1
00:00:00,060 --> 00:00:01,770
The following
content is provided

2
00:00:01,770 --> 00:00:04,019
under a creative
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 --> 00:00:17,209
from hundreds of MIT courses,
visit MIT OpenCourseWare

7
00:00:17,209 --> 00:00:17,834
at ocw.mit.edu.

8
00:00:26,112 --> 00:00:27,570
PROFESSOR: Welcome
everyone, today.

9
00:00:27,570 --> 00:00:30,000
Today we're going to be talking
about advanced concepts.

10
00:00:30,000 --> 00:00:31,230
These are kind of
like what I would

11
00:00:31,230 --> 00:00:33,021
consider the next
generation of solar cells

12
00:00:33,021 --> 00:00:34,310
if these ideas pan out.

13
00:00:34,310 --> 00:00:36,268
Some of them are very
near and dear to my heart

14
00:00:36,268 --> 00:00:39,400
because it's what my research
is mainly focused on.

15
00:00:39,400 --> 00:00:41,467
Also another quick
realization I had last week.

16
00:00:41,467 --> 00:00:43,550
Probably Tonio mentioned
this to me at some point,

17
00:00:43,550 --> 00:00:45,880
but do you guys know
where 626 comes from,

18
00:00:45,880 --> 00:00:49,280
why the course is called that?

19
00:00:49,280 --> 00:00:58,770
So 6.26 times 10 to the minus
34th is Planck's constant.

20
00:00:58,770 --> 00:01:01,350
So haha, funny.

21
00:01:01,350 --> 00:01:03,670
Thought that was kind of cool.

22
00:01:03,670 --> 00:01:05,120
It's a little joke put in there.

23
00:01:05,120 --> 00:01:11,241
So anyhow without further ado,
one of the cool things-- cells

24
00:01:11,241 --> 00:01:11,740
are done.

25
00:01:11,740 --> 00:01:12,290
Yay.

26
00:01:12,290 --> 00:01:13,180
This is great news.

27
00:01:13,180 --> 00:01:15,880
I was very, very happy.

28
00:01:15,880 --> 00:01:18,872
So it didn't go quite
as well as expected.

29
00:01:18,872 --> 00:01:20,330
I think most of
you are finding out

30
00:01:20,330 --> 00:01:22,580
that these are incredibly
[INAUDIBLE] limited.

31
00:01:22,580 --> 00:01:25,657
I'm going to talk about why
that might be in a second.

32
00:01:25,657 --> 00:01:27,240
Hopefully that will
help you guide you

33
00:01:27,240 --> 00:01:28,656
through your
analysis, which we'll

34
00:01:28,656 --> 00:01:31,170
be doing for the final
part of the quiz 2.3,

35
00:01:31,170 --> 00:01:35,590
which hopefully I'll be able to
post by tomorrow afternoon-ish.

36
00:01:35,590 --> 00:01:38,650
I'm still working
on that right now.

37
00:01:38,650 --> 00:01:41,280
But just to give you guys
kind of the processing

38
00:01:41,280 --> 00:01:43,320
that you didn't get
to see, this is what

39
00:01:43,320 --> 00:01:44,630
the contact firing looks like.

40
00:01:44,630 --> 00:01:47,080
It's a very, very,
very fast process.

41
00:01:47,080 --> 00:01:49,600
It's really remarkable
to me that this

42
00:01:49,600 --> 00:01:51,530
combines so many
photlithographic steps

43
00:01:51,530 --> 00:01:54,220
that if you wanted to
make this in the lab using

44
00:01:54,220 --> 00:01:57,010
photlithography, and
condenses these six or seven

45
00:01:57,010 --> 00:01:59,190
steps into one process.

46
00:01:59,190 --> 00:02:00,860
So it's pretty remarkable.

47
00:02:00,860 --> 00:02:03,630
So DuPont, the people who
made the pace-- so we used

48
00:02:03,630 --> 00:02:07,630
PV16A if you guys are curious
for our front contacts

49
00:02:07,630 --> 00:02:08,961
for silver paste.

50
00:02:08,961 --> 00:02:10,919
And we actually get--
the really important part

51
00:02:10,919 --> 00:02:13,089
is this peaking
temperature, and you

52
00:02:13,089 --> 00:02:15,380
can see we actually got pretty
close to the ramp times.

53
00:02:15,380 --> 00:02:18,680
Where it really starts
to deviate is here,

54
00:02:18,680 --> 00:02:22,690
and we hold it around 400,
a little bit longer, which

55
00:02:22,690 --> 00:02:25,540
is several reasons for that, and
I'll go into that in a second.

56
00:02:25,540 --> 00:02:27,990
But what's important
is that all these are

57
00:02:27,990 --> 00:02:29,650
fired in an oxygen atmosphere.

58
00:02:29,650 --> 00:02:31,420
So normally in
these belt furnaces,

59
00:02:31,420 --> 00:02:33,920
and this is again what's used
in industry-- this is actually

60
00:02:33,920 --> 00:02:36,720
the same exact tool
that I use at Harvard--

61
00:02:36,720 --> 00:02:39,390
it actually looks like this with
the giant CRT monitor on top.

62
00:02:39,390 --> 00:02:41,270
The only way to get the data
off is with a three and a half

63
00:02:41,270 --> 00:02:42,076
floppy drive.

64
00:02:42,076 --> 00:02:43,450
I can't tell you
how difficult it

65
00:02:43,450 --> 00:02:45,930
was to find a working three
and a half floppy drive.

66
00:02:45,930 --> 00:02:47,380
Most of them are
demagnetized at this point.

67
00:02:47,380 --> 00:02:48,330
I had to buy new ones.

68
00:02:48,330 --> 00:02:49,420
They're still being
made, by the way.

69
00:02:49,420 --> 00:02:50,753
If you want them, go to Staples.

70
00:02:53,180 --> 00:02:56,300
So anyhow, it's done in
an oxygen atmosphere,

71
00:02:56,300 --> 00:03:00,030
so they generally force airflow
into these giant belt furnaces,

72
00:03:00,030 --> 00:03:01,230
and these are really long.

73
00:03:01,230 --> 00:03:03,380
Like that's probably
about a foot

74
00:03:03,380 --> 00:03:05,690
wide, maybe 18 inches wide.

75
00:03:05,690 --> 00:03:10,220
So this goes on for tens
of feet, or several meters,

76
00:03:10,220 --> 00:03:12,250
depending on where you're from.

77
00:03:12,250 --> 00:03:14,510
And it's done an oxygen
atmosphere, an air atmosphere,

78
00:03:14,510 --> 00:03:17,445
because it needs to burn
off a lot of the binders

79
00:03:17,445 --> 00:03:19,570
and some of the organics
that might still be there.

80
00:03:19,570 --> 00:03:21,944
And so those organics burn
off, and what you're left with

81
00:03:21,944 --> 00:03:26,410
are the little metal particles,
the frit, and some metal oxide

82
00:03:26,410 --> 00:03:26,910
glass.

83
00:03:26,910 --> 00:03:29,430
It's usually lead oxide,
although DuPont will never

84
00:03:29,430 --> 00:03:33,040
tell you, but from
papers I've read

85
00:03:33,040 --> 00:03:36,690
this is generally what's used.

86
00:03:36,690 --> 00:03:39,930
So when that happens-- this
is that burn-off period.

87
00:03:39,930 --> 00:03:42,414
When it spikes, the
frit will actually

88
00:03:42,414 --> 00:03:44,080
burn through your
silicon nitride layer.

89
00:03:44,080 --> 00:03:45,954
You guys didn't have a
silicon nitride layer,

90
00:03:45,954 --> 00:03:48,262
but if it was there, it
would help eat through it.

91
00:03:48,262 --> 00:03:50,220
And so that way, you've
removed the dielectrics

92
00:03:50,220 --> 00:03:53,460
so you can actually make metal
contact with your silicon

93
00:03:53,460 --> 00:03:55,880
surface underneath.

94
00:03:55,880 --> 00:04:00,400
That firing will-- during
that peak also simultaneously,

95
00:04:00,400 --> 00:04:02,850
these middle particles
will melt and form

96
00:04:02,850 --> 00:04:05,710
this triple eutectic point
with the silicon underneath it.

97
00:04:05,710 --> 00:04:08,790
And so the triple comes from
the silver, the silicon,

98
00:04:08,790 --> 00:04:09,342
and the lead.

99
00:04:09,342 --> 00:04:11,440
The lead actually
dissociates from the glass.

100
00:04:11,440 --> 00:04:13,810
This lowers the melting
temperature of that mixture,

101
00:04:13,810 --> 00:04:16,130
and you can make good ohmic
contact with the surface.

102
00:04:16,130 --> 00:04:18,720
So it's actually a pretty
remarkable, incredible process.

103
00:04:18,720 --> 00:04:20,380
It's still almost
kind of magic to me

104
00:04:20,380 --> 00:04:22,560
on how the whole thing
works, and a lot of it's

105
00:04:22,560 --> 00:04:24,972
just kind of guess and check
and still very proprietary.

106
00:04:24,972 --> 00:04:26,930
So the science is still
a little lacking there.

107
00:04:26,930 --> 00:04:30,430
You don't find a lot of
good articles explaining

108
00:04:30,430 --> 00:04:31,180
the science of it.

109
00:04:31,180 --> 00:04:32,846
I have a few if you
guys are interested.

110
00:04:32,846 --> 00:04:36,381
NREL's put out a lot of really
interesting stuff, and some

111
00:04:36,381 --> 00:04:37,880
of the ways they
actually figure out

112
00:04:37,880 --> 00:04:40,540
the actual profile underneath
the contacts versus you

113
00:04:40,540 --> 00:04:43,690
measure the temperature on
the side of the contacts,

114
00:04:43,690 --> 00:04:47,840
so actually touching the silicon
or the silicon nitride surface,

115
00:04:47,840 --> 00:04:51,150
is you actually measure hydrogen
diffusion underneath it.

116
00:04:51,150 --> 00:04:53,530
And so they'll actually
etch off the contacts

117
00:04:53,530 --> 00:04:57,110
and using secondary ion
mass spectroscopy measure

118
00:04:57,110 --> 00:04:59,531
the hydrogen concentration
off the contact

119
00:04:59,531 --> 00:05:01,530
and underneath the contact,
and that'll actually

120
00:05:01,530 --> 00:05:04,280
tell you what temperature it
saw based on that diffusion

121
00:05:04,280 --> 00:05:05,100
profile.

122
00:05:05,100 --> 00:05:07,230
So there is some good
science going on,

123
00:05:07,230 --> 00:05:11,080
but there's not a lot
of great papers on it.

124
00:05:11,297 --> 00:05:12,880
So anyhow, that just
kind of gives you

125
00:05:12,880 --> 00:05:14,505
what I was working
on over the weekend,

126
00:05:14,505 --> 00:05:16,100
trying to get these cells fired.

127
00:05:16,100 --> 00:05:18,940
So I aggregated a
lot of the results

128
00:05:18,940 --> 00:05:21,464
and tried to make sense
of what was going on.

129
00:05:21,464 --> 00:05:23,880
I think I showed some of you
different firing temperatures

130
00:05:23,880 --> 00:05:25,050
and the fill factors
that we got and there

131
00:05:25,050 --> 00:05:28,150
was kind of just noisy data, so
I couldn't get any real trends.

132
00:05:28,150 --> 00:05:31,219
The best trend I could
find was that if you

133
00:05:31,219 --> 00:05:33,260
take the median of all
the short-circuit currents

134
00:05:33,260 --> 00:05:34,430
of the cells and
you look at the ones

135
00:05:34,430 --> 00:05:36,100
with four millimeter and
two millimeter spacing,

136
00:05:36,100 --> 00:05:38,220
the four millimeter spacing
has a slightly higher

137
00:05:38,220 --> 00:05:39,994
short-circuit
current, presumably

138
00:05:39,994 --> 00:05:40,910
due to shading losses.

139
00:05:40,910 --> 00:05:42,950
If you look at the relative
areas that each of these

140
00:05:42,950 --> 00:05:44,658
cover up, you would
expect a 4% increase.

141
00:05:44,658 --> 00:05:48,440
This is about 3%, so
I'll take that as fact.

142
00:05:48,440 --> 00:05:50,717
It's certainly within
the noise and the error

143
00:05:50,717 --> 00:05:52,300
of the number of
samples that we have,

144
00:05:52,300 --> 00:05:55,730
but I thought it was
interesting nonetheless.

145
00:05:55,730 --> 00:05:58,090
Additionally, I plotted
the series resistance

146
00:05:58,090 --> 00:06:01,110
that I calculated from
your dark IV curves

147
00:06:01,110 --> 00:06:02,854
and plotted your maximum power.

148
00:06:02,854 --> 00:06:04,770
I also removed all the
cells that were broken.

149
00:06:04,770 --> 00:06:06,437
I'm sure if I actually
normalized them--

150
00:06:06,437 --> 00:06:08,228
some of the cells that
were broken actually

151
00:06:08,228 --> 00:06:10,650
performed a lot better,
and I'll get into maybe why

152
00:06:10,650 --> 00:06:13,492
that was the case in a second.

153
00:06:13,492 --> 00:06:14,950
But really the
take-home message is

154
00:06:14,950 --> 00:06:16,825
that our really best
performing cells are all

155
00:06:16,825 --> 00:06:18,300
clumped over here.

156
00:06:18,300 --> 00:06:22,290
And so you can see that our best
cells have a very low series

157
00:06:22,290 --> 00:06:22,915
resistance.

158
00:06:22,915 --> 00:06:24,040
That's not always the case.

159
00:06:24,040 --> 00:06:25,790
There's some outliers,
like this guy here,

160
00:06:25,790 --> 00:06:28,869
and I don't know
how that happened.

161
00:06:28,869 --> 00:06:30,660
So some really interesting
things going on,

162
00:06:30,660 --> 00:06:32,500
and I'm still trying
to sift through it.

163
00:06:32,500 --> 00:06:34,874
So I talked about
this I think yesterday

164
00:06:34,874 --> 00:06:37,040
in the analysis section,
but there's several sources

165
00:06:37,040 --> 00:06:37,980
of series resistance.

166
00:06:37,980 --> 00:06:41,033
Any of you guys familiar
with what they are?

167
00:06:41,033 --> 00:06:44,327
AUDIENCE: [INAUDIBLE] resistance
and emitter sheet resistance.

168
00:06:44,327 --> 00:06:46,285
PROFESSOR: Emitter sheet
resistance, and then--

169
00:06:46,285 --> 00:06:47,590
AUDIENCE: [INAUDIBLE].

170
00:06:47,590 --> 00:06:49,480
PROFESSOR: So there's line
resistance along the fingers

171
00:06:49,480 --> 00:06:50,480
and then there's
contact resistance,

172
00:06:50,480 --> 00:06:52,760
so I'm going to draw
this out really quickly.

173
00:06:52,760 --> 00:06:55,915
So if you have--
this is your emitter.

174
00:06:58,940 --> 00:07:03,080
You have a contact here
and a contact here.

175
00:07:05,690 --> 00:07:11,430
If you generate an electron
here, it'll diffuse,

176
00:07:11,430 --> 00:07:13,450
hit the junction.

177
00:07:13,450 --> 00:07:16,029
Then it has to flow through
the emitter to the contact.

178
00:07:16,029 --> 00:07:17,820
There's an associated
resistance with that.

179
00:07:17,820 --> 00:07:19,870
That's the series
resistance that we

180
00:07:19,870 --> 00:07:22,510
taught in class that has to do
with the emitter resistance.

181
00:07:22,510 --> 00:07:25,571
Additionally to hop over that
metal semiconductor junction,

182
00:07:25,571 --> 00:07:27,070
there's actually a
resistance there.

183
00:07:27,070 --> 00:07:28,528
That's called a
contact resistance.

184
00:07:32,020 --> 00:07:45,010
So this is emitter and
this is our contact,

185
00:07:45,010 --> 00:07:48,304
and then it has to travel from
the line to the busbar here,

186
00:07:48,304 --> 00:07:49,595
and that's our line resistance.

187
00:07:55,177 --> 00:07:56,760
And I think what's
really limiting us,

188
00:07:56,760 --> 00:08:02,182
because when I looked at both
the shallow and deep emitters,

189
00:08:02,182 --> 00:08:03,640
the series resistance
didn't really

190
00:08:03,640 --> 00:08:05,579
have any noticeable trends.

191
00:08:05,579 --> 00:08:08,120
So I really think that this is
what's limiting us completely,

192
00:08:08,120 --> 00:08:11,752
is our contact resistance,
which is kind of too bad.

193
00:08:11,752 --> 00:08:13,460
So one of the reasons
I think that that's

194
00:08:13,460 --> 00:08:15,810
the case additionally
is that the cells

195
00:08:15,810 --> 00:08:19,497
that performed the best-- I know
Joel's did particularly well.

196
00:08:19,497 --> 00:08:20,580
What was your fill factor?

197
00:08:20,580 --> 00:08:21,270
Do you remember?

198
00:08:21,270 --> 00:08:24,596
AUDIENCE: Well, on the one side
it was 0.5, and on the other

199
00:08:24,596 --> 00:08:26,470
it was 0.64.

200
00:08:26,470 --> 00:08:28,290
PROFESSOR: OK, so
one side was better.

201
00:08:28,290 --> 00:08:31,252
So again, the inhomogeneity
of doing the measurement

202
00:08:31,252 --> 00:08:33,460
is also another indication
that our series resistance

203
00:08:33,460 --> 00:08:36,630
across the contacts
is really poor.

204
00:08:36,630 --> 00:08:39,760
So some of the other
contacts if you look on the--

205
00:08:39,760 --> 00:08:43,151
where's my good piece of chalk?

206
00:08:43,151 --> 00:08:43,650
Right.

207
00:08:43,650 --> 00:08:47,820
If here's our busbar, some
people saw delamination.

208
00:08:47,820 --> 00:08:49,600
Who saw their
contacts delaminated?

209
00:08:49,600 --> 00:08:51,380
Yeah, so a couple people.

210
00:08:51,380 --> 00:08:53,940
And so what that means is
that these contacts actually

211
00:08:53,940 --> 00:08:58,010
peeled up, and that's usually
showing that the-- what's

212
00:08:58,010 --> 00:09:02,020
funny, though, is that
these areas in the middle

213
00:09:02,020 --> 00:09:02,800
didn't delaminate.

214
00:09:02,800 --> 00:09:05,580
It was only the edges
here, and so that's

215
00:09:05,580 --> 00:09:08,370
showing that some parts
are being properly fired

216
00:09:08,370 --> 00:09:11,887
and some parts are
being underfired

217
00:09:11,887 --> 00:09:13,720
because when I fired
at higher temperatures,

218
00:09:13,720 --> 00:09:15,110
I was getting shunting.

219
00:09:15,110 --> 00:09:17,810
And that was due to shunting
through the center area.

220
00:09:17,810 --> 00:09:21,230
So I think that the RTA
that I was using-- this

221
00:09:21,230 --> 00:09:23,870
wasn't the case on some of the
earlier cells I had processed,

222
00:09:23,870 --> 00:09:26,480
but is rather
inhomogeneous in terms

223
00:09:26,480 --> 00:09:30,244
of the temperature profile
it's delivering to our cells.

224
00:09:30,244 --> 00:09:32,493
So I think that could be
part of our problem, as well.

225
00:09:32,493 --> 00:09:34,910
AUDIENCE: How many cells did
he fire up at the same time?

226
00:09:34,910 --> 00:09:35,860
Just one, or--

227
00:09:35,860 --> 00:09:37,871
PROFESSOR: Yeah,
it's one at a time,

228
00:09:37,871 --> 00:09:40,370
and there's some variability
in the peak firing temperature.

229
00:09:42,867 --> 00:09:44,450
So I've learned a
lot about when we're

230
00:09:44,450 --> 00:09:48,310
going to buy an RTA for
our lab, what to look for.

231
00:09:48,310 --> 00:09:49,980
The temporal
resolution of the data

232
00:09:49,980 --> 00:09:53,500
you get out of it and how it
records it is 0.6 seconds.

233
00:09:53,500 --> 00:09:56,080
And when this thing's
ramping at 100 C per second,

234
00:09:56,080 --> 00:09:58,484
your variability is quite large.

235
00:09:58,484 --> 00:10:00,900
So the temperature I read is
not just from the data point,

236
00:10:00,900 --> 00:10:03,150
but I do a linear interpolation
to actually figure out

237
00:10:03,150 --> 00:10:07,600
what the peak is on
both sides of that peak,

238
00:10:07,600 --> 00:10:10,150
and it ends up being about
anywhere from 20 to 40

239
00:10:10,150 --> 00:10:13,660
degrees Celsius higher
using that technique.

240
00:10:13,660 --> 00:10:17,050
And so the variability varies.

241
00:10:17,050 --> 00:10:18,960
It's about 20
degrees plus or minus

242
00:10:18,960 --> 00:10:21,770
10 degrees Celsius in terms
of what I want the peak to be

243
00:10:21,770 --> 00:10:23,910
and what it actually is.

244
00:10:23,910 --> 00:10:27,490
So the RTA that we're using now
is not incredibly reputable.

245
00:10:27,490 --> 00:10:30,130
Also, better RTAs will
have a different heater

246
00:10:30,130 --> 00:10:32,620
on top and bottom, and
so you can actually

247
00:10:32,620 --> 00:10:34,250
heat one side more
than the other.

248
00:10:34,250 --> 00:10:40,160
And generally belt furnaces,
so commercial ones,

249
00:10:40,160 --> 00:10:42,840
will have that same
kind of control.

250
00:10:42,840 --> 00:10:46,780
I know at 1366, they've
actually done optimization

251
00:10:46,780 --> 00:10:48,906
to do heating the top
and bottom differently

252
00:10:48,906 --> 00:10:51,280
because they use different
pace, they have different heat

253
00:10:51,280 --> 00:10:53,225
requirements,
thermal requirements.

254
00:10:53,225 --> 00:10:55,770
And so nailing that
temperature and that profile

255
00:10:55,770 --> 00:10:58,025
actually takes a lot of
iteration and a lot of time.

256
00:10:58,025 --> 00:11:00,780
And it's really just learning
the tell as opposed to,

257
00:11:00,780 --> 00:11:02,460
the again, good
science behind it.

258
00:11:02,460 --> 00:11:07,270
So it can be rather difficult
and time-consuming to do.

259
00:11:07,270 --> 00:11:10,242
So that's what I have
to say about the cells.

260
00:11:10,242 --> 00:11:12,450
I think overall, though, I
think the efficiencies are

261
00:11:12,450 --> 00:11:15,250
on the order of 6% to I think
9% for some of the cells,

262
00:11:15,250 --> 00:11:18,660
so that's not too bad.

263
00:11:18,660 --> 00:11:21,850
I'm still rather
pleased with that.

264
00:11:21,850 --> 00:11:24,927
Last time I think
Tonio stopped right

265
00:11:24,927 --> 00:11:27,260
before he was going to talk
about performance of modules

266
00:11:27,260 --> 00:11:31,710
in the field and kind
of what's the difference

267
00:11:31,710 --> 00:11:35,070
between your cell efficiency
and your module efficiency.

268
00:11:35,070 --> 00:11:36,480
Modules are put out in the sun.

269
00:11:36,480 --> 00:11:37,480
They heat up.

270
00:11:37,480 --> 00:11:41,440
They can get shaded by
snow, rain, clouds, trees,

271
00:11:41,440 --> 00:11:44,050
and also the modulus are made
up of many different cells that

272
00:11:44,050 --> 00:11:45,460
are connected in
series and parallel,

273
00:11:45,460 --> 00:11:47,120
and how does that affect
their overall performance

274
00:11:47,120 --> 00:11:48,110
of the module?

275
00:11:48,110 --> 00:11:51,230
And so that's what we're going
to talk about really quickly.

276
00:11:51,230 --> 00:11:53,050
So first off, why does
temperature matter?

277
00:11:53,050 --> 00:11:56,114
So all solar cells when
they're measured either at NREL

278
00:11:56,114 --> 00:11:58,280
or in our lab, are measured
under standard operating

279
00:11:58,280 --> 00:12:00,405
conditions-- or standard
testing conditions, sorry.

280
00:12:00,405 --> 00:12:03,390
Certainly not standard
operating conditions.

281
00:12:03,390 --> 00:12:06,060
Standard testing conditions
are done at 25 C.

282
00:12:06,060 --> 00:12:08,280
Generally the cells
are actively cooled

283
00:12:08,280 --> 00:12:11,290
to maintain that temperature
of room temperature,

284
00:12:11,290 --> 00:12:13,460
and most semiconductor
simulations you perform

285
00:12:13,460 --> 00:12:15,280
are generally done
at 300 Kelvin.

286
00:12:15,280 --> 00:12:17,350
I know all the calculations
I did for fitting

287
00:12:17,350 --> 00:12:20,100
your [INAUDIBLE] diode
curve were done at 300

288
00:12:20,100 --> 00:12:23,640
K. Typical operation can
actually be pretty hot,

289
00:12:23,640 --> 00:12:25,264
so 50 to 65 degrees Celsius.

290
00:12:25,264 --> 00:12:27,805
I actually don't know if that
is in Fahrenheit, but it's hot.

291
00:12:32,390 --> 00:12:35,240
It's hotter than
Phoenix, and I'm

292
00:12:35,240 --> 00:12:37,250
sure they get even
hotter in Phoenix.

293
00:12:43,292 --> 00:12:47,800
So the effective temperature--
so how does it affect Voc?

294
00:12:47,800 --> 00:12:50,245
If you recall, this
is our diad equation.

295
00:12:50,245 --> 00:12:51,620
We have our
illumination current,

296
00:12:51,620 --> 00:12:52,950
which we'll get to
how that's affected

297
00:12:52,950 --> 00:12:53,890
by temperature in a second.

298
00:12:53,890 --> 00:12:55,140
But for now, just ignore that.

299
00:12:55,140 --> 00:12:57,270
Assume it stays constant.

300
00:12:57,270 --> 00:13:01,080
So one of things that's
really affecting your Voc-- so

301
00:13:01,080 --> 00:13:05,930
remember, your Voc is going
to make your current go to 0,

302
00:13:05,930 --> 00:13:09,300
and the main thing
that's being altered

303
00:13:09,300 --> 00:13:12,060
is your saturation current,
so your intrinsic carrier

304
00:13:12,060 --> 00:13:14,580
concentration, your diffusivity.

305
00:13:14,580 --> 00:13:16,590
So who here thinks
your Voc is going

306
00:13:16,590 --> 00:13:18,720
to go up with temperature?

307
00:13:18,720 --> 00:13:20,480
Anyone?

308
00:13:20,480 --> 00:13:23,690
And who thinks it'll go down?

309
00:13:23,690 --> 00:13:26,620
Who has no idea?

310
00:13:26,620 --> 00:13:30,050
OK, so who raised their
hand for it'll go down?

311
00:13:30,050 --> 00:13:30,550
Sorry.

312
00:13:30,550 --> 00:13:32,800
Ben, do you have any ideas
why that might be the case?

313
00:13:35,288 --> 00:13:36,767
AUDIENCE: No.

314
00:13:36,767 --> 00:13:38,740
[LAUGHTER]

315
00:13:38,740 --> 00:13:40,735
PROFESSOR: So what's
going to happen to your I0

316
00:13:40,735 --> 00:13:41,860
with increased temperature?

317
00:13:42,892 --> 00:13:44,975
AUDIENCE: The intrinsic
carrier concentration, ni,

318
00:13:44,975 --> 00:13:48,010
is going to increase, and
that should increase I0?

319
00:13:48,010 --> 00:13:50,060
PROFESSOR: That
should increase I0,

320
00:13:50,060 --> 00:13:54,320
so therefore it's going to take
a smaller voltage to counteract

321
00:13:54,320 --> 00:13:57,160
I sub L to make I equal to 0.

322
00:13:57,160 --> 00:13:59,490
And so why does N sub I
increase with temperature?

323
00:14:02,960 --> 00:14:06,330
AUDIENCE: Greater
thermal energy present

324
00:14:06,330 --> 00:14:10,170
increases the film probability
that electrons can be

325
00:14:10,170 --> 00:14:12,100
excited from the [INAUDIBLE]?

326
00:14:12,100 --> 00:14:13,933
PROFESSOR: Right, so
it increases the number

327
00:14:13,933 --> 00:14:16,230
of thermally-promoted carriers.

328
00:14:16,230 --> 00:14:19,270
And so that effect can
actually be pretty stark,

329
00:14:19,270 --> 00:14:21,150
and now we go to a demo.

330
00:14:21,150 --> 00:14:22,640
Yay.

331
00:14:22,640 --> 00:14:26,260
All right, so what we have here
is our favorite solar cell.

332
00:14:26,260 --> 00:14:29,960
This came, again, from
those small little toy

333
00:14:29,960 --> 00:14:32,740
cars that we pulled off, and
again, these solar cells,

334
00:14:32,740 --> 00:14:34,730
the cars will
assume were nothing,

335
00:14:34,730 --> 00:14:37,150
and these are about $10
apiece, so pretty cheap.

336
00:14:37,150 --> 00:14:39,566
Probably cheaper than your
solar cells that you guys made.

337
00:14:41,322 --> 00:14:43,030
And, well, certain I
know they are cheap.

338
00:14:43,030 --> 00:14:45,180
The wafers themselves, by
the way, were $16 apiece,

339
00:14:45,180 --> 00:14:47,388
so saying they were a dollar
was a little inaccurate.

340
00:14:49,800 --> 00:14:53,020
So anyhow, we have this
hooked up to a multimeter

341
00:14:53,020 --> 00:14:56,754
and we're going to measure
the voltage off of it.

342
00:14:56,754 --> 00:14:58,670
I'm going to illuminate
it with our desk lamp,

343
00:14:58,670 --> 00:15:00,669
and we can actually get
a pretty decent voltage.

344
00:15:00,669 --> 00:15:02,660
I can't actually see this.

345
00:15:02,660 --> 00:15:05,830
So it's about 0.57
volts, and now we're

346
00:15:05,830 --> 00:15:07,380
going to subject
it to temperature.

347
00:15:10,134 --> 00:15:11,800
This is obviously a
gross overestimation

348
00:15:11,800 --> 00:15:13,670
of what's actually
going to happen,

349
00:15:13,670 --> 00:15:15,360
but can someone read
off-- Joel, can you

350
00:15:15,360 --> 00:15:16,610
read the temperature for me?

351
00:15:16,610 --> 00:15:18,360
AUDIENCE: The temperature,
or the voltage?

352
00:15:18,360 --> 00:15:19,200
PROFESSOR: Sorry, the voltage.

353
00:15:19,200 --> 00:15:19,825
AUDIENCE: Yeah.

354
00:15:19,825 --> 00:15:32,300
We're at 0.561 and decreasing
down to 0.55, 0.54, 0.53--

355
00:15:32,300 --> 00:15:34,410
PROFESSOR: So it's going down.

356
00:15:34,410 --> 00:15:36,700
So congratulations
the people who

357
00:15:36,700 --> 00:15:38,015
said it would go down, correct.

358
00:15:41,040 --> 00:15:43,080
Wow, that's actually quite hot.

359
00:15:43,080 --> 00:15:45,420
Anyhow, so that's what
happens in the field.

360
00:15:45,420 --> 00:15:48,590
So when these cells heat up,
you actually lose on your Voc.

361
00:15:51,630 --> 00:15:54,220
And if you really want to, you
can go through all the math.

362
00:15:54,220 --> 00:15:56,450
You know what the
equation is for your Voc.

363
00:15:56,450 --> 00:16:02,030
It's shown here, and you can
go through this if you like.

364
00:16:02,030 --> 00:16:05,110
It's all on PVCDROM if you want
to go through the derivation.

365
00:16:05,110 --> 00:16:07,820
You then can take the derivative
with respect to temperature.

366
00:16:07,820 --> 00:16:10,260
And if you plug in values
for crystalline silicon,

367
00:16:10,260 --> 00:16:13,730
it comes to about 2.2
millivolts per degree Celsius,

368
00:16:13,730 --> 00:16:15,270
which sounds like
a small number,

369
00:16:15,270 --> 00:16:18,065
but it's at around 0.1
volts if you go up to 65 C,

370
00:16:18,065 --> 00:16:21,360
so that's actually quite
a substantial amount.

371
00:16:21,360 --> 00:16:26,000
And again, when you're
going from 0.6 to 0.59,

372
00:16:26,000 --> 00:16:30,780
these small margins can make or
break a lot of installations,

373
00:16:30,780 --> 00:16:32,670
so that's an important one.

374
00:16:32,670 --> 00:16:36,810
OK, now this one's
a little harder.

375
00:16:36,810 --> 00:16:41,270
What do you think will
happen with your illumination

376
00:16:41,270 --> 00:16:43,799
current, or your ISC?

377
00:16:43,799 --> 00:16:45,090
Who thinks it's going to go up?

378
00:16:47,650 --> 00:16:48,150
Anyone?

379
00:16:48,150 --> 00:16:51,400
Who thinks it's
going to go down?

380
00:16:51,400 --> 00:16:54,269
Who thinks it's going
to stay the same?

381
00:16:54,269 --> 00:16:56,560
Who has absolutely no idea
and didn't raise their hand?

382
00:16:56,560 --> 00:16:58,560
OK, that's totally fine.

383
00:16:58,560 --> 00:17:01,420
This confused me,
as well, by the way.

384
00:17:01,420 --> 00:17:04,060
So when we think about it, what
is our short-circuit current

385
00:17:04,060 --> 00:17:05,329
proportional to generally?

386
00:17:07,540 --> 00:17:08,040
Anyone?

387
00:17:08,040 --> 00:17:09,172
AUDIENCE: [INAUDIBLE].

388
00:17:09,172 --> 00:17:11,630
PROFESSOR: Right, so generally
your illumination intensity.

389
00:17:11,630 --> 00:17:14,846
And what photons generally
are we collecting?

390
00:17:14,846 --> 00:17:16,130
AUDIENCE: Super bandgap.

391
00:17:16,130 --> 00:17:17,250
PROFESSOR: Super bandgap.

392
00:17:17,250 --> 00:17:20,130
So what could be happening is
our bandgap could be changing.

393
00:17:20,130 --> 00:17:23,050
I sub L, as Ben pointed
out, should increase

394
00:17:23,050 --> 00:17:26,829
the flux of photons
above the bandgap,

395
00:17:26,829 --> 00:17:30,940
and EG actually decreases
the temperature.

396
00:17:30,940 --> 00:17:33,730
So your bandgap actually
usually increases--

397
00:17:33,730 --> 00:17:35,946
the true bandgap of
silicon defined--

398
00:17:35,946 --> 00:17:37,320
most properties
of semiconductors

399
00:17:37,320 --> 00:17:39,864
are actually
defined at 0 Kelvin,

400
00:17:39,864 --> 00:17:41,280
and semiconductors
are technically

401
00:17:41,280 --> 00:17:43,790
insulators because they
don't conduct electricity

402
00:17:43,790 --> 00:17:45,142
at 0 Kelvin.

403
00:17:45,142 --> 00:17:47,350
That just means semiconductor
is a class of materials

404
00:17:47,350 --> 00:17:50,870
that have a small enough bandgap
that they can thermally promote

405
00:17:50,870 --> 00:17:53,370
carriers that they can actually
conduct some electricity

406
00:17:53,370 --> 00:17:54,970
at room temperature.

407
00:17:54,970 --> 00:17:57,900
But the true bandgap of
silicon's around 1.17.

408
00:17:57,900 --> 00:17:59,630
And as you increase
temperature, you

409
00:17:59,630 --> 00:18:01,030
get a reduction of the bandgap.

410
00:18:01,030 --> 00:18:03,279
And so what should happen
is your ISC should increase,

411
00:18:03,279 --> 00:18:06,060
but only very, very slightly.

412
00:18:06,060 --> 00:18:10,750
So if you go over the derivation
for the actual decrease

413
00:18:10,750 --> 00:18:14,520
in the overall efficiency,
there's a few things.

414
00:18:14,520 --> 00:18:19,950
One, remember your
maximum power is

415
00:18:19,950 --> 00:18:22,630
your Voc times your ISC
times your fill factor.

416
00:18:22,630 --> 00:18:24,730
So if you want the
derivative of your max power

417
00:18:24,730 --> 00:18:26,596
with respect to
temperature, you need

418
00:18:26,596 --> 00:18:28,470
to do the partial
derivatives and sum them up

419
00:18:28,470 --> 00:18:30,053
for all of those
different components.

420
00:18:30,053 --> 00:18:33,600
So that's what this
calculation's doing.

421
00:18:33,600 --> 00:18:36,540
And again, this is what
it is for fill factor.

422
00:18:36,540 --> 00:18:38,880
This is all in Martin
Green's paper down here

423
00:18:38,880 --> 00:18:41,960
and also on PVCDROM if you
guys are interested in that.

424
00:18:41,960 --> 00:18:45,230
Martin Green's
tabulated a lot of this,

425
00:18:45,230 --> 00:18:47,140
making these general
expressions and also

426
00:18:47,140 --> 00:18:51,990
fitting it to experimental
data to get that information.

427
00:18:51,990 --> 00:18:55,800
And then it ends up being
not a negligible percent,

428
00:18:55,800 --> 00:19:00,940
but about half a percent per
degree Celsius for silicon.

429
00:19:00,940 --> 00:19:03,160
So if you go up
by 40 degrees, you

430
00:19:03,160 --> 00:19:07,985
can see that temperature's a
non-negligible effect in terms

431
00:19:07,985 --> 00:19:08,610
of performance.

432
00:19:11,306 --> 00:19:13,529
AUDIENCE: [INAUDIBLE]
silicon's kind

433
00:19:13,529 --> 00:19:18,220
of below the ideal bandgap for
a single-junction solar cell?

434
00:19:18,220 --> 00:19:18,850
PROFESSOR: Yes.

435
00:19:18,850 --> 00:19:20,975
AUDIENCE: If you had a
semi-conductor with one that

436
00:19:20,975 --> 00:19:23,098
was above that, would an
increase in temperature

437
00:19:23,098 --> 00:19:25,223
increase the efficiency
because it's getting simply

438
00:19:25,223 --> 00:19:26,890
closer to that bandgap?

439
00:19:26,890 --> 00:19:28,349
PROFESSOR: Oh,
that's a good point.

440
00:19:28,349 --> 00:19:30,723
Generally temperature's a lot
worse because you're really

441
00:19:30,723 --> 00:19:31,690
going to hurt your Voc.

442
00:19:31,690 --> 00:19:34,754
You would increase your J0
tremendously with temperature

443
00:19:34,754 --> 00:19:36,670
because you're creating
more thermal carriers,

444
00:19:36,670 --> 00:19:39,336
so your diffusion current that's
counteracting your illumination

445
00:19:39,336 --> 00:19:41,470
current is going to
increase dramatically,

446
00:19:41,470 --> 00:19:44,410
so temperature is
always your enemy.

447
00:19:44,410 --> 00:19:46,360
There was a great
picture on PVCDROM

448
00:19:46,360 --> 00:19:49,550
of a solar install in
Antarctica and its perform

449
00:19:49,550 --> 00:19:52,250
well above spec,
which is kind of cool.

450
00:19:54,960 --> 00:19:56,870
So yeah, temperature's
generally an enemy.

451
00:19:56,870 --> 00:19:58,590
And so a lot of people's
ideas have often

452
00:19:58,590 --> 00:20:01,075
been like to do active cooling
and use that for waste heat,

453
00:20:01,075 --> 00:20:03,450
and there's a lot of reasons
why that doesn't necessarily

454
00:20:03,450 --> 00:20:05,260
economically make sense.

455
00:20:05,260 --> 00:20:07,370
But when you do, let's
say concentrated solar,

456
00:20:07,370 --> 00:20:08,910
for example, and
I think Tonio will

457
00:20:08,910 --> 00:20:11,010
talk about this in
the next lecture,

458
00:20:11,010 --> 00:20:13,290
but concentrated solar is
you're putting in not just

459
00:20:13,290 --> 00:20:15,460
one sun but about
1,000, maybe 10,000 suns

460
00:20:15,460 --> 00:20:17,206
on one small device.

461
00:20:17,206 --> 00:20:19,580
And the idea there is you can
have a really small device,

462
00:20:19,580 --> 00:20:22,082
have it be really expensive
but incredibly efficient,

463
00:20:22,082 --> 00:20:23,790
on the order of 40%
efficient because you

464
00:20:23,790 --> 00:20:26,040
have these stacks of
difference semiconductors

465
00:20:26,040 --> 00:20:30,590
that absorb different regions
of the light preferentially.

466
00:20:30,590 --> 00:20:33,410
And those heat up tremendously
when you're being subjective.

467
00:20:33,410 --> 00:20:37,235
It's like an ant under a
magnifying glass or something.

468
00:20:37,235 --> 00:20:37,860
It can burn it.

469
00:20:37,860 --> 00:20:40,770
So they actually active
cooling to cool those cells.

470
00:20:40,770 --> 00:20:43,020
And they also have to track
the sun, as well, in order

471
00:20:43,020 --> 00:20:45,590
to concentrate it.

472
00:20:45,590 --> 00:20:48,430
And the last thing which you
guys did for your exam number

473
00:20:48,430 --> 00:20:51,870
one was the effect of light
intensity, where you looked

474
00:20:51,870 --> 00:20:53,500
at light intensity
throughout the day,

475
00:20:53,500 --> 00:20:56,590
assumed some sinusoid
for the incident light,

476
00:20:56,590 --> 00:20:58,940
and you measured how
your efficiency changes.

477
00:20:58,940 --> 00:21:03,620
And you remember, your Voc
decreases with light intensity

478
00:21:03,620 --> 00:21:07,350
with your JSC, and that's due
to the decrease in the photon

479
00:21:07,350 --> 00:21:10,090
flux.

480
00:21:10,090 --> 00:21:11,970
And your efficiency
goes down according

481
00:21:11,970 --> 00:21:15,735
to those two equations there,
and I think the derating factor

482
00:21:15,735 --> 00:21:18,350
if you assume a sinusoid
that hits at one sun

483
00:21:18,350 --> 00:21:21,699
and then declines, you
should derate your--

484
00:21:21,699 --> 00:21:23,240
if you use instead
of using peak sun,

485
00:21:23,240 --> 00:21:25,161
you want to derate
that by about 20%.

486
00:21:25,161 --> 00:21:26,660
So it's actually
pretty substantial.

487
00:21:26,660 --> 00:21:30,230
I was kind of surprised
at that finding.

488
00:21:30,230 --> 00:21:32,540
So when V cells go
out into the field,

489
00:21:32,540 --> 00:21:36,760
they are not just
one individual cell.

490
00:21:36,760 --> 00:21:41,910
So oftentimes you string these
in series and then in parallel.

491
00:21:41,910 --> 00:21:44,850
So right now we're just going to
look at series and in parallel

492
00:21:44,850 --> 00:21:48,480
solely, and what you see here
are your three different cells

493
00:21:48,480 --> 00:21:49,460
all mounted in series.

494
00:21:49,460 --> 00:21:50,840
So the top of this
cell is connected

495
00:21:50,840 --> 00:21:52,285
to the bottom of this
cell, the top of this cell

496
00:21:52,285 --> 00:21:54,290
is connected to the
bottom of this cell.

497
00:21:54,290 --> 00:21:58,080
Our bad cell, marked here with
the I guess the denim pattern.

498
00:21:58,080 --> 00:22:00,786
I don't know what that
was in PowerPoint.

499
00:22:00,786 --> 00:22:02,160
But what's important
here is that

500
00:22:02,160 --> 00:22:03,550
because the current
flowing to this one

501
00:22:03,550 --> 00:22:05,760
is also flowing to this
one is flowing to this one,

502
00:22:05,760 --> 00:22:09,120
the voltages add and the
currents are all matched.

503
00:22:09,120 --> 00:22:12,620
So the current out of
this stack of cells

504
00:22:12,620 --> 00:22:14,540
is limited by your weakest cell.

505
00:22:14,540 --> 00:22:17,030
It can't exceed your
short-circuit current

506
00:22:17,030 --> 00:22:18,420
of your weakest cell.

507
00:22:18,420 --> 00:22:22,390
So you can see that your
operating point for these three

508
00:22:22,390 --> 00:22:24,610
cells is not ideal for this one.

509
00:22:24,610 --> 00:22:27,157
It's definitely way off
the peak operating point

510
00:22:27,157 --> 00:22:28,990
for this cell and way
off the peak operating

511
00:22:28,990 --> 00:22:30,720
point for that cell.

512
00:22:30,720 --> 00:22:32,550
So it can be a pretty
detrimental effect,

513
00:22:32,550 --> 00:22:34,630
and this is why
often in industry

514
00:22:34,630 --> 00:22:37,570
when you do a series of
testing on your cells,

515
00:22:37,570 --> 00:22:41,080
each cell gets tested either
using suns Voc or a very, very

516
00:22:41,080 --> 00:22:44,740
fast IV measurement.

517
00:22:44,740 --> 00:22:47,000
You can [? bin ?] cells
based on their performance

518
00:22:47,000 --> 00:22:50,457
and then match them so that
if you all your cells are

519
00:22:50,457 --> 00:22:52,873
perfectly matched, they're all
operating at the peak power

520
00:22:52,873 --> 00:22:55,590
point and you're getting the
most out of all those cells.

521
00:22:55,590 --> 00:23:01,150
And so that's why reducing the
variability within your process

522
00:23:01,150 --> 00:23:03,397
can really, really increase
your module performance,

523
00:23:03,397 --> 00:23:05,230
and that's something a
lot of companies work

524
00:23:05,230 --> 00:23:07,730
really, really hard at.

525
00:23:07,730 --> 00:23:10,040
Now, the other reason why
series is important to study

526
00:23:10,040 --> 00:23:11,530
is the effect of shading.

527
00:23:11,530 --> 00:23:13,940
So when you decrease--
so let's say

528
00:23:13,940 --> 00:23:16,931
a leaf falls on your solar
panel or it's partially shaded

529
00:23:16,931 --> 00:23:18,180
because you put it near trees.

530
00:23:18,180 --> 00:23:22,240
If you look at the BigBelly
Solar ones right near campus,

531
00:23:22,240 --> 00:23:23,420
they're always in shade.

532
00:23:23,420 --> 00:23:25,747
It makes no sense to me.

533
00:23:25,747 --> 00:23:27,580
I think they might get
one hour of sun a day

534
00:23:27,580 --> 00:23:30,210
in certain times of the
Anyhow, And so if you partially

535
00:23:30,210 --> 00:23:32,060
shade one of the
cells, you can see

536
00:23:32,060 --> 00:23:36,840
that the IV curve drops
down, and now this cell

537
00:23:36,840 --> 00:23:39,026
is running in reverse bias.

538
00:23:39,026 --> 00:23:42,800
And when you're in reverse bias,
you're running a lot of current

539
00:23:42,800 --> 00:23:44,520
through your shunts.

540
00:23:44,520 --> 00:23:46,880
And for those who were
here this morning,

541
00:23:46,880 --> 00:23:48,630
we know that shunts
are really, really bad

542
00:23:48,630 --> 00:23:50,963
because that's where all your
current's running through.

543
00:23:50,963 --> 00:23:53,107
And if your current's
running through one localized

544
00:23:53,107 --> 00:23:54,940
spot, that's where it's
heating up the most,

545
00:23:54,940 --> 00:23:57,106
and it can heat up to
temperatures that can actually

546
00:23:57,106 --> 00:23:59,750
melt the encapsulant or
completely destroy the cell,

547
00:23:59,750 --> 00:24:01,700
so this can actually
be rather destructive.

548
00:24:01,700 --> 00:24:07,750
And generally once your cells
are made and encapsulated

549
00:24:07,750 --> 00:24:12,230
in the module, it's really
hard to remove them and replace

550
00:24:12,230 --> 00:24:12,730
that module.

551
00:24:12,730 --> 00:24:15,620
Pretty much the whole
module's kaput at that point,

552
00:24:15,620 --> 00:24:16,810
so it's pretty destructive.

553
00:24:16,810 --> 00:24:20,280
So what people
generally do for-- yeah?

554
00:24:20,280 --> 00:24:23,320
AUDIENCE: So in maintenance,
like yearly maintenance,

555
00:24:23,320 --> 00:24:26,700
you should clean the
panel because of the dirt,

556
00:24:26,700 --> 00:24:31,560
or is there any other thing
that you should do [INAUDIBLE]?

557
00:24:31,560 --> 00:24:33,060
PROFESSOR: There's
important things.

558
00:24:33,060 --> 00:24:35,982
So you don't want--
it helps if they're

559
00:24:35,982 --> 00:24:38,550
angled so snow and other
things will fall off of it.

560
00:24:38,550 --> 00:24:40,520
You don't want to
put it in areas

561
00:24:40,520 --> 00:24:43,103
next to a chimney, for example,
where certain parts of the day

562
00:24:43,103 --> 00:24:47,810
it will be shaded,
or things like that.

563
00:24:47,810 --> 00:24:51,870
So I've seen some
studies on cleaning,

564
00:24:51,870 --> 00:24:54,080
and it all depends on if it
economically makes sense.

565
00:24:54,080 --> 00:24:56,132
If you want to pay someone
to clean your panels,

566
00:24:56,132 --> 00:24:56,840
that's expensive.

567
00:24:56,840 --> 00:24:59,710
Is it worth it for them
to do that if you're only

568
00:24:59,710 --> 00:25:00,802
saving just a little bit?

569
00:25:00,802 --> 00:25:03,260
Obviously, the shading's so
bad that it could detrimentally

570
00:25:03,260 --> 00:25:04,470
ruin and destroy your module.

571
00:25:04,470 --> 00:25:07,420
It's really-- you
want to do that?

572
00:25:07,420 --> 00:25:10,120
Ways to prevent the destruction
is actually a lot of cells

573
00:25:10,120 --> 00:25:12,470
will have bypass diodes.

574
00:25:12,470 --> 00:25:15,456
So if a string goes
down, the current

575
00:25:15,456 --> 00:25:17,080
will just [? flow ?]
through that diode

576
00:25:17,080 --> 00:25:20,790
and won't destroy the
cell, and so that's one way

577
00:25:20,790 --> 00:25:22,740
of kind of fixing that problem.

578
00:25:22,740 --> 00:25:23,740
That's a great question.

579
00:25:29,980 --> 00:25:33,140
So now we're stacking
ourselves in parallel.

580
00:25:33,140 --> 00:25:35,380
Normally, again, in a module,
you have them in series

581
00:25:35,380 --> 00:25:36,940
and then these stacks
of cells in series

582
00:25:36,940 --> 00:25:38,856
are actually put in
parallel, but there's just

583
00:25:38,856 --> 00:25:41,670
an easy illustration.

584
00:25:41,670 --> 00:25:44,740
Again, cells in parallel,
you're matching in voltage

585
00:25:44,740 --> 00:25:47,580
and the currents
are what add in.

586
00:25:47,580 --> 00:25:51,190
And so if one cell let's say
has a low output voltage,

587
00:25:51,190 --> 00:25:57,375
it's going to shift
the operating power

588
00:25:57,375 --> 00:26:00,175
point for the other two cells
off of their maximum power

589
00:26:00,175 --> 00:26:05,660
point and you're going
to get a reduced output.

590
00:26:05,660 --> 00:26:10,450
And so this is actually
very analogous to this idea

591
00:26:10,450 --> 00:26:12,930
that when you make
your cells larger,

592
00:26:12,930 --> 00:26:15,930
generally their efficiency
and performance go down.

593
00:26:15,930 --> 00:26:18,046
You have a higher chance
of inhomogeneities

594
00:26:18,046 --> 00:26:20,420
and other things that might
detrimentally hurt your cell.

595
00:26:20,420 --> 00:26:22,290
And this is kind of
analogous to saying

596
00:26:22,290 --> 00:26:25,470
when you hook a whole
bunch of cells in parallel,

597
00:26:25,470 --> 00:26:27,124
you're limited by
your worst cell.

598
00:26:27,124 --> 00:26:29,040
And in this case, you're
limited by your worst

599
00:26:29,040 --> 00:26:31,690
region of that
cell, and so that's

600
00:26:31,690 --> 00:26:33,670
the same kind of analogy.

601
00:26:33,670 --> 00:26:35,540
And I think it's kind
of cool, and so you

602
00:26:35,540 --> 00:26:36,710
can see this for a lot
of different cells.

603
00:26:36,710 --> 00:26:38,550
So these are actually
getting pretty big.

604
00:26:38,550 --> 00:26:41,230
So when you're getting
to-- I don't know

605
00:26:41,230 --> 00:26:42,700
if they make cells this big.

606
00:26:42,700 --> 00:26:45,540
But it's 100 by 100.

607
00:26:45,540 --> 00:26:46,545
That's a meter squared.

608
00:26:46,545 --> 00:26:48,170
Wow, so maybe Cad-tel
can get that big.

609
00:26:48,170 --> 00:26:48,920
Actually, they do.

610
00:26:48,920 --> 00:26:51,484
Look, here's First
Solar right here.

611
00:26:51,484 --> 00:26:52,900
AUDIENCE: [INAUDIBLE].

612
00:26:52,900 --> 00:26:53,935
PROFESSOR: I'm sorry?

613
00:26:53,935 --> 00:26:55,026
AUDIENCE: [INAUDIBLE].

614
00:26:55,026 --> 00:26:56,650
PROFESSOR: Oh, OK.
[INAUDIBLE] modules.

615
00:26:56,650 --> 00:26:59,300
So some of the
cells you can see.

616
00:26:59,300 --> 00:27:02,612
So here's cells--
amorphous, submodule.

617
00:27:02,612 --> 00:27:04,720
Anyhow, so the idea
is that you're, again,

618
00:27:04,720 --> 00:27:11,140
limited by your most
defective region in the cell.

619
00:27:11,140 --> 00:27:12,730
And again, illustrated
by this idea

620
00:27:12,730 --> 00:27:17,720
here is that if you take a cell,
you split it up into a grid,

621
00:27:17,720 --> 00:27:21,090
and you model it as a bunch
of many cells operating

622
00:27:21,090 --> 00:27:28,410
in parallel, you're again
limited by that bad region.

623
00:27:28,410 --> 00:27:33,010
And so this looks
a little outdated.

624
00:27:33,010 --> 00:27:34,940
So this is 2006.

625
00:27:34,940 --> 00:27:39,430
So this is cell performance
versus module performance,

626
00:27:39,430 --> 00:27:42,200
and you can see that it's
also quite significant.

627
00:27:42,200 --> 00:27:44,820
One of the other things
that's different about cells

628
00:27:44,820 --> 00:27:47,930
and modules is that you
have glass in front.

629
00:27:47,930 --> 00:27:50,754
So when you optimize--
for example, in class

630
00:27:50,754 --> 00:27:52,420
we optimized our
silicon nitride coating

631
00:27:52,420 --> 00:27:55,680
for an air-silicon interface.

632
00:27:55,680 --> 00:28:00,720
Modules, you optimize at silicon
nitride coating for a silicon

633
00:28:00,720 --> 00:28:02,320
to glass interface.

634
00:28:02,320 --> 00:28:06,950
And the glass, which has an
index refraction of about 1.5,

635
00:28:06,950 --> 00:28:09,340
you're automatically going
to reflect 4% of the light,

636
00:28:09,340 --> 00:28:11,050
so that light's lost.

637
00:28:11,050 --> 00:28:13,270
But it does help with
your reflection losses

638
00:28:13,270 --> 00:28:16,430
because you now have a
graded index of refraction.

639
00:28:16,430 --> 00:28:18,567
So it does reduce some of
those reflection losses

640
00:28:18,567 --> 00:28:19,400
from that interface.

641
00:28:19,400 --> 00:28:19,899
Yeah?

642
00:28:19,899 --> 00:28:23,114
AUDIENCE: Is this just silicon?

643
00:28:23,114 --> 00:28:24,780
PROFESSOR: I'm guessing
yes because this

644
00:28:24,780 --> 00:28:26,250
is from Richard Swanson.

645
00:28:26,250 --> 00:28:27,547
This is Tonio's slide.

646
00:28:27,547 --> 00:28:29,630
My guess would be yes, but
don't quote me on that.

647
00:28:30,586 --> 00:28:34,897
AUDIENCE: So for the modules,
how is voltage and current

648
00:28:34,897 --> 00:28:35,397
controlled?

649
00:28:35,397 --> 00:28:37,757
I'm assuming that they control
it at a specific point.

650
00:28:37,757 --> 00:28:39,840
PROFESSOR: Yes, so that's
a great, great question,

651
00:28:39,840 --> 00:28:40,340
actually.

652
00:28:40,340 --> 00:28:42,370
So your modules are
hooked up to an inverter.

653
00:28:42,370 --> 00:28:45,420
The output of your solar
cells are direct current,

654
00:28:45,420 --> 00:28:48,260
and so you have to convert
that to alternating current.

655
00:28:48,260 --> 00:28:50,700
And the inverters are
actually rather intelligent.

656
00:28:50,700 --> 00:28:53,831
They'll actually constantly scan
at a very high frequency where

657
00:28:53,831 --> 00:28:55,830
it should be operating
and taking the derivative

658
00:28:55,830 --> 00:28:59,032
and finding where
that optimal point is.

659
00:28:59,032 --> 00:29:00,865
And so one of the other
ideas is that if you

660
00:29:00,865 --> 00:29:02,630
have a bunch of
modules, usually those

661
00:29:02,630 --> 00:29:04,049
are also hooked up in parallel.

662
00:29:04,049 --> 00:29:05,840
And the idea is to have
instead of just one

663
00:29:05,840 --> 00:29:09,190
inverter, a microinverter-- so
small inverters for each one.

664
00:29:09,190 --> 00:29:11,960
So that way, each of those are
operating at the maximum power

665
00:29:11,960 --> 00:29:13,360
point.

666
00:29:13,360 --> 00:29:16,310
And I think problem is that
inverters are expensive

667
00:29:16,310 --> 00:29:17,930
and the larger
they are, generally

668
00:29:17,930 --> 00:29:20,710
the more efficient they
are and cheaper can be.

669
00:29:20,710 --> 00:29:24,675
So I don't know how far
that idea's been realized,

670
00:29:24,675 --> 00:29:26,127
but it's a good question.

671
00:29:26,127 --> 00:29:26,960
Does that answer it?

672
00:29:29,540 --> 00:29:33,280
So we have-- good,
plenty of time.

673
00:29:33,280 --> 00:29:37,710
So now we're moving
onto advanced concepts,

674
00:29:37,710 --> 00:29:40,952
which, again, area
that I'm working in.

675
00:29:40,952 --> 00:29:41,660
Kind of exciting.

676
00:29:41,660 --> 00:29:46,550
These are ideas that are
very, very high efficiency,

677
00:29:46,550 --> 00:29:49,836
potentially very
low-cost, and that's

678
00:29:49,836 --> 00:29:50,960
kind of the allure to them.

679
00:29:50,960 --> 00:29:53,460
And they're great science
and research projects,

680
00:29:53,460 --> 00:29:56,946
so they're really
exciting to be in.

681
00:29:56,946 --> 00:29:58,570
So one of things that
kind of motivates

682
00:29:58,570 --> 00:30:01,766
this work-- I think you guys
have seen this equation.

683
00:30:01,766 --> 00:30:02,890
I know I've used it before.

684
00:30:02,890 --> 00:30:07,410
But you know that the defining
metric for the performance

685
00:30:07,410 --> 00:30:09,310
of a cell in terms
of its economic cost

686
00:30:09,310 --> 00:30:13,100
is dollars per watt, so
dollars per watt peak.

687
00:30:13,100 --> 00:30:15,930
This is all defined at
your peak sun elimination.

688
00:30:15,930 --> 00:30:17,671
This makes it
location-blind, right?

689
00:30:17,671 --> 00:30:20,170
The really important thing when
integrating it onto the grid

690
00:30:20,170 --> 00:30:22,100
is how much are you
paying in terms of cents

691
00:30:22,100 --> 00:30:22,880
per kilowatt hour.

692
00:30:25,480 --> 00:30:29,600
But to make it
location-blind, you

693
00:30:29,600 --> 00:30:32,252
put it into dollars per
watt, and dollars per watt

694
00:30:32,252 --> 00:30:33,460
is a function of many things.

695
00:30:33,460 --> 00:30:37,100
It's how much does it cost
to make the module itself?

696
00:30:37,100 --> 00:30:40,030
So how many dollars
per given area?

697
00:30:40,030 --> 00:30:41,400
You divide it by the insulation.

698
00:30:41,400 --> 00:30:42,900
For dollars per
watt peak, you'd set

699
00:30:42,900 --> 00:30:45,790
that to aim 1.5 or 1,000
kilowatts per meter squared.

700
00:30:45,790 --> 00:30:47,690
Multiply that times
your efficiency,

701
00:30:47,690 --> 00:30:51,190
and then by your yield, which
is a manufacturing parameter.

702
00:30:51,190 --> 00:30:54,980
So for some cells where if you
can reduce this number quite

703
00:30:54,980 --> 00:30:56,890
dramatically, so
for example, suppose

704
00:30:56,890 --> 00:31:00,100
you decide to go to very,
very thin wafers for silicon.

705
00:31:00,100 --> 00:31:03,690
You want to move to 20 micron
wafers, let's say, which

706
00:31:03,690 --> 00:31:07,220
is something that a lot of
companies now are working on.

707
00:31:07,220 --> 00:31:09,792
What happens is that those
become very, very fragile.

708
00:31:09,792 --> 00:31:12,250
And when you handle them and
using all these pick and place

709
00:31:12,250 --> 00:31:14,166
operations to move them
to different processes

710
00:31:14,166 --> 00:31:16,870
within the cell fab,
oftentimes they'll break.

711
00:31:16,870 --> 00:31:19,257
And so that yield parameter
can come very important

712
00:31:19,257 --> 00:31:21,340
and [INAUDIBLE] interplayed
with this number here,

713
00:31:21,340 --> 00:31:22,990
and that's why it's there.

714
00:31:22,990 --> 00:31:25,851
So anyhow, what this
graph is is this

715
00:31:25,851 --> 00:31:27,540
is the dollars
per meter squared,

716
00:31:27,540 --> 00:31:30,500
so the cost of
producing that cell.

717
00:31:30,500 --> 00:31:32,990
And then you have
this efficiency here,

718
00:31:32,990 --> 00:31:35,980
and each of these
lines represents

719
00:31:35,980 --> 00:31:38,300
a certain dollars per watt.

720
00:31:38,300 --> 00:31:41,800
And so I think the DOE has
wanted us to get to $1.00 per

721
00:31:41,800 --> 00:31:44,909
watt, and you can do that by
either producing incredibly

722
00:31:44,909 --> 00:31:46,950
efficient cells, around
50%, which seems a little

723
00:31:46,950 --> 00:31:48,990
unreasonable.

724
00:31:48,990 --> 00:31:51,450
And they can cost $500
per meter squared,

725
00:31:51,450 --> 00:31:55,406
and that would be
actually a cheap panel.

726
00:31:55,406 --> 00:31:57,780
Or the other idea is to go to
very, very low cost and low

727
00:31:57,780 --> 00:32:00,810
efficiency, and you can try
to hover around in here.

728
00:32:00,810 --> 00:32:03,752
And so this nomenclature
I'm about to use

729
00:32:03,752 --> 00:32:06,210
is a little outdated, but I
think still some people use it,

730
00:32:06,210 --> 00:32:08,820
although rather loosely.

731
00:32:08,820 --> 00:32:11,220
What's known as
first-generation-- and again,

732
00:32:11,220 --> 00:32:12,842
these terms I've
read papers now they

733
00:32:12,842 --> 00:32:14,550
claim that they're
third-generation where

734
00:32:14,550 --> 00:32:16,550
they used to be classified
as second-generation.

735
00:32:16,550 --> 00:32:19,300
It doesn't really matter,
but the ideas behind them

736
00:32:19,300 --> 00:32:20,550
are still relevant.

737
00:32:20,550 --> 00:32:22,550
So the first-generation
with the single bandgap,

738
00:32:22,550 --> 00:32:25,499
it's crystalline silicon still
[INAUDIBLE] the market leader,

739
00:32:25,499 --> 00:32:27,540
and they're on the order
of 15% to 20% efficient,

740
00:32:27,540 --> 00:32:29,580
but they're still
relatively high-cost.

741
00:32:29,580 --> 00:32:30,996
And there's actually
a lot of work

742
00:32:30,996 --> 00:32:33,980
now that says that this
can move in this direction,

743
00:32:33,980 --> 00:32:38,800
moving to thinner wafers,
increased laser processing.

744
00:32:38,800 --> 00:32:43,170
Just tighter manufacturing can
really bring that cost down.

745
00:32:43,170 --> 00:32:45,495
So your second-generation
are your thin film.

746
00:32:45,495 --> 00:32:50,120
So this is either CIGS,
amorphous silicon, organics.

747
00:32:50,120 --> 00:32:52,830
A lot of these very, very
cheap to deposit, cheap

748
00:32:52,830 --> 00:32:54,510
to manufacture,
but they generally

749
00:32:54,510 --> 00:32:58,210
suffer in efficiency, so they
hover around in this area.

750
00:32:58,210 --> 00:33:00,170
And so far, this
idea hasn't really

751
00:33:00,170 --> 00:33:02,170
fulfilled itself except
for maybe cad-tel,

752
00:33:02,170 --> 00:33:03,625
which is a thin film material.

753
00:33:03,625 --> 00:33:05,250
And then what we're
about to talk about

754
00:33:05,250 --> 00:33:07,340
is our third
generation of cells.

755
00:33:07,340 --> 00:33:11,640
So these are potentially
low-cost and potentially

756
00:33:11,640 --> 00:33:12,585
incredibly efficient.

757
00:33:15,090 --> 00:33:19,520
So we're going to talk about
only a few types today.

758
00:33:19,520 --> 00:33:20,700
I encourage you to look up.

759
00:33:20,700 --> 00:33:24,650
There's a whole different types
of solar cells you can look up,

760
00:33:24,650 --> 00:33:27,340
ideas, research being done.

761
00:33:27,340 --> 00:33:29,550
But I'm going talk--
this is what I work on,

762
00:33:29,550 --> 00:33:32,180
our intermediate-band
solar cells, specifically

763
00:33:32,180 --> 00:33:34,300
impurity-band photovoltaics.

764
00:33:34,300 --> 00:33:37,190
And I'm sure you're all
familiar with this picture,

765
00:33:37,190 --> 00:33:41,920
but if you want to create
free mobile charges

766
00:33:41,920 --> 00:33:43,530
within your
semiconductor, you have

767
00:33:43,530 --> 00:33:46,180
to shine light on
it that has energies

768
00:33:46,180 --> 00:33:48,230
greater than the bandgap.

769
00:33:48,230 --> 00:33:52,480
And so you lose these
low-energy photons.

770
00:33:52,480 --> 00:33:54,957
They're not collected,
and what's a way

771
00:33:54,957 --> 00:33:56,040
that we can collect these?

772
00:33:56,040 --> 00:34:01,350
Well, one idea this
guy had was if you

773
00:34:01,350 --> 00:34:04,120
can create a material that
has a density of states that

774
00:34:04,120 --> 00:34:06,790
looks like this, you can
create a stepping stone

775
00:34:06,790 --> 00:34:09,654
by putting a half-filled
band within your bandgap that

776
00:34:09,654 --> 00:34:12,070
allows you to promote to that
band and then from that band

777
00:34:12,070 --> 00:34:13,330
into the conduction band.

778
00:34:13,330 --> 00:34:15,500
So the idea is that your
voltage output can still

779
00:34:15,500 --> 00:34:18,625
be maintained for
the host material,

780
00:34:18,625 --> 00:34:20,250
but you can collect
that extra current.

781
00:34:20,250 --> 00:34:22,679
So your ISC will
increase substantially,

782
00:34:22,679 --> 00:34:23,860
and so that's the promise.

783
00:34:23,860 --> 00:34:26,360
And if you do some theoretical
calculations similar

784
00:34:26,360 --> 00:34:29,900
what Shockley and Queisser did,
assuming the radiative lifetime

785
00:34:29,900 --> 00:34:31,659
is your limiting
factor-- and again,

786
00:34:31,659 --> 00:34:34,560
that's certainly not the case
in some of the materials that

787
00:34:34,560 --> 00:34:37,520
are worked on for this
idea-- and the idea

788
00:34:37,520 --> 00:34:41,010
is you can actually get up
to around 63% efficient.

789
00:34:41,010 --> 00:34:44,194
And so what this graph is here,
here's your single gap limit.

790
00:34:44,194 --> 00:34:46,860
This is a calculation similar to
what Shockley and Queisser did.

791
00:34:46,860 --> 00:34:48,984
I think this is using
blackbody radiation, which is

792
00:34:48,984 --> 00:34:50,900
why the curves are so smooth.

793
00:34:50,900 --> 00:34:53,210
And it can actually
outperform a tandem cell,

794
00:34:53,210 --> 00:34:55,445
which is again a
multi-junction cell that

795
00:34:55,445 --> 00:34:58,410
are two cells stacked.

796
00:34:58,410 --> 00:35:01,090
And it turns out the
best is something

797
00:35:01,090 --> 00:35:03,810
that has a gap level
that's about 0.7 EV

798
00:35:03,810 --> 00:35:08,410
from the conduction band and
has a larger band gap, so

799
00:35:08,410 --> 00:35:11,520
that distance between
your valence band

800
00:35:11,520 --> 00:35:16,490
and your conduction band of
around 1.9 electron volts,

801
00:35:16,490 --> 00:35:18,070
and that's kind of
the ideal material

802
00:35:18,070 --> 00:35:21,320
for an intermediate-band
solar cell.

803
00:35:21,320 --> 00:35:23,480
So theoretically, this
idea is a great idea.

804
00:35:23,480 --> 00:35:26,930
Problem is, how do you
make a material that

805
00:35:26,930 --> 00:35:29,540
has this band structure?

806
00:35:29,540 --> 00:35:31,090
And that's really
the challenge where

807
00:35:31,090 --> 00:35:34,060
people are working right
now, and it's actually

808
00:35:34,060 --> 00:35:36,150
rather difficult. And so
there's three approaches.

809
00:35:36,150 --> 00:35:39,700
One is the impurity
band, which I personally

810
00:35:39,700 --> 00:35:42,890
like because it's a
much cheaper method.

811
00:35:42,890 --> 00:35:45,390
Other one is this band
anti-crossing idea

812
00:35:45,390 --> 00:35:47,980
where you can actually
split the conduction band,

813
00:35:47,980 --> 00:35:50,660
and it's generally done with
these mixed metal oxides

814
00:35:50,660 --> 00:35:52,160
and these highly
mismatched alloys,

815
00:35:52,160 --> 00:35:53,160
and they're really cool.

816
00:35:53,160 --> 00:35:56,100
And this is currently the
most successful material there

817
00:35:56,100 --> 00:35:57,570
for intermediate-band
solar cells.

818
00:35:57,570 --> 00:36:01,310
And then there's quantum dot
arrays and quantum structures

819
00:36:01,310 --> 00:36:04,930
that are also a possible
fabrication method.

820
00:36:04,930 --> 00:36:10,006
OK, so idea behind an
impurity photovoltaic

821
00:36:10,006 --> 00:36:12,130
is you start with let's
say a material like silicon

822
00:36:12,130 --> 00:36:14,100
or some other high-bandgap
gap semiconductor,

823
00:36:14,100 --> 00:36:19,340
you put in impurities that
have these deep-level states.

824
00:36:19,340 --> 00:36:20,520
Iron would be an example.

825
00:36:20,520 --> 00:36:22,270
So iron in very
low concentrations

826
00:36:22,270 --> 00:36:23,850
can be incredibly detrimental.

827
00:36:23,850 --> 00:36:25,467
But as you increase
that concentration

828
00:36:25,467 --> 00:36:27,050
above some critical
limit often called

829
00:36:27,050 --> 00:36:29,530
the [INAUDIBLE]
transition limit,

830
00:36:29,530 --> 00:36:34,280
you can actually form a
band within the bandgap.

831
00:36:34,280 --> 00:36:37,180
And so the idea what's going
on here is that each of atoms

832
00:36:37,180 --> 00:36:40,190
this has some
electron wave function

833
00:36:40,190 --> 00:36:41,477
and some radius that it sees.

834
00:36:41,477 --> 00:36:43,310
And then as you increase
that concentration,

835
00:36:43,310 --> 00:36:45,000
these wave functions
overlap and you

836
00:36:45,000 --> 00:36:48,960
can have conduction through
those mid-level states.

837
00:36:48,960 --> 00:36:51,880
So that's one
idea, and I like it

838
00:36:51,880 --> 00:36:54,680
because you use what I could
consider very dirty materials.

839
00:36:54,680 --> 00:36:56,790
The concentrations
you can put in here

840
00:36:56,790 --> 00:37:01,390
around are one atomic percent,
so your silicon's only 99.9%

841
00:37:01,390 --> 00:37:02,340
pure at this point.

842
00:37:06,186 --> 00:37:09,810
Other idea is to use either
quantum dots or these quantum

843
00:37:09,810 --> 00:37:10,310
wells.

844
00:37:10,310 --> 00:37:13,760
So I think the idea here
is that these are localized

845
00:37:13,760 --> 00:37:15,290
and then this
becomes delocalized.

846
00:37:15,290 --> 00:37:19,895
So as you bring these quantum
well structures or quantum dots

847
00:37:19,895 --> 00:37:21,270
closer together,
you can actually

848
00:37:21,270 --> 00:37:23,390
get tunneling
through these states,

849
00:37:23,390 --> 00:37:25,150
and it basically
essentially forms

850
00:37:25,150 --> 00:37:27,740
a band within your bandgap.

851
00:37:27,740 --> 00:37:29,242
Similar idea is to
do this locally--

852
00:37:29,242 --> 00:37:30,700
is that you have
one photon promote

853
00:37:30,700 --> 00:37:33,490
to one of these confined states
in your quantum structure,

854
00:37:33,490 --> 00:37:35,700
and then another photon
to keep promoting up

855
00:37:35,700 --> 00:37:40,004
into your conduction
band to create carriers.

856
00:37:40,004 --> 00:37:42,236
AUDIENCE: [INAUDIBLE]
would there

857
00:37:42,236 --> 00:37:44,468
be quantum [INAUDIBLE]
embedded in the silicon,

858
00:37:44,468 --> 00:37:47,795
or these are solar cells
made out of [INAUDIBLE]?

859
00:37:47,795 --> 00:37:50,420
PROFESSOR: It's usually stacked
between two other semiconductor

860
00:37:50,420 --> 00:37:50,920
materials.

861
00:37:50,920 --> 00:37:54,000
So you have semiconductor,
quantum dot, semiconductor.

862
00:37:54,000 --> 00:37:57,560
If you look up this paper
among-- if you look up

863
00:37:57,560 --> 00:38:01,940
anything-- so Antonio
Luque-- sorry, yeah.

864
00:38:01,940 --> 00:38:04,470
Luque and Marti are
two people in Spain

865
00:38:04,470 --> 00:38:07,249
who've kind of promoted a lot
of the theory around this.

866
00:38:07,249 --> 00:38:08,790
And look up their
quantum dot papers,

867
00:38:08,790 --> 00:38:10,430
and they'll really go
into the device structure

868
00:38:10,430 --> 00:38:11,110
if you're curious.

869
00:38:11,110 --> 00:38:12,410
I'd recommend doing that, Joel.

870
00:38:12,410 --> 00:38:16,206
AUDIENCE: So the x-axis
on these figures are--

871
00:38:16,206 --> 00:38:18,070
PROFESSOR: Real space.

872
00:38:18,070 --> 00:38:21,240
AUDIENCE: So there will be
some portions of the device

873
00:38:21,240 --> 00:38:26,348
where this is possible in--
it kind of depends upon

874
00:38:26,348 --> 00:38:29,059
how many quantum
wells you actually

875
00:38:29,059 --> 00:38:30,058
put into your structure.

876
00:38:30,058 --> 00:38:32,224
PROFESSOR: Right, and that's
one of limiting things,

877
00:38:32,224 --> 00:38:34,640
is the EQE of these devices
because they can't grow

878
00:38:34,640 --> 00:38:37,157
so many are pretty limited.

879
00:38:37,157 --> 00:38:38,740
And again, these
also provide pathways

880
00:38:38,740 --> 00:38:40,114
for traps and
other things, which

881
00:38:40,114 --> 00:38:42,560
is again a huge problem
for these types of devices

882
00:38:42,560 --> 00:38:43,241
and ideas.

883
00:38:43,241 --> 00:38:44,740
And people are
working on that quite

884
00:38:44,740 --> 00:38:47,406
extensively from the theory side
and from the experimental side,

885
00:38:47,406 --> 00:38:49,865
and trying to merge the two.

886
00:38:49,865 --> 00:38:50,365
OK.

887
00:38:52,910 --> 00:38:53,859
Sorry, Ben?

888
00:38:53,859 --> 00:38:55,442
AUDIENCE: The quantum
well [INAUDIBLE]

889
00:38:55,442 --> 00:38:57,150
on the right, where
[INAUDIBLE] functions

890
00:38:57,150 --> 00:39:01,600
don't overlap, do those act
as recombination centers?

891
00:39:01,600 --> 00:39:04,640
PROFESSOR: It seems to me
that they would act as traps.

892
00:39:04,640 --> 00:39:08,210
Again, I'm not an expert on
these types of materials,

893
00:39:08,210 --> 00:39:10,940
but that would be
my inclination.

894
00:39:10,940 --> 00:39:13,360
I guess the idea is that you
have a very, very short well

895
00:39:13,360 --> 00:39:14,818
and then you have
a field across it

896
00:39:14,818 --> 00:39:17,720
so there's not something
else to hop to.

897
00:39:17,720 --> 00:39:20,061
And it just pulls it
right out, so that

898
00:39:20,061 --> 00:39:21,310
would be one device structure.

899
00:39:21,310 --> 00:39:24,830
Again, these are generally
just like a few layers

900
00:39:24,830 --> 00:39:25,740
of quantum dots.

901
00:39:32,390 --> 00:39:34,135
Oh, I didn't put
up the [INAUDIBLE].

902
00:39:34,135 --> 00:39:35,760
So this is the
band-anticrossing model.

903
00:39:35,760 --> 00:39:38,720
So this is they're using
these quaternary alloys.

904
00:39:38,720 --> 00:39:40,979
So these are, again,
very, very hard to grow,

905
00:39:40,979 --> 00:39:42,770
and they've done these
awesome measurements

906
00:39:42,770 --> 00:39:45,700
where they can show your
valence band, conduction

907
00:39:45,700 --> 00:39:47,450
band, and your impurity
bands, and they're

908
00:39:47,450 --> 00:39:51,960
using a technique called
photomoduled reflectance,

909
00:39:51,960 --> 00:39:55,860
so basically the reflectance
under oscillating fields.

910
00:39:55,860 --> 00:39:57,710
And from that, you
can get these kind

911
00:39:57,710 --> 00:40:00,506
of resonance points between
the different transitions

912
00:40:00,506 --> 00:40:02,630
and you can figure out
where your energy state lie.

913
00:40:02,630 --> 00:40:04,430
It's a really cool technique.

914
00:40:04,430 --> 00:40:06,520
But what they did is
they've done some very, very

915
00:40:06,520 --> 00:40:10,276
careful engineering and grown
these very, very carefully.

916
00:40:10,276 --> 00:40:11,650
And they can
actually demonstrate

917
00:40:11,650 --> 00:40:15,130
the sub-bandgap response
where if they have,

918
00:40:15,130 --> 00:40:18,420
let's say a photon that's
able to drive this transition

919
00:40:18,420 --> 00:40:20,890
but it's too low in energy
to drive that transition,

920
00:40:20,890 --> 00:40:22,300
they will get no current out.

921
00:40:22,300 --> 00:40:27,060
And then as soon as
they add the higher one,

922
00:40:27,060 --> 00:40:29,040
the current will increase.

923
00:40:29,040 --> 00:40:32,560
And then if they turn off
that low-energy photon source,

924
00:40:32,560 --> 00:40:35,306
the current
decreases but doesn't

925
00:40:35,306 --> 00:40:37,180
stop because the other
high-energy photon can

926
00:40:37,180 --> 00:40:38,980
do both promotions.

927
00:40:38,980 --> 00:40:41,460
So this is actually one of
the only successful devices

928
00:40:41,460 --> 00:40:43,280
of impurity-band
solar cells, and it

929
00:40:43,280 --> 00:40:47,480
was quite an awesome feat.

930
00:40:47,480 --> 00:40:52,340
So I think one of
the few ones we'll

931
00:40:52,340 --> 00:40:55,740
talk about for another advanced
concept is hot carrier cells.

932
00:40:55,740 --> 00:40:59,420
This is something I think
Martin Green pioneered,

933
00:40:59,420 --> 00:41:01,810
and I believe it
still working on.

934
00:41:01,810 --> 00:41:03,570
I haven't heard much
about them recently,

935
00:41:03,570 --> 00:41:06,180
so I don't know what
the progress is there.

936
00:41:06,180 --> 00:41:08,580
So the idea is that one
of the biggest losses,

937
00:41:08,580 --> 00:41:11,030
as I'm sure you guys know
from I think homework 1 or 2,

938
00:41:11,030 --> 00:41:13,200
is thermalization
in your solar cells.

939
00:41:13,200 --> 00:41:15,890
When you promote an electron
well up into the conduction

940
00:41:15,890 --> 00:41:21,734
band, it then gives
off heat or phonons,

941
00:41:21,734 --> 00:41:23,400
which are just lattice
vibrations, which

942
00:41:23,400 --> 00:41:26,140
is another way of
thinking about heat.

943
00:41:26,140 --> 00:41:28,490
In that process, the problem
is that's incredibly fast.

944
00:41:28,490 --> 00:41:29,990
If you look at the
time scales here,

945
00:41:29,990 --> 00:41:32,240
so you come and you promote
this electron way up

946
00:41:32,240 --> 00:41:34,780
into your conduction band.

947
00:41:34,780 --> 00:41:37,090
This is what it looks like
prior at that promotion,

948
00:41:37,090 --> 00:41:40,450
and then it slowly decays,
and that decay happens over

949
00:41:40,450 --> 00:41:42,360
about a picosecond.

950
00:41:42,360 --> 00:41:45,722
So if you want to move your
carrier in a picosecond,

951
00:41:45,722 --> 00:41:47,180
if you know what
your field is, you

952
00:41:47,180 --> 00:41:49,880
can see how long that length
is, and that length scale

953
00:41:49,880 --> 00:41:51,700
is also incredibly short.

954
00:41:51,700 --> 00:41:57,220
And so one of the ideas
is, OK, how either we

955
00:41:57,220 --> 00:42:00,110
can decrease our path
length, or we can somehow

956
00:42:00,110 --> 00:42:03,390
slow down thermalization,
and that's actually one idea.

957
00:42:03,390 --> 00:42:06,840
And to slow down
thermalization is something

958
00:42:06,840 --> 00:42:10,360
that they call carrier cooling.

959
00:42:10,360 --> 00:42:13,990
And basically making certain
types of device structures

960
00:42:13,990 --> 00:42:17,110
or material structures, you can
inhibit certain phonon modes.

961
00:42:17,110 --> 00:42:19,240
So when these
electrons get promoted,

962
00:42:19,240 --> 00:42:20,410
they want to give off heat.

963
00:42:20,410 --> 00:42:22,201
They basically give
off lattice vibrations,

964
00:42:22,201 --> 00:42:24,280
so they want to shake
the atoms around them

965
00:42:24,280 --> 00:42:28,430
and distribute that motion.

966
00:42:28,430 --> 00:42:30,410
If you prevent that
from happening somehow,

967
00:42:30,410 --> 00:42:33,840
through certain
types of structures,

968
00:42:33,840 --> 00:42:37,666
you can allow that
process to go slower,

969
00:42:37,666 --> 00:42:40,220
and that's one of the ideas.

970
00:42:40,220 --> 00:42:42,430
Extracting the carriers--
also kind of difficult.

971
00:42:42,430 --> 00:42:46,160
You need to have contacts
that are selective that

972
00:42:46,160 --> 00:42:49,070
can take the hot carriers at
all the different energies

973
00:42:49,070 --> 00:42:51,070
that they're promoted
at, and that's

974
00:42:51,070 --> 00:42:52,740
also a really difficult idea.

975
00:42:52,740 --> 00:42:54,823
And some of the ideas that
they've been working on

976
00:42:54,823 --> 00:42:57,510
are these resonant
tunneling contacts.

977
00:42:57,510 --> 00:43:00,780
So again, this is
kind of hairy stuff.

978
00:43:00,780 --> 00:43:02,284
I think these slides
are rather old

979
00:43:02,284 --> 00:43:04,200
and I don't know what's
come of this research.

980
00:43:04,200 --> 00:43:08,200
So it's a really exciting
idea and I encourage

981
00:43:08,200 --> 00:43:11,840
you guys to look more into it.

982
00:43:11,840 --> 00:43:13,840
OK, so now we're going
to move a little bit away

983
00:43:13,840 --> 00:43:17,739
from advanced concepts and talk
about kind of bulk thin film

984
00:43:17,739 --> 00:43:19,280
materials, and I
think we're actually

985
00:43:19,280 --> 00:43:23,780
going to end quite
early, which is fine.

986
00:43:23,780 --> 00:43:30,280
So the most common
commercially-available type

987
00:43:30,280 --> 00:43:32,812
of material are these
wafer-based materials,

988
00:43:32,812 --> 00:43:34,270
so monocrystalline,
which you know,

989
00:43:34,270 --> 00:43:35,769
which is what our
cells are made of.

990
00:43:35,769 --> 00:43:37,760
This CZ, or Czochralski
growth, probably

991
00:43:37,760 --> 00:43:39,200
pronouncing that wrong.

992
00:43:39,200 --> 00:43:41,616
Silicon,
multichrystalline silicon,

993
00:43:41,616 --> 00:43:43,990
which I think Tonio was showing
you what that looks like.

994
00:43:43,990 --> 00:43:45,910
You can see the
different grains.

995
00:43:45,910 --> 00:43:47,320
They're actually quite pretty.

996
00:43:47,320 --> 00:43:49,820
Ribbon silicon, which was
pioneered by Evergreen,

997
00:43:49,820 --> 00:43:53,180
and I think now that
technology's gone.

998
00:43:53,180 --> 00:43:55,680
Evergreen went out of
business, so I don't think

999
00:43:55,680 --> 00:43:58,090
that technology's still around.

1000
00:43:58,090 --> 00:44:00,610
The [INAUDIBLE] still
might be making modules.

1001
00:44:00,610 --> 00:44:04,264
Anyhow, so thin films.

1002
00:44:04,264 --> 00:44:06,055
Cad-tel is still
currently-- or First Solar

1003
00:44:06,055 --> 00:44:08,740
is still one of the
cheapest module makers.

1004
00:44:08,740 --> 00:44:12,096
Their process was so cheap
that their efficiency really

1005
00:44:12,096 --> 00:44:13,470
suffered because
they didn't even

1006
00:44:13,470 --> 00:44:15,840
have an ARC coating for
the first 5 or 10 years

1007
00:44:15,840 --> 00:44:16,700
of development.

1008
00:44:16,700 --> 00:44:18,908
And they said we didn't need
it, and they were right.

1009
00:44:18,908 --> 00:44:20,890
They were still
outperforming silicon people.

1010
00:44:20,890 --> 00:44:23,389
That's becoming less so the
case and one of the big problems

1011
00:44:23,389 --> 00:44:26,840
with Cad-tel, as you guys
know, is the [INAUDIBLE]

1012
00:44:26,840 --> 00:44:31,010
and also the toxicity,
which some people are

1013
00:44:31,010 --> 00:44:33,931
really concerned about.

1014
00:44:33,931 --> 00:44:34,430
Yeah?

1015
00:44:34,430 --> 00:44:35,680
AUDIENCE: Just a note on that.

1016
00:44:35,680 --> 00:44:37,470
So if we're debating
in this class

1017
00:44:37,470 --> 00:44:40,840
whether the should be allowed
to be imported into Japan.

1018
00:44:40,840 --> 00:44:42,690
And my understanding,
from the reading,

1019
00:44:42,690 --> 00:44:46,070
was that Japan actually does
not allow cadmium telluride

1020
00:44:46,070 --> 00:44:46,850
imports.

1021
00:44:46,850 --> 00:44:48,910
Is that correct?

1022
00:44:48,910 --> 00:44:51,260
AUDIENCE: I guess so, then yes.

1023
00:44:51,260 --> 00:44:52,970
I don't actually know.

1024
00:44:52,970 --> 00:44:55,990
I do know that Cad-tel
is a very stable compound

1025
00:44:55,990 --> 00:44:57,826
and I know people
who work with it here

1026
00:44:57,826 --> 00:44:59,450
and they actually
work on recycling it,

1027
00:44:59,450 --> 00:45:03,820
so how to dissolve it and
separate the two elements.

1028
00:45:03,820 --> 00:45:06,750
And they're not worried
about the hazards

1029
00:45:06,750 --> 00:45:09,650
of the actual raw
material because it's

1030
00:45:09,650 --> 00:45:12,610
a pretty stable compound, but a
lot of people do worry about it

1031
00:45:12,610 --> 00:45:14,530
and it's certainly
a valid concern.

1032
00:45:15,640 --> 00:45:19,591
But one of the thin film
replacements, well amorphic

1033
00:45:19,591 --> 00:45:21,340
silicon-- you think
you guys know that is,

1034
00:45:21,340 --> 00:45:23,380
but that's deposited silicon.

1035
00:45:23,380 --> 00:45:27,110
It has no real like lattice
structure to speak of.

1036
00:45:27,110 --> 00:45:29,029
It's kind of a disordered
mess, which means

1037
00:45:29,029 --> 00:45:30,320
it has a lot of dangling bonds.

1038
00:45:30,320 --> 00:45:36,430
It has very, very low
lifetimes as a result,

1039
00:45:36,430 --> 00:45:39,680
and also in [INAUDIBLE]
low mobilities.

1040
00:45:39,680 --> 00:45:42,450
You also deposit it and
you deposit with hydrogen

1041
00:45:42,450 --> 00:45:46,261
to passivate all of
those dangling bonds,

1042
00:45:46,261 --> 00:45:48,760
and that's usually done with
either plasma-enhanced chemical

1043
00:45:48,760 --> 00:45:51,990
vapor deposition, and it's
done on either metal or glass.

1044
00:45:51,990 --> 00:45:56,120
And so it has the
potential to be very cheap.

1045
00:45:56,120 --> 00:45:58,300
And then one of
things I think was

1046
00:45:58,300 --> 00:46:01,570
aimed to replace Cad-tel
was CIGS-- so copper,

1047
00:46:01,570 --> 00:46:03,970
indium, gallium, diselenide.

1048
00:46:03,970 --> 00:46:07,720
I'll get to some of the
problems with CIGS in a second,

1049
00:46:07,720 --> 00:46:09,340
but there are few
startups around it.

1050
00:46:09,340 --> 00:46:13,784
I think Nanosolar
is one, HelioVolt,

1051
00:46:13,784 --> 00:46:16,200
and I think Solyndra was one,
although I don't really want

1052
00:46:16,200 --> 00:46:17,638
to say their name out loud.

1053
00:46:21,510 --> 00:46:23,180
So problem with
CIGS is if you're

1054
00:46:23,180 --> 00:46:25,280
looking for
Earth-abundant films,

1055
00:46:25,280 --> 00:46:28,010
again if we want to scale-- if
we want solar energy to scale

1056
00:46:28,010 --> 00:46:30,530
to terawatt levels, if
we want it to provide all

1057
00:46:30,530 --> 00:46:32,450
of human energy
needs-- then we're

1058
00:46:32,450 --> 00:46:34,360
going to need to use
elements and materials

1059
00:46:34,360 --> 00:46:37,430
that are Earth-abundant, cheap
to find, cheap to produce,

1060
00:46:37,430 --> 00:46:39,660
and CIGS isn't going
to get us there.

1061
00:46:39,660 --> 00:46:41,340
It's got indium and gallium.

1062
00:46:41,340 --> 00:46:43,182
Indium, which is highly
used and the price

1063
00:46:43,182 --> 00:46:44,640
has skyrocketed
now because I think

1064
00:46:44,640 --> 00:46:48,910
it's used in your
displays and televisions.

1065
00:46:48,910 --> 00:46:50,148
There's indium in here.

1066
00:46:50,148 --> 00:46:51,064
AUDIENCE: [INAUDIBLE].

1067
00:46:53,057 --> 00:46:55,140
PROFESSOR: OK, so that's
where it comes from then.

1068
00:46:55,140 --> 00:46:56,110
Thank you.

1069
00:46:56,110 --> 00:46:58,780
So what people have
been working or trying

1070
00:46:58,780 --> 00:47:03,320
to look at replacements--
so this is CZTS.

1071
00:47:03,320 --> 00:47:07,740
So it's copper, zinc, tin,
it's sometimes sulfide.

1072
00:47:07,740 --> 00:47:09,700
This is selenide,
but they are also

1073
00:47:09,700 --> 00:47:12,240
working on replacing the
selenium with sulfur, something

1074
00:47:12,240 --> 00:47:14,110
a little more Earth-abundant.

1075
00:47:14,110 --> 00:47:19,410
And this was done at IBM and
it was pretty remarkable,

1076
00:47:19,410 --> 00:47:22,260
actually, because I think
within a very, very short period

1077
00:47:22,260 --> 00:47:24,300
of time, they were able
to get efficiencies

1078
00:47:24,300 --> 00:47:28,240
around 9.6%, which for
a fledgling material

1079
00:47:28,240 --> 00:47:31,610
is incredible, and
I'll get to why that

1080
00:47:31,610 --> 00:47:33,987
was so incredible in a second.

1081
00:47:33,987 --> 00:47:36,320
But again, when thinking about
Earth-abundant materials,

1082
00:47:36,320 --> 00:47:38,153
I really recommend you
guys read this paper.

1083
00:47:38,153 --> 00:47:40,200
This is a paper by Cyrus Wadia.

1084
00:47:40,200 --> 00:47:42,220
It's "Environmental
Science and Technology"

1085
00:47:42,220 --> 00:47:44,261
and it looks all of these
different semiconductor

1086
00:47:44,261 --> 00:47:46,300
materials and it
looks at in yellow,

1087
00:47:46,300 --> 00:47:48,800
this is what you
could get if you

1088
00:47:48,800 --> 00:47:54,010
look at annual production of
that semiconductor material

1089
00:47:54,010 --> 00:47:56,360
now in terms of what is
being produced to produce

1090
00:47:56,360 --> 00:47:58,140
the raw materials that make it.

1091
00:47:58,140 --> 00:48:00,610
And then the known
economic reserves,

1092
00:48:00,610 --> 00:48:03,840
what can we mine and
get economically today,

1093
00:48:03,840 --> 00:48:05,220
how much could it produce?

1094
00:48:05,220 --> 00:48:07,990
And your worldwide
consumption is on this line,

1095
00:48:07,990 --> 00:48:11,280
and you can see Cad-tel is
just barely eking it out.

1096
00:48:11,280 --> 00:48:13,590
And the ones of note
for this discussion

1097
00:48:13,590 --> 00:48:16,320
are Cad-tel, CIGS, which
is slightly better,

1098
00:48:16,320 --> 00:48:19,430
and then CZTS, which is
rather plentiful even

1099
00:48:19,430 --> 00:48:23,057
with reserves that are
currently being mined now,

1100
00:48:23,057 --> 00:48:24,890
and so that's kind of
the take-home message.

1101
00:48:24,890 --> 00:48:27,200
The other thing also
really interesting

1102
00:48:27,200 --> 00:48:29,620
is the actual raw material cost.

1103
00:48:29,620 --> 00:48:33,200
Cad-tel, turns out,
it's not that great

1104
00:48:33,200 --> 00:48:36,860
when you compare it
to CIGS and then CZTS.

1105
00:48:36,860 --> 00:48:39,360
It's a whole lot cheaper.

1106
00:48:39,360 --> 00:48:44,180
So one of the other things
about CZTS that's so cool

1107
00:48:44,180 --> 00:48:47,030
is that there's basically
this huge growth parameter

1108
00:48:47,030 --> 00:48:49,300
space for making these
quaternary alloys,

1109
00:48:49,300 --> 00:48:52,910
and IBM on their first shot
just kind of nailed it.

1110
00:48:52,910 --> 00:48:57,090
So they didn't really look at
this whole parameter space.

1111
00:48:57,090 --> 00:49:00,510
It was just kind of their
first shot was right there,

1112
00:49:00,510 --> 00:49:03,160
and so it might already be
optimized, and a lot of people

1113
00:49:03,160 --> 00:49:05,316
are kind of
concerned about that.

1114
00:49:05,316 --> 00:49:07,440
But I think it's certainly
a very interesting field

1115
00:49:07,440 --> 00:49:11,744
and really cool to
work in, so I think

1116
00:49:11,744 --> 00:49:12,910
hopefully there's hope here.

1117
00:49:12,910 --> 00:49:15,240
So actually we're
ending way early,

1118
00:49:15,240 --> 00:49:17,040
but that concludes
the lecture today.

1119
00:49:17,040 --> 00:49:19,206
If you guys want to stick
around for more questions,

1120
00:49:19,206 --> 00:49:22,010
that's totally fine, but that's
all I had prepared for today.

1121
00:49:22,010 --> 00:49:23,980
So thank you.