Tut 13
Tut 13
1 Recap
Parametrized surfaces
• Let 𝐷 be a bounded region in ℝ2 .
• Parametrized surface: a vector function 𝐷 → ℝ3 ,
#»
𝑟 (𝑢, 𝑣) = ⟨𝑥(𝑢, 𝑣), 𝑦(𝑢, 𝑣), 𝑧(𝑢, 𝑣)⟩,
one-to-one, and first order partial derivatives (e.g. #»
𝑟 𝑥 ) are bounded componentwise.
• Surface: the image of a parametrized surface.
• E.g. the graph of 𝑓 ∶ ℝ3 → ℝ has parametrization
𝑓(𝑢, 𝑣) = ⟨𝑢, 𝑣, 𝑓(𝑢, 𝑣)⟩
in the domain of 𝑓, so it is a surface.
• Normal vector of #»𝑟 at (𝑢0 , 𝑣0 ):
#» #»
#» , 𝑣 ) = 𝜕 𝑟 (𝑢 , 𝑣 ) × 𝜕 𝑟 (𝑢 , 𝑣 ),
𝑛(𝑢0 0
𝜕𝑢 0 0 𝜕𝑣 0 0
#» , 𝑣 ) ≠ #»
provided 𝑛(𝑢 0.
0 0
Area(𝑆) ≔ ∬ ‖
‖⟨1, 0, 𝑓𝑢 ⟩ × ⟨0, 1, 𝑓𝑣 ⟩‖
‖ d𝐴 = ∬ ‖
‖⟨−𝑓𝑢 , −𝑓𝑣 , 1⟩‖
2 2
‖ d𝐴 = ∬ √1 + (𝑓𝑢 ) + (𝑓𝑣 ) d𝐴 ,
𝐷 𝐷 𝐷
1
Surface integral
• Like line integrals, there are two types of surface integrals.
• For scalar functions and for vector fields.
#»
• Interpretation: the sum of effective quantities (those normal to the surface 𝑆) of 𝐹 over
the surface 𝑆.
• Independent of parametrization.
• Sign of integral depends on orientation, i.e. which of
𝜕 #»
𝑟 𝜕 #»𝑟
±( × ).
𝜕𝑢 𝜕𝑣
– Usually given by the problem.
– For closed surface, usually take outwards as positive.
2
Stokes’ theorem
• Stokes’ Theorem:
– Let 𝑆 be oriented, piecewise-smooth.
– 𝑆 bounded by a simple, closed, piecewise-smooth, positively oriented boundary curve
𝐶.
#»
– Let 𝐹 be 𝐶 1 componentwise.
– Then
#» #» #» #»
∮ 𝐹⋅ d 𝑟 = ∬ (∇ × 𝐹 )⋅ d 𝑆 .
𝐶 𝑆
Figure 1: Orientations of 𝐶 (Red), given the positive orientation of surface (Blue) is down-
wards (Green)
Interpretation of curl
• Interpretation of curl: the local circulation of a vector field at a fixed point.
• Different from global circulation.
• E.g. the vector field
#» 𝑦 𝑥
𝐹 =⟨ ,− , 0⟩
𝑥 2 + 𝑦2 𝑥 2 + 𝑦2
#» #»
in Fig. 2 appears to be clockwise, yet ∇ × 𝐹 = 0 except at the origin.
3
Figure 2: Vector field ⟨𝑦/(𝑥2 + 𝑦2 ), −𝑥/(𝑥2 + 𝑦2), 0⟩
• Reason: For a circle not containing the origin, the clockwise push is cancelled out by the
counterclockwise push.
4
2 Problems
1 (Ex 16.6, Q 22 & 24). Find a parametrization of
(a). The part of the cylinder 𝑥2 + 𝑦2 = 9 that lines above the 𝑥𝑦-plane and between the
planes 𝑦 = −4, 𝑦 = 4.
(b). The part of the ellipsoid 𝑥2 + 2𝑦2 + 3𝑧2 = 1 that lies to the left of the 𝑥𝑧-plane.
2 (Ex 16.6, Q 42). Find the area of the surface 𝑥 = 𝑧2 + 𝑦 that lies between the planes
𝑦 = 0, 𝑦 = 2, 𝑧 = 0 and 𝑧 = 2.
3 (Fitzpartick, Ex 20.2, Q 3). Find the surface area of the conical set
∬ (𝑥2 + 𝑦2 ) d𝑆 ,
𝑆
where
𝑆 = {(𝑥, 𝑦, 𝑧) ∶ 𝑥, 𝑦 > 0, 0 < 𝑧 < 3, 𝑧2 = 3(𝑥2 + 𝑦2 )}.
5
8 (Fitzpartick, Ex 20.3, Q 6). Verify Stokes’ Theorem for
#»
𝐹 = ⟨3𝑦, −𝑥𝑧, 𝑦𝑧2 ⟩,
2𝑧 = 𝑥2 + 𝑦2
10 (Ex 16.8, Q 20). Let 𝐶 be a simple, closed, smooth curve that lies in the plane
𝑥 + 𝑦 + 𝑧 = 1. Show that the line integral
∮ 𝑧 d𝑥 − 2𝑥 d𝑦 + 3𝑦 d𝑧
𝐶
depends only on the area of the region enclosed by 𝐶 and not on the shape of 𝐶 or its location
in the plane.
6
3 Solutions
1. (a). Although it is tempting to represent the polar coordinates to represent the cylinder,
note that this would take two variables, and we still need one to represent 𝑦. So three
variables are needed, and this approach does not work.
Instead, observe that one variable must be reserved to 𝑦, as it is independent of 𝑥 and 𝑧. To
represent the circle in the cylinder, we would need to use the dependencies on 𝑥 and 𝑧. For
example, we can take
𝑧 = √9 − 𝑥 2 ,
as the circle above the 𝑥𝑦-plane.
The parametrization thus is
#»
𝑟 (𝑢, 𝑣) = ⟨𝑢, 𝑣, √9 − 𝑢2 ⟩, −3 ≤ 𝑢 ≤ 3 and − 4 ≤ 𝑣 ≤ 4.
(b). It may not be apparent on how to parametrize a ellipsoid, yet we know how to parametrize
a sphere. So we can first transform it into a sphere then parametrize it.
Observe that if we let
𝑦 𝑧
(𝑎, 𝑏, 𝑐) = (𝑥, , ),
√2 √2
then the equation becomes
𝑎 2 + 𝑏 2 + 𝑐2 = 1,
which is the equation of a sphere.
To find a parametrization for a sphere, remember in the polar coordinates of triple integrals,
we use the triple (𝜌, 𝜃, 𝜙) to represent the whole ℝ3 . For a fixed sphere though, 𝜌 is fixed, so
it can be represented by two variables. In details, we have
0 ≤ 𝜃 ≤ 2𝜋 and 0 ≤ 𝜙 ≤ 𝜋.
However, in this case, we are asked to parametrize the part of the sphere left to the 𝑥𝑧-plane
only, meaning 𝑦 ≤ 0. To do so we need to restrict 𝜃. Therefore, by replacing (𝜃, 𝜙) by (𝑢, 𝑣),
the parametrization is
#» 1 1
𝑟 (𝑢, 𝑣) = ⟨sin 𝑣 cos 𝑢, sin 𝑣 sin 𝑢, cos 𝑣⟩, 𝜋 ≤ 𝑢 ≤ 2𝜋 and 0 ≤ 𝑣 ≤ 𝜋.
√2 √3
7
(a) Plot of (a) (b) Plot of (b)
arctan 2√2
Area(𝑆) = 2 ∫ sec3 𝜃 d𝜃 .
0
Figure 5: Plot of 𝑥 = 𝑧2 + 𝑦, 0 ≤ 𝑦, 𝑧 ≤ 2
ℎ 2 𝑥 2 + ℎ 2 𝑦2 + 𝑧 2 = ℎ 2 𝑟 2 ,
9
which 𝑆 in the 𝑎𝑏𝑐-space is a sphere of radius ℎ𝑟. A parametrization for the surface in the
𝑎𝑏𝑐-space is
#»
𝑠 (𝑢, 𝑣) = ⟨ℎ𝑟 sin 𝑣 cos 𝑢, ℎ𝑟 sin 𝑣 sin 𝑢, ℎ𝑟 cos 𝑣⟩,
so a parametrization for 𝑆 is
#»
𝑟 (𝑢, 𝑣) = ⟨𝑟 sin 𝑣 cos 𝑢, 𝑟 sin 𝑣 sin 𝑢, ℎ𝑟 cos 𝑣⟩.
Since we are only considering the portion of the ellipsoid in the first octant (𝑥, 𝑦, 𝑧 > 0),
𝜋 𝜋
0≤𝑢≤ and 0≤𝑣≤ .
2 2
𝜕 #»
𝑟 𝜕 #»
𝑟
= ⟨−𝑟 sin 𝑣 sin 𝑢, 𝑟 sin 𝑣 cos 𝑢, 0⟩ and = ⟨𝑟 cos 𝑣 cos 𝑢, 𝑟 cos 𝑣 sin 𝑢, −ℎ𝑟 sin 𝑣⟩.
𝜕𝑢 𝜕𝑣
Their cross product are
𝜕 #»
𝑟 𝜕 #»
𝑟 ⎡ ı⃗ ȷ⃗ k⃗ ⎤
× = det −𝑟 sin 𝑣 sin 𝑢 𝑟 sin 𝑣 cos 𝑢 0
𝜕𝑢 𝜕𝑣
⎣ 𝑟 cos 𝑣 cos 𝑢 𝑟 cos 𝑣 sin 𝑢 −𝑟ℎ sin 𝑣⎦
= ⟨−𝑟 2 ℎ sin2 𝑣 cos 𝑢, −𝑟 2 ℎ sin2 𝑣 sin 𝑣, −𝑟 2 sin 𝑣 cos 𝑣⟩,
and
‖
‖ 𝜕 #»
𝑟 𝜕 #»
𝑟‖ √ 4 2 4
‖ 𝜕𝑢 × 𝜕𝑣 ‖
4 2 2
‖ = 𝑟 ℎ sin 𝑣 cos2 𝑢 + 𝑟 4 ℎ2 sin 𝑣 sin 𝑢 + 𝑟 4 sin 𝑣 cos2 𝑣
‖ ‖
= 𝑟 2 sin 𝑣√ℎ2 sin2 𝑣 + cos2 𝑣
= 𝑟 2 sin 𝑣√ℎ2 + (1 − ℎ2 ) cos2 𝑣. (1)
This integral is of the form √𝑥2 ± 𝑎 2 , so we should use trigonometric substitution. However,
the complication is that depending on the sign of 1 − ℎ2 , the integral is in different form.
10
Case 1: 1 − ℎ2 > 0 ⟺ ℎ < 1.
In this case the integrand is √ℎ2 + (1 − ℎ2 )𝑡 2 , 1 − ℎ2 > 0, so we let
𝜋 𝜋 ℎ
𝑡√1 − ℎ2 = ℎ tan 𝜃, − <𝜃< ⟹ d𝑡 = sec2 𝜃 d𝜃 .
2 2 √1 − ℎ 2
Then
√1−ℎ2
arctan
𝜋 ℎ ℎ
Area(𝑆) = 𝑟 2 ∫ ℎ sec3 𝜃 d𝜃 .
2 √1 − ℎ 2 0
𝜋 2 ℎ 2 ⎛ √1 − ℎ 2 || 2 + √1 − ℎ2 ||⎞
= 𝑟 + ln || || (See Fig. 6a)
4 √1 − ℎ2 ⎝ ℎ2 | ℎ |⎠
𝜋 ⎛ ℎ2 1 + √1 − ℎ 2 ⎞
= 𝑟2 1 + ln .
4 ⎝ √1 − ℎ2 ℎ ⎠
𝜋 𝜋 ℎ
𝑡√ℎ2 − 1 = ℎ sin 𝜃, − ≤𝜃≤ ⟹ d𝑡 = cos 𝜃 d𝜃 .
2 2 √ℎ 2 − 1
Then
√ℎ2 −1
arcsin
𝜋 ℎ ℎ
Area(𝑆) = 𝑟 2 ℎ cos2 𝜃 d𝜃
2 √ℎ2 − 1 ∫0
√ℎ2 −1
arcsin
𝜋 ℎ2 ℎ 1 + cos 2𝜃
= 𝑟2 ∫ d𝜃
2 √ℎ 2 − 1 0 2
√ℎ2 −1
𝜋 ℎ2 sin 2𝜃 arcsin ℎ
= 𝑟2 𝜃 +
4 √ℎ 2 − 1 [ 2 ]0
𝜋 ℎ2 ⎛ √ℎ 2 − 1 √ℎ 2 − 1 ⎞
= 𝑟2 arcsin + (See Fig. 6b)
4 √ℎ 2 − 1 ⎝ ℎ ℎ2 ⎠
𝜋 ⎛ ℎ2 √ℎ 2 − 1 ⎞
= 𝑟2 1 + arcsin .
4 ⎝ √ℎ 2 − 1 ℎ ⎠
11
Case 3: ℎ = 1.
In this case, the integrand is just √ℎ2 = ℎ = 1, so
𝜋
Area(𝑆) = 𝑟 2 .
2
Note that in this case, 𝑆 is just the sphere of radius 𝑟, so this value, which is one-eighth of
the surface area of the sphere, makes sense.
In conclusion,
𝜋 2
⎧
⎪ 𝑟 if ℎ=1
⎪
⎪ 2
⎪
⎪
⎪
⎪
⎪
⎪ 𝜋 2⎛ ℎ2 1 + √1 − ℎ 2 ⎞
𝑟 1+ ln if ℎ>1
Area(𝑆) = ⎨ 4 ⎝ √ 1 − ℎ 2 ℎ ⎠
⎪
⎪
⎪
⎪
⎪
⎪ 𝜋 ⎛ ℎ2 √ℎ 2 − 1 ⎞
⎪
⎪
⎪ 𝑟2 1 + arcsin if ℎ < 1.
⎩4 ⎝ √ℎ 2 − 1 ℎ ⎠
ℎ ℎ
1 √ℎ2 − 1
𝜃 𝜃
√ℎ 2 − 1 1
(a) Case 1 (b) Case 2
12
Proceeding in this way, you should obtain an integral of the form
ℎ2 ∞ 1
∫ d𝑠 .
4 0 (𝑠 + ℎ2 − 1)2
2
Nonetheless, you still need to separately consider the cases ℎ > 1 and ℎ < 1.
Another thing to note is that for the case ℎ > 1, you may obtain
𝜋 ℎ 2 ⎛ √ℎ 2 − 1 𝜋 1 ⎞
Area(𝑆) = + − arctan .
4 √ℎ 2 − 1 ⎝ ℎ 2 2 √ℎ 2 − 1 ⎠
Even though this form is different form what we have shown, they are indeed the same. This
is due to the following identities
1 𝜋 𝑥
arctan = − arctan 𝑥 and arctan 𝑥 = arcsin .
𝑥 2 √1 + 𝑥 2
Remark 2. As observe from this example, the formula for general ellipsoid is complicated
in a sense that it is case-dependent and contains inverse trigonometric functions. An
approximation formula, given by Thomsen (2004), for general ellipsoid
𝑥2 𝑦2 𝑧2
+ + =1
𝑎 2 𝑏 2 𝑐2
is
𝑟 𝑎 𝑝 𝑏 𝑝 + 𝑏 𝑝 𝑐𝑝 + 𝑎 𝑝 𝑐𝑝
Area(𝑆) ≈ 4𝜋√ , 𝑝 ≈ 1.6075.
3
In our problem,
1 1
𝑎=𝑏= and 𝑐 = 2 2 ,
𝑟 2 ℎ𝑟
so the approximation of our desired area (need to divide by 8) is
𝜋√ 𝑝
𝑝 1 + 2𝑥
.
2 3
By plotting the exact and the approximation formula in Fig. 8, we can see that the two
formulae give very close results.
In addition, as ℎ → 0+ , the area approaches 𝜋𝑟 2 /4, since the surface is getting closer to two
ellipses, which the area of an ellipse is 𝜋𝑎𝑏.
𝑦
𝜋
2
𝜋
4
𝑥
0 1
13
4. This is a surface integral of a scalar function.
Step 1: Find parametrization.
Since we are working with 𝑧 > 0, we can take
𝑧 = 𝑓(𝑥, 𝑦) = √3√𝑥2 + 𝑦2 .
So a simple parametrization is
#»
𝑟 (𝑥, 𝑦) = ⟨𝑥, 𝑦, √3√𝑥2 + 𝑦2 ⟩.
‖
‖ 𝜕 #»
𝑟 𝜕 #»
𝑟‖ 2
‖ 𝜕𝑥 × 𝜕𝑦 ‖
2
‖ = √1 + (𝑓𝑥 ) + (𝑓𝑦 )
‖ ‖
√
√ ⎛ √3𝑥 ⎞2 ⎛ √3𝑦 ⎞2
= √1 + +
√ ⎝ √𝑥 2 + 𝑦 2 ⎠ ⎝ √𝑥 2 + 𝑦 2 ⎠
3𝑥2 + 3𝑦2
= √1 +
𝑥 2 + 𝑦2
= 2.
d𝑥 d𝑦 = 𝑟 d𝑟 d𝜃 ,
hence
𝜋/2 √3
∬ 𝑥 2 + 𝑦2 d𝑆 = 2 ∬ 𝑥 2 + 𝑦2 d𝐴 = 2 ∫ ∫ 𝑟 2 ⋅ 𝑟 d𝑟 d𝜃
𝑆 𝐷 0 0
1 4 √3
= 𝜋[ 𝑟 ]
4 0
9𝜋
= ,
4
14
as required.
𝜕 #»
𝑟 𝜕 #»
𝑟
= ⟨2𝑣, 2𝑢, 2𝑢⟩ and = ⟨2𝑢, −2𝑣, 2𝑣⟩.
𝜕𝑢 𝜕𝑣
Thus
𝜕 #»
𝑟 𝜕 #»
𝑟 ⎡ ı⃗ ȷ⃗ k⃗ ⎤
× = det 2𝑣 2𝑢 2𝑢
𝜕𝑢 𝜕𝑣
⎣2𝑢 −2𝑣 2𝑣 ⎦
= ⟨8𝑢𝑣, 4(𝑢2 − 𝑣 2 ), −4(𝑢2 + 𝑣 2 )⟩,
and
‖
‖ 𝜕 #»
𝑟 𝜕 #»
𝑟‖
‖ 𝜕𝑢 × 𝜕𝑣 ‖‖ = √64𝑢2 𝑣 2 + 16(𝑢2 − 𝑣 2 )2 + 16(𝑢2 + 𝑣 2 )2
‖ ‖
= 4√4𝑢2 𝑣 2 + (𝑢4 − 2𝑢2 𝑣 2 + 𝑣 4 ) + (𝑢4 + 2𝑢2 𝑣 2 + 𝑣 4 )
= 4√2𝑢4 + 4𝑢2 𝑣 2 + 2𝑣 4
= 4√2(𝑢2 + 𝑣 2 ).
15
Step 3: Evaluate the integral.
After replacing the integrand by the parametrization, the surface integral is
2 + 𝑦2 ) d𝑆 = 2 ‖ #» #» ‖
2 − 𝑣 2 )2 ]‖ 𝜕 𝑟 × 𝜕 𝑟 ‖ d𝐴
∬ (𝑥 ∬ [(2𝑢𝑣) + (𝑢 ‖ 𝜕𝑢 𝜕𝑣 ‖
𝑆 𝑆 ‖ ‖
3
= 4√2 ∬ (𝑢2 + 𝑣 2 ) d𝐴 .
𝑆
It is apparent that polar coordinates is the easier approach. The domain of integration is
𝑢2 + 𝑣 2 ≤ 1 ⟹ 𝑟 ≤ 1.
16
Step 1: Find a parametrization.
The surface we are integrating over is
𝑧 = √𝑥2 + 𝑦2 ,
𝜕 #»
𝑟 𝑥 𝜕 #»
𝑟 𝑦
= ⟨1, 0, ⟩ and = ⟨0, 1, ⟩.
𝜕𝑥 √𝑥 2 + 𝑦 2 𝜕𝑦 √𝑥 2 + 𝑦 2
Their cross product is
⎡ ı⃗ ȷ⃗ k⃗ ⎤
#»
𝜕𝑟 𝜕𝑟 #» 1 0 𝑥
𝑥 𝑦
× = det ⎢
⎢ ⎥
√𝑥2 +𝑦2 ⎥ = ⟨− ,− , 1⟩. (3)
𝜕𝑥 𝜕𝑦 ⎢ 𝑦 ⎥ √𝑥 2 + 𝑦 2 √𝑥 2 + 𝑦 2
0 1
⎣ √𝑥2 +𝑦2 ⎦
Different from the integral of scalar functions, we only need the vector, not its norm. Another
issue is the orientation. We are specifically asked to consider the downward orientation, yet
the vector in Eq. (3) is pointing upward as the 𝑧-component is postive. Hence we shall take
𝜕 #»
𝑟 𝜕 #»
𝑟 𝑥 𝑦
× =⟨ , , −1⟩,
𝜕𝑦 𝜕𝑥 √𝑥2 + 𝑦2 √𝑥2 + 𝑦2
as the normal vector.
Step 3: Evaluate the integral.
The integrand is
#» 𝜕 #»
𝑟 𝜕 #»
𝑟 𝑥 𝑦
𝐹 ( #»
𝑟 (𝑥, 𝑦)) ⋅ ( × ) = ⟨−𝑥, −𝑦, (𝑥2 + 𝑦2 )3/2 ⟩ ⋅ ⟨ , , −1⟩
𝜕𝑦 𝜕𝑥 √𝑥 2 + 𝑦 2 √𝑥 2 + 𝑦 2
−(𝑥2 + 𝑦2 ) − (𝑥2 + 𝑦2 )2
= .
√𝑥 2 + 𝑦 2
Again, doing integral in polar coordinates simplifies the calculation. The domain of integra-
tion in Eq. (2) becomes
1 ≤ 𝑟 ≤ 3.
17
Hence
2
#» #» (𝑥2 + 𝑦2 ) + (𝑥2 + 𝑦2 )
∬ 𝐹 ⋅ 𝑆 d𝑆 = − ∬ d𝑥 d𝑦
𝑆 𝐷 √𝑥 2 + 𝑦 2
2𝜋 3 2
𝑟 + 𝑟4
= −∫ ∫ ⋅ 𝑟 d𝑟 d𝜃
0 1 𝑟
1 1 3
= −2𝜋[ 𝑟 3 + 𝑟 5 ]
3 5 1
35 1 1
= −2𝜋(9 + − − )
5 3 5
1712
=− 𝜋.
15
#» #»
(a) Without vector field 𝐹 (b) With vector field 𝐹
7. This is a surface integral of a vector field. The complication here is that the thetrahedron
has 4 faces, and to find the total integral we need to consider each face then add the integrals
over each of them up.
In addition, since the surface is a boundary of a solid, the outwards directions are considered
as the positive orientation.
In the following I will not list out the steps, but mention which surface are being integrated.
One should be reminded that the procedures for calculating surface integrals are
1. Find a parametrization;
2. Calculate the normal vector;
3. Evaluate the integral.
Surface 𝑆1 : (0, 0, 0), (1, 0, 0) and (0, 1, 0).
This surface is a triangle on the 𝑥𝑦-plane. A parametrization is
Then
𝜕 𝑟#»1 𝜕 𝑟#»1
= ⟨1, 0, 0⟩ and = ⟨0, 1, 0⟩,
𝜕𝑥 𝜕𝑦
𝜕 𝑟#» 𝜕 𝑟#»
⟹ 1 × 2 = ⟨0, 0, 1⟩.
𝜕𝑥 𝜕𝑦
18
But this is upwards-pointing, so we take
𝜕 𝑟#»1 𝜕 𝑟#»1
× = ⟨0, 0, −1⟩.
𝜕𝑦 𝜕𝑥
The integral is
#» #» 𝜕 𝑟#»1 𝜕 𝑟#»1
∬ 𝐹( 𝑟1 ) ⋅ ( 𝜕𝑦 × 𝜕𝑥 ) d𝐴 = ∬ ⟨𝑦, −𝑦, 𝑥⟩ ⋅ ⟨0, 0, −1⟩ d𝐴
𝑆1 𝑆1
1 1−𝑥
=∫ ∫ −𝑥 d𝑦 d𝑥
0 0
1
= ∫ −𝑥 + 𝑥2 d𝑥
0
1
1 1
= [− 𝑥2 + 𝑥3 ]
2 3 0
1
=− .
6
19
Then
𝜕 𝑟#»3 𝜕 𝑟#»3
= ⟨1, 0, 0⟩ and = ⟨0, 0, 1⟩,
𝜕𝑥 𝜕𝑧
𝜕 𝑟#» 𝜕 𝑟#»
⟹ 3 × 3 = ⟨0, −1, 0⟩.
𝜕𝑥 𝜕𝑧
This is outwards-pointing, as required. Thus the integral is
#» #» 𝜕 #»𝑟 𝜕 #»
𝑟
∬ 𝐹( 𝑟3 ) ⋅ ( 𝜕𝑥 × 𝜕𝑧 ) d𝐴 = ∬ ⟨0, 𝑧, 𝑥⟩ ⋅ ⟨0, −1, 0⟩ d𝐴
𝑆3 𝑆3
1 1−𝑧
=∫ ∫ −𝑧 d𝑥 d𝑧
0 0
1
= ∫ −𝑧 + 𝑧2 d𝑧
0
1 1 1
= [− 𝑧2 + 𝑧3 ]
2 3 0
1
=− .
6
Then
𝜕 𝑟#»4 𝜕 𝑟#»4
= ⟨1, 0, −1⟩ and = ⟨0, 1, −1⟩,
𝜕𝑥 𝜕𝑦
𝜕 𝑟#» 𝜕 𝑟#»
⟹ 4 × 4 = ⟨1, 1, 1⟩,
𝜕𝑥 𝜕𝑦
which is outwards-pointing. The domain of integration is the triangle on the 𝑥𝑦-plane
0≤𝑦≤1 and 0 ≤ 𝑥 ≤ 1 − 𝑦.
Thus the integral is
#» #» 𝜕 𝑟#»4 𝜕 𝑟#»4
∬ 𝐹( 𝑟4 ) ⋅ ( × ) d𝐴 = ∬ ⟨𝑦, 1 − 𝑥 − 2𝑦, 𝑥⟩ ⋅ ⟨1, 1, 1⟩ d𝐴
𝑆4 𝜕𝑥 𝜕𝑦 𝑆4
1 1−𝑦
=∫ ∫ 1 − 𝑦 d𝑥 d𝑦
0 0
1
= ∫ (1 − 𝑦)2 d𝑦
0
1 3 1
=[ 𝑦 − 𝑦2 + 𝑦
3 ]0
5
= .
6
20
By adding up the integral over each of the surfaces,
#» #» 4 #» #» 1 1 1 5 1
∬ 𝐹⋅ d 𝑆 = ∑ ∬ 𝐹 ⋅ d 𝑆 = − 6 − 6 − 6 + 6 = 3 .
𝑆 𝑖=1 𝑆𝑖
Remark 4. Although I presented the solution in a way that the domain of the parametriza-
tion is determined before integration, it should be determined by whichever order is easier
to compute.
8. The Stokes’ Theorem tells us that for 𝑆 an oriented, piecewise-smooth surface, with
#»
a simple, closed, piecewise-smooth boundary 𝐶 with positive orientation. Then if 𝐹 is a
vector field having continuous partial derivatives on an open set containing 𝑆, then
#» #» #» #»
∫ 𝐹⋅ d 𝑟 = ∬ (∇ × 𝐹 )⋅ d 𝑆 .
𝐶 𝑆
#»
For 𝐹 = ⟨3𝑦, −𝑥𝑧, 𝑦𝑧2 ⟩ and 𝑆 a paraboloid, the requirements are satisfied, so Stokes’ Theo-
rem should apply.
Since the orientation is not explicitly given, we will take the upwards-pointing direction
as the positive orientation, and so the boundary has the counterclockwise direction as
positive.
Surface integral.
Step 1: Find a parametrization.
We can easily rewrite 𝑧 as a graph of 𝑥, 𝑦, so we can take
𝑥 2 + 𝑦2 𝑥 2 + 𝑦2
𝑟#»1 (𝑥, 𝑦) = ⟨𝑥, 𝑦, ⟩, 0≤𝑧= ≤ 2.
2 2
21
This normal vector is upwards-pointing, as desired.
Step 3: Compute the curl.
The given vector field is
#»
𝐹 = ⟨3𝑦, −𝑥𝑧, 𝑦𝑧2 ⟩.
Thus
⎡ ı⃗ ȷ⃗ k⃗ ⎤
#» 𝜕 𝜕 𝜕
∇ × 𝐹 = det ⎢ 𝜕𝑥 𝜕𝑦 𝜕𝑘 ⎥
= ⟨𝑥 + 𝑧2 , 0, −𝑧 − 3⟩.
⎣3𝑦 −𝑥𝑧 𝑦𝑧2 ⎦
Before working on the integral, notice that the integration domain is a circle, so working in
polar coordinates is easier. The domain in polar coordinates is
𝑟2
0≤ ≤2 ⟺ 0 ≤ 𝑟 ≤ 2.
2
#» #» #» 𝜕 #» 𝑟 𝜕 #»
𝑟
∬ (∇ × 𝐹 )⋅ d 𝑆 = ∬ (∇ × 𝐹 ) ⋅ ( 𝜕𝑥 × 𝜕𝑦 ) d𝐴
𝑆 𝑆
2
𝑥 2 + 𝑦2 𝑥 2 + 𝑦2
= ∬ ⟨𝑥 + ( ) , 0, − − 3⟩ ⋅ ⟨−𝑥, −𝑦, 1⟩ d𝐴
𝑆 2 2
2
𝑥 2 + 𝑦2 𝑥 2 + 𝑦2
=∬ −𝑥2 − 𝑥( ) − − 3 d𝐴
𝑆 2 2
2𝜋 2
𝑟5 𝑟2
= − ∫ ∫ (𝑟 2 cos2 𝜃 + cos 𝜃 + + 3)𝑟 d𝑟 d𝜃
0 0 4 2
2𝜋 2
𝑟6 𝑟3
= − ∫ ∫ (𝑟 3 cos2 𝜃 + cos 𝜃 + + 3𝑟) d𝑟 d𝜃
0 0 4 2
2𝜋 2
1 𝑟7 𝑟4 3
= − ∫ [ 𝑟 4 cos2 𝜃 + cos 𝜃 + + 𝑟 2 ] d𝜃
0 4 4⋅7 8 2 0
2𝜋
32
= − ∫ 4 cos2 𝜃 + cos 𝜃 + 2 + 6 d𝜃
0 7
2𝜋
32
= − ∫ 2(1 + cos 2𝜃) + cos 𝜃 + 8 d𝜃
0 7
2𝜋
32
= −[10𝜃 + sin 2𝜃 + sin 𝜃]
7 0
= −20𝜋.
22
Line integral.
Step 1: Find a parametrization.
It is clear that the boundary of the surface is the circle 𝑟 = 2. If we want the line to traverse
in the counterclockwise direction, we can take
In conclusion, we verified the conclusion of Stokes’ Theorem indeed holds in this case.
#» #»
(a) Without ∇ × 𝐹 (b) With ∇ × 𝐹
23
(a) Viewed from first octant (b) Viewed from top
#»
Figure 14: Plot of boundary and vector field 𝐹
𝜕 𝑟#»1 𝜕 𝑟#»1
= ⟨1, 0, 2𝑥⟩ and = ⟨0, 1, 2𝑦⟩,
𝜕𝑥 𝜕𝑦
𝜕 𝑟#» 𝜕 𝑟#»
⟹ 1 × 1 = ⟨−2𝑥, −2𝑦, 1⟩.
𝜕𝑥 𝜕𝑦
This is upwards-pointing, as desired.
#»
Step 3: Compute the curl of 𝐹 .
Remember
#»
𝐹(𝑥, 𝑦, 𝑧) = ⟨𝑥3 − 𝑧, 𝑥𝑦, 𝑦 + 𝑧2 ⟩.
Its curl is given by
⎡ ı⃗ ȷ⃗ k⃗ ⎤
#» 𝜕 𝜕 𝜕
∇ × 𝐹 = det ⎢ 𝜕𝑥 𝜕𝑦 𝜕𝑧 ⎥ = ⟨1, −1, 𝑦⟩.
⎣𝑥 − 𝑧 𝑥𝑦 𝑦 + 𝑧 ⎦
3 2
#» #» 𝜕 𝑟#»1 𝜕 𝑟#»1
∬ (∇ × 𝐹( 𝑟1 )) ⋅ ( 𝜕𝑥 × 𝜕𝑦 ) d𝐴 = ∬ ⟨1, −1, 𝑦⟩ ⋅ ⟨−2𝑥, −2𝑦, 1⟩ d𝐴
𝑆 𝑆
= ∬ −2𝑥 + 2𝑦 + 𝑦 d𝐴
𝑆
= ∬ −2𝑥 + 3𝑦 d𝐴
𝑆
𝜋/2 cos 𝜃
=∫ ∫ (−2𝑟 cos 𝜃 + 3𝑟 sin 𝜃) ⋅ 𝑟 d𝑟 d𝜃
−𝜋/2 0
𝜋/2
1 3 cos 𝜃
=∫ [ 𝑟 ]0 ⋅ (−2 cos 𝜃 + 3 sin 𝜃) d𝜃
−𝜋/2 3
𝜋/2
2
= ∫ − cos4 𝜃 + sin 𝜃 cos3 𝜃 d𝜃
−𝜋/2 3
2 𝜋/2
= − ∫ cos4 𝜃 d𝜃 ,
3 −𝜋/2
1 + cos 2𝜃 2
cos4 𝜃 = ( )
2
1
= (1 + 2 cos 2𝜃 + cos2 2𝜃)
4
1 1 + cos 4𝜃
= [1 + 2 cos 2𝜃 + ]
4 2
1 3 1
= ( + 2 cos 2𝜃 + cos 4𝜃).
4 2 2
25
Substituting this back to the integral,
𝜕 𝑟#»2 𝜕 𝑟#»2
= ⟨1, 0, 1⟩ and = ⟨0, 1, 1⟩,
𝜕𝑥 𝜕𝑦
𝜕 𝑟#» 𝜕 𝑟#»
⟹ 2 × 2 = ⟨−1, 0, 1⟩.
𝜕𝑥 𝜕𝑦
26
The integral is
#» #» 𝜕 𝑟#»2 𝜕 𝑟#»2
∬ (∇ × 𝐹( 𝑟2 )) ⋅ ( 𝜕𝑥 × 𝜕𝑦 ) d𝐴 = ∬ ⟨1, −1, 𝑦⟩ ⋅ ⟨−1, 0, 1⟩ d𝐴
𝑆 𝑆
= ∬ −1 + 𝑦 d𝐴
𝑆
𝜋/2 cos 𝜃
=∫ ∫ (−1 + 𝑟 sin 𝜃) ⋅ 𝑟 d𝑟 d𝜃
−𝜋/2 0
𝜋/2 cos 𝜃
𝑟2 1 3
=∫ [ 2 + 3 𝑟 sin 𝜃]0 d𝜃
−
−𝜋/2
𝜋/2
1
= ∫ − cos2 𝜃 + sin 𝜃 cos3 𝜃 d𝜃
−𝜋/2 2
1 𝜋/2 1 + cos 2𝜃
=− ∫ d𝜃 , (sin 𝜃 cos3 𝜃 is odd)
2 −𝜋/2 2
1 sin 2𝜃 𝜋/2
= − [𝜃 +
4 2 ]−𝜋/2
𝜋
=− .
4
In general, Stokes’ Theorem tells us that when doing surface integral of the curl of a vector
field, we can consider another surface, as long as this and the original surface shares the
same boundary and orientation. It facilitates us to consider a surface that is easier the
integrate over. However, it does not assert such a relation for general vector fields.
Method 3: Compute the line integral directly.
Step 1: Find a parametrization of the line.
We have shown that the intersection of the plane is the polar equation
𝑟 = cos 𝜃.
Using this on the polar coordinates for 𝑥, 𝑦 and 𝑧 = 𝑥, we obtain a parametrization of the
line
1
𝑟#»3 (𝑡) = ⟨𝑟 cos 𝑡, 𝑟 sin 𝑡, 𝑟 cos 𝑡⟩ = ⟨cos2 𝑡, sin 2𝑡, cos2 𝑡⟩,
2
with
𝜋 𝜋
− ≤𝑡≤ .
2 2
You can check that the orientation of this parametrization is counterclockwise.
Step 2: Calculate the line element.
Taking derivative,
27
Step 3: Evaluate the line integral.
Recall that
#»
𝐹 = ⟨𝑥3 − 𝑧, 𝑥𝑦, 𝑦 + 𝑧2 ⟩,
so under the parametrization 𝑟#»(𝑡), 3
#» 1 1
𝐹 ( #»
𝑟 3 (𝑡)) = ⟨cos6 𝑡 − cos2 𝑡, sin 2𝑡 cos2 𝑡, sin 2𝑡 + cos4 𝑡⟩.
2 2
The integrand is
#»
𝐹 ( 𝑟#»(𝑡)) ⋅ 𝑟#»′ (𝑡)
3 3
1 1
= ⟨cos6 𝑡 − cos2 𝑡, sin 2𝑡 cos2 𝑡, sin 2𝑡 + cos4 𝑡⟩ ⋅ ⟨− sin 2𝑡, cos 2𝑡, − sin 2𝑡⟩
2 2
1 1 1
= − sin 2𝑡(cos6 𝑡 − cos2 𝑡 + cos2 𝑡 cos 2𝑡 − cos4 𝑡) − sin2 2𝑡.
2 2 2
However, notice that the first term is odd.
The integral thus is,
𝜋/2
#» #» #» #» #»
∮ 𝐹⋅ d 𝑟 = ∫ 𝐹 ( 𝑟3 (𝑡)) ⋅ 𝑟3 ′ (𝑡) d𝑡
𝐶 −𝜋/2
𝜋/2
1 1 1
=∫ − sin 2𝑡(cos6 𝑡 − cos2 𝑡 + cos2 𝑡 cos 2𝑡 − cos4 𝑡) − sin2 2𝑡 d𝑡
−𝜋/2 2 2 2
𝜋/2
1
= − ∫ sin2 2𝑡 d𝑡 (The first term is odd)
2 −𝜋/2
1 𝜋/2
= − ∫ 1 − cos 4𝑡 d𝑡
4 −𝜋/2
𝜋/2
1 1
= − [𝑡 − sin 4𝑡]
4 4 −𝜋/2
𝜋
=− .
4
In conclusion, all three methods gives the same value, which is guaranteed by Stokes’
Theorem.
10. The desciption of the problem already suggested us using Stokes’ Theorem, as we want
a double integral to calculate area.
Remember that the given line integral is equivalent to
#»
∫ ⟨𝑧, −2𝑥, 3𝑦⟩⋅ d 𝑟 ,
𝐶
⎡ ı⃗ ȷ⃗ k⃗ ⎤
#» 𝜕 𝜕 𝜕
∇ × 𝐹 = det ⎢ 𝜕𝑥 𝜕𝑦 𝜕𝑧 ⎥ = ⟨3, 1, −2⟩.
⎣ 𝑧 −2𝑥 3𝑦⎦
28
The surface has a parametrization
#»
𝑟 (𝑥, 𝑦) = ⟨𝑥, 𝑦, 1 − 𝑥 − 𝑦⟩,
so
𝜕 #»
𝑟 𝜕 #»
𝑟
= ⟨1, 0, −1⟩ and = ⟨0, 1, −1⟩,
𝜕𝑥 𝜕𝑦
𝜕 #»
𝑟 𝜕 #»
𝑟
⟹ × = ⟨1, 1, 1⟩.
𝜕𝑥 𝜕𝑦
Since no instructions on the orientation is given, we just take the direction given by the
above normal vector as positive.
By Stokes’ Theorem,
#» 𝜕 #» 𝑟 𝜕 #»
𝑟
∫ 𝑧 d𝑥 − 2𝑥 d𝑦 + 3𝑦 d𝑧 = ∬ (∇ × 𝐹 ) ⋅ ( 𝜕𝑥 × 𝜕𝑦 ) d𝐴
𝐶 𝑆
= 3 ∬ d𝐴 .
𝑆
This shows that the given line integral depends only on the area of the region bounded by 𝐶.
4 Demos
Interactive 3D plots created in Math3D are availbale below.
Recap
• Orientation of surface and boundary
#»
• Vector field with curl 0
Problems
1(a). https://www.math3d.org/PvS3VxIfT
1(b). https://www.math3d.org/6dkT5CK7Y
2. https://www.math3d.org/l7csRHcEq
3. https://www.math3d.org/2rr25RSox
4. https://www.math3d.org/s4xnNQYfr
5. https://www.math3d.org/tw3wSPnzq
6. https://www.math3d.org/5Vx8AAxk1
7. https://www.math3d.org/E9JJhbNPX
8. https://www.math3d.org/pUv7wAHhL
9. https://www.math3d.org/keWJdPmpM
29