2 Surface integrals
2.1 Surfaces in R
3
Figure 2.1:
a. There are several possibilities to describe surface in the space.
The simplest way is to use the explicit form, that is to deﬁne
surface as a graph of some function z = f(x, y).
Example 2.1 Graph of f(x, y) = e
−x
2
−y
2
sin xy, x ∈ [−1, 1],
y ∈ [−1, 1] determines the surface of ﬁg. 2.1.
φ
ψ
r
Figure 2.2:
Another standard way is to use the implicit form. In this case
we deﬁne surface with the help of the equation F(x, y, z) = 0.
Example 2.2 The ellipsoid with radii a, b and c in R
3
is given
implicitly by
x
2
a
2
+
y
2
b
2
+
z
2
c
2
= 1,
see ﬁg. 2.2.
Figure 2.3:
Example 2.3 The cylinder with radii a and b is given implicitly
by
x
2
a
2
+
y
2
b
2
= 1,
see ﬁg. 2.3.
Finally, we can use the parametric form to describe a surface.
Deﬁnition 2.1 Let
r(u, v) = i x(u, v) +j y(u, v) +k z(u, v), (2.1)
be a vector valued function deﬁned in a set D ⊂ R
2
. The range S
of r(u, v) is called the surface traces by r. The set D and the
equation (2.1) constitutes a parametrization of the surface. The
variables u and v are called the parameters, and D is called the
parameter domain.
Example 2.4 Find parametrization of the sphere x
2
+ y
2
+ z
2
= a
2
.
Solution: In spherical coordinates any point on the sphere is given by
x(φ, ψ) = a cos φ sin ψ, y(φ, ψ) = a sin φ sinψ, z(φ, ψ) = a cos ψ,
where φ ∈ [0, 2π] and ψ ∈ [0, π]. Hence, the parametrization is given by
r(φ, ψ) = i a cos φ sin ψ +j a sin φ sinψ + k a cos ψ, φ ∈ [0, 2π], ψ ∈ [0, π].
37
Example 2.5 Find parametrization of the ellipsoid
x
2
a
2
+
y
2
b
2
+
z
2
c
2
= a
2
,
see ﬁg. 2.2.
Solution: A point on the ellipsoid has the coordinates
x(φ, ψ) = a cos φ sin ψ, y(φ, ψ) = b sin φ sin ψ, z(φ, ψ) = c cos ψ,
where φ ∈ [0, 2π] and ψ ∈ [0, π]. Hence, the parametrization is given by
r(φ, ψ) = i a cos φ sinψ +j b sin φ sin ψ +k c cos ψ, φ ∈ [0, 2π], ψ ∈ [0, π].
Example 2.6 Find parametrization of the cone
z
2
=
x
2
a
2
+
y
2
b
2
,
where z ∈ [0, h], see ﬁg. 2.4.
φ
z
r
Figure 2.4:
Solution: A typical point on the cone has the coordinates
x(r, φ) = ar cos φ, y(r, φ) = br sin φ, z(r, φ) =
_
x
2
a
2
+
y
2
b
2
= r,
where r ∈ [0, h] and φ ∈ [0, 2π]. Hence, the parametrization is given by
r(r, φ) = i ar cos φ +j br sin φ +k r, r ∈ [0, h], φ ∈ [0, 2π].
b. In the course we deal with smooth surfaces.
Deﬁnition 2.2 A parametric surface r(u, v) = i x(u, v) +j y(u, v) +k z(u, v) is smooth
if r
u
and r
v
are continuous and r
u
×r
v
is non zero on the parameter domain.
The condition r
u
×r
v
= 0 implies that two vectors r
u
and r
v
are not parallel and never
zero, so they always span a tangent plane to the surface.
Example 2.7 A sphere, a paraboloid and a cylinder are smooth surfaces. The cone of
example 2.6 is not smooth since r
r
×r
φ
= −i br cos φ +j ar sin φ +k abr = 0 when r = 0.
38
2.2 Surface integrals
x
y
z
s
∗
i
σ
i
S
Figure 2.5:
a. The notion of surface integrals is introduced in the
same way as the notion of line integrals with respect to
arc length.
Deﬁnition 2.3 Let S ⊂ R
3
be a surface and let f :
S → R be a real valued function. Consider a ﬁnite
partition P = {σ
0
, σ
1
, . . . , σ
n
} of S into n patches (see
ﬁg. 2.5). Let △σ
i
denote the area of ith patch and let
P = max
1≤i≤n
△σ
i
denote the norm of the partition.
For each patch we choose arbitrary point s
∗
i
and form
the integral sum
S
n
=
n
i=1
f(s
∗
i
)△σ
i
.
The surface integral of f over the surface S is given by
lim
P→0
S
n
= lim
P→0
n
i=1
f(s
∗
i
)△σ
i
=
__
S
fdσ, (2.2)
if this limit exists.
Remark 2.1 In the formula (2.2) symbol S indicates the domain of integration, f is the
integrand and dσ denotes the area element on S.
Remark 2.2 A surface integral (2.2) is additive, i.e. if a surface S made by joining of a
ﬁnite number of surfaces S = S
1
∪ S
2
∪ . . . ∪ S
n
=
n
i=1
S
i
, then the integral of a function f
over S is the sum of integrals over the surfaces that make it up
__
S
fdσ =
n
i=1
__
S
i
fdσ.
Remark 2.3 If we let f(x, y, z) = 1, then surface integral (2.2) gives the area A of the
surface S, i.e.
A =
__
S
dσ.
b. In order to make abstract deﬁnition 2.3 practical we must answer two important
questions. The ﬁrst question is: what conditions guarantee the existence of a surface
integral (2.2)? The second one is: how to evaluate this integral? To answer the ﬁrst
question it is suﬃcient to assume that S is a smooth surface and f is continuous in S. The
answer to the second question depends on the form of deﬁnition of S. We begin our study
with the implicitly deﬁned surfaces.
39
Theorem 2.1 Let S ⊂ R
3
be a smooth surface deﬁned implicitly by the equation F(x, y, z) =
c, let D ⊂ R
2
be a projection of S onto the (x, y)plane, and let f : S →R be a continuous
function, then the surface integral of f over S exists and its value is given by the formula:
__
S
fdσ =
__
D
f
∇F
∇F · k 
dxdy, (2.3)
where ∇F · k  = 0.
x y
z
σ
∗
i
σ
i
P
i
S
D
∇F
k
γ
i
Figure 2.6:
Proof: By the deﬁnition of surface integrals we have
__
S
fdσ = lim
P→0
n
i=1
f(s
∗
i
)△σ
i
,
where P = σ
0
, σ
1
, . . . , σ
n
is a partition of S. In order
to prove the theorem we have to evaluate the area △σ
i
of the patch σ
i
, see ﬁg. 2.6.
First, we observe that any ﬁnite partition of S in
duces the ﬁnite partition P
∗
= {σ
∗
0
, σ
∗
1
, . . . , σ
∗
n
} of the
shadow region D in such a way that σ
∗
i
⊂ D is a shadow
region for the patch σ
i
⊂ S.
Second, let s
∗
i
= (x
i
, y
i
, z
i
) be a point on the patch
σ
i
. If the norm of the partition P is small then the area of the patch is approximately
equal to the area △P
i
of the parallelogram P
i
on the tangent plane of S at s
∗
i
, see ﬁg. 2.6.
From the basic geometrical considerations it follows that
△P
i
 cos γ
i
 = σ
∗
i
,
where γ
i
is the angle between the normal vector to the surface S at s
∗
i
and the normal
vector to the patch σ
∗
i
, see ﬁg. 2.6. Observe that gradient ∇F(s
∗
i
) is normal to S, and k
is normal to σ
∗
i
. Since ∇F(s
∗
i
) · k  = ∇F(s
∗
i
)k  cos γ
i
 = ∇F(s
∗
i
) cos γ
i
, we obtain
△σ
i
≈ △P
i
=
σ
∗
i
 cos γ
i

=
∇F(s
∗
i
)
∇F(s
∗
i
) · k 
σ
∗
i
.
Third, substituting this expression into the integral sum we obtain
S
n
≈
n
i=1
f(s
∗
i
)
∇F(s
∗
i
)
∇F(s
∗
i
) · k 
σ
∗
i
.
The right hand side of the last formula is a Riemann sum of f
∇F
∇F · k 
with respect to
the partition P
∗
of D. Taking a limit of S
n
as P →0 we obtain
__
S
fdσ =
__
D
f
∇F
∇F · k 
dxdy.
The double integral over D exists because f
∇F
∇F · k 
is continuous on D.
40
Example 2.8 Find integral of f(x, y, z) = z over the surface cut from the paraboloid
x
2
+ y
2
−z = 0 by the plane z = 4, see ﬁg. 2.7.
Figure 2.7:
Solution: The shadow area of the paraboloid is given by the
circle x
2
+ y
2
≤ 4. The gradient is
∇F = ∇(x
2
+ y
2
−z) = i 2x +j 2y −k ,
norm of the gradient
∇F =
_
4x
2
+ 4y
2
+ 1,
and ∇F · k  =  −1 = 1. For the paraboloid z = x
2
+ y
2
, thus, using (2.3) we obtain
__
S
zdσ =
__
D
(x
2
+ y
2
)
_
4x
2
+ 4y
2
+ 1dxdy =
_
2
0
_
2π
0
r
3
√
1 + 4r
2
dφdr
= 2π
_
2
0
r
3
√
1 + 4r
2
dr =
π
16
_
17
1
(u −1)
√
udu =
π
16
_
2
5
u
5/2
−
2
3
u
3/2
_
¸
¸
¸
¸
17
1
= π
391
√
17 + 1
60
.
Note that we have used polar coordinates to evaluate the double integral above.
Example 2.9 Find the area of the surface cut from the hemisphere x
2
+ y
2
+ z
2
= 2,
z ≥ 0, by the plane z = 1, see ﬁg. 2.8.
Figure 2.8:
Solution: The projection of the surface S onto (x, y)plane is the
annulus 1 ≤ x
2
+y
2
≤ 2. The gradient of F(x, y, z) = x
2
+y
2
+z
2
−2
is
∇F = ∇(x
2
+ y
2
+ z
2
−2) = i 2x +j 2y +k 2z,
norm of the gradient
∇F =
_
4x
2
+ 4y
2
+ 4z
2
= 2
_
x
2
+ y
2
+ z
2
= 2
√
2,
and ∇F · k  = 2z = 2z. For the hemisphere we have z =
_
2 −x
2
−y
2
, and using (2.3)
we obtain
A =
__
S
dσ =
√
2
__
D
dxdy
z
=
√
2
__
D
dxdy
_
2 −x
2
−y
2
=
√
2
_
√
2
1
_
2π
0
r
√
2 −r
2
dφdr
= 2
√
2π
_
√
2
1
r
√
2 −r
2
dr = 2
√
2π
_
−(2 −r
2
)
1/2
_¸
¸
¸
√
2
1
= 2
√
2π.
Note that we have used polar coordinates to evaluate the double integral above.
41
In theorem 2.1 we have assumed that the shadow region (projection) of S belongs to the
(x, y)plane. But we can also project S onto (x, z), (y, z) or any other planes. If D denotes
a projection of S onto arbitrary plane then formula (2.3) must be modiﬁed as follows:
__
S
fdσ =
__
D
f
∇F
∇F · p
dA. (2.3
∗
)
where p is a unit vector normal to D and dA is the area element in D.
Example 2.10 Find the area of the hemisphere x
2
+ y
2
+ z
2
= 1, y ≥ 0, see ﬁg. 2.9.
Figure 2.9:
Solution: Projecting the hemisphere onto (x, z)plane we see
that the shadow region is given by the disk x
2
+ z
2
≤ 1. The
gradient of F(x, y, z) = x
2
+ y
2
+ z
2
−1 is
∇F = ∇(x
2
+ y
2
+ z
2
−1) = i 2x +j 2y +k 2z,
norm of the gradient
∇F =
_
4x
2
+ 4y
2
+ 4z
2
= 2
_
x
2
+ y
2
+ z
2
= 2.
Since the shadow area D belongs to (x, z)plane the unit vector normal to D is given by
j , for the area element we have dA = dxdz. In this example ∇F · j  = 2y = 2y and
y =
√
1 −x
2
−z
2
. Using formula (2.3
∗
) we obtain
A =
__
S
dσ =
__
D
dxdz
y
=
__
D
dxdz
√
1 −x
2
−z
2
=
_
1
0
_
2π
0
r
√
1 −r
2
dφdr
= 2π
_
1
0
r
√
1 −r
2
dr = 2π
_
−(1 −r
2
)
1/2
_¸
¸
¸
1
0
= 2π.
Here again we have used change to the polar coordinates.
c. Let us assume now that surface S is deﬁned parametrically. In this situation we have
the following
Theorem 2.2 Let S ⊂ R
3
be a smooth surface determined parametrically
r(u, v) = i x(u, v) +j y(u, v) +k z(u, v),
let D ⊂ R
2
be a parameter domain, and let f : S → R be a continuous function, then the
surface integral of f over S exists and its value is given by the formula:
__
S
fdσ =
__
D
f(x(u, v)), y(u, v), z(u, v))r
u
×r
v
dudv. (2.4)
42
σ
∗
i
△v
i
△u
i
(u
i
, v
i
)
u
v
x y
z
σ
i
P
i
S
r
v
△v
i
r
u
△u
i
Figure 2.10:
Proof: The idea of the proof is similar to that we have used to prove theorem 2.1. Given
a ﬁnite partition P = σ
0
, σ
1
, . . . , σ
n
of S, in order to prove the theorem we have to evaluate
the area △σ
i
of the ith patch, see ﬁg. 2.10.
First, we observe that any ﬁnite partition of S induces the ﬁnite partition P
∗
=
σ
∗
0
, σ
∗
1
, . . . , σ
∗
n
of the parameter domain D in such a way that r(σ
∗
i
) = σ
i
.
Second, let σ
∗
i
be a small rectangle in D with sides △u
i
and △v
i
. In this case we obtain
that σ
∗
i
= △u
i
△v
i
. Each side of σ
∗
i
maps to a curve on the surface S and point (u
i
, v
i
)
maps to s
∗
i
= (x
i
, y
i
, z
i
), see ﬁg. 2.10. The vectors r
u
(u
i
, v
i
) and r
v
(u
i
, v
i
) are tangent to
S at (x
i
, y
i
, z
i
), their cross product r
u
(u
i
, v
i
) ×r
v
(u
i
, v
i
) is orthogonal to S and never zero
because S is smooth. If the norm of the partition P is small then the area of the patch σ
i
is approximately equal to the area △P
i
of the parallelogram on the tangent plane spanned
by vectors r
u
(u
i
, v
i
)△u
i
and r
v
(u
i
, v
i
)△v
i
. Using properties of cross product we obtain
△σ
i
≈ △P
i
= r
u
(u
i
, v
i
)△u
i
×r
v
(u
i
, v
i
)△v
i
= r
u
(u
i
, v
i
) ×r
v
(u
i
, v
i
)△u
i
△v
i
.
Third, substituting this expression into the integral sum we obtain
S
n
≈ lim
P→0
n
i=1
f(r(u
i
, v
i
))r
u
(u
i
, v
i
) ×r
v
(u
i
, v
i
)△u
i
△v
i
,
The right hand side of the last formula is a Riemann sum of fr
u
× r
v
with respect to
the partition P
∗
of D. Taking a limit of S
n
as P →0 we obtain
__
S
fdσ =
__
D
f(x(u, v), y(u, v), z(u, v))r
u
×r
v
dudv.
The double integral over D exists because S is smooth and f is continuous.
43
Example 2.11 Find the area A of the sphere x
2
+ y
2
+ z
2
= a
2
.
Solution: Parametrization of the sphere is given in the example 2.4. Let us evaluate
r
φ
×r
ψ
, we have
r
φ
= −i a sin φ sinψ +j a cos φ sin ψ,
r
ψ
= i a cos φ cos ψ +j a sin φ cos ψ −k a sin ψ,
r
φ
×r
ψ
=
¸
¸
¸
¸
¸
¸
i j k
−a sin φ sinψ a cos φ sinψ 0
a cos φ cos ψ a sin φ cos ψ −a sin ψ
¸
¸
¸
¸
¸
¸
= −i a
2
cos φ sin
2
ψ −j a
2
sin φ sin
2
ψ −k a
2
sin ψ cos ψ,
r
φ
×r
ψ
=
_
a
4
cos
2
φ sin
4
ψ + a
4
sin
2
φ sin
4
ψ + a
4
sin
2
ψ cos
2
ψ
= a
2
sin ψ
_
cos
2
φ sin
2
ψ + sin
2
φ sin
2
ψ + cos
2
ψ = a
2
sin ψ.
Using formula 2.4 we obtain
A =
__
S
dσ = a
2
_
2π
0
_
π
0
sin ψdφdψ = 2πa
2
_
π
0
sin ψdψ = 4πa
2
.
Example 2.12 Evaluate the integral
__
S
zdσ, where S is a surface cut from the cylinder
x
2
+ y
2
= 1 by the planes z = 0 and z = x + 1, see ﬁg. 2.11.
Figure 2.11:
Solution: The standard parametrization of the cylinder is given
by
r(φ, z) = i cos φ +j sin φ +k z.
Since 0 ≤ z ≤ x + 1 = cos φ + 1 the parameter domain D is given
by φ ∈ [0, 2π] and z ∈ [0, cos φ + 1]. Let us evaluate r
φ
×r
z
, we
have
r
φ
= −i sin φ +j cos φ, r
z
= k ,
r
φ
×r
z
=
¸
¸
¸
¸
¸
¸
i j k
−sin φ cos φ 0
0 0 1
¸
¸
¸
¸
¸
¸
= i cos φ +j sin φ,
r
φ
×r
z
=
_
cos
2
φ + sin
2
φ = 1.
Using (2.4) we obtain
__
S
zdσ =
__
D
zdφdz =
_
2π
0
_
cos φ+1
0
zdzdφ =
1
2
_
2π
0
(cos φ + 1)
2
dφ
=
1
2
_
2π
0
_
1 + 2 cos φ +
1
2
(1 + cos 2φ)
_
dφ
=
1
2
_
3
2
φ + 2 sin φ +
1
4
sin 2φ
_
¸
¸
¸
¸
2π
0
=
3
2
π.
44
2.2.1 Applications
Mass. If the density (mass per unit area) of a thin shell S is given by ρ(x, y, z), then its
mass M is
M =
__
S
ρdσ. (2.5)
First moments about the coordinate planes.
M
xy
=
__
S
zρdσ, M
xz
=
__
S
yρdσ, M
yz
=
__
S
xρdσ. (2.6)
Coordinates of the center of mass.
x = M
yz
/M, y = M
xz
/M, z = M
xy
/M. (2.7)
Moment of inertia and radius of gyration about a line. If r(x, y, z) denotes the
distance from (x, y, z) to a line ℓ, then moment of inertia and radius of gyration about the
line are given by
I
ℓ
=
__
S
r
2
ρdσ and R
ℓ
=
_
I
ℓ
/M. (2.8)
Note that for axes x, y and z we have r
x
(x, y, z) =
_
y
2
+ z
2
, r
y
(x, y, z) =
√
x
2
+ z
2
, and
r
z
(x, y, z) =
_
x
2
+ y
2
respectively, thus,
I
x
=
__
S
(y
2
+ z
2
)ρdσ, I
y
=
__
S
(x
2
+ z
2
)ρdσ, I
z
=
__
S
(x
2
+ y
2
)ρdσ. (2.9)
Example 2.13 Evaluate mass of the cone z
2
= x
2
+y
2
, z ∈ [0, 1] if its density is given by
ρ(x, y, z) = z.
Solution: The standard parametrization of the cone is given by
r(φ, z) = i z cos φ +j z sin φ +k z, φ ∈ [0, 2π], z ∈ [0, 1].
We have
r
φ
= −i z sin φ +j z cos φ, r
z
= i cos φ +j sin φ +k ,
r
φ
×r
z
=
¸
¸
¸
¸
¸
¸
i j k
−z sin φ z cos φ 0
cos φ sin φ 1
¸
¸
¸
¸
¸
¸
= i z cos φ +j z sin φ −k z,
r
φ
×r
z
=
_
z
2
cos
2
φ + z
2
sin
2
φ + z
2
=
√
2z,
and
M =
__
S
zdσ =
√
2
_
1
0
_
2π
0
z
2
dφdz = 2
√
2π
_
1
0
z
2
dz =
2
√
2
3
π.
45
Example 2.14 Determine center of mass of the hemisphere x
2
+ y
2
+ z
2
= a
2
, z ≥ 0.
Assume that density is constant ρ.
Solution: Radial symmetry of S about zaxis implies that x = 0 and y = 0. Let us
evaluate z. Parametrization of the sphere is given in example 2.4. For the upper hemisphere
we must assume that φ ∈ [0, 2π] and ψ ∈ [0, π/2], using the result of example 2.11 we obtain
that
M =
__
S
ρdσ = ρ
__
S
dσ = 2πa
2
ρ.
For the moment we have:
M
xy
=
__
S
zρdσ = ρ
__
S
zdσ = ρ
_
2π
0
_
π/2
0
a
3
sin ψ cos ψdψdφ
=
a
3
ρ
4
_
2π
0
−cos 2ψ
¸
¸
¸
π/2
0
dφ =
πa
3
ρ
2
,
ﬁnally,
z =
M
xy
M
=
πa
3
ρ
4πa
2
ρ
=
a
4
.
46
b. ψ ∈ [0. a2 b r φ ∈ [0.4: x2 y 2 + 2. 2π]. 2. v) = i x(u. φ ∈ [0.
where φ ∈ [0. a2 b c see ﬁg. see ﬁg. so they always span a tangent plane to the surface. The cone of example 2. 2π]. Example 2. Solution: A point on the ellipsoid has the coordinates x(φ. y(r. the parametrization is given by r(r. φ) = br sin φ. the parametrization is given by r(φ.6 Find parametrization of the cone z2 = where z ∈ [0. 2π] and ψ ∈ [0. ψ) = b sin φ sin ψ.5 Find parametrization of the ellipsoid x2 y 2 z 2 + 2 + 2 = a2 . 2π]. Solution: A typical point on the cone has the coordinates x(r. h] and φ ∈ [0. r ∈ [0. Hence.4. φ) = ar cos φ. ψ) = a cos φ sin ψ.2. y(φ. In the course we deal with smooth surfaces. z(r. The condition ru × rv = 0 implies that two vectors ru and rv are not parallel and never zero. z
where r ∈ [0. a2 b φ Figure 2. Hence. v) + j y(u. ψ) = i a cos φ sin ψ + j b sin φ sin ψ + k c cos ψ. z(φ. π]. π]. φ) = i ar cos φ + j br sin φ + k r. Deﬁnition 2. h]. v) + k z(u. h]. φ) = x2 y 2 + 2 = r.7 A sphere. v) is smooth if ru and rv are continuous and ru × rv is non zero on the parameter domain.Example 2. ψ) = c cos ψ.
38
.6 is not smooth since rr × rφ = −i br cos φ + j ar sin φ + k abr = 0 when r = 0. a paraboloid and a cylinder are smooth surfaces.2 A parametric surface r(u. Example 2. 2.
39
. The ﬁrst question is: what conditions guarantee the existence of a surface integral (2. ∪ Sn =
n
Si .2) gives the area A of the surface S.e.
1≤i≤n
For each patch we choose arbitrary point s∗ and form i the integral sum
n
Sn =
i=1
f (s∗ )△σi .2
Surface integrals
z S y s∗ i σi x Figure 2.2) symbol S indicates the domain of integration. Remark 2.
b.2 A surface integral (2. In order to make abstract deﬁnition 2. i
The surface integral of f over the surface S is given by
n P →0
lim Sn = lim
P →0
f (s∗ )△σi = i
i=1
f dσ. The answer to the second question depends on the form of deﬁnition of S.2. then the integral of a function f
i=1
over S is the sum of integrals over the surfaces that make it up f dσ =
S i=1 Si
f dσ.2)? The second one is: how to evaluate this integral? To answer the ﬁrst question it is suﬃcient to assume that S is a smooth surface and f is continuous in S.3 practical we must answer two important questions. then surface integral (2.3 Let S ⊂ R3 be a surface and let f : S → R be a real valued function. . y. i.2)
if this limit exists.3 If we let f (x.
Remark 2. . σ1 . f is the integrand and dσ denotes the area element on S.
S
(2.2) is additive. . We begin our study with the implicitly deﬁned surfaces. .e. Let △σi denote the area of ith patch and let P = max △σi denote the norm of the partition. 2. z) = 1. i. if a surface S made by joining of a
n
ﬁnite number of surfaces S = S1 ∪ S2 ∪ . . The notion of surface integrals is introduced in the same way as the notion of line integrals with respect to arc length.1 In the formula (2. A=
S
dσ. . Remark 2.5:
a. σn } of S into n patches (see ﬁg. Consider a ﬁnite partition P = {σ0 . Deﬁnition 2.5).
6. y)plane.3)
Proof: By the deﬁnition of surface integrals we have f dσ = lim
S P →0
f (s∗ )△σi . Since ∇F (s∗) · k  = ∇F (s∗ ) k  cos γi  = ∇F (s∗ )  cos γi . . i
i=1
where P = σ0 . let D ⊂ R2 be a projection of S onto the (x. substituting this expression into the integral sum we obtain Sn ≈ f (s∗ ) i
i=1
∇F (s∗) i σ∗ . i From the basic geometrical considerations it follows that where γi is the angle between the normal vector to the surface S at s∗ and the normal i ∗ vector to the patch σi . see ﬁg. ∇F · k 
40
. zi ) be a point on the patch i σi . yi. z) = c. ∇F · k  k z ∇F Pi γi σi
(2. If the norm of the partition P is small then the area of the patch is approximately equal to the area △Pi of the parallelogram Pi on the tangent plane of S at s∗ . then the surface integral of f over S exists and its value is given by the formula: f dσ =
S D
f
where ∇F · k  = 0. . 2.6. △σi ≈ △Pi = ∗  cos γi  ∇F (si ) · k  n ∗ △Pi  cos γi  = σi . see ﬁg. σn } of the x ∗ shadow region D in such a way that σi ⊂ D is a shadow D ∗ σi region for the patch σi ⊂ S. σ1 . In order to prove the theorem we have to evaluate the area △σi S of the patch σi .6: Second. Figure 2. . we observe that any ﬁnite partition of S in∗ ∗ ∗ y duces the ﬁnite partition P ∗ = {σ0 . σn is a partition of S.Theorem 2.
Third. and k i ∗ is normal to σi . Taking a limit of Sn as P → 0 we obtain The right hand side of the last formula is a Riemann sum of f ∇F dxdy.
n
∇F dxdy. . y. First. . . σ1 .1 Let S ⊂ R3 be a smooth surface deﬁned implicitly by the equation F (x. ∇F · k  f dσ =
S D
f
The double integral over D exists because f
∇F is continuous on D. . 2. Observe that gradient ∇F (s∗ ) is normal to S. 2. let s∗ = (xi . . ∇F (s∗) · k  i i
∇F with respect to ∇F · k  the partition P ∗ of D. and let f : S → R be a continuous function. see ﬁg.6. we obtain i i i
∗ σi ∇F (s∗ ) i ∗ = σi .
41
. Solution: The projection of the surface S onto (x. Example 2. by the plane z = 1. For the hemisphere we have z = we obtain A =
S
dσ =
√ 1
√
2
2
D
dxdy √ = 2 z
dxdy
D
√ = 2 2π
√
√ r dr = 2 2π −(2 − r 2 )1/2 2 − r2
2−
x2
−
y2
√ 1 2
=
2
1
√
√ = 2 2π. y. see ﬁg. z) = x2 +y 2 +z 2 −2 is ∇F = ∇(x2 + y 2 + z 2 − 2) = i 2x + j 2y + k 2z.7:
and ∇F · k  =  − 1 = 1.8: 2 − x2 − y 2 . 2. norm of the gradient ∇F = 4x2 + 4y 2 + 1.3) √
√ 2 0 2π
and ∇F · k  = 2z = 2z. see ﬁg. norm of the gradient ∇F = √ 4x2 + 4y 2 + 4z 2 = 2 x2 + y 2 + z 2 = 2 2. The gradient of F (x. For the paraboloid z = x2 + y 2 . and using (2.8 Find integral of f (x.Example 2.9 Find the area of the surface cut from the hemisphere x2 + y 2 + z 2 = 2. y)plane is the annulus 1 ≤ x2 +y 2 ≤ 2. Solution: The shadow area of the paraboloid is given by the circle x2 + y 2 ≤ 4.8. y.
r dφdr 2 − r2
Note that we have used polar coordinates to evaluate the double integral above. using (2.7. Figure 2. The gradient is ∇F = ∇(x2 + y 2 − z) = i 2x + j 2y − k . Figure 2. 2. thus. 60 = 2π
√ π (u − 1) udu = 16
Note that we have used polar coordinates to evaluate the double integral above. z) = z over the surface cut from the paraboloid x2 + y 2 − z = 0 by the plane z = 4.3) we obtain
2 2π 0
zdσ =
S D
(x + y )
2
2
2
4x2
+
4y 2
+ 1dxdy =
0 17 1
√ r 3 1 + 4r 2 dφdr 2 5/2 2 3/2 u − u 5 3
17 1
√ π r 3 1 + 4r 2 dr = 16 0 √ 391 17 + 1 = π . z ≥ 0.
v). v). z(u. 2. Since the shadow area D belongs to (x.2 Let S ⊂ R3 be a smooth surface determined parametrically r(u. v)). ∇F · p
(2. y ≥ 0. v) = i x(u. The gradient of F (x. z) or any other planes. In this example ∇F · j  = 2y = 2y and √ y = 1 − x2 − z 2 . y. Solution: Projecting the hemisphere onto (x. z). But we can also project S onto (x. y)plane. for the area element we have dA = dxdz. and let f : S → R be a continuous function. v) + j y(u. Example 2. v)) ru × rv dudv. c.In theorem 2. Using formula (2.3) must be modiﬁed as follows: f dσ =
S D
f
∇F dA. If D denotes a projection of S onto arbitrary plane then formula (2. let D ⊂ R2 be a parameter domain. (y. then the surface integral of f over S exists and its value is given by the formula: f dσ =
S D
f (x(u.
Here again we have used change to the polar coordinates. z) = x2 + y 2 + z 2 − 1 is ∇F = ∇(x2 + y 2 + z 2 − 1) = i 2x + j 2y + k 2z. v) + k z(u.1 we have assumed that the shadow region (projection) of S belongs to the (x.3∗ ) we obtain A =
S 1
dσ =
D
dxdz = y
D
√
= 2π
0
√
r dr = 2π −(1 − r 2 )1/2 2 1−r
dxdz = 1 − x2 − z 2
1 0
1 0 0
2π
√
r dφdr 1 − r2
= 2π.4)
42
. see ﬁg. Let us assume now that surface S is deﬁned parametrically. z)plane we see that the shadow region is given by the disk x2 + z 2 ≤ 1. y(u. In this situation we have the following Theorem 2. norm of the gradient Figure 2.9.3∗ )
where p is a unit vector normal to D and dA is the area element in D.9: ∇F = 4x2 + 4y 2 + 4z 2 = 2 x2 + y 2 + z 2 = 2. z)plane the unit vector normal to D is given by j .10 Find the area of the hemisphere x2 + y 2 + z 2 = 1.
(2.
vi ) is orthogonal to S and never zero because S is smooth. . σ1 . substituting this expression into the integral sum we obtain
n
Sn ≈ lim
P →0
i=1
f (r(ui. . vi )△ui and rv (ui .
43
. vi ) △ui△vi . vi ) × rv (ui . σn of S. 2. in order to prove the theorem we have to evaluate the area △σi of the ith patch. . y(u. vi ) △ui△vi. The vectors ru (ui . vi )
Proof: The idea of the proof is similar to that we have used to prove theorem 2. vi )△ui × rv (ui . . z(u. vi )) ru (ui. yi.10:
Pi σi y S
(ui . Taking a limit of Sn as P → 0 we obtain f dσ =
S D
f (x(u. . ∗ Second. Using properties of cross product we obtain △σi ≈ △Pi = ru (ui. vi ) × rv (ui . yi . vi ) are tangent to i S at (xi . . see ﬁg.10.z rv △vi v △ui
∗ σi
ru △ui △vi x u Figure 2.10. v). let σi be a small rectangle in D with sides △ui and △vi . 2. zi ). see ﬁg. vi )△vi. First. v)) ru × rv dudv.1. we observe that any ﬁnite partition of S induces the ﬁnite partition P ∗ = ∗ ∗ ∗ ∗ σ0 . If the norm of the partition P is small then the area of the patch σi is approximately equal to the area △Pi of the parallelogram on the tangent plane spanned by vectors ru (ui . Each side of σi maps to a curve on the surface S and point (ui . Third. . In this case we obtain ∗ ∗ that σi = △ui △vi . their cross product ru (ui. zi ). vi ) and rv (ui . σ1 . v). . vi ) × rv (ui .
The right hand side of the last formula is a Riemann sum of f ru × rv with respect to the partition P ∗ of D. vi )△vi = ru (ui . Given a ﬁnite partition P = σ0 . vi ) maps to s∗ = (xi . σn of the parameter domain D in such a way that r(σi ) = σi .
The double integral over D exists because S is smooth and f is continuous.
11 Find the area A of the sphere x2 + y 2 + z 2 = a2 . cos φ + 1]. i j k − sin φ cos φ 0 rφ × rz = 0 0 1 rφ × rz Using (2. see ﬁg.Example 2. 2 4 2 0 44
. i j k −a sin φ sin ψ a cos φ sin ψ 0 rφ × rψ = a cos φ cos ψ a sin φ cos ψ −a sin ψ =
= −i a2 cos φ sin2 ψ − j a2 sin φ sin2 ψ − k a2 sin ψ cos ψ. we have rφ = −i a sin φ sin ψ + j a cos φ sin ψ. Let us evaluate rφ × rψ . 2. Solution: The standard parametrization of the cylinder is given by r(φ.
=
cos2 φ + sin2 φ = 1.4 we obtain A=
S
dσ = a2
0 0
sin ψdφdψ = 2πa2
0
sin ψdψ = 4πa2 .
2π π π
rφ × rψ
= a2 sin ψ Using formula 2.
Example 2.12 Evaluate the integral S zdσ. Let us evaluate rφ × rz .4) we obtain
2π cos φ+1 0
rz = k . 2π] and z ∈ [0. we have rφ = −i sin φ + j cos φ. Solution: Parametrization of the sphere is given in the example 2. z) = i cos φ + j sin φ + k z. 1 zdzdφ = 2
2π
zdσ =
S D
zdφdz =
0 2π
(cos φ + 1)2 dφ
0
=
1 2 1 = 2
1 1 + 2 cos φ + (1 + cos 2φ) dφ 2 0 2π 3 1 3 φ + 2 sin φ + sin 2φ = π.11:
= i cos φ + j sin φ. where S is a surface cut from the cylinder x2 + y 2 = 1 by the planes z = 0 and z = x + 1. rψ = i a cos φ cos ψ + j a sin φ cos ψ − k a sin ψ. Since 0 ≤ z ≤ x + 1 = cos φ + 1 the parameter domain D is given by φ ∈ [0.
Figure 2. a4 cos2 φ sin4 ψ + a4 sin2 φ sin4 ψ + a4 sin2 ψ cos2 ψ cos2 φ sin2 ψ + sin2 φ sin2 ψ + cos2 ψ = a2 sin ψ.4.11.
z = Mxy /M. and
Note that for axes x.
z ∈ [0.
Iz =
S
(x2 + y 2 )ρdσ. z) = x2 + y 2 respectively. and M=
S
φ ∈ [0. 3
2
45
. 1]. then moment of inertia and radius of gyration about the line are given by Iℓ =
S
r 2 ρdσ
and Rℓ =
Iℓ /M. y. z). rz = i cos φ + j sin φ + k .8) x2 + z 2 . ry (x.
Iy =
S
(x2 + z 2 )ρdσ. y = Mxz /M. (2.1
Applications
Mass. z) = z. y.6)
Coordinates of the center of mass. y.
Mxz =
S
yρdσ. √
(2. y. We have rφ = −i z sin φ + j z cos φ. y.2.7)
Moment of inertia and radius of gyration about a line. rφ × rz = cos φ sin φ 1 √ rφ × rz = z 2 cos2 φ + z 2 sin2 φ + z 2 = 2z. Mxy =
S
zρdσ. Solution: The standard parametrization of the cone is given by r(φ. If the density (mass per unit area) of a thin shell S is given by ρ(x. 2π]. x = Myz /M.5)
S
First moments about the coordinate planes.
Myz =
S
xρdσ.2. i j k −z sin φ z cos φ 0 = i z cos φ + j z sin φ − k z.9)
Example 2. z ∈ [0. z) = i z cos φ + j z sin φ + k z. 1] if its density is given by ρ(x.
zdσ =
√
1
2π 2 0
2
0
z dφdz = 2 2π
0
√
1
√ 2 2 z dz = π. z) to a line ℓ. z) denotes the distance from (x. z) =
(y 2 + z 2 )ρdσ.
(2. thus. y. y and z we have rx (x.13 Evaluate mass of the cone z 2 = x2 + y 2 . z) = rz (x. Ix =
S
y 2 + z 2 . y. (2. If r(x.
(2. then its mass M is M= ρdσ.
14 Determine center of mass of the hemisphere x2 + y 2 + z 2 = a2 . Assume that density is constant ρ.Example 2. Parametrization of the sphere is given in example 2. z ≥ 0.4.
S S
For the moment we have:
2π π/2
Mxy =
S
zρdσ = ρ
S 2π 0
zdσ = ρ
0 π/2 0 0
a3 sin ψ cos ψdψdφ
a3 ρ = 4 ﬁnally. Let us evaluate z. 2π] and ψ ∈ [0. π/2].11 we obtain that M= ρdσ = ρ dσ = 2πa2 ρ.
− cos 2ψ
πa3 ρ dφ = . Solution: Radial symmetry of S about zaxis implies that x = 0 and y = 0. using the result of example 2. M 4πa2 ρ 4
46
. 2
z=
πa3 ρ a Mxy = = . For the upper hemisphere we must assume that φ ∈ [0.