School of Mathematics and Statistics, The University of Sydney
 11. The vector product
Main menu Section menu   Previous section Next section The vector product in terms of components The algebraic rules of the vector product
Glossary
Examples
Quiz

Derivation of the formula

Page 3 of 7 

Unless one (or both) of the vectors is zero, it is not obvious how to get the formula to compute the vector product from the geometric definition of the vector product. If you are interested in how to obtain it, here is a possible way.

The procedure depends on looking at projections of the parallelograms spanned by the vectors u and v into the three coordinate planes. We proceed in three steps.

Step 1

We first look at vectors in the xy-plane. The magnitude of u × v is

|u ×  v|=  |u ||v|sin h,

so u×v = ±(|u||v| sin h)k. The sign is positive if moving u to v through the smaller angle is counterclockwise and negative in the other case. Hence if we let h positive in the former and negative in the latter case we get

u × v = (|u||v|sinh)k.

(We use that sin(-h) = - sin h.) Now let u _L be the vector obtained by rotating u counterclockwise by 90°. As |u _L | = |u| it follows from the definition of the scalar product that

                         o          _L  |u||v |sin h = |u||v |cos(90 -  h) = u  · v.

Two possible situations are shown below

             _L             u    o             h         u  90 -  h v u_ L       90o - h      u                  h                           v

We therefore get u × v = (u _L · v)k.

Now assume that u = u1i + u2j and v = v1i + v2j. Then from the figure below u _L = -u 2i + u1j.

              y                u1         u _L            u2                           u   - u2                        u1    x

Using the formula for the scalar product we get

u _L  · v = (- u2i + u1j) · (v1i + v2j) = - u2v1 + u1v2 = u1v2 - u2v1

and therefore u × v = (u _L · v)k = (u 1v2 - u2v1)k as claimed.

Step 2

Let now u and v be two non-zero and non-parallel vectors in space. Denote by P the parallelogram they span when placed tail-to-tail. By definition of the vector product the area of P is A = |u × v|.

We now project P along the z-axis onto the xy-plane. That projection, P', is spanned by the projections u' and v' of u and v onto the xy-plane.

It will be shown in Step 3 that the area, A', of P' is given by

A'=  A cosa,

where a is the acute angle between the planes containing P and P'.

Observe that a is also the acute angle between the z-axis and w := u × v, which are perpendicular to the respective planes.

Writing w = w1i + w2j + w3k it follows that

cosa =  |w3-| = --|w3-|-         |w |   |u × v|

as you see from the figures below, representing a cross section along w and the z-axis.

     z w3       a   w = u ×  v             a       xy -plane                  plane determined by u and v      z          plane determined  by u and v          a                   xy-plane       a   w  w3

If w3 > 0 the triple u', v', k is right handed, otherwise it is left handed. Hence by the definition of the vector product uv' = w3k. If we note that u' = u1i + u2j and v' = v1i + v2j we conclude from Step 1 that w3 = u1v2 - u2v1.

We get the other components w1 and w2 by the cyclic permutation i --> j --> k --> i. If we do so we get w1 = u2v3 - u3v2 and w2 = u3v1 - u1v3, proving the formula for u × v.

Step 3

We determine how the area of a parallelogram is changed when it is projected onto a plane. Suppose that E1 and E2 are two planes intersecting at a line l. If R is a rectangle on E2 with area A and one edge parallel to l, then the area of its projection, R', orthogonal to E1 is A cos a, where a is the acute angle between the two planes:

                          R                                 R'                   a   l

Now look at an arbitrary parallelogram P on E2 spanned by vectors u and v. We can transform it into a rectangle of the same area such that one edge of that rectangle is parallel to l as shown below

v        u                            l

The projections of the black, green and red parallelograms into E1 have all the same area as the picture looks very similar in E1. Suppose the area of the original black parallelogram is A2 and that of its projection is A1. Then the area of the red rectangle is A2 and that of its projection is A1. From the previous considerations we know that

A1  = A2 cosa

with a being the acute angle between E1 and E2.

Feedback
Main menu Section menu   Previous section Next section The vector product in terms of components The algebraic rules of the vector product