1
00:00:07,395 --> 00:00:10,353
[MUSIC PLAYING]

2
00:00:16,266 --> 00:00:17,390
PROFESSOR: Hello, everyone.

3
00:00:17,390 --> 00:00:20,010
Today we'll be taking a
look at how light interacts

4
00:00:20,010 --> 00:00:21,940
with the surface
of a solar cell.

5
00:00:21,940 --> 00:00:24,760
Right now I'm standing next
to a solar module made up

6
00:00:24,760 --> 00:00:27,935
of individual
silicon solar cells.

7
00:00:27,935 --> 00:00:31,950
If you look closely, these
cells actually appear black.

8
00:00:31,950 --> 00:00:35,120
And they appear black for
a very important reason.

9
00:00:35,120 --> 00:00:38,310
Solar engineers work very hard
to make their solar cells as

10
00:00:38,310 --> 00:00:40,310
efficient as possible.

11
00:00:40,310 --> 00:00:43,660
Reflected light is lost
energy, so good engineers

12
00:00:43,660 --> 00:00:47,360
will want to minimize the total
amount of reflected light.

13
00:00:47,360 --> 00:00:50,610
To make solar cells absorb
as much light as possible,

14
00:00:50,610 --> 00:00:54,210
and appear black, solar
engineers do two things.

15
00:00:54,210 --> 00:00:57,570
First, they grow
this very thin film

16
00:00:57,570 --> 00:01:00,170
of a dielectric
layer on the surface.

17
00:01:00,170 --> 00:01:04,160
This layer is aptly called
an anti-reflection coating.

18
00:01:04,160 --> 00:01:06,940
Second, they texture the wafer.

19
00:01:06,940 --> 00:01:09,900
And today we'll demonstrate
how texturing is performed,

20
00:01:09,900 --> 00:01:14,820
and quantify its enhancement
for reducing light reflection.

21
00:01:14,820 --> 00:01:17,020
Silicon wafers don't
start out black.

22
00:01:17,020 --> 00:01:19,260
In fact, they appear gray.

23
00:01:19,260 --> 00:01:20,960
Polished wafers even
look mirror-like,

24
00:01:20,960 --> 00:01:23,430
and reflect quite
a bit of light.

25
00:01:23,430 --> 00:01:26,720
Here we see a polished wafer,
which reflects around 1/3

26
00:01:26,720 --> 00:01:28,370
of the light off its surface.

27
00:01:28,370 --> 00:01:31,300
And to create the rough surface
that reflects less light,

28
00:01:31,300 --> 00:01:33,750
solar engineers immerse
their silicon wafers

29
00:01:33,750 --> 00:01:36,370
into a hot, wet
chemical bath, which

30
00:01:36,370 --> 00:01:38,910
helps create tiny
surface features.

31
00:01:38,910 --> 00:01:40,900
In this example,
we use a solution

32
00:01:40,900 --> 00:01:43,880
of potassium hydroxide,
or KOH, which

33
00:01:43,880 --> 00:01:46,820
is heated to around
80 degrees Celsius.

34
00:01:46,820 --> 00:01:49,000
This violent
reaction is actually

35
00:01:49,000 --> 00:01:51,010
etching into the
silicon, and carving out

36
00:01:51,010 --> 00:01:53,190
little tiny pyramids
on the surface.

37
00:01:53,190 --> 00:01:56,400
And the result is a wafer that
loses its shiny appearance,

38
00:01:56,400 --> 00:01:57,940
and appears to
have a dull finish.

39
00:01:57,940 --> 00:02:00,470
The textured wafer is
left with a surface that

40
00:02:00,470 --> 00:02:02,920
is covered with microscopic
pyramids, whose base

41
00:02:02,920 --> 00:02:05,380
is around a micron,
or about 1/50

42
00:02:05,380 --> 00:02:06,680
of the width of a human hair.

43
00:02:06,680 --> 00:02:09,430
It turns out that this wafer
only reflects about 1/3

44
00:02:09,430 --> 00:02:11,420
as much light as
it previously did.

45
00:02:11,420 --> 00:02:14,430
The reason this KOH which
bath work so well at texturing

46
00:02:14,430 --> 00:02:16,410
the silicon surface
is due to the fact

47
00:02:16,410 --> 00:02:19,550
that these silicon wafers
are large crystals, which

48
00:02:19,550 --> 00:02:22,290
in this case means that the
atoms are formed in an ordered,

49
00:02:22,290 --> 00:02:23,760
repeating pattern.

50
00:02:23,760 --> 00:02:27,430
I have a model of the silicon
crystal structure right here.

51
00:02:27,430 --> 00:02:31,010
On our model, I've highlighted
the surface in red.

52
00:02:31,010 --> 00:02:34,160
Note that each silicon
atom below the surface

53
00:02:34,160 --> 00:02:37,750
is actually bonded to
four other silicon atoms,

54
00:02:37,750 --> 00:02:40,120
with four covalent bonds.

55
00:02:40,120 --> 00:02:42,400
Note that the surface
atoms are only

56
00:02:42,400 --> 00:02:45,610
bonded to two other
silicon atoms,

57
00:02:45,610 --> 00:02:48,100
and it has two bonds
that are unbonded.

58
00:02:48,100 --> 00:02:50,870
The alkaline etch is able
to remove silicon atoms more

59
00:02:50,870 --> 00:02:54,980
rapidly, when they have fewer
bonds holding to the lattice.

60
00:02:54,980 --> 00:02:58,790
Hence, a solution quickly
removes atoms on the surface.

61
00:02:58,790 --> 00:03:01,580
Now if I were acting
as the KOH solution,

62
00:03:01,580 --> 00:03:03,230
I would remove
all the atoms that

63
00:03:03,230 --> 00:03:05,990
only have two covalent bonds.

64
00:03:05,990 --> 00:03:07,910
Let's remove a few atoms.

65
00:03:07,910 --> 00:03:10,570
So this one only has
two covalent bonds.

66
00:03:10,570 --> 00:03:11,300
It gets removed.

67
00:03:17,240 --> 00:03:19,660
These three atoms
on the surface only

68
00:03:19,660 --> 00:03:21,700
have two covalent
bonds holding them,

69
00:03:21,700 --> 00:03:23,245
so I'll remove them as well.

70
00:03:23,245 --> 00:03:23,870
Let's go ahead.

71
00:03:32,960 --> 00:03:35,770
Now we can see that after
we have removed a few atoms,

72
00:03:35,770 --> 00:03:41,360
we have created some atoms below
our original surface that only

73
00:03:41,360 --> 00:03:44,140
have two covalent bonds
holding to the lattice.

74
00:03:44,140 --> 00:03:47,560
These atoms will also
get removed by the KOH.

75
00:03:47,560 --> 00:03:49,320
Now if we continue
this process, it

76
00:03:49,320 --> 00:03:50,860
would look something like this.

77
00:03:57,330 --> 00:03:59,710
So what we're going to do
is measure the reflectivity

78
00:03:59,710 --> 00:04:03,210
of both a flat and
a textured wafer.

79
00:04:03,210 --> 00:04:05,080
First, we'll measure
the flat wafer,

80
00:04:05,080 --> 00:04:06,820
and use a laser pointer
as a light source

81
00:04:06,820 --> 00:04:10,500
to simulate sunlight coming
from very far away point.

82
00:04:10,500 --> 00:04:12,220
We'll shine light
down on the surface,

83
00:04:12,220 --> 00:04:13,720
and measure the
amount of light that

84
00:04:13,720 --> 00:04:15,580
gets reflected or
bounced back away

85
00:04:15,580 --> 00:04:23,120
from the surface,
which we'll label as R.

86
00:04:23,120 --> 00:04:25,700
So I'm standing next to our
first experiment, which I just

87
00:04:25,700 --> 00:04:27,490
outlined in the previous sketch.

88
00:04:27,490 --> 00:04:30,760
We'll be using
this laser pointer,

89
00:04:30,760 --> 00:04:36,872
shining it onto this silicon
wafer, and into our photodiode.

90
00:04:36,872 --> 00:04:39,080
For those of you who don't
know what a photodiode is,

91
00:04:39,080 --> 00:04:42,060
it is a tool that can
measure the amount of light

92
00:04:42,060 --> 00:04:43,490
hitting its surface.

93
00:04:43,490 --> 00:04:45,469
The current that's read
off of this ammeter

94
00:04:45,469 --> 00:04:47,260
will be proportional
to the amount of light

95
00:04:47,260 --> 00:04:48,650
hitting our photodiode.

96
00:04:48,650 --> 00:04:52,060
To help visualize the beam path,
we're going to use some steam.

97
00:04:58,740 --> 00:05:03,230
Now if we turn on our laser
pointer, hits our photodiode,

98
00:05:03,230 --> 00:05:04,480
and we can get a good reading.

99
00:05:04,480 --> 00:05:08,370
And right now we see that it's
reading around 0.9 milliamps.

100
00:05:08,370 --> 00:05:11,710
But how does this compare
to a textured wafer?

101
00:05:11,710 --> 00:05:13,260
Let's find out.

102
00:05:13,260 --> 00:05:15,320
Now it's hard to
measure the reflectivity

103
00:05:15,320 --> 00:05:17,510
of a textured wafer
with a laser pointer,

104
00:05:17,510 --> 00:05:22,080
because it bounces the light
off at several different angles,

105
00:05:22,080 --> 00:05:26,140
and we can't measure the entire
beam with a photodiode alone.

106
00:05:26,140 --> 00:05:29,440
However, we can simulate
what this would be like.

107
00:05:29,440 --> 00:05:31,000
Let's go to our
sketch board to show

108
00:05:31,000 --> 00:05:33,420
how we can approximate
this measurement.

109
00:05:33,420 --> 00:05:35,960
To measure the reflectivity
of a textured surface,

110
00:05:35,960 --> 00:05:39,000
we'll create a 10,000
to one scale model.

111
00:05:39,000 --> 00:05:40,590
We'll approximate
our textured surface

112
00:05:40,590 --> 00:05:43,080
using two pieces of silicon.

113
00:05:43,080 --> 00:05:46,130
Again, we use a laser
pointer as our light source,

114
00:05:46,130 --> 00:05:48,230
and shine it on one
side of the pyramid.

115
00:05:48,230 --> 00:05:50,270
It'll bounce off that
next surface, and off

116
00:05:50,270 --> 00:05:52,190
the adjacent side,
and then we'll

117
00:05:52,190 --> 00:05:54,792
measure the amount
of light reflected.

118
00:05:54,792 --> 00:05:56,250
Let's go to our
experimental setup.

119
00:06:01,280 --> 00:06:03,540
So let's clarify our set up.

120
00:06:03,540 --> 00:06:08,785
The two angled lines in our
drawing, the adjacent sides,

121
00:06:08,785 --> 00:06:15,740
on our atomic models, and the
two angled, non-textured wafers

122
00:06:15,740 --> 00:06:18,590
in our set up, are all
representing the pyramid

123
00:06:18,590 --> 00:06:21,310
structure of a
textured wafer, just

124
00:06:21,310 --> 00:06:24,270
like the one visible in our
scanning electron microscope

125
00:06:24,270 --> 00:06:24,980
image.

126
00:06:24,980 --> 00:06:27,579
Let's use some steam to
visualize the beam path.

127
00:06:32,740 --> 00:06:34,280
Now that the steam
has settled, we

128
00:06:34,280 --> 00:06:36,550
can get an accurate
reading of the reflectivity

129
00:06:36,550 --> 00:06:39,190
of our modeled silicon surface.

130
00:06:39,190 --> 00:06:42,260
And I turn on our laser pointer.

131
00:06:42,260 --> 00:06:43,930
So now we can see
that our photodiode

132
00:06:43,930 --> 00:06:46,440
is reading around
0.33 milliamps.

133
00:06:46,440 --> 00:06:49,782
This corresponds to
around a 9% reflectivity,

134
00:06:49,782 --> 00:06:51,490
which is quite a huge
reduction from what

135
00:06:51,490 --> 00:06:54,370
we had before, by
about a factor of 3.

136
00:06:54,370 --> 00:06:56,300
So in summary,
today we learned how

137
00:06:56,300 --> 00:06:58,860
silicon is etched
using a KOH solution,

138
00:06:58,860 --> 00:07:00,630
and how the resultant
pyramids increase

139
00:07:00,630 --> 00:07:02,260
the efficiency of
our solar cells,

140
00:07:02,260 --> 00:07:04,460
by reducing the amount
of reflective losses

141
00:07:04,460 --> 00:07:06,569
by a factor of 3.

142
00:07:06,569 --> 00:07:08,110
If you found this
interesting, please

143
00:07:08,110 --> 00:07:10,360
watch our other solar
demos to learn more about

144
00:07:10,360 --> 00:07:12,410
how these exciting devices work.

145
00:07:12,410 --> 00:07:15,694
I'm Joe Sullivan from MIT,
and thanks for watching.

146
00:07:15,694 --> 00:07:16,360
I'm out of here.

147
00:07:19,860 --> 00:07:22,910
[MUSIC PLAYING]