PARAMETRIC EQUATIONS OF A 3D SURFACE

z
π/2

u
-π π O
x y

x = cos u sin v
 π
y = sin u sin v [u , v ]∈ [− π , π ] × 0, 
 2
z = cos v
 SIMPLE 3D-SURFACES

M is a planar area bounded by a closed regular curve ∂M

ϕ(u,v), ψ(u,v), χ(u,v) are one-to-one mappings of M into R3

ϕ(u,v), ψ(u,v), χ(u,v) have continuous first partial derivatives

on M

We will call the set

S={ [x,y,z] | x = ϕ(u,v), y = ψ(u,v), z = χ(u,v) }

a simple 3D-surface
A 3D-surface S is closed if it divides R3 into at least two
contiguous parts that cannot be connected by a continuous
line without crossing S.
SPHERE TORUS
x = cos u sin v x = cos v(4 + cos u )
y = sin u sin v [u, v]∈ [0,2π ]× [0,π ] y = sin v(4 + cos u ) [u, v]∈ [0,2π ]× [0,2π ]
z = cos v z = sin u
 If we require the mappings ϕ(u,v), ψ(u,v), χ(u,v) to be one-to-
one only on M – ∂M, that is, not including the boundary, the
surfaces may also be closed. For example, with the sphere in
the previous picture,

The red segment is all taken

to the north pole
The green sides are taken
to the Greenwich meridian

0
2π u
The blue segment is all taken
to the south pole
 Tangent vectors
x = ϕ (u , v ), y = ψ (u , v ), z = χ (u , v )
z
u = u0 tv
v
[ϕ (u0 , v ),ψ (u0 , v ), χ (u0 , v )]
tu
v = v0
v0

[ϕ (u0 , v ),ψ (u0 , v ), χ (u0 , v )]

y

u0 u x

tu = (ϕ 'u (u0 , v0 ),ψ 'u (u0 , v0 ), χ 'u (u0 , v0 ))

t v = (ϕ 'v (u0 , v0 ),ψ 'v (u0 , v0 ), χ 'v (u0 , v0 ))

 Normal
x = ϕ (u , v ), y = ψ (u , v ), z = χ (u , v )
z
u = u0 tv
v n
[ϕ (u0 , v ),ψ (u0 , v ), χ (u0 , v )]

tu
v = v0
v0

[ϕ (u0 , v ),ψ (u0 , v ), χ (u0 , v )]

y

u0 u x

n = tu ×tv =(ψ'u χ'v −ψ'v χ'u , χ'u ϕ'v −χ'v ϕ'u ,ϕ'uψ'v −ϕ'vψ'u )
 AREA OF 3D-SURFACES

To define the area of a 3D surface we could try to proceed in a

way similar to defining the length of a curve. For a curve, we
took all polygons inscribed into it and defined the length as the
lub of the lengths of all such polygons. This could be achieved
by letting the norm of the polygons tend to zero.

B B

A A
For a 3D-surface S, we could similarly take a triangulation
consisting of triangles inscribed in S and let its norm, that is,
the area of the largest triangle, tend to zero. The limit of the
sum of the inscribed triangles would then be taken for the
area of S.
S S
Surprisingly, this is not possible even for such relatively simple
surfaces as a cylindrical surface. The following example due to
Schwarz illustrates this point
n sectors

m layers
 V
h
n VC =
m
B SA = SB = SC = r
S π
D
C ∠BSC = ∠CSA =
n
A
π
2 2
DV = r 2 1 − cos  + 2
ABV = AD DV h
π  n m
AD = SA sin ∠CSA = r sin
n
π π
2
2
2
DV = DC + VC
2
ABV = r sin   h
r 2 1 − cos  + 2
n  n m
π
DC = SC − SD = r − r cos
n
 The area S of the cylinder surface:

S = lim 2mn ABV =

m →∞
n →∞

π π
2 2

r 2 1 − cos  + 2 =
h
= lim 2mnr sin
m →∞
n →∞
n  n m

π π
2

= lim 2nr sin r 2 m 2 1 − cos  + h 2 =

m →∞
n →∞
n  n

π
sin
π π
2 2

= lim 2πr n r 2 m 2 1 − cos + h 2 = lim 2πr r 2 m 2 1 − cos  + h 2

π  
m →∞
n →∞
 n m →∞
n →∞
 n
n
However, the last limit does not exist.
Indeed, since m and n may tend to infinity in an arbitrary way,
we can assume that lim m = q
m →∞ n 2
n →∞

for some constant q. This means that, for large values of n, we

can replace m by qn2 in the limit:

π π
2 2
 
lim 2πr r 2 m 2 1 − cos  + h 2 = lim 2πr r 2 q 2 n 4 1 − cos  + h 2 =
m →∞
n →∞
 n n →∞  n

π
r q π sin
2 2 4 4
π
2
 n + h2
= lim 2πr r 2 q 2 n 4 1 − cos  + h 2 = lim 2πr
 n π π
n →∞ n →∞ 2 4
 
 1 + cos  4
 n  n
This means that

π
r q π sin
2 2 4 4

n r qπ
2 2 4
S = lim 2πr + h 2
= 2πr + h 2

π π
n →∞ 2 4
  4
1 + cos  4
 n n

and so the result depends on q. Thus the surface cannot be

determined using this method.
The reason why this method fails when calculating surface
areas whereas it is successful in calculating the length of a
curve is the following. As can be seen in the picture below,
with the length of the polygon segments becoming smaller,
their direction tends to that of the appropriate tangent vectors:

B
A
 However, in Schwartz’s
V example, the unit normals of
n the approximating triangles
do not tend to those of the
B cylindrical surface. For
S example for the triangle on
C
D the left, the coordinates of
A the unit normal are:

2 − rπ 2 q m
n= i+0j+ k where lim = q.
4+π r q
4 2 2
4+π r q
4 2 2 m →∞ n 2
n →∞

Only for q = 0 do we get n = 1i + 0 j + 0k , which is the unit normal

of the cylindrical surface. However, for q → ∞, we have

n = 0i + 0 j − 1k which is perpendicular to it.
To calculate the area of a surface we will use the concept of a
tangent plane.
z
tangent vector

(ϕ 'v ,ψ 'v , χ 'v ) v

S
M

y
tangent vector
u
(ϕ 'u ,ψ 'u , χ 'u )
x
the area of the piece will be approximated by that of the “peeling off”
red tile made from the tangent plane
 tangent vector
n = ( A, B, C ) = tu × t v
tv=(ϕv, ψv, χv)
z
normal i j k
n = (A,B,C) S tu × tv = ϕ u ψ u χu
ϕv ψ v χv
ψu χu χu ϕu
A= B=
ψv χv χv ϕv

y C = ϕu ψ u
ϕv ψ v
Area of the tile = | n |
x tangent vector
tu=(ϕu, ψu, χu) n = A2 + B 2 + C 2
The area |S| of a surface S defined by the parametric equations
x = ϕ (u , v ), y = ψ (u , v ), z = χ (u , v ), [u, v]∈ M
can be calculated using the following formula

ψ 'u χ 'u χ 'u ϕ 'u ϕ 'u ψ ' u

2 2 2

S = ∫∫ + + du dv
M
ψ 'v χ 'v χ 'v ϕ 'v ϕ 'v ψ 'v
For a 3D-surface S expressed by the explicit function

z = f ( x, y ), [ x, y ] ∈ M

we get the following formula

2 2 2
0 z 'u z 'u 1 1 0
S = ∫∫ + + du dv
M
1 z 'v z 'v 0 0 1

which yields

S = ∫∫ z 'u + z 'v +1 du dv
2 2

M
Example

Calculate the area of a sphere with a radius of r.

We will choose a sphere with the centre at [0,0,0],

which has the parametric equations

x = r cos u sin v
y = r sin u sin v [u, v]∈ [0,2π ]× [0,π ]
y = r cos v

Clearly, for reasons of symmetry, we can consider an

M’ : [0,π/2]×[0,π/2] multiplying the result by 8.
We have
t u = (− r sin u sin v, r cos u sin v,0 )
t v = (r cos u cos v, r sin u cos v,−r sin v )
and so

A = − r 2 cos u sin 2 v, B = − r 2 sin u sin 2 v, C = −r 2 sin 2 u sin v cos v

which yields

A 2 + B 2 + C 2 = r 4 (sin 4 v + sin 2 v cos 2 v )and

A 2 + B 2 + C 2 = r 2 sin v
π π
2 2
π
S = 8r 2
∫ sin v du dv = 8r
2
∫ du ∫ sin v dv = 8r
2
(1) = 4πr 2
M' 0 0
2
 Orientating a surface

To orientate a surface means to say which side is the „upper one“.

We do this by orienting the normals. This then makes it possible to
deter- mine the orientation of closed curves. (right-handed and left-
handed threads).
 S

P C

For the definition of orientation to be correct, the following

condition must be true for any closed curve C in S:

After moving the normal continuously along C and

(*)
returning to the starting point P, it must still have the same
direction.
 Orientable surfaces

3D-surfaces for which condition (*) is satisfied are called

orientable.
There are 3D surfaces that do not meet this condition and are not
orientable.
An example of a non-orientable 3D-surface is the Möbius band

u
x = cos u − v sin u sin
2
u ∈ [0,2π ), v ∈ [− 0.1,0.1]
u
y = sin u + v cos u sin
2
u
z = v cos
2