Ch 10 Functions in 3D

Volume Integrals

A volume integral is an integral in which the region of integration is three dimensional; i.e. there are
three independent integration variables. It is represented by the compact notation:

𝐼 = ∫! 𝑓𝑑𝑣 Eq 1

where V is the volume of integration, f is the function that is integrated, and dv is called the volume
element. As in the double integral, the form of dv depends on the type of variables used in evaluating
I. When cartesian variables are employed, dv = dxdydz.

Evaluation of the triple integral in Eq 1 is simple when the following two conditions are fulfilled.

(i) Volume of integration V has the shape of a rectangular box; i.e. the lower and upper limits
of all three integration variables are constants.

(ii) The function f has a product form: f(x,y,z) = p(x)q(y)r(z), or when f consists of a sum of
terms each of which has a product form.

e.g.1 Evaluate the integral

𝐼 = ∫! 𝑥𝑦 " 𝑧 # 𝑑𝑥𝑑𝑦𝑑𝑧 , 𝑉: the box 0£ 𝑥 £ 𝑎, 0£ 𝑦 £ 𝑏, 0£ z £ c

where a, b, c are constants (you should regard these constants as given quantities).

This integral satisfies the two conditions above, and leads to a factorization of I.

𝐼 = 𝐼$ 𝐼" 𝐼#

where
% ' )
𝑎" "
𝑏# 𝑐(
𝐼$ = 9 𝑥𝑑𝑥 = , 𝐼" = 9 𝑦 𝑑𝑦 = , 𝐼# = 9 𝑧 # 𝑑𝑥 =
2 3 4
& & &

Hence,

𝑎" 𝑏 # 𝑐 ( 𝑎" 𝑏 # 𝑐 (
𝐼 = > ?> ?> ? =
2 3 4 24

1
The volume of integration may be infinite as in the next example.

e.g.2 Evaluate

𝐼 = 9 𝑒 *(,-#.-/0) 𝑑𝑥𝑑𝑦𝑑𝑧, 𝑉: the infinite region 𝑥 ≥ 0, 𝑦 ≥ 0, 𝑧 ≥ 0

!

The lower limit of each of the three variables is zero, and their upper limits are infinity which
may be regarded as a constant. Thus V can be viewed as a rectangular box with infinite
volume. Furthermore, the function that is integrated has a product form:
𝑒 *(,-#.-/0) = 𝑒 *, 𝑒 *#. 𝑒 */0 . Thus, the two simplicity conditions above are satisfied, and the
integral I factorizes, as in e.g.1.

𝐼 = 𝐼$ 𝐼" 𝐼#

where
2 2 2
*, *#.
1 1
𝐼$ = 9 𝑒 𝑑𝑥 = 1, 𝐼" = 9 𝑒 𝑑𝑥 = , 𝐼# = 9 𝑒 */0 𝑑𝑥 =
3 5
& & &

Hence,

1 1 1
𝐼 = (1) K L K L = = 0.0667
3 5 15

e.g.3 Evaluate

𝐼 = 9 (𝑥 " + 𝑦 " + 𝑧 " ) 𝑑𝑥𝑑𝑦𝑑𝑧, 𝑉: the sphere of radius 𝑎, center at the origin
!

The function that is integrated has the desired form: it is a sum of three terms each of which
has a product form in cartesian variables. However, the shape of V is not that of a rectangular
box with the cartesian variables; i.e. the lower and upper limits of the three integration
variables x, y, and z can not all be taken as constants. Thus the simplicity conditions are not
both satisfied leading us to conclude that the cartesian variables are not suitable for
evaluating this integral. We need to find new variables with which the simplicity conditions
are fulfilled. In doing this, it should be kept in mind that the number of variables can not
change, it must be 3; only the type of variables may be changed.

2
 The spherical boundary surface of V in this example suggests that one of the new variables
may be taken as the radial distance, r, of a point from the origin. In the spherical V, this
variable will change from 0 to a so that both the lower and the upper limits of r are
constants, as required by the simplicity conditions. We need to define two more variables.
They are taken as two angular variables. The new set of variables is an extension of the 2D
polar variables to 3D.

Spherical polar coordinates

The location of a point in 3D space may be specified either by giving its cartesian coordinates (x,y,z),
or by providing the values of the new variables (r,θ,φ). The relationship among the two sets of
variables is depicted in Figure 1a.

z
z
r sin q
P(x,y,z)
r cos q r
z q
q
0 y
f x 0 y

x y
x
(a) (b)

Figure 1

The coordinate surfaces in the new variables are families of: (1) concentric spheres about the origin
(r = const.); (2) right circular cones with apex at the origin and axis along z (θ = const.), as shown in
Figure 1b; and (3) half-planes from the z axis (φ = const.)

The two sets of variables are related by the equations (see Figure 1a)

𝑥 = 𝑟 sin θ cos φ

𝑦 = 𝑟 sin θ sin φ

𝑧 = 𝑟 cos θ Eq 2a

3
The inverse relations are

𝑟 = X𝑥 " + 𝑦 " + 𝑧 "

𝑧
θ = cos*$ > ?
X𝑥 " + 𝑦 " + 𝑧 "

.
𝑡𝑎𝑛*$ ], ^ if 𝑥 > 0
ϕ=Z Eq 2b
.
𝑡𝑎𝑛*$ ] ^ + 𝜋 if 𝑥 < 0
,

Range of spherical polar variables for a sphere of radius a

As stated before, range of r is from 0 to a. The angle φ is measured from the positive x axis. It
corresponds to anticlockwise rotation about the z axis of a point in the xy-plane by an angle φ. Full
range of φ is from 0 to 2π, and it means a complete rotation about the z axis.

The angle θ is measured from the positive z axis. As shown in Figure 1b, all points that are on the
lateral surface of the cone have the same θ value. Value of θ is zero on the positive z axis where the
cone degenerates into a line. Now start increasing θ. For θ=90⁰, the cone becomes a disc (of radius a)
in the xy plane on which every point has θ=π/2. Increasing θ to π, the cone once again degenerates
into a line, but this time along the negative z axis. All points on the negative z axis have θ=π.
Increasing θ value to larger than π does not give different cones from the ones before. Therefore,
the full range of θ is from 0 to π.

In summary, every point inside a sphere of radius a can be traversed without repetition by varying r,
θ, and φ independently, by the following amounts.

𝑟: 0 to 𝑎

θ: 0 to π

φ: 0 to 2π Eq 3

Returning now to the integral in e.g.3, we see that in terms of the spherical polar variables, the
volume of integration will have a rectangular box shape in the sense that all three variables r, θ, and
φ will have lower and upper limits all of which are constants. Thus the first condition for the
simplicity of a multiple integral is satisfied. The second condition is also satisfied. The function that is

4
integrated in e.g.3, f(x,y,z) = x2+y2+z2 becomes f = r2 in the new variables (see Eq 2b), and satisfies the
condition of having a product form (in the new variables). What remains is the appropriate
expression for the volume element, dv, in terms of r, θ, and φ. As will be explained below, the
required expression is

dv = r2 sin(θ) dr dθ dφ Eq 4

The jacobian of the transformation is: J = r2 sin(θ). As pointed out in the previous chapter, J must be
positive at all points in V, and this is ensured by the fact that for 0 < θ < π , sin(θ) is positive.

In terms of the spherical polar variables, our integral factorizes

𝐼 = 9 𝑟 ( sin( θ) 𝑑𝑟𝑑θ𝑑ϕ = 𝐼$ 𝐼" 𝐼#

!

where
% 3 "3
𝑎/ (
𝐼$ = 9 𝑟 𝑑𝑟 = , 𝐼" = 9 sin 𝜃 𝑑𝜃 = 2, 𝐼# = 9 𝑑𝜙 = 2𝜋
5
& & &

giving

𝑎/ 4𝜋𝑎/
𝐼 = > ? (2)(2𝜋) =
5 5

for the value of the integral in e.g.3.

e.g.4 Evaluate

$ ! -. ! -0 !
𝐼 = 9 𝑒 *#4, 𝑥 " 𝑑𝑣, 𝑉: all space
!

In cartesian variables, the integrand does not have a product form due to the square root in
the argument of the exponential function. Reference to Eq 2a,b shows that there will be no
square root problem if we use spherical polar variables instead. The integrand then becomes
$ ! -. ! -0 !
𝑒 *#4, 𝑥 " = 𝑒 *5/# 𝑟 " 𝑠𝑖𝑛" θ 𝑐𝑜𝑠 " ϕ

5
and it has the desired product form in (r, θ,φ) variables. Volume of integration is "all space",
which may be regarded as a sphere with an infinite radius; i.e. in Eq 3, we set a = ¥. Inserting
the volume element from Eq 4, the integral I becomes

𝐼 = 9 𝑒 *5/# 𝑟 " 𝑠𝑖𝑛" θ 𝑐𝑜𝑠 " ϕ 𝑟 " 𝑠𝑖𝑛 θ d𝑟dθdϕ = 𝐼$ 𝐼" 𝐼#

!

where
"3

𝐼$ = 9 𝑐𝑜𝑠 " ϕ dϕ = π
&

3
4
𝐼" = 9 𝑠𝑖𝑛# θ dθ =
3
&

𝐼# = 9 𝑒 *5/# 𝑟 ( 𝑑𝑟 = 3/ 4! = 5832
&

In evaluating the last integral we used a result from integral tables.

2 7!
∫& 𝑒 *', 𝑥 7 𝑑𝑥 = '"#$ , 𝑏 > 0, 𝑛 = 0,1,2, … Eq 5

The value of integral I is thus

4
𝐼 = (π) K L (5832) = 7776π = 24429.0
3

6
e.g.5 Describe the volume, V, of 1/8 of a sphere of radius a in which y³0, z³0, but x£0 by stating
the lower and upper limits of the spherical polar variables r, θ, and φ.

Since V is part of the sphere, range of r does not change from that for the full sphere; it
ranges from 0 to a. Turning to the angles, we see from Figure 1a that in order to ensure z³0,
the value of θ should not exceed π/2; thus the range of θ is from 0 to π/2. For the φ angle,
Figure 1a shows that for φ between π/2 and 3π/2, the cartesian variable x has a negative
value. Hence the lower limit of φ is π/2. For the upper limit of φ we must make sure that the
variable y should not become negative which occurs when φ exceeds π. Thus, the range of φ
is from π/2 to π. In summary,

r: 0 to a

θ: 0 to π/2

φ: π/2 to π

The Laplacian operator in orthogonal coordinate systems

An orthogonal coordinate system is one in which the coordinate lines intersect at right angles to one
another. The standard cartesian coordinate system (x,y,z) is an example. The coordinate lines in
general orthogonal systems are not necessarily straight lines. For example, the coordinate lines in the
2D or 3D polar coordinate systems intersect at right angles even though not all are straight lines.

The Laplacian operator in 3D cartesian variables is defined as an extension to the 2D definition by

adding the contribution from the z variable.

9! 9! 9!
∇" = 9, ! + 9. ! + 90 ! Eq 6

Note that there are no mixed second derivatives in this expression.

7
e.g. 6 Simplify ∇" 𝑓 for f = x2y3z4

𝜕" " # ( 𝜕" 𝜕" " # (

∇" (𝑥 " 𝑦 # 𝑧 ( ) = (𝑥 𝑦 𝑧 ) + (𝑥 " # ()
𝑦 𝑧 + (𝑥 𝑦 𝑧 )
𝜕𝑥 " 𝜕𝑦 " 𝜕𝑧 "

= 2𝑦 # 𝑧 ( + 6𝑥 " 𝑦𝑧 ( + 12𝑥 " 𝑦 # 𝑧 "

Finding the form of the Laplacian operator in other types of variables is relatively simple when the
new variables form an orthogonal set. Let the cartesian variables x, y, and z be related to three new
variables by the equations

𝑥 = 𝑥(𝑞$ , 𝑞" , 𝑞# ), 𝑦 = 𝑦(𝑞$ , 𝑞" , 𝑞# ), 𝑧 = 𝑧(𝑞$ , 𝑞" , 𝑞# ) Eq 7

It will be assumed that the variables q1, q2, and q3 form an orthogonal coordinate system. An
example is the set of spherical polar variables r, θ, φ defined in Eq 2a. The square of the distance
between two adjacent points is

𝑑𝑠 " = 𝑑𝑥 " + 𝑑𝑦 " + 𝑑𝑧 " = ℎ$" 𝑑𝑞$" + ℎ"" 𝑑𝑞"" + ℎ#" 𝑑𝑞#"

where

9, " 9. " 90 "

ℎ:" = ] ^ +] ^ +] ^ , 𝑖 = 1,2,3 Eq 8
9;% 9;% 9;%

e.g. 7 Find the quantities hr, hθ , and hφ for the spherical polar variables Eq 2a.
Associating q1, q2, and q3 with r, θ, and φ; one obtains, using Eq 8:

hr = 1, hθ = r, and hφ = r sinθ

The utilty of these quantities lies in the following expressions for the volume element and the
Laplacian operator in terms of the new variables.

8
The volume element in orthogonal coordinates

𝑑𝑣 = 𝐽 𝑑𝑞$ 𝑑𝑞" 𝑑𝑞# , where 𝐽 = ℎ$ ℎ" ℎ# Eq 9

Thus the jacobian of the transformation is easily obtained. For the spherical polar variables,

𝐽 = ℎ5 ℎ< ℎ= = (1)(𝑟)(𝑟 sin θ) = 𝑟 " sin θ Eq 10

It is this expression for the jacobian that was used in Eq 4 above.

The Laplacian operator in orthogonal coordinates

$ 9 ?! ?& 9 9 ?$ ?& 9 9 ?$ ?! 9
∇" = s ] ^+ ] ^+ ] ^ t Eq 11
> 9;$ ?$ 9;$ 9;! ?! 9;! 9;& ?& 9;&

Let us compare this expression with the simpler cartesian one in Eq 6. Eq 11 consists of a sum of
three terms such that only one variable is involved in each term; i.e. there are no mixed second
derivatives. This feature is the same as in the cartesian expression. The main difference from the
cartesian expression is the presence of h1, h2, and h3 between the first and the second derivatives of
each term; these are all equal to 1 in the cartesian case. Note that in the first term involving variable
q1, h1 is in the denominator with h2 and h3 being in the numerator. The second and the third terms
also have similar properties. In each term, the index of h in the denominator is the same as that of
the variable while the remaining two h quantities occur in the numerator.

Since there are no mixed second derivatives in Eq 11, for a function that has a product form:
𝑓(𝑞$ , 𝑞" , 𝑞# ) = 𝑅(𝑞$ )𝑇(𝑞" )𝐹(𝑞# ), ∇" 𝑓 satisfies the useful relation

∇" (𝑅𝑇𝐹) = 𝑇𝐹 ∇" 𝑅 + 𝑅𝐹 ∇" 𝑇 + 𝑅𝑇 ∇" 𝐹 Eq 12

This relation helps in simplifying ∇" 𝑓 especially when the Laplacian operator is given in noncartesian
variables. An example is given below in e.g.9.

9
e.g. 8 Write the Laplacian operator in spherical polar variables.
We have: hr = 1, hθ = r, and hφ = r sinθ. Also, 𝐽 = 𝑟 " sin θ. Then from Eq 11

$ 9 5 ! @AB C 9 9 5 @AB C 9 9 5 9
∇" = 5 ! @AB C s ] ^ + 9C ] ^ + 9D ]5 @AB C 9D^ t
95 $ 95 5 9C

Simplifying yields

$ 9 9 $ 9 9 $ 9!
∇" = 5 ! 95 ]𝑟 " 95^ + 5 ! @AB C 9C ]sin 𝜃 9C^ + 5 ! @AB! C 9D! Eq 13a

An equivalent expression may be obtained by expanding the first term on the right. For an arbitrary
function f(r, θ, φ), we have

$ 9 9 $ 9 9E " 9E 9! E " 9 9!
5! 95
]𝑟 " 95^ 𝑓 = 5 ! 95 ]𝑟 " 95 ^ = 5 95 + 95 ! = ]5 95 + 95 ! ^ 𝑓

Hence, Eq 13a may be written in the alternative form

9! " 9 $ 9 9 $ 9!
∇" = 95 ! + 5 95 + 5 ! @AB C 9C ]sin 𝜃 9C ^ + 5 ! @AB! C 9D! Eq 13b

Eq 9 for the volume element, and Eq 11 for the Laplacian can also be applied to the 2D case where
there are two variables. One discards the term involving the derivative of z from Eq 8, discards h3 and
dq3 from Eq 9, and drops h3 and the third term from Eq 11. For example, for the 2D polar variables

x = r cos θ, y = r sin θ

𝜕𝑥 " 𝜕𝑦 "
ℎ5" = K L + K L = 1, ℎ5 = 1
𝜕𝑟 𝜕𝑟

𝜕𝑥 " 𝜕𝑦 "
ℎ<" = K L + K L = 𝑟 " , ℎ< = 𝑟
𝜕θ 𝜕θ

Hence, J = (1)(r) = r, so that dA = r drdθ. Similary, the expression of the 2D Laplacian in these
coordinates is from Eq 11 (dropping the third term and h3)

10
 $ 9 9 9 $ 9
∇" = s ]𝑟 ^+ ] ^t
5 95 95 9C 5 9C

$ 9 9 $ 9!
= ]𝑟 ^+ 2D Eq 14a
5 95 95 5 ! 9C !

Similar to going from Eq 13a to Eq 13b, the first term on the left can be expanded to give an
equivalent expression

9! $ 9 $ 9!
∇" = 95 ! + 5 95 + 5 ! 9C! 2D Eq 14b

This expression for the Laplacian operator in 2D polar variables was obtained in the previous chapter
by another procedure.

e.g. 9 Evaluate ∇" 𝑓 for 𝑓(𝑟, 𝜃, 𝜙) = 𝑟𝑒 *5/" sin 𝜃 cos 𝜙

The function given has a product form: 𝑓(𝑟, 𝜃, 𝜙) = 𝑅(𝑟)𝑇(𝜃)𝐹(𝜙) where

𝑅 = 𝑟𝑒 *5/" , 𝑇 = sin 𝜃 , 𝐹 = cos 𝜙

Hence we may use the property Eq 12 for the simplification of ∇" 𝑓

∇" (𝑅𝑇𝐹) = 𝑇𝐹 ∇" 𝑅 + 𝑅𝐹 ∇" 𝑇 + 𝑅𝑇 ∇" 𝐹 Eq 15

We have
'
F!G " FG $ " $ " "
∇" 𝑅 = F5 !
+ 5 F5 = ]( 𝑟 − 2 + 5 ^ 𝑒 *! = ]( − 5 + 5 ! ^ 𝑅 Eq 16a

$ F F $ F
∇" 𝑇 = 5 ! @AB C FC ]sin 𝜃 FC^ sin 𝜃 = 5 ! @AB C FC (sin 𝜃 cos 𝜃)

)HI ! C*I:7! C )HI ! C*I:7! C

= 5 ! @AB C
= 5 ! @AB! C
𝑇 Eq 16b

$ F! JK@ D $
∇" 𝐹 = 5 ! @AB! C FD! (cos 𝜙) = − 5 ! @AB! C = − 5 ! @AB! C 𝐹 Eq 16c

11
 Expressions Eq 16a-c are inserted into Eq 15. After simplification, one obtains, as the final
result

1 2
∇" 𝑓 = K − L 𝑓
4 𝑟

Cylindrical coordinates
The three independent variables in this system are the two plane polar variables in the xy-plane, and
the third one is the cartesian variable z (Figure 2). They form an orthogonal system, and are useful in
problems with cylindrical symmetry.

P(x,y,z)

0 y
r
f x

y
x

Figure 2

The coordinate surfaces are: (1) right circular cylinders with axis along the z axis, which form families
of concentric circles about the origin in the xy-plane ( r = const. ); (2) half-planes from the z axis
(φ=const. ); (3) planes parallel to the xy-plane ( z = const. ). A point P(x,y,z) is located by giving the
distance r from the origin to the cylinder on which the point lies, the angle φ in the xy-plane, and the
distance on the z axis from this plane to the point P. From Figure 2

12
 𝑥 = 𝜌 cos 𝜙

𝑦 = 𝜌 sin 𝜙

𝑧=𝑧 Eq 17

ℎL = 1, ℎD = 𝜌, ℎ0 = 1 Eq 18

𝑑𝑣 = 𝜌 𝑑𝜌 𝑑𝜙 𝑑𝑧 Volume element Eq 19

The expression for the Laplacian operator is obtained from Eq 14a or b (with appropriate relabeling
of variables), and adding to it the contribution from the cartesian variable z.

9! $ 9 $ 9! 9!
∇" = 9L! + L 9L + L! 9D! + 90 ! Eq 20

e.g. 10 Evaluate the integral

𝐼 = 9 𝑥 " 𝑧 𝑑𝑣
!

where V is the volume of a cylinder with axis along z, radius a, and z ranging from 0 to
3.

The two simplicity conditions for a triple integral are not both satisfied here if we attempt to
use the cartesian variables, but they are fulfilled if we switch to cylindrical coordinates. The
ranges of the cylindrical variables in V are

r: 0 to a

φ: 0 to 2π

z: 0 to 3

It is seen that all lower limits as well as all upper limits are constants as required by the first
simplicity condition. The second condition, f having a product form, is also satisfied

𝑥 " 𝑧 = 𝜌" cos" 𝜙 𝑧

13
Using the volume element Eq 19, the integral becomes

𝐼 = 9 𝜌" cos" 𝜙 𝑧 𝜌 𝑑𝜌 𝑑𝜙 𝑑𝑧 = 𝐼$ 𝐼" 𝐼#

!

where
% # "3
#
𝑎( 9
𝐼$ = 9 𝜌 𝑑𝜌 = , 𝐼" = 9 𝑧 𝑑𝑧 = , 𝐼# = 9 cos" 𝜙 𝑑𝜙 = 𝜋
4 2
& & &

Hence the final result

𝑎( 9 9𝜋𝑎(
𝐼 = > ?K L𝜋 =
4 2 8

14