Div, Grad, Curl

Charles Byrne (Charles [email protected]) Department of Mathematical Sciences University of Massachusetts Lowell Lowell, MA 01854, USA February 1, 2009
Abstract When we begin to study vector calculus, we encounter a number of new concepts, divergence, gradient, curl, and so on, all related to the del operator, . Shortly thereafter, we are hit with a blizzard of formulas relating these concepts. It is all rather abstract and students easily lose their way. It occurred to Prof. Schey of MIT to present these ideas to his students side-by-side with the basics of electrostatics, which, after all, was one of the main applications that drove the development of the vector calculus in the rst place. Eventually, he wrote a small book [1], which is now a classic. These notes are based, in part, on that book.

The Electric Field

The basic principles of the electrostatics are the following: 1. there are positive and negative electrical charges, and like charges repel, unlike charges attract; 2. the force is a central force, that is, the force that one charge exerts on another is directed along the ray between them and, by Coulombs Law, its strength falls o as the square of the distance between them; 3. super-position holds, which means that the force that results from multiple charges is the vector sum of the forces exerted by each one separately. Apart from the rst principle, this is a good description of gravity and magnetism as well. According to Newton, every massive body exerts a gravitational force of attraction on every other massive body. A space craft heading to the moon feels the 1

attractive force of both the earth and the moon. For most of the journey, the craft is trying to escape the earth, and the eect of the moon pulling the craft toward itself is small. But, at some point in the journey, the attraction of the moon becomes stronger than that of the earth, and the craft is mainly being pulled toward the moon. Even before the space craft was launched, something existed up there in space, waiting for a massive object to arrive and experience attractive force. This something is the gravitational eld due to the totality of massive bodies doing the attracting. Einstein and others showed that gravity is a bit more complicated than that, but this is a story for another time and a dierent teller. Faraday, working in England in the rst half of the nineteenth century, was the rst to apply this idea of a eld to electrostatics. He reasoned that a distribution of electrical charges sets up something analogous to a gravitational eld, called an electric eld, such that, once another charge is placed within that eld, it has a force exerted on it. The important idea here is that something exists out there even when there is no charge present to experience this force, just as with the gravitational eld. There are also magnetic elds, and the study of the interaction of electric and magnetic elds is the focus of electromagnetism.

The Electric Field Due To A Single Charge

Suppose there is charge q at the origin in three-dimensional space. The electric eld resulting from this charge is E(x, y, z) = where u(x, y, z) = ( x2 x2 q u(x, y, z), + y2 + z2 (2.1)

x y z , 2 , 2 ) 2 + z2 2 + z2 +y x +y x + y2 + z2

is the unit vector pointing from (0, 0, 0) to (x, y, z). The electric eld can be written in terms of its component functions, that is, E(x, y, z) = (E1 (x, y, z), E2 (x, y, z), E3 (x, y, z)), where qx , + y 2 + z 2 )3/2 qy E2 (x, y, z) = 2 , (x + y 2 + z 2 )3/2 E1 (x, y, z) = (x2

and E3 (x, y, z) = (x2

qz . + y 2 + z 2 )3/2

It is helpful to note that these component functions are the three rst partial derivatives of the function (x, y, z) = x2 q . + y2 + z2 (2.2)

Gradients and Potentials

Because of the super-position principle, even when the electric eld is the result of multiple charges it will still be true that the component functions of the eld are the three partial derivatives of some scalar-valued function (x, y, z). This function is called the potential function for the eld. For any scalar-valued function f (x, y, z), the gradient of f at the point (x, y, z) is the vector of its rst partial derivatives at (x, y, z), that is, f (x, y, z) = ( the vector-valued function f f f (x, y, z), (x, y, z), (x, y, z)); x y z

f is called the gradient eld of f . Therefore, the electric

eld E is the gradient eld of its potential function.

Gausss Law

Lets begin by looking at Gausss Law, and then well try to gure out what it means.

Gausss Law: E n dS = 4
S V

dV.

(4.1)

The integral on the left side is the integral over the surface S, while the integral on the right side is the triple integral over the volume V enclosed by the surface S. We must remember to think of integrals as summing, so on the left we are summing something over the surface, while on the right we are summing something else over the enclosed volume.

4.1

The Charge Density Function

The function = (x, y, z) assigns to each point in space a number, the charge density at that point. The vector n = n(x, y, z) is the outward unit normal vector to the surface at the point (x, y, z) on the surface, that is, it is a unit vector pointing directly out of the surface at the point (x, y, z).

4.2

The Flux
E n = E(x, y, z) n(x, y, z)

The dot product

is the size of the component of the electric eld vector E(x, y, z) that points directly out of the surface. The surface integral on the left side of Equation (4.1) is a measure of the outward ux of the electric eld through the surface. If there were no charges inside the surface S there would be no outward ux. Gausss Law tells us that the total outward ux that does exist is due to how much charge there is inside the surface, that is, to the totality of charge density inside the surface. Our goal is to nd a convenient way to determine the electric eld everywhere, assuming we know the charge density function everywhere. Gausss Law is only a partial answer, since it seems to require lots of surface and volume integrals.

A Local Gausss Law and Divergence

Gausss Law involves arbitrary surfaces and the volumes they enclose. It would be more helpful if the law could be expressed locally, at each point in space separately. To achieve this, we consider a xed point (x, y, z) in space, and imagine this point to be the center of a sphere. We apply Gausss Law to this sphere and get the ux through its surface. Now we imagine shrinking the sphere down to its center point. In the limit, the ratio of the ux to the volume of the sphere, as the radius of the sphere goes to zero, is the number div E(x, y, z) = E1 E2 E3 + + . x y z (5.1)

When we apply the same limiting process to the integral on the right side of Gausss Law, we just get 4(x, y, z). Therefore, the local or dierential form of Gausss Law becomes div E(x, y, z) = 4(x, y, z). 4 (5.2)

This is also the rst of the four Maxwells Equations. Our goal is to determine the electric eld from knowledge of the charge density function . The partial dierential equation in (5.2) is not enough, by itself, since it involves three dierent unknown functions, E1 , E2 , and E3 , and only one known function . The next step in solving the problem involves the potential function for the electric eld.

5.1

The Laplacian

For a scalar-valued function f (x, y, z) the Laplacian is

f (x, y, z) =

2f 2f 2f + 2 + 2 = x2 y z

( f) =

For a vector-valued function F(x, y, z) = (F1 (x, y, z), F2 (x, y, z), F3 (x, y, z)), the symbol
2

F is the vector-valued function whose components are the Laplacians

of the individual F1 , F2 , and F3 , that is,

F=(

F1 ,

F2 ,

F3 ).

Poissons Equation and Harmonic Functions

As we discussed earlier, the component functions of the electric eld are the three rst partial derivatives of a single function, (x, y, z), the electrostatic potential function. Our goal then is to nd the potential function. When we calculate the divergence of the electric eld using we nd that div E(x, y, z) = 2 2 2 + + 2 = x2 y 2 z ( ) =
2

Therefore, the dierential form of Gausss Law can be written as

(x, y, z) = 4(x, y, z);

(6.1)

this is called Poissons Equation. In any region of space where there are no charges, that is, where (x, y, z) = 0, we have
2

(x, y, z) = 0.

(6.2)

Functions that satisfy Equation (6.2) are called harmonic functions. The reader may know that both the real and imaginary parts of a complex-valued analytic function are harmonic functions of two variables. This connection between electrostatics and complex analysis motivated the (ultimately fruitless) search for a three-dimensional extension of complex analysis.

The Curl

The divergence of a vector eld is a local measure of the ux, which we may think of as outward ow of something. The curl is a measure of the rotation of the something. For any vector eld F(x, y, z) = (F1 (x, y, z), F2 (x, y, z), F3 (x, y, z)), the curl of F is the vector eld F3 F2 F1 F3 F2 F1 curl F(x, y, z) = F = ( , , ). (7.1) y z z x x y A useful identity involving the curl is the following: ( F) = ( F)
2

F.

(7.2)

7.1

An Example

Imagine that the points (x, y, z) in three-dimensional space are moving in time, revolving counter-clockwise around the z-axis in such a way that the velocity vector at each point is v(x, y, z) = (y, x, 0), where is the constant angular speed. Then curl v(x, y, z) = (0, 0, 2). The divergence of the velocity eld is div v(x, y, z) = 0. The motion here is rotational; there is no outward ow of anything. Here the curl describes how fast the rotation is, and indicates the axis of rotation; the fact that there is no outward ow is indicated by the divergence being zero.

7.2

Solenoidal Fields

When the divergence of a vector eld is zero, the eld is said to be solenoidal; the velocity eld in the previous example is solenoidal. The second of Maxwells four equations is that the magnetic eld is solenoidal. 6

7.3

The Curl of the Electrostatic Field

We can safely assume that the mixed second partial derivatives of the potential function satisfy 2 2 = , xy yx 2 2 = , xz zx and 2 2 = . zy yz It follows, therefore, that, because the electrostatic eld has a potential, its curl is zero. The third of Maxwells Equations (for electrostatics)is curl E(x, y, z) = 0. (7.3)

The Magnetic Field

We denote by B(x, y, z) a magnetic eld. In the static case, in which neither the magnetic eld nor the electric eld is changing with respect to time, there is no connection between them. The equations that describe this situation are Maxwells Equations for the Static Case: 1. div E = 4; 2. curl E = 0; 3. div B = 0; 4. curl B = 0. It is what happens in the dynamic case, when the electric and magnetic elds change with time, that is interesting. Ampere discovered that a wire carrying a current acts like a magnet. When the electric eld changes with time, there is a current density vector eld J proportional to the rate of change of the electric eld, and Item 4 above is replaced by Amperes Law: curl B = a E , t

where a is some constant. Therefore, the curl of the magnetic eld is proportional to the rate of change of the electric eld with respect to time. Faraday (and also Henry) discovered that moving a magnet inside a wire coil creates a current in the wire. When the magnetic eld is changing with respect to time, the electric eld has a non-zero curl proportional to the rate at which the magnetic eld is changing. Then Item 2 above is replaced by curl E = b B , t

where b is some constant. Therefore, the curl of the electric eld is proportional to the rate of change of the magnetic eld. It is this mutual dependence that causes electromagnetic waves: as the electric eld changes, it creates a changing magnetic eld, which, in turn, creates a changing electric eld, and so on.

Electro-magnetic Waves

We consider now the behavior of electric and magnetic elds that are changing with time, in a region of space where there are no charges or currents. Maxwells Equations are then 1. div E = 0; 2. curl E = b B ; t 3. div B = 0; 4. curl B = a E . t We then have ( E) = b( B ) = b ( t t B) = ab E 2E ( ) = ab 2 . t t t

Using Equation (7.2), we can also write ( Therefore, we have 2E , t2 which means that, for each i = 1, 2, 3, the component function Ei satises the three2

E) =

E)

E = div E

E=

E.

E = ab

dimensional wave equation 2 Ei = c2 t2 8

Ei .

The same is true for the component functions of the magnetic eld. Here the constant c is the speed of propagation of the wave, which turns out to be the speed of light. It was this discovery that suggested to Maxwell that light is an electromagnetic phenomenon.

References
[1] Schey, H.M. (1973) Div, Curl, Grad, and All That, W.W. Norton.