Chapter 3 Div, Grad and Curl

Vector Fields
One interesting quality of vectors is that they have no position - a vector by itself can always be pictured as starting at the origin (This is in fact why you are allowed to add them by placing the tip of one against the tail of another). In physics however we often nd that a vector quantity (such as force) will vary from place to place. A function which attributes a unique vector to each point in space is called a vector eld. Ive drawn a couple of vector elds below. Clearly it is not possible to represent every vector in the eld, since there is a dierent vector at every point and so they are packed innitely densely.

Vector elds are extremely useful in physics and can represent anything from the velocity of uids, the displacement caused by an elastic vibration, or an electric eld.

15

16 Dierentiation of vector elds

CHAPTER 3. DIV, GRAD AND CURL

Happily, dierentiation of vectors follows the same set of rules which we use dierentiation of scalars. If a vector F depends on some parameter t, such that F(t) = Fx (t) i + Fy (t) j + Fz (t)k then the derivative with respect to t is dF dFx dFy dFz = i+ j+ k. dt dt dt dt (3.1)

In this way we can see that vector dierentiation inherits all the rules for scalar dierentiation. These are:

Addition:

d dF dG (F + G)= + dt dt dt d dF du (F(u(t)))= dt dt dt d dF (cF)=c dt dt d du dF (uF)= F + u dt dt dt

(3.2)

The chain rule:

(3.3)

Multiplication by a constant:

(3.4)

Product rules:

d dF dG (F G)= G+F dt dt dt d dF dG (F G)= G+F dt dt dt

(3.5)

Note that in the product rules we have specied the extra rules for using the dot () and cross () products.

17

The gradient
Dierentiation describes the rate of change of a function with respect to a given parameter. If this parameter is one of the coordinates x, y , or z , then the partial derivative f (x, y, z )/x denotes the slope of the function, as measured parallel to the x axis. That is, f (x, y, z )/x measures the rate of change of the function f in the x direction. We will now introduce a generalization of dierentiation to three dimensions. Because the function f (x, y, z ) depends on each of three coordinates, we require a vector eld to describe the slope of f in each of the three coordinate directions. The simplest way to contain all this information is to put all the partial derivatives of the function into a vector. This vector eld is known as the gradient of the function. Given a scalar function f (x, y, z ), we dene the gradient of f to be the vector f = The gradient has the following properties: f is a vector eld f measures the rate of increase of the scalar function f in each of the three coordinate directions f points in the direction in which f increases the most.
1 0.5 0 -0.5 -1 1 0.75 0.5 0.25 0 -2 0 2 4
-1 -2 -1 0 1 2 3 4 -0.5 0 0.5 1

f f f , , x y z

(3.6)

To illustrate the last point, we consider the (two-dimensional) scalar function above, which has been represented both as a surface (left) and as a contour graph (right). Over the top of the contour function have been drawn the vectors representing the scalar eld f . Each vector points in the direction in which the function f increases the most, and the length of each vector is proportional to the steepness of the slope.

18

CHAPTER 3. DIV, GRAD AND CURL

Problems
1. Given a scalar eld f (r) = rn , and r = |r|, show that f = nrn2 r. where r = x i + y j + zk

2. Find the gradient of the eld f (x, y, z ) = where q is a constant. q x2 + y 2 + z 2 ,

19

The operator
The gradient f has the form of a vector multiplying a scalar function: f = , , x y z f .

The object in parentheses is not in fact a vector, although behaves very much like one. Instead it is an example of what is known as a vector operator, and we can re-write it as = or, equivalently, , , x y z , (3.7)

= i + j +k (3.8) x y z An operator is something which operates on a function to produce another function. In this case is perhaps best looked at as an instruction to dierentiate a function in each of the three coordinate directions. As with an ordinary vector, we can use to multiply in three dierent ways: 1. Multiplication by a scalar function f : the result is the vector f (the gradient). 2. Dot product with a vector function v: the result is a scalar v, known as the divergence 3. Cross product with a vector function v: the result is a vector v, known as the curl

The Divergence of a vector eld

For a vector eld v = (vx , vy , vz ), the divergence of the eld at any point is dened by the number v = , , (vx , vy , vz ) x y z vx vy vz + + . = x y z

(3.9)

Note that this is clearly a scalar quantity. You will often see the divergence written div v = v .

20

CHAPTER 3. DIV, GRAD AND CURL

The divergence is essentially a measure of the expansion of a vector eld at any given point. That is, it gives the amount that the vector eld spreads out (or diverges) from a particular point.

Example: Consider the vector function v(x, y, z ) = x i + y j + zk The divergence of this function is ) (vx ) v = ( i + j +k i + vy j + vz k x y z ) (x ) = ( i + j +k i + y j + zk x y z = (x) + (y ) + (z ) = 1 + 1 + 1 = 3 . x y z Thus the function has a positive divergence, or spreading out at every point, as is illustrated

-1

-2 -2 -1 0 1 2

by the gure above.

21

The curl of a vector eld

From the denition of the vector, or cross product, we have: i v =
x

j
y

k
z

(3.10)

vx vy vz vz vx vz vy j = i y z x z vy vx +k . x y

(3.11)

The curl is a measure of how much (and in which direction) a vector eld curls about itself. Vector functions which twist dramatically typically have large curls in the vicinity of the twisting, whereas smooth, laminar functions tend to have small values of curl. To illustrate:

Example (i) Consider the vector function v(x, y, z ) = y i + x j+k The curl of this function is

-1

-2 -2 -1 0 1 2

22

CHAPTER 3. DIV, GRAD AND CURL

i v =
x

j
y

k
z

y x 1 (1) (x) (1) (y ) = i j y z x z (x) (y ) +k x y (1 + 1) = 2k . = i(0 0) + j(0 0) + k Example (ii) By way of contrast, consider the vector function we looked at previously v(x, y, z ) = x i + y j + zk The curl of this function is i v =
x

j
y

k
z

x y z (z ) (x) (z ) (y ) j = i y z x z (y ) (x) +k x y (0 0) = 0 . = i(0 0) + j(0 0) + k This function has zero curl. Vector elds of this type are known as irrotational.

23

Problems
1. Find the gradients of the following functions: (a) f (x, y, z ) = 2x2 + y 2 + 3z 3 (b) f (x, y, z ) = x2 y 3 z 4 (c) f (x, y, z ) = ez sin x ln(y ) 2. Find (i) the divergence, and (ii) the curl of the following vector functions: (a) v = x2 i + 2xy j + 3z 2 k (b) v = xy i + 2yz j + 3zxk 3. Calculate the divergence of the function 1 v= r , r , and r = |r|. where r = x i + y j + zk

24

CHAPTER 3. DIV, GRAD AND CURL

Second derivatives
We can form more complicated operations by combining the dierent permutations of gradient, divergence and curl. If is a scalar eld, and v = (vx , vy , vz ) is a vector eld, we obtain:

1. The divergence of a gradient:

, , , , x y z x y z 2 2 2 = + + 2 x2 y 2 z This operation is known as the Laplacian of , and is usually written () = 2 2 + + . (3.12) x2 y 2 z 2 This dierential operation is one of the most widely used in the entire eld of physics. Important examples are: 2 = 2 = 0 Laplaces equation 1 2 = Heat conduction, or diusion equation t 1 2u 2 u = 2 2 Wave equation c t Note that the Laplacian is a scalar quantity.

2. The curl of a gradient:

j +k y z z y x z z x x y y x 2 2 2 2 2 2 = i j +k yz zy xz zx xy yx =0 = 0 i + 0 j + 0k (3.13)

() = i

So this is in general true: the curl of a gradient is always zero. Vector elds for which the curl is zero are called irrotational.

25

3. The gradient of a divergence ( v):

This second derivative has no special name, but appears in the equations of elasticity.

4. The divergence of a curl:

( v) = 0 . (3.14)

The proof of this identity is along the same lines as that for calculating the curl of a gradient, given above. Vector elds which have zero divergence everywhere are called solenoidal.

5. The curl of a curl:

( v) = ( v) 2 v
2

(3.15)

The proof of this identity is left as an exercise. Note that the Laplacian operator operates on each of the three components of the vector v separately. The curl of a curl is a vector quantity. Example (i) Calculate the Laplacian of the function f = x2 + 2xy 2 + 4z 3 . Solution: 2 f = 2f 2f 2f + + x2 y 2 z 2 = 2 + 4x + 24z

Example (ii) Given r = x2 + y 2 + z 2 , show that the scalar function = 1 r

26 obeys Laplaces equation, given above. Solution: 2 (r1 ) = 2 x x = x (x2

CHAPTER 3. DIV, GRAD AND CURL

1 2 2 1 x = 3 r r

x + + z 2 )1/2 2x 1 2 2 2 2 3 / 2 2 (x + y + z ) (x + y + z 2 )1/2 y2

Similarly,

y2 1 2 (r1 ) = , y 2 r3 r z2 1 2 (r1 ) = . z 2 r3 r 2 (r1 ) 2 (r1 ) 2 (r1 ) + + x2 y 2 z 2 2 2 2 2 x +y +z 1 r 1 = = =0 . r3 r r2 r

and Hence, 2 =

27

Problems
1. Calculate the Laplacian ( 2 ) of the following functions: (a) = x2 + (y + 1)2 + xz 2 (b) = cos x cos y cos z 2*. Prove the identity (3.14), i.e. ( v) = 0 . 3*. From the denition of , prove the product identity ( ) = + . given that and are both scalar elds.

28

CHAPTER 3. DIV, GRAD AND CURL