1
00:00:03,350 --> 00:00:10,350
What do rain, bees, and vectors have in common?
They're all ways that MIT physicists visualize
2
00:00:10,600 --> 00:00:16,160
the concept of "flux." In this video, we'll
explore flux as it relates to Gauss' law.
3
00:00:16,160 --> 00:00:20,480
If you've ever had trouble choosing the right
Gaussian surface, getting a quantity out of
4
00:00:20,480 --> 00:00:25,970
an integral, or just wanted to know what flux
is, get ready: this video is for you.
5
00:00:25,970 --> 00:00:29,330
This video is part of the Derivatives and
Integrals video series.
6
00:00:29,330 --> 00:00:34,190
Derivatives and integrals are used to analyze
the properties of a system. Derivatives describe
7
00:00:34,190 --> 00:00:38,999
local properties of systems, and integrals
quantify their cumulative properties.
8
00:00:38,999 --> 00:00:44,249
"Hello. My name is Peter Fisher. I am a professor
in the physics department at MIT, and today
9
00:00:44,249 --> 00:00:48,239
I'll be talking with you about electric flux
and Gauss' Law.
10
00:00:48,239 --> 00:00:53,149
By this time in your course you've seen and
used Gauss' Law quite a bit. You have a good
11
00:00:53,149 --> 00:00:57,899
handle on the electric field, what it does
and how it works. We will also assume that
12
00:00:57,899 --> 00:01:02,559
you know how to take an area integral, a skill
that is important to Gauss' Law but which
13
00:01:02,559 --> 00:01:04,510
is not taught in this video.
14
00:01:04,510 --> 00:01:09,180
Our end goal is to improve your ability to
use Gauss' Law. To do this, we'll help you
15
00:01:09,180 --> 00:01:14,030
develop an organized view of electric flux
and what goes into it. We'll also spend a
16
00:01:14,030 --> 00:01:19,549
lot of time talking about symmetry and how
to use it to your advantage.
17
00:01:19,549 --> 00:01:24,020
We're going to start by thinking about what
goes into flux and the interrelationships
18
00:01:24,020 --> 00:01:29,049
between these quantities, so that we can use
flux more effectively.
19
00:01:29,049 --> 00:01:33,509
As a refresher, here are the two equations
we'll be working with today: the definition
20
00:01:33,509 --> 00:01:38,729
of flux and Gauss' Law. You may have seen
these written with a double integral, or with
21
00:01:38,729 --> 00:01:43,320
slightly different notation, so take a quick
look to make sure you're familiar with all
22
00:01:43,320 --> 00:01:46,770
the pieces.
23
00:01:46,770 --> 00:01:51,609
To really use symmetry to our best advantage
in Gauss' Law, we need to understand a few
24
00:01:51,609 --> 00:01:57,810
more basic things. We need to be able to use
vector areas, to categorize charge distributions,
25
00:01:57,810 --> 00:02:02,109
and to determine when and how the two can
work together.
26
00:02:02,109 --> 00:02:07,359
The underlying reason we want to use symmetry
is that integrals are difficult, significantly
27
00:02:07,359 --> 00:02:12,050
more difficult than derivatives. There are
many situations in which we cannot determine
28
00:02:12,050 --> 00:02:17,140
the flux integral. Therefore, we seek to simplify.
29
00:02:17,140 --> 00:02:21,450
The integrals we're dealing with are surface
integrals, also called area integrals. There
30
00:02:21,450 --> 00:02:26,970
are two basic kinds, open and closed, as you
can see on the screen. An open surface uses
31
00:02:26,970 --> 00:02:32,440
an open integral, and a closed surface uses
a closed integral. Flux can be defined for
32
00:02:32,440 --> 00:02:37,000
any surface, but Gauss' Law uses only closed
integrals.
33
00:02:37,000 --> 00:02:43,220
When we take the integral, we use dA, a tiny
piece of the area. dA is a vector piece of
34
00:02:43,220 --> 00:02:49,500
the area, so it has a direction to it as well.
dA always points perpendicular to the surface.
35
00:02:49,500 --> 00:02:54,740
With closed surfaces we usually choose dA
to point outward from the surface.
36
00:02:54,740 --> 00:03:01,740
dA usually gets represented as two other differentials,
such as dx times dy. Which two depends on
37
00:03:02,060 --> 00:03:07,840
the coordinate system you choose, with some
being more complex than others. Sometimes
38
00:03:07,840 --> 00:03:14,120
we will need these, so it's good to have them.
The good part is that if we use symmetry correctly,
39
00:03:14,120 --> 00:03:18,970
we can avoid having to do an integral at all.
40
00:03:18,970 --> 00:03:25,090
Let's pursue our goal of not doing an integral.
We have this integral of E dot dA. Let's write
41
00:03:25,090 --> 00:03:30,470
the dot product as a cosine. Under certain
circumstances, we can remove the E from the
42
00:03:30,470 --> 00:03:36,200
integral. We can only do that when our electric
field has uniform magnitude over the entire
43
00:03:36,200 --> 00:03:43,000
surface. You can see some examples on the
bottom of the screen.
44
00:03:43,000 --> 00:03:48,230
We can remove the cosine(theta) from our integral
only when the angle between the area and the
45
00:03:48,230 --> 00:03:53,340
electric field doesn't change. The left-hand
example shows a field that always points in
46
00:03:53,340 --> 00:03:58,829
the same direction, and an area that always
points in the same direction. The middle example
47
00:03:58,829 --> 00:04:03,800
has a field whose direction changes, but it
always points outward, just like the area
48
00:04:03,800 --> 00:04:10,390
vector from the cylinder points outward. It
still works. In the right-hand example, the
49
00:04:10,390 --> 00:04:15,800
field has a constant direction, but the area
vector does not. The angle between them changes,
50
00:04:15,800 --> 00:04:19,370
so we couldn't remove the cosine from the
integral.
51
00:04:19,370 --> 00:04:26,350
If we want to move both the field and the
cosine out of the integral, we need to fulfill
52
00:04:26,350 --> 00:04:32,470
both conditions. The left-hand and center
examples only fill one condition each. The
53
00:04:32,470 --> 00:04:38,500
example on the right has both, and would be
ideal to use.
54
00:04:38,500 --> 00:04:43,930
Before we move on, a reminder about flux and
angles. Only electric field lines that actually
55
00:04:43,930 --> 00:04:49,010
pass through a surface provide flux. When
the surface is parallel to the field, no flux
56
00:04:49,010 --> 00:04:54,080
is provided. This is particularly useful for
some three-dimensional surfaces, such as the
57
00:04:54,080 --> 00:04:58,730
empty cylinder to the right. If we remember
this, we can really simplify our integrals.
58
00:04:58,730 --> 00:04:59,690
Another thing to remember: Flux is a scalar.
As long as the field strength is the same,
59
00:04:59,690 --> 00:05:00,660
and the angle between the field and surface
stays the same, the flux will be the same.
60
00:05:00,660 --> 00:05:01,560
We don't have to worry about the flux pointing
in a particular direction, because it's just
61
00:05:01,560 --> 00:05:01,840
a number with no direction.
62
00:05:01,840 --> 00:05:06,720
Now we know how we would like the field and
the area to line up. Next we need to understand
63
00:05:06,720 --> 00:05:11,680
the circumstances that make that possible.
We need to look at electric charge distributions
64
00:05:11,680 --> 00:05:13,400
and their symmetry.
65
00:05:13,400 --> 00:05:19,070
Electric charge can come as a single point,
as in an electron or proton, but it can also
66
00:05:19,070 --> 00:05:24,040
take other shapes. It might be stretched into
a line, spread throughout a volume, or spread
67
00:05:24,040 --> 00:05:30,480
over a surface. Not a Gaussian surface - remember,
that's something we make up to solve a problem.
68
00:05:30,480 --> 00:05:35,490
Charge is spread on a physical object. You
can stretch the charge out into lines, spread
69
00:05:35,490 --> 00:05:40,970
it on thick wires or spherical shells, build
a solid ball out of it, or even just smear
70
00:05:40,970 --> 00:05:43,220
it around on an object.
71
00:05:43,220 --> 00:05:47,800
An object has symmetry when it has exactly
similar parts that are either facing each
72
00:05:47,800 --> 00:05:53,290
other, or arranged around an axis. On the
screen are some pictures of objects that have
73
00:05:53,290 --> 00:05:54,510
symmetry.
74
00:05:54,510 --> 00:06:01,310
There are three categories that we'll investigate,
and three types of symmetry to consider. Our
75
00:06:01,310 --> 00:06:05,460
hope is to have a highly symmetric charge
distribution, with any of the three types
76
00:06:05,460 --> 00:06:09,389
of symmetry. That's when Gauss' Law is easiest
to apply.
77
00:06:09,389 --> 00:06:13,490
These examples have little or no symmetry
to them, and are not the sort of thing for
78
00:06:13,490 --> 00:06:16,280
which we can use Gauss' Law.
79
00:06:16,280 --> 00:06:21,330
These three charge distributions have visible
symmetry, but it's not enough. We need a charge
80
00:06:21,330 --> 00:06:26,580
distribution so symmetric that the field it
generates is also symmetric.
81
00:06:26,580 --> 00:06:32,290
Let us show why. Let's say we have a charged
cube, which is certainly a symmetric object.
82
00:06:32,290 --> 00:06:37,669
We can draw the field from it fairly easily.
If we place a cubical Gaussian surface around
83
00:06:37,669 --> 00:06:43,310
the charge, we can see the problem: the field
and the surface have different angles in different
84
00:06:43,310 --> 00:06:48,139
places. There's some symmetry here, but not
enough to allow us to pull everything out
85
00:06:48,139 --> 00:06:50,870
of the integral.
86
00:06:50,870 --> 00:06:57,639
These cases are ideal: a point charge, a spherical
shell, an infinite sheet of charge, or an
87
00:06:57,639 --> 00:07:03,419
infinite cylindrical shell of charge. No doubt
you've used Gauss' Law with these charge distributions
88
00:07:03,419 --> 00:07:09,639
before. How can we describe such distributions
and find others like them?
89
00:07:09,639 --> 00:07:14,630
All three of those have a special quality:
if you move in two directions, they look exactly
90
00:07:14,630 --> 00:07:20,510
the same. No matter how far you move across
the plane in the x or y directions, it looks
91
00:07:20,510 --> 00:07:26,699
identical. Move along the cylinder or around
it, and it looks the same. Rotate around the
92
00:07:26,699 --> 00:07:31,990
sphere around the equator or the poles, and
you can't tell the difference.
93
00:07:31,990 --> 00:07:37,580
We can also add similar objects together or
scale them however we like. By integrating
94
00:07:37,580 --> 00:07:43,199
the fields from a lot of shells, we can create
a solid object, just like these Russian nesting
95
00:07:43,199 --> 00:07:44,400
dolls.
96
00:07:44,400 --> 00:07:50,210
Of course, perfectly symmetric charge distributions
are rare in real life. This is probably because
97
00:07:50,210 --> 00:07:56,510
infinitely long cylinders are so hard to find.
Thus we often use Gauss' Law in cases where
98
00:07:56,510 --> 00:08:02,040
the charge distributions are close to one
of our symmetric cases. We can usually obtain
99
00:08:02,040 --> 00:08:05,800
excellent estimates of the true electric field
in this way.
100
00:08:05,800 --> 00:08:10,699
Two approximations are particularly common:
when the object is so large that we can treat
101
00:08:10,699 --> 00:08:16,430
it as being infinite, or when we are so close
to the object that its large-scale features
102
00:08:16,430 --> 00:08:23,150
become unimportant. In these cases, we often
treat the object as being an infinite plane.
103
00:08:23,150 --> 00:08:28,370
Here's an example of the first approximation.
Compare the distance from your point of interest
104
00:08:28,370 --> 00:08:34,250
to the object, to the size of the object itself.
If the smallest dimension of the object is
105
00:08:34,250 --> 00:08:39,999
ten times your distance from it, you can rely
on a very good estimation from Gauss' Law,
106
00:08:39,999 --> 00:08:42,938
even if your plane isn't really infinite.
107
00:08:42,938 --> 00:08:48,089
To use the other approximation, we want a
situation where the curvature is very small
108
00:08:48,089 --> 00:08:52,709
in the place we care about. For example, it
would be very difficult for us to find the
109
00:08:52,709 --> 00:08:58,699
field at this point. However, if we zoom in,
we can reach a point where the surface looks
110
00:08:58,699 --> 00:09:05,699
like a flat plane. Finding an estimate for
the field at this point will be much easier.
111
00:09:06,589 --> 00:09:11,290
Now we have the background we need. It's time
to put all the pieces together and pick our
112
00:09:11,290 --> 00:09:12,699
Gaussian surface.
113
00:09:12,699 --> 00:09:18,249
We'll do a quick review in case you missed
something, and give you an example or two.
114
00:09:18,249 --> 00:09:23,430
After that, it's your turn to pick the best
surface for a given situation.
115
00:09:23,430 --> 00:09:28,350
In order to simplify our integrals as much
as possible, we seek out situations with as
116
00:09:28,350 --> 00:09:33,110
much symmetry as we can get. The symmetry
in the charge distribution isn't something
117
00:09:33,110 --> 00:09:38,439
we can control, but our choice of surface
will determine whether there is symmetry in
118
00:09:38,439 --> 00:09:43,410
how the electric field meets the Gaussian
surface. And that's always the key: you choose
119
00:09:43,410 --> 00:09:47,980
the surface to make the problem easier for
you. Finally, it's important to remember that
120
00:09:47,980 --> 00:09:53,480
in some cases, Gauss' Law won't be helpful,
and it's good to look for another method.
121
00:09:53,480 --> 00:10:00,480
Here are three charge distributions: a long
cylinder, a sphere, and a flat plane. To maximize
122
00:10:00,699 --> 00:10:06,499
the effects of symmetry, we choose surfaces
that match up well with the charge distributions.
123
00:10:06,499 --> 00:10:11,629
To find the field near a cylinder, we surround
it with another cylinder. To find the field
124
00:10:11,629 --> 00:10:17,619
near a sphere, we encase it in a sphere. To
find the field near a charged plane, we use
125
00:10:17,619 --> 00:10:22,949
a box-shaped surface. Sometimes people will
use a cylindrical surface instead, and count
126
00:10:22,949 --> 00:10:29,100
the flux that goes through the top of the
cylinder. Both approaches are valid.
127
00:10:29,100 --> 00:10:35,490
Here is your first challenge: choose the Gaussian
surface that will best fit this charge distribution.
128
00:10:35,490 --> 00:10:41,879
We want to know the electric field at a certain
distance from the center of this solid sphere.
129
00:10:41,879 --> 00:10:46,220
What surface would you choose?
130
00:10:46,220 --> 00:10:53,220
A sphere inside our existing sphere will be
easiest. This example helps us remember that
131
00:10:54,920 --> 00:10:59,170
the Gaussian surface can be inside an object.
The area of the surface will be the usual
132
00:10:59,170 --> 00:11:03,379
four thirds pi r cubed. We'll need to remember
to use only the charge inside our volume,
133
00:11:03,379 --> 00:11:07,980
not the whole charge of the sphere.
134
00:11:07,980 --> 00:11:13,240
Here's challenge number two. We would like
to know the electric charge contained in a
135
00:11:13,240 --> 00:11:17,059
thundercloud. Let's say that the inside of
the cloud has an electric field that looks
136
00:11:17,059 --> 00:11:18,860
like this.
137
00:11:18,860 --> 00:11:25,860
What surface would you choose?
138
00:11:27,689 --> 00:11:33,829
A box-shaped surface will take advantage of
the Cartesian symmetry in this electric field.
139
00:11:33,829 --> 00:11:37,959
Because of the alignment between the sides
of the box and the electric field, they will
140
00:11:37,959 --> 00:11:43,139
not contribute to the flux integral. We can
use just the area of the top and bottom of
141
00:11:43,139 --> 00:11:50,139
the box, and the field at those locations,
to find the flux. This is a common technique.
142
00:11:51,249 --> 00:11:55,540
Here's challenge number three: finding the
electric field at a certain distance from
143
00:11:55,540 --> 00:11:58,869
the center of charged disc.
144
00:11:58,869 --> 00:12:05,869
What surface would you choose?
145
00:12:06,959 --> 00:12:11,600
No surface will work! The problem is that
there's not enough symmetry in this situation
146
00:12:11,600 --> 00:12:13,959
to pick an appropriate surface.
147
00:12:13,959 --> 00:12:19,739
If we can't get an exact solution, we would
like to approximate the field. Unfortunately,
148
00:12:19,739 --> 00:12:24,199
if we look at the distances involved, we are
not close enough to the disc to make a good
149
00:12:24,199 --> 00:12:25,730
approximation.
150
00:12:25,730 --> 00:12:30,899
In the end, this is not a good Gauss' Law
problem. It is solved much more easily through
151
00:12:30,899 --> 00:12:35,929
other techniques, such as integration and
Coulomb's Law. An example can be found in
152
00:12:35,929 --> 00:12:39,319
the supplemental materials for this video.
153
00:12:39,319 --> 00:12:44,300
Here is the final challenge. A top and side
view of the field are shown.
154
00:12:44,300 --> 00:12:51,300
What surface would align well with this field?
155
00:12:52,589 --> 00:12:57,769
A cylindrical surface would work very well
with this circular field. You can see that
156
00:12:57,769 --> 00:13:03,350
the field is perpendicular to the area vector
at all points. Our dot product is now giving
157
00:13:03,350 --> 00:13:09,569
us a cosine of ninety degrees at all times.
This means a zero value for the flux at all
158
00:13:09,569 --> 00:13:14,589
points. We don't even need to determine the
area of the cylinder.
159
00:13:14,589 --> 00:13:20,559
If our flux is zero, the total charge inside
must also be zero. We might guess that positive
160
00:13:20,559 --> 00:13:26,149
and negative charges together could create
this field, but as it turns out, they cannot.
161
00:13:26,149 --> 00:13:30,529
Fields that look like this are created by
a changing magnetic field, using Faraday's
162
00:13:30,529 --> 00:13:34,559
Law, which you will learn more about later
in your course.
163
00:13:34,559 --> 00:13:38,749
There are some times when you'll just have
to integrate; there's no way around it. Not
164
00:13:38,749 --> 00:13:44,139
every situation is perfectly symmetric. When
you come to these situations, look for the
165
00:13:44,139 --> 00:13:50,649
following things to help you out. First, an
area vector that you can define easily. Second,
166
00:13:50,649 --> 00:13:56,959
an electric field that you can easily find
or that is defined for you in a problem. Third,
167
00:13:56,959 --> 00:14:01,730
see if you can create a surface where either
the magnitude of the field or its angle is
168
00:14:01,730 --> 00:14:07,809
uniform. The more of these you can take advantage
of, the better.
169
00:14:07,809 --> 00:14:08,939
Let's review.
170
00:14:08,939 --> 00:14:13,519
The more symmetry we can take advantage of
in our problem, the easier time we'll have
171
00:14:13,519 --> 00:14:20,209
with Gauss' Law and flux. Ideally, we want
a situation where the charge, the field it
172
00:14:20,209 --> 00:14:25,049
creates, and the Gaussian surface we choose
all have the same symmetry, so that we can
173
00:14:25,049 --> 00:14:29,600
simplify the integral as much as possible.
174
00:14:29,600 --> 00:14:35,660
For the final segment of this video, we wanted
to answer a common question about electromagnetism.
175
00:14:35,660 --> 00:14:40,549
What is flux? What does it mean? How can we
understand it?
176
00:14:40,549 --> 00:14:45,499
We took a video camera into the physics department
at MIT to find out how physics professors
177
00:14:45,499 --> 00:14:52,499
and students think about flux. Here's what
we found out.
178
00:14:59,110 --> 00:15:06,110
Thank you
179
00:15:19,879 --> 00:15:26,879
for watching this video. We
hope that it has
180
00:17:32,190 --> 00:17:36,620
improved your understanding of flux and Gauss'
Law. Good luck in your future exploration
181
00:17:36,620 --> 00:17:42,490
of electromagnetism.