1
00:00:00,050 --> 00:00:02,500
The following content is
provided under a Creative

2
00:00:02,500 --> 00:00:04,019
Commons license.

3
00:00:04,019 --> 00:00:06,350
Your support will help
MIT OpenCourseWare

4
00:00:06,350 --> 00:00:10,720
continue to offer high quality
educational resources for free.

5
00:00:10,720 --> 00:00:13,340
To make a donation or
view additional materials

6
00:00:13,340 --> 00:00:17,226
from hundreds of MIT courses,
visit MIT OpenCourseWare

7
00:00:17,226 --> 00:00:17,851
at ocw.mit.edu.

8
00:00:26,180 --> 00:00:27,740
PROFESSOR: So I'm
going to describe

9
00:00:27,740 --> 00:00:30,120
the basic classification of
solar cell characterization

10
00:00:30,120 --> 00:00:31,306
methods.

11
00:00:31,306 --> 00:00:33,680
And then I'll describe some
of the characterization tools

12
00:00:33,680 --> 00:00:35,850
that are used to
measure Jsc losses

13
00:00:35,850 --> 00:00:38,900
and other tools are used
to measure VOC and fill

14
00:00:38,900 --> 00:00:40,340
factor losses.

15
00:00:40,340 --> 00:00:42,990
So we're getting a sense
of the lay of the land.

16
00:00:42,990 --> 00:00:48,140
Let me ask you how would you
characterize-- how would you

17
00:00:48,140 --> 00:00:51,914
create a taxonomy of solar cell
characterization techniques?

18
00:00:51,914 --> 00:00:54,080
Those of you might have a
little bit more experience

19
00:00:54,080 --> 00:00:56,496
might actually have used a few
characterization techniques

20
00:00:56,496 --> 00:00:57,430
in your laboratories.

21
00:00:57,430 --> 00:00:59,260
What taxonomy would you use?

22
00:00:59,260 --> 00:01:02,530
How would you slice the pie
of all these characterization

23
00:01:02,530 --> 00:01:03,100
techniques?

24
00:01:03,100 --> 00:01:03,600
Go.

25
00:01:06,910 --> 00:01:10,020
What are ways to divide the
characterization universe

26
00:01:10,020 --> 00:01:11,220
into distinct groupings?

27
00:01:16,310 --> 00:01:17,980
How about we start-- go ahead.

28
00:01:17,980 --> 00:01:19,032
Yeah?

29
00:01:19,032 --> 00:01:21,190
AUDIENCE: Performance
and performance losses.

30
00:01:21,190 --> 00:01:23,380
PROFESSOR: Yeah
so we can describe

31
00:01:23,380 --> 00:01:27,240
based on the economic
variables like efficiency

32
00:01:27,240 --> 00:01:29,680
and then within
efficiency Jsc, Voc,

33
00:01:29,680 --> 00:01:33,660
fill factor, mechanical
yield, reliability.

34
00:01:33,660 --> 00:01:36,220
These are all parameters that
can lump into a cost model

35
00:01:36,220 --> 00:01:39,140
and ultimately impact the
cost-- dollars per watt,

36
00:01:39,140 --> 00:01:40,630
let's say of a PV panel.

37
00:01:40,630 --> 00:01:43,610
Excellent, so we can describe
it based on performance.

38
00:01:43,610 --> 00:01:46,700
If we dive into
performance a little more,

39
00:01:46,700 --> 00:01:49,930
what are some of the properties
that affect performance?

40
00:01:49,930 --> 00:01:54,790
We have electrical properties,
optical properties, mechanical,

41
00:01:54,790 --> 00:01:56,270
thermal.

42
00:01:56,270 --> 00:01:59,110
So we can break up
characterization techniques

43
00:01:59,110 --> 00:02:01,310
based on the properties
that they probe.

44
00:02:01,310 --> 00:02:03,295
And we can also
characterize-- we

45
00:02:03,295 --> 00:02:06,380
can create a taxonomy of
characterization techniques

46
00:02:06,380 --> 00:02:08,487
based on how fast they
are, whether it's something

47
00:02:08,487 --> 00:02:11,070
that you have to sit around and
twiddle your thumbs for a half

48
00:02:11,070 --> 00:02:13,770
or two hours to get done
or something that gets done

49
00:02:13,770 --> 00:02:15,610
in 10 milliseconds
of which there

50
00:02:15,610 --> 00:02:17,620
are several
characterization techniques.

51
00:02:17,620 --> 00:02:21,499
What would the 10 millisecond
variety enable you to do?

52
00:02:21,499 --> 00:02:22,790
In a manufacturing environment?

53
00:02:22,790 --> 00:02:23,230
Yeah?

54
00:02:23,230 --> 00:02:24,064
AUDIENCE: Sort them.

55
00:02:24,064 --> 00:02:24,938
PROFESSOR: Sort them.

56
00:02:24,938 --> 00:02:25,480
Test them.

57
00:02:25,480 --> 00:02:27,750
Measure them-- every
single cell coming through.

58
00:02:27,750 --> 00:02:29,970
So you might have
a bar code like one

59
00:02:29,970 --> 00:02:31,890
of those fancy
two-dimensional bar codes

60
00:02:31,890 --> 00:02:34,345
laser marked on each single
wafer going through your line

61
00:02:34,345 --> 00:02:37,640
so you can trace it back in
your MES system all the way back

62
00:02:37,640 --> 00:02:41,290
to the crystal that was grown
or from the thin film growth

63
00:02:41,290 --> 00:02:43,270
chamber where it was deposited.

64
00:02:43,270 --> 00:02:47,370
So, yes, you have
inline and shall

65
00:02:47,370 --> 00:02:51,910
we say, the inline
characterization techniques

66
00:02:51,910 --> 00:02:54,070
and what are called
offline characterization

67
00:02:54,070 --> 00:02:56,200
techniques, which tend
to be lower throughput.

68
00:02:56,200 --> 00:02:58,491
You can think of the inline
characterization techniques

69
00:02:58,491 --> 00:03:00,250
as being in the line
of the manufacturing

70
00:03:00,250 --> 00:03:02,770
environment and the offline
as being those techniques that

71
00:03:02,770 --> 00:03:06,560
are sitting in your laboratory
waiting to be used in the R&D

72
00:03:06,560 --> 00:03:11,000
lab, which might be next door
to the manufacturing line.

73
00:03:11,000 --> 00:03:13,080
Device performance
metric affected--

74
00:03:13,080 --> 00:03:16,102
that addresses the
efficiency point up here.

75
00:03:16,102 --> 00:03:18,310
And then by property tested--
electrical, structural,

76
00:03:18,310 --> 00:03:19,460
optical, mechanical.

77
00:03:19,460 --> 00:03:23,140
So if you talk to somebody
who works in a PV company,

78
00:03:23,140 --> 00:03:25,270
she will likely give
you a break down

79
00:03:25,270 --> 00:03:27,402
based on number
three right here.

80
00:03:27,402 --> 00:03:29,860
OK, these are the techniques
that we have in our inspection

81
00:03:29,860 --> 00:03:31,850
system and those are
the techniques that I

82
00:03:31,850 --> 00:03:33,450
have over there in the R&D lab.

83
00:03:33,450 --> 00:03:36,566
If you talk to somebody who
is giving a fundamental course

84
00:03:36,566 --> 00:03:38,940
in materials science, they
are likely to pick part number

85
00:03:38,940 --> 00:03:41,480
one way up there and give
you the breakdown based

86
00:03:41,480 --> 00:03:43,210
on the properties
that are probed.

87
00:03:43,210 --> 00:03:46,045
I'm going to opt for
number two today in lecture

88
00:03:46,045 --> 00:03:48,420
not because it's any better
or worse than all the others,

89
00:03:48,420 --> 00:03:50,750
but just because that's
the metric that we care

90
00:03:50,750 --> 00:03:54,360
about right now as we're trying
to probe dollars per watt that

91
00:03:54,360 --> 00:03:57,500
relates to our quiz
and our homework,

92
00:03:57,500 --> 00:04:03,600
but also relates to the ultimate
economic driver of solar.

93
00:04:03,600 --> 00:04:07,090
So we will keep
this in mind that we

94
00:04:07,090 --> 00:04:09,510
might need to ferret
out references

95
00:04:09,510 --> 00:04:12,040
for the different techniques
in different textbooks

96
00:04:12,040 --> 00:04:14,625
or different papers based
on whether they're probing

97
00:04:14,625 --> 00:04:16,750
electrical structural
optical mechanical properties

98
00:04:16,750 --> 00:04:18,149
of our solar cells.

99
00:04:18,149 --> 00:04:21,399
But we're going to be
focused on efficiency--

100
00:04:21,399 --> 00:04:24,020
short circuit, current, open
circuit, voltage, fill factor.

101
00:04:24,020 --> 00:04:27,010
So let's go ahead
and dive into that.

102
00:04:27,010 --> 00:04:29,020
We are going to first
start with techniques

103
00:04:29,020 --> 00:04:31,790
to measure Jsc or short
circuit current losses.

104
00:04:31,790 --> 00:04:34,405
And some of these slides
are going to be repeats.

105
00:04:34,405 --> 00:04:35,780
And the reason
they're repeats is

106
00:04:35,780 --> 00:04:38,430
because the first time you
saw it OK, you sort of kind of

107
00:04:38,430 --> 00:04:38,980
got it.

108
00:04:38,980 --> 00:04:42,080
You went to the lab, made
your devices, tested them.

109
00:04:42,080 --> 00:04:45,510
And now all of a sudden, you
have a much stronger background

110
00:04:45,510 --> 00:04:47,300
with which to understand.

111
00:04:47,300 --> 00:04:49,190
We're going to
start by discussing

112
00:04:49,190 --> 00:04:51,857
the optical components
just very briefly

113
00:04:51,857 --> 00:04:54,190
and then spectra response and
minority carrier diffusion

114
00:04:54,190 --> 00:04:55,920
length, revisiting
some of these concepts

115
00:04:55,920 --> 00:04:57,702
that we have already
seen, but now

116
00:04:57,702 --> 00:05:00,160
with the benefit of having all
of our background knowledge.

117
00:05:00,160 --> 00:05:04,690
The spectrophotometer
measures specular and diffuse

118
00:05:04,690 --> 00:05:07,090
reflectance and transmission.

119
00:05:07,090 --> 00:05:08,720
All right, let's
break that down.

120
00:05:08,720 --> 00:05:13,370
Specular reflectance--
specu lar-- Latin.

121
00:05:13,370 --> 00:05:17,230
So specular reflectance
means light comes in and out

122
00:05:17,230 --> 00:05:21,180
pretty much at the same angle
relative to the surface normal.

123
00:05:21,180 --> 00:05:22,950
So if you come in
from here, it's

124
00:05:22,950 --> 00:05:25,590
going to bounce out light
there right at the same angle

125
00:05:25,590 --> 00:05:28,040
relative to the surface normal.

126
00:05:28,040 --> 00:05:30,490
Diffuse reflectance
means that if you

127
00:05:30,490 --> 00:05:31,990
shine light in a
certain angle, it's

128
00:05:31,990 --> 00:05:34,840
going to reflect back not
necessarily at that angle.

129
00:05:34,840 --> 00:05:36,330
You could have a distribution.

130
00:05:36,330 --> 00:05:39,740
A Lambertian scatterer
might qualify.

131
00:05:39,740 --> 00:05:42,370
And reflectance
in transmission--

132
00:05:42,370 --> 00:05:44,930
we talked about different
optical losses of a solar cell

133
00:05:44,930 --> 00:05:45,690
material.

134
00:05:45,690 --> 00:05:48,160
Reflectance means that light
comes back off the surface.

135
00:05:48,160 --> 00:05:50,887
Transmission means that
light did not get absorbed.

136
00:05:50,887 --> 00:05:53,220
So it went through the material
and didn't get absorbed.

137
00:05:53,220 --> 00:05:54,990
That is an optical loss as well.

138
00:05:54,990 --> 00:05:57,210
So the spectrophotometer
is useful for measuring

139
00:05:57,210 --> 00:05:59,020
these different loss mechanisms.

140
00:05:59,020 --> 00:06:01,320
And it can tease
apart the specular

141
00:06:01,320 --> 00:06:03,410
from the diffuse
reflectance, giving you

142
00:06:03,410 --> 00:06:07,650
some indication, some idea,
of how the surface is behaving

143
00:06:07,650 --> 00:06:10,200
and what you might
do to improve it.

144
00:06:10,200 --> 00:06:13,210
So the spectrophotometer
is useful in that regard.

145
00:06:13,210 --> 00:06:14,910
In terms of
increasing absorption,

146
00:06:14,910 --> 00:06:17,250
we talked about various
methods to increase absorption.

147
00:06:17,250 --> 00:06:19,000
Those who attended Eli
Yablonovitch's talk

148
00:06:19,000 --> 00:06:21,670
heard about many more.

149
00:06:21,670 --> 00:06:25,460
The goal here is essentially
to increase the optical path

150
00:06:25,460 --> 00:06:28,890
length by texturing your
surface, for instance.

151
00:06:28,890 --> 00:06:31,729
The physical thickness
can remain very low.

152
00:06:31,729 --> 00:06:33,270
And, again, just to
refresh ourselves

153
00:06:33,270 --> 00:06:35,830
if we decrease the
thickness of our devices

154
00:06:35,830 --> 00:06:38,040
but manage to have very
good like trapping,

155
00:06:38,040 --> 00:06:40,280
what happens to our
excess carrier density?

156
00:06:40,280 --> 00:06:44,140
It goes up, right?

157
00:06:44,140 --> 00:06:46,450
Because now we
have more carriers

158
00:06:46,450 --> 00:06:48,640
being generated in
a smaller volume

159
00:06:48,640 --> 00:06:51,230
so the carrier
density increases.

160
00:06:51,230 --> 00:06:53,260
And as the carry
density goes up,

161
00:06:53,260 --> 00:06:56,242
that means the separation of the
quasi-Fermi energies increases,

162
00:06:56,242 --> 00:06:58,700
which means the maximum voltage
extractable from or devices

163
00:06:58,700 --> 00:06:59,920
increases as well.

164
00:06:59,920 --> 00:07:01,720
So there's a strong
push right now

165
00:07:01,720 --> 00:07:03,710
in the field to go thinner
and thinner devices.

166
00:07:03,710 --> 00:07:05,560
That also has the added
benefit economically

167
00:07:05,560 --> 00:07:07,090
of using less material.

168
00:07:07,090 --> 00:07:09,400
So we want to
decrease the thickness

169
00:07:09,400 --> 00:07:13,970
by improving our optical
trapping-- our light trapping.

170
00:07:13,970 --> 00:07:17,230
Another benefit is it allows
carriers to be absorbed closer

171
00:07:17,230 --> 00:07:21,970
to the junction, which increases
the probability of collection.

172
00:07:21,970 --> 00:07:25,610
And we talked about this during
the very beginning of class

173
00:07:25,610 --> 00:07:28,640
how you might
texture your surface.

174
00:07:28,640 --> 00:07:32,790
But now we actually saw
it in producing our cells.

175
00:07:32,790 --> 00:07:38,010
So this is an example-- an SCM
image-- of textured silicon.

176
00:07:38,010 --> 00:07:41,110
In that particular case,
this was an alkali etch

177
00:07:41,110 --> 00:07:44,120
on a single crystalline
sample probably of 1 0

178
00:07:44,120 --> 00:07:48,082
0 orientation so that these
edges of the pyramids are 1 1

179
00:07:48,082 --> 00:07:50,110
1 planes.

180
00:07:50,110 --> 00:07:51,720
You could also
achieve a similar,

181
00:07:51,720 --> 00:07:54,000
although not identically,
looking result

182
00:07:54,000 --> 00:07:57,060
if you have performed
an acidic etch, which

183
00:07:57,060 --> 00:08:00,940
would be isotropic in nature.

184
00:08:00,940 --> 00:08:04,210
So the light comes in
for the textured surface.

185
00:08:04,210 --> 00:08:06,360
Some of it goes into the device.

186
00:08:06,360 --> 00:08:08,220
Notice that Snell's
law is in effect.

187
00:08:08,220 --> 00:08:10,916
So the light bent
or was refracted.

188
00:08:10,916 --> 00:08:13,040
And some of the light is
reflected off the surface.

189
00:08:13,040 --> 00:08:14,456
Now, because of
the texturization,

190
00:08:14,456 --> 00:08:16,840
you get that
second-chance absorption.

191
00:08:16,840 --> 00:08:19,680
So the light has two
bounces before leaving,

192
00:08:19,680 --> 00:08:22,350
which means that the probability
of the light getting absorbed

193
00:08:22,350 --> 00:08:25,580
is 1 minus r quantity squared.

194
00:08:25,580 --> 00:08:27,370
And so you get an
enhanced absorption.

195
00:08:27,370 --> 00:08:29,730
And for those who
looked at the samples

196
00:08:29,730 --> 00:08:31,190
before and after
texturization, you

197
00:08:31,190 --> 00:08:34,090
could visibly see that
they looked darker--

198
00:08:34,090 --> 00:08:35,740
the reflectivity had gone down.

199
00:08:35,740 --> 00:08:36,960
I see some smiles over here.

200
00:08:36,960 --> 00:08:39,630
There are probably some
initial etching processes

201
00:08:39,630 --> 00:08:41,880
that didn't quite work out,
but eventually the process

202
00:08:41,880 --> 00:08:45,010
was controlled and
worked out fine.

203
00:08:45,010 --> 00:08:47,350
So we have as well
other mechanisms

204
00:08:47,350 --> 00:08:51,190
to trap the light besides just
texturing our front surface.

205
00:08:51,190 --> 00:08:54,520
These we didn't get to
do an actual cell fab.

206
00:08:54,520 --> 00:08:58,060
But we could envision putting
a reflective or defuse scatter

207
00:08:58,060 --> 00:08:59,535
on the back and
reflective layer so

208
00:08:59,535 --> 00:09:01,910
that the light that goes all
the way through the material

209
00:09:01,910 --> 00:09:05,330
once gets bounced back
in perhaps at an angle

210
00:09:05,330 --> 00:09:07,740
to increase the
trapping in the back.

211
00:09:07,740 --> 00:09:10,350
And that I think is represented
on the previous slide.

212
00:09:10,350 --> 00:09:11,310
It is indeed.

213
00:09:11,310 --> 00:09:13,660
So we have a textured
back surface.

214
00:09:13,660 --> 00:09:17,220
And we also have this layer--
probably in this case,

215
00:09:17,220 --> 00:09:19,490
it would most likely be
a dielectric material

216
00:09:19,490 --> 00:09:22,020
with the refractive index
that is significantly

217
00:09:22,020 --> 00:09:24,809
different than the
absorber itself.

218
00:09:24,809 --> 00:09:26,600
So if your absorber
material is an organic,

219
00:09:26,600 --> 00:09:29,290
you might have a
refractive index

220
00:09:29,290 --> 00:09:32,487
somewhere between 1.5 and
2 maybe on the high end.

221
00:09:32,487 --> 00:09:34,070
And if you have an
inorganic material,

222
00:09:34,070 --> 00:09:37,800
you could have refractive
indices as high as 3, 3.5.

223
00:09:37,800 --> 00:09:39,500
And the material
you put in the back

224
00:09:39,500 --> 00:09:41,580
would probably be as close
as you could possibly

225
00:09:41,580 --> 00:09:46,360
get to air to get as
a large reflection

226
00:09:46,360 --> 00:09:49,270
off the back as possible.

227
00:09:49,270 --> 00:09:52,317
In the case of silicon
devices, oftentimes you

228
00:09:52,317 --> 00:09:54,400
have a dielectric material
with a refractive index

229
00:09:54,400 --> 00:09:57,450
somewhere between 1.5 and 2.

230
00:09:57,450 --> 00:10:00,000
And that serves also to
passivate the surface

231
00:10:00,000 --> 00:10:02,250
to prevent carriers from
recombining in the back side.

232
00:10:05,090 --> 00:10:08,470
So a collection probability--
we talked about this earlier now

233
00:10:08,470 --> 00:10:10,340
that we've performed
quantum efficiency

234
00:10:10,340 --> 00:10:11,862
and we've tested
our own devices,

235
00:10:11,862 --> 00:10:13,320
I'm going to walk
through it again.

236
00:10:13,320 --> 00:10:15,280
It's an important
concept to really grasp.

237
00:10:15,280 --> 00:10:17,860
I want to make sure
that everybody got it.

238
00:10:17,860 --> 00:10:20,640
We have here a p-n junction.

239
00:10:20,640 --> 00:10:22,070
Here's our P-type material.

240
00:10:22,070 --> 00:10:24,000
Here's our N-type material.

241
00:10:24,000 --> 00:10:27,770
Shown at the blue little
dots-- those are our holes--

242
00:10:27,770 --> 00:10:31,160
free holes-- mobile holes able
to move around the material.

243
00:10:31,160 --> 00:10:34,449
And on the n-type side,
we have here the red dots.

244
00:10:34,449 --> 00:10:36,490
Those are also free to
move around the materials.

245
00:10:36,490 --> 00:10:39,000
Those are electron--
negative charged carriers.

246
00:10:39,000 --> 00:10:42,410
Omitted from this diagram--
omitted from this diagram

247
00:10:42,410 --> 00:10:47,250
is the fixed charge associated
with each connectivity type.

248
00:10:47,250 --> 00:10:50,130
We have, for example,
fixed negative charge

249
00:10:50,130 --> 00:10:53,710
from the ionized
acceptors and the p-type

250
00:10:53,710 --> 00:10:56,010
and fixed positive charge
from the ionized donors.

251
00:10:56,010 --> 00:10:58,820
And the n-type-- we've
omitted it for clarity.

252
00:10:58,820 --> 00:11:01,510
We note that an electric
field builds up here

253
00:11:01,510 --> 00:11:02,870
at the junction between.

254
00:11:02,870 --> 00:11:06,190
And that sweeps carriers
into one side or another

255
00:11:06,190 --> 00:11:09,680
so that if light comes in and
generates a pair of charges--

256
00:11:09,680 --> 00:11:11,270
due to charge
neutrality, it generates

257
00:11:11,270 --> 00:11:15,180
a pair-- a positive and
a negative-- the charge

258
00:11:15,180 --> 00:11:17,730
carrier type will be
swept across the junction

259
00:11:17,730 --> 00:11:20,430
and the other will remain
inside of the material.

260
00:11:20,430 --> 00:11:24,440
And so to be more precise
about the exact motion that

261
00:11:24,440 --> 00:11:26,790
occurs with the carriers,
we have at first a diffusion

262
00:11:26,790 --> 00:11:30,290
process until the electric field
starts becoming large and then

263
00:11:30,290 --> 00:11:33,140
finally drift across
that junction.

264
00:11:33,140 --> 00:11:35,760
And that's what generates the
current inside of a solar cell

265
00:11:35,760 --> 00:11:36,680
device.

266
00:11:36,680 --> 00:11:39,990
So collection
probability means that

267
00:11:39,990 --> 00:11:42,830
light-- a light-generated--
a photo-generated minority

268
00:11:42,830 --> 00:11:45,670
carrier can readily recombine.

269
00:11:45,670 --> 00:11:48,610
But if the carrier reaches
the edge of the space charge

270
00:11:48,610 --> 00:11:52,290
region, the minority carrier can
be swept across and collected.

271
00:11:52,290 --> 00:11:57,300
So the probability of collection
means for every electron hole

272
00:11:57,300 --> 00:11:59,380
pair that I generate at
a certain depth, what

273
00:11:59,380 --> 00:12:01,754
is the probability that the
minority carrier will make it

274
00:12:01,754 --> 00:12:02,660
across a junction.

275
00:12:02,660 --> 00:12:06,250
And this is what we're
probing in spectral response.

276
00:12:06,250 --> 00:12:07,690
Yes?

277
00:12:07,690 --> 00:12:12,117
AUDIENCE: How long is the
exciton diffusion in silicon?

278
00:12:12,117 --> 00:12:13,200
PROFESSOR: Great question.

279
00:12:13,200 --> 00:12:17,457
The exciton in silicon has a
binding energy less than kt,

280
00:12:17,457 --> 00:12:19,290
which means at room
temperature, the exciton

281
00:12:19,290 --> 00:12:21,670
readily dissociates.

282
00:12:21,670 --> 00:12:24,400
So the exciton diffusion
length can only

283
00:12:24,400 --> 00:12:26,910
be measured at
lower temperatures.

284
00:12:26,910 --> 00:12:28,750
At room temperature,
you essentially

285
00:12:28,750 --> 00:12:30,850
just discuss minority carriers.

286
00:12:30,850 --> 00:12:32,740
And the minority carrier
diffusion length--

287
00:12:32,740 --> 00:12:34,660
let's revert it in
terms of lifetime

288
00:12:34,660 --> 00:12:37,050
and then do the translation
to diffusion length.

289
00:12:37,050 --> 00:12:38,980
In terms of lifetimes,
one typically

290
00:12:38,980 --> 00:12:40,475
has a range between
a microsecond

291
00:12:40,475 --> 00:12:42,940
to a millisecond-- millisecond
being on the high end.

292
00:12:42,940 --> 00:12:45,500
It can go as high as
five milliseconds even.

293
00:12:45,500 --> 00:12:47,870
And so if you revert that
into a diffusion length,

294
00:12:47,870 --> 00:12:50,180
one microsecond lifetime
would correspond

295
00:12:50,180 --> 00:12:51,904
to a 50-micron diffusion length.

296
00:12:51,904 --> 00:12:53,820
And then you can do the
rest of the conversion

297
00:12:53,820 --> 00:12:56,140
from there just taking the
square root sign into account.

298
00:12:56,140 --> 00:12:58,420
So it can be rather long on
the order of the thickness

299
00:12:58,420 --> 00:12:59,790
of the solar cell devices.

300
00:12:59,790 --> 00:13:03,890
Now those length skills, which
we were talking about hundreds

301
00:13:03,890 --> 00:13:07,250
of microns-- maybe even
a thousand microns.

302
00:13:07,250 --> 00:13:12,830
Compare that now to
10-nanometer diffusion length

303
00:13:12,830 --> 00:13:17,492
say for an exciton in a
polymer blend material.

304
00:13:17,492 --> 00:13:19,700
Then you have to reconsider
your device architecture.

305
00:13:19,700 --> 00:13:21,616
And how you're going to
actually collect them.

306
00:13:21,616 --> 00:13:23,350
So this is a very
generic picture, which

307
00:13:23,350 --> 00:13:24,516
could easily be substituted.

308
00:13:24,516 --> 00:13:26,900
Instead of n-type, p-type, it
could be easily substituted

309
00:13:26,900 --> 00:13:30,650
by say polymer 1 and polymer 2
that just have the right band

310
00:13:30,650 --> 00:13:32,750
alignment to separate charge.

311
00:13:32,750 --> 00:13:35,800
Then it gets a little
more complicated as well.

312
00:13:35,800 --> 00:13:38,770
Here we're assuming
that the carriers that

313
00:13:38,770 --> 00:13:40,550
are swept across the
junction can move

314
00:13:40,550 --> 00:13:42,640
easily away from the junction.

315
00:13:42,640 --> 00:13:44,810
In a polymer, not always so.

316
00:13:44,810 --> 00:13:46,590
You might have charge
accumulation right

317
00:13:46,590 --> 00:13:48,850
at the edge of the space
charge region, which

318
00:13:48,850 --> 00:13:51,410
creates its own field,
which counteracts

319
00:13:51,410 --> 00:13:55,010
the built-in field here and
also inhibits current flow.

320
00:13:55,010 --> 00:13:57,310
So things get a lot
more complicated

321
00:13:57,310 --> 00:14:00,080
as you venture away from
the simple case of say

322
00:14:00,080 --> 00:14:02,760
a well-behaved
what it was called

323
00:14:02,760 --> 00:14:05,550
in this case a homo junction,
meaning a p-n junction created

324
00:14:05,550 --> 00:14:07,690
from the same material
on both sides-- silicon

325
00:14:07,690 --> 00:14:09,080
just n-type an p-type.

326
00:14:09,080 --> 00:14:10,980
If you have a hetero
junction comprised

327
00:14:10,980 --> 00:14:14,450
of two different
materials, you may have

328
00:14:14,450 --> 00:14:16,050
additional effects occurring.

329
00:14:16,050 --> 00:14:18,750
But it's always helpful to think
about the motion of carriers

330
00:14:18,750 --> 00:14:21,160
from the perspective
of drift and diffusion.

331
00:14:21,160 --> 00:14:24,300
And if you're ever in doubt,
you can begin thinking about

332
00:14:24,300 --> 00:14:28,189
the process in terms of delta
T's-- small increments of time

333
00:14:28,189 --> 00:14:30,730
where you imagine light coming
in generating an electron hole

334
00:14:30,730 --> 00:14:33,970
pair, imagining what happens
as that carrier moves

335
00:14:33,970 --> 00:14:36,350
to the junction then
moves across the junction.

336
00:14:36,350 --> 00:14:37,390
And then what happens?

337
00:14:37,390 --> 00:14:38,806
Does it immediately
get collected?

338
00:14:38,806 --> 00:14:39,670
Does it stay there?

339
00:14:39,670 --> 00:14:41,950
What happens next?

340
00:14:41,950 --> 00:14:44,510
Is there a potential--
an attractive potential

341
00:14:44,510 --> 00:14:47,670
between the carriers on
either side of that junction.

342
00:14:47,670 --> 00:14:49,270
So asking those
sorts of questions

343
00:14:49,270 --> 00:14:52,080
can help you walk through and
troubleshoot what exactly is

344
00:14:52,080 --> 00:14:55,640
going wrong with your device.

345
00:14:55,640 --> 00:14:59,870
Collection probability--
so we take this diagram

346
00:14:59,870 --> 00:15:02,280
that we just on the previous
slide-- this one right here.

347
00:15:02,280 --> 00:15:06,160
And we say, if light were
to come in and generate

348
00:15:06,160 --> 00:15:08,840
an electron hole
pair right here,

349
00:15:08,840 --> 00:15:12,290
then I would expect very high
probability of collection,

350
00:15:12,290 --> 00:15:14,865
meaning the minority carrier
would be swept across.

351
00:15:14,865 --> 00:15:16,740
The majority carrier
would stay on this side.

352
00:15:16,740 --> 00:15:19,825
And presuming that they can
get from here to the contacts--

353
00:15:19,825 --> 00:15:21,450
big assumption-- but
assuming that they

354
00:15:21,450 --> 00:15:25,020
can-- then the probability of
collection of those carries

355
00:15:25,020 --> 00:15:27,140
will be very, very high.

356
00:15:27,140 --> 00:15:29,760
But if my light comes in
deep within the device

357
00:15:29,760 --> 00:15:31,600
and generates an
electron hole pair,

358
00:15:31,600 --> 00:15:35,189
the minority carrier will
have to cruise through-- you

359
00:15:35,189 --> 00:15:35,980
can think about it.

360
00:15:35,980 --> 00:15:39,150
It may be a nice, tasty chicken
in the Everglades National

361
00:15:39,150 --> 00:15:42,447
Park trying to make its
way over from that side

362
00:15:42,447 --> 00:15:44,280
all the way over to the
junction and across.

363
00:15:44,280 --> 00:15:46,520
The minority carrier is
at risk of recombining.

364
00:15:46,520 --> 00:15:48,870
And many of them don't make it.

365
00:15:48,870 --> 00:15:51,530
So we have the following graph--
the collection probability

366
00:15:51,530 --> 00:15:54,870
versus distance into the
device where this yellow region

367
00:15:54,870 --> 00:15:57,980
represents the space charge
region-- where the probability

368
00:15:57,980 --> 00:15:59,200
of collection is very high.

369
00:15:59,200 --> 00:16:02,060
And as you move away from
the space charge region,

370
00:16:02,060 --> 00:16:04,830
the probability of
collection decreases.

371
00:16:04,830 --> 00:16:08,160
And you can see it
decreases exponentially

372
00:16:08,160 --> 00:16:10,500
as a function of distance
from the space charge region.

373
00:16:10,500 --> 00:16:13,264
The reason you have almost--
not exactly symmetric--

374
00:16:13,264 --> 00:16:14,930
these are two different
decay constants,

375
00:16:14,930 --> 00:16:17,580
but they're both exponential
functions on either side--

376
00:16:17,580 --> 00:16:21,397
is because over here on what
ostensibly would be the-- what

377
00:16:21,397 --> 00:16:24,330
did we call it--
the p-type side--

378
00:16:24,330 --> 00:16:26,282
the electrons would be
the minority carriers.

379
00:16:26,282 --> 00:16:27,740
And over the n-type
side, the holes

380
00:16:27,740 --> 00:16:28,920
are the minority carriers.

381
00:16:28,920 --> 00:16:31,030
And each one is trying
to get to the other side.

382
00:16:31,030 --> 00:16:33,880
So the collection
probability is actually

383
00:16:33,880 --> 00:16:36,330
the collection probability
of minority carriers--

384
00:16:36,330 --> 00:16:37,840
the precise
definition of which is

385
00:16:37,840 --> 00:16:40,320
changing from one side of
the junction to the other.

386
00:16:40,320 --> 00:16:41,680
On one side of the
junction, the electrons

387
00:16:41,680 --> 00:16:42,400
are the minority carriers.

388
00:16:42,400 --> 00:16:44,858
And on the other side, the
holes are the minority carriers.

389
00:16:46,870 --> 00:16:52,480
So by probing at different
wavelengths of light,

390
00:16:52,480 --> 00:16:55,824
we essentially change the
optical absorption coefficient

391
00:16:55,824 --> 00:16:57,990
of the material, which means
that the light will get

392
00:16:57,990 --> 00:16:59,740
absorbed at different depths.

393
00:16:59,740 --> 00:17:02,360
So we might have a very short
wavelength light that probes

394
00:17:02,360 --> 00:17:03,420
up here.

395
00:17:03,420 --> 00:17:05,974
And then, say light is
coming in from this side--

396
00:17:05,974 --> 00:17:07,640
so short wavelengths
light gets absorbed

397
00:17:07,640 --> 00:17:08,869
near the front surface.

398
00:17:08,869 --> 00:17:11,240
And then longer, longer,
longer, and longer, and longer

399
00:17:11,240 --> 00:17:14,400
wavelengths until we're
deep within the material.

400
00:17:14,400 --> 00:17:18,420
And we can flush out exactly
what the current collection

401
00:17:18,420 --> 00:17:21,839
probability would be
as a function of depth.

402
00:17:21,839 --> 00:17:25,710
Then we essentially
take that data

403
00:17:25,710 --> 00:17:29,990
knowing the generation rate
and the collection probability.

404
00:17:29,990 --> 00:17:35,564
We can tease out the minority
carrier diffusion length

405
00:17:35,564 --> 00:17:36,730
using the spectrophotometer.

406
00:17:36,730 --> 00:17:40,620
So let me show you examples
of good and bad cells.

407
00:17:40,620 --> 00:17:43,550
This would be a bad solar cell
and a good solar cell device

408
00:17:43,550 --> 00:17:44,960
right over here.

409
00:17:44,960 --> 00:17:48,010
The good solar cell has a high
internal quantum efficiency

410
00:17:48,010 --> 00:17:49,870
out to longer wavelengths.

411
00:17:49,870 --> 00:17:52,292
And as we know from the very
second lecture of our class,

412
00:17:52,292 --> 00:17:53,750
the longer wavelengths
get absorbed

413
00:17:53,750 --> 00:17:57,180
deeper into the device
further from the junction.

414
00:17:57,180 --> 00:18:00,257
And so you can see this
tail off occurring.

415
00:18:00,257 --> 00:18:02,340
Actually, the tail off is
somewhere in this region

416
00:18:02,340 --> 00:18:05,049
right here for this device
is occurring within the bulk.

417
00:18:05,049 --> 00:18:07,090
At some point, the device
just isn't thick enough

418
00:18:07,090 --> 00:18:08,580
to absorb all the light.

419
00:18:08,580 --> 00:18:12,470
So there is an inevitable drop
toward longer wavelengths.

420
00:18:12,470 --> 00:18:15,176
But the bad cell doesn't
an even poorer job

421
00:18:15,176 --> 00:18:16,800
of collecting these
longer wavelengths.

422
00:18:16,800 --> 00:18:19,270
And that's in part because
the minority carrier diffusion

423
00:18:19,270 --> 00:18:22,535
length of that particular
cell was much lower

424
00:18:22,535 --> 00:18:24,410
or save the minority
carrier diffusion length

425
00:18:24,410 --> 00:18:26,630
of the material which
comprise the cell is

426
00:18:26,630 --> 00:18:28,320
much lower than the good cell.

427
00:18:28,320 --> 00:18:30,570
So if you come in with a
longer wavelength for light--

428
00:18:30,570 --> 00:18:35,177
let's say maybe 933
nanometers in silicon

429
00:18:35,177 --> 00:18:37,635
that would give an absorption
depth of around 100 microns--

430
00:18:37,635 --> 00:18:40,220
that's pretty far
from the junction--

431
00:18:40,220 --> 00:18:43,770
now you would have a much lower
probability in the bad cell

432
00:18:43,770 --> 00:18:45,230
of collecting than
the good cell.

433
00:18:45,230 --> 00:18:49,930
And that's precisely because
the diffusion length is lower.

434
00:18:49,930 --> 00:18:52,950
So integrated over all
of the wavelengths,

435
00:18:52,950 --> 00:18:56,920
one obtains the
short-circuit current.

436
00:18:56,920 --> 00:19:00,270
And the method of measurement
was the spectrophotometer.

437
00:19:00,270 --> 00:19:02,200
Now, I understand
most of you opted

438
00:19:02,200 --> 00:19:05,460
to choose Sun's Voc for your
one characterization tool

439
00:19:05,460 --> 00:19:07,150
to apply to your cells.

440
00:19:07,150 --> 00:19:10,267
Did anybody choose
a spectrophotometer?

441
00:19:10,267 --> 00:19:10,850
Show of hands.

442
00:19:10,850 --> 00:19:11,620
Anybody?

443
00:19:11,620 --> 00:19:12,620
AUDIENCE: It was broken.

444
00:19:12,620 --> 00:19:15,220
PROFESSOR: It was broken, yeah,
so the filter wheel was down.

445
00:19:15,220 --> 00:19:16,270
That was it.

446
00:19:16,270 --> 00:19:18,570
So let's see, the filter
wheel-- the filter wheel.

447
00:19:18,570 --> 00:19:20,400
Let's go fixing this.

448
00:19:20,400 --> 00:19:21,140
Here it is.

449
00:19:21,140 --> 00:19:23,890
So the filter wheel was broken.

450
00:19:23,890 --> 00:19:27,390
So essentially, the
polychromatic light source

451
00:19:27,390 --> 00:19:29,640
was shining through a filter
wheel which was selecting

452
00:19:29,640 --> 00:19:31,496
one wavelength of light.

453
00:19:31,496 --> 00:19:32,870
The monochrometer,
of course, was

454
00:19:32,870 --> 00:19:37,090
helping as well and eventually
on to the solar cell

455
00:19:37,090 --> 00:19:40,940
sample, which was shown
in front of the light.

456
00:19:40,940 --> 00:19:45,920
So that's the way in principle
spectral response works.

457
00:19:45,920 --> 00:19:50,390
And one variant of
spectral response--

458
00:19:50,390 --> 00:19:51,540
that's for the full device.

459
00:19:51,540 --> 00:19:52,900
You typically have
an illuminated area

460
00:19:52,900 --> 00:19:55,400
of a few millimeters squared
or maybe even a few centimeters

461
00:19:55,400 --> 00:19:55,910
squared.

462
00:19:55,910 --> 00:19:58,895
But let's say that wasn't
good enough for you.

463
00:19:58,895 --> 00:20:01,176
You knew you had an
inhomogeneous material

464
00:20:01,176 --> 00:20:02,800
and you want to probe
the inhomogeneity

465
00:20:02,800 --> 00:20:03,770
across your device.

466
00:20:03,770 --> 00:20:05,630
You had a large
device maybe about

467
00:20:05,630 --> 00:20:08,900
that big-- a few 10s
of centimeters squared.

468
00:20:08,900 --> 00:20:10,855
And you want to probe
the distribution

469
00:20:10,855 --> 00:20:13,230
of minority carrier diffusion
lengths across your device.

470
00:20:13,230 --> 00:20:14,430
How would you do that?

471
00:20:14,430 --> 00:20:17,410
Well, in one incarnation,
you would use a much more

472
00:20:17,410 --> 00:20:18,244
finely-focused beam.

473
00:20:18,244 --> 00:20:20,618
Instead of having something
on the order of a millimeter,

474
00:20:20,618 --> 00:20:22,300
you might shrink the
beam spot size down

475
00:20:22,300 --> 00:20:24,900
to the range of single microns.

476
00:20:24,900 --> 00:20:29,062
And then use an xy stepper motor
to scan across your device.

477
00:20:29,062 --> 00:20:31,020
And that's precisely what
this does right here.

478
00:20:31,020 --> 00:20:35,050
This white piece
is the xy stage.

479
00:20:35,050 --> 00:20:39,120
And this black head
right here is essentially

480
00:20:39,120 --> 00:20:43,300
comprised of several lasers that
will shine on to the sample.

481
00:20:43,300 --> 00:20:47,140
Now, you need a few different
wavelengths of light

482
00:20:47,140 --> 00:20:50,390
to really flesh out this curve
to determine the minority

483
00:20:50,390 --> 00:20:52,160
carrier diffusion
length from this curve.

484
00:20:52,160 --> 00:20:54,450
And so the laser light
is typically chosen

485
00:20:54,450 --> 00:21:00,380
at very auspicious wavelengths
to maximize utility

486
00:21:00,380 --> 00:21:03,430
in this particular
type of evaluation.

487
00:21:03,430 --> 00:21:07,600
And so we have usually
four-- a minimum three.

488
00:21:07,600 --> 00:21:09,350
You can always fit a
line through two data

489
00:21:09,350 --> 00:21:13,010
points-- so a minimum three
laser wavelengths and usually

490
00:21:13,010 --> 00:21:16,910
about four or more to process
the quantum efficiency

491
00:21:16,910 --> 00:21:18,290
as a function of position.

492
00:21:18,290 --> 00:21:22,140
And so the xy stage moves the
sample around and the laser

493
00:21:22,140 --> 00:21:24,310
head-- or moves the
laser head around

494
00:21:24,310 --> 00:21:26,620
while the sample stays
fixed and you map out

495
00:21:26,620 --> 00:21:30,120
the current response at
each point on the device,

496
00:21:30,120 --> 00:21:32,990
obtaining a map that
looks much like this right

497
00:21:32,990 --> 00:21:36,510
here where this is on the y
scale or the z scale, sorry,

498
00:21:36,510 --> 00:21:40,580
you have from 0 to
120-micron diffusion length

499
00:21:40,580 --> 00:21:43,320
on that particular
solar cell device.

500
00:21:43,320 --> 00:21:49,300
And points 3, 1, and 2 represent
regions in which the minority

501
00:21:49,300 --> 00:21:51,530
carrier diffusion
with was calculated

502
00:21:51,530 --> 00:21:54,460
using the method that we have
discussed in the previous

503
00:21:54,460 --> 00:21:56,979
lectures-- the
Paul Basore method.

504
00:21:56,979 --> 00:21:58,895
So we have a map of
minority carrier diffusion

505
00:21:58,895 --> 00:22:00,780
length across our device.

506
00:22:00,780 --> 00:22:02,750
Some regions are
underperforming.

507
00:22:02,750 --> 00:22:04,900
Other regions are
performing rather well.

508
00:22:04,900 --> 00:22:08,920
And we can see that in general,
these lower-cost materials

509
00:22:08,920 --> 00:22:12,930
tend to be fairly inhomogeneous
in terms of their performance

510
00:22:12,930 --> 00:22:13,430
response.

511
00:22:16,370 --> 00:22:19,520
And if you really want
to get fancy and say,

512
00:22:19,520 --> 00:22:23,300
my goodness, if I spend all
this time mapping my device,

513
00:22:23,300 --> 00:22:26,210
I'm not going to get through
many devices during my Ph.D.

514
00:22:26,210 --> 00:22:29,550
I'm going tour have to make
do with very poor statistics.

515
00:22:29,550 --> 00:22:31,224
How do I speed this up?

516
00:22:31,224 --> 00:22:32,890
Some very clever
people have figured out

517
00:22:32,890 --> 00:22:35,730
that if you fire your
laser diode simultaneously

518
00:22:35,730 --> 00:22:38,040
but with different
frequencies-- each of them

519
00:22:38,040 --> 00:22:40,402
with its own frequency--
you can deconvolute

520
00:22:40,402 --> 00:22:41,860
the current response
of your device

521
00:22:41,860 --> 00:22:45,205
using a Fourier transform to
pick out the current response

522
00:22:45,205 --> 00:22:49,147
at each portion of the frequency
space-- in other words,

523
00:22:49,147 --> 00:22:50,980
decouple the different
wavelengths of light.

524
00:22:50,980 --> 00:22:53,438
So it's a little clever method
involving Fourier transforms

525
00:22:53,438 --> 00:22:56,710
and lock-in method to speed up
the measurement a little bit--

526
00:22:56,710 --> 00:23:00,600
small aisde-- small
footnote there.

527
00:23:00,600 --> 00:23:02,700
Important-- minority
carrier diffusion length.

528
00:23:02,700 --> 00:23:04,790
That's of utmost importance.

529
00:23:04,790 --> 00:23:07,320
So we can relate the
diffusion length directly back

530
00:23:07,320 --> 00:23:11,890
to our IV curve via the
saturation current-- via the J

531
00:23:11,890 --> 00:23:14,330
nought or I nought, depending
on whether you're looking

532
00:23:14,330 --> 00:23:16,440
at density or absolute amount.

533
00:23:16,440 --> 00:23:17,970
Let's shift gears a bit.

534
00:23:17,970 --> 00:23:21,500
I want to talk about
the Voc measurements

535
00:23:21,500 --> 00:23:23,420
and fill factor
loss measurements.

536
00:23:23,420 --> 00:23:27,380
So how do you determine
losses in Voc and fill factor?

537
00:23:27,380 --> 00:23:29,030
Before I jump into
this, does anybody

538
00:23:29,030 --> 00:23:32,840
have any questions about
Jsc loss mechanisms?

539
00:23:32,840 --> 00:23:34,400
So sad that the
spectrophotometer

540
00:23:34,400 --> 00:23:35,760
wasn't up and running.

541
00:23:35,760 --> 00:23:38,100
Sniff.

542
00:23:38,100 --> 00:23:41,120
It should be fixed
relatively soon, right Joe?

543
00:23:41,120 --> 00:23:42,400
It's fixed now?

544
00:23:42,400 --> 00:23:43,930
All right.

545
00:23:43,930 --> 00:23:47,935
So we have I think David Berney
Needleman is back on board,

546
00:23:47,935 --> 00:23:48,550
right?

547
00:23:48,550 --> 00:23:49,342
AUDIENCE: Rupac is.

548
00:23:49,342 --> 00:23:50,633
PROFESSOR: Rupac as well, yeah.

549
00:23:50,633 --> 00:23:52,300
So if we have folks
who still want

550
00:23:52,300 --> 00:23:54,950
to measure their devices using
spectrophotometer measurement,

551
00:23:54,950 --> 00:23:56,610
we can probably
make arrangements

552
00:23:56,610 --> 00:23:58,630
for say what, 10
cells to be measured?

553
00:23:58,630 --> 00:23:59,721
Something in that range?

554
00:23:59,721 --> 00:24:00,897
AUDIENCE: There were
only three people

555
00:24:00,897 --> 00:24:02,105
who want to do it originally.

556
00:24:04,780 --> 00:24:07,010
PROFESSOR: OK, let's
fit them in-- the three

557
00:24:07,010 --> 00:24:08,760
people who wanted to do them.

558
00:24:08,760 --> 00:24:10,532
And we'll have
our full data set.

559
00:24:10,532 --> 00:24:11,710
Good.

560
00:24:11,710 --> 00:24:16,090
So let's describe fill
factor and VOC losses.

561
00:24:16,090 --> 00:24:22,520
And this is venturing
into operating conditions.

562
00:24:22,520 --> 00:24:25,110
Under short circuit
conditions, how much power

563
00:24:25,110 --> 00:24:27,440
is running through
the solar cell device?

564
00:24:27,440 --> 00:24:30,545
Power defined as
current voltage product.

565
00:24:30,545 --> 00:24:31,204
AUDIENCE: Zero.

566
00:24:31,204 --> 00:24:32,412
PROFESSOR: Zero, why is that?

567
00:24:32,412 --> 00:24:33,287
AUDIENCE: No voltage.

568
00:24:33,287 --> 00:24:34,810
PROFESSOR: No voltage, right?

569
00:24:34,810 --> 00:24:38,790
So you're not testing your
solar cell under ideal operating

570
00:24:38,790 --> 00:24:39,400
conditions.

571
00:24:39,400 --> 00:24:45,260
It's an artificial
operating condition simply

572
00:24:45,260 --> 00:24:49,820
to probe a bulk property
characteristics-- to minimize

573
00:24:49,820 --> 00:24:51,640
the effect of the junction.

574
00:24:51,640 --> 00:24:53,600
Now, if you really
want to see how does

575
00:24:53,600 --> 00:24:57,080
is the junction behaving-- how
that charge separation going--

576
00:24:57,080 --> 00:25:00,380
you might want to venture
into forward bias conditions

577
00:25:00,380 --> 00:25:02,780
and eventually into open
circuit voltage conditions

578
00:25:02,780 --> 00:25:06,200
to really test what the junction
quality is instead of just

579
00:25:06,200 --> 00:25:08,580
probing bulk properties.

580
00:25:08,580 --> 00:25:12,280
So let's talk about some of
the measurement techniques

581
00:25:12,280 --> 00:25:15,910
to really get to the heart of
how are solar cells performing.

582
00:25:15,910 --> 00:25:17,690
It's helpful to do
the Jsc measurements.

583
00:25:17,690 --> 00:25:21,680
It helps you predict what might
be some of the loss mechanisms

584
00:25:21,680 --> 00:25:22,210
later on.

585
00:25:22,210 --> 00:25:24,460
But for this really
gets at the meat of it.

586
00:25:24,460 --> 00:25:26,030
IV curve measurements-- check.

587
00:25:26,030 --> 00:25:27,300
We studied that.

588
00:25:27,300 --> 00:25:28,760
We did that in the laboratory.

589
00:25:28,760 --> 00:25:30,385
And I think we have
a pretty good sense

590
00:25:30,385 --> 00:25:32,580
of what's going on there now.

591
00:25:32,580 --> 00:25:35,660
Series resistance losses-- we
talked earlier in the class

592
00:25:35,660 --> 00:25:37,520
about context and sheet
resistance losses.

593
00:25:37,520 --> 00:25:40,090
Now we'll go back
to it again just

594
00:25:40,090 --> 00:25:44,170
to revisit so we have the
materials necessary to write up

595
00:25:44,170 --> 00:25:46,040
the report here on this project.

596
00:25:46,040 --> 00:25:49,540
Shunt resistance-- specifically
in shunt resistance

597
00:25:49,540 --> 00:25:51,090
we'll talk about
lock-in thermography

598
00:25:51,090 --> 00:25:52,420
and electroluminescence.

599
00:25:52,420 --> 00:25:55,182
And finally, we'll have
a small slide on Suns-Voc

600
00:25:55,182 --> 00:25:56,640
and talk about that
as well since I

601
00:25:56,640 --> 00:25:58,650
know the majority
of you selected that

602
00:25:58,650 --> 00:26:01,850
as your measurement tool.

603
00:26:01,850 --> 00:26:06,171
All right, so open
circuit voltage--

604
00:26:06,171 --> 00:26:07,670
reading straight
off the slide here.

605
00:26:07,670 --> 00:26:09,419
"If the collected
light-generated carriers

606
00:26:09,419 --> 00:26:11,940
are not extracted in the solar
cell, but instead remain,

607
00:26:11,940 --> 00:26:14,260
then a charge separation
exists across,

608
00:26:14,260 --> 00:26:17,250
meaning there's a
buildup of charge.

609
00:26:17,250 --> 00:26:20,610
And the charge separation
reduces the electric field

610
00:26:20,610 --> 00:26:22,550
in the depletion
region, which reduces

611
00:26:22,550 --> 00:26:25,870
the barrier to diffusion
current and causes the diffusion

612
00:26:25,870 --> 00:26:27,000
current to flow.

613
00:26:27,000 --> 00:26:30,330
In words, if you have light
coming into your device,

614
00:26:30,330 --> 00:26:33,559
but you're not able to extract
the charges from either side

615
00:26:33,559 --> 00:26:35,350
to make sure that the
chemical potential is

616
00:26:35,350 --> 00:26:37,550
equal on both sides,
the chemical potential

617
00:26:37,550 --> 00:26:38,880
will change on either side.

618
00:26:38,880 --> 00:26:40,740
You'll have a
buildup of one charge

619
00:26:40,740 --> 00:26:42,720
type on one side of the
junction and build up

620
00:26:42,720 --> 00:26:45,303
of the opposite charge type on
the other side of the junction.

621
00:26:45,303 --> 00:26:48,465
And so that will
shift the Fermi energy

622
00:26:48,465 --> 00:26:49,840
on the other side
of the junction

623
00:26:49,840 --> 00:26:52,400
and result in a
bias of your device

624
00:26:52,400 --> 00:26:55,900
and eventually counteract
the built-in electric field

625
00:26:55,900 --> 00:26:59,930
to the point where the diffusion
current is now enabled to flow.

626
00:26:59,930 --> 00:27:01,525
All this sounding
familiar to folks.

627
00:27:01,525 --> 00:27:02,700
It might be a little rusty.

628
00:27:02,700 --> 00:27:05,020
But it's this all there.

629
00:27:05,020 --> 00:27:08,750
So the idea now is to
begin probing that junction

630
00:27:08,750 --> 00:27:09,450
condition.

631
00:27:09,450 --> 00:27:12,660
We talked about the
ideal diode equation.

632
00:27:12,660 --> 00:27:15,510
And now we're going to
piece through it once again.

633
00:27:15,510 --> 00:27:19,120
This is your current
density here.

634
00:27:19,120 --> 00:27:23,460
And I believe if you really
want to be precise about this,

635
00:27:23,460 --> 00:27:26,060
you should use straight
current and not

636
00:27:26,060 --> 00:27:28,390
current density--
the reason being

637
00:27:28,390 --> 00:27:31,110
if you look at current
resistance product, that

638
00:27:31,110 --> 00:27:32,230
gives you voltage.

639
00:27:32,230 --> 00:27:33,970
But current density
resistance product

640
00:27:33,970 --> 00:27:35,670
gives you essentially
a voltage density,

641
00:27:35,670 --> 00:27:38,490
which is not a unit
that we typically use.

642
00:27:38,490 --> 00:27:41,270
So just keep that in mind
as a small, little asterisk.

643
00:27:41,270 --> 00:27:44,430
This equation is
often used in PV.

644
00:27:44,430 --> 00:27:47,720
That little unit conversion
issue is generally ignored,

645
00:27:47,720 --> 00:27:50,830
but you might want to flag
it for your own benefit.

646
00:27:50,830 --> 00:27:53,590
So we have the current
output of our device

647
00:27:53,590 --> 00:27:55,130
as a function of voltage.

648
00:27:55,130 --> 00:27:57,910
And this is a rather complicated
expression going well

649
00:27:57,910 --> 00:27:59,640
beyond the ideal
diode equation--

650
00:27:59,640 --> 00:28:02,800
the ideal diode equation, which
would consist of the J sub L--

651
00:28:02,800 --> 00:28:04,050
the illumination current.

652
00:28:04,050 --> 00:28:06,567
This is the light coming in
creating the carriers that are

653
00:28:06,567 --> 00:28:07,900
then swept across the junction .

654
00:28:07,900 --> 00:28:13,530
The J sub L is going to
be the integral current

655
00:28:13,530 --> 00:28:17,870
under a spectral response curve
at each wavelength weighted

656
00:28:17,870 --> 00:28:20,540
for the wave length of the
incoming light or the intensity

657
00:28:20,540 --> 00:28:24,060
at that particular bandwidth
of the incoming light.

658
00:28:24,060 --> 00:28:25,460
Figure it this way.

659
00:28:25,460 --> 00:28:27,952
You're measuring the
spectral response

660
00:28:27,952 --> 00:28:29,410
at each particular
wavelength. what

661
00:28:29,410 --> 00:28:30,680
the collection probability is.

662
00:28:30,680 --> 00:28:32,805
Then if you multiply by
the total number of photons

663
00:28:32,805 --> 00:28:35,820
in that wavelength range, you're
determining the total number

664
00:28:35,820 --> 00:28:39,230
of carriers generated within
that frequency range of light.

665
00:28:39,230 --> 00:28:41,460
And if you add up
all the little bins,

666
00:28:41,460 --> 00:28:44,512
you get all of the carriers
generated by all of the light.

667
00:28:44,512 --> 00:28:46,720
And you can adjust depending
what spectral conditions

668
00:28:46,720 --> 00:28:48,990
you have-- say 8.15 spectrum.

669
00:28:48,990 --> 00:28:53,280
So the J sub L here-- this one,
is our short circuit current

670
00:28:53,280 --> 00:28:56,800
effectively and what is derived
from a spectral response

671
00:28:56,800 --> 00:28:57,970
measurement.

672
00:28:57,970 --> 00:29:02,180
The rest of these terms
here are of interest.

673
00:29:02,180 --> 00:29:05,130
So the J-- let's break
down into pieces here.

674
00:29:05,130 --> 00:29:07,450
This term over
here is essentially

675
00:29:07,450 --> 00:29:09,710
to take into account
shunt resistance losses.

676
00:29:09,710 --> 00:29:11,370
If we have shunts
in our device, we're

677
00:29:11,370 --> 00:29:14,420
going to be reducing
the total power output.

678
00:29:14,420 --> 00:29:16,710
These two components--
what are usually

679
00:29:16,710 --> 00:29:19,310
called diffusion current
and recombination current--

680
00:29:19,310 --> 00:29:21,980
this one, you have to
take me for my word at it

681
00:29:21,980 --> 00:29:25,040
right now-- this is the
recombination within the space

682
00:29:25,040 --> 00:29:26,880
charge region.

683
00:29:26,880 --> 00:29:28,640
And this over here
is a recombination

684
00:29:28,640 --> 00:29:31,660
within the bulk
of the solar cell.

685
00:29:31,660 --> 00:29:33,780
Here's a good way
to think about it.

686
00:29:33,780 --> 00:29:36,320
If our solar cell is
not forward biased--

687
00:29:36,320 --> 00:29:38,970
if it's very, very slightly
forward biased-- just

688
00:29:38,970 --> 00:29:43,010
a little bit-- there's going to
be a large barrier for majority

689
00:29:43,010 --> 00:29:45,420
carriers one side the junction
to overcome-- to drive

690
00:29:45,420 --> 00:29:47,530
a drift current--
sorry, diffusion

691
00:29:47,530 --> 00:29:50,180
current-- a very large
barrier to overcome.

692
00:29:50,180 --> 00:29:51,880
And hence, the
recombination is most

693
00:29:51,880 --> 00:29:54,264
likely to occur within
the space charge region.

694
00:29:54,264 --> 00:29:56,430
The carriers aren't likely
to overcome that barrier,

695
00:29:56,430 --> 00:29:58,850
get into the bulk,
and recombine there.

696
00:29:58,850 --> 00:30:01,440
Whereas if we are under large
forward bias conditions,

697
00:30:01,440 --> 00:30:03,750
now the carriers can very
easily go over that barrier

698
00:30:03,750 --> 00:30:05,350
and recombine in the bulk.

699
00:30:05,350 --> 00:30:07,850
So this term right
here dominates

700
00:30:07,850 --> 00:30:09,830
under low forward-bias
conditions.

701
00:30:09,830 --> 00:30:12,520
And this term over here
dominates under larger forward

702
00:30:12,520 --> 00:30:13,900
bias conditions.

703
00:30:13,900 --> 00:30:17,950
And that's why you see when
you plot your IV curve on log

704
00:30:17,950 --> 00:30:22,130
linear scale, you see
essentially two flat points

705
00:30:22,130 --> 00:30:22,910
on your IV curve.

706
00:30:22,910 --> 00:30:25,560
Let me forward that to
the next slide right here.

707
00:30:25,560 --> 00:30:29,110
So this is log of current
versus linear voltage.

708
00:30:29,110 --> 00:30:31,580
And you see one flat
portion right here

709
00:30:31,580 --> 00:30:33,760
and another flat
portion right here

710
00:30:33,760 --> 00:30:36,100
before the series resistance
begins to dominate

711
00:30:36,100 --> 00:30:37,400
at too high voltage.

712
00:30:37,400 --> 00:30:40,670
And your shunt resistance
dominates at too low voltage.

713
00:30:40,670 --> 00:30:44,060
This flat portion right there
is being driven by recombination

714
00:30:44,060 --> 00:30:46,580
within the space charge region.

715
00:30:46,580 --> 00:30:47,680
Why?

716
00:30:47,680 --> 00:30:49,200
Well, the forward
bias voltage isn't

717
00:30:49,200 --> 00:30:51,217
enough to really lower
that barrier enough

718
00:30:51,217 --> 00:30:53,550
to allow the carriers deep
within the bulk to recombine.

719
00:30:53,550 --> 00:30:56,390
So the carriers are recombining
in the space charge region.

720
00:30:56,390 --> 00:31:00,440
Whereas here, they are
recombining within the bulk.

721
00:31:00,440 --> 00:31:05,120
So this is a more
complicated expression

722
00:31:05,120 --> 00:31:09,700
for the IV characteristics
of a solar cell device.

723
00:31:09,700 --> 00:31:12,975
Not always are these
precise mechanisms at work.

724
00:31:12,975 --> 00:31:14,850
If you have a new material
system that you're

725
00:31:14,850 --> 00:31:16,940
working with, there might
be different re combination

726
00:31:16,940 --> 00:31:17,731
mechanisms at work.

727
00:31:17,731 --> 00:31:20,810
There might be charged
accumulation effects at work.

728
00:31:20,810 --> 00:31:23,690
But they can all be encapsulated
in some form of current voltage

729
00:31:23,690 --> 00:31:25,010
expression.

730
00:31:25,010 --> 00:31:28,570
Think of the IV characteristics
of the solar cell

731
00:31:28,570 --> 00:31:30,880
similar to say the
constitutive relations

732
00:31:30,880 --> 00:31:34,710
of a complex
viscoelastic material

733
00:31:34,710 --> 00:31:39,920
where you have say a nice,
linear elastic component,

734
00:31:39,920 --> 00:31:41,170
a dashpot component.

735
00:31:41,170 --> 00:31:43,240
So you have a very
complex expression

736
00:31:43,240 --> 00:31:46,161
describing the stress-strain
relationship in the material.

737
00:31:46,161 --> 00:31:48,160
A similar thing goes on
here with the solar cell

738
00:31:48,160 --> 00:31:50,493
where you're describing the
current voltage relationship

739
00:31:50,493 --> 00:31:53,080
if you understand all of
the mechanisms driving

740
00:31:53,080 --> 00:31:56,270
carrier recombination in your
material and charge separation.

741
00:31:56,270 --> 00:31:57,710
You can come up
with an expression

742
00:31:57,710 --> 00:32:01,350
that describes very precisely
the IV characteristics.

743
00:32:01,350 --> 00:32:03,890
But if you change
material systems,

744
00:32:03,890 --> 00:32:06,590
you may not necessarily be able
to transition the same models

745
00:32:06,590 --> 00:32:07,090
over.

746
00:32:07,090 --> 00:32:08,430
It's helpful to start with them.

747
00:32:08,430 --> 00:32:10,430
But sooner or later you
might decide, well, gee,

748
00:32:10,430 --> 00:32:14,670
I have to make
some modifications.

749
00:32:14,670 --> 00:32:17,460
All right, OK,
series resistance--

750
00:32:17,460 --> 00:32:21,720
we already studied this before,
but just a quick refresher.

751
00:32:21,720 --> 00:32:23,940
We have the series
resistance of the bulk.

752
00:32:23,940 --> 00:32:26,030
We have the series
resistance of the emitter.

753
00:32:26,030 --> 00:32:29,590
And then we have the series
resistance of the contacts

754
00:32:29,590 --> 00:32:31,840
as we contact the
device and then

755
00:32:31,840 --> 00:32:33,840
the series resistance
within the contact itself.

756
00:32:33,840 --> 00:32:37,070
So at least four different
components of series resistance

757
00:32:37,070 --> 00:32:41,540
in our device that are all
lumped together when we

758
00:32:41,540 --> 00:32:43,440
express our IV curve
in this manner.

759
00:32:43,440 --> 00:32:46,237
We all lump them together
in that R sub S. Question?

760
00:32:46,237 --> 00:32:48,320
AUDIENCE: Without going
through the math, is there

761
00:32:48,320 --> 00:32:50,980
a simple explanation
where the 2 comes from?

762
00:32:50,980 --> 00:32:54,060
PROFESSOR: Yeah so simple
explanation, if you would like,

763
00:32:54,060 --> 00:32:57,374
is mostly where the Fermi
energy is sitting relative

764
00:32:57,374 --> 00:32:58,790
to the conduction
or valence band.

765
00:32:58,790 --> 00:33:01,250
And the space charge region--
the quasi neutral region

766
00:33:01,250 --> 00:33:05,910
is closer to mid gap as
opposed to closer to bandage.

767
00:33:05,910 --> 00:33:09,392
Hand wavy-- if you really
want to get into it,

768
00:33:09,392 --> 00:33:10,850
I spent about three
months studying

769
00:33:10,850 --> 00:33:13,630
this as a graduate student.

770
00:33:13,630 --> 00:33:16,530
The 1 and the 2--
the ideality factor,

771
00:33:16,530 --> 00:33:18,930
which is to show more precisely.

772
00:33:18,930 --> 00:33:21,760
It's the number that comes
before the kT for each

773
00:33:21,760 --> 00:33:22,340
of these.

774
00:33:22,340 --> 00:33:24,240
That's called the
ideality factor.

775
00:33:24,240 --> 00:33:27,660
It typically expresses
N sub 1 or N sub 2.

776
00:33:27,660 --> 00:33:31,460
In this case I just substituted
straight out for a 1 and a 2.

777
00:33:31,460 --> 00:33:34,330
Those numbers aren't
always 1 and 2.

778
00:33:34,330 --> 00:33:36,330
As a matter of fact,
especially the G nought 1--

779
00:33:36,330 --> 00:33:38,038
the bulk recombination
current, depending

780
00:33:38,038 --> 00:33:40,060
on where the defect
levels lie, it

781
00:33:40,060 --> 00:33:44,120
can be somewhere between
say 1.1 and 1.4 typically.

782
00:33:44,120 --> 00:33:51,380
And beautiful thesis work
done by a fellow in Australia

783
00:33:51,380 --> 00:33:55,230
hold on-- Keith McIntosh
And the title of the thesis

784
00:33:55,230 --> 00:33:58,240
is called "Humps,
Bumps, and Lumps."

785
00:33:58,240 --> 00:34:01,800
It's all about non idealities
within an IV curve.

786
00:34:01,800 --> 00:34:03,870
So if you're really
curious about that,

787
00:34:03,870 --> 00:34:07,042
"Humps, Bumps, and
Lumps," by McIntosh

788
00:34:07,042 --> 00:34:08,000
I believe was the name.

789
00:34:11,300 --> 00:34:13,500
So we talked about series
resistance already.

790
00:34:13,500 --> 00:34:14,969
Hold that-- put
that in your RAM.

791
00:34:14,969 --> 00:34:17,710
That's our R sub
S right over here.

792
00:34:17,710 --> 00:34:19,357
That's our R sub
S. And our shunt--

793
00:34:19,357 --> 00:34:21,440
you want to hold this other
component in your RAM.

794
00:34:21,440 --> 00:34:23,409
So that's the picture
to have of our series.

795
00:34:23,409 --> 00:34:25,540
This is our shunt.

796
00:34:25,540 --> 00:34:28,960
This is a p-n junction.

797
00:34:28,960 --> 00:34:31,080
This is your n plus region.

798
00:34:31,080 --> 00:34:32,150
And this is your p.

799
00:34:32,150 --> 00:34:36,790
And this is looking at
the electron energy in 3D.

800
00:34:36,790 --> 00:34:41,020
So we typically take a slice
out like that just like this.

801
00:34:41,020 --> 00:34:44,139
And we draw the electron
energy as a function

802
00:34:44,139 --> 00:34:47,260
of position in a
nice little wavy form

803
00:34:47,260 --> 00:34:49,706
that we typically do to
describe the p-n junction.

804
00:34:49,706 --> 00:34:51,080
What's being
described right here

805
00:34:51,080 --> 00:34:55,560
is adding a leader dimension--
a spatial dimension in what's

806
00:34:55,560 --> 00:34:57,680
shown as the
abscissa in this plot

807
00:34:57,680 --> 00:35:00,590
and showing what is a
weakness in the p-n junction.

808
00:35:00,590 --> 00:35:02,390
It could be a localized defect.

809
00:35:02,390 --> 00:35:04,900
This could be a
shard of metal that

810
00:35:04,900 --> 00:35:07,807
happened to lie in the wrong
place in your device and fire

811
00:35:07,807 --> 00:35:09,640
through, essentially
when you did the firing

812
00:35:09,640 --> 00:35:12,334
step to adhere the metal
to create the contact.

813
00:35:12,334 --> 00:35:14,500
It could have gone straight
through the p-n junction

814
00:35:14,500 --> 00:35:16,630
and contacted the other side.

815
00:35:16,630 --> 00:35:19,280
It could be a
charged dislocation.

816
00:35:19,280 --> 00:35:23,580
It could be structural
defect, but a local reduction

817
00:35:23,580 --> 00:35:24,849
in the barrier height.

818
00:35:24,849 --> 00:35:27,390
Now where is the electron going
to go if you have a diffusion

819
00:35:27,390 --> 00:35:28,060
current?

820
00:35:28,060 --> 00:35:29,990
It's going to be crowding
through that little,

821
00:35:29,990 --> 00:35:32,710
localized, reduction
in the energy barrier.

822
00:35:32,710 --> 00:35:34,820
And you're going to have
a higher current flowing

823
00:35:34,820 --> 00:35:36,580
through the specific region.

824
00:35:36,580 --> 00:35:38,610
So how to measure that?

825
00:35:38,610 --> 00:35:41,630
Well, if I make sure my
full IV curve like this,

826
00:35:41,630 --> 00:35:44,710
it might manifest itself as
some reduced shunt resistance.

827
00:35:44,710 --> 00:35:46,970
So you'll have a larger
current flowing through it

828
00:35:46,970 --> 00:35:49,200
at lower bias voltages.

829
00:35:49,200 --> 00:35:51,230
But how do I know
what's going on?

830
00:35:51,230 --> 00:35:54,070
Without some spatial-- a
spatially-resolved measurement

831
00:35:54,070 --> 00:35:56,670
tool, I am in the
dark-- no pun intended.

832
00:35:56,670 --> 00:36:00,300
So I have to figure out how
to measure the current flow

833
00:36:00,300 --> 00:36:02,320
as a function of position.

834
00:36:02,320 --> 00:36:04,040
Now somebody thought
about this a while.

835
00:36:04,040 --> 00:36:06,123
And they said, ha, well,
let me get this straight.

836
00:36:06,123 --> 00:36:08,930
If I apply a known bias
voltage to my device,

837
00:36:08,930 --> 00:36:11,020
I'm going to get some
current flowing through it.

838
00:36:11,020 --> 00:36:14,550
And that current-- what I
read out of the whole device

839
00:36:14,550 --> 00:36:18,260
microscopically has some
spatial distribution to it.

840
00:36:18,260 --> 00:36:19,260
Let me think about this.

841
00:36:19,260 --> 00:36:23,610
OK, so current times
voltage is power.

842
00:36:23,610 --> 00:36:25,340
And power generates heat.

843
00:36:25,340 --> 00:36:29,710
So if I have a current crowding
at some portion of my device,

844
00:36:29,710 --> 00:36:32,110
I'm going to be generating
a large amount of heat

845
00:36:32,110 --> 00:36:35,679
as the carriers flow through
that specific region.

846
00:36:35,679 --> 00:36:37,220
So if I have some
method of measuring

847
00:36:37,220 --> 00:36:40,380
the heat-- the heat
distribution across my device

848
00:36:40,380 --> 00:36:43,420
is very, very sensitive--
I might be able to divide

849
00:36:43,420 --> 00:36:45,750
that heat measurement--
if calibrated properly

850
00:36:45,750 --> 00:36:47,870
to determine total
power-- I might

851
00:36:47,870 --> 00:36:52,030
be able to divide that
heat measurement by voltage

852
00:36:52,030 --> 00:36:54,090
so I know the applied
bias voltage to my device.

853
00:36:54,090 --> 00:36:55,840
I take my power, divide
it by the voltage,

854
00:36:55,840 --> 00:36:58,640
and I get the current as
a function of xy position.

855
00:36:58,640 --> 00:37:02,940
So I can map out my dark IV
curve at any arbitrary bias

856
00:37:02,940 --> 00:37:04,260
voltage condition.

857
00:37:04,260 --> 00:37:06,330
Let's say I want to do
the measurement at 0.4

858
00:37:06,330 --> 00:37:07,850
volts forward bias.

859
00:37:07,850 --> 00:37:11,212
I apply 0.4 volts
across my device.

860
00:37:11,212 --> 00:37:12,920
And I just measure
the heat distribution.

861
00:37:12,920 --> 00:37:14,500
And in the calibrated
measurement,

862
00:37:14,500 --> 00:37:16,420
you can see how
the dark current is

863
00:37:16,420 --> 00:37:20,060
distributed-- how the diffusion
current is distributed.

864
00:37:20,060 --> 00:37:23,880
And that's what lock-in
thermography is all about.

865
00:37:23,880 --> 00:37:28,990
This is a somewhat
typical image.

866
00:37:28,990 --> 00:37:31,700
It's still very noisy because
it was a very fast image.

867
00:37:31,700 --> 00:37:33,450
But you have here a
solar cell device.

868
00:37:33,450 --> 00:37:35,991
You can barely make out the
contact grid right here.

869
00:37:35,991 --> 00:37:37,490
And then the fingers
are moving down

870
00:37:37,490 --> 00:37:38,865
separated by about this much.

871
00:37:38,865 --> 00:37:40,240
On this particular
image, you can

872
00:37:40,240 --> 00:37:42,281
see the fingers or the
dark shadow of the fingers

873
00:37:42,281 --> 00:37:43,060
right here.

874
00:37:43,060 --> 00:37:46,285
And you see bright
spots 1, 2, 3, 4, 5, 6,

875
00:37:46,285 --> 00:37:50,120
7 across the device,
indicating regions

876
00:37:50,120 --> 00:37:52,070
where more current is flowing.

877
00:37:52,070 --> 00:37:52,570
Aha!

878
00:37:52,570 --> 00:37:55,050
So I'm beginning to
figure out that, well,

879
00:37:55,050 --> 00:37:57,460
this dark IV curve that
I have right here--

880
00:37:57,460 --> 00:37:59,710
this large amount of current
flowing through my device

881
00:37:59,710 --> 00:38:01,050
at lower bias voltages.

882
00:38:01,050 --> 00:38:03,480
That's not all
uniformly distributed.

883
00:38:03,480 --> 00:38:06,080
But it's rather concentrated
in specific areas

884
00:38:06,080 --> 00:38:10,390
where I have weaknesses
in my p-n junction

885
00:38:10,390 --> 00:38:12,790
under low forward bias voltages.

886
00:38:12,790 --> 00:38:14,700
And then at larger
forward bias voltages

887
00:38:14,700 --> 00:38:17,969
once this barrier is
significantly reduced,

888
00:38:17,969 --> 00:38:20,010
I may have more carriers
going through in regions

889
00:38:20,010 --> 00:38:21,884
of smaller minority
carrier diffusion length.

890
00:38:21,884 --> 00:38:23,290
Why?

891
00:38:23,290 --> 00:38:25,870
Well, because if the carrier
goes in and recombines quickly

892
00:38:25,870 --> 00:38:27,830
because the diffusion
length is very short,

893
00:38:27,830 --> 00:38:29,840
that means there's one
less carrier in there,

894
00:38:29,840 --> 00:38:31,756
which means that now
there's a greater driving

895
00:38:31,756 --> 00:38:32,920
force for diffusion.

896
00:38:32,920 --> 00:38:34,900
And another care
is going to come in

897
00:38:34,900 --> 00:38:38,110
and recombine too and another
carrier and recombine too.

898
00:38:38,110 --> 00:38:40,394
And so now you have a
larger diffusion current

899
00:38:40,394 --> 00:38:42,810
going into the regions of lower
minority carrier diffusion

900
00:38:42,810 --> 00:38:43,462
length.

901
00:38:43,462 --> 00:38:45,420
Let's look at a few
lock-in thermography images

902
00:38:45,420 --> 00:38:48,960
and see how this plays out.

903
00:38:48,960 --> 00:38:54,810
So under low-bias conditions--
so I'm still-- still

904
00:38:54,810 --> 00:38:57,670
have a very large barrier there
in the p-n junction region.

905
00:38:57,670 --> 00:39:01,640
I'm typically going to see
isolated hot spots-- not

906
00:39:01,640 --> 00:39:03,970
always, but typically
isolated hot spots.

907
00:39:03,970 --> 00:39:08,342
And these most often are
defects within the p-n junction.

908
00:39:08,342 --> 00:39:10,050
Somebody must have
scratched the junction

909
00:39:10,050 --> 00:39:14,180
or maybe dropped their tweezer
on it or maybe a piece of metal

910
00:39:14,180 --> 00:39:16,210
during the contact
metalization fired through

911
00:39:16,210 --> 00:39:17,640
at that particular region.

912
00:39:17,640 --> 00:39:19,994
And so we have what are
typically these point shunts.

913
00:39:19,994 --> 00:39:21,660
Sometimes you see
shunts around the edge

914
00:39:21,660 --> 00:39:24,790
as well if you didn't
isolate the edge well.

915
00:39:24,790 --> 00:39:27,660
And now as we forward bias,
forward bias, forward bias,

916
00:39:27,660 --> 00:39:30,150
those barriers become
less important.

917
00:39:30,150 --> 00:39:32,956
And now what's driving the
current-- the dark current

918
00:39:32,956 --> 00:39:34,830
in the device is a
recombination of the bulk.

919
00:39:34,830 --> 00:39:37,590
So the regions of
higher recombination

920
00:39:37,590 --> 00:39:39,800
will be regions of
higher current flow.

921
00:39:39,800 --> 00:39:41,590
They'll drive more
recombination current.

922
00:39:41,590 --> 00:39:45,260
So this is a higher bias
voltage-- 560 millivolts

923
00:39:45,260 --> 00:39:48,680
or 0.56 volts as
opposed to 0.36 volts.

924
00:39:48,680 --> 00:39:51,540
We can still see the shadow
of the big three right here.

925
00:39:51,540 --> 00:39:53,850
But they're much diminished.

926
00:39:53,850 --> 00:39:56,660
Now instead what we see
are these wispier features

927
00:39:56,660 --> 00:39:58,250
right here.

928
00:39:58,250 --> 00:40:02,330
Now I think gee, if that
really is-- if those really

929
00:40:02,330 --> 00:40:05,254
are regions of lower minority
carrier diffusion length--

930
00:40:05,254 --> 00:40:07,670
I just learned about a methods
to measure minority carrier

931
00:40:07,670 --> 00:40:09,420
diffusion length, didn't I?

932
00:40:09,420 --> 00:40:11,160
What was that?

933
00:40:11,160 --> 00:40:14,334
Laser beam induced
current method?

934
00:40:14,334 --> 00:40:16,250
Spatially-resolved laser
beam in this current?

935
00:40:16,250 --> 00:40:18,720
So I have at least one method
of measuring minority carrier

936
00:40:18,720 --> 00:40:19,520
diffusion length.

937
00:40:19,520 --> 00:40:21,980
Let's put one and one
together and see what we get.

938
00:40:21,980 --> 00:40:24,870
So we'll take this
image right here

939
00:40:24,870 --> 00:40:27,620
and put it aside,
rotate it, and put it

940
00:40:27,620 --> 00:40:30,210
right beside the minority
carrier diffusion length.

941
00:40:30,210 --> 00:40:34,011
And voila, you see how
those wispier features

942
00:40:34,011 --> 00:40:36,260
are corresponding to regions
of lower minority carrier

943
00:40:36,260 --> 00:40:39,260
diffusion length-- once
again, the explanation.

944
00:40:39,260 --> 00:40:41,740
If you have large
forward bias condition,

945
00:40:41,740 --> 00:40:43,080
the barrier is lower.

946
00:40:43,080 --> 00:40:44,982
Carriers can easily
diffuse into the bulk.

947
00:40:44,982 --> 00:40:46,440
The carriers defusing
into the bulk

948
00:40:46,440 --> 00:40:48,610
will go a certain distance
before recombining.

949
00:40:48,610 --> 00:40:50,140
If the diffusion
length is short,

950
00:40:50,140 --> 00:40:51,800
they'll recombine
quickly, which means

951
00:40:51,800 --> 00:40:54,710
now there's no more carrier--
which means the diffusion

952
00:40:54,710 --> 00:40:57,130
current will push another
carrier into that spot, which

953
00:40:57,130 --> 00:40:58,561
means they'll recombine.

954
00:40:58,561 --> 00:41:00,560
More current is flowing
through if the diffusion

955
00:41:00,560 --> 00:41:01,820
length is shorter.

956
00:41:01,820 --> 00:41:03,335
The alternative is if the
diffusion length is long

957
00:41:03,335 --> 00:41:04,970
and one carrier makes
it over and then takes

958
00:41:04,970 --> 00:41:06,470
its time getting
all the way across

959
00:41:06,470 --> 00:41:08,350
before recombination occurs.

960
00:41:08,350 --> 00:41:11,600
So what you're seeing
here is essentially

961
00:41:11,600 --> 00:41:16,010
an effect-- the current
the dark, forward current,

962
00:41:16,010 --> 00:41:18,550
meaning you're measuring
the IV curve in the dark.

963
00:41:18,550 --> 00:41:20,050
You're obtaining an IV curve.

964
00:41:20,050 --> 00:41:21,010
Here it is.

965
00:41:21,010 --> 00:41:23,650
You're obtaining an IV
curve for your device.

966
00:41:23,650 --> 00:41:25,900
You're measuring in the dark
under larger forward bias

967
00:41:25,900 --> 00:41:28,750
conditions somewhere around
here where bulk recombination is

968
00:41:28,750 --> 00:41:29,740
dominating.

969
00:41:29,740 --> 00:41:32,990
And you're able to
visualize the current loss

970
00:41:32,990 --> 00:41:35,000
mechanisms in your device.

971
00:41:35,000 --> 00:41:38,650
So you say, well, gee
if I want in the dark--

972
00:41:38,650 --> 00:41:41,090
if I want this to be
as small as possible

973
00:41:41,090 --> 00:41:43,542
so that when I transpose
this in the light

974
00:41:43,542 --> 00:41:46,000
and I shift everything into
negative-- into fourth quadrant

975
00:41:46,000 --> 00:41:49,950
territory because I illuminate
it now-- if I want there to be

976
00:41:49,950 --> 00:41:53,250
a small, dark forward
current as possible

977
00:41:53,250 --> 00:41:56,820
so as I maximize my fill factor
when I illuminate my device,

978
00:41:56,820 --> 00:42:01,720
I want all of these current
loss mechanisms to go away.

979
00:42:01,720 --> 00:42:06,637
Let me repeat that one
more time so people get it.

980
00:42:06,637 --> 00:42:08,470
So what I'm doing right
now is I'm measuring

981
00:42:08,470 --> 00:42:10,050
this device in the dark.

982
00:42:10,050 --> 00:42:12,890
And this is my IV
curve right here.

983
00:42:12,890 --> 00:42:18,370
And I have an example of
bad device-- good device.

984
00:42:18,370 --> 00:42:23,814
The bad device has--
so bad and good.

985
00:42:23,814 --> 00:42:25,230
The reason this
is bad and good is

986
00:42:25,230 --> 00:42:28,390
because when I transpose this
curve under illumination--

987
00:42:28,390 --> 00:42:34,300
this is now under illumination--
my good and bad right here--

988
00:42:34,300 --> 00:42:37,350
you'll see that the
bad has a lower Voc.

989
00:42:37,350 --> 00:42:40,450
So the intersection
between these illuminated

990
00:42:40,450 --> 00:42:44,750
curves-- this is
now illuminated.

991
00:42:44,750 --> 00:42:47,050
And this is dark.

992
00:42:47,050 --> 00:42:49,540
Right over here--
these are dark curves.

993
00:42:49,540 --> 00:42:51,094
These here are
illuminated curves.

994
00:42:51,094 --> 00:42:53,010
The intersection between
the illuminated curve

995
00:42:53,010 --> 00:42:56,600
and the x-axis denotes
the open circuit voltage.

996
00:42:56,600 --> 00:43:00,620
And you see that the larger
this current is in the dark

997
00:43:00,620 --> 00:43:04,370
the earlier you're going to
intersect with your x-axis--

998
00:43:04,370 --> 00:43:07,280
the lower the voltage output of
your device will be-- the lower

999
00:43:07,280 --> 00:43:10,670
the maximum power point will be.

1000
00:43:10,670 --> 00:43:13,530
So want in the dark when
you measure the IV curve,

1001
00:43:13,530 --> 00:43:16,480
you want that current to be as
small as possible in the dark.

1002
00:43:16,480 --> 00:43:18,350
And, obviously, when
you illuminate it,

1003
00:43:18,350 --> 00:43:20,619
you want this jump to
be as large as possible.

1004
00:43:20,619 --> 00:43:22,410
And you want this to
almost look like a box

1005
00:43:22,410 --> 00:43:25,190
to have a large fill factor.

1006
00:43:25,190 --> 00:43:28,180
So if we say, OK, we
want this to be small

1007
00:43:28,180 --> 00:43:30,034
because we want a
large fill factor,

1008
00:43:30,034 --> 00:43:31,950
somehow we have to know
where the current loss

1009
00:43:31,950 --> 00:43:33,400
mechanisms are occurring.

1010
00:43:33,400 --> 00:43:35,950
And that's where we have lock-in
thermography available to us

1011
00:43:35,950 --> 00:43:37,490
we can visualize it.

1012
00:43:37,490 --> 00:43:41,140
You say, well, to do
lock-in thermography well,

1013
00:43:41,140 --> 00:43:44,601
I'm going to need an-- indium
antimonide-- sorry, yeah that

1014
00:43:44,601 --> 00:43:45,100
would be it.

1015
00:43:45,100 --> 00:43:49,514
It would be an indium antimonide
camera sensitive to the 3 to 5

1016
00:43:49,514 --> 00:43:50,930
micron wavelength
range that might

1017
00:43:50,930 --> 00:43:54,580
cost $70,000 to get a good
one with high frame rates.

1018
00:43:54,580 --> 00:43:57,810
The lock-in system-- I don't
know if I have that money.

1019
00:43:57,810 --> 00:44:00,150
I can go over and borrow
Professor Buonassisi's system

1020
00:44:00,150 --> 00:44:01,690
and use it there.

1021
00:44:01,690 --> 00:44:03,410
Or if my shunts
are big enough, I

1022
00:44:03,410 --> 00:44:06,795
can just use straight-out
thermionic sheets

1023
00:44:06,795 --> 00:44:09,190
or thermochromic
sheets, rather, sorry.

1024
00:44:09,190 --> 00:44:10,920
So you can take
these liquid crystals

1025
00:44:10,920 --> 00:44:12,740
and slap them
because your device

1026
00:44:12,740 --> 00:44:15,702
and measure the heat
distribution straight up just

1027
00:44:15,702 --> 00:44:18,430
by making sure you have
good thermal contact

1028
00:44:18,430 --> 00:44:20,880
between your cell and your
thermochromic sheets, which

1029
00:44:20,880 --> 00:44:21,690
cost about $100.

1030
00:44:21,690 --> 00:44:23,940
Now the reason we use the
lock-in thermogrpahy and not

1031
00:44:23,940 --> 00:44:27,670
these thermochromic sheets
is this curve right here.

1032
00:44:27,670 --> 00:44:29,910
This is lock-in
thermography sensitivity

1033
00:44:29,910 --> 00:44:31,080
versus integration time.

1034
00:44:31,080 --> 00:44:33,840
And we're getting into
the 10 microKelvin range,

1035
00:44:33,840 --> 00:44:35,950
which is rattlesnake territory.

1036
00:44:35,950 --> 00:44:38,877
That's how the rattlesnakes are
able to sense heat and reach

1037
00:44:38,877 --> 00:44:41,210
out to you is because they
have those little organs that

1038
00:44:41,210 --> 00:44:44,150
look like little, black--
they're integrating spheres

1039
00:44:44,150 --> 00:44:48,110
to be honest-- small,
little spot to open up.

1040
00:44:48,110 --> 00:44:51,410
Rattlesnake-- that's why
the sidewinder missile.

1041
00:44:51,410 --> 00:44:52,220
Anybody?

1042
00:44:52,220 --> 00:44:53,310
F 16s?

1043
00:44:53,310 --> 00:44:53,810
Hornets?

1044
00:44:53,810 --> 00:44:54,820
No.

1045
00:44:54,820 --> 00:44:58,100
If you're into aviation,
there's a special type

1046
00:44:58,100 --> 00:44:59,880
of heat seeking
missile developed

1047
00:44:59,880 --> 00:45:05,400
I think in the 1950s
that uses a device not

1048
00:45:05,400 --> 00:45:08,860
dissimilar from the rattlesnakes
heat sensing organs.

1049
00:45:08,860 --> 00:45:10,690
Regardless, I
digress, we're looking

1050
00:45:10,690 --> 00:45:14,120
at about a 10
microKelvin sensitivity

1051
00:45:14,120 --> 00:45:16,220
in this lock-in thermography
technique, which

1052
00:45:16,220 --> 00:45:18,920
is very sensitive-- can
measure under low forward bias

1053
00:45:18,920 --> 00:45:21,040
conditions and hence useful.

1054
00:45:25,152 --> 00:45:27,970
So that pretty much sums
up lock-in thermography.

1055
00:45:27,970 --> 00:45:32,990
It is a difficult technique
to wrap your brain around.

1056
00:45:32,990 --> 00:45:37,820
So for those of you who got
75% of it, congratulations.

1057
00:45:37,820 --> 00:45:39,520
That was really,
really well done.

1058
00:45:39,520 --> 00:45:42,090
For those who got about
25% of my explanations

1059
00:45:42,090 --> 00:45:43,730
today, don't worry.

1060
00:45:43,730 --> 00:45:45,120
You're not alone.

1061
00:45:45,120 --> 00:45:47,220
It just takes a
lot of thinking it

1062
00:45:47,220 --> 00:45:50,940
through exactly where is the
current flowing as a function

1063
00:45:50,940 --> 00:45:52,240
of bias condition?

1064
00:45:52,240 --> 00:45:53,844
What does that do
to my IV curve?

1065
00:45:53,844 --> 00:45:56,010
And then what does it do
in a two-dimensional regime

1066
00:45:56,010 --> 00:45:58,030
when I'm measuring heat output?

1067
00:45:58,030 --> 00:45:59,880
It's is a very
complicated thing.

1068
00:45:59,880 --> 00:46:02,260
Thankfully there is an entire
book lock-in thermography

1069
00:46:02,260 --> 00:46:04,380
written by Otwin Breitenstein.

1070
00:46:04,380 --> 00:46:07,690
I gave to you today
a paper-- in one

1071
00:46:07,690 --> 00:46:10,700
of the papers that I
distributed today--

1072
00:46:10,700 --> 00:46:15,710
the one that has the PSSA
written on it-- this one.

1073
00:46:15,710 --> 00:46:19,719
That paper is by
Otwin Breitenstein

1074
00:46:19,719 --> 00:46:22,010
from the Max Planck Institute
of Microstructure Physics

1075
00:46:22,010 --> 00:46:24,680
in Halle, Germany about two
a half and a half hours south

1076
00:46:24,680 --> 00:46:25,690
of Berlin.

1077
00:46:25,690 --> 00:46:28,690
And this describes in more
detail that which I attempted

1078
00:46:28,690 --> 00:46:31,240
to get across in class today.

1079
00:46:31,240 --> 00:46:33,420
He with Martin
Langenkamp have also

1080
00:46:33,420 --> 00:46:35,140
written a book on
lock-in thermography.

1081
00:46:35,140 --> 00:46:37,210
So if you're interested
in more information,

1082
00:46:37,210 --> 00:46:38,920
that's where to go.

1083
00:46:38,920 --> 00:46:44,510
In your handout
set today, we also

1084
00:46:44,510 --> 00:46:48,200
have a classic paper on
series resistance effects

1085
00:46:48,200 --> 00:46:49,910
of solar cell measurements.

1086
00:46:49,910 --> 00:46:53,530
This one is from
'63-- yes, 1963.

1087
00:46:53,530 --> 00:46:56,410
Classic paper worth
the read if you're

1088
00:46:56,410 --> 00:46:58,260
in that regime
making devices that

1089
00:46:58,260 --> 00:47:00,650
are series resistance limited.

1090
00:47:00,650 --> 00:47:02,590
It's more optional
reading for folks.

1091
00:47:02,590 --> 00:47:06,780
And let's venture forward.

1092
00:47:06,780 --> 00:47:09,640
Folks who did these scanning
Elbit techniques for a while

1093
00:47:09,640 --> 00:47:12,590
really found value in
measuring as a function

1094
00:47:12,590 --> 00:47:15,520
of spatially-resolved
manner down

1095
00:47:15,520 --> 00:47:19,030
to a micron in
spatial resolution

1096
00:47:19,030 --> 00:47:20,931
the electrical output
of a solar cell device.

1097
00:47:20,931 --> 00:47:22,680
And then they thought
to themselves, well,

1098
00:47:22,680 --> 00:47:26,340
goodness, if we measure
with one micron spot size,

1099
00:47:26,340 --> 00:47:28,930
it's going to take
us a week to get

1100
00:47:28,930 --> 00:47:32,550
a measurement across a full area
solar device, if not longer.

1101
00:47:32,550 --> 00:47:35,070
And if we measure say
with the 200 nanometer

1102
00:47:35,070 --> 00:47:37,880
spot size using some
sort of X-ray technique

1103
00:47:37,880 --> 00:47:40,090
and we accumulate for
say two seconds a point,

1104
00:47:40,090 --> 00:47:42,550
it can take 10,000 years to
map out a full-size solar cell

1105
00:47:42,550 --> 00:47:44,720
device, literally, do the math.

1106
00:47:44,720 --> 00:47:47,110
And they said is there
an easier way for us

1107
00:47:47,110 --> 00:47:51,030
to acquire information that does
not rely on scanning-- on point

1108
00:47:51,030 --> 00:47:52,557
by point, step by
step measurements.

1109
00:47:52,557 --> 00:47:54,890
And they thought about it
long and hard they said, well,

1110
00:47:54,890 --> 00:47:57,875
gee, these new, handheld,
digital cameras coming out

1111
00:47:57,875 --> 00:47:59,160
on the market are pretty cool.

1112
00:47:59,160 --> 00:48:01,050
What technology do they rely on?

1113
00:48:01,050 --> 00:48:04,820
Oh, it's a CCD-- Charge
Couple Display that visualizes

1114
00:48:04,820 --> 00:48:09,300
or images over a large area
the output of a device.

1115
00:48:09,300 --> 00:48:12,820
And that's exactly how the
lock-in thermography techniques

1116
00:48:12,820 --> 00:48:16,510
were developed here with this
imaging camera-- in that case,

1117
00:48:16,510 --> 00:48:18,930
an indium antimonice CCD.

1118
00:48:18,930 --> 00:48:21,620
And they said, well,
what other wavelengths

1119
00:48:21,620 --> 00:48:24,740
emit-- in what other wavelength
does a solar cell emit?

1120
00:48:24,740 --> 00:48:26,640
We know that it emits heat.

1121
00:48:26,640 --> 00:48:29,820
And heat is typically in the 3
to 5 micron wavelengths regime,

1122
00:48:29,820 --> 00:48:33,550
maybe even as high as 10 microns
depending on what temperature.

1123
00:48:33,550 --> 00:48:37,660
And we know that as well we
can have band to band emission

1124
00:48:37,660 --> 00:48:40,430
inside of a solar cell that
would emit around the band gap

1125
00:48:40,430 --> 00:48:41,430
energy.

1126
00:48:41,430 --> 00:48:43,640
So let's set up
a camera to image

1127
00:48:43,640 --> 00:48:47,340
the band-to-band
recombination intensity

1128
00:48:47,340 --> 00:48:49,300
or the band-to-band
illumination intensity.

1129
00:48:49,300 --> 00:48:50,910
And that's what
this paper is about.

1130
00:48:56,030 --> 00:48:57,160
This is wavelength.

1131
00:48:57,160 --> 00:48:58,590
This is intensity.

1132
00:48:58,590 --> 00:49:01,530
This is the emission
spectrum of a solar cell.

1133
00:49:01,530 --> 00:49:04,880
So if you were to measure what
colors-- what color light does

1134
00:49:04,880 --> 00:49:08,050
a solar cell emit after
you pump current into it,

1135
00:49:08,050 --> 00:49:09,840
this isn't is more
or less the spectrum.

1136
00:49:09,840 --> 00:49:14,450
So we have the thermal
radiation shown over here.

1137
00:49:14,450 --> 00:49:16,034
This is longer wavelength light.

1138
00:49:16,034 --> 00:49:17,450
That's the heat
that it gives off.

1139
00:49:17,450 --> 00:49:20,510
That's what indicates current
flow inside of the device.

1140
00:49:20,510 --> 00:49:24,880
There is the band to band
luminescence right here.

1141
00:49:24,880 --> 00:49:26,845
And we'll neglect the
other two for now.

1142
00:49:26,845 --> 00:49:28,220
Those are more
advanced concepts.

1143
00:49:28,220 --> 00:49:29,470
But the band to
band luminescence

1144
00:49:29,470 --> 00:49:31,011
is essentially just
the recombination

1145
00:49:31,011 --> 00:49:32,780
of a carrier from
the conduction band

1146
00:49:32,780 --> 00:49:35,930
to the valence band--
recombination--

1147
00:49:35,930 --> 00:49:37,110
very straightforward.

1148
00:49:37,110 --> 00:49:41,860
And the recombination
intensity is inversely

1149
00:49:41,860 --> 00:49:46,140
proportional to the
minority carrier lifetime.

1150
00:49:46,140 --> 00:49:48,240
Why is that?

1151
00:49:48,240 --> 00:49:52,470
This was that quote remember
who attended Eli Yablonovich's

1152
00:49:52,470 --> 00:49:52,970
talk.

1153
00:49:52,970 --> 00:49:55,176
He was saying, it's
completely counter intuitive.

1154
00:49:55,176 --> 00:49:56,550
Do you remember
specifically what

1155
00:49:56,550 --> 00:49:58,550
he was referring to in this?

1156
00:49:58,550 --> 00:50:00,897
Let me show you an image here.

1157
00:50:00,897 --> 00:50:02,355
AUDIENCE: That you
want your device

1158
00:50:02,355 --> 00:50:05,740
to emit more light for
higher device performance.

1159
00:50:05,740 --> 00:50:09,860
PROFESSOR: Exactly, yeah,
so what is going on here?

1160
00:50:09,860 --> 00:50:12,220
We have our tau-- one
of our tau effective--

1161
00:50:12,220 --> 00:50:15,490
or one of over lifetime
what we're measuring.

1162
00:50:15,490 --> 00:50:17,720
This is our effective
lifetime that we're measuring.

1163
00:50:17,720 --> 00:50:20,140
And we have one over tau.

1164
00:50:20,140 --> 00:50:22,240
Let's call it non-radiative.

1165
00:50:22,240 --> 00:50:24,720
So we have non-radiative
recombination mechanisms.

1166
00:50:24,720 --> 00:50:26,096
This comprises
Shockley-Read-Hall

1167
00:50:26,096 --> 00:50:27,636
recombination--
Auger recombination--

1168
00:50:27,636 --> 00:50:29,590
a bunch of different
recombination mechanisms.

1169
00:50:29,590 --> 00:50:33,530
And then we have one of our tau
radiative, which is essentially

1170
00:50:33,530 --> 00:50:36,990
this recombination
mechanism shown right here,

1171
00:50:36,990 --> 00:50:40,280
which is just the
band-to-band recombination.

1172
00:50:40,280 --> 00:50:42,260
So if you have
defects around, that

1173
00:50:42,260 --> 00:50:45,430
will lower the lifetime--
the non-radioactive lifetime.

1174
00:50:45,430 --> 00:50:47,220
If you have tons
and tons of defects,

1175
00:50:47,220 --> 00:50:49,990
the carriers will most likely
recombine through the defects

1176
00:50:49,990 --> 00:50:51,530
and not through the band-to-band

1177
00:50:51,530 --> 00:50:53,590
They will only recombine
with band-to-band

1178
00:50:53,590 --> 00:50:55,950
if there is nothing--
no other recombination

1179
00:50:55,950 --> 00:50:59,160
path left available to them.

1180
00:50:59,160 --> 00:51:00,660
It's a question of probability.

1181
00:51:00,660 --> 00:51:03,940
So if this is really,
really short--

1182
00:51:03,940 --> 00:51:05,440
non-radiative
recombination lifetime

1183
00:51:05,440 --> 00:51:07,440
is really, really short
because a defect density

1184
00:51:07,440 --> 00:51:09,200
is very, very large.

1185
00:51:09,200 --> 00:51:12,590
So the lifetime due to
non-radiative recombination

1186
00:51:12,590 --> 00:51:15,470
mechanisms is very short,
then that's tau effective

1187
00:51:15,470 --> 00:51:18,260
is also going to be very short.

1188
00:51:18,260 --> 00:51:23,260
And very few carriers are
going to recombine radiatively.

1189
00:51:23,260 --> 00:51:26,640
Conversely, if we have a very
clean material-- virtually

1190
00:51:26,640 --> 00:51:29,620
defect-free-- this
non-radiative lifetime

1191
00:51:29,620 --> 00:51:31,552
is then going to
be out the roof.

1192
00:51:31,552 --> 00:51:33,260
And now we're going
to be limited instead

1193
00:51:33,260 --> 00:51:35,280
by the radiative lifetime.

1194
00:51:35,280 --> 00:51:37,800
And so tau effective will be
more similar to radiative.

1195
00:51:37,800 --> 00:51:39,550
Now an interesting
thing happens when

1196
00:51:39,550 --> 00:51:41,280
you have a radiative
recombination event.

1197
00:51:41,280 --> 00:51:43,681
As the name would
suggest, it emits light.

1198
00:51:43,681 --> 00:51:46,180
And that light can be detected
by a CCD camera on your body.

1199
00:51:46,180 --> 00:51:54,110
So what you have-- there you go.

1200
00:51:54,110 --> 00:51:56,450
What you have right
here is a solar cell

1201
00:51:56,450 --> 00:52:00,090
device that is being
biased so current

1202
00:52:00,090 --> 00:52:02,300
is flowing into the device.

1203
00:52:02,300 --> 00:52:06,930
And now the map
what you're seeing

1204
00:52:06,930 --> 00:52:09,457
is indicative of radiative
recombination, why?

1205
00:52:09,457 --> 00:52:11,165
Because nearby the
cell, it's essentially

1206
00:52:11,165 --> 00:52:12,930
a very similar mechanism
or similar set up

1207
00:52:12,930 --> 00:52:14,990
to what we saw
right here where we

1208
00:52:14,990 --> 00:52:19,030
had bias voltage across
the device and imaging

1209
00:52:19,030 --> 00:52:21,890
not with an infrared camera,
but now with a camera matched

1210
00:52:21,890 --> 00:52:23,542
to the band gap energy.

1211
00:52:23,542 --> 00:52:25,500
This could be an indium
gallium arsenide camera

1212
00:52:25,500 --> 00:52:29,224
or say silicon camera, depending
on the precise band gap.

1213
00:52:29,224 --> 00:52:31,140
And you're imaging the
radiative recombination

1214
00:52:31,140 --> 00:52:35,090
as a function of xy
position across your wafer.

1215
00:52:35,090 --> 00:52:36,840
So these darker regions
here on the bottom

1216
00:52:36,840 --> 00:52:40,180
and on the right-- sorry
left-- the other right.

1217
00:52:40,180 --> 00:52:43,390
They indicate regions of lower
minority carrier diffusion

1218
00:52:43,390 --> 00:52:43,950
length.

1219
00:52:43,950 --> 00:52:44,450
Why?

1220
00:52:44,450 --> 00:52:46,820
Because there's less
radiative recombination

1221
00:52:46,820 --> 00:52:48,880
because this term
in those regions

1222
00:52:48,880 --> 00:52:52,180
is very, very small so that
there's not much radiative

1223
00:52:52,180 --> 00:52:53,380
recombination going on.

1224
00:52:53,380 --> 00:52:55,910
Whereas these other regions
here that appear brighter

1225
00:52:55,910 --> 00:53:00,469
have a lot more radiative
recombination going on.

1226
00:53:00,469 --> 00:53:01,510
Any questions about that?

1227
00:53:01,510 --> 00:53:03,001
Yeah?

1228
00:53:03,001 --> 00:53:05,983
AUDIENCE: What fraction of
these radiative recombinations

1229
00:53:05,983 --> 00:53:07,971
are reabsorbed by the material?

1230
00:53:07,971 --> 00:53:10,704
Is it possible to figure
that out to a high degree

1231
00:53:10,704 --> 00:53:14,840
so you can get a quantitative
sense of how much is radiating?

1232
00:53:14,840 --> 00:53:15,620
PROFESSOR: Sure.

1233
00:53:15,620 --> 00:53:18,240
The question pertains
to photon recycling

1234
00:53:18,240 --> 00:53:22,220
and how many photons that
are radiatively emitted

1235
00:53:22,220 --> 00:53:25,010
get reabsorbed by the material.

1236
00:53:25,010 --> 00:53:27,340
That would depend on
the optical path length

1237
00:53:27,340 --> 00:53:29,180
inside of the material
of carriers being

1238
00:53:29,180 --> 00:53:31,280
generated in the material.

1239
00:53:31,280 --> 00:53:33,950
And you assume that
carrier generation

1240
00:53:33,950 --> 00:53:36,180
is fairly isotropic in
nature-- that there is not

1241
00:53:36,180 --> 00:53:39,700
any preferred direction unless
you have a dyed molecule that's

1242
00:53:39,700 --> 00:53:41,380
designed in a certain
way to emit light

1243
00:53:41,380 --> 00:53:43,174
in a certain orientation.

1244
00:53:43,174 --> 00:53:44,590
So you can assume
that the carrier

1245
00:53:44,590 --> 00:53:46,632
mission in most materials
as isotropic in nature.

1246
00:53:46,632 --> 00:53:48,131
And then it's just
a question of ray

1247
00:53:48,131 --> 00:53:50,610
tracing to figure out what
the optical path length is.

1248
00:53:50,610 --> 00:53:52,780
Compare that to the optical
absorption coefficient

1249
00:53:52,780 --> 00:53:55,219
in the material, and you get
your answer-- what fraction

1250
00:53:55,219 --> 00:53:56,260
of light gets reabsorbed.

1251
00:54:01,200 --> 00:54:05,030
All right, I can sense where
we're reaching threshold here

1252
00:54:05,030 --> 00:54:07,090
in terms of new
information gathered.

1253
00:54:07,090 --> 00:54:11,696
I'm just going to briefly go
over the Suns-Voc at the very

1254
00:54:11,696 --> 00:54:13,740
end .

1255
00:54:13,740 --> 00:54:17,470
And the Suns-Voc
technique is useful

1256
00:54:17,470 --> 00:54:22,090
as we saw during class because
as this paper right here by Ron

1257
00:54:22,090 --> 00:54:24,910
Sinton and Andres
Cuevas-- this one--

1258
00:54:24,910 --> 00:54:28,010
I know you definitely
have this one.

1259
00:54:28,010 --> 00:54:30,820
This paper essentially
describes the functioning

1260
00:54:30,820 --> 00:54:31,890
of the technique.

1261
00:54:31,890 --> 00:54:36,512
On page number two on the upper,
left-hand side, so figure 2,

1262
00:54:36,512 --> 00:54:37,970
we essentially have
the figure that

1263
00:54:37,970 --> 00:54:40,740
was determined on the screen
during the measurements.

1264
00:54:40,740 --> 00:54:42,700
So for those of you close
enough to the monitor

1265
00:54:42,700 --> 00:54:44,574
to actually watch the
measurement being taken

1266
00:54:44,574 --> 00:54:48,300
in the Suns-Voc, you
saw the light intensity

1267
00:54:48,300 --> 00:54:51,150
decaying exponentially.

1268
00:54:51,150 --> 00:54:53,640
So the light intensity
is a flash bulb, poof,

1269
00:54:53,640 --> 00:54:55,740
and decayed exponentially.

1270
00:54:55,740 --> 00:54:58,510
The open circuit voltage
measured by the device--

1271
00:54:58,510 --> 00:55:00,970
so you had this solar cell
device sitting on the platen

1272
00:55:00,970 --> 00:55:03,610
with a little probe
coming in and a 50 megaohm

1273
00:55:03,610 --> 00:55:06,930
resistor connected in series--
very little current flowing

1274
00:55:06,930 --> 00:55:09,009
through that external circuit.

1275
00:55:09,009 --> 00:55:10,300
But it's measuring the voltage.

1276
00:55:10,300 --> 00:55:12,216
So it's essentially
maintaining the solar cell

1277
00:55:12,216 --> 00:55:14,140
in open circuit
voltage conditions

1278
00:55:14,140 --> 00:55:16,599
while the flash is decaying.

1279
00:55:16,599 --> 00:55:18,390
And it's measuring the
open circuit voltage

1280
00:55:18,390 --> 00:55:20,820
across the device
at each illumination

1281
00:55:20,820 --> 00:55:23,300
intensity as the pulse decays.

1282
00:55:23,300 --> 00:55:25,360
So you can see the
time scale here

1283
00:55:25,360 --> 00:55:27,490
is on the order of a few
10s of milliseconds--

1284
00:55:27,490 --> 00:55:29,050
so a very fast measurement.

1285
00:55:29,050 --> 00:55:31,990
The electronics response
time has to be very fast.

1286
00:55:31,990 --> 00:55:35,680
It tells you what the RC time
constant of the circuit is.

1287
00:55:35,680 --> 00:55:37,930
And you can notice that the
voltage-- the open circuit

1288
00:55:37,930 --> 00:55:42,720
voltage decays linearly when the
illumination intensity decays

1289
00:55:42,720 --> 00:55:43,429
exponentially.

1290
00:55:43,429 --> 00:55:43,970
Why is that ?

1291
00:55:47,148 --> 00:55:48,510
AUDIENCE: Natural log.

1292
00:55:48,510 --> 00:55:49,926
PROFESSOR: This
is a natural log--

1293
00:55:49,926 --> 00:55:52,095
so what Joe showed to you
yesterday or on Thursday

1294
00:55:52,095 --> 00:55:56,280
in class is how the voltage
output of a solar cell

1295
00:55:56,280 --> 00:55:58,800
varies by illumination
intensity.

1296
00:55:58,800 --> 00:56:00,840
So there's a logarithmic
dependence there.

1297
00:56:00,840 --> 00:56:04,160
So that's why you take the log
of the exponentially decaying

1298
00:56:04,160 --> 00:56:07,010
light intensity and you obtain
a linear relation that's

1299
00:56:07,010 --> 00:56:11,660
the open circuit voltage
decline as a function of time.

1300
00:56:11,660 --> 00:56:15,970
So we can derive an
IV characteristic

1301
00:56:15,970 --> 00:56:21,340
by translating that open circuit
voltage into an implied bias

1302
00:56:21,340 --> 00:56:22,680
voltage.

1303
00:56:22,680 --> 00:56:26,410
And knowing the short circuit
current density up front

1304
00:56:26,410 --> 00:56:30,470
essentially by measuring it
by measuring the solar cell

1305
00:56:30,470 --> 00:56:33,744
device in a solar simulator,
we can obtain the short circuit

1306
00:56:33,744 --> 00:56:34,410
current density.

1307
00:56:34,410 --> 00:56:38,320
We pin both curves up here
at the measured short circuit

1308
00:56:38,320 --> 00:56:39,780
current density.

1309
00:56:39,780 --> 00:56:43,530
And then we plot both
curves out-- the IV curve

1310
00:56:43,530 --> 00:56:45,110
and the Suns-Voc curve.

1311
00:56:45,110 --> 00:56:48,030
Now the Suns-Voc curve
is similar to the IV

1312
00:56:48,030 --> 00:56:49,590
curve with only one difference.

1313
00:56:49,590 --> 00:56:51,569
There's no current or
very, very little current

1314
00:56:51,569 --> 00:56:53,360
flowing through the
external circuit, which

1315
00:56:53,360 --> 00:56:55,890
means there isn't any
real power flowing

1316
00:56:55,890 --> 00:56:59,410
through the external circuit,
which means if we go back

1317
00:56:59,410 --> 00:57:03,180
to our expression right here--
if we have no current flowing

1318
00:57:03,180 --> 00:57:07,120
through, that J
is equal to zero.

1319
00:57:07,120 --> 00:57:10,304
And now that R sub
S term disappears.

1320
00:57:10,304 --> 00:57:11,720
So we've come up
with a clever way

1321
00:57:11,720 --> 00:57:15,509
to drop all of our R sub
S out of that expression.

1322
00:57:15,509 --> 00:57:17,300
And the only thing that
matters essentially

1323
00:57:17,300 --> 00:57:20,030
is to some degree shunt.

1324
00:57:20,030 --> 00:57:24,590
But this is essentially
the highest IV curve

1325
00:57:24,590 --> 00:57:27,530
that one could possibly obtain
in the absence of series

1326
00:57:27,530 --> 00:57:28,470
resistance.

1327
00:57:28,470 --> 00:57:30,300
And that's what
you see right here.

1328
00:57:30,300 --> 00:57:33,534
That's this delta in
between your illuminated IV

1329
00:57:33,534 --> 00:57:35,700
curve-- what is measured
using the solar simulator--

1330
00:57:35,700 --> 00:57:37,760
and your Suns-Voc
curve, which is

1331
00:57:37,760 --> 00:57:42,610
measured in the absence
of series resistance.

1332
00:57:42,610 --> 00:57:45,010
Now this works well
with most materials.

1333
00:57:45,010 --> 00:57:48,430
If you have a very high
capacitance in your device,

1334
00:57:48,430 --> 00:57:53,450
you'll want to look very
carefully at this decay--

1335
00:57:53,450 --> 00:57:55,220
at the light intensity decay.

1336
00:57:55,220 --> 00:57:57,410
If your capacitance and
your device is very large,

1337
00:57:57,410 --> 00:58:00,444
it might actually flatten
out the voltage response.

1338
00:58:00,444 --> 00:58:02,610
So you want to make sure
that the time-- the RC time

1339
00:58:02,610 --> 00:58:07,490
constant of your device is
matched to the flash bulb decay

1340
00:58:07,490 --> 00:58:08,012
time.

1341
00:58:08,012 --> 00:58:09,720
That's just one word
of warning for those

1342
00:58:09,720 --> 00:58:12,425
working on high voltage
or organic materials.

1343
00:58:15,180 --> 00:58:18,930
Any questions on the
Sun-Voc measurement?

1344
00:58:18,930 --> 00:58:22,820
This gives you what's
called an implied IV curve.

1345
00:58:22,820 --> 00:58:28,930
And that is-- we could say in
the best case scenario what

1346
00:58:28,930 --> 00:58:31,360
you would get from
the solar cell

1347
00:58:31,360 --> 00:58:34,688
in the absence of any
series resistance losses.

1348
00:58:34,688 --> 00:58:37,118
AUDIENCE: The last
question-- Voc--

1349
00:58:37,118 --> 00:58:41,990
that's not a solar
simulator in any way is it?

1350
00:58:41,990 --> 00:58:43,122
PROFESSOR: No.

1351
00:58:43,122 --> 00:58:44,570
AUDIENCE: Is that important?

1352
00:58:44,570 --> 00:58:46,180
PROFESSOR: Yeah,
it is important.

1353
00:58:46,180 --> 00:58:49,260
It would be
different-- in essence,

1354
00:58:49,260 --> 00:58:53,680
it would be a different
current response

1355
00:58:53,680 --> 00:58:58,590
if you didn't have the
same exact solar spectrum.

1356
00:58:58,590 --> 00:59:00,570
Now, that's a very
important point.

1357
00:59:00,570 --> 00:59:03,590
If you're venturing outside
of, say, silicon-based devices,

1358
00:59:03,590 --> 00:59:05,965
and you're more sensitive
to the infrared or the UV,

1359
00:59:05,965 --> 00:59:07,840
you really want to make
sure that you measure

1360
00:59:07,840 --> 00:59:12,530
the intensity as a function of
wavelength output of that bulb.

1361
00:59:12,530 --> 00:59:16,160
Otherwise, you might be higher
or lower during the Suns-Voc

1362
00:59:16,160 --> 00:59:17,470
measurement.

1363
00:59:17,470 --> 00:59:20,590
For a silicon-based
device, the majority

1364
00:59:20,590 --> 00:59:23,320
falls within the
specified regime.

1365
00:59:23,320 --> 00:59:27,125
Down here-- actually right
here on this diagram--

1366
00:59:27,125 --> 00:59:28,780
that a little spot
right there-- that

1367
00:59:28,780 --> 00:59:31,200
is a small
calibration solar cell

1368
00:59:31,200 --> 00:59:33,720
of known short-circuit current
and open circuit voltage.

1369
00:59:33,720 --> 00:59:36,355
That's used to calibrate
for other silicon devices

1370
00:59:36,355 --> 00:59:39,510
if you're moving away from
silicon, that becomes an issue.

1371
00:59:45,420 --> 00:59:47,800
Correctable-- you can make
sure the spectral response--

1372
00:59:47,800 --> 00:59:51,760
or the spectral radiance
of that light source

1373
00:59:51,760 --> 00:59:53,590
and compare it to
the solar simulator

1374
00:59:53,590 --> 00:59:56,460
and do the normalization
accordingly.

1375
00:59:56,460 --> 00:59:57,780
Here's what I recommend.

1376
00:59:57,780 --> 00:59:59,680
With the last 15
minutes of class,

1377
00:59:59,680 --> 01:00:01,320
I wanted to catch
each of the teams

1378
01:00:01,320 --> 01:00:02,990
and talk about your
class projects just

1379
01:00:02,990 --> 01:00:05,550
to make sure that things
were rolling along.

1380
01:00:05,550 --> 01:00:07,990
I have feelers out to
some of the mentors who

1381
01:00:07,990 --> 01:00:09,140
haven't responded yet.

1382
01:00:09,140 --> 01:00:11,590
And so I just wanted to
touch base with each of you

1383
01:00:11,590 --> 01:00:14,320
for about three minutes just
make sure everything is rolling

1384
01:00:14,320 --> 01:00:15,870
along.