Introduction to Electromagnetism

Gradient and Divergence

(2-6, 2-7, 2-8)

Yoonchan Jeong
School of Electrical Engineering, Seoul National University
Tel: +82 (0)2 880 1623, Fax: +82 (0)2 873 9953
Email: yoonchan@snu.ac.kr
Gradient of a Scalar Field (1)
Let us consider a scalar function of space coordinates: V (u1 , u2 , u3 )
Suppose that there are two surfaces on which V is constant:
→ V = V1 & V = V1 + dV
← Very close to each other!

D. K. Cheng, Field and Wave Electromagnetics, 2nd ed., Addison-Wesley, 1989.

Normal vector
Space rate of change: dV / dl → Maximum when: dl // dn
Gradient of a scalar: The vector of the maximum space rate of increase
of the scalar
dV
grad V ≡ a n ← a n : Unit normal vector
dn
dV
→ ∇V ≡ a n
dn 2
Gradient of a Scalar Field (2)
Directional derivative along dl :
dV dV dn dV
= = cos α
dl dn dl dn
dV
= a n ⋅ a l = (∇V ) ⋅ a l
dn D. K. Cheng, Field and Wave Electromagnetics, 2nd ed., Addison-Wesley, 1989.

→ dV = (∇V ) ⋅ dl ← dl = a u1 dl1 + a u2 dl2 + a u3 dl3

= a u1 (h1du1 ) + a u2 (h2 du2 ) + a u3 (h3 du3 )
Expressed in terms of the differential changes in coordinates:
∂V ∂V ∂V
→ dV = dl1 + dl2 + dl3 ← Chain rule
∂l1 ∂l2 ∂l3
 ∂V ∂V ∂V 
 ⋅ (a u1 dl1 + a u2 dl2 + a u3 dl3 )
??

=  a u1 + a u2 + a u3
 ∂l1 ∂l2 ∂l3 
 ∂V ∂V ∂V 

=  a u1 + a u2 + a u3  ⋅ dl
 ∂l1 ∂l2 ∂l3 
∂V ∂V ∂V ∂V ∂V ∂V
∴ ∇V = a u1 + a u2 + a u3 = a u1 + a u2 + a u3
∂l1 ∂l2 ∂l3 h1∂u1 h2 ∂u2 h3∂u3 3
Gradient of a Scalar Field (3)
Recall:
∂V ∂V ∂V ∂V ∂V ∂V
∇V = a u1 + a u2 + a u3 = a u1 + a u2 + a u3
∂l1 ∂l2 ∂l3 h1∂u1 h2 ∂u2 h3∂u3

Vector differential operator (del):

∂ ∂ ∂
∇ ≡ a u1 + a u2 + a u3
h1∂u1 h2 ∂u 2 h3∂u3
In Cartesian coordinates:
∂ ∂ ∂
∇ ≡ ax + ay + az
∂x ∂y ∂z

4
Divergence of a Vector Field (1)
Directed field lines for vector fields: Flux lines or streamlines

D. K. Cheng, Field and Wave Electromagnetics, 2nd ed., Addison-Wesley, 1989.

Consider a volume with an enclosed surface:

1. An excess of outward flow through the surface:
→ A net positive divergence → The volume contains a source.
2. An excess of inward flow through the surface:
→ A net negative divergence → The volume contains a sink.
3. No excess of outward or inward flow through the surface:
→ A net zero divergence → Neither a source nor a sink
5
Divergence of a Vector Field (2)
Definition of the divergence a vector field A at a point:
The net outward flux of A per unit volume as the volume about
the point tends to zero:
div A ≡ lim
∫S
A ⋅ ds
∆v →0 ∆v
Consider a differential volume of sides ∆x, ∆y and ∆z in Cartesian coordinates:
A ≡ a x Ax + a y Ay + a z Az

∫ A ⋅ ds = [ ∫
S
front
face
+ ∫back + ∫right + ∫left + ∫top + ∫bottom ] A ⋅ ds
face face face face face

∫front
face
A ⋅ ds = A front ⋅ ∆s front = A front ⋅ a x (∆y∆z )
face face face
∆x
= Ax ( x0 +
, y0 , z0 )∆y∆z
D. K. Cheng, Field and Wave Electromagnetics, 2nd ed., Addison-Wesley, 1989.
2
→ Taylor series expansion:
∆x ∆x ∂Ax
Ax ( x0 + , y0 , z0 ) = Ax ( x0 , y0 , z0 ) + + H.O.T.
2 2 ∂x ( x0 , y0 , z0 )
6
 Divergence of a Vector Field (3)
Continued:
∫back
face
A ⋅ ds = A back ⋅ ∆s back = A back ⋅ (−a x ∆y∆z )
face face face

∆x
= − Ax ( x0 − , y0 , z0 )∆y∆z
2
∆x ∆x ∂Ax
Ax ( x0 − , y0 , z0 ) = Ax ( x0 , y0 , z0 ) + (− ) + H.O.T.
2 2 ∂x ( x0 , y 0 , z 0 )

 ∂A 
→ [ ∫front + ∫back ] A ⋅ ds =  x + H.O.T. 
D. K. Cheng, Field and Wave Electromagnetics, 2nd ed., Addison-Wesley, 1989.

∆x∆y∆z
face face  ∂x  ( x0 , y 0 , z 0 )
Similarly:
 ∂Ay 
→ [ ∫right + ∫left ] A ⋅ ds =  + H.O.T.  ∆x∆y∆z
face face  ∂y  ( x0 , y 0 , z 0 )

 ∂A 
→ [ ∫top + ∫botton ] A ⋅ ds =  z + H.O.T.  ∆x∆y∆z
face face  ∂z  ( x0 , y 0 , z 0 )
In result:
 ∂Ax ∂Ay ∂Az 
∫SA ⋅ ds =  ∂x + ∂y + ∂z  ∆x∆y∆z + H.O.T.
7
( x0 , y 0 , z 0 )
Divergence of a Vector Field (4)
Recall:
div A ≡ lim
∫ A ⋅ ds
S
∆v →0 ∆v
 ∂Ax ∂Ay ∂Az 
∫SA ⋅ ds =  ∂x + ∂y + ∂z  ∆x∆y∆z + H.O.T.
div A = lim
( x0 , y 0 , z 0 )
∆v →0 ∆v
∂Ax ∂Ay ∂Az ∂ ∂ ∂
div A = + + ← ∇ = ax + ay + ay
∂x ∂y ∂z ∂x ∂y ∂z

∇ ⋅ A ≡ div A

In general orthogonal curvilinear coordinates:

1  ∂ ∂ ∂ 
∇⋅A ≡  (h2 h3 A1 ) + (h1h3 A2 ) + (h1h2 A3 )
h1h2 h3  ∂u1 ∂u2 ∂u3 
→ HW 8
Divergence Theorem
The volume integral of the divergence of a vector field equals the total
outward flux of the vector through the surface that bounds the volume:

∫ ∇ ⋅ Adv = ∫ A ⋅ ds
V S
← A: Smooth or piecewise smooth!

Divide the volume into small differential volume elements:

(∇ ⋅ A) j ∆v j = ∫ A ⋅ ds ← By definition
Sj

N 
∫V∇ ⋅ Adv = ∆lim
v j →0
∑ (∇ ⋅ A ) j ∆v j 
 j =1 
N 
= lim ∑ ∫ A ⋅ ds 
∆v j →0
 j =1 j 
S
Carl Friedrich Gauss

= ∫ A ⋅ ds
D. K. Cheng, Field and Wave Electromagnetics, 2nd ed., Addison-Wesley, 1989.
(1777 - 1855)
S
What if not continuously smooth?
A volume integral of the divergence of a vector can be converted into
a closed surface integral of the vector, and vice versa!
9