1
00:00:00,070 --> 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 --> 00:00:13,330
To make a donation or
view additional materials

6
00:00:13,330 --> 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
00:00:25,080 --> 00:00:25,746
PROFESSOR: Good.

9
00:00:25,746 --> 00:00:27,579
Well, why don't we go
ahead and get started.

10
00:00:27,579 --> 00:00:29,790
We're going to be discussing
photovoltaic efficiency,

11
00:00:29,790 --> 00:00:31,150
measurement, and
theoretical limits.

12
00:00:31,150 --> 00:00:32,900
And there will be
plenty of natural breaks

13
00:00:32,900 --> 00:00:34,560
over the course of
today's presentation

14
00:00:34,560 --> 00:00:36,580
for us to have our debate.

15
00:00:36,580 --> 00:00:39,410
This is a fun lecture
because we start out

16
00:00:39,410 --> 00:00:42,710
by talking about how to measure
solar cell device efficiency.

17
00:00:42,710 --> 00:00:46,520
Later, we will discuss
the theoretical efficiency

18
00:00:46,520 --> 00:00:49,160
limits of solar cells.

19
00:00:49,160 --> 00:00:52,850
Why do we focus an entire
lecture on efficiency?

20
00:00:52,850 --> 00:00:54,500
Well first, as we
discussed previously,

21
00:00:54,500 --> 00:00:57,890
efficiency is a very strong
determining factor for cost.

22
00:00:57,890 --> 00:01:01,310
The rationale, again, is that
if you have low efficiency,

23
00:01:01,310 --> 00:01:03,690
you're going to need
more commodity materials

24
00:01:03,690 --> 00:01:05,650
to make a given watt peak.

25
00:01:05,650 --> 00:01:10,550
That means you'll need a
larger area of solar module

26
00:01:10,550 --> 00:01:12,070
to make a certain
amount of power,

27
00:01:12,070 --> 00:01:14,111
which means you'll need
more glass, encapsulance,

28
00:01:14,111 --> 00:01:14,790
and so forth.

29
00:01:14,790 --> 00:01:16,790
So efficiency is a
strong lever determining

30
00:01:16,790 --> 00:01:19,470
cost of all downstream
components except

31
00:01:19,470 --> 00:01:22,810
for the area independent
factors like the inverter.

32
00:01:22,810 --> 00:01:25,396
Secondly, efficiency is
tricky to measure accurately.

33
00:01:25,396 --> 00:01:28,020
That is why there are only a few
laboratories around the world,

34
00:01:28,020 --> 00:01:31,570
a handful, that are certified
to measure solar cell

35
00:01:31,570 --> 00:01:33,200
efficiencies.

36
00:01:33,200 --> 00:01:35,130
These are the
efficiencies that could

37
00:01:35,130 --> 00:01:38,794
be reported in, say, the
efficiency compendiums,

38
00:01:38,794 --> 00:01:40,710
an example of which
you've just picked up here

39
00:01:40,710 --> 00:01:42,850
as one of the readings.

40
00:01:42,850 --> 00:01:46,960
The reality is that we can
measure efficiency or get

41
00:01:46,960 --> 00:01:49,960
a pretty close value for
an efficiency of a device

42
00:01:49,960 --> 00:01:51,540
within our own laboratories.

43
00:01:51,540 --> 00:01:54,590
But there are a number
of possible errors

44
00:01:54,590 --> 00:01:57,560
that can creep up and nip us in
the heel if we're not careful.

45
00:01:57,560 --> 00:02:00,960
And that's why we spend
some time in today's lecture

46
00:02:00,960 --> 00:02:02,946
discussing those
potential pitfalls.

47
00:02:02,946 --> 00:02:04,570
And thirdly, there
are new technologies

48
00:02:04,570 --> 00:02:06,750
that are being
promised right and left

49
00:02:06,750 --> 00:02:08,889
to overcome some of
the fundamental limits

50
00:02:08,889 --> 00:02:12,510
of traditional solar cell
devices, like this one right

51
00:02:12,510 --> 00:02:13,820
here.

52
00:02:13,820 --> 00:02:16,290
And we have to understand
what those limits are so

53
00:02:16,290 --> 00:02:20,020
that we can design better
ways to overcome them.

54
00:02:20,020 --> 00:02:21,770
So learning objectives.

55
00:02:21,770 --> 00:02:22,950
Bit of a small font here.

56
00:02:22,950 --> 00:02:25,190
But the idea is--
our very first point

57
00:02:25,190 --> 00:02:28,370
is to identify the sources of
record solar cell efficiencies

58
00:02:28,370 --> 00:02:30,860
to understand where one
goes to look them up.

59
00:02:30,860 --> 00:02:33,760
How do you find the record
efficiency of say, a silicon

60
00:02:33,760 --> 00:02:35,820
device or a [INAUDIBLE] device?

61
00:02:35,820 --> 00:02:37,570
Eventually, we'll
talk about measurement

62
00:02:37,570 --> 00:02:39,360
of solar cell efficiencies.

63
00:02:39,360 --> 00:02:43,320
And finally, the
theoretical or fundamental

64
00:02:43,320 --> 00:02:45,790
limits to solar
cell efficiencies.

65
00:02:45,790 --> 00:02:48,380
So the key concepts for--
learning objective number one,

66
00:02:48,380 --> 00:02:51,770
to identify a source
for record efficiencies.

67
00:02:51,770 --> 00:02:55,550
My go to place is a
Progress in Photovoltaics,

68
00:02:55,550 --> 00:02:57,680
it's a journal and
in the PB field.

69
00:02:57,680 --> 00:03:02,320
And every six
months, PiP comes out

70
00:03:02,320 --> 00:03:05,000
with solar cell
efficiency tables

71
00:03:05,000 --> 00:03:08,260
led by one of their editors,
Martin Green, professor

72
00:03:08,260 --> 00:03:10,500
at University of New
South Wales in Australia.

73
00:03:10,500 --> 00:03:13,880
The latest addition that I
could find was version 38.

74
00:03:13,880 --> 00:03:17,180
I believe this was from
July or June this summer.

75
00:03:17,180 --> 00:03:19,990
And every six months or so, they
come out with a new version.

76
00:03:19,990 --> 00:03:22,160
And what you'll find
inside of that paper--

77
00:03:22,160 --> 00:03:25,560
this is one of the four
handouts that you have today.

78
00:03:25,560 --> 00:03:28,260
One of the three
articles that you have.

79
00:03:28,260 --> 00:03:32,140
What you'll find inside of a
typical Martin Green solar cell

80
00:03:32,140 --> 00:03:35,700
efficiency table is a
listing of-- in table one,

81
00:03:35,700 --> 00:03:37,140
a listing of individual cells.

82
00:03:37,140 --> 00:03:38,800
In table two of modules.

83
00:03:38,800 --> 00:03:42,660
And within the cells,
any new record efficiency

84
00:03:42,660 --> 00:03:45,720
will be shown in bold.

85
00:03:45,720 --> 00:03:49,130
Over the six months preceding
the release of the latest

86
00:03:49,130 --> 00:03:52,910
version, there were indeed
four record efficiencies

87
00:03:52,910 --> 00:03:54,720
that had been made.

88
00:03:54,720 --> 00:03:58,390
And that's pretty
impressive advance.

89
00:03:58,390 --> 00:04:01,420
Note one thing which
we'll come back to later.

90
00:04:01,420 --> 00:04:03,170
Note the plus minus
appearing here

91
00:04:03,170 --> 00:04:06,800
after the efficiency number.

92
00:04:06,800 --> 00:04:09,140
Is anybody surprised
at that number?

93
00:04:09,140 --> 00:04:12,634
Ashley, did you expect
it to be that big?

94
00:04:12,634 --> 00:04:14,032
AUDIENCE: No, not that big.

95
00:04:14,032 --> 00:04:14,720
PROFESSOR: Not that big, right?

96
00:04:14,720 --> 00:04:15,345
AUDIENCE: Yeah.

97
00:04:15,345 --> 00:04:18,970
But say the record efficiency
of crystal silicon device, 25%

98
00:04:18,970 --> 00:04:21,360
plus or minus 0.5.

99
00:04:21,360 --> 00:04:23,540
Pretty large delta.

100
00:04:23,540 --> 00:04:24,930
We'll explain some
of the reasons

101
00:04:24,930 --> 00:04:27,770
why that error bar is so large.

102
00:04:27,770 --> 00:04:29,280
Another thing to keep in mind.

103
00:04:29,280 --> 00:04:33,495
Look at gallium arsenide
at 28.1%, just achieved

104
00:04:33,495 --> 00:04:37,320
by Alta Devices in March.

105
00:04:37,320 --> 00:04:39,940
Keep that number,
28.1 plus or minus

106
00:04:39,940 --> 00:04:42,860
20% in your mind, at
least the first one, 28.1

107
00:04:42,860 --> 00:04:46,900
We'll march on to the module
efficiency tables right here.

108
00:04:46,900 --> 00:04:50,580
So now the gallium
arsenide module efficiency

109
00:04:50,580 --> 00:05:00,100
record for the efficiency tables
right here is 21%, or 21.1%.

110
00:05:00,100 --> 00:05:04,360
Crystalline silicon has dropped
from 25 to 23 and so forth.

111
00:05:04,360 --> 00:05:08,120
And this is fairly typical
that record module efficiencies

112
00:05:08,120 --> 00:05:11,380
are in the order of 2% to 7%
lower than record efficiency

113
00:05:11,380 --> 00:05:12,760
cells.

114
00:05:12,760 --> 00:05:14,510
Can anybody guess
why that might be?

115
00:05:17,336 --> 00:05:19,974
AUDIENCE: [INAUDIBLE].

116
00:05:19,974 --> 00:05:20,640
PROFESSOR: Yeah.

117
00:05:20,640 --> 00:05:22,750
You have a mix of
different performers.

118
00:05:22,750 --> 00:05:25,120
And when you connect them
in series and in parallel,

119
00:05:25,120 --> 00:05:29,330
you're going to be limited by
the lowest voltage or current,

120
00:05:29,330 --> 00:05:31,870
respectively.

121
00:05:31,870 --> 00:05:32,370
Yes.

122
00:05:32,370 --> 00:05:32,890
OK.

123
00:05:32,890 --> 00:05:34,920
So certainly there
are mismatches

124
00:05:34,920 --> 00:05:37,810
between the individual
devices inside of a module.

125
00:05:37,810 --> 00:05:39,390
That's where the
majority of that

126
00:05:39,390 --> 00:05:42,200
comes from for, say,
crystalline silicon

127
00:05:42,200 --> 00:05:45,020
or discrete monolithic wafers.

128
00:05:45,020 --> 00:05:48,430
But how about for some of
these thin film devices?

129
00:05:48,430 --> 00:05:52,910
They are deposited using these
large chemical vapor deposition

130
00:05:52,910 --> 00:05:55,800
reactors, for example,
or PVD reactors.

131
00:05:55,800 --> 00:05:58,910
And you deposit a uniform
thin film over a large area,

132
00:05:58,910 --> 00:06:01,680
use lasers to cut little
trenches in the films

133
00:06:01,680 --> 00:06:04,550
and discretize the
devices that way.

134
00:06:04,550 --> 00:06:06,880
So how come there
are differences

135
00:06:06,880 --> 00:06:10,000
between record cell and
record module for thin film?

136
00:06:14,989 --> 00:06:15,780
PROFESSOR: Exactly.

137
00:06:15,780 --> 00:06:19,320
So if you have inhomogeneities
in thickness or in composition,

138
00:06:19,320 --> 00:06:21,600
or even in surface
quality from region

139
00:06:21,600 --> 00:06:24,400
to region in that
large area, you're

140
00:06:24,400 --> 00:06:27,360
likely to reduce
your performance.

141
00:06:27,360 --> 00:06:29,945
One analogy, since
I know many of you

142
00:06:29,945 --> 00:06:32,290
are mechanical engineers,
one analogy to this

143
00:06:32,290 --> 00:06:35,570
is when you're doing tensile
tests with ceramics or brittle

144
00:06:35,570 --> 00:06:40,900
specimens at room temperature,
and you pull on your specimen,

145
00:06:40,900 --> 00:06:44,830
and you obtain a certain
fracture stress, you then take

146
00:06:44,830 --> 00:06:47,080
that smaller piece, pull again.

147
00:06:47,080 --> 00:06:48,970
Now the fracture
stress is higher.

148
00:06:48,970 --> 00:06:51,320
Pull on that smaller
piece that broke off.

149
00:06:51,320 --> 00:06:53,860
Yet again, another fracture
stress is even higher.

150
00:06:53,860 --> 00:06:56,360
In other words, in
that large specimen,

151
00:06:56,360 --> 00:06:59,720
there was one point
that was extremely weak,

152
00:06:59,720 --> 00:07:01,940
another point that
was sort of weak,

153
00:07:01,940 --> 00:07:04,070
and another point
that was mildly weak.

154
00:07:04,070 --> 00:07:06,800
And as you increase the
size of your specimen,

155
00:07:06,800 --> 00:07:10,440
the likelihood of having one of
these failure points increases.

156
00:07:10,440 --> 00:07:14,110
That's an analogy, let's say,
to a large area module, as well,

157
00:07:14,110 --> 00:07:16,800
if we could have pinholes
or other manufacturing

158
00:07:16,800 --> 00:07:18,440
defects inside of
a large area module

159
00:07:18,440 --> 00:07:20,190
that could reduce the
performance locally,

160
00:07:20,190 --> 00:07:22,140
and everything is
interconnected,

161
00:07:22,140 --> 00:07:25,430
it tends to drop the
performance overall.

162
00:07:25,430 --> 00:07:28,840
So we have the record
laboratory efficiencies.

163
00:07:28,840 --> 00:07:32,660
This is another
reference source.

164
00:07:32,660 --> 00:07:36,350
I am not personally aware
where Larry Kazmerski publishes

165
00:07:36,350 --> 00:07:37,420
this on a regular basis.

166
00:07:37,420 --> 00:07:39,670
I know he maintains this table.

167
00:07:39,670 --> 00:07:41,760
And if you email
him very nicely,

168
00:07:41,760 --> 00:07:44,710
he'll email you back with
the most updated version.

169
00:07:44,710 --> 00:07:47,580
But I'm not aware of
any publication outlet

170
00:07:47,580 --> 00:07:50,270
where this is
regularly appearing.

171
00:07:50,270 --> 00:07:52,770
But nevertheless, it captures
the record efficiencies

172
00:07:52,770 --> 00:07:55,160
versus time the same
way that you, too,

173
00:07:55,160 --> 00:07:58,310
could do if you went to
the Martin Green records.

174
00:07:58,310 --> 00:08:01,870
And I believe we're at,
was it, number 39 now?

175
00:08:01,870 --> 00:08:02,370
38.

176
00:08:02,370 --> 00:08:03,260
We're at Version 38.

177
00:08:03,260 --> 00:08:06,610
And if you went back in time
to all of the different tables

178
00:08:06,610 --> 00:08:08,270
and tabulated the
results versus time,

179
00:08:08,270 --> 00:08:10,478
you would get a plot that
looked very similar to this

180
00:08:10,478 --> 00:08:13,400
for each technology.

181
00:08:13,400 --> 00:08:16,290
So next up we're going
to identify the sources

182
00:08:16,290 --> 00:08:18,240
of standard solar spectrum.

183
00:08:18,240 --> 00:08:20,220
So we know the
record efficiencies.

184
00:08:20,220 --> 00:08:21,920
We're taking their
word for it right now

185
00:08:21,920 --> 00:08:23,370
that they did everything right.

186
00:08:23,370 --> 00:08:25,730
Now we're going to learn
what that everything is

187
00:08:25,730 --> 00:08:27,860
and how to do it
right, or at least some

188
00:08:27,860 --> 00:08:30,976
of the pieces of doing it right.

189
00:08:30,976 --> 00:08:33,059
We're going to identify,
first of all, the sources

190
00:08:33,059 --> 00:08:34,809
of standard solar spectrum.

191
00:08:34,809 --> 00:08:36,740
So in one scenario
you could say, well,

192
00:08:36,740 --> 00:08:39,020
let me just go outside and
measure the solar spectrum.

193
00:08:39,020 --> 00:08:41,412
If you go outside today,
it's kind of cloudy.

194
00:08:41,412 --> 00:08:43,620
The spectrum would look very
different than tomorrow.

195
00:08:43,620 --> 00:08:44,994
And so it would
be very difficult

196
00:08:44,994 --> 00:08:47,184
for you to compare the
performance of your device

197
00:08:47,184 --> 00:08:48,850
against somebody who
might be in Germany

198
00:08:48,850 --> 00:08:50,349
or somebody who
might be in Finland,

199
00:08:50,349 --> 00:08:51,970
or somebody who
might be in Brazil.

200
00:08:51,970 --> 00:08:54,940
So that's why we come up
with the standard reference

201
00:08:54,940 --> 00:08:56,060
solar spectrum.

202
00:08:56,060 --> 00:08:57,910
And it's shown right here.

203
00:08:57,910 --> 00:09:01,560
The ASTM standard
reference spectrum

204
00:09:01,560 --> 00:09:06,250
is shown on the nice NREL
website for air mass zero.

205
00:09:06,250 --> 00:09:10,660
That's in the outer reaches of
our earth's atmosphere and air

206
00:09:10,660 --> 00:09:13,640
mass 1.5, which is assumed
to be a standard in, say,

207
00:09:13,640 --> 00:09:16,160
temperate climates.

208
00:09:16,160 --> 00:09:20,170
So there have been numerous
revisions to the standard.

209
00:09:20,170 --> 00:09:22,330
You might want to use
the latest one just

210
00:09:22,330 --> 00:09:26,940
to make sure that everything
is up to proper spec.

211
00:09:26,940 --> 00:09:32,680
The latest standard
here, essentially they,

212
00:09:32,680 --> 00:09:35,490
in very pedantic
detail, walk through all

213
00:09:35,490 --> 00:09:39,230
of the possible scattering
mechanisms in the atmosphere.

214
00:09:39,230 --> 00:09:44,620
And this is their justification
for the specific solar spectrum

215
00:09:44,620 --> 00:09:45,760
that they're measuring.

216
00:09:45,760 --> 00:09:52,230
And so you have a nice
explanation in much detail.

217
00:09:52,230 --> 00:09:55,230
Let me just show you once
again the standard spectrum.

218
00:09:55,230 --> 00:09:56,230
You've seen this before.

219
00:09:56,230 --> 00:09:57,938
You've worked with it
in homework number,

220
00:09:57,938 --> 00:09:59,460
I believe it was one or two.

221
00:09:59,460 --> 00:10:04,880
And we see AM0, AM1.5
direct, and AM1.5 global,

222
00:10:04,880 --> 00:10:08,680
AM1.5 global being the capture
of sunlight from a full

223
00:10:08,680 --> 00:10:11,540
hemisphere, direct being
looking very closely,

224
00:10:11,540 --> 00:10:14,480
in a very small solid angle,
directly at the sun on a sunny

225
00:10:14,480 --> 00:10:15,290
day.

226
00:10:15,290 --> 00:10:17,720
Given these
atmospheric conditions

227
00:10:17,720 --> 00:10:21,260
with a small amount of
atmospheric scattering,

228
00:10:21,260 --> 00:10:25,792
obtaining an integrated power
density, watts per square meter

229
00:10:25,792 --> 00:10:28,000
if you integrate over all
wavelengths, it's about 90%

230
00:10:28,000 --> 00:10:32,280
of the global full
hemisphere measurement.

231
00:10:32,280 --> 00:10:34,900
Any questions so far about this?

232
00:10:34,900 --> 00:10:35,400
OK.

233
00:10:35,400 --> 00:10:38,070
So that's pretty straightforward
and the measurement devices

234
00:10:38,070 --> 00:10:42,250
for the global and
direct measurement today.

235
00:10:42,250 --> 00:10:47,070
Obviously we didn't have
those fancy contraptions 60,

236
00:10:47,070 --> 00:10:49,680
70 years ago in quite the same
way, with the same materials

237
00:10:49,680 --> 00:10:52,250
and the same design and
the same quality of glass.

238
00:10:52,250 --> 00:10:55,360
But we had other
measurement devices.

239
00:10:55,360 --> 00:10:59,080
And you'll see in a few slides
how the solar constant has been

240
00:10:59,080 --> 00:11:00,970
varying as a function of time.

241
00:11:00,970 --> 00:11:03,280
Granted, the solar
output also varies,

242
00:11:03,280 --> 00:11:07,331
but our ability to measure the
solar output varies as well.

243
00:11:07,331 --> 00:11:09,830
So if you look at the evolution
of the solar constant versus

244
00:11:09,830 --> 00:11:12,170
time, this is the
number of watts

245
00:11:12,170 --> 00:11:15,760
per meter squared in the outer
reaches of our atmosphere,

246
00:11:15,760 --> 00:11:17,070
so AM0.

247
00:11:17,070 --> 00:11:21,500
And you can see that
our values have changed.

248
00:11:21,500 --> 00:11:24,200
So if you were to
ask, is this because

249
00:11:24,200 --> 00:11:26,230
of the change of our
measurement capabilities

250
00:11:26,230 --> 00:11:29,550
or because of the change of the
solar output, the people who

251
00:11:29,550 --> 00:11:33,480
study the sun will tell you that
the variation in solar output

252
00:11:33,480 --> 00:11:36,480
is expected to be rather small.

253
00:11:36,480 --> 00:11:39,250
So the likely origin of
this large fluctuation,

254
00:11:39,250 --> 00:11:42,990
pre-1960, is most likely
due to our ability

255
00:11:42,990 --> 00:11:46,380
to measure the
spectrum accurately.

256
00:11:46,380 --> 00:11:50,070
So next we'll describe how to
simulate the solar spectrum

257
00:11:50,070 --> 00:11:51,750
in the laboratory,
and we'll describe

258
00:11:51,750 --> 00:11:53,300
how a solar simulator works.

259
00:11:53,300 --> 00:11:54,520
So great.

260
00:11:54,520 --> 00:11:57,680
We have an idealized
solar spectrum right here.

261
00:11:57,680 --> 00:12:00,350
How do we recreate this in
a laboratory environment?

262
00:12:00,350 --> 00:12:04,160
How do we obtain a
light source that

263
00:12:04,160 --> 00:12:06,690
follows this profile exactly?

264
00:12:06,690 --> 00:12:11,190
Well, barring the ability to
recreate a small fusion source

265
00:12:11,190 --> 00:12:13,190
in the laboratory
and the ability

266
00:12:13,190 --> 00:12:17,400
to introduce exactly
the right Fraunhofer

267
00:12:17,400 --> 00:12:20,020
lines into our
spectrum, we are not

268
00:12:20,020 --> 00:12:22,030
going to be able to
reproduce that exactly.

269
00:12:22,030 --> 00:12:25,400
But we have several techniques
that come fairly close.

270
00:12:25,400 --> 00:12:27,480
So the solar simulator.

271
00:12:27,480 --> 00:12:33,810
This is a schematic coming
off of the Newport website.

272
00:12:33,810 --> 00:12:35,820
Many of the solar simulators
in the laboratories

273
00:12:35,820 --> 00:12:38,180
here at MIT, as you'll
see as you walk around,

274
00:12:38,180 --> 00:12:41,810
be Newport Oriole
or related brands,

275
00:12:41,810 --> 00:12:44,190
we have the light
source back here.

276
00:12:44,190 --> 00:12:46,080
Note the type of
the lamp right here.

277
00:12:46,080 --> 00:12:48,480
It's a xenon arc lamp.

278
00:12:48,480 --> 00:12:54,680
Mirrors, essentially
a series of optics

279
00:12:54,680 --> 00:12:58,160
to create the right
form factor of the light

280
00:12:58,160 --> 00:13:03,340
and ultimately work our way
toward a planar incoming beam.

281
00:13:03,340 --> 00:13:06,860
And the what's called spectral
correction filter, which

282
00:13:06,860 --> 00:13:08,970
is right in the middle
of the optics train.

283
00:13:08,970 --> 00:13:10,780
And that's
essentially to correct

284
00:13:10,780 --> 00:13:14,260
for variances between the
emission of the xenon arc

285
00:13:14,260 --> 00:13:19,280
clamp and the ASTM solar
spectrum as defined right here.

286
00:13:22,160 --> 00:13:27,100
And we have a shutter, as
well, to block the light.

287
00:13:27,100 --> 00:13:30,580
For example, if your
device or cell or material

288
00:13:30,580 --> 00:13:34,850
is photosensitive, you
might want to not expose it

289
00:13:34,850 --> 00:13:37,050
for long periods of time.

290
00:13:37,050 --> 00:13:40,780
There's also-- when I say
photosensitive I don't mean

291
00:13:40,780 --> 00:13:43,670
that the device will
stop working under light,

292
00:13:43,670 --> 00:13:45,370
but that the
performance will change.

293
00:13:45,370 --> 00:13:47,210
There are materials, for
example, amorphous silicon

294
00:13:47,210 --> 00:13:49,376
that we discussed, where
the performance does change

295
00:13:49,376 --> 00:13:53,210
as a function of illumination,
cumulative illumination

296
00:13:53,210 --> 00:13:54,420
intensity.

297
00:13:54,420 --> 00:13:56,890
So another interesting
thing to note here

298
00:13:56,890 --> 00:13:59,880
is that we have
quasi-planar light coming

299
00:13:59,880 --> 00:14:03,750
in at the end of
this optics train.

300
00:14:03,750 --> 00:14:05,870
But if we're going to be
measuring, for example,

301
00:14:05,870 --> 00:14:08,680
concentrating solar
cell apparatus,

302
00:14:08,680 --> 00:14:12,770
it's much more important to have
a higher degree of planarity

303
00:14:12,770 --> 00:14:16,680
of the incoming light, and then
other optics would be used.

304
00:14:16,680 --> 00:14:18,390
One might envision,
for instance,

305
00:14:18,390 --> 00:14:21,140
increasing the distance
between the light source

306
00:14:21,140 --> 00:14:23,350
and the actual sample,
or the optical path

307
00:14:23,350 --> 00:14:29,320
length between the collimating
mirror and the sample.

308
00:14:29,320 --> 00:14:33,650
So now that we have light
approaching our sample right

309
00:14:33,650 --> 00:14:36,420
here, so light is
incoming on the sample,

310
00:14:36,420 --> 00:14:39,170
we have three things, broadly,
that we have to worry about.

311
00:14:39,170 --> 00:14:40,590
We have to worry
about uniformity

312
00:14:40,590 --> 00:14:43,870
of the light, the
uniformity from small region

313
00:14:43,870 --> 00:14:45,902
to region of the
illuminated area.

314
00:14:45,902 --> 00:14:47,610
We need to worry about
spectral fidelity.

315
00:14:47,610 --> 00:14:52,760
That means, how closely does
the spectrum of our lamp

316
00:14:52,760 --> 00:14:55,930
match the ASTM standard
solar spectrum?

317
00:14:55,930 --> 00:14:57,820
And thirdly, temporal stability.

318
00:14:57,820 --> 00:15:01,120
That means if I turn on
the lamp this morning

319
00:15:01,120 --> 00:15:03,250
and want to take in a
measurement immediately,

320
00:15:03,250 --> 00:15:04,080
is it stable yet?

321
00:15:04,080 --> 00:15:04,580
OK.

322
00:15:04,580 --> 00:15:07,750
Let me wait a half hour for
things to-- for example,

323
00:15:07,750 --> 00:15:10,580
the thermal loads inside of
the system to reach equilibrium

324
00:15:10,580 --> 00:15:12,120
with environment.

325
00:15:12,120 --> 00:15:13,540
Now I'm going to measure it.

326
00:15:13,540 --> 00:15:15,330
What if I come back
in an hour and a half?

327
00:15:15,330 --> 00:15:16,770
Will I still get
the same result?

328
00:15:16,770 --> 00:15:19,080
What if I come back
tomorrow or next month?

329
00:15:19,080 --> 00:15:21,600
Temporal stability is
another major concern

330
00:15:21,600 --> 00:15:24,090
for solar simulators.

331
00:15:24,090 --> 00:15:26,300
On the right-hand side
is just an example

332
00:15:26,300 --> 00:15:29,360
of the radiance versus
wavelength of a given light

333
00:15:29,360 --> 00:15:32,200
source versus time.

334
00:15:32,200 --> 00:15:37,080
And you can see there's a new
lamp after a certain working

335
00:15:37,080 --> 00:15:41,230
period of 1,200 hours.

336
00:15:41,230 --> 00:15:45,090
So we have non-ideal
matches, several examples

337
00:15:45,090 --> 00:15:48,880
of light sources that
don't quite get it right

338
00:15:48,880 --> 00:15:53,620
or have several spikes
in the output spectrum.

339
00:15:53,620 --> 00:15:56,900
So in all cases, here
the AM1 direct spectrum

340
00:15:56,900 --> 00:16:00,569
is shown in this dash dot line.

341
00:16:00,569 --> 00:16:02,110
And different light
sources are shown

342
00:16:02,110 --> 00:16:04,220
either solid or dashed lines.

343
00:16:04,220 --> 00:16:05,520
So not great.

344
00:16:05,520 --> 00:16:07,870
And then finally, we
reach our xenon arc lamp

345
00:16:07,870 --> 00:16:09,670
with air mass filters.

346
00:16:09,670 --> 00:16:13,710
And the filters are to
suppress certain peaks

347
00:16:13,710 --> 00:16:16,090
and certain general
portions of the spectrum

348
00:16:16,090 --> 00:16:21,490
so that we have an approximation
of our ASTM standard reference

349
00:16:21,490 --> 00:16:22,390
spectrum.

350
00:16:22,390 --> 00:16:24,700
And you notice that
it's not perfect.

351
00:16:24,700 --> 00:16:29,570
You notice that there are spikes
in the output of the xenon arc

352
00:16:29,570 --> 00:16:33,130
lamp, as you might expect
from the physics involved

353
00:16:33,130 --> 00:16:34,660
in the light source.

354
00:16:34,660 --> 00:16:38,480
This is of some
concern if your device

355
00:16:38,480 --> 00:16:41,670
is particularly sensitive to
a region, a spectral region

356
00:16:41,670 --> 00:16:43,540
where those might be present.

357
00:16:43,540 --> 00:16:47,630
So it's not a bad idea
to take a measurement

358
00:16:47,630 --> 00:16:49,760
of your light source
and actually understand

359
00:16:49,760 --> 00:16:52,810
how it compares, how it
matches up against the ASTM

360
00:16:52,810 --> 00:16:55,370
standard spectrum.

361
00:16:55,370 --> 00:16:58,530
In terms of standards, or
ranking different types

362
00:16:58,530 --> 00:17:03,020
of solar simulators,
there are three classes

363
00:17:03,020 --> 00:17:04,849
according to the IEC standards.

364
00:17:04,849 --> 00:17:07,810
There are, as well,
other common standards.

365
00:17:07,810 --> 00:17:10,079
There's a standard
used in Japan,

366
00:17:10,079 --> 00:17:12,390
and the ASTM standards as well.

367
00:17:12,390 --> 00:17:14,589
But let's focus
on the IEC 904-9.

368
00:17:17,200 --> 00:17:20,540
These are the requirements
for solar simulators measuring

369
00:17:20,540 --> 00:17:23,470
crystalline silicon
single junction devices.

370
00:17:23,470 --> 00:17:27,150
So it's a very
specific standard.

371
00:17:27,150 --> 00:17:28,980
And in a few slides
we'll explain

372
00:17:28,980 --> 00:17:30,480
what the potential
differences are

373
00:17:30,480 --> 00:17:33,500
when you're measuring other
types of solar cell materials.

374
00:17:33,500 --> 00:17:36,040
We have the spectral match,
or spectral fidelity.

375
00:17:36,040 --> 00:17:39,740
We have the non-uniformity
and the temporal instability

376
00:17:39,740 --> 00:17:42,430
in this case, since
we're defining

377
00:17:42,430 --> 00:17:44,470
a relatively small parameter.

378
00:17:44,470 --> 00:17:49,050
Class A solar simulators
have relatively tight specs.

379
00:17:49,050 --> 00:17:51,230
But you'll still notice
here that the non-uniformity

380
00:17:51,230 --> 00:17:54,690
of the irradiance plus or minus
2%, the temporal instability

381
00:17:54,690 --> 00:17:56,240
plus or minus 2%.

382
00:17:56,240 --> 00:17:58,890
This is where you start to
see some of those error bars

383
00:17:58,890 --> 00:18:01,110
on the record efficiency
measurements, right?

384
00:18:01,110 --> 00:18:07,690
So if a laboratory has a very
good handle on its reference

385
00:18:07,690 --> 00:18:11,280
solar simulator, it will be
able to calculate these effects

386
00:18:11,280 --> 00:18:14,299
and estimate what their impact
is on the actual solar cell

387
00:18:14,299 --> 00:18:15,090
efficiency measure.

388
00:18:17,610 --> 00:18:19,610
Note that the
temporal instability

389
00:18:19,610 --> 00:18:21,660
for the Japanese standard
is a little bit more

390
00:18:21,660 --> 00:18:23,630
stringent than the IEC test.

391
00:18:23,630 --> 00:18:27,350
Minor detail, but depending
on who your collaborators are,

392
00:18:27,350 --> 00:18:28,810
where they are in
the world, they

393
00:18:28,810 --> 00:18:31,780
might be using a different
standard than you.

394
00:18:31,780 --> 00:18:32,930
Just keep that in mind.

395
00:18:32,930 --> 00:18:37,149
The solar simulator
downstairs in Building 35,

396
00:18:37,149 --> 00:18:38,690
in the laboratory
that several of you

397
00:18:38,690 --> 00:18:41,189
have seen yesterday when you
did the phosphorus diffusions--

398
00:18:41,189 --> 00:18:44,600
I poked my head in and
saw everybody there.

399
00:18:44,600 --> 00:18:46,290
So about a dozen
of you might have

400
00:18:46,290 --> 00:18:48,373
walked past the solar
simulator in the laboratory.

401
00:18:48,373 --> 00:18:52,980
That's a large area, as
in the illuminated area

402
00:18:52,980 --> 00:18:57,290
at the working plane is around
20 by 20 centimeter squared.

403
00:18:57,290 --> 00:19:01,380
And it is a Class
AAB solar simulator.

404
00:19:01,380 --> 00:19:07,330
So I believe that would be
spectral match, non-uniformity,

405
00:19:07,330 --> 00:19:09,140
and then temporal stability.

406
00:19:09,140 --> 00:19:12,490
So you'll typically see
solar simulators rated

407
00:19:12,490 --> 00:19:18,220
in this way, AAA, Triple A, or
Class B solar simulator, or AAB

408
00:19:18,220 --> 00:19:20,620
and so forth.

409
00:19:20,620 --> 00:19:22,272
So again, note the
significant figures.

410
00:19:22,272 --> 00:19:23,480
That's where that comes from.

411
00:19:23,480 --> 00:19:26,840
Pretty straightforward.

412
00:19:26,840 --> 00:19:29,570
Next we're going to
describe how to accurately

413
00:19:29,570 --> 00:19:31,640
measure and report
cell efficiency

414
00:19:31,640 --> 00:19:34,380
and some common pitfalls to
avoid when actually measuring

415
00:19:34,380 --> 00:19:35,470
the cells.

416
00:19:35,470 --> 00:19:37,190
So this is really
directed toward people

417
00:19:37,190 --> 00:19:40,100
who are doing active research
right now in the field of PV.

418
00:19:40,100 --> 00:19:43,710
For those who aren't, enjoy.

419
00:19:43,710 --> 00:19:46,430
And we'll come back as soon
as this little section is

420
00:19:46,430 --> 00:19:47,490
over to the debate.

421
00:19:47,490 --> 00:19:50,560
And finally some topics
of general interest.

422
00:19:50,560 --> 00:19:55,160
So this is just a
small subset of things

423
00:19:55,160 --> 00:19:58,740
to keep in mind when you're
measuring an actual solar cell.

424
00:19:58,740 --> 00:20:00,770
This is by no means
a comprehensive list.

425
00:20:00,770 --> 00:20:04,040
There are, indeed, people
who spend their entire lives

426
00:20:04,040 --> 00:20:07,940
optimizing and perfecting the
art of measuring solar cells.

427
00:20:07,940 --> 00:20:10,480
So the very first thing
that you might want to do

428
00:20:10,480 --> 00:20:20,730
is have a reference solar cell
encapsulated, and then mailed

429
00:20:20,730 --> 00:20:26,760
to you after a measurement is
performed at a certification

430
00:20:26,760 --> 00:20:27,560
laboratory.

431
00:20:27,560 --> 00:20:30,700
So this is an example of a
very small crystalline silicon

432
00:20:30,700 --> 00:20:35,040
solar cell device inside
of an encapsulated frame.

433
00:20:35,040 --> 00:20:37,970
And the encapsulation
is meant to prevent

434
00:20:37,970 --> 00:20:40,280
any degradation
to the solar cell,

435
00:20:40,280 --> 00:20:43,350
as well as damage that might
incur during accidental use.

436
00:20:43,350 --> 00:20:47,310
And this cell is called
a reference cell.

437
00:20:47,310 --> 00:20:49,350
It was tested at NREL.

438
00:20:49,350 --> 00:20:53,140
The current voltage, the
short circuit current,

439
00:20:53,140 --> 00:20:56,700
open circuit voltage, and
fill factor of that device

440
00:20:56,700 --> 00:21:00,560
is well-known and reported,
and is essentially

441
00:21:00,560 --> 00:21:03,190
sent with that device
back to our laboratory.

442
00:21:03,190 --> 00:21:08,140
Now whenever we want to make
a new measurement on any cell

443
00:21:08,140 --> 00:21:10,100
that we want to measure
inside the laboratory,

444
00:21:10,100 --> 00:21:11,920
we'll measure the
reference cell first just

445
00:21:11,920 --> 00:21:14,760
to make sure that our solar
simulator is well-behaved.

446
00:21:14,760 --> 00:21:16,784
When I say well-behaved,
what could happen?

447
00:21:16,784 --> 00:21:18,700
Well, one of the many
things that could happen

448
00:21:18,700 --> 00:21:22,350
is that the lamp
intensity decreases.

449
00:21:22,350 --> 00:21:24,820
That's probably the most
common thing that can happen.

450
00:21:24,820 --> 00:21:26,390
You'll just have an
overall reduction

451
00:21:26,390 --> 00:21:28,200
in the output of your
lamp, in which case

452
00:21:28,200 --> 00:21:30,880
you'll notice a reduction
in what cell parameter?

453
00:21:30,880 --> 00:21:34,196
Current, voltage,
or fill factor?

454
00:21:34,196 --> 00:21:34,820
Current, right?

455
00:21:34,820 --> 00:21:36,426
So the light
intensity decreases.

456
00:21:36,426 --> 00:21:37,800
You'll have a
reduction primarily

457
00:21:37,800 --> 00:21:41,720
in your current, logarithmic
reduction in voltage.

458
00:21:41,720 --> 00:21:47,190
So you have your standard
calibrated reference cell.

459
00:21:47,190 --> 00:21:49,860
This is definitely something
that each laboratory that

460
00:21:49,860 --> 00:21:52,360
is going to be serious
about measuring efficiency

461
00:21:52,360 --> 00:21:53,340
should have.

462
00:21:53,340 --> 00:21:56,150
This is also an apparatus that
doesn't leave our laboratory

463
00:21:56,150 --> 00:21:58,300
except for class purposes.

464
00:21:58,300 --> 00:22:00,110
It's something we
treat very carefully,

465
00:22:00,110 --> 00:22:02,980
since I think the cost
ran in the few hundreds

466
00:22:02,980 --> 00:22:05,850
or thousands of dollars.

467
00:22:05,850 --> 00:22:08,930
Avoid extraneous--
basically, avoid

468
00:22:08,930 --> 00:22:10,980
light from coming
in from outside

469
00:22:10,980 --> 00:22:12,490
of your solar simulator.

470
00:22:12,490 --> 00:22:14,326
It is often shown, or
the solar simulators

471
00:22:14,326 --> 00:22:16,700
are often shown in this manner
right here, where you have

472
00:22:16,700 --> 00:22:17,540
everything out in the open.

473
00:22:17,540 --> 00:22:19,430
That's to show you
what's going on inside.

474
00:22:19,430 --> 00:22:20,650
When you actually
take the measurement,

475
00:22:20,650 --> 00:22:22,470
typically you have a
small black curtain

476
00:22:22,470 --> 00:22:27,380
that's light tight around your
apparatus, or maybe even a box.

477
00:22:27,380 --> 00:22:30,680
Ensure 25 degrees C
measurement conditions.

478
00:22:30,680 --> 00:22:33,560
Remember that the open
circuit voltage can change,

479
00:22:33,560 --> 00:22:34,660
depending on the band gap.

480
00:22:34,660 --> 00:22:36,701
The larger the band gap,
the smaller this effect.

481
00:22:36,701 --> 00:22:39,260
The smaller the band gap,
the larger this effect.

482
00:22:39,260 --> 00:22:42,740
The VOC can change as
the temperature changes.

483
00:22:42,740 --> 00:22:45,480
And so it's important
to have a good handle

484
00:22:45,480 --> 00:22:46,620
of your temperature.

485
00:22:46,620 --> 00:22:49,220
That means that you might, for
instance, have active heating

486
00:22:49,220 --> 00:22:50,910
and cooling on your chuck.

487
00:22:50,910 --> 00:22:54,600
And you'll also, at the minimum,
be measuring the temperature

488
00:22:54,600 --> 00:22:59,090
on top of the chuck, where the
actual solar cell will sit,

489
00:22:59,090 --> 00:23:02,720
not far away removed through
some layers of insulation

490
00:23:02,720 --> 00:23:04,060
away from the chuck.

491
00:23:04,060 --> 00:23:06,740
Even if your temperature is
not precisely at 25 degrees

492
00:23:06,740 --> 00:23:09,880
C, if you know the temperature
dependence of your solar cell,

493
00:23:09,880 --> 00:23:13,010
you can correct for it or
account for it to pull it back

494
00:23:13,010 --> 00:23:15,700
to 25 degrees C.

495
00:23:15,700 --> 00:23:17,970
Next, choose your
probe locations

496
00:23:17,970 --> 00:23:21,240
judiciously to avoid
series resistance losses.

497
00:23:21,240 --> 00:23:22,890
Let's think that
through for a second.

498
00:23:22,890 --> 00:23:24,455
So if we have our solar
cell device right here

499
00:23:24,455 --> 00:23:26,621
and we're going to be
measuring its cell efficiency,

500
00:23:26,621 --> 00:23:29,630
if I put one little probe
right here in the corner,

501
00:23:29,630 --> 00:23:32,354
then the current that's
being generated over here

502
00:23:32,354 --> 00:23:34,770
has to travel a very large
distance through a lot of metal

503
00:23:34,770 --> 00:23:36,010
to reach that point.

504
00:23:36,010 --> 00:23:39,210
But if I have probes that
essentially will consist

505
00:23:39,210 --> 00:23:41,490
of many different
individual metal points,

506
00:23:41,490 --> 00:23:45,780
and they come down on either
bus bar like shown right here--

507
00:23:45,780 --> 00:23:48,850
where you see the green right
here, and off of the green are

508
00:23:48,850 --> 00:23:51,090
many little probe tips,
individual probe tips that

509
00:23:51,090 --> 00:23:52,548
will make contact
with the bus bar,

510
00:23:52,548 --> 00:23:54,700
so you have essentially
one coming down right here,

511
00:23:54,700 --> 00:23:57,860
another coming down right here--
your series resistance losses

512
00:23:57,860 --> 00:24:00,040
will be much less.

513
00:24:00,040 --> 00:24:02,120
But one thing to keep
in mind is that when

514
00:24:02,120 --> 00:24:03,990
you do have these
probes sticking

515
00:24:03,990 --> 00:24:05,650
on top of your
solar cell device,

516
00:24:05,650 --> 00:24:08,090
they will scatter some
of the light as well.

517
00:24:08,090 --> 00:24:10,700
So there are best practices
in terms of what color

518
00:24:10,700 --> 00:24:16,620
they should be and how tall
they should be, as well.

519
00:24:16,620 --> 00:24:19,010
So choosing the
probe location is

520
00:24:19,010 --> 00:24:21,570
a large step toward
achieving high efficiencies.

521
00:24:24,230 --> 00:24:27,040
I don't under
emphasize that point.

522
00:24:27,040 --> 00:24:28,620
It really does make
a huge difference

523
00:24:28,620 --> 00:24:32,290
where you put your probe
tips, especially for cells

524
00:24:32,290 --> 00:24:34,320
with high series
resistance, which

525
00:24:34,320 --> 00:24:36,030
can be several of the
new materials that

526
00:24:36,030 --> 00:24:37,810
are being developed.

527
00:24:37,810 --> 00:24:40,210
Account for spectral mismatch
between calibration cell

528
00:24:40,210 --> 00:24:41,140
and your cell.

529
00:24:41,140 --> 00:24:42,970
Let me drive that point home.

530
00:24:42,970 --> 00:24:46,720
So I have a crystal and silicon
calibration cell right here.

531
00:24:46,720 --> 00:24:48,330
I know its spectral response.

532
00:24:48,330 --> 00:24:52,190
I know at what wavelengths
it responds the strongest.

533
00:24:52,190 --> 00:24:58,120
And so it will detect any
mismatch between the ASTM

534
00:24:58,120 --> 00:25:01,120
standard and the actual
light source there

535
00:25:01,120 --> 00:25:03,130
where it responds most strongly.

536
00:25:03,130 --> 00:25:05,870
Now let's imagine
that I'm designing

537
00:25:05,870 --> 00:25:09,030
a new organic material
that responds really well.

538
00:25:09,030 --> 00:25:11,390
Pick something, the infrared.

539
00:25:11,390 --> 00:25:14,200
And so now silicon
will stop responding,

540
00:25:14,200 --> 00:25:16,520
let's say, at around
1,100 nanometers.

541
00:25:16,520 --> 00:25:20,810
So anything beyond here,
silicon won't be able to detect.

542
00:25:20,810 --> 00:25:22,660
But let's say my device
is very sensitive

543
00:25:22,660 --> 00:25:24,430
in that spectral region.

544
00:25:24,430 --> 00:25:26,940
My silicon calibration
cell says go ahead,

545
00:25:26,940 --> 00:25:28,830
take your measurement,
everything's fine.

546
00:25:28,830 --> 00:25:30,742
It matches your ASTM standard.

547
00:25:30,742 --> 00:25:33,200
But then when I stick my cell
in there, all of a sudden I'm

548
00:25:33,200 --> 00:25:34,857
getting a super high efficiency.

549
00:25:34,857 --> 00:25:36,940
I just made an organic
device, and my efficiency's

550
00:25:36,940 --> 00:25:41,090
11%, which would
be a world record.

551
00:25:41,090 --> 00:25:41,910
I'm ecstatic.

552
00:25:41,910 --> 00:25:47,080
I'm really happy until I do
the quantum efficiency of both

553
00:25:47,080 --> 00:25:51,240
the standard calibration
cell and my new cell,

554
00:25:51,240 --> 00:25:52,620
and I realize, wait a second.

555
00:25:52,620 --> 00:25:54,350
They're nowhere near
each other in terms

556
00:25:54,350 --> 00:25:57,380
of their responsivity, what
region of the solar spectrum

557
00:25:57,380 --> 00:25:58,930
they can respond well at.

558
00:25:58,930 --> 00:26:01,900
And the reason I'm
bringing that up

559
00:26:01,900 --> 00:26:04,726
is, again, to
emphasize we have, say,

560
00:26:04,726 --> 00:26:06,850
for example, an amorphous
silicon device, a gallium

561
00:26:06,850 --> 00:26:09,930
arsenide device over here,
a SiGs device showing

562
00:26:09,930 --> 00:26:12,080
the different regions
of the solar spectrum,

563
00:26:12,080 --> 00:26:14,880
the solar spectrum shown in this
dark orange in the background

564
00:26:14,880 --> 00:26:18,390
right here, peaking
at around 550.

565
00:26:18,390 --> 00:26:20,080
We can see that the
different materials

566
00:26:20,080 --> 00:26:22,200
are responding to different
regions of the solar spectrum.

567
00:26:22,200 --> 00:26:23,700
And crystalline
silicon would be in

568
00:26:23,700 --> 00:26:26,084
between the amorphous
silicon-- actually,

569
00:26:26,084 --> 00:26:27,500
it would be a
little further over,

570
00:26:27,500 --> 00:26:33,510
be starting up at around between
1,200 and 1,100 nanometers.

571
00:26:33,510 --> 00:26:36,280
So we can see how
different materials

572
00:26:36,280 --> 00:26:38,990
are more sensitive to different
regions of the solar spectrum.

573
00:26:38,990 --> 00:26:40,649
And there's even a
standard test method

574
00:26:40,649 --> 00:26:42,190
for determining the
spectral mismatch

575
00:26:42,190 --> 00:26:44,980
parameter between your device
and the reference cell, an ASTM

576
00:26:44,980 --> 00:26:45,890
standard for it.

577
00:26:45,890 --> 00:26:47,870
And the typical way
to account for this

578
00:26:47,870 --> 00:26:50,190
is to measure the
spectral irradiance

579
00:26:50,190 --> 00:26:52,400
as a function of wavelength
of your light source,

580
00:26:52,400 --> 00:26:55,590
measure the quantum efficiency,
meaning the responsivity versus

581
00:26:55,590 --> 00:26:58,520
wavelength of your device, and
then measure your calibration

582
00:26:58,520 --> 00:26:59,310
cell.

583
00:26:59,310 --> 00:27:01,040
And using that
math, you can really

584
00:27:01,040 --> 00:27:04,630
begin to normalize for
these extraneous effects.

585
00:27:04,630 --> 00:27:05,940
Why do I bring that up?

586
00:27:05,940 --> 00:27:08,280
I bring that up
because there was

587
00:27:08,280 --> 00:27:11,100
an example, several examples,
of folks in the literature--

588
00:27:11,100 --> 00:27:13,200
I've taken off their
names, so protecting

589
00:27:13,200 --> 00:27:15,450
the innocent here-- folks
in the literature who

590
00:27:15,450 --> 00:27:18,880
report very high efficiency
devices, or at the time

591
00:27:18,880 --> 00:27:23,100
was a near-record
efficiency device.

592
00:27:23,100 --> 00:27:26,730
What they showed were the
QE curves of the devices,

593
00:27:26,730 --> 00:27:29,440
and then the IV curves
of their devices.

594
00:27:29,440 --> 00:27:33,340
And you guys, in your homeworks,
calculated the short circuit

595
00:27:33,340 --> 00:27:35,750
current, which is shown here--
essentially the intercept

596
00:27:35,750 --> 00:27:38,170
with the y-axis-- you
calculated the short circuit

597
00:27:38,170 --> 00:27:40,030
current from the QE.

598
00:27:40,030 --> 00:27:42,180
So you know how to
perform that calculation.

599
00:27:42,180 --> 00:27:43,070
Well, guess what?

600
00:27:43,070 --> 00:27:45,450
Folks at NREL also know
how to do that calculation,

601
00:27:45,450 --> 00:27:46,830
and a whole lot more.

602
00:27:46,830 --> 00:27:50,560
So the folks at NREL did that
and said, well, wait a second.

603
00:27:50,560 --> 00:27:52,700
When I do the
integration of your QE,

604
00:27:52,700 --> 00:27:55,340
I'm not getting these short
circuit currents over here.

605
00:27:55,340 --> 00:27:58,990
I suspect what happened was,
in your solar simulator,

606
00:27:58,990 --> 00:28:00,887
you were using a
silicon reference cell.

607
00:28:00,887 --> 00:28:03,220
But your cells were more
sensitive to a different region

608
00:28:03,220 --> 00:28:05,850
of the solar spectrum--
in this particular case

609
00:28:05,850 --> 00:28:09,270
they were more sensitive to
the shorter wavelengths--

610
00:28:09,270 --> 00:28:13,520
and you didn't have a properly
calibrated solar simulator.

611
00:28:13,520 --> 00:28:16,470
So you're over reporting
your current outputs.

612
00:28:16,470 --> 00:28:19,310
And of course that's very
embarrassing for a group.

613
00:28:19,310 --> 00:28:23,010
In this particular case there
were merits on both sides.

614
00:28:23,010 --> 00:28:27,290
There was a rebuttal
to the rebuttal.

615
00:28:27,290 --> 00:28:29,850
So it's not a simple
black and white case

616
00:28:29,850 --> 00:28:34,220
for this particular story,
although the logic does

617
00:28:34,220 --> 00:28:37,340
fall more strongly on one side.

618
00:28:37,340 --> 00:28:41,220
So this is to say, avoid
this sort of controversy.

619
00:28:41,220 --> 00:28:43,510
Perform your
measurements properly,

620
00:28:43,510 --> 00:28:45,410
don't over report
your efficiencies,

621
00:28:45,410 --> 00:28:49,400
and when in doubt, you can
always ship your cells to NREL

622
00:28:49,400 --> 00:28:52,910
or [INAUDIBLE] or another
certified testing center

623
00:28:52,910 --> 00:28:55,890
and get a certified
cell efficiency.

624
00:28:55,890 --> 00:28:59,234
Then you can place your IV curve
inside of your publication,

625
00:28:59,234 --> 00:29:00,650
and in the little
corner over here

626
00:29:00,650 --> 00:29:03,880
it'll have the figure
of NREL or [INAUDIBLE]

627
00:29:03,880 --> 00:29:07,134
and the properly
certified information.

628
00:29:07,134 --> 00:29:08,550
Now that can be
rather complicated

629
00:29:08,550 --> 00:29:10,800
if you're, for example,
growing organic materials that

630
00:29:10,800 --> 00:29:12,310
degrade quickly.

631
00:29:12,310 --> 00:29:14,480
Somehow you have to
transport it over there

632
00:29:14,480 --> 00:29:16,020
without it degrading.

633
00:29:16,020 --> 00:29:20,300
And that's where arranging
in advance the transfer

634
00:29:20,300 --> 00:29:24,320
of the materials, and
perhaps even looking

635
00:29:24,320 --> 00:29:25,790
into what sort of
transfer chamber

636
00:29:25,790 --> 00:29:29,850
you're going to encapsulate your
device in could be of interest.

637
00:29:29,850 --> 00:29:32,420
But at the very least,
please, please, please

638
00:29:32,420 --> 00:29:34,690
remember spectral
response mismatch

639
00:29:34,690 --> 00:29:38,440
when you're doing
efficiency measurements.

640
00:29:38,440 --> 00:29:39,320
Any questions so far?

641
00:29:42,490 --> 00:29:49,406
AUDIENCE: So then does NREL
also make non-silicon reference

642
00:29:49,406 --> 00:29:50,690
cells?

643
00:29:50,690 --> 00:29:53,500
PROFESSOR: So NREL actually
doesn't make these cells.

644
00:29:53,500 --> 00:29:55,220
They're local
companies that live

645
00:29:55,220 --> 00:29:57,770
right around NREL that are
manufacturing these and putting

646
00:29:57,770 --> 00:29:58,930
them together.

647
00:29:58,930 --> 00:30:01,090
NREL will test those cells.

648
00:30:01,090 --> 00:30:03,740
Since it takes a few
months, typically,

649
00:30:03,740 --> 00:30:06,824
to get turnaround on a cell
efficiency measurement--

650
00:30:06,824 --> 00:30:08,990
unless you're fast tracked
in because your device is

651
00:30:08,990 --> 00:30:11,090
degrading or you've
arranged in advance

652
00:30:11,090 --> 00:30:14,200
and kind of put your place
in line-- because it takes

653
00:30:14,200 --> 00:30:16,020
a few months turnaround,
there's actually

654
00:30:16,020 --> 00:30:17,670
a premium on inventory.

655
00:30:17,670 --> 00:30:19,880
And so these companies will
manufacture the devices,

656
00:30:19,880 --> 00:30:21,754
send them in, get them
tested and calibrated,

657
00:30:21,754 --> 00:30:24,646
or get the calibrated
measurements performed at NREL,

658
00:30:24,646 --> 00:30:26,270
and then bring them
back to the company

659
00:30:26,270 --> 00:30:28,645
and put them on the shelf
until you put pick up the phone

660
00:30:28,645 --> 00:30:31,430
and call them and say,
I'd like a certified cell.

661
00:30:31,430 --> 00:30:33,830
You can actually
make your own, too,

662
00:30:33,830 --> 00:30:37,590
if you follow a set
of standard protocols

663
00:30:37,590 --> 00:30:39,430
that you can receive
from the folks at NREL.

664
00:30:39,430 --> 00:30:43,910
I believe Keith Emery might be
a good point contact at first.

665
00:30:43,910 --> 00:30:45,990
There are a very
specific set of protocols

666
00:30:45,990 --> 00:30:47,406
that you should
follow if you want

667
00:30:47,406 --> 00:30:49,050
to make your own in the lab.

668
00:30:49,050 --> 00:30:50,630
In other words,
you have to design

669
00:30:50,630 --> 00:30:53,350
the contacts a certain way,
the encapsule in a certain way,

670
00:30:53,350 --> 00:30:58,000
make sure that the materials
comprising the remainder

671
00:30:58,000 --> 00:31:01,070
are black so that they don't
reflect the light back in,

672
00:31:01,070 --> 00:31:05,140
little details that are
only gathered by experience.

673
00:31:05,140 --> 00:31:06,740
If you follow all
of those parameters,

674
00:31:06,740 --> 00:31:09,444
and the folks over there,
Keith Emery or Paul [INAUDIBLE]

675
00:31:09,444 --> 00:31:11,110
look at it, and they
inspect it and say,

676
00:31:11,110 --> 00:31:14,990
yeah, yeah, looks good,
then you can get that tested

677
00:31:14,990 --> 00:31:17,130
and serve as a
calibration cell as well.

678
00:31:17,130 --> 00:31:19,860
So you can circumvent some
of the cost associated

679
00:31:19,860 --> 00:31:24,340
with buying a certified
cell from a company.

680
00:31:24,340 --> 00:31:26,000
But there is a premium to them.

681
00:31:26,000 --> 00:31:29,610
They don't tend to be
particularly cheap.

682
00:31:32,910 --> 00:31:34,615
So yeah, if you have
another material

683
00:31:34,615 --> 00:31:36,770
and you'd like to
have it calibrated,

684
00:31:36,770 --> 00:31:39,690
you can make your own
calibration reference standard

685
00:31:39,690 --> 00:31:42,430
as long as it doesn't degrade.

686
00:31:42,430 --> 00:31:43,315
Question.

687
00:31:43,315 --> 00:31:45,690
AUDIENCE: Putting
the multiple contact

688
00:31:45,690 --> 00:31:48,540
in yours while you're
measuring the efficiency.

689
00:31:48,540 --> 00:31:50,730
Is that very representative
of the devices that

690
00:31:50,730 --> 00:31:51,855
actually work in the field?

691
00:31:53,560 --> 00:31:56,400
PROFESSOR: So there
are a lot of things--

692
00:31:56,400 --> 00:32:00,740
let me spend a minute waxing
poetic about the discrepancies

693
00:32:00,740 --> 00:32:02,510
between these cell
efficiency measurements

694
00:32:02,510 --> 00:32:05,340
and what actual cells
experience in the field.

695
00:32:05,340 --> 00:32:08,380
We understand the logic behind
cell efficiency measurements.

696
00:32:08,380 --> 00:32:11,070
We understand we have to have
a universal way of comparing

697
00:32:11,070 --> 00:32:15,670
a cell in Japan, in Boston,
in Freiburg, Germany.

698
00:32:15,670 --> 00:32:18,840
So we understand that we
need some standard method

699
00:32:18,840 --> 00:32:20,980
of cell measurement.

700
00:32:20,980 --> 00:32:23,180
This cell measurement
proves very useful when

701
00:32:23,180 --> 00:32:25,340
you measure the output
in terms of peak watts,

702
00:32:25,340 --> 00:32:28,030
because then you can multiply
by the number of peak hours

703
00:32:28,030 --> 00:32:30,910
of sunlight per day and estimate
the energy output as a function

704
00:32:30,910 --> 00:32:32,310
of location on the earth.

705
00:32:32,310 --> 00:32:33,180
So it has its uses.

706
00:32:33,180 --> 00:32:35,900
Now as to its drawbacks.

707
00:32:35,900 --> 00:32:38,520
We're measuring at
25 degrees Celsius.

708
00:32:38,520 --> 00:32:40,810
These cells are typically
operating around 60.

709
00:32:40,810 --> 00:32:44,940
We know that there's a voltage
drop, primarily a voltage drop

710
00:32:44,940 --> 00:32:48,410
with increasing temperature.

711
00:32:48,410 --> 00:32:49,460
That's point one.

712
00:32:49,460 --> 00:32:52,880
Point number two, the
contacting scheme.

713
00:32:52,880 --> 00:32:55,360
Typically on these
devices in a module,

714
00:32:55,360 --> 00:32:58,510
you'll have soldered
contacts in the front,

715
00:32:58,510 --> 00:33:01,480
and so they make contact
with the entire bus bar,

716
00:33:01,480 --> 00:33:04,770
but really, depending on
your soldering machine,

717
00:33:04,770 --> 00:33:06,820
only in a few locations
where there's really

718
00:33:06,820 --> 00:33:10,980
a good electrical contact, and
on the back in three locations.

719
00:33:10,980 --> 00:33:13,970
So yes, there are differences
in how the cell is

720
00:33:13,970 --> 00:33:15,720
contacted in real life.

721
00:33:15,720 --> 00:33:18,335
Could be one of the reasons for
discrepancies between module

722
00:33:18,335 --> 00:33:22,230
and cell efficiency, probably
one of the minor ones compared

723
00:33:22,230 --> 00:33:23,820
to homogeneities.

724
00:33:23,820 --> 00:33:26,530
Why don't we take a
quick little pause

725
00:33:26,530 --> 00:33:30,460
right here before we dive
into describing the efficiency

726
00:33:30,460 --> 00:33:33,210
limitations of a typical
cell and get some cool demos

727
00:33:33,210 --> 00:33:34,430
at the very end.

728
00:33:34,430 --> 00:33:38,100
What we're going to do is
have ourselves a quick debate.

729
00:33:38,100 --> 00:33:41,450
So I'd like to call forward
at the front of the room

730
00:33:41,450 --> 00:33:46,660
the representatives of the
two teams, one of which

731
00:33:46,660 --> 00:33:51,020
is going to debate in favor
of the development of novel

732
00:33:51,020 --> 00:33:53,190
materials for solar cells.

733
00:33:53,190 --> 00:33:55,930
Let me give you an example
of one material that's

734
00:33:55,930 --> 00:33:58,940
attracted quite a bit of
attention, which is pyrite.

735
00:33:58,940 --> 00:34:04,400
I can pass this around as folks
are coming up to the front.

736
00:34:04,400 --> 00:34:09,420
Here's an example of an
iron sulfide-based mineral

737
00:34:09,420 --> 00:34:14,699
which is purported to
have a very high degree

738
00:34:14,699 --> 00:34:17,710
of manufacturability because of
the large resource abundance,

739
00:34:17,710 --> 00:34:20,030
and also the large
refining capacity

740
00:34:20,030 --> 00:34:23,920
for the respective
elemental constituents.

741
00:34:23,920 --> 00:34:26,350
There is a gap between the
performance of current pyrite

742
00:34:26,350 --> 00:34:30,909
cells and their theoretical
record efficiencies, quite

743
00:34:30,909 --> 00:34:33,540
substantial, the
record efficiency being

744
00:34:33,540 --> 00:34:36,524
up closer to 20%, the
actual efficiency being down

745
00:34:36,524 --> 00:34:39,060
at around two.

746
00:34:39,060 --> 00:34:41,909
That is not unlike, for
example, tin sulfide,

747
00:34:41,909 --> 00:34:44,459
or other related materials.

748
00:34:44,459 --> 00:34:46,250
I'd say copper zinc
tin sulfide is probably

749
00:34:46,250 --> 00:34:49,389
the most advanced at around
10%, but still about half

750
00:34:49,389 --> 00:34:50,797
of its theoretical limit.

751
00:34:50,797 --> 00:34:52,880
And so the big question
that we're going to debate

752
00:34:52,880 --> 00:34:57,870
is, does it make sense
to invest a lot of funds

753
00:34:57,870 --> 00:35:00,280
to come up with these
earth abundant alternatives

754
00:35:00,280 --> 00:35:01,910
for our existing
solar cell materials,

755
00:35:01,910 --> 00:35:05,120
our cadtel, our copper
indium gallium selenide,

756
00:35:05,120 --> 00:35:07,350
cognizant that
the supply of some

757
00:35:07,350 --> 00:35:09,940
of these heavier elements,
tellurium and indium,

758
00:35:09,940 --> 00:35:10,940
is limited.

759
00:35:10,940 --> 00:35:13,520
And we may not have enough of
these elements in the earth's

760
00:35:13,520 --> 00:35:16,560
crust to scale up to
the terawatts level.

761
00:35:16,560 --> 00:35:19,050
And so we'll hear
two points of view,

762
00:35:19,050 --> 00:35:23,280
one in favor of development of
new materials, and one against.

763
00:35:23,280 --> 00:35:26,490
And just to situate ourselves
in a position or a location

764
00:35:26,490 --> 00:35:28,820
where these debates
actually do happen,

765
00:35:28,820 --> 00:35:30,600
you can imagine
yourself, for example,

766
00:35:30,600 --> 00:35:33,260
in the Office of Science
and Technology Policy,

767
00:35:33,260 --> 00:35:35,790
which reports directly
to Barack Obama.

768
00:35:35,790 --> 00:35:40,730
These are about 50 PhDs
who are all in an office

769
00:35:40,730 --> 00:35:44,389
under the direction
of a director,

770
00:35:44,389 --> 00:35:46,680
thinking deeply about some
of the scientific challenges

771
00:35:46,680 --> 00:35:50,100
that our nation faces and the
proper scientific response,

772
00:35:50,100 --> 00:35:52,480
coordinating amongst many
agencies, including the DOD,

773
00:35:52,480 --> 00:35:54,360
DOE, NSF, and so forth.

774
00:35:54,360 --> 00:35:56,430
And so you can picture
yourselves in a debate,

775
00:35:56,430 --> 00:35:59,530
in a lively discussion,
all friendly,

776
00:35:59,530 --> 00:36:03,420
but really with the potential
to influence national policy.

777
00:36:03,420 --> 00:36:07,590
Do we direct resource funds
to develop these novel, earth

778
00:36:07,590 --> 00:36:11,270
abundant alternatives that we
might need in 10 years' time?

779
00:36:11,270 --> 00:36:13,692
Or do we focus and allocate
resources elsewhere?

780
00:36:13,692 --> 00:36:15,900
So I'll welcome the two
participants up to the front.

781
00:36:22,020 --> 00:36:24,960
So folks want some
insight into where

782
00:36:24,960 --> 00:36:26,880
the Office of Science
and Technology Policy

783
00:36:26,880 --> 00:36:30,400
actually decided to go.

784
00:36:30,400 --> 00:36:34,410
A gentleman who took a version
of this course in Berkeley

785
00:36:34,410 --> 00:36:40,970
in 2003, I think it was,
his name is Cyrus Wadia.

786
00:36:40,970 --> 00:36:42,900
Graduated with his
PhD from Berkeley

787
00:36:42,900 --> 00:36:44,890
and actually went off to
join Office of Science

788
00:36:44,890 --> 00:36:45,860
and Technology Policy.

789
00:36:45,860 --> 00:36:48,120
He took his class project
for the PB course, which

790
00:36:48,120 --> 00:36:50,770
was analyzing
alternative materials

791
00:36:50,770 --> 00:36:55,200
and published that
paper that you

792
00:36:55,200 --> 00:36:58,570
read by Cyrus Wadia
on resource abundance

793
00:36:58,570 --> 00:37:02,100
that was published in 2009,
for that work earned himself

794
00:37:02,100 --> 00:37:07,220
a TR 35 reward from Tech
Review, and then joined

795
00:37:07,220 --> 00:37:10,600
the Obama administration's
OSTP, and has

796
00:37:10,600 --> 00:37:15,040
been developing the Materials
Genome Project within OSTP.

797
00:37:15,040 --> 00:37:19,750
The first solicitation
for proposals was issued,

798
00:37:19,750 --> 00:37:22,230
I believe, a week
and a half ago.

799
00:37:22,230 --> 00:37:24,630
So that's funneled
through the NSF.

800
00:37:24,630 --> 00:37:26,760
But it's a larger
effort to develop

801
00:37:26,760 --> 00:37:29,920
some of these materials
involving NSFD, DOD, DOE,

802
00:37:29,920 --> 00:37:30,950
and so forth.

803
00:37:30,950 --> 00:37:35,360
So it actually did come
to fruition through OSTP.

804
00:37:37,980 --> 00:37:40,174
Budgeting is always the
big question, though,

805
00:37:40,174 --> 00:37:41,590
because that, of
course, gets done

806
00:37:41,590 --> 00:37:44,180
through the
committees in Congress

807
00:37:44,180 --> 00:37:47,660
and ultimately reconciled
between the House

808
00:37:47,660 --> 00:37:49,750
and the Senate, and has
to make it through OMB

809
00:37:49,750 --> 00:37:52,295
and finally to the
individual directorates.

810
00:37:52,295 --> 00:37:55,170
So that's how things
actually get done.

811
00:37:55,170 --> 00:37:57,490
But it takes the vision
of OSTP, sometimes,

812
00:37:57,490 --> 00:37:59,350
to drive these larger
projects forward.

813
00:37:59,350 --> 00:38:01,757
So we'll wish him the best.

814
00:38:01,757 --> 00:38:03,840
We're going to describe
the efficiency limitations

815
00:38:03,840 --> 00:38:05,990
of a typical solar cell now.

816
00:38:05,990 --> 00:38:10,300
And what I'll do is I'll pass
around some of these books

817
00:38:10,300 --> 00:38:12,790
just to, again,
situate ourselves.

818
00:38:12,790 --> 00:38:14,620
Did I grab Jenny
Nelson's on the way out?

819
00:38:14,620 --> 00:38:15,610
Thought I did.

820
00:38:15,610 --> 00:38:16,600
Hm.

821
00:38:16,600 --> 00:38:17,280
OK.

822
00:38:17,280 --> 00:38:18,020
Well.

823
00:38:18,020 --> 00:38:20,700
Peter Wurfel and--
oh, there it is.

824
00:38:20,700 --> 00:38:22,300
Yes, absolutely.

825
00:38:22,300 --> 00:38:26,190
So there's two books
that are entitled

826
00:38:26,190 --> 00:38:27,320
Physics of Solar Cells.

827
00:38:27,320 --> 00:38:29,820
You can see it's a
very popular topic.

828
00:38:29,820 --> 00:38:32,180
And one is The
Photovoltaic Handbook.

829
00:38:32,180 --> 00:38:35,810
All refer to some aspect
of efficiency limits.

830
00:38:35,810 --> 00:38:38,320
And as you look
through your sheets,

831
00:38:38,320 --> 00:38:40,110
you'll see at the
bottom, typically

832
00:38:40,110 --> 00:38:44,620
the different pages of each
of the books are listed here.

833
00:38:44,620 --> 00:38:47,890
So the very first thing
that we should consider

834
00:38:47,890 --> 00:38:50,180
before we get into
PV technology,

835
00:38:50,180 --> 00:38:54,050
we just have some solar
device looking at the sun.

836
00:38:54,050 --> 00:39:00,650
And the sun is radiating at it
at 6,000 Kelvin, 5,800 Kelvin.

837
00:39:00,650 --> 00:39:03,920
And it, the solar contraption
on the surface the earth,

838
00:39:03,920 --> 00:39:07,790
is radiating back at the sun
with black body radiation

839
00:39:07,790 --> 00:39:09,210
at 300 Kelvin.

840
00:39:09,210 --> 00:39:12,410
So much lower power from
Stefan Boltzmann's Law

841
00:39:12,410 --> 00:39:14,020
we know exactly
the amount of power

842
00:39:14,020 --> 00:39:15,350
being emitted by that device.

843
00:39:15,350 --> 00:39:17,090
So the two are
radiating at each other,

844
00:39:17,090 --> 00:39:19,330
and they're in equilibrium.

845
00:39:19,330 --> 00:39:24,172
And that results
in the very first,

846
00:39:24,172 --> 00:39:25,630
we'll call it
blackbody efficiency,

847
00:39:25,630 --> 00:39:27,780
or maximum solar heat
engine efficiency,

848
00:39:27,780 --> 00:39:30,280
which would be around 86%.

849
00:39:30,280 --> 00:39:33,570
So that's our first
fundamental cutoff.

850
00:39:33,570 --> 00:39:36,680
Now that 86% is averaged
over all wavelengths.

851
00:39:36,680 --> 00:39:41,350
There tends to be a wavelength
dependence to this as well.

852
00:39:41,350 --> 00:39:43,760
So the next step was
to say, OK, well,

853
00:39:43,760 --> 00:39:46,240
we know that we
can't do better than

854
00:39:46,240 --> 00:39:49,470
the theoretical
thermodynamic limit of one

855
00:39:49,470 --> 00:39:52,930
object looking at another.

856
00:39:52,930 --> 00:39:56,230
But what is the actual
limit of a solar cell?

857
00:39:56,230 --> 00:39:59,160
And this was a question
that the first developers

858
00:39:59,160 --> 00:40:00,710
of the solar cell
asked themselves.

859
00:40:00,710 --> 00:40:03,580
So Prince was one of the
three of the team that

860
00:40:03,580 --> 00:40:07,300
in 1954 published on the crystal
silicon solar cell device,

861
00:40:07,300 --> 00:40:09,700
and very quickly followed
it up with another article

862
00:40:09,700 --> 00:40:13,420
here in JP focused on the
theoretical efficiency limit.

863
00:40:13,420 --> 00:40:19,050
And a curve was proposed,
looking much like this,

864
00:40:19,050 --> 00:40:21,760
with two data points for
germanium and silicon,

865
00:40:21,760 --> 00:40:24,880
theoretical limits, that
is, and power density.

866
00:40:24,880 --> 00:40:28,120
So not a conversion efficiency
per se, but a power density.

867
00:40:28,120 --> 00:40:30,850
And of course, assuming a
certain input power density,

868
00:40:30,850 --> 00:40:33,950
one can calculate an
efficiency from there.

869
00:40:33,950 --> 00:40:39,220
And along came--
well, at this time,

870
00:40:39,220 --> 00:40:45,000
H-J Queisser had just moved
over from Germany to Palo Alto,

871
00:40:45,000 --> 00:40:49,030
to The Apricot Barn, and
working with Shockley,

872
00:40:49,030 --> 00:40:52,540
the esteemed
Shockley at the time,

873
00:40:52,540 --> 00:40:55,000
to develop a detailed
balance model

874
00:40:55,000 --> 00:40:57,420
for describing how a
solar cell performs,

875
00:40:57,420 --> 00:41:00,540
or what its ultimate theoretical
efficiency limit could be.

876
00:41:00,540 --> 00:41:03,780
Now what a detailed
balance model does

877
00:41:03,780 --> 00:41:06,080
is just basically
accounting, accounting

878
00:41:06,080 --> 00:41:10,567
for all of the photons coming
in and out of the device.

879
00:41:10,567 --> 00:41:12,400
Through the photons,
the electron hole pairs

880
00:41:12,400 --> 00:41:13,340
are generated.

881
00:41:13,340 --> 00:41:16,050
This is assuming a very
high quality material

882
00:41:16,050 --> 00:41:19,700
that does not have any
form of recombination other

883
00:41:19,700 --> 00:41:21,410
than radiative recombination.

884
00:41:21,410 --> 00:41:24,100
So we talked about the different
methods of limiting lifetime.

885
00:41:24,100 --> 00:41:27,710
Radiative recombination
is one of the methods

886
00:41:27,710 --> 00:41:30,911
of limiting performance
that occurs in high quality

887
00:41:30,911 --> 00:41:31,410
materials.

888
00:41:31,410 --> 00:41:32,620
If you have a poor
quality material,

889
00:41:32,620 --> 00:41:34,286
you'll have non-radiative
recombination,

890
00:41:34,286 --> 00:41:36,110
say, Shockley-Read-Hall
recombination.

891
00:41:36,110 --> 00:41:38,260
But in the detailed
balance model,

892
00:41:38,260 --> 00:41:40,327
only radiative
recombination was assumed.

893
00:41:40,327 --> 00:41:42,410
That way you can count
number of photons coming in

894
00:41:42,410 --> 00:41:45,620
and number of photons
going out of your device.

895
00:41:45,620 --> 00:41:48,930
Furthermore, they assumed
that the mobility of carriers

896
00:41:48,930 --> 00:41:51,844
was infinite inside
of their material.

897
00:41:51,844 --> 00:41:54,010
A few laughs over here
coming from the folks who are

898
00:41:54,010 --> 00:41:55,370
working on organic materials.

899
00:41:55,370 --> 00:41:56,060
But it's true.

900
00:41:56,060 --> 00:41:59,080
They assumed that the
mobility was infinite,

901
00:41:59,080 --> 00:42:01,480
such that the separation
of the quasi Fermi

902
00:42:01,480 --> 00:42:05,000
energies throughout the
entire device was equal.

903
00:42:05,000 --> 00:42:06,412
So if you had a
limited mobility,

904
00:42:06,412 --> 00:42:08,370
if you had a certain
higher density of carriers

905
00:42:08,370 --> 00:42:09,786
in the front than
toward the back,

906
00:42:09,786 --> 00:42:13,249
you'd have a difference
in the chemical potential

907
00:42:13,249 --> 00:42:14,790
through the thickness
of your device.

908
00:42:14,790 --> 00:42:17,110
They assumed infinite mobility.

909
00:42:17,110 --> 00:42:20,780
So it was a very simple,
yet elegantly insightful

910
00:42:20,780 --> 00:42:26,470
model that, upon first
submission, was rejected.

911
00:42:26,470 --> 00:42:29,060
So the first time they submitted
this model for publication

912
00:42:29,060 --> 00:42:31,850
it was rejected outright.

913
00:42:31,850 --> 00:42:36,140
And it took them another several
months of edits, probably

914
00:42:36,140 --> 00:42:38,550
about a year, and
they resubmitted it.

915
00:42:38,550 --> 00:42:40,300
And then it became
a sleeper paper.

916
00:42:40,300 --> 00:42:41,970
It wasn't cited that much.

917
00:42:41,970 --> 00:42:43,846
If you go to the Web of
Science, for example,

918
00:42:43,846 --> 00:42:45,303
and look at this
manuscript, you'll

919
00:42:45,303 --> 00:42:47,440
see that the number of
citations in the early years

920
00:42:47,440 --> 00:42:49,100
was rather limited.

921
00:42:49,100 --> 00:42:51,319
Nowadays there
isn't a talk about

922
00:42:51,319 --> 00:42:53,860
the fundamental efficiency of
a solar cell without mentioning

923
00:42:53,860 --> 00:42:55,960
the Shockley-Queisser
efficiency limit.

924
00:42:55,960 --> 00:42:58,050
That's coming directly
from this paper right here.

925
00:42:58,050 --> 00:43:01,250
And that serves as a
motivational story for you.

926
00:43:01,250 --> 00:43:04,390
If your paper is rejected,
just remember that now you're

927
00:43:04,390 --> 00:43:07,520
along with some several
esteemed individuals who

928
00:43:07,520 --> 00:43:09,800
have set precedence in
the field of photovoltaics

929
00:43:09,800 --> 00:43:11,890
and are Nobel worthy.

930
00:43:11,890 --> 00:43:14,980
So that's my personal opinion.

931
00:43:14,980 --> 00:43:18,500
If you fall into that category,
don't feel discouraged.

932
00:43:18,500 --> 00:43:21,800
Regroup and find a way to
make your paper better.

933
00:43:21,800 --> 00:43:23,280
You have a copy
of this manuscript

934
00:43:23,280 --> 00:43:27,200
right here, by the way, in
your handouts for today.

935
00:43:27,200 --> 00:43:30,840
And interestingly, through
their detailed balance model,

936
00:43:30,840 --> 00:43:37,140
they obtained an ultimate
efficiency versus wavelength,

937
00:43:37,140 --> 00:43:42,860
or energy in this
case, curve that looks

938
00:43:42,860 --> 00:43:45,970
very similar to the curve that
you derived in your earlier

939
00:43:45,970 --> 00:43:49,490
homeworks, where you just
assumed two loss mechanisms.

940
00:43:49,490 --> 00:43:51,910
One was non-absorption
of light and the other

941
00:43:51,910 --> 00:43:54,680
was thermalization of carriers.

942
00:43:54,680 --> 00:43:57,780
So the detailed
balance model, the way

943
00:43:57,780 --> 00:44:00,890
they reached this point,
is different, fundamentally

944
00:44:00,890 --> 00:44:01,390
different.

945
00:44:01,390 --> 00:44:05,300
The physics that was
assumed is basically

946
00:44:05,300 --> 00:44:06,910
listed out right here.

947
00:44:06,910 --> 00:44:10,430
But the end result is fairly
similar to the rough back

948
00:44:10,430 --> 00:44:12,450
of the envelope calculation
that you performed

949
00:44:12,450 --> 00:44:13,950
in an earlier homework.

950
00:44:13,950 --> 00:44:16,280
So again, just to
go up, they assumed

951
00:44:16,280 --> 00:44:18,645
that the photons with energies
greater than the band gap

952
00:44:18,645 --> 00:44:21,490
are absorbed, create
one electron hole pair.

953
00:44:21,490 --> 00:44:22,800
They assume thermalization.

954
00:44:22,800 --> 00:44:24,760
So these were your
two assumptions

955
00:44:24,760 --> 00:44:26,140
that you did in the homework.

956
00:44:26,140 --> 00:44:28,030
Then they do a few
additional things.

957
00:44:28,030 --> 00:44:29,760
They assume that
radiative losses

958
00:44:29,760 --> 00:44:32,390
occur within the material,
that not every carrier is

959
00:44:32,390 --> 00:44:37,140
collected-- that's
important-- so you have

960
00:44:37,140 --> 00:44:40,430
radiative losses in
your device, so that

961
00:44:40,430 --> 00:44:41,860
reduces the performance.

962
00:44:41,860 --> 00:44:44,200
And then they assume
further that there

963
00:44:44,200 --> 00:44:48,800
is a thermodynamic loss in
their device, in other words,

964
00:44:48,800 --> 00:44:51,570
that you don't extract the
full band gap of energy,

965
00:44:51,570 --> 00:44:53,760
but that there's a
thermodynamic loss

966
00:44:53,760 --> 00:44:56,720
as you go from band to band,
which would be the band gap

967
00:44:56,720 --> 00:45:00,140
energy, to the difference or
separation of the quasi Fermi

968
00:45:00,140 --> 00:45:03,280
energies, which is
this delta mu here,

969
00:45:03,280 --> 00:45:05,340
which is the change in
the chemical potential

970
00:45:05,340 --> 00:45:08,620
from the front side to the
back side of the device.

971
00:45:08,620 --> 00:45:12,380
So there are some additional
loss terms that were included.

972
00:45:12,380 --> 00:45:14,760
For a full description of
the detailed balance limit,

973
00:45:14,760 --> 00:45:16,530
and all the math, I
definitely encourage

974
00:45:16,530 --> 00:45:18,430
you go to this website.

975
00:45:18,430 --> 00:45:22,050
As well, Peter Wurfel's
text does a wonderful job

976
00:45:22,050 --> 00:45:23,550
of describing this.

977
00:45:23,550 --> 00:45:25,280
And of course, you
have your paper here.

978
00:45:25,280 --> 00:45:27,430
You can read through the
original paper yourself,

979
00:45:27,430 --> 00:45:29,263
or perhaps suggest it
at an upcoming journal

980
00:45:29,263 --> 00:45:31,570
meeting for your group.

981
00:45:31,570 --> 00:45:35,550
Here is the calculation of the
detailed balance limit for AM0

982
00:45:35,550 --> 00:45:40,030
and AM1.5, essentially as
a function of band gap,

983
00:45:40,030 --> 00:45:42,110
coming from the PVCDROM.

984
00:45:42,110 --> 00:45:44,990
And you can see how silicon
and gallium arsenide are pretty

985
00:45:44,990 --> 00:45:49,340
close to the theoretical
maximum in terms

986
00:45:49,340 --> 00:45:51,820
of the theoretical maximum
efficiency for a single band

987
00:45:51,820 --> 00:45:54,390
gap semiconductor material.

988
00:45:54,390 --> 00:45:57,320
So this curve right here,
again, is just representing one

989
00:45:57,320 --> 00:46:00,820
band gap material, a
single band gap material.

990
00:46:00,820 --> 00:46:03,760
So let's take it from
here and venture forward

991
00:46:03,760 --> 00:46:07,430
into some more realistic
performance reduction effects.

992
00:46:07,430 --> 00:46:09,660
We can, for instance, take
into account recombination

993
00:46:09,660 --> 00:46:13,470
mechanisms that aren't
only-- for example,

994
00:46:13,470 --> 00:46:15,780
that aren't only radiative.

995
00:46:15,780 --> 00:46:17,470
We can take Auger recombination.

996
00:46:17,470 --> 00:46:20,070
We can take Shockley-Read-Hall
recombination into account.

997
00:46:20,070 --> 00:46:25,070
We can also take what's called
photon recycling into account.

998
00:46:25,070 --> 00:46:26,280
So what is photon recycling?

999
00:46:26,280 --> 00:46:28,738
Photon recycling is when you
have a radiative recombination

1000
00:46:28,738 --> 00:46:31,550
event, and that photon gets
trapped within the material.

1001
00:46:31,550 --> 00:46:33,010
It's not allowed to escape.

1002
00:46:33,010 --> 00:46:35,320
But it gets trapped because,
for example, off the top,

1003
00:46:35,320 --> 00:46:37,660
there's an index of
refraction mismatch.

1004
00:46:37,660 --> 00:46:41,840
Or the angle at which it
tries to exit is too oblique,

1005
00:46:41,840 --> 00:46:44,950
and so you have total internal
reflection within your device.

1006
00:46:44,950 --> 00:46:47,970
And you eventually
have reabsorption.

1007
00:46:47,970 --> 00:46:49,510
So that's called
photon recycling

1008
00:46:49,510 --> 00:46:51,610
because you have a radiative
recombination event.

1009
00:46:51,610 --> 00:46:53,736
It emits a photon inside
of your solar cell device,

1010
00:46:53,736 --> 00:46:55,693
and that photon bounces
around a few more times

1011
00:46:55,693 --> 00:46:57,070
until it's
reabsorbed, generating

1012
00:46:57,070 --> 00:46:59,470
another electron hole pair.

1013
00:46:59,470 --> 00:47:04,010
And that is, essentially,
the major boost in some

1014
00:47:04,010 --> 00:47:06,590
of these ultra thin, high
performance solar cell

1015
00:47:06,590 --> 00:47:10,640
materials such as the Alta
Devices record efficiency 28%

1016
00:47:10,640 --> 00:47:12,344
gallium arsenide
cell has the ability

1017
00:47:12,344 --> 00:47:13,510
to do this photon recycling.

1018
00:47:17,380 --> 00:47:21,220
Now how do you modify the
detailed balance limit

1019
00:47:21,220 --> 00:47:23,426
to account for finite mobility?

1020
00:47:23,426 --> 00:47:24,800
I suspect this
corner of the room

1021
00:47:24,800 --> 00:47:26,820
is going to want to hear this.

1022
00:47:26,820 --> 00:47:29,660
There is a beautiful piece
of work done by [INAUDIBLE].

1023
00:47:33,160 --> 00:47:36,300
The PhD thesis is
even more insightful

1024
00:47:36,300 --> 00:47:38,550
than the manuscript in terms
of actually breaking down

1025
00:47:38,550 --> 00:47:39,760
each individual component.

1026
00:47:39,760 --> 00:47:41,880
A great deal of modeling went
into this, including photon

1027
00:47:41,880 --> 00:47:43,090
recycling, including
some things that

1028
00:47:43,090 --> 00:47:44,360
are very difficult to model.

1029
00:47:44,360 --> 00:47:45,990
And the effect of
finite mobility

1030
00:47:45,990 --> 00:47:51,260
was calculated on a modified
detailed balance limit model.

1031
00:47:51,260 --> 00:47:55,990
And you can see
number one, that curve

1032
00:47:55,990 --> 00:47:57,610
that's starting here at the 20.

1033
00:47:57,610 --> 00:48:02,819
That is for a mobility that
is close to the optimal.

1034
00:48:02,819 --> 00:48:04,360
If you drop by an
order of magnitude,

1035
00:48:04,360 --> 00:48:06,110
or two orders of
magnitude, relative

1036
00:48:06,110 --> 00:48:11,550
to the maximum mobility
inside of a material,

1037
00:48:11,550 --> 00:48:14,410
your performance begins
to degrade considerably.

1038
00:48:14,410 --> 00:48:16,680
And now you can
understand why silicon,

1039
00:48:16,680 --> 00:48:19,970
or crystalline
silicon, which has,

1040
00:48:19,970 --> 00:48:22,550
say, hole mobilities
somewhere in the range

1041
00:48:22,550 --> 00:48:28,136
of a few hundreds of centimeters
squared per volt second.

1042
00:48:28,136 --> 00:48:29,510
So keep that number
in your mind,

1043
00:48:29,510 --> 00:48:31,134
on the order of
hundreds of centimeters

1044
00:48:31,134 --> 00:48:32,590
squared per volt hole mobility.

1045
00:48:32,590 --> 00:48:35,160
Now you compare to
amorphous silicon,

1046
00:48:35,160 --> 00:48:37,930
which has something in the
range of 10 to the minus 3

1047
00:48:37,930 --> 00:48:41,070
to 10 to the minus 1
centimeters per volt second.

1048
00:48:41,070 --> 00:48:44,160
And you begin to see why those
materials with poor mobilities

1049
00:48:44,160 --> 00:48:47,110
are really impacted in
terms of their performance.

1050
00:48:47,110 --> 00:48:51,010
There is a method to calculate
the impact of limited mobility

1051
00:48:51,010 --> 00:48:52,280
on device performance.

1052
00:48:52,280 --> 00:48:54,760
And if that intrigues
you, I would definitely

1053
00:48:54,760 --> 00:48:56,970
refer you to this manuscript.

1054
00:48:56,970 --> 00:48:59,110
One very simple way
to think about it

1055
00:48:59,110 --> 00:49:02,390
is, at least for
a band conductor,

1056
00:49:02,390 --> 00:49:04,490
you would have a certain
diffusion length.

1057
00:49:04,490 --> 00:49:07,780
At least a first order
would be the diffusivity

1058
00:49:07,780 --> 00:49:10,610
times the lifetime
and the diffusivity

1059
00:49:10,610 --> 00:49:16,200
would be related to the mobility
inside of your material.

1060
00:49:16,200 --> 00:49:18,884
So your mobility is factoring
into your diffusion length,

1061
00:49:18,884 --> 00:49:20,800
and the diffusion length
is affecting how many

1062
00:49:20,800 --> 00:49:22,860
charge carriers are collected.

1063
00:49:22,860 --> 00:49:26,860
So that's a simple way to
think about the problem.

1064
00:49:26,860 --> 00:49:30,610
A similar related approach that
kind of pulls all this together

1065
00:49:30,610 --> 00:49:34,410
in a graphical form is
at the very last page--

1066
00:49:34,410 --> 00:49:38,050
some of the very last pages
of Peter Wurfel's text.

1067
00:49:38,050 --> 00:49:41,110
And this is going through
the efficiency calculation

1068
00:49:41,110 --> 00:49:43,660
that we just saw, starting
from a certain amount of light

1069
00:49:43,660 --> 00:49:45,110
that's not absorbed.

1070
00:49:45,110 --> 00:49:47,360
The light that is absorbed--
so that's efficiency loss

1071
00:49:47,360 --> 00:49:49,535
mechanism number one--
for a thin device,

1072
00:49:49,535 --> 00:49:51,160
let's say light
trapping isn't perfect.

1073
00:49:51,160 --> 00:49:54,780
We're at around 74%
efficiency of light trapping,

1074
00:49:54,780 --> 00:49:56,180
of absorption of light.

1075
00:49:56,180 --> 00:49:59,200
Let's say now that we have
a thermalization event that

1076
00:49:59,200 --> 00:50:03,970
results in about a 33%
loss, so we're down to 67%

1077
00:50:03,970 --> 00:50:06,160
efficiency for the
thermalization of charge

1078
00:50:06,160 --> 00:50:07,536
carriers, just a step.

1079
00:50:07,536 --> 00:50:08,910
So now our combined
efficiency is

1080
00:50:08,910 --> 00:50:12,280
going to be the multiplicative
product of those two.

1081
00:50:12,280 --> 00:50:16,140
So just the thermalization
of charge carriers

1082
00:50:16,140 --> 00:50:19,410
results in a 33% drop
of an efficiency.

1083
00:50:19,410 --> 00:50:22,244
Now there's a delta between,
say, the band to band

1084
00:50:22,244 --> 00:50:23,910
and the actual chemical
potential inside

1085
00:50:23,910 --> 00:50:26,360
of our material once we consider
the ensemble of carriers, not

1086
00:50:26,360 --> 00:50:28,651
just those free carriers,
but the ensemble of carriers,

1087
00:50:28,651 --> 00:50:31,865
which is going to dictate
the ultimate potential.

1088
00:50:31,865 --> 00:50:34,240
So we have some carriers that
are excited and others that

1089
00:50:34,240 --> 00:50:35,160
aren't.

1090
00:50:35,160 --> 00:50:40,640
These thermodynamic losses
results in another 36% drop.

1091
00:50:40,640 --> 00:50:42,570
And finally, fill
factor losses, which

1092
00:50:42,570 --> 00:50:44,830
represent the solar cell
and practical operation.

1093
00:50:44,830 --> 00:50:48,470
When we have series resistance
and so forth, another 11%.

1094
00:50:48,470 --> 00:50:51,570
So if we multiply these
four numbers together,

1095
00:50:51,570 --> 00:50:54,260
we drop down to 28.

1096
00:50:54,260 --> 00:50:56,610
And the beauty of doing
breaking things out like this--

1097
00:50:56,610 --> 00:50:58,610
and what Peter Wurfel
does very nicely

1098
00:50:58,610 --> 00:51:01,180
is dives into each of
these in great detail

1099
00:51:01,180 --> 00:51:03,904
and explains to you exactly
how those numbers are derived.

1100
00:51:03,904 --> 00:51:05,820
The beauty of doing
something like this is you

1101
00:51:05,820 --> 00:51:08,030
can pick the lowest
number, say, 64 and 67,

1102
00:51:08,030 --> 00:51:10,466
and say, I want
to work on those.

1103
00:51:10,466 --> 00:51:12,590
I want to make my PhD thesis
about those parameters

1104
00:51:12,590 --> 00:51:14,340
because that has
the biggest impact.

1105
00:51:14,340 --> 00:51:19,931
And that's what you can do
with an analysis like this.

1106
00:51:19,931 --> 00:51:21,180
Now you can take this further.

1107
00:51:21,180 --> 00:51:23,340
You can say, well,
these are only four

1108
00:51:23,340 --> 00:51:25,530
of the many parameters
that impact performance.

1109
00:51:25,530 --> 00:51:30,700
I would like to look
at many, many more.

1110
00:51:30,700 --> 00:51:36,200
And that's what one has done
for crystalline silicon, which

1111
00:51:36,200 --> 00:51:40,640
is arguably the most researched
and, from the point of view

1112
00:51:40,640 --> 00:51:44,240
of physical understanding,
advanced solar cell technology.

1113
00:51:44,240 --> 00:51:46,819
We can see a variety of
performance loss mechanisms

1114
00:51:46,819 --> 00:51:49,110
that have been taken into
account inside of the crystal

1115
00:51:49,110 --> 00:51:50,280
silicon devices.

1116
00:51:50,280 --> 00:51:53,950
This was an invited
talk by Dick Swanson,

1117
00:51:53,950 --> 00:51:57,360
the founder of Sun Power,
former professor at Stanford,

1118
00:51:57,360 --> 00:52:01,115
who pulled this together,
a very nice presentation.

1119
00:52:01,115 --> 00:52:03,740
He speaks with authority because
Sun Power produces the highest

1120
00:52:03,740 --> 00:52:07,360
efficiency crystalline silicon
solar cell device commercially.

1121
00:52:07,360 --> 00:52:11,700
And interestingly,
right here, this

1122
00:52:11,700 --> 00:52:16,570
is the thermodynamic limit and
the calculated limit efficiency

1123
00:52:16,570 --> 00:52:17,610
versus time.

1124
00:52:17,610 --> 00:52:18,680
So we start from Prince.

1125
00:52:18,680 --> 00:52:20,300
That was the very first one.

1126
00:52:20,300 --> 00:52:22,940
And we have Shockley-Queisser
up there, and so forth.

1127
00:52:22,940 --> 00:52:25,780
So the theoretical limit
of solar cell performance

1128
00:52:25,780 --> 00:52:28,510
has also changed with time.

1129
00:52:28,510 --> 00:52:30,160
And their calculations vary.

1130
00:52:30,160 --> 00:52:32,410
So it's important to be
able to understand this

1131
00:52:32,410 --> 00:52:34,800
as well as you go into it.

1132
00:52:34,800 --> 00:52:37,800
And here is the actual best
laboratory performance.

1133
00:52:37,800 --> 00:52:42,467
So if we had, for example, a
25% efficient cell over there,

1134
00:52:42,467 --> 00:52:45,050
it would already be higher than
some of the earlier efficiency

1135
00:52:45,050 --> 00:52:47,890
calculations for
crystalline silicon.

1136
00:52:47,890 --> 00:52:51,300
What Dick Swanson thinks is the
practical limit is right there.

1137
00:52:51,300 --> 00:52:53,740
That's the Swanson
prediction, that we

1138
00:52:53,740 --> 00:52:58,820
won't get much above 26%
for crystalline silicon.

1139
00:52:58,820 --> 00:53:01,890
So loss mechanisms
visualized right here,

1140
00:53:01,890 --> 00:53:04,910
and several good readings
on efficiency limits.

1141
00:53:04,910 --> 00:53:07,110
What I'm going to
do is pause here.

1142
00:53:07,110 --> 00:53:09,490
There is still some
material in your text,

1143
00:53:09,490 --> 00:53:11,740
and some really cool
demos that we're

1144
00:53:11,740 --> 00:53:14,560
going to have to wait
until next week to see.