figure 8 we can see that
p a cos

but since is the angle between a and n̂ , we can see from (8) that
a cos a nˆ (recalling that nˆ 1 ), leading to another form of
the equation of a plane,

E E n O T n
p
r nˆ p. (14)

Note that in (14) the right hand side is only the perpendicular
distance from the plane to the origin because n̂ is a unit vector.
Note also that while in (13) reversing the direction of n̂ did not
change the plane, in (14) it would push the plane to the opposite
side of the origin.
We can also write the equation of the plane in component form.

© Stuart Dalziel (Michaelmas, 2021) 34

. Vectors The equation of a plane

Suppose r ( x y z ) and nˆ (l m n) with nˆ 1.

nˆ r lx my nz p.

Since

nˆ r nˆ ˆix nˆ ˆjy nˆ kˆ z cos x x cos y y cos z z

then l , m, n are the direction cosines: angle between the normal and
each of the axes, while

nˆ r nˆ r cos r cos p

Three points
We can also find the equation of the plane if we know three points
in the plane. Suppose that a , b and c are the position vectors of
three points A , B and C , respectively, in the plane.

C r a P
c a
A
b a
B
a c r
b

O
Figure 9: A plane containing the points a b c and the general point r .
The vectors c a and b a therefore both lie entirely within the
plane (see figure 9). The plane itself is two-dimensional, and that
means that any other vector which lies in the plane can be expressed
as linear combination of c a and b a . In particular, the vector
r a lies within the plane, so r a must be the sum of a scalar
multiple of b a plus a scalar multiple of c a , i.e.

© Stuart Dalziel (Michaelmas, 2021) 35

. Vectors The equation of a plane

r a (b a) (c a) , , (15)

This expression is yet another form for the equation of a plane, and
the idea is that the scalar parameters range over all real values
so as to describe all possible points on the plane. [Recall that the
equivalent form for the equation of a line, (5), contained only one
scalar parameter, corresponding to the fact that a line is a one-
dimensional object.]
What happens if the points A, B and C all lie on a line?
What does this mean about the vectors c a and b a ?
Can they define a two-dimensional object (i.e. a plane)?

Example 8: Plane passing through three points

Find the equation of the plane which passes through the points
(1 0 0) , (2 1 0) and (1 1 1) . What is the minimum distance of this
plane from the origin?
The equation for the plane can be expressed as
r a (b a) (c a)
r 1,0,0 (1,1,0) (0,1, 1) for real , .
The components are then x 1 , y and z .
Eliminating y z .
Eliminating x y z 1.
1 1 1
n 1, 1, 1 nˆ , ,
3 3 3
The plane can be written as is may be written as
1 1 1 1 1
x y z , , x, y , z nˆ r .
3 3 3 3 3
1
The minimum distance to the origin is therefore .
3

© Stuart Dalziel (Michaelmas, 2021) 36

. Vectors The equation of a plane

Example 9: Intersection of two planes

Find the line of intersection of the two planes with normals parallel
to the vectors (1 0 2) and ( 1 1 1) , respectively, which both pass
through the point (1 1 0) . What angles does this line make with the
coordinate axes?
11

w
T

© Stuart Dalziel (Michaelmas, 2021) 37

. Vectors The equation of a plane

T
a

© Stuart Dalziel (Michaelmas, 2021) 38

. Vectors The equation of a plane

Example 10: Visibility over a wall

Can an observer A at (4,5,1) see object at the point B at (2 3 3 2)
when there is an intervening wall with top given by the line
r (1 2 1) (1 1 1) (32)
[the z axis is the upward vertical]?
12 – Moodle?

I
T
t
R

© Stuart Dalziel (Michaelmas, 2021) 39

. Vectors Equations of other objects

1.5 Equations of other objects

1.5.1 Sphere
Consider a sphere of radius and centre whose position vector is
a . Then a point r on the surface of the sphere is by definition a
distance from the centre a , and it follows that the equation of a
sphere is
r a . (16)

1.5.2 Cylinder

at OI top
13
n̂

direction
of OF A
L G Tin
P
scalar RQ
magnitude
r

an
Op It CT.mil O r

Figure 10: Vector definition of a cylinder.

R
Consider a circular cylinder of radius R whose axis is parallel to the
unit vector n̂ . The origin O lies on the cylinder axis, the point P is
a general point on the cylinder surface with OP r , and the point
Q is defined by OQ being the projection of OP onto the cylinder
axis (see figure 10). In the right-angled triangle OPQ we see that

OQ r cos r nˆ ,

projection of I
© Stuart Dalziel (Michaelmas, 2021) 40
. Vectors Equations of other objects

where the last step has been completed using (8). Since OQ is
parallel to the cylinder axis, we therefore have

OQ fifth
r nˆ nˆ . direction
We can also see that QP OP OQ , and that QP R (the cylinder
radius), and putting all this information together leads us to the
vector equation of the cylinder

r r nˆ nˆ R. (17)

To have axis through a point A , use r r a so that

r a (r a) nˆ nˆ R.

1.5.3 Cone
14

n̂

of
r
r q O
q
Q

Figure 11: Vector definition of a cone of semi-angle .

© Stuart Dalziel (Michaelmas, 2021) 41

. Vectors Vector product

Suppose that the cone has semi-angle , its axis is parallel to the
unit vector n̂ and its vertex is at the point Q with position vector q .
Again, the point P is a general point on the cone surface with
OP r (see figure 11).
The defining property of the cone is that the line from the vertex to
any point on the surface always makes an angle with axis. This
fact can be expressed mathematically using the scalar product (8) as
on
any point r q nˆ r q cos , (18)

surface
which is the vector equation ofvertex
the cone.

Hot product some constant

1.6 Vector product i K
y q
While the scalar product ‘multiplies’ two vectors together to create
a scalar, the vector product multiplies two vectors together to yield a
new vector with special properties.
1.6.1 Vector product definition
15

© Stuart Dalziel (Michaelmas, 2021) 42

. Vectors Vector product

Both a b and a b are widely used, often without explanation.

Different fields, and indeed different publishers, tend to adopt
different versions of this notation. Here we will use both.

a b b a b

a
a
b

a b
b

Figure 12: By convention, the vector product forms a right-handed system.

1.6.2 Properties of the vector product

a, b, a b right-handed system

b, a, b a right-handed system
a, b, b a left-handed system
a, b, b a right-handed system

© Stuart Dalziel (Michaelmas, 2021) 43

. Vectors Vector product

16

a
Z

1.6.3 Component form for vector product

We can now evaluate a b in component form. We have

a b (ax ˆi a y ˆj az kˆ ) (bx ˆi by ˆj bz kˆ )

The brackets are expanded using the distributive property (21), and
we then remove all terms involving cross products of vectors with
themselves thanks to (23). This leaves

© Stuart Dalziel (Michaelmas, 2021) 44

. Vectors Vector product

a b axby ˆi ˆj axbz ˆi kˆ a y bx ˆj ˆi
a y bz ˆj kˆ az bxkˆ ˆi a b kˆ
z y
ˆj

Finally, we can use (20) to collect terms in pairs (e.g. ˆi ˆj ˆj ˆi )

and then use (24) to give

a b a y bz az by ˆi az bx ax bz ˆj a x by a y bx kˆ . (25)

1.6.4 Vector product as a determinant

The formula (25) is quite cumbersome and hard to remember, but
can be expressed much more compactly using the notation of the
determinant of a matrix. Matrices and determinants will be covered
in detail in the Easter term course, but for now think of a matrix as a
table of elements
a b c
M d e f ,
g h i

and its determinant

det(M) M

as a measure of the ‘scale’ of the matrix. Here, M is said to be a

3 3 matrix as it contains three rows and three columns. This is
useful when dealing with 3D Cartesian coordinates.

However, we need to know how to compute M .

We will present two ways, but you only need to know one – or, for
this part of the course, just be able to remember the form of (25),
although it is better to be able to use one of the two methods here.

© Stuart Dalziel (Michaelmas, 2021) 45

. Vectors Vector product

Method 1: Laplace expansion

One way is to construct the 2 2 submatrices (‘minors’) with their

elements retaining their left-to-right order and switch the sign on
consecutive determinants.
17

N
l

© Stuart Dalziel (Michaelmas, 2021) 46

. Vectors Vector product

Method 2: Rule of Sarrus

Based on the Leibniz formula for permutations.
18

© Stuart Dalziel (Michaelmas, 2021) 47

. Vectors Vector product

Vector product
We extend the idea of a matrix to allow it to have vector elements
and introduce the notation for a b as

ˆi ˆj kˆ
a b ax ay az , (26)
bx by bz

and the right-hand side is expanded using one of the patterns

described above to recover (25).
Laplace expansion:
+ – +

ˆi ˆj kˆ
ay az ax ay
a b ax ay az ˆi ˆj ax az
kˆ
by bz bx bz bx by
bx by bz
ˆi a b az by ˆj a b axbz kˆ axby a y bx
y z z x

Rule of Sarrus:

ˆi ˆj kˆ ˆi ˆj kˆ
ax ay az ax ay az ˆia b ˆja b kˆ ax by bx a y kˆ bx az ˆi bz a y ˆj
y z z x

bx by bz bx by bz
ˆi a b az by ˆj a b ax bz kˆ axby a y bx
y z z x

or more succinctly

ˆi ˆj kˆ ˆi ˆj kˆ
ax ay az ax ay az ˆi a b a z by ˆj a b ax bz kˆ ax by a y bx
y z z x

bx by bz bx by bz

© Stuart Dalziel (Michaelmas, 2021) 48

. Vectors Vector product

Later, you are likely to come across the Levi-Civita symbol to

provide a compact notation. This symbol is defined as
1 if i, j , k , are an even permutation of 1,2,3,
ijk 1 if i, j , k , are an odd permutation of 1,2,3,
0 otherwise

then c a b becomes ci ijk a j bk .

Example 11: Calculate vector product

Calculate a b when a (1,2,3) and b (4,5,6) .
19

© Stuart Dalziel (Michaelmas, 2021) 49

. Vectors Vector product

Example 12: Angle between vectors

Use the cross product to find the angle between the vectors
a (2 1 0) and b (3 0 1) . Find a unit vector that is perpendicular
to both these vectors.
a b
Recall a b a b sin nˆ , thus sin .
ab
ˆi ˆj kˆ ˆi ˆj kˆ
Now a b ax ay az 2 1 0
bx by bz 3 0 1
Using rule of Sarrus:
ˆi ˆj kˆ ˆi ˆj kˆ
2 1 0 2 1 0 2 1
3 0 1 3 0 1 3 0
ˆi 1 0 ˆj 0 2 kˆ 0 3 1, 2, 3
a b 1, 2, 3 is perpendicular to both a and b .

Noting that a b 12 22 33 14 , then the unit vector is

a b 1
nˆ 1, 2, 3 .
a b 14
The angle between a and b is given by
a b 14 7 1 7
sin sin 31.95 .
ab 5 10 5 5

But much easier to use dot product a b a b cos

1 a b 1 6 1 6
cos cos cos 31.95
a b 5 10 5 2

© Stuart Dalziel (Michaelmas, 2021) 50

. Vectors Vector product

Example 13: Spinning body

A rigid body spins with angular speed about an axis n̂ passing
through the point Q. Find a vector expression for the velocity of any
point in the body.
The speed of motion is equal to the angular speed multiplied by
the distance from the axis about which it is rotating.

n̂

u
P

p r
Q
q
O

The direction of the motion is perpendicular to the plane

containing the point P and the rotation axis. The vector QP p lies
in this plane, and so the velocity is in the direction of nˆ p .
Now, if û is the unit vector in this direction, then the velocity is u uˆ
Recalling that the vector product is
nˆ p nˆ p sin uˆ uˆ
and that p r q , then the velocity is given everywhere by
u nˆ p nˆ r q .

© Stuart Dalziel (Michaelmas, 2021) 51

. Vectors Vector product

Vector product equation for a line

We can use the vector product to derive B

alternative equations for a line and plane. For P

instance, consider the line described in figure 7.

As we can see, the vector r a is parallel to the A
r
b

line b a for any point r on the line. It a

therefore follows from property 4 above that O

r a b a 0, (27)

which is yet another form for the equation of a line through the
points given by position vectors a and b .

Vector product equation for a plane

For the equation of a plane, let us look again at figure 9.
20

C r a P
c a
A
b a
B
a c r
b

We see that the vectors b a and c a lie in the plane. By the

definition of the vector product,
(b a) (c a)

is perpendicular to both b a and c a , and is therefore the normal

(but not necessarily the unit normal) to the plane. For any point r in
the plane, the vector r a lies in the plane, and is therefore
perpendicular to the normal to the plane. It then follows that Using
property 4 of the scalar product, it therefore follows that

r a b a c a 0. (28)

Eat 5 0
© Stuart Dalziel (Michaelmas, 2021) 52
E a no
. Vectors Application to calculating distances

1.7 Application to calculating distances

1.7.1 Shortest distance of a point from a line
21
Q
B
d

q–a R

b–a
b
A q

Figure 13. Shortest distance of the point Q from the line AB .

Recall (5), the vector equation of a line,

r a b a ,

and consider the point Q with position vector q (see figure 13). The
shortest distance, d , between Q and the line corresponds to the
distance along the perpendicular to the line, i.e. QR in the figure.
Considering the right-angled triangle AQR we see that

d q a sin ,
not the projection
with the angle between AQ q a and AB b a . Using the
definition of the vector product (19), we can therefore write

q a b a
d . (29)
b a

Note that the factor b a in the denominator of this expression.

normalizing
© Stuart Dalziel (Michaelmas, 2021) 53
. Vectors Application to calculating distances

If instead of the general form for the line (5) we use the parametric
form (6) with unit vector tˆ b a b a then (29) simplifies to

d q a tˆ .

1.7.2 Shortest distance of a point from a plane

22
P
p a
d n̂
An
Q
A

a
p

O
Figure 14: The shortest distance between the point P and the plane defined by the
point A and normal n̂ .

Consider the plane that passes through point A (given by position

vector a ) and that has unit normal n̂ . From (13), points r satisfy

r a nˆ 0 .

The closest point on the plane to a point P, given by position vector

p , is the point Q, where QP is normal to the plane and QP d .
The noting that APQ is a right-angle triangle, then

QP d p a cos .

Recalling the definition of the scalar product (8), we find that the
shortest distance is given by

d p a nˆ . (55)

© Stuart Dalziel (Michaelmas, 2021) 54

. Vectors Application to calculating distances

1.7.3 Shortest distance of a line from a line

Q L1
u
B

t d
u
t u L2
P
q
A b
p

t a
O

Figure 15: Shortest distance between non-intersecting lines in three dimensions.

Consider line L1 passing through point A and line L2 passing through

point B. The lines are parallel to the vectors t and u , respectively.
The lines can therefore be expressed parametrically as the sets of
points r for L1 and s for L2 satisfying
r a t and s b u

for real parameters and .

If the lines do not intersect, then there are no solutions to r s , but

we can still identify the point where the lines are closest. Suppose
that the shortest distance between the two lines is between the
points P and Q. Since this is the shortest distance, the line joining P
and Q will be perpendicular to both lines.

The first question is: What direction is PQ in? The lines are parallel
to t and u , respectively, and PQ is perpendicular to both of these.
The vector which is perpendicular to both t and u is simply t u ,
and so PQ is parallel to t u .

© Stuart Dalziel (Michaelmas, 2021) 55

. Vectors Scalar Triple Product

The second question is: what is d PQ ? This is easy to answer

now we know what direction PQ is in, because we can work out
one vector joining the two lines, which is b a , and then the
shortest distance between the two lines is the component of b a in
the direction t u , and we can find this simply using the scalar
product as

b a t u
d , (30)
t u

remembering the need to normalise t u .

This expression gives us an easy way of finding out if two lines
intersect or not. If they do intersect then their closest distance is
obviously zero, at the point of intersection, so all we need is to set
the quantity on the right of (30) to zero. The condition for two lines
to intersect is therefore

b a t u 0. (31)

The quantity in (31) is actually three vectors multiplied together, in

this case the vector product of t u is taken first, and then the
scalar product is taken with b a . This is a triple product, and we
will consider these in the next couple of sections.

1.8 Scalar Triple Product

The quantity a b c is a scalar, and is called the scalar triple
product of a b c . The only way that this product makes sense is if
the vector product is taken first, to give the vector b c , and then
the scalar product is taken of a with b c . (If we were to attempt to
compute a b c then we would find a b is a scalar, but
operator is not defined between a scalar and a vector.) For this
reason the brackets are unnecessary, and we often write

© Stuart Dalziel (Michaelmas, 2021) 56

. Vectors Scalar Triple Product

a b c a b c [a, b, c] . (32)

The third notation in (32) is sometimes also used as a shorthand.

Given a b c in component form, we use equation (25) to work out
b c , and then use (12) to give

a b c a x by c z bz c y a y bz cx bx cz az bx c y by c x (33)

By rearranging this formula, we can find the following results:

a b c b c a c a b
a c b b a c. (34)
c b a

In other words, the value of a b c is unchanged if we make an

even permutation of the order of a b c , but the sign changes if we
make an odd permutation of the order of a b c .

Here, a permutation means swapping pairs of elements. An even

permutation means we swap an even number of pairs, while an odd
permutation means we swap an odd number of pairs. For (34) this
means that if the cyclic order of a b c remains the same (i.e., either
a b c , b c, a or c, a b ) then the sign remains the same, but if the
cyclic order is changed (i.e. c, b, a , b, a, c or a, c, b ) then the sign
changes.
23

One important property is that if two (or more) of a b c are parallel

then their triple scalar product is zero. So, for example, suppose that
a and b are parallel, then using (34)
a b c c a b
c 0
a bae
0
c c brat
since for parallel lines a b 0 (property 4 in section 1.6).
c carb
© Stuart Dalziel (Michaelmas, 2021) 57
. Vectors Scalar Triple Product

The formula (33) is hard to remember, but can be expressed much

more compactly using the notation of determinants introduced
already in section 1.6. We write

ax ay az
a b c bx by bz , detca I (35)E
cx cy cz

which has been formed from taking the components of a b c as the

three rows. The determinant can now be expanded using the pattern
we have seen previously:
The Laplace expansion:

ax ay az
by bz b bz bx by
a b c bx by bz ax ay x az
cy cz cx cz cx cy
cx cy cz
ax by cz bz c y a y bx cz bz cx az bx c y by cx

or the rule of Sarrus:

ax ay az ax ay
a b c bx by bz bx by
cx cy cz cx cy

ax by cz bz c y a y bz cx bx cz az bx c y by cx

© Stuart Dalziel (Michaelmas, 2021) 58

. Vectors Scalar Triple Product

Example 14: Scalar triple product

Evaluate the scalar triple product of the vectors (2 3 1) , (1 2 2) and
(4 5 1) .

Write in the form of a determinant

2 3 1
1 2 2
4 5 1
and expand (rule of Sarrus)
2 3 1 2 3
1 2 2 1 2 2 2 1 2 5 3 2 4 1 1 11 5 2 4
4 5 1 4 5
2 8 3 7 1 3 16 21 3
2
Volume of parallelepiped
One important application of the scalar triple product is to the
calculation of the volume of a parallelepiped. Consider the
parallelepiped formed by the vectors a b c .
24
b c

a
h
c

Figure 16: Parallelepiped formed from the vectors a b c.

© Stuart Dalziel (Michaelmas, 2021) 59

. Vectors Scalar Triple Product

Earlier, we saw that the diagonal is given by a b c .

The area of the base is

Abase b c sin b c. (36)

where we have recalled the definition of the vector product (19).

The height h of the parallelepiped corresponds to the component of
the vector a in the direction perpendicular to the base. The unit
vector perpendicular to the base is
b c
b c
and therefore the height h is
b c
h a . (37)
b c

The volume of the parallelepiped is the area of its base times its
height, and putting together (36) and (37) we find that the volume is
V a b c, (38)
i.e. the scalar triple product.

© Stuart Dalziel (Michaelmas, 2021) 60

. Vectors Scalar Triple Product

Example 15: Scalar triple product

Show that a a c 0 . What does this result mean geometrically?
What is the condition for three vectors to be coplanar?
(i) We can show this the long way:
ax ay az
ax ay az ax a y cz az c y a y a z cx a x cz az ax c y a y cx
cx cy cz
a x a y cz ax az c y a y a z cx a y a x cz az ax c y az a y cx
0
25

la
a anc c (
la
d
O (

(
t
c
(
v
z

© Stuart Dalziel (Michaelmas, 2021) 61

. Vectors Scalar Triple Product

A zero scalar triple product, a b c 0 indicates that the vectors

are co-planar and so only span a two-dimensional space. We shall
use this idea further later.

Example 16: Volume of a rhomboid

Find the volume of the rhomboid drawn in a face-centred cubic
lattice.
z
z

O D
O

C
A
y A B y
x
x

Figure 17: Cubic lattice, side length a . The small atoms are located
at the centred of each face and the large atoms at the corners.
We start by finding the vectors defining the rhomboid (the ‘cubic’
version of a parallelepiped). Noting that the atoms A, B, C and D
are located at
rA a,0,0 ,
rB 1
2 a, 12 a,0 ,
rC a, 12 a, 12 a ,
rD 1
2 a,0, 12 a ,
then the volume of the rhomboid is defined by the vectors
rAB rB rA 1
2 a, 12 a,0 ,
rAC rC rA 0, 12 a, 12 a
rAD rD rA 1
2 a,0, 12 a .

© Stuart Dalziel (Michaelmas, 2021) 62

. Vectors Scalar Triple Product

Thus the volume of the rhomboid is

V rAB rAC rAD
1
2 a 1
2 a 0
0 1
2a 1
2 a
1
2 a 0 2a
1

1
2 a 1
2 a 1
2 a 1
2 a 1
2 a 1
2 a 0
1 3
a
4
Of course, the volume is not really negative. The minus sign is the
result of us choosing a left-handed system for our atoms A, B, C
and D. The volume is really
1 3
V a .
4

26

note
axe bx c I can b XC

© Stuart Dalziel (Michaelmas, 2021) 63

. Vectors Vector triple product

1.9 Vector triple product

In the last section we took the scalar product of two vectors, a r ,
where one of those vectors happened to be the result of taking the
vector product of two other vectors, r b c . An obvious extension
to this idea is to take the vector product rather than the scalar
product, i.e. a r , where r b c .

a
b c
r a (b c)
Hes the
Figure 18: Geometrical representationinof the vectorsame plane as I AT
triple product. The blue vector
b c is perpendicular to both b and c (which define the orientation of the green
plane, and so the blue vector is normal to that plane). The dark green vector
r a (b c) is the perpendicular to both the purple vector a and the blue
vector b c , and thus to the purple plane that is defined by those two vectors..

The quantity a b c is a vector, and is called the vector triple

product of a b c . Unlike the scalar triple product a b c , the
position of the brackets in a b c is crucial, and we will see
below that
a b c a b c. (39)

We can use our expression (46) for the vector product in

components to evaluate a (b c) . First, use (25) to evaluate b c ,
so that

© Stuart Dalziel (Michaelmas, 2021) 64

 . Vectors Vector triple product

a (b c) (ax ˆi a y ˆj az kˆ ) (by cz bz c y )ˆi (bz cx bx cz )ˆj (bx c y by cx )kˆ

and then using the distributive property (21) to expand the brackets
and the results (23) and (24) for the coordinate vectors to evaluate
the cross products, giving

a (b c) ( a y kˆ az ˆj)(by cz bz c y )
(ax kˆ az ˆi )(bz cx bx cz )
( ax ˆj a y ˆi )(bx c y by cx )

Gathering together the terms in ˆi ˆj kˆ , we find the final result

gives a linear mutant
a (b c) (a c)b (a b)c . (40)
of
combinationWe can proceed in exactly the same wayFustat
to find (a b) c in the
vectors I and
form
o (a b) c (c a)b (c b)a . (41)

By comparing (40) and (41), it becomes clear that

in the same
a (b c) (a b) c . P plane as Edf
Lutte
27

save T

Encina plane as
Ice s
a
w
I at Eta

© Stuart Dalziel (Michaelmas, 2021) 65

. Vectors Vector triple product

Example 17: Multiple ways of calculating vector products

(i) Calculate (ˆi ˆj) ˆi in two ways. (ii) Calculate (a b) (c r) .
28

R
i u
o

ii let me txt v Ext d

Carbon
unc brat
c a b cu b a

Ext at b III baa

W
equivalently
VAC Carr
carb C
carb a

v r C V c r Incl a

L xbl.r I c Carb CI r

ear b a ear albtL ib.r C

r
Lamb C
r Cbncl at Cr canal otter Cann
a clone

© Stuart Dalziel (Michaelmas, 2021) 66

. Vectors Vector triple product

(ii) We have a number of options for calculating (a b) (c r) .

We could do three vector products, or one vector product and two
scalar products. The latter is clearly less prone to error!
We shall compute
a b u u a b u b a
where u c r . Expanding
a b u c r a b c r b a. (A)
Equivalently, we could compute
v c r v r c v c r,
where v a b . Expanding
v c r a b r c a b c r. (B)
Since both (A) and (B) are equal, it says we can write
a b r c a b c r c r a b c r b a
c r b a c r a b a b r c
r
a b c
The above example shows that we can write any vector r in terms
of three other vectors a, b, c provided they span the space (i.e. they
are not zero-length, co-linear or co-planar and hence a b c 0 ):

r b c a r c a b r a b c
r
a b c
We can write this as
r a b c

where
r b c d c a d a b
, , .
a b c a b c a b c

© Stuart Dalziel (Michaelmas, 2021) 67

 . Vectors
similar to
vector
the concept
Basis vectors
of
spacelinearly independent
1.10 Basis vectors

Ik
Definition and properties
30
far Cartesian 29
so orthogonal
S
SD N D
what about 3D UD W

For N D we need N basis vectors N

er er r
en T
and C FO
To be a ban's
det to 1
1 Vectors linearly independent 2
2 No of vectors must equal the no
of
dimensions in space
L
Linear independence requires
Rie t the t anew so
then the only solution is N da p no
042
There's only 1
way of describing a
point
2 X e t ther t r
t Xn en
C

there's only 1 set of ti tu In W

contribution
e each vector provides a
unique
combination not mean
any for S

egget © Stuart Dalziel (Michaelmas, 2021)

men all
liver
68
orthogonal they
independence
are linearly
if E are
 3

. Vectors Basis vectors

E basis
is the set I
j where a étj a
D
i ALE 7 0
D t Dr
Cj t
Tj
T
it 23 T t Hat 23 j 20
b

R Nz 123 T

there is infinite nor

of solutions o
we can that
Alternatively argue
FOR The cartesian
point 5 Cl 1.0
3
we could write
5 at
j
I
a
city t CI a a
for any 2 ER
infinite no
of sets of th Xz As

if I 0.0 1
i
then we can't represent it using j a

© Stuart Dalziel (Michaelmas, 2021) 69

 . Vectors Basis vectors

Reciprocal basis
In Example 17, we saw in three-dimensions that we could use
vector triple products to express any vector in terms of three other
linearly independent vectors a, b, c . (Linear independence requires
that the scalar triple product a b c [a, b, c] is non-zero.
constant terms
If a, b, c form a basis such that a general vector r is given by
resultant
vector
00J I
r a b c, recall that
v bid
water
then the reciprocal basis is given by a
b c c a a b
a bae
A , B , C . d
a, b, c a, b , c a, b, c scalar

d r A
f
This gives, for example, Toone
quantity

B a A 1, a B 0, a C 0 ,
Bar
C so that a only has a component in the direction of A , similarly for
X r
b and c .
note dot product
Noting that
is different to A r A a A b A c

multiplication you B r , C r
CAN T cared
tens on
then the reciprocal basis provides a tool for determining the
both sides coefficients when using the basis.

1
Specifically, r a b c requires

r b c r c a r a b
the , , ,
adjusting a b c a b c a b c
coefficient which we can interpret as the ratio of projections of r and one of
the basis vectors onto the normal of the other two basis vectors.
an indical

© Stuart Dalziel (Michaelmas, 2021) 70

In
much they

common
. Vectors Basis vectors

b c
A a
a b c

d a as the ratio of the

Figure 19: Geometrical interpretation of the coefficient
projection of d onto the vector product b c to the projection of a onto the onto
the vector product b c .

If a, b, c is right-handed orthonormal basis, then the reciprocal basis

A, B, C is identical to a, b, c .

if orthonormal
D A I air

© Stuart Dalziel (Michaelmas, 2021) 71

. Vectors Basis vectors

Example 18: Sets of vectors as a basis

Check that: (a) ˆi ˆj , ˆj kˆ , kˆ ˆi is a basis for three-dimensional
space; (b) the vectors (1 1 0 0) , (0 1 1 0) , (1 0 1 0) , (1 0 0 1) form a
basis for four-dimensional space. Are the vectors normal? Are the
vectors unit length?
(a) Earlier we saw that a non-zero scalar triple product indicates
that a set of three vectors span a volume, and hence that set can
describe any vector in three-dimensional space. Thus we need to
compute
1 1 0
ˆi ˆj ˆj kˆ kˆ ˆi 0 1 1
1 0 1
(A)
11 0 11 0 0 0 1
1 1
2
and so the vectors are linearly independent. As there are three
linearly independent vectors in three-dimensions, then they form a
basis.
Harder way
Alternatively, we can use the fundamental definition in (42) and
solve
e
1 1 e
2 2 3 3 e 0,
with e1 ˆi ˆj , e ˆj kˆ , e kˆ ˆi . Thus
2 3

ˆi ˆj ˆj kˆ kˆ ˆi
1 2 3

ˆi ˆj kˆ 0
1 3 1 2 2 3

© Stuart Dalziel (Michaelmas, 2021) 72

. Vectors Basis vectors

This gives us three simultaneous equations

1 3 0,
1 2 0, (B)
2 3 0.
Substituting 3 1 into the third gives 2 3 1 , while
substituting this into the second then gives 1 0 and thus the
unique solution is 1 2 3 0 , and the vectors are
independent.
Note that we could have written the system of equations (B) as the
matrix equation
1 1 0 1 0
0 1 1 2 0 ,
1 1 0 3 0
and that the matrix has the same structure as the determinant in
(A). Since the determinant of the matrix does not vanish, then the
only solution is 1 2 3 0.
We can check whether the vectors are normal and/or unit length
using the scalar products:
ˆi ˆj ˆi ˆj ˆj kˆ ˆj kˆ kˆ ˆi kˆ ˆi 2 not unit
length,
ˆi ˆj ˆj kˆ ˆj kˆ kˆ ˆi kˆ ˆi ˆi ˆj 1 not normal.

(b) To test the four-dimensional vectors, we again use (42), taking

e1 (1 1 0 0) , e2 (0 1 1 0) , e3 (1 0 1 0) , e4 (1 0 0 1) to get

1 1,1,0,0 2 0,1,1,0 3 1,0,1,0 4 1,0,0,1

1 3 4 , 1 2 , 2 3 , 4 (C)
0,0,0,0

© Stuart Dalziel (Michaelmas, 2021) 73

. Vectors Basis vectors

This gives us four simultaneous equations. Starting with the last of

these immediately gives 4 0 , leaving
1 3 0,
1 2 0,
2 3 0,
as we had for (i). The solution is therefore 1 2 3 4 0.
Note that the structure of the simultaneous equations implied by
(C) suggests that our solution i 0 i is true if and only if
e1 e1 p e1q e1r e1s
e2 e2 p e2 q e2 r e2 s
0
e3 e3 p e3q e3r e3s
e4 e4 p e4 q e4 s e4t

Matrices, determinants and linear independence are discussed

further in Easter term.

Again, we can test for normality and unit length using the vector
products
e1 e1 e2 e2 e3 e3 e4 e4 2
and
e1 e 2 e1 e3 e1 e 4 e 2 e3 e3 e 4 1
e2 e4 0
Hence, only e2 and e4 are orthogonal, and none of the basis
vectors have unit length.

© Stuart Dalziel (Michaelmas, 2021) 74

. Vectors Basis vectors

As we can see from Example 18(i), it is not necessary for the basis
vectors to be normal (perpendicular) to each other. However, if the
basis vectors are mutually perpendicular to each other, then the
basis is said to be orthogonal. If the basis vectors also have unit
length then the basis is orthonormal (we have already mentioned
these terms for three dimensions in section 1.3).
30
orthonormal for N linearly independent O
vectors in N D and d

if i
j
ei
ej o
if i
j
if orthonormal in ND then if I

q ar ar e t r
t an en
en't a
h s bi g t be t t bn en
t

at É aj by

If the basis is not orthonormal, then computing the various vector

products is more complex. We shall not cover this here.

© Stuart Dalziel (Michaelmas, 2021) 75

 . Vectors Other orthogonal coordinates

1.11 Other orthogonal coordinates

The Cartesian x, y, z coordinates are a very familiar example of an
orthogonal coordinate system in three dimensions. However, there
are other examples that can be very useful in specific applications,
and we will describe the three most commonly used ones here.
1.11.1 Cylindrical polar coordinates
31

V O Z
T
T
radius fazimuthal
weight

angle

Figure 20: Cylindrical polar coordinates of the point P relative to the origin O .
The point Q is the projection of P onto the plane z 0 , with view from above on
the right.

We introduce the cylindrical polar coordinates r z shown in

figure 20. The z coordinate is the same as for Cartesian coordinates,
but the r coordinate corresponds to the distance OQ and
corresponds to the angle between the x axis and the vector OQ . By
convention, we take 0 2 , but an alternative is . It
is also very important to note that we must restrict r 0 to ensure
that there is only one way to represent every point (although this
breaks down at the origin).

© Stuart Dalziel (Michaelmas, 2021) 76

 . Vectors Other orthogonal coordinates

in 3D
point P given by E 32
For some
P
Cr O
y 71
Coordinates ctr or Z
C
Restrict r 20 Of Oz 22 or TO EN
R
I
Actually still have many ways of describing Io as u
o would do
any T

Tn roose
p
yr since C
To Z z
the position vector of point
a
in here refers to

I Creo so rsino Z
Cartesian bass vertus

ear er
d ez cont of the page
agitation a
vector
of o of R

eat'siightaTidinger
I
as
they are

eo since it an
op orthogonal
handed
ez I right
U

I enter ez
ler l Stuart Dalzielteal
©teal (Michaelmas, 2021) 77

er ea ea ez ez er o
 . Vectors Other orthogonal coordinates

P
Pythagoras
R
I 21 Jr't z2
Position vector I
R
VE t
Zee t
does not explicitly include 0

but recall that P

o e
er cos t
sing
P
eat since it cos
of 1
2
These are curvilinear coordinates C
not constant but depend on
Basis vectors are

position

This complicates things

vectors when
Use cartesian other constant basis
vector calculations dot prodent
doing e.g
will still use other coord systems
However we

© Stuart Dalziel (Michaelmas, 2021) 78

 combo i a
basis is a
unique of p
. Vectors H
Other orthogonal coordinates

Warning: Although the cylindrical polar basis vectors eˆ r , eˆ , eˆ z are

orthogonal, their orientation depends on . Suppose ex
a ar eˆ r a eˆ az eˆ z and b br er be bz e z
for
where eˆ r , eˆ , eˆ z and er , e , e z are the orthonormal bases for a, b ,
respectively, then when we compute the scalar product

aj b ar eˆ r a eˆ az eˆ z br e r be bz e z
ar br eˆ r e r a b eˆ e az bz eˆ z e z
ar b eˆ r e ar bz eˆ r e z a br eˆ e r
a bz eˆ e z az br eˆ z e r az b eˆ z e

we find this is not simply a b ar br a b az bz as although

eˆ z e z , this will not in general be true for the other basis vectors as
eˆ r , eˆ , er , e all depend on and so we cannot do the normal
simplification. Specifically, 1 eˆ r er 1, 1 eˆ e 1,

Etter
1 eˆ r e 1, 1 eˆ er 1 rather than being equal to 1, 1, 0, 0,
respectively.

Eat
if dit
For NST1A, you are best to always compute scalar and vector
products using Cartesian coordinates!

have orientation
direction
1.11.2 Plane polar coordinates
If we just work in the xy plane, i.e. in two dimensions, then we can
use plane polar coordinates r . These are shown on the right of
figure 20, and are simply cylindrical polar coordinates but with the
z direction suppressed. So for instance in two dimensions

r r cos ˆi r sin ˆj reˆ r (51)

Again, we restrict r 0 and limit 0 2 or and

note that eˆ r is a function of .

© Stuart Dalziel (Michaelmas, 2021) 79

 . Vectors Other orthogonal coordinates

1.11.3 Spherical coordinates

33
z
r O d
d er
d azimuthal Q P e
radius
distance angle r
e
y
from 07 polar
I inclination
angle agle O

Z x
n
T r stull

wig Igraine rsiulsing

y
y y so sa
Figure 21: Spherical polar coordinates of the point P relative to the origin O . The
point Q is the projection of P ontopthe z axis. p
r since cost
We introduce the spherical polar coordinates r shown in figure
21. The r coordinate is the distance of the point P from the origin.
The coordinate is the angle between the upward vertical axis and
the line OP , measured in the direction down from the North Pole,
so that the North Pole has 0 , the equator has 2 and the
South Pole has . The coordinate is the angle measured from
the fixed x axis round to the direction QP in the anticlockwise
sense shown on the right of Figure 16. It is very important to note
the ranges of the coordinates:
0 r,

signs
0 , (52)
yer sincesing 0 (or )

same as the polar coordinates

© Stuart Dalziel (Michaelmas, 2021) 80
 . Vectors Other orthogonal coordinates

Note that the definitions of r and in spherical polar coordinates

are fundamentally different from r and in cylindrical polar
coordinates.

The general point P has position vector r ; from the right-angled

triangle OPQ we see that for P

PQ r sin , OQ kˆ z r cos OQ z r cos .

Then via simple trigonometry, we can see that

x r sin cos ,
Er since cost it
y r sin sin ,

so that sinojl
r r sin cos ˆi r sin sin ˆj r cos kˆ . (53)

By resolving the components of (53) we can express the unit vectors

eˆ r , eˆ , eˆ in terms of ˆi ˆj kˆ . To do this it is best to first note the
vector QP in figure 21 is parallel to cos ˆi sin ˆj , and we can then
resolve this vector into the eˆ r and eˆ directions. This gives

bundle eˆo r sin (cos ˆi sin ˆj) cos kˆ ,

position eˆ cos (cos ˆi sin ˆj) sin kˆ , (54)
eˆ sin ˆi cos ˆj.
34

The three unit vectors eˆ r , eˆ , eˆ are perpendicular to each other, and

form a right-handed orthogonal system. In particular,

eˆ eˆ r eˆ , eˆ r eˆ eˆ , eˆ eˆ eˆ r (55)

We need to be careful, however, as the orientation of these change

with the location!

© Stuart Dalziel (Michaelmas, 2021) 81

 . Vectors Other orthogonal coordinates

Finally, note that in spherical polar coordinates the position vector

of P is very simple:
magnitude
r r eˆ r ,

but that eˆ r is a function of and . direction

Conventions – You need to be aware of this
r O O spherical
We need to be careful using spherical polar coordinates as there
is no universally accepted convention for the order in which the
coordinates are specified.
Cri O Z cylindrical
Here we use the order radius ( r ), inclination angle ( ), azimuthal
angle ( ). This ordering is common practice in physics and is
specified by ISO 31-11 (which defines mathematical signs and
G ol plane polar symbols for use in physical sciences and technology).
However, it is not uncommon to find the order of the two angles
reversed!
The same ISO standard suggests that for cylindrical polar
coordinates we should use rather than r as the radius and
rather than as the angle, giving , , z . Here we have
adopted r , , z as this is very widely used due to the obvious
connection with two-dimensional polar coordinates.
In an exam, read the question carefully to determine what each of
the symbols means. If you introduce a different coordinate system
in your answer, make it clear what means what.

© Stuart Dalziel (Michaelmas, 2021) 82

 . Vectors Other orthogonal coordinates

Example 19: Changing coordinate systems

Express the coordinates of the point (1 1 2) in (a) cylindrical polar
coordinates, (b) spherical polar coordinates.
7
a pi (a) x, y, z 1,1, 2 in cylindrical polar coordinates:

i my r x2 y2 12 12 2
1 y
and tan tan 1 1 45 rad
x 4

r, , z 2, , 2 .
4
d
can noteonly
length 1 x 1 1
(Alternatively, that x r cos cos cos .)
r 2
(b) x, y, z 1,1, 2 in spherical polar coordinates:

r x2 y2 z2 12 12 22 6
Since z r cos , then
1 z 1 2
cos cos 0.615 rad ,
r 6
1 y
and tan tan 1 1 rad
x 4
1 2
r, , 6, cos , .
6 4

© Stuart Dalziel (Michaelmas, 2021) 83

 . Vectors Vector area

1.12 Vector area

35
Area objectives of an object A
isindependent of
orientation
different p.o.us

area can have an orientation

subjective

n I
In B

vector area AT projection of the objate

area onto a
plane
121 A

Lent term vector calculus surface integrals

e.g

Is folds as n'cen da

© Stuart Dalziel (Michaelmas, 2021) 84

 . Vectors Vector area

Z n
shadow
C

y TE
light shine through
II

as if it's divided into

x the 3 perspectives
q
has components is represented as a
of area
vector
Sy j si ti
Cartesian SI i t t

n t Act Il É
A Cnn is it A c
j j
t Ain
Since projection of
Sy A cnn.is in a
if Idifiojat
the light C
I i i
made
i direction how
direction to the S
I
mould pan though
the adjusting factor
I Aee
C
in ti da donations
g A
where
I I Ty Nz C

I d i t E j p ten I I
C

Convention outward normal for a closed body

© Stuart Dalziel (Michaelmas, 2021) 85

 . Vectors Vector area

C
Ease sit sit sit i

EE
si L

Ii
toe entire vector area for And closed
cancels out A
body is o everything eventually
T

Lent term: we will calculate the vector area of a curved surface by

splitting the surface into small plane patches - in the limit of these
patches shrinking to zero size the finite sum is then replaced by an
integral over the whole surface to give the vector area. Calculating
such surface integrals will be described next term, but certain
examples are possible now using the fact that the vector area of a
closed surface is zero.

Example 20: Vector area of a pyramid

Calculate the vector area of the open pyramid with vertices
O (0 0 0) , A (0 1 0) , B (1 1 0) , C (1 0 0) and D ( 12 1 1
2 2 ).

y
D
z A B

O C x

© Stuart Dalziel (Michaelmas, 2021) 86

. Vectors Vector area

Now, the vector area for the entire closed body is S 0.

As the vector area of the whole is the sum
S SOABC SOAD S ABD S BCD SCOD
and the vector area of the base (area 1 and normal in the kˆ
direction) is
SOABC k,
then the vector area of the upper four sides is
Stop SOAD S ABD S BCD SCOD kˆ .
Harder way:
Can readily work out vector area for the individual faces. For
example, the face OAD has normal nˆ OAD 1
2
, 0, 12 and area
1
AOAD 1
2 1 1
2
, then the vector area is
2 2
SOAD 1
4 ,0, 14 .
We can doing this in one step by recalling our expression for the
area of the base of a parallelipeped, and noting that the area of a
triangle is half this and so
ˆi ˆj kˆ
1 ˆ ˆj 0 0
SOAD 1
2 OD OA 1 1
2 2
1
2
1
2 i 0 1
2 kˆ 1
2 0
2
0 1 0
1
4 ,0, 14 ,
as before. Note that we need to be careful in choosing the order
OD OA and not OA OD to ensure we have the correct
(outward) direction for the normal.
It is obvious that the vertical component of SOAD had to be 14 since
there are four faces that add up to a vertical component of 1. The
horizontal component of SOAD is balanced by S BCD 4 ,0, 4 .
1 1

© Stuart Dalziel (Michaelmas, 2021) 87

 . Vectors Vector area

no need calculate
to vector oven individed
of
ten 21: Vector areathem
faces a Example
adding up
of a hemispherical shell
Calculate the vector area of the outside of a hemispherical shell of
radius a .
use the
a property of cloned body
definition
to advantage
objects gas

36

closed Q T

F
shell s dosed base

O ta c k
A
act I A

O
B C

© Stuart Dalziel (Michaelmas, 2021) 88

. Vectors Worked Examples on Vectors

1.13 Worked Examples on Vectors

Note that Tripos questions generally indicate the approximate
number of marks for each part. This often gives some indication on
the relative difficulty or effort required, but not always!

Example 22: Tripos question 2005, Paper 1, 5C

The vertices of a tetrahedron O P Q R have coordinates (0 0 0) ,
(2 1 1) , (1 2 2) and (0 0 3) respectively. Find by vector methods
(a) the angle between the faces OPR and OQR [5]
(b) the angle between the vector OP and the normal to the face
PQR [5]
(c) the area of the face PQR [5]
(d) the shortest distance from the origin to the plane containing P ,
Q and R . [5]
37

© Stuart Dalziel (Michaelmas, 2021) 89

. Vectors Worked Examples on Vectors

(
B
t
o

A
s

© Stuart Dalziel (Michaelmas, 2021) 90

. Vectors Worked Examples on Vectors

a
T

© Stuart Dalziel (Michaelmas, 2021) 91

. Vectors Worked Examples on Vectors

T
P
i

© Stuart Dalziel (Michaelmas, 2021) 92

. Vectors Worked Examples on Vectors

Example 23: Tripos 2003, paper 1, question 1A

The vectors a b c and A B C form reciprocal sets, defined such
that
b c c a a b
A B C
a (b c) a (b c) a (b c)
and a (b c) 0.
(a) Show that:
(i) A a B b C c 1
(ii) A b A c 0
(iii) A B C are non-coplanar. [8]

(b) Show that the vectors a ( 1 1 0) and b (0 2 1) and

c (1 0 1) are non-coplanar. Find the reciprocal basis, and hence
write the vector d (2 1 1) in terms of the basis a b c . [12]
b c a b c
(a)(i) A a a 1.
a b c a b c
c a b c a a b c
Similarly B b b 1,
a b c a b c a b c
where we have used the cyclic property for the scalar triple
product.
Likewise,
a b b c a a b c
C c b 1
a b c a b c a b c

© Stuart Dalziel (Michaelmas, 2021) 93

. Vectors Worked Examples on Vectors

(a)(ii) Since b c is perpendicular to both b and c , then

b c
A b b 0
a b c
b c
Similarly, A c c 0
a b c
(a)(iii) If A B C are co-planar, then the scalar triple product will be
zero. Check:
b c c a a b
A B C
a b c a b c a b c
3
1
b c c a a b
a b c
Noting that p q r p r q p q r and identifying
p c a , q a and r b , then
3
1
A B C b c c a b a c a a b .
a b c
Noting that since a, b, c are not co-planar, and c a a 0 , then
3
1
A B C b c c a b a
a b c
3
1
c a b b c a
a b c
3
1
a b c a b c
a b c
1
0
a b c

© Stuart Dalziel (Michaelmas, 2021) 94

. Vectors Worked Examples on Vectors

(b) Show that the vectors a ( 1 1 0) and b (0 2 1) and

c (1 0 1) are non-coplanar.
Check the scalar triple product
1 1 0
a b c 0 2 1 1 2 0 11 0 2 1 3.
1 0 1
Since this is nonzero, the vectors are not co-planar.

Find the reciprocal basis, and hence write the vector d (2 1 1) in

terms of the basis a b c . [12]
Noting that
ˆi ˆj kˆ
a b 1 1 0 1,1, 2
0 2 1
It is worth checking that this is orthogonal to a and b , etc.
ˆi ˆj kˆ
b c 0 2 1 2,1, 2
1 0 1
ˆi ˆj kˆ
c a 1 0 1 1,1,1
1 1 0

And recalling a b c 3 , then

b c
A , ,
2 1
3 3
2
3 ,
a b c
c a
B , ,
1 1 1
3 3 3 ,
a b c

© Stuart Dalziel (Michaelmas, 2021) 95

. Vectors Worked Examples on Vectors

a b
C 1 1
3 3, , 2
3 .
a b c
If we write d (2 1 1) in terms of a ( 1 1 0) , b (0 2 1) and
c (1 0 1) as
d ada bdb cdc
then
2 da dc (A)
1 da 2d b (B)
1 db dc (C)
Taking (A)+(B) 3 2db dc and adding (C)
2
2 3db db
3
1
da 1 2d b
3
5
dc db 1
3
1 2 5
d a b c
3 3 3

The smarter way is to note that

da A d 2 1
3 3, , 2
3 2,1, 1 1
3 4 1 2 1
3

db B d , ,
1 1 1
3 3 3 2,1, 1 1
3 2 1 1 2
3

da C d , ,
1 1
3 3
2
3 2,1, 1 1
3 2 1 2 5
3

1 2 5
d a b c
3 3 3

© Stuart Dalziel (Michaelmas, 2021) 96

. Vectors Worked Examples on Vectors

Example 24: Tripos 2004, paper 1, question 1A

The components of the vectors a b c are (a1 a2 a3 ) , (b1 b2 b3 ) and
(c1 c2 c3 ) , respectively.
(a) Write down, in terms of the components, the scalar triple
product a (b c) . Show that
a (b c) (a b) c [4]
(b) Using vector methods, find the equation, in the form
x y z , of the plane passing through the points (2, 1, 2) ,
(1 2 3) and (4 1 0) . [12]
(c) Let i j and k be mutually orthogonal unit vectors. Find the
volume of the parallelepiped, three of whose edges are i 3j ,
i 3j 2k and j k . [4]
(a) Components of scalar triple product:
a1 a2 a3
b2 b3 b3 b1 b1 b2
a (b c) b1 b2 b3 a1 a2 a3
c2 c3 c3 c1 c1 c2
c1 c2 c3
a1 b2 c3 b3c2 a2 b3c1 b1c3 a3 b1c2 b2c1
a1b2 c3 a2b3c1 a3b1c2 a1b3c2 a2b1c3 a3b2 c1
c1 c2 c3
(a b) c c (a b) a1 a2 a3
b1 b2 b3
c1 a2b3 a3b2 c2 a3b1 a1b3 c3 a1b2 a2b1
a1b2 c3 a2b3c1 a3b1c2 a1b3c2 a2b1c3 a3b2c1
Hence a (b c) (a b) c , QED.
(b) Equation of a plane can be written as
(r a) n 0
r n a n

© Stuart Dalziel (Michaelmas, 2021) 97

. Vectors Worked Examples on Vectors

where r x, y, z is any point on the plane, a is a known point on

the plane and n , , is normal to the plane such that
a n . Thus r n x y z a n.
Here, we select a 2, 1,2 , b 1,2,3 and c 4,1,0 so that
the vectors b a and c a are parallel to the plane. Hence,
n b-a c a 1,3,1 2,2, 2
ˆi ˆj kˆ
1 3 1 ˆi 6 2 ˆj 2 2 kˆ 2 6
2 2 2
8,0, 8
is normal to the plane and
a n 2, 1,2 8,0, 8 16 16 32
Thus the equation of the plane can be written as
8x 8z 32
x z 4.
(c) The volume of a parallelepiped is
V a b c,
where here a i 3j 1,3,0 , b i 3j 2k 1,3, 2 and
c j k 0,1,1 . Thus
1 3 0
V a b c 1 3 2
0 1 1
13 2 3 0 1 0 1 0
8
(As this is positive, we don’t need to worry about taking the
absolute value.)

© Stuart Dalziel (Michaelmas, 2021) 98