Lecture 3

Vector Differential
Calculus
Grad, Div, Curl
The central tool in the discussion in this lecture is the ‘del’ operator; also called the ‘nabla’
operator, defined as
𝜕 𝜕 𝜕
∇= 𝐢 +𝐣 + 𝐢 (1)
𝜕𝑥 𝜕𝑦 𝜕𝑧

1. Gradient of a Scalar field

Imagine we have a scalar field – let us say the temperature in some region in a Cartesian
coordinate. We ask the question ‘What is the change in temperature?’. Immediately we would
enquire ‘In which direction?’. This means, even if the temperature is a scalar field, the change in
temperature would be a vector quantity. This vector quantity is denoted by the gradient.

Formally, we are given a scalar function f(x, y, z) that is defined and differentiable in a domain in
3D-space with Cartesian coordinates x, y, z. We denote the gradient of that function by grad f or
f, and is defined as
𝜕𝑓 𝜕𝑓 𝜕𝑓
grad 𝑓 = ∇𝑓 = 𝐢 +𝐣 + 𝐢 (2)
𝜕𝑥 𝜕𝑦 𝜕𝑧
Directional derivative

To understand the physical meaning of the gradient of a scalar function, we define a term known
as directional derivative.

Directional derivative is defined as the rate of

C change of a scalar function f(x,y,z) in a direction
given by a unit vector b. An example could be that
s a particle is moving in a charged field in a
Q
b particular direction b. We are interested to find the
rate of change of the charge that will be
encountered by the particle.
P

Figure 01. Directional derivative Referring to Figure 01, let the particle be moving
along the curve C. Let b be a vector along the
curve C. Further, let s be the distance along the curve C. Since f is a function of x, y and z, the
rate of change of f(x, y, z) along s is given by
𝑑𝑓 𝜕𝑓 𝜕𝑥 𝜕𝑓 𝜕𝑦 𝜕𝑓 𝜕𝑧
= + + (3)
𝑑𝑠 𝜕𝑥 𝜕𝑠 𝜕𝑦 𝜕𝑠 𝜕𝑧 𝜕𝑠
The position vector of any point on the curve C is given by
r(s) = x(s) i + y(s) j + z(s) k = P + sb (4)
Differentiating this, we obtain
𝑑𝑟 𝑑𝑥 𝑑𝑦 𝑑𝑥
= 𝐢+ 𝐢+ 𝐤=𝐛 (5)
𝑑𝑠 𝑑𝑠 𝑑𝑠 𝑑𝑠
Finally, df/ds from equation (3) can be written in terms of dot product as
 𝑑𝑓 𝜕𝑓 𝜕𝑓 𝜕𝑓 𝜕𝑥 𝜕𝑥 𝜕𝑥
= 𝐢+ 𝐣+ 𝐤 ∙ 𝐢+ 𝐣+ 𝐤 (6)
𝑑𝑠 𝑑𝑥 𝑑𝑥 𝑑𝑥 𝜕𝑠 𝜕𝑠 𝜕𝑠
We recognize the quantity in the first parenthesis as grad f, and from (5), quantity in the second
parenthesis as the vector b. Denoting the directional derivative as Dbf, (6) can be written as
𝐷 𝑓 = 𝐠𝐫𝐚𝐝 𝑓 ∙ 𝐛 (7)
Therefore, using the gradient of a scalar function we can determine the rate of change of the
scalar function in a particular direction.

Gradient characterizes maximum increase

From the definition of the directional derivative, Dbf = grad f  b. Using the definition of the dot
product, we can write.
Dbf = grad f  b = ||grad f|| ||b|| cos 
where  is the angle between grad f and b. Since b is a unit vector, ||b|| = 1. Therefore,
Dbf = ||grad f|| cos 
Now, Dbf would be maximum when cos = 1. Therefore, (Dbf)max = ||grad f||. This gives us an
interpretation of the gradient of a scalar field. The norm of the gradient vector gives the
maximum rate of change of the scalar field at a particular point.

Gradient as surface normal

Once again let f(x,y,z) be a scalar field. The expression f(x,y,z) = c, where c is a constant, will
denote a surface within the scalar field. Any point on the surface will have a position vector
r(s) = x(s) i + y(s) j + z(s) k.
The first derivative of r is
r′(s) = x′(s) i + y′(s) j + z′(s) k
The rate of change of f is written as
𝑑𝑓 𝜕𝑓 𝜕𝑥 𝜕𝑓 𝜕𝑦 𝜕𝑓 𝜕𝑧
= + + = 𝐠𝐫𝐚𝐝 𝑓 ∙ 𝐫′
𝑑𝑠 𝜕𝑥 𝜕𝑠 𝜕𝑦 𝜕𝑠 𝜕𝑧 𝜕𝑠

But as f(x,y,z) is a constant surface, df/ds = 0. Which means that (grad f) and r′ are perpendicular
to each other. We know that r′ is a tangent to the surface f(x,y,z) = c. Therefore, (grad f) is
normal to the surface.

Therefore, the gradient of a scalar field at a point gives us the maximum rate of change
of the scalar field at that point, and the direction of the maximum rate of change is normal to the
surface.

Example 01
Given f(x, y, z) = 3x2 –y2z2, find f at the point (1, –2, –1)
𝜕 𝜕 𝜕
∇𝑓 = 𝐢+ 𝐣+ 𝐤 (3𝑥 𝑦 − 𝑦 𝑧 )
𝜕𝑥 𝜕𝑥 𝜕𝑥
𝜕 𝜕 𝜕
= 𝐢 (3𝑥 𝑦 − 𝑦 𝑧 ) + 𝐣 (3𝑥 𝑦 − 𝑦 𝑧 ) + 𝐤 (3𝑥 𝑦 − 𝑦 𝑧 )
𝜕𝑥 𝜕𝑦 𝜕𝑧
= 6xy i + (3x2 – 3y2z2) j – y3z2 k
= 6(1)(–2) i + [3(1)2 – 3(–2)2(–1)2] j –2(–2)2(–1) k
= 12 i – 9 j – 16 k

Example 02
1
Find ∇𝑓 if (𝑎) 𝑓 = ln‖𝐫‖ , (𝑏) 𝑓 =
𝑟
(a) r(t) = xi + yj + zk. The length 𝑟 = ‖𝐫‖ = 𝑥 +𝑦 +𝑧 .
1
𝑓 = ln‖𝐫‖ = ln(𝑥 + 𝑦 + 𝑧 )
2
1 𝜕 𝜕 𝜕
∇𝑓 = 𝐢 ln(𝑥 + 𝑦 + 𝑧 ) + 𝐣 ln(𝑥 + 𝑦 + 𝑧 ) + 𝐤 ln(𝑥 + 𝑦 + 𝑧 )
2 𝜕𝑥 𝜕𝑥 𝜕𝑥
1 2𝑥 2𝑦 2𝑧
= 𝐢 +𝐣 +𝐤
2 𝑥 +𝑦 +𝑧 𝑥 +𝑦 +𝑧 𝑥 +𝑦 +𝑧
𝑥𝐢 + 𝑦𝐣 + 𝑧𝐤 𝐫
= =
𝑥 +𝑦 +𝑧 𝑟

1 1 /
(b) ∇𝑓 = ∇ =∇ = ∇ (𝑥 + 𝑦 + 𝑧 )
𝑟 𝑥 +𝑦 +𝑧
1 2𝑥 1 2𝑥 1 2𝑥
=𝐢 − +𝐣 − +𝐤 −
2 2 2
(𝑥 + 𝑦 + 𝑧 ) (𝑥 + 𝑦 + 𝑧 ) (𝑥 + 𝑦 + 𝑧 )
−𝑥𝐢 − 𝑦𝐣 − 𝑧𝐤 𝐫
= =−
𝑟
(𝑥 + 𝑦 + 𝑧 )

Example 03
Find a unit normal to the surface –x2yz2 + 2xy2z = 1 at the point P(1, 1, 1).

We have learnt in our discussion of the gradient, that the gradient to a surface at any point is
normal to the surface.

Considering f = –x2yz2 + 2xy2z, f = (–2xyz2)i + (–x2z2 + 4xyz)j + (–2x2yz + 2xy2)k

At P(1, 1, 1), f(1, 1, 1) = 3j, ||f(1, 1, 1)|| = ||3j|| = 3.
Therefore, at point P(1, 1, 1), the unit normal is j.

Example 04
Find an equation of the tangent plane to the surface f = x2yz – 4xyz2 = –6 at the point P(1, 2, 1).

We find the normal to the surface at the given point

f = (2xyz – 4yz2)i + (x2z –4xz2)j + (x2y – 8xyz)k
The normal at P(1, 2, 1) is –4i –3j –14k.
From our discussion in the last lecture, a surface perpendicular to this normal will have an
equation 4x + 3y + 14z = k. Substituting the point P into this equation, we get the value of k = 24.
Therefore, the equation of the plane is 4x + 3y + 14z = 24.

Example 05
Let f = x2yz – 4xyz2. Find the directional derivative of f at P(1, 3, 1) in the direction 2i – j – 2k.

First we find
f = (2xyz – 4yz2)i + (x2z – 4xz2)j + (x2y – 8xyz)k .
Then
f(1, 3, 1) = –6i – 3j – 21k.
The unit vector in the direction 2i – j – 2k is
2𝐢 − 𝐣 − 2𝐤 2 1 2
𝐚= = 𝐢− 𝐣− 𝐤
(2) + (−1) + (−2) 3 3 3
The required directional derivative is
2 1 2
∇𝑓(1, 3, 1) ∙ 𝐚 = (– 6𝐢 – 3𝐣 – 21𝐤) ∙ 𝐢 − 𝐣 − 𝐤 = −4 + 1 + 14 = 11
3 3 3

Exercise

1. Consider f = 2x4 – x2y. Find f and ||f|| at point (2, –2, –1)
2. Suppose a vector A = 2x2i – 3yzj + xz2k and f = 2z – x2y. Find A  f and A  f at point
(1, –1, 1).
3. Find ∇𝑓 where 𝑓 = (𝑥 + 𝑦 + 𝑧 )𝑒
4. Find a vector that is perpendicular to the surface of the paraboloid of revolution
z = x2 + y2 at the point (1, 2, 5).
5. Find the unit outward normal to the surface (x – 1)2 + y2 + (z + 2)2 = 9 at the point (3, 1,
4).
6. Find an equation of the tangent plane to the surface xz2 + x2y = z – 1 at the point (1, –3,
2).
 7. Find an equation of the normal plane and an equation of normal line to the surface
z = x2 + y2 at the point (2, –1, 5).
8. Find the directional derivative of f = 4xz3 – 3x2y2z at (2, –1, 2) in the direction 2i – 3j +
6k.
9. Find the directional derivative of P = 4e2x – y + z at the point (1, 1, –1) in a direction
toward the point (–3, 5, 6).
10. For two functions u and v to be functionally related u  v = 0. Determine whether
u = tan–1x + tan–1y and v = (x + y)/(1 – xy) are functionally related.

2. Divergence of a Vector field

Let v(x, y, z) be a differentiable vector function, where x, y, z are Cartesian coordinates, and let
v1, v2, v3 be the components of v. The divergence of the vector function is obtained by taking the
dot product of the del operator, , with the vector v. The divergence of the v is indicated by
  v, or div v.
𝜕 𝜕 𝜕
div 𝐯 = ∇ ∙ 𝐯 = 𝐢 +𝐣 + 𝐢 ∙ (𝑣 𝐢 + 𝑣 𝐣 + 𝑣 𝐤)
𝜕𝑥 𝜕𝑦 𝜕𝑧

𝜕𝑣 𝜕𝑣 𝜕𝑣
= + + (8)
𝜕𝑥 𝜕𝑦 𝜕𝑧

Example 06
For v = 3xz i + 2xy j – yz2 k, div v = 3z + 2x – 2yz

Physical significance of the divergence

Divergence gives us a method to check for the conservation of any quantity that flows. These
include heat, fluid, charge, momentum. To understand this, let us consider an element of size x,
y, z, as shown in Figure 02. Let the element be experiencing a flow of a quantity with a flux v
given by the vector v = v1i + v2 j + v3k. We are also assuming that the element is neither
consuming, nor generating the quantity that is flowing. For example, if the quantity of flow is
heat, the element is neither generating, nor consuming heat.

We will start the derivation of the governing equation by performing a conservation of the
quantity that is flowing. To do this, we ascertain the flow of quantity through all the faces of the
block.

In x direction, the flux at entry is v1. Therefore, the amount of quantity entering through that face
is (flux  area)
𝑣 = 𝑣 ∆𝑦∆𝑧 (9.1)
 𝜕𝑣
𝑣 + ∆𝑧
𝜕𝑧
v2 ∆𝑥 v1

z ∆𝑧 𝜕𝑣
𝑣 + ∆𝑦
𝜕𝑦

∆𝑦
𝜕𝑣
𝑣 + ∆𝑥
𝜕𝑥
v3
y
x

Figure 02. Analysis of flow through a rectangular element x, y, z

To ascertain the flux at exit, we expand v1 at entry in Taylor series. Therefore
𝜕𝑣 𝜕 𝑣 (∆𝑥)
𝑣 =𝑣 + ∆𝑥 + +∙∙∙ (9.2)
𝜕𝑥 𝜕𝑥 2!
We will ignore the second and higher-order terms. Therefore, the net amount of flow through the
block in x-direction can be obtained by the difference of (9.2) and (9.1). We obtain
𝜕𝑣
𝑣 = 𝑣 + ∆𝑥 ∆𝑦∆𝑧 (10.1)
𝜕𝑥
Similarly, net flux transfer through y and z directions are
𝜕𝑣
𝑣 = 𝑣 + ∆𝑦 ∆𝑥∆𝑧 (10.2)
𝜕𝑦
𝜕𝑣
𝑣 = 𝑣 + ∆𝑧 ∆𝑥∆𝑦 (10.3)
𝜕𝑧
Total net transfer can be obtained by adding (10.1), (10.2) and (10.3).
𝜕𝑣 𝜕𝑣 𝜕𝑣
𝑣 = + + ∆𝑥∆𝑦∆𝑧 (11)
𝜕𝑥 𝜕𝑦 𝜕𝑧
Since there is no net consumption or generation of flux within the element, the net flux through
the element would be zero. This gives us
𝜕𝑣 𝜕𝑣 𝜕𝑣
+ + =0
𝜕𝑥 𝜕𝑦 𝜕𝑧
or div v = 0
Therefore, we observe that the divergence of a vector quantity gives us a way to ascertain the
conservation of a flow quantity in a conservative field. (Conservative field is where there is no
loss or gain of the quantity that flows).
Example 07
Suppose A = x2z2i – 2y2z2j + xy2zk. Find div A at P(1, –1, 1).

𝜕 𝜕 𝜕
div 𝐀 = 𝑥 𝑧 + (−2𝑦 𝑧 ) + 𝑥𝑦 𝑧 = 2𝑥𝑧 − 4𝑦𝑧 + 𝑥𝑦
𝜕𝑥 𝜕𝑦 𝜕𝑧
At P(1, –1, 1) div A = 2(1)(1)2 – 4(–1)(1)2 + (1)(–1)2 = 7

Example 08
Suppose f = 6x3y2z. Find  f

First f = 18x2y2z i + 12x3yz j + 6x3y2 k

Then  f = 36xy2z + 24x3z

 f is also known as 2f – Laplacian of f.

Exercise

1. Let A = 3xyz2 i + 2xy3 j – x2yz k and f = 3x2 – yz. At the point P(1, –1, 1), find (a)  A,
(b) A f, (c)  (fA), (d)  (f).
2. Evaluate div(2x2z i – xy2z j + 3yz2 k).
3. For v = x2 i + 4y2 j + 9z2 k, find div v at P(–1, 0, ½)
4. For v = cos xyz j + sin xyz k, find div v at P(2, /2, 0)
5. For v = x2y2z2(x i + y j + z k) find div v at P(3, –1, 4)
6. For v = (x2 + y2 + z2)–3/2(x i + y j + z k) find div v.
7. For f = z – (x2 + y2)–1/2, find 2f.
8. For f = 1/( x2 + y2 + z2), find 2f.
9. Find 2(ln r).
10. Let F = (3x2y – z) i + (xz3 + y4) j – 2x3z2 k. Find ( F) at the point (2, –1, 0).

3. Curl of a vector field

Let v = v1 i + v2 j + v3 k be a differentiable vector function of the Cartesian coordinates x, y, z.

Then the curl of the vector function v or the vector field given by v is defined by the ‘del’
operator operating on the vector v with a cross-product. Earlier we saw that the cross-product can
be indicated by a determinant. Therefore, the curl of a vector can be symbolically written by the
determinant
 𝐢 𝐣 𝐤
𝜕 𝜕 𝜕
𝐜𝐮𝐫𝐥 𝐯 = ∇ × 𝐯 = (12.1)
𝜕𝑥 𝜕𝑦 𝜕𝑧
𝑣 𝑣 𝑣
𝜕𝑣 𝜕𝑣 𝜕𝑣 𝜕𝑣 𝜕𝑣 𝜕𝑣
= − 𝐢+ − 𝐣+ − 𝐤 (12.2)
𝜕𝑦 𝜕𝑧 𝜕𝑧 𝜕𝑥 𝜕𝑥 𝜕𝑦

Example 09

Let A = yz i + 3zx j + z k. Find curl A.

We can use the relations (12.1) or (12.2).

𝐢 𝐣 𝐤
𝜕 𝜕 𝜕
𝐜𝐮𝐫𝐥 𝐀 = = −3𝑥𝐢 + 𝑦𝐣 + (3𝑧 − 𝑧)𝐤 = – 3𝑥𝐢 + 𝑦𝐣 + 2𝑧𝒌
𝜕𝑥 𝜕𝑦 𝜕𝑧
𝑦𝑧 𝑧𝑥 𝑧

The physical significance of the curl of a vector is not as obvious as grad and div. But curl
appears in rotation of systems, as in fluid, turbines, compressors; and rotation of system in a
rotating frame, as in cyclones.

Exercise

1. Let A = 2xz2i – yzj + 3xz2k and f = x2yz. Find (a)   A, (b)   (fA), (c)   (  A),
(d) (A    A) (e)   (fA)
2. Let F = x2yz, G = xy – 3z2. Find (a) [(F)  (G)] (b)  [(F)  (G)]
(c)   [(F)  (G)]
3. Given A = x2zi + yz3j – 3xyk, B = y2i – yzj + 2xk and f = 2x2 + yz. Find (a) A  (f),
(b) (A  )f, (c) (A  )B, (d) B(A  ), (e) (  A)B, (f) A  (  B), (g) (A  )  B
4. Prove (v  )v = ½v2 – v  (  v)
5. Prove   (A  B) = B  (  A) – A  (  B)
6. Prove   (A  B) = (B  )A – B(  A) – (A  )B + A(  B)