WEBVTT

00:00:00.060 --> 00:00:02.500
The following content is
provided under a Creative

00:00:02.500 --> 00:00:04.010
Commons license.

00:00:04.010 --> 00:00:06.350
Your support will help
MIT OpenCourseWare

00:00:06.350 --> 00:00:10.720
continue to offer high quality
educational resources for free.

00:00:10.720 --> 00:00:13.340
To make a donation or
view additional materials

00:00:13.340 --> 00:00:17.209
from hundreds of MIT courses,
visit MIT OpenCourseWare

00:00:17.209 --> 00:00:17.834
at ocw.mit.edu.

00:00:24.967 --> 00:00:26.550
PROFESSOR: But today
we're going to be

00:00:26.550 --> 00:00:28.500
talking about crystalline
silicon solar cells.

00:00:28.500 --> 00:00:32.460
Now, for those of you who do not
work in crystalline silicon PV,

00:00:32.460 --> 00:00:33.970
the reason this
topic is important

00:00:33.970 --> 00:00:37.010
is because crystalline
silicon comprises about 90%

00:00:37.010 --> 00:00:39.110
of all solar cells
manufactured today.

00:00:39.110 --> 00:00:41.195
It's the dominant technology,
and the technologies

00:00:41.195 --> 00:00:43.070
that you're working on
are going to displace,

00:00:43.070 --> 00:00:45.110
or are aiming to displace
crystalline silicon,

00:00:45.110 --> 00:00:46.829
so it's good to know your enemy.

00:00:46.829 --> 00:00:48.870
For those who are working
on crystalline silicon,

00:00:48.870 --> 00:00:51.017
this is meant to be
a background of all

00:00:51.017 --> 00:00:52.850
of the different aspects--
the entire supply

00:00:52.850 --> 00:00:54.600
chain of crystalline
silicon-- so that you

00:00:54.600 --> 00:00:56.840
gain insight into the areas
that you're not currently

00:00:56.840 --> 00:00:57.890
focused on.

00:00:57.890 --> 00:01:00.880
You're getting a perspective
of the bigger picture.

00:01:00.880 --> 00:01:04.989
Crystalline silicon PV has
been around since 1954.

00:01:04.989 --> 00:01:08.160
The original-- well, in
its current incarnation.

00:01:08.160 --> 00:01:11.070
That was when Bell
Laboratories announced

00:01:11.070 --> 00:01:14.690
the development of the modern
crystalline silicon PV cell,

00:01:14.690 --> 00:01:17.810
and that was 6%
efficiency in 1954,

00:01:17.810 --> 00:01:19.490
published in general
applied physics,

00:01:19.490 --> 00:01:23.460
and the cell architecture,
it's obviously

00:01:23.460 --> 00:01:24.940
evolved over the
years but it's not

00:01:24.940 --> 00:01:27.870
entirely dissimilar
from what we have today

00:01:27.870 --> 00:01:32.740
as a cell architecture
for our modern PV cells.

00:01:32.740 --> 00:01:35.160
So, over the course
of-- it's almost

00:01:35.160 --> 00:01:37.680
been 60 years of development
of crystalline silicon

00:01:37.680 --> 00:01:39.660
photovoltaic technology.

00:01:39.660 --> 00:01:42.060
That means both the cell
itself, the materials

00:01:42.060 --> 00:01:44.570
that go into it, and also the
manufacturing, or the methods

00:01:44.570 --> 00:01:46.995
to produce said
materials and device,

00:01:46.995 --> 00:01:48.370
over the course
of those, almost,

00:01:48.370 --> 00:01:51.010
60 years much
innovation has happened

00:01:51.010 --> 00:01:54.780
both in terms of
manufacturing and technology.

00:01:54.780 --> 00:01:59.100
So today, we'll be going
over kind of a status quo

00:01:59.100 --> 00:02:02.980
snapshot of where crystalline
silicon stands and we brought

00:02:02.980 --> 00:02:05.210
in a number of
show and tell items

00:02:05.210 --> 00:02:09.660
so that you can see as we talk.

00:02:12.790 --> 00:02:14.780
So just for the
show and tell, we're

00:02:14.780 --> 00:02:17.340
going to be moving from the
feedstock materials over here

00:02:17.340 --> 00:02:20.615
finally into wafers
and cells on that side.

00:02:20.615 --> 00:02:21.350
All right.

00:02:23.890 --> 00:02:27.290
So, these lecture
notes are going

00:02:27.290 --> 00:02:29.090
to be valid for both 10 and 11.

00:02:29.090 --> 00:02:32.140
We're going to split this up
over two classes to really

00:02:32.140 --> 00:02:34.364
dive into some of the details.

00:02:34.364 --> 00:02:35.780
The first question
is why silicon?

00:02:35.780 --> 00:02:38.989
Why did silicon evolve
as what is currently

00:02:38.989 --> 00:02:40.780
the dominant technology,
which is currently

00:02:40.780 --> 00:02:44.340
90 percent of the PV market,
and I think it boils down

00:02:44.340 --> 00:02:47.770
to a couple of reasons.

00:02:47.770 --> 00:02:49.750
One is scalability.

00:02:49.750 --> 00:02:53.680
If you look at the elemental
abundance, on the vertical axis

00:02:53.680 --> 00:02:56.210
it's abundance, atoms
of the element per 10

00:02:56.210 --> 00:02:58.955
to the 16 atoms of silicon.

00:02:58.955 --> 00:03:01.080
The reason that everything
is normalized to silicon

00:03:01.080 --> 00:03:04.565
is because there is,
well, quite a lot of it

00:03:04.565 --> 00:03:05.440
in the earth's crust.

00:03:05.440 --> 00:03:08.620
As you can see, it's the
second most abundant element

00:03:08.620 --> 00:03:10.550
on the Earth's crust.

00:03:10.550 --> 00:03:14.260
It just so happens that, out of
all the stardust that is here

00:03:14.260 --> 00:03:17.630
on the planet, we have a
high percentage of silicon

00:03:17.630 --> 00:03:19.710
like the moon and like
many other planets

00:03:19.710 --> 00:03:23.250
in our solar system--
at least the hard ones.

00:03:23.250 --> 00:03:24.817
You can see oxygen
is probably the,

00:03:24.817 --> 00:03:27.400
well, oxygen is the only element
with higher natural abundance

00:03:27.400 --> 00:03:30.660
in the earth's crust, the
upper crust, than silicon

00:03:30.660 --> 00:03:32.820
and we go down as
we go to higher

00:03:32.820 --> 00:03:34.330
and higher atomic number.

00:03:34.330 --> 00:03:38.280
The probability of formation due
to subsequent fusion reactions

00:03:38.280 --> 00:03:40.560
in stars decreases
and, hence, it

00:03:40.560 --> 00:03:43.080
follows this almost a
power law distribution

00:03:43.080 --> 00:03:44.530
as you can see there.

00:03:44.530 --> 00:03:46.747
So it's scalable.

00:03:46.747 --> 00:03:48.830
It is present in the Earth
in high enough capacity

00:03:48.830 --> 00:03:50.210
to reach terawatt scales.

00:03:50.210 --> 00:03:53.890
It's nontoxic and, as
Don Sadoway likes to say,

00:03:53.890 --> 00:03:55.700
if you want
batteries dirt-cheap,

00:03:55.700 --> 00:03:57.400
you have to make
them out of dirt.

00:03:57.400 --> 00:03:59.730
A similar expression is used
in the crystalline silicon

00:03:59.730 --> 00:04:00.650
community.

00:04:00.650 --> 00:04:03.630
I believe the quote
in 1366 is, "It's not

00:04:03.630 --> 00:04:06.770
only good for the planet,
it is the planet."

00:04:06.770 --> 00:04:09.250
A variety of riffs off
of this particular chart

00:04:09.250 --> 00:04:11.550
right here, but from a
technological point of view,

00:04:11.550 --> 00:04:14.530
why did silicon evolve to
the point where it is today?

00:04:14.530 --> 00:04:17.589
It forms a very
tenacious surface oxide.

00:04:17.589 --> 00:04:21.399
So, if you were to expose a
piece of pure silicon to air,

00:04:21.399 --> 00:04:26.190
the surface oxide that forms
is very, very strong and very

00:04:26.190 --> 00:04:28.410
resistant, and very dense.

00:04:28.410 --> 00:04:30.920
So, unlike some
materials that corrode

00:04:30.920 --> 00:04:33.240
when exposed to
atmosphere, silicon

00:04:33.240 --> 00:04:37.180
oxidizes maybe the first few 10s
of angstroms, 100 of angstroms,

00:04:37.180 --> 00:04:42.530
and then it peters out so it's
diffusion-limited oxide growth

00:04:42.530 --> 00:04:44.470
mechanism that
eventually stabilizes

00:04:44.470 --> 00:04:46.760
at a very thin but very
dense and very protective

00:04:46.760 --> 00:04:48.000
oxide layer.

00:04:48.000 --> 00:04:51.340
So the risk of having a
silicon wafer degrade inside

00:04:51.340 --> 00:04:52.901
of a solar module is very low.

00:04:52.901 --> 00:04:55.150
Furthermore, that oxide layer
from an electrical point

00:04:55.150 --> 00:04:56.820
of view it's very passivating.

00:04:56.820 --> 00:04:58.710
So as we studied
on, as we solved

00:04:58.710 --> 00:05:01.800
in the exam, those interface
states or those surface states,

00:05:01.800 --> 00:05:03.450
the surface of
semiconductor, those

00:05:03.450 --> 00:05:05.480
can be reduced or
minimized by the presence

00:05:05.480 --> 00:05:07.600
of certain passivating
layers, and it just so

00:05:07.600 --> 00:05:11.610
happens that by the
benevolence of nature,

00:05:11.610 --> 00:05:15.310
the silicon oxide, which is
shown in these red triangles

00:05:15.310 --> 00:05:19.980
right here, has a very
low surface recombination

00:05:19.980 --> 00:05:22.560
velocity, passivates
a surface very well,

00:05:22.560 --> 00:05:24.880
and results in
high-performing devices.

00:05:24.880 --> 00:05:27.570
In this particular case, they're
plotting emitter saturation

00:05:27.570 --> 00:05:30.150
current density in
femtoamps per centimeter

00:05:30.150 --> 00:05:33.290
squared-- this is very, very
low-- versus sheet resistance.

00:05:33.290 --> 00:05:36.030
This is essentially the dopant
concentration in the emitter,

00:05:36.030 --> 00:05:38.830
so they're looking at how the
passivation quality changes

00:05:38.830 --> 00:05:41.560
as a function of dopant
density and silicon oxide

00:05:41.560 --> 00:05:44.340
works pretty well, and it's an
effective diffusion barrier.

00:05:44.340 --> 00:05:46.020
And, probably most
significantly, those

00:05:46.020 --> 00:05:49.220
are maybe one looking
forward rationale

00:05:49.220 --> 00:05:52.940
one technological or
scientific rationale

00:05:52.940 --> 00:05:54.890
and as far as the
field is concerned,

00:05:54.890 --> 00:05:57.320
as far as engineering
community is concerned,

00:05:57.320 --> 00:05:59.040
silicon has a lot of momentum.

00:05:59.040 --> 00:06:02.290
It's the most common
semiconductor material, silicon

00:06:02.290 --> 00:06:05.280
and germanium were both
purified, more or less,

00:06:05.280 --> 00:06:08.930
around the same decades but,
because silicon has a wider

00:06:08.930 --> 00:06:12.210
band gap, you have a lower
thermal carrier concentration,

00:06:12.210 --> 00:06:13.800
lower intrinsic
carrier concentration,

00:06:13.800 --> 00:06:15.920
folks were able to make
transistors and devices

00:06:15.920 --> 00:06:19.130
with lower noise out of
silicon as opposed to germanium

00:06:19.130 --> 00:06:22.230
and silicon technology
really took off

00:06:22.230 --> 00:06:25.330
in terms of the PV
industry benefited a lot

00:06:25.330 --> 00:06:26.551
by that cross-pollination.

00:06:26.551 --> 00:06:28.800
Many technologies came in
from the integrated circuits

00:06:28.800 --> 00:06:32.336
industry to assist or give
a boost to the PV industry.

00:06:32.336 --> 00:06:33.710
This number is a
little outdated,

00:06:33.710 --> 00:06:34.876
it's now about $100 billion.

00:06:34.876 --> 00:06:39.840
Hard to keep up with things
growing at 68% a year.

00:06:39.840 --> 00:06:44.320
Technology acceptance results
in lower interest rates.

00:06:44.320 --> 00:06:47.080
So if you have a
technology that is

00:06:47.080 --> 00:06:50.160
well-accepted by the market
then you go to a bank and say,

00:06:50.160 --> 00:06:51.960
hey, I want to install
some of those things

00:06:51.960 --> 00:06:53.751
and the bank says what
are those things you

00:06:53.751 --> 00:06:55.330
say oh, hundreds of
thousands of them

00:06:55.330 --> 00:06:56.180
have been installed already.

00:06:56.180 --> 00:06:56.700
It's OK.

00:06:56.700 --> 00:06:57.610
It's a proven technology.

00:06:57.610 --> 00:06:59.651
The bank says OK, I'll
lower your interest rates.

00:06:59.651 --> 00:07:01.790
That means you pay
less money on interest.

00:07:01.790 --> 00:07:04.030
Your capital is more cheap.

00:07:04.030 --> 00:07:07.490
It works better in your
favor, and the opposite

00:07:07.490 --> 00:07:10.170
is true with an entirely new
technology that's unproven.

00:07:10.170 --> 00:07:13.170
So that's really
summing up why silicon.

00:07:13.170 --> 00:07:15.490
Momentum, forward
motion if you will,

00:07:15.490 --> 00:07:18.500
some inherent intrinsic
technological advantages,

00:07:18.500 --> 00:07:21.030
some of which are
listed here, and I'll

00:07:21.030 --> 00:07:23.067
get to that in a
second, scalability.

00:07:23.067 --> 00:07:24.900
To get back to the
technological advantages,

00:07:24.900 --> 00:07:27.290
I think it's important to
recognize what they are so

00:07:27.290 --> 00:07:30.350
that when you're thinking
of a new material,

00:07:30.350 --> 00:07:32.492
you can cross check and
say, gee, do I have these

00:07:32.492 --> 00:07:33.450
or do I not have these.

00:07:33.450 --> 00:07:35.040
If I don't have them, it's
not the end of the world.

00:07:35.040 --> 00:07:37.530
You might have other advantages
that overcome the ones

00:07:37.530 --> 00:07:40.740
that silicon doesn't have.

00:07:40.740 --> 00:07:43.360
Let's add some more
into this list.

00:07:43.360 --> 00:07:45.940
Just stream of consciousness.

00:07:45.940 --> 00:07:48.390
Silicon has a very
high refractive index

00:07:48.390 --> 00:07:50.030
near the band gap edge.

00:07:50.030 --> 00:07:52.330
So, near the band gap
edge, it's absorbing light

00:07:52.330 --> 00:07:53.170
less efficiently.

00:07:53.170 --> 00:07:53.670
Right?

00:07:53.670 --> 00:07:56.660
It has a larger
attenuation length

00:07:56.660 --> 00:07:59.350
of the light, a smaller optical
absorption coefficient right

00:07:59.350 --> 00:08:01.180
as you approach the band gap.

00:08:01.180 --> 00:08:04.200
So silicon absorbs poorly
in the infrared because it's

00:08:04.200 --> 00:08:05.840
an indirect band
gap semiconductor,

00:08:05.840 --> 00:08:08.730
but it also has a
very large optical,

00:08:08.730 --> 00:08:10.720
sorry, a very large
real component

00:08:10.720 --> 00:08:12.355
of the refractive index.

00:08:12.355 --> 00:08:15.020
Does anybody remember
what that refers to?

00:08:15.020 --> 00:08:17.360
Real component of
refractive index.

00:08:17.360 --> 00:08:18.540
Lesson number two.

00:08:18.540 --> 00:08:20.185
What does that dictate?

00:08:20.185 --> 00:08:21.060
AUDIENCE: Reflection.

00:08:21.060 --> 00:08:22.390
PROFESSOR: Reflection, exactly.

00:08:22.390 --> 00:08:25.691
So, if I were to tailor and
index of refraction grading

00:08:25.691 --> 00:08:27.190
on the front side
of my device, so I

00:08:27.190 --> 00:08:30.470
allow the light to be absorbed
efficiently, on the backside

00:08:30.470 --> 00:08:33.230
I can put a very large
index of refraction mismatch

00:08:33.230 --> 00:08:35.020
so that the light bounces back.

00:08:35.020 --> 00:08:36.929
In other words, the
light trapping silicon

00:08:36.929 --> 00:08:38.320
is benefited by
the fact that you

00:08:38.320 --> 00:08:41.289
have this awesome
reflection capability.

00:08:41.289 --> 00:08:43.319
The refractive
index is around 3.6,

00:08:43.319 --> 00:08:45.110
the real component of
the refractive index,

00:08:45.110 --> 00:08:47.600
in the infrared at
around 1070 nanometers.

00:08:47.600 --> 00:08:49.590
Which means that if you
design your cell right,

00:08:49.590 --> 00:08:51.714
you can get an extension
of the optical path length

00:08:51.714 --> 00:08:54.700
by a factor of 50
over the thickness.

00:08:54.700 --> 00:08:56.414
So if your thickness
of the device is d,

00:08:56.414 --> 00:08:58.330
the optical path length
can be increased up to

00:08:58.330 --> 00:09:01.700
about [? 51d. ?]
That's as a result

00:09:01.700 --> 00:09:03.880
of this great reflectance.

00:09:03.880 --> 00:09:07.090
Many other materials that are
being explored as PV materials

00:09:07.090 --> 00:09:08.760
have refractive
indices around two,

00:09:08.760 --> 00:09:10.676
which would mean that
your optical path length

00:09:10.676 --> 00:09:12.080
extension is around 16.

00:09:12.080 --> 00:09:13.580
So that's one thing
to keep in mind.

00:09:13.580 --> 00:09:15.663
even though it doesn't
absorb light quite as well,

00:09:15.663 --> 00:09:16.870
it traps light fairly well.

00:09:16.870 --> 00:09:19.310
Another advantage of
silicon is that it

00:09:19.310 --> 00:09:23.240
forms sp3 hybridized
orbitals, for chemists, it

00:09:23.240 --> 00:09:25.207
forms-- it's
tetrahedrally coordinated,

00:09:25.207 --> 00:09:27.040
in other words bond to
four other neighbors,

00:09:27.040 --> 00:09:30.000
and most 3D transition
metals don't do that.

00:09:30.000 --> 00:09:32.020
They don't bond in
that configuration.

00:09:32.020 --> 00:09:35.890
Some do but many don't
and, as a result,

00:09:35.890 --> 00:09:39.210
the solid's solubility in
other words, the ability

00:09:39.210 --> 00:09:42.280
to incorporate impurities into a
growing silicon crystal is low.

00:09:42.280 --> 00:09:45.790
It rejects the impurities
from the solid into the melt,

00:09:45.790 --> 00:09:48.552
and you're able to purify the
material very efficiently.

00:09:48.552 --> 00:09:50.510
That's not always the
case with most materials.

00:09:50.510 --> 00:09:52.670
Sometimes they incorporate
impurities very readily,

00:09:52.670 --> 00:09:54.296
up to a few atomic percents.

00:09:54.296 --> 00:09:56.170
The typical impurity
concentration of silicon

00:09:56.170 --> 00:09:58.870
is in the order of parts per
million, parts per billion,

00:09:58.870 --> 00:09:59.820
parts per trillion.

00:09:59.820 --> 00:10:01.660
Still can be enough, as you
learned during your homework

00:10:01.660 --> 00:10:03.785
assignments, still could
be enough to affect device

00:10:03.785 --> 00:10:06.220
performance but is very low.

00:10:06.220 --> 00:10:08.050
It would be a lot
worse if silicon

00:10:08.050 --> 00:10:10.710
were able to absorb more
impurities and so forth.

00:10:10.710 --> 00:10:14.120
So, there are a number of
reasons why the silicon PV

00:10:14.120 --> 00:10:18.150
technology has gained the
foothold that it has so

00:10:18.150 --> 00:10:21.000
to bump it out of its
leadership position,

00:10:21.000 --> 00:10:24.290
one really has to be clever
and the parameter of merit

00:10:24.290 --> 00:10:27.360
is performance per unit cost.

00:10:27.360 --> 00:10:31.340
Kilowatt hours per
dollar, if you will.

00:10:31.340 --> 00:10:33.940
So, we're going to talk about
the current manufacturing

00:10:33.940 --> 00:10:36.290
methods and materials
because this will give you

00:10:36.290 --> 00:10:39.240
an insight into the
dollars per kilowatt

00:10:39.240 --> 00:10:40.740
hour, the kilowatt
hours per dollar.

00:10:40.740 --> 00:10:44.110
Essentially, the cost
per unit energy produced.

00:10:44.110 --> 00:10:45.790
You can begin to
seize opportunities

00:10:45.790 --> 00:10:48.331
within the crystal silicon world
to improve the manufacturing

00:10:48.331 --> 00:10:50.560
process or you can begin
to say OK, you know what,

00:10:50.560 --> 00:10:51.982
this is way too complicated.

00:10:51.982 --> 00:10:53.690
Let me take a completely
different route.

00:10:53.690 --> 00:10:55.270
I'm going to develop a
new technology instead

00:10:55.270 --> 00:10:57.436
that will overcome these
manufacturing difficulties.

00:10:57.436 --> 00:10:59.090
So let's explore them in detail.

00:10:59.090 --> 00:11:00.950
First, the market.

00:11:00.950 --> 00:11:06.300
This is the evolution of market
share from 1980 to mid 2000s.

00:11:06.300 --> 00:11:10.620
After mid 2000s, the market just
continues growing at 68% a year

00:11:10.620 --> 00:11:12.570
and you really lose
resolution to this portion

00:11:12.570 --> 00:11:16.660
down here so it's to 2006 so
that we can actually see what's

00:11:16.660 --> 00:11:19.150
going on in the earlier days.

00:11:19.150 --> 00:11:22.110
In the earlier days,
1980, let's pick 1985,

00:11:22.110 --> 00:11:24.910
the market was split about
a third-third-third between

00:11:24.910 --> 00:11:27.730
thin films, amorphous
silicon namely,

00:11:27.730 --> 00:11:29.540
monocrystalline
silicon, and a material

00:11:29.540 --> 00:11:31.280
called multicrystalline silicon.

00:11:31.280 --> 00:11:32.630
Now let's go piece by piece.

00:11:32.630 --> 00:11:36.320
What is monocrystalline silicon,
multicrystalline silicon

00:11:36.320 --> 00:11:37.270
and thin films?

00:11:37.270 --> 00:11:40.230
Well thin films are
materials that are usually

00:11:40.230 --> 00:11:43.530
between a few hundred
nanometers up to about three,

00:11:43.530 --> 00:11:45.410
maybe five microns thick.

00:11:45.410 --> 00:11:47.070
To give you size
perspective, your hair

00:11:47.070 --> 00:11:48.850
is about 50 microns
in diameter, so we're

00:11:48.850 --> 00:11:51.480
talking about 1/50 the
width of your hair.

00:11:51.480 --> 00:11:53.590
That's the active absorber
layer and of course

00:11:53.590 --> 00:11:55.470
the plastics and encapsulates
and everything else that

00:11:55.470 --> 00:11:57.150
go around them make
it a bit thicker,

00:11:57.150 --> 00:11:59.130
but the absorber
layer is very thin

00:11:59.130 --> 00:12:02.434
and so you're not spending
much on your absorber layer.

00:12:02.434 --> 00:12:03.850
It absorbs light
very efficiently,

00:12:03.850 --> 00:12:07.240
has a very large
absorption coefficient,

00:12:07.240 --> 00:12:09.490
and is able to absorb
photons efficiently.

00:12:09.490 --> 00:12:11.751
Crystalline silicon,
on the other hand,

00:12:11.751 --> 00:12:14.125
does not absorb light as well
as many thin film materials

00:12:14.125 --> 00:12:15.930
so we need about an
order of magnitude

00:12:15.930 --> 00:12:19.380
to two orders of magnitude
thicker substrates,

00:12:19.380 --> 00:12:22.212
and the crystalline
silicon substrates today

00:12:22.212 --> 00:12:23.920
in commercial
manufacturing are typically

00:12:23.920 --> 00:12:30.150
between 160 to 190 microns,
with an average around 170, 180.

00:12:30.150 --> 00:12:32.870
So about four times the
thickness of your hair.

00:12:32.870 --> 00:12:34.887
Monocrystalline
silicon and multi.

00:12:34.887 --> 00:12:36.470
Let's talk about the
difference there.

00:12:36.470 --> 00:12:38.670
So, monocrystalline
silicon, folks

00:12:38.670 --> 00:12:41.390
are probably familiar seeing
pictures, at least something

00:12:41.390 --> 00:12:41.890
like this.

00:12:41.890 --> 00:12:42.730
Right?

00:12:42.730 --> 00:12:46.900
So this right here is an example
of a Cherkofsky silicon wafer.

00:12:49.690 --> 00:12:51.980
Appropriate for
integrated circuit work.

00:12:51.980 --> 00:12:54.780
I'll pass this around so
folks can get a sense.

00:12:54.780 --> 00:12:59.400
So this is an example of
a monocrystalline silicon

00:12:59.400 --> 00:13:01.730
wafer for the integrated
circuits industry.

00:13:01.730 --> 00:13:04.020
Let's analyze it in a
little bit more detail.

00:13:04.020 --> 00:13:07.757
So, the front surface is
polished, nicely polished.

00:13:07.757 --> 00:13:09.590
Polished to, I think,
somewhere in the order

00:13:09.590 --> 00:13:12.280
of a few nanometers
mean surface roughness.

00:13:12.280 --> 00:13:14.800
Using a chemical mechanical
polishing mechanism.

00:13:14.800 --> 00:13:18.840
The thickness is around
700-- or 675 microns.

00:13:18.840 --> 00:13:20.120
Somewhere in that range.

00:13:20.120 --> 00:13:21.710
So very, very thick wafer.

00:13:21.710 --> 00:13:23.690
The objective is not to break.

00:13:23.690 --> 00:13:24.260
Right?

00:13:24.260 --> 00:13:25.830
If you're making
integrated circuit,

00:13:25.830 --> 00:13:27.450
this entire wafer that
I'm holding right here

00:13:27.450 --> 00:13:29.580
could be worth a few 10s or
100s thousands of dollars

00:13:29.580 --> 00:13:31.220
by the end of the
processing sequence,

00:13:31.220 --> 00:13:33.594
so if one of these breaks,
that's an awful lot of revenue

00:13:33.594 --> 00:13:34.680
that the company's losing.

00:13:34.680 --> 00:13:36.812
So the substrate is
thick because they

00:13:36.812 --> 00:13:37.770
don't want it to break.

00:13:37.770 --> 00:13:40.450
Silicon is brittle
at room temperature.

00:13:40.450 --> 00:13:42.660
If you were to manufacture
solar cell out of this,

00:13:42.660 --> 00:13:44.210
you could but it would
be very expensive.

00:13:44.210 --> 00:13:45.584
The chemical
mechanical polishing

00:13:45.584 --> 00:13:47.260
that they use to
flatten the surface out

00:13:47.260 --> 00:13:49.720
costs a lot of money,
it's very time intensive,

00:13:49.720 --> 00:13:52.250
and the thickness of the
silicon is above and beyond

00:13:52.250 --> 00:13:54.290
what is necessary to
absorb light well.

00:13:54.290 --> 00:13:55.990
If anything, increasing
the thickness

00:13:55.990 --> 00:13:58.215
is just increasing your
emitter saturation current,

00:13:58.215 --> 00:14:00.340
since you have a higher
recombination current being

00:14:00.340 --> 00:14:01.822
driven by bulk recombination.

00:14:01.822 --> 00:14:03.280
You have more
recombination centers

00:14:03.280 --> 00:14:04.821
because you have a
greater thickness,

00:14:04.821 --> 00:14:06.640
and it's driving a
larger diffusion current

00:14:06.640 --> 00:14:08.740
from the emitter into the base.

00:14:08.740 --> 00:14:11.680
So making it this thick
really doesn't make sense.

00:14:11.680 --> 00:14:15.310
So I'll pass this around so
folks can kind of get a sense.

00:14:15.310 --> 00:14:17.390
Make sure this gets
the entire round.

00:14:17.390 --> 00:14:20.397
I'll be recycling those platens.

00:14:20.397 --> 00:14:22.230
Please hold, if you're
going to take it out,

00:14:22.230 --> 00:14:24.220
which you're welcome
to do, please hold it

00:14:24.220 --> 00:14:26.900
like a photograph.

00:14:26.900 --> 00:14:31.720
What I don't want to have happen
is folks put their fingerprints

00:14:31.720 --> 00:14:33.430
all over it.

00:14:33.430 --> 00:14:38.960
The wafers that are
used in the PV industry

00:14:38.960 --> 00:14:40.850
are cut from the
same ingot like that

00:14:40.850 --> 00:14:44.320
one, except that the
ingots, essentially,

00:14:44.320 --> 00:14:47.770
if you were to pack circular
wafers into a module,

00:14:47.770 --> 00:14:49.890
it would look
something like this.

00:14:49.890 --> 00:14:52.259
Here's your module
and, mind you,

00:14:52.259 --> 00:14:54.050
you're spending a lot
of money on the glass

00:14:54.050 --> 00:14:56.760
and the encapsulates
and the aluminum framing

00:14:56.760 --> 00:15:01.469
and so forth, and now your
solar cells look like that.

00:15:01.469 --> 00:15:03.760
There's probably more of them
that you can put in here,

00:15:03.760 --> 00:15:05.760
but what do you
notice about this?

00:15:05.760 --> 00:15:07.810
What is the packing density,
or packing fraction.

00:15:07.810 --> 00:15:08.830
It's very low, right?

00:15:08.830 --> 00:15:10.746
You're losing all of
this material in between.

00:15:10.746 --> 00:15:13.370
All that space is just
going to be blank space.

00:15:13.370 --> 00:15:16.690
Some of the earliest PV modules
actually use circular wafers,

00:15:16.690 --> 00:15:18.390
but the more modern
ones, what they do

00:15:18.390 --> 00:15:20.770
is a very complicated
cost analysis

00:15:20.770 --> 00:15:29.550
where they say, OK, if I were to
chop off the edges of my wafer

00:15:29.550 --> 00:15:32.330
and completely remove them,
I'd be losing a lot of silicon

00:15:32.330 --> 00:15:34.970
but I'd be increasing
the packing fraction.

00:15:34.970 --> 00:15:37.660
So in the limit that my
module materials, the glass,

00:15:37.660 --> 00:15:40.080
the encapsulant, the
framing materials

00:15:40.080 --> 00:15:43.070
are infinitely expensive and
my silicon costs nothing,

00:15:43.070 --> 00:15:44.740
I want to do this.

00:15:44.740 --> 00:15:47.300
In the limit that my
module materials are free

00:15:47.300 --> 00:15:51.010
and installation is free but
the silicon is super expensive,

00:15:51.010 --> 00:15:52.980
I want to keep
full round wafers,

00:15:52.980 --> 00:15:55.210
and the reality is that
we're somewhere in between.

00:15:55.210 --> 00:15:57.760
And so, instead of making
one or the other extreme,

00:15:57.760 --> 00:15:59.940
typically what you'll
see is something

00:15:59.940 --> 00:16:02.750
like this chopped
off, like that,

00:16:02.750 --> 00:16:06.480
where you have a pseudo-square.

00:16:06.480 --> 00:16:09.570
The wafer itself has
flat edges on the sides

00:16:09.570 --> 00:16:13.000
but it also has kind of
pseudo-rounded corners here,

00:16:13.000 --> 00:16:15.950
and Joe did we bring
any of those in?

00:16:15.950 --> 00:16:18.810
The psuedo-squares, the
monocrystalline psuedo-squares.

00:16:18.810 --> 00:16:20.190
These ones.

00:16:20.190 --> 00:16:20.690
OK.

00:16:20.690 --> 00:16:21.189
All right.

00:16:21.189 --> 00:16:21.776
No worries.

00:16:21.776 --> 00:16:23.150
I'll show them to
you next class.

00:16:23.150 --> 00:16:25.390
So, the idea is to
make-- cut it out

00:16:25.390 --> 00:16:27.200
of the same ingot
as that one right

00:16:27.200 --> 00:16:31.550
there, but make it thinner, on
the order of 170 microns thick,

00:16:31.550 --> 00:16:36.620
and to chop off part of the
edge, and how much you chop off

00:16:36.620 --> 00:16:40.060
depends on the dynamic pricing
of silicon versus module

00:16:40.060 --> 00:16:42.660
materials and installation
and whether or not

00:16:42.660 --> 00:16:44.230
you can sell the
module, if there's

00:16:44.230 --> 00:16:45.700
a certain threshold
of performance

00:16:45.700 --> 00:16:47.908
that it needs to reach
because obviously, if you have

00:16:47.908 --> 00:16:50.980
a bunch of dead space in here,
you're losing that to-- you're

00:16:50.980 --> 00:16:52.430
not producing power out of that.

00:16:52.430 --> 00:16:55.230
So if somebody wants a
module that's yay efficient,

00:16:55.230 --> 00:16:57.440
you might want to increase
the packing density.

00:16:57.440 --> 00:16:59.490
So that's
monocrystalline silicon.

00:16:59.490 --> 00:17:01.590
Multicrystalline silicon.

00:17:01.590 --> 00:17:03.110
Let's put it this way for now.

00:17:03.110 --> 00:17:05.026
We'll describe how
multicrystalline silicon is

00:17:05.026 --> 00:17:06.569
made, but for now
I'm going to say

00:17:06.569 --> 00:17:11.180
that multicrystalline silicon
is a crystalline silicon

00:17:11.180 --> 00:17:14.099
variety that is comprised
of many small grains.

00:17:14.099 --> 00:17:16.950
So if you look at a
multicrystalline silicon wafer,

00:17:16.950 --> 00:17:19.250
something like,
let's say, oh this

00:17:19.250 --> 00:17:21.430
is a perfect example
right in here.

00:17:21.430 --> 00:17:23.720
If you look at a
multicrystalline silicon wafer,

00:17:23.720 --> 00:17:30.050
you can see that it looks
nice and-- here maybe, that's

00:17:30.050 --> 00:17:31.670
probably an OK view of it.

00:17:31.670 --> 00:17:32.920
You can see individual grains.

00:17:32.920 --> 00:17:33.510
Right?

00:17:33.510 --> 00:17:35.600
If you look closely at it.

00:17:35.600 --> 00:17:38.520
And those are grains
of crystalline material

00:17:38.520 --> 00:17:40.387
that are joined by
grain boundaries.

00:17:40.387 --> 00:17:41.970
So the grain orientation
in one region

00:17:41.970 --> 00:17:43.780
might be pointing in
this direction, the grain

00:17:43.780 --> 00:17:45.240
orientation in the
neighboring region like that,

00:17:45.240 --> 00:17:47.220
and they come together
at a grain boundary

00:17:47.220 --> 00:17:50.220
and, when we have
polycrystalline materials

00:17:50.220 --> 00:17:55.310
like this, it's generally
indicative of some faster

00:17:55.310 --> 00:17:58.550
growth that didn't allow
for a nice homogeneous

00:17:58.550 --> 00:18:00.074
single crystal
material to evolve,

00:18:00.074 --> 00:18:02.490
and that's indeed what happens
during the multicrystalline

00:18:02.490 --> 00:18:03.680
silicon ingot growth.

00:18:03.680 --> 00:18:08.140
It's occurring under a slightly
modified growth condition then,

00:18:08.140 --> 00:18:11.140
say, that beautiful single
crystalline piece over there,

00:18:11.140 --> 00:18:14.160
and we'll explain how
they're made in a second.

00:18:14.160 --> 00:18:16.930
So those are the technologies
in general, the base absorber

00:18:16.930 --> 00:18:19.200
materials, and then there's
ribbon silicon which

00:18:19.200 --> 00:18:21.180
is a really, really
small fraction

00:18:21.180 --> 00:18:23.940
of the total production in
decreasing, but at one time,

00:18:23.940 --> 00:18:27.019
ribbon silicon was viewed as
the up and coming technology.

00:18:27.019 --> 00:18:29.060
Still today, there are
about 20 startup companies

00:18:29.060 --> 00:18:31.450
around the United States
working on some aspect of this

00:18:31.450 --> 00:18:34.569
and probably about a dozen
more around the world.

00:18:34.569 --> 00:18:35.069
Yeah.

00:18:35.069 --> 00:18:36.488
AUDIENCE: I had a
question about the multi.

00:18:36.488 --> 00:18:37.154
PROFESSOR: Yeah.

00:18:37.154 --> 00:18:39.892
AUDIENCE: So for
the multi and micro

00:18:39.892 --> 00:18:43.110
and poly, is that
different grain sizes?

00:18:43.110 --> 00:18:43.940
PROFESSOR: Sort of.

00:18:43.940 --> 00:18:47.140
So, multicrystalline silicon
is a polycrystalline silicon

00:18:47.140 --> 00:18:48.000
material.

00:18:48.000 --> 00:18:49.750
The definition of
multicrystalline silicon

00:18:49.750 --> 00:18:53.280
is that the average grain size
is about a centimeter squared,

00:18:53.280 --> 00:18:56.670
or larger, and that's where
multicrystalline came about.

00:18:56.670 --> 00:18:59.630
Polycrystalline silicon,
in the silicon community,

00:18:59.630 --> 00:19:01.290
has a very specific meaning.

00:19:01.290 --> 00:19:05.430
It means, usually a
plasma-enhanced chemical vapor

00:19:05.430 --> 00:19:08.670
deposited layer, so
PCVD-deposited layer

00:19:08.670 --> 00:19:12.520
of silicon, that has on the
order of one to five micron

00:19:12.520 --> 00:19:13.500
diameter grains.

00:19:13.500 --> 00:19:15.530
So very, very small
grain material.

00:19:15.530 --> 00:19:17.830
About 1/50 the
width of your hair.

00:19:17.830 --> 00:19:19.450
Maybe 1/10 the
width of your hair

00:19:19.450 --> 00:19:21.575
and, to distinguish it from
that really small grain

00:19:21.575 --> 00:19:23.570
material that will
perform very poorly,

00:19:23.570 --> 00:19:27.140
one calls this
multicrystalline silicon.

00:19:27.140 --> 00:19:29.309
AUDIENCE: And is there
microcrystalline silicon?

00:19:29.309 --> 00:19:31.350
PROFESSOR: There is also
microcrystalline silicon

00:19:31.350 --> 00:19:34.270
and microcrystalline
silicon is actually

00:19:34.270 --> 00:19:37.670
at the phase transition between
amorphous and polycrystalline

00:19:37.670 --> 00:19:38.530
silicon.

00:19:38.530 --> 00:19:40.880
So as you're going from
an amorphous material

00:19:40.880 --> 00:19:43.320
increasing the temperature,
let's say, of growth

00:19:43.320 --> 00:19:47.250
or increasing other parameters
during the deposition process,

00:19:47.250 --> 00:19:49.870
as you begin to evolve
from an amorphous material

00:19:49.870 --> 00:19:51.740
into a crystalline
material, you transition

00:19:51.740 --> 00:19:53.400
through this
microcrystalline regime

00:19:53.400 --> 00:19:54.937
which is a bit of a hybrid.

00:19:54.937 --> 00:19:56.520
It has some regions
that are amorphous

00:19:56.520 --> 00:19:58.720
and other regions
that are crystalline.

00:19:58.720 --> 00:20:02.190
In your assigned readings,
this book was assigned,

00:20:02.190 --> 00:20:05.820
and I believe in the syllabus
it says read chapter X.

00:20:05.820 --> 00:20:07.636
Unfortunately, there
is no chapter X.

00:20:07.636 --> 00:20:09.700
I guess you could
interpret it as 10,

00:20:09.700 --> 00:20:12.560
but the essence was that there
are two versions of the book.

00:20:12.560 --> 00:20:14.630
One is version three,
which was published

00:20:14.630 --> 00:20:16.840
about seven years ago, and
the newest version just

00:20:16.840 --> 00:20:18.330
came out last year.

00:20:18.330 --> 00:20:20.630
The newest addition
is addition three.

00:20:20.630 --> 00:20:24.450
So the chapters have rearranged
slightly, but what I'll do

00:20:24.450 --> 00:20:26.830
is I'll highlight crystalline
silicon solar cells

00:20:26.830 --> 00:20:28.660
and modules in here
so that you can

00:20:28.660 --> 00:20:31.430
get a sense of what
is in the chapter

00:20:31.430 --> 00:20:34.150
and you're welcome to
go back and have a look.

00:20:34.150 --> 00:20:36.840
So I'll go ahead and highlight
this chapter right here

00:20:36.840 --> 00:20:37.785
and pass it around.

00:20:37.785 --> 00:20:39.660
Feel free to glance
through the book as well.

00:20:39.660 --> 00:20:40.440
It's a great read.

00:20:40.440 --> 00:20:41.940
It dives into great
detail into each

00:20:41.940 --> 00:20:44.680
of the different technologies.

00:20:44.680 --> 00:20:45.180
OK.

00:20:45.180 --> 00:20:47.060
So, let's talk about
feedstock refining.

00:20:47.060 --> 00:20:49.390
We're going to start
the silicon value

00:20:49.390 --> 00:20:51.610
chain from the raw
materials and work our way

00:20:51.610 --> 00:20:55.050
all the way to the
final module at the end.

00:20:55.050 --> 00:20:57.620
So we'll start with the
feedstocks themselves.

00:20:57.620 --> 00:21:00.420
Down here is a rough
cost breakdown.

00:21:00.420 --> 00:21:03.170
Kind of think of
it as wafer, cell,

00:21:03.170 --> 00:21:05.060
module being like
a third-third-third

00:21:05.060 --> 00:21:07.410
of the total module cost
and then balance the system

00:21:07.410 --> 00:21:09.340
components beyond that.

00:21:09.340 --> 00:21:11.190
So we'll start
from our feedstocks

00:21:11.190 --> 00:21:12.690
and the raw materials
in the ground,

00:21:12.690 --> 00:21:14.340
we'll wind up with
systems on the roof,

00:21:14.340 --> 00:21:15.900
and we'll walk through
each of the different steps

00:21:15.900 --> 00:21:17.760
of current
manufacturing process.

00:21:17.760 --> 00:21:19.540
So raw materials.

00:21:19.540 --> 00:21:24.050
Shown here is quartz and
coal, for a very good reason.

00:21:24.050 --> 00:21:28.280
The way feedstock refining
occurs at the very first stage

00:21:28.280 --> 00:21:32.690
is to take oxidized silicon,
silicon dioxide, quartz

00:21:32.690 --> 00:21:37.400
and to reduce it to
silicon, say, silicon zero.

00:21:37.400 --> 00:21:41.440
Unoxidized silicon, which is
also called silicon metal.

00:21:41.440 --> 00:21:43.797
It's called a metal because
it is very low resistivity.

00:21:43.797 --> 00:21:46.005
It's very low resistivity
because there's a very high

00:21:46.005 --> 00:21:47.210
impurity content still.

00:21:47.210 --> 00:21:49.920
The purity of this material
coming out here is around 99,

00:21:49.920 --> 00:21:51.984
99.9% here.

00:21:51.984 --> 00:21:53.900
So, it sounds like a
high purity but, if we're

00:21:53.900 --> 00:21:56.010
talking about parts per
million of impurities,

00:21:56.010 --> 00:21:58.520
we have some further refining
steps to do after this.

00:21:58.520 --> 00:21:59.880
So let's walk through this.

00:21:59.880 --> 00:22:02.740
We start with the raw
materials in the upper left.

00:22:02.740 --> 00:22:04.360
It says raw material inputs.

00:22:04.360 --> 00:22:06.350
Carbon and SiO2.

00:22:06.350 --> 00:22:08.740
The SiO2 forms, usually, quartz.

00:22:08.740 --> 00:22:12.990
That can be some of high
purity pegmatite, it could be,

00:22:12.990 --> 00:22:16.277
for example, a hydrothermal
quartz, higher purity

00:22:16.277 --> 00:22:17.110
varieties of quartz.

00:22:17.110 --> 00:22:21.850
You could even use, maybe, a
metamorphic quartzite material.

00:22:21.850 --> 00:22:22.980
Let me explain.

00:22:22.980 --> 00:22:25.550
So, some of the highest
purity materials

00:22:25.550 --> 00:22:29.180
are coming from these veins
of magma that float up

00:22:29.180 --> 00:22:33.430
and then phase separated
during millennia.

00:22:33.430 --> 00:22:35.290
Some of the lowest
purity quartz is

00:22:35.290 --> 00:22:38.310
coming from sand,
essentially crushed rock that

00:22:38.310 --> 00:22:40.630
made its way into, say,
a beach-like environment

00:22:40.630 --> 00:22:43.010
and then rock was
deposited on top of that,

00:22:43.010 --> 00:22:46.510
pressure was increased,
and this whole mixture

00:22:46.510 --> 00:22:50.710
of mica, feldspar,
and of quartz got

00:22:50.710 --> 00:22:53.280
pushed together and
formed a solid block.

00:22:53.280 --> 00:22:57.084
That would be your
metamorphic quartz materials,

00:22:57.084 --> 00:22:58.750
and so you'd have a
much higher impurity

00:22:58.750 --> 00:23:00.750
content in the metamorphic
quartz than you would

00:23:00.750 --> 00:23:04.760
in, say, a high purity pegmatite
or hydrothermal quartz.

00:23:04.760 --> 00:23:07.280
Regardless, depending on the
feedstock source of the quartz,

00:23:07.280 --> 00:23:08.780
and there are people
who study this.

00:23:08.780 --> 00:23:10.738
Believe it or not, there
are entire departments

00:23:10.738 --> 00:23:13.111
dedicated to mining
quartz and figuring out

00:23:13.111 --> 00:23:15.360
where the different veins
of the highest purity quartz

00:23:15.360 --> 00:23:18.320
are, where you get them from.

00:23:18.320 --> 00:23:21.020
That's the SiO2 input
and the C input over here

00:23:21.020 --> 00:23:23.350
on the left hand side, Carbon.

00:23:23.350 --> 00:23:26.130
So, typically what is
used in the PV industry

00:23:26.130 --> 00:23:31.350
is either a fast-growing
wood source like eucalyptus

00:23:31.350 --> 00:23:34.240
or southern pine, right?

00:23:34.240 --> 00:23:36.460
Northern pine tends
to be slower growing,

00:23:36.460 --> 00:23:38.220
but eucalyptus and
southern pine both

00:23:38.220 --> 00:23:40.050
tend to be fairly fast-growing.

00:23:40.050 --> 00:23:43.240
You can tell by the
spacing in the rings,

00:23:43.240 --> 00:23:48.540
if you chop the tree down and
do a cross section, or coal.

00:23:48.540 --> 00:23:50.380
So carbon, essentially.

00:23:50.380 --> 00:23:54.910
And the two react inside
of this furnace right here

00:23:54.910 --> 00:23:57.860
and this furnace, just to
give you a sense of scale,

00:23:57.860 --> 00:23:59.420
here's a human being.

00:23:59.420 --> 00:24:00.510
This is the furnace.

00:24:00.510 --> 00:24:03.430
So it's about five stories
tall, 12 meters in diameter.

00:24:03.430 --> 00:24:05.180
It's a big, big, big creature.

00:24:05.180 --> 00:24:09.390
This furnace right here
is what is producing

00:24:09.390 --> 00:24:11.640
the reduced silicon
and what's happening

00:24:11.640 --> 00:24:14.770
is these feedstock chunks are
being thrown in at the top

00:24:14.770 --> 00:24:19.800
and there's an arc going
between the electrodes, usually

00:24:19.800 --> 00:24:23.950
some carbon-bearing
material, and a base contact,

00:24:23.950 --> 00:24:26.510
and so that arc creates
a very high temperature.

00:24:26.510 --> 00:24:28.610
Something in the order
of up to 2000 degrees

00:24:28.610 --> 00:24:30.850
Celsius, near the arc,
and the temperature

00:24:30.850 --> 00:24:33.050
decreases as you go
further and further away,

00:24:33.050 --> 00:24:34.852
so up near the top
here it might be even

00:24:34.852 --> 00:24:36.560
below the melting
temperature of silicon,

00:24:36.560 --> 00:24:38.140
somewhere around 1,200 degrees.

00:24:38.140 --> 00:24:41.020
So this is an extremely
inhomogeneous, messy system.

00:24:41.020 --> 00:24:44.750
This metallurgical grade silicon
refining furnace right here,

00:24:44.750 --> 00:24:49.510
this arc furnace, also called a
carbothermic reduction furnace,

00:24:49.510 --> 00:24:52.330
a very busy place.

00:24:52.330 --> 00:24:52.932
Lots going on.

00:24:52.932 --> 00:24:54.390
Extremely inhomogeneous
if you were

00:24:54.390 --> 00:24:56.882
to take a cross section
also in terms of temperature

00:24:56.882 --> 00:24:58.340
and in terms of
the chemical states

00:24:58.340 --> 00:25:00.040
of the different
constituents species,

00:25:00.040 --> 00:25:02.280
but the general
reaction that happens

00:25:02.280 --> 00:25:03.900
is the carbon would
much rather bond

00:25:03.900 --> 00:25:06.384
to the oxygen than silicon,
and so the carbon steals

00:25:06.384 --> 00:25:08.800
the oxygen from the silicon
reduces the silicon to silicon

00:25:08.800 --> 00:25:11.520
metal and CO2 is released.

00:25:11.520 --> 00:25:12.770
We'll get to that in a second.

00:25:12.770 --> 00:25:13.400
Flag that.

00:25:13.400 --> 00:25:14.525
Put an asterisk next to it.

00:25:14.525 --> 00:25:16.490
We'll come back to
that in a second.

00:25:16.490 --> 00:25:18.030
Other byproducts
of this reaction,

00:25:18.030 --> 00:25:20.220
so this is the liquid
silicon metal coming out here

00:25:20.220 --> 00:25:20.803
at the bottom.

00:25:20.803 --> 00:25:23.840
It's essentially liquid
molten silicon reduced,

00:25:23.840 --> 00:25:27.360
so silicon zero, not a silicon
oxide, reduced silicon metal,

00:25:27.360 --> 00:25:30.530
and then finally it's
poured into these buckets,

00:25:30.530 --> 00:25:34.660
also called ladles and
solidified, crushed up to size,

00:25:34.660 --> 00:25:36.477
and then distributed at the end.

00:25:36.477 --> 00:25:38.310
Other byproducts coming
out of this reaction

00:25:38.310 --> 00:25:40.761
include-- this is
liquid silicon up hear.

00:25:40.761 --> 00:25:42.260
It's very high
temperature and there

00:25:42.260 --> 00:25:45.870
are gases and a lot of oxygen
because of the reduction

00:25:45.870 --> 00:25:51.730
process, and so silica, or
SiO gas, can be produced

00:25:51.730 --> 00:25:55.470
and silica gas can begin
aggravating and forming

00:25:55.470 --> 00:25:59.710
very small particles,
almost like shards,

00:25:59.710 --> 00:26:03.030
of silicon oxide
material, and these

00:26:03.030 --> 00:26:05.910
can be on the order
of one to five microns

00:26:05.910 --> 00:26:08.320
and very rough and
jaggedy around the edges.

00:26:08.320 --> 00:26:11.020
Now, who here has studied public
health and knows anything about

00:26:11.020 --> 00:26:13.967
PM1 or PM1.5 denominations.

00:26:13.967 --> 00:26:14.800
Do they ring a bell?

00:26:14.800 --> 00:26:15.716
What are those Ashley?

00:26:15.716 --> 00:26:20.903
AUDIENCE: It's the
size of particles that

00:26:20.903 --> 00:26:22.200
can get stuck in your lungs.

00:26:22.200 --> 00:26:22.850
PROFESSOR: Exactly!

00:26:22.850 --> 00:26:23.349
Right?

00:26:23.349 --> 00:26:26.630
So PM1 or PM1.5 would refer
to the micron diameter,

00:26:26.630 --> 00:26:29.470
1 or 1.5 micron diameter
particle that would get stuck

00:26:29.470 --> 00:26:32.000
in the [INAUDIBLE] and
result, eventually,

00:26:32.000 --> 00:26:36.751
in edema or, probably, more of
water filling up in the lungs

00:26:36.751 --> 00:26:38.750
as a result of the body
trying to expunge these,

00:26:38.750 --> 00:26:40.150
and because they're
jaggedy and pointy,

00:26:40.150 --> 00:26:42.900
they get stuck in there and they
don't come out and eventually

00:26:42.900 --> 00:26:45.590
the people can even
affixate as a result.

00:26:45.590 --> 00:26:48.149
So, before in the past, when
we had these big smokestacks

00:26:48.149 --> 00:26:50.440
sitting on the top of these
metallurgical grade silicon

00:26:50.440 --> 00:26:53.530
refineries that would just spew
the silica dust into the air,

00:26:53.530 --> 00:26:55.650
the folks downstream
would be affected

00:26:55.650 --> 00:26:59.010
and this actually did happen, to
some degree, in, for, example,

00:26:59.010 --> 00:27:02.960
Kristiansand in Norway
and, as a result,

00:27:02.960 --> 00:27:06.810
the refineries began
putting in filters over here

00:27:06.810 --> 00:27:09.800
to prevent the silica
dust from getting thrown

00:27:09.800 --> 00:27:11.580
and spewed out
into the atmosphere

00:27:11.580 --> 00:27:14.430
and the filters are a very
interesting contraption.

00:27:14.430 --> 00:27:17.000
A lot of work went into
designing them just right

00:27:17.000 --> 00:27:19.900
to allow the air to go out
but the particulate matter

00:27:19.900 --> 00:27:22.480
to stay behind and once
every delta t, maybe

00:27:22.480 --> 00:27:26.360
in the order of an hour so,
the airflow direction inverse

00:27:26.360 --> 00:27:28.810
and all the dust comes
crashing down to the bottom

00:27:28.810 --> 00:27:30.569
and then gets collected
inside of here.

00:27:30.569 --> 00:27:33.110
It's kind of like pushing air
through the different direction

00:27:33.110 --> 00:27:35.970
through a sock, and all the
dust comes out to the bottom,

00:27:35.970 --> 00:27:39.050
you collect it, and
it's sold to the--?

00:27:39.050 --> 00:27:43.382
AUDIENCE: The footwear
industry for absorbing--

00:27:43.382 --> 00:27:44.340
PROFESSOR: It might be.

00:27:44.340 --> 00:27:46.400
I don't know, but I know
that the majority of it

00:27:46.400 --> 00:27:52.100
goes to the cement industry
and so, depending on the market

00:27:52.100 --> 00:27:54.830
rates of silicon, here at the
bottom metallurgical grade

00:27:54.830 --> 00:27:58.230
silicon, versus what the cement
industry is willing to pay,

00:27:58.230 --> 00:28:00.020
you might tune your
process to optimize

00:28:00.020 --> 00:28:01.790
for one industry or another.

00:28:01.790 --> 00:28:05.520
So, this is to say that early
on in refining processes,

00:28:05.520 --> 00:28:08.170
you're serving multiple
industries with one plant

00:28:08.170 --> 00:28:11.177
and volatility of
pricing is affected,

00:28:11.177 --> 00:28:13.260
in part, by what those
other industries are doing.

00:28:13.260 --> 00:28:14.600
What the demand there is.

00:28:14.600 --> 00:28:15.850
It's something to be aware of.

00:28:15.850 --> 00:28:18.399
Let's go back to the CO2 real
quick that's being emitted.

00:28:18.399 --> 00:28:20.440
So that is one of the
byproducts of the reaction.

00:28:20.440 --> 00:28:23.530
In terms of total CO2
content from the production

00:28:23.530 --> 00:28:27.680
of solar cells, the CO2 produced
during the reduction process

00:28:27.680 --> 00:28:30.777
is a small percentage, I think
something under 5% or 10%

00:28:30.777 --> 00:28:32.360
is the number I
pulled out of my head,

00:28:32.360 --> 00:28:35.260
it's a small percentage
of the total CO2 emitted

00:28:35.260 --> 00:28:38.850
during solar cell manufacturing
because the electricity that

00:28:38.850 --> 00:28:41.610
goes into producing the rest
of the solar cells coming

00:28:41.610 --> 00:28:43.350
from fossil fuel based
sources comprises

00:28:43.350 --> 00:28:46.270
the majority of CO2
emissions during fabrication

00:28:46.270 --> 00:28:47.380
of these devices.

00:28:47.380 --> 00:28:49.570
The electricity used to
run these electrodes,

00:28:49.570 --> 00:28:51.520
for instance, the
electricity used

00:28:51.520 --> 00:28:55.140
to melt this silicon
byproduct here,

00:28:55.140 --> 00:28:57.830
or to gasify it in the
subsequent reactions, that

00:28:57.830 --> 00:29:00.570
is the majority of the CO2
coming out of the process.

00:29:00.570 --> 00:29:02.740
Any questions so far about this?

00:29:02.740 --> 00:29:03.980
They're fun plants to see.

00:29:03.980 --> 00:29:07.460
We don't have too many
of them in the US.

00:29:07.460 --> 00:29:10.680
Majority of these carbothermic
reduction furnaces

00:29:10.680 --> 00:29:12.810
are either in China, Norway.

00:29:12.810 --> 00:29:15.460
Norway has a lot
of cheap hydropower

00:29:15.460 --> 00:29:19.030
so the hydroplant is usually
only a few 10s of kilometers

00:29:19.030 --> 00:29:21.690
away from the
refinery and if you

00:29:21.690 --> 00:29:24.380
go to, say, [INAUDIBLE]
in Norway, where they have

00:29:24.380 --> 00:29:25.980
a number of these
plants, you'll see

00:29:25.980 --> 00:29:27.479
not only silicon
being refined there

00:29:27.479 --> 00:29:31.080
but also magnesium,
other elements, aluminum

00:29:31.080 --> 00:29:33.579
being smelted in
the same peninsula--

00:29:33.579 --> 00:29:34.620
the same industrial park.

00:29:34.620 --> 00:29:40.847
AUDIENCE: When general mining
of silicon happens or silica,

00:29:40.847 --> 00:29:43.156
the Chinese have--

00:29:43.156 --> 00:29:44.530
PROFESSOR: The
reduction process,

00:29:44.530 --> 00:29:47.210
this carbothermic
reduction process here,

00:29:47.210 --> 00:29:49.340
the majority of it
happens at the same places

00:29:49.340 --> 00:29:53.140
like Norway or China-- places
that have cheap electricity.

00:29:53.140 --> 00:29:55.239
There's also a
feedstock refinery.

00:29:55.239 --> 00:29:57.030
I don't know if it
extends all the way back

00:29:57.030 --> 00:29:58.880
to the metallurgical
grade silicon refining,

00:29:58.880 --> 00:30:00.440
but there's a feedstock
refining facility

00:30:00.440 --> 00:30:01.981
going up in the
Middle East right now

00:30:01.981 --> 00:30:05.960
in Qatar, as a result of
the cheap natural gas.

00:30:05.960 --> 00:30:08.204
So, wherever you have
cheap access to energy,

00:30:08.204 --> 00:30:10.120
you can set one of these
plants up and get off

00:30:10.120 --> 00:30:11.910
and running and
your CO2 intensity

00:30:11.910 --> 00:30:14.650
will be dictated by the fuel
source that you're using.

00:30:14.650 --> 00:30:16.570
Hydro, in that case,
it might be low

00:30:16.570 --> 00:30:18.170
unless you take
methane into account

00:30:18.170 --> 00:30:20.760
that might be emitted
in the reservoir,

00:30:20.760 --> 00:30:24.250
if you have decaying biomass
underneath the water,

00:30:24.250 --> 00:30:26.497
but if you would
exclude that and if you

00:30:26.497 --> 00:30:28.580
look at the CO2 intensity
of the fossil fuels that

00:30:28.580 --> 00:30:30.177
are being burned,
it might be better

00:30:30.177 --> 00:30:32.760
to do it in, say, Norway, from
an environmental point of view,

00:30:32.760 --> 00:30:35.386
than to, say, manufacture
this stuff in China.

00:30:35.386 --> 00:30:35.886
Yeah.

00:30:35.886 --> 00:30:37.261
AUDIENCE: How many
kilowatt hours

00:30:37.261 --> 00:30:38.874
are we talking [INAUDIBLE]?

00:30:38.874 --> 00:30:40.640
PROFESSOR: Okay, so
what is the energy

00:30:40.640 --> 00:30:44.780
intensity of this process
right here, in other words.

00:30:44.780 --> 00:30:46.876
Well, why don't I
put a flag on that.

00:30:46.876 --> 00:30:48.750
Why don't we put a flag
on that and come back

00:30:48.750 --> 00:30:51.990
with specific numbers for
this process right here.

00:30:51.990 --> 00:30:54.232
I don't want to say something
and regret it later.

00:30:54.232 --> 00:30:56.065
AUDIENCE: Well, we know
the energy intensity

00:30:56.065 --> 00:30:57.350
of the solar panel itself.

00:30:57.350 --> 00:30:58.033
PROFESSOR: Yeah.

00:30:58.033 --> 00:30:59.116
AUDIENCE: But the energy--

00:30:59.116 --> 00:31:01.680
PROFESSOR: But
specifically what fraction

00:31:01.680 --> 00:31:04.760
comes from the MGSi
refining, I'd rather not

00:31:04.760 --> 00:31:07.200
pull something out of my head.

00:31:07.200 --> 00:31:09.200
Any other questions?

00:31:09.200 --> 00:31:11.340
OK.

00:31:11.340 --> 00:31:14.000
So somewhere in the order
of two million metric tons

00:31:14.000 --> 00:31:17.034
of metallurgical grade
silicon are produced annually.

00:31:17.034 --> 00:31:18.950
Probably somewhere in
the order of 10% of that

00:31:18.950 --> 00:31:21.600
is destined for the PV industry.

00:31:21.600 --> 00:31:23.950
The remainder gets
split among a variety

00:31:23.950 --> 00:31:24.970
of different industries.

00:31:24.970 --> 00:31:26.428
So what I'm talking
about here when

00:31:26.428 --> 00:31:29.080
I say metallurgical silicon, I'm
referring to this right here.

00:31:29.080 --> 00:31:30.250
This stuff coming out.

00:31:30.250 --> 00:31:34.670
It has about 99% or 99.9%
purity and it gets used

00:31:34.670 --> 00:31:36.030
in a variety of industries.

00:31:36.030 --> 00:31:38.174
So those industries
are: the PV industry,

00:31:38.174 --> 00:31:40.090
and we'll explain how
the rest of the refining

00:31:40.090 --> 00:31:43.370
happens, the integrated circuits
industry, that's the wafer that

00:31:43.370 --> 00:31:45.680
just went around that's
made its way back up here,

00:31:45.680 --> 00:31:51.510
and silicones those are--
so, a pet peeve of mine

00:31:51.510 --> 00:31:56.040
is hearing the word silicon and
silicone used interchangeably.

00:31:56.040 --> 00:31:59.710
Silicon is this element-- is an
element on the periodic table

00:31:59.710 --> 00:32:02.570
and it's the element that
comprises this wafer right

00:32:02.570 --> 00:32:03.250
here.

00:32:03.250 --> 00:32:06.642
Silicone, on the other
hand, is an organelle,

00:32:06.642 --> 00:32:09.100
I guess you could say, it's
not exactly organelle metallic,

00:32:09.100 --> 00:32:12.300
silicon isn't a
metal, but it would

00:32:12.300 --> 00:32:18.520
be a molecule that is comprised
of carbon atoms and silicon--

00:32:18.520 --> 00:32:21.300
silicon being in the middle and
the carbon being on the sides--

00:32:21.300 --> 00:32:25.440
and that is used as caulking or
sealing agent in your showers,

00:32:25.440 --> 00:32:29.110
for instance, or in
plumbing, round windows.

00:32:29.110 --> 00:32:32.320
It tends to be very flexible,
compliant but yet impermeable,

00:32:32.320 --> 00:32:34.940
preventing the inflow of gases.

00:32:34.940 --> 00:32:37.140
So silicones,
they're metal alloys

00:32:37.140 --> 00:32:39.430
including steel and aluminum.

00:32:39.430 --> 00:32:42.230
Why would you silicon there?

00:32:42.230 --> 00:32:45.085
What does it have to do
with steel or aluminum?

00:32:45.085 --> 00:32:45.910
Let me ask this.

00:32:45.910 --> 00:32:48.760
Has anyone ever played
with pure aluminum?

00:32:48.760 --> 00:32:52.300
Highly refined, ultra
high purity aluminum.

00:32:52.300 --> 00:32:54.030
Say five nines or six nines.

00:32:54.030 --> 00:32:54.850
Yes!

00:32:54.850 --> 00:32:56.470
What happens to
ultra-pure aluminum?

00:32:56.470 --> 00:32:57.941
AUDIENCE: It's really flexible.

00:32:57.941 --> 00:32:59.440
PROFESSOR: It's
really flexible, you

00:32:59.440 --> 00:33:01.127
can dent it with
your fingernail,

00:33:01.127 --> 00:33:02.710
and it wouldn't make
very great boxes.

00:33:02.710 --> 00:33:03.590
Right?

00:33:03.590 --> 00:33:07.650
So we need it to be stronger
and scratch-resistant and so

00:33:07.650 --> 00:33:09.990
we have these additives into
the aluminum, silicon being

00:33:09.990 --> 00:33:12.364
one of them, that increases
the strength of the aluminum,

00:33:12.364 --> 00:33:13.870
essentially
preventing plasticity

00:33:13.870 --> 00:33:16.780
or preventing a dislocation
flow into the material.

00:33:16.780 --> 00:33:20.630
So that's more or less how
silicon-- metallurgical grade

00:33:20.630 --> 00:33:25.170
silicon, also called MGSi as
shown up here at the very top--

00:33:25.170 --> 00:33:27.770
that's how MGSi gs is
distributed worldwide

00:33:27.770 --> 00:33:29.710
and that's the
current production.

00:33:29.710 --> 00:33:31.440
Now let me ask another question.

00:33:31.440 --> 00:33:36.170
Steel and aluminum, where
are those used the most?

00:33:36.170 --> 00:33:39.022
What industry uses
steel, aluminum the most?

00:33:39.022 --> 00:33:39.980
AUDIENCE: Construction.

00:33:39.980 --> 00:33:43.110
PROFESSOR: Constructive
industry, automotive industry.

00:33:43.110 --> 00:33:44.940
How fast are those
growing annually?

00:33:48.540 --> 00:33:50.630
Let's estimate it from GDP.

00:33:50.630 --> 00:33:51.880
Annual-- worldwide GDP.

00:33:51.880 --> 00:33:54.270
What's the worldwide
GDP growth look like.

00:33:54.270 --> 00:33:55.380
US is around 1%.

00:33:55.380 --> 00:33:55.942
China 8%.

00:33:55.942 --> 00:33:57.650
Let's pick a number
somewhere in between.

00:33:57.650 --> 00:33:58.550
Four, right?

00:33:58.550 --> 00:33:59.050
All right.

00:33:59.050 --> 00:34:01.630
So, let's say 4%, 5% worldwide.

00:34:01.630 --> 00:34:03.487
Silicone's probably
on that order.

00:34:03.487 --> 00:34:04.570
How about the PV industry.

00:34:04.570 --> 00:34:07.090
How fast is it going right now?

00:34:07.090 --> 00:34:09.615
Somewhere in the order of,
it's a volatile year right now,

00:34:09.615 --> 00:34:11.489
this one year, but in
the past, historically,

00:34:11.489 --> 00:34:14.510
it's been around
40% to 60% a year.

00:34:14.510 --> 00:34:16.560
So, where do you think
the price pressure

00:34:16.560 --> 00:34:18.810
for metallurgical grade
silicon is going to come from?

00:34:18.810 --> 00:34:19.585
What industry?

00:34:19.585 --> 00:34:20.710
It's going to come from PV.

00:34:20.710 --> 00:34:22.459
It's a small fraction
of the pie right now

00:34:22.459 --> 00:34:23.630
but it's growing fast.

00:34:23.630 --> 00:34:25.510
Something to keep in mind.

00:34:25.510 --> 00:34:29.420
So that's why, if you look at
pricing of metallurgical grade

00:34:29.420 --> 00:34:31.150
silicon, yes.

00:34:31.150 --> 00:34:33.870
Superimposed upon pricing
is a function of time.

00:34:33.870 --> 00:34:37.010
You have the global
macroeconomic situation.

00:34:37.010 --> 00:34:37.510
Right?

00:34:37.510 --> 00:34:40.020
So that's kind of the dampening
function on top of it all,

00:34:40.020 --> 00:34:43.850
but there's just this general
trend toward rising prices

00:34:43.850 --> 00:34:48.210
as you put increasing price
pressure on metallurgical grade

00:34:48.210 --> 00:34:48.811
silicon.

00:34:48.811 --> 00:34:50.310
So additional
refining capacity will

00:34:50.310 --> 00:34:54.070
be needed if the current growth
keeps up in this industry.

00:34:54.070 --> 00:34:56.630
So let me talk about going
from metallurgical grade

00:34:56.630 --> 00:35:00.400
silicon about two nines
to three nines pure.

00:35:00.400 --> 00:35:02.970
What I mean two nines
means 99%, three nines

00:35:02.970 --> 00:35:06.590
would be 99.9% pure,
to silicon that we

00:35:06.590 --> 00:35:08.670
can use for solar
cells, which typically

00:35:08.670 --> 00:35:11.730
has to be about six nines pure.

00:35:11.730 --> 00:35:15.010
And so this is called
the Siemens process which

00:35:15.010 --> 00:35:18.087
is purification through
gaseous distillation,

00:35:18.087 --> 00:35:19.920
and that's the method
that is currently used

00:35:19.920 --> 00:35:22.132
to make most of our silicon.

00:35:22.132 --> 00:35:23.590
So the way this
process works is we

00:35:23.590 --> 00:35:26.760
start with metallurgical
grade silicon at the top,

00:35:26.760 --> 00:35:29.400
represented by a little sack
of metallurgical grade silicon

00:35:29.400 --> 00:35:30.290
chunks.

00:35:30.290 --> 00:35:33.340
We produce silane gas out
of that metallurgical grade

00:35:33.340 --> 00:35:34.150
silicon.

00:35:34.150 --> 00:35:37.800
We essentially-
silane gas is SiH4.

00:35:37.800 --> 00:35:43.150
So it would essentially
be this right here.

00:35:43.150 --> 00:35:46.320
So you'd have a silicon atom
here, tetrahedrally coordinated

00:35:46.320 --> 00:35:50.720
with-- tetrahedrally meaning
four bonds with hydrogen atoms

00:35:50.720 --> 00:35:56.170
on the side-- and this is
silane gas-- well, silane--

00:35:56.170 --> 00:35:58.070
which, at room
temperature, is a gas

00:35:58.070 --> 00:36:01.660
and that's what happens
in this step right here.

00:36:01.660 --> 00:36:05.170
We're forming-- we're
gasifying the silicon.

00:36:05.170 --> 00:36:09.350
This process is the
distillation process.

00:36:09.350 --> 00:36:11.650
To extract the pure
silane gas, it's

00:36:11.650 --> 00:36:13.150
the distillation
process that's used

00:36:13.150 --> 00:36:16.689
in large towers similar to
fractional distillation where

00:36:16.689 --> 00:36:18.730
we might heat up the
material and then, depending

00:36:18.730 --> 00:36:23.490
on its mass, it settles down to
a certain height in that tower

00:36:23.490 --> 00:36:25.060
and we're able to extract it.

00:36:25.060 --> 00:36:27.210
The silane gas
here has been sold

00:36:27.210 --> 00:36:28.680
to the photovoltaics industry.

00:36:28.680 --> 00:36:30.050
LCD.

00:36:30.050 --> 00:36:31.790
Liquid crystal display.

00:36:31.790 --> 00:36:32.390
Right?

00:36:32.390 --> 00:36:34.914
Thin film industries as
well, they use silane.

00:36:34.914 --> 00:36:37.080
If you're depositing the
polycrystalline and silicon

00:36:37.080 --> 00:36:39.880
for your LCDs or if you're
making amorphous silicon

00:36:39.880 --> 00:36:41.954
solar cells, they
use silane as well.

00:36:41.954 --> 00:36:43.370
So this little
truck here might go

00:36:43.370 --> 00:36:44.995
to three different
companies, depending

00:36:44.995 --> 00:36:47.730
on who's willing to pay more.

00:36:47.730 --> 00:36:50.840
Most of the silane is
used for polysilicon.

00:36:50.840 --> 00:36:53.700
The gas has to be converted
back into a solid,

00:36:53.700 --> 00:36:56.600
and that's where this
particular process here,

00:36:56.600 --> 00:36:58.290
the Siemens process is used.

00:36:58.290 --> 00:37:00.370
Again, you have
a current passing

00:37:00.370 --> 00:37:02.970
through some seed
material and the gas

00:37:02.970 --> 00:37:04.910
is being cracked onto that seed.

00:37:04.910 --> 00:37:06.320
You form these rods.

00:37:06.320 --> 00:37:08.170
The rods are then
cracked into chunks

00:37:08.170 --> 00:37:10.980
and then the chunks are
loaded into ingot crucibles.

00:37:10.980 --> 00:37:11.480
Yes.

00:37:11.480 --> 00:37:17.560
AUDIENCE: So the silane gas is
shipped as a gas in the trucks.

00:37:17.560 --> 00:37:18.466
PROFESSOR: Sure.

00:37:18.466 --> 00:37:20.410
AUDIENCE: Or on rails?

00:37:20.410 --> 00:37:23.550
PROFESSOR: Well it's
pyrophoric, as you

00:37:23.550 --> 00:37:26.140
can guess from just glancing
at this chemical structure

00:37:26.140 --> 00:37:26.640
right here.

00:37:26.640 --> 00:37:27.890
It's highly reactive.

00:37:27.890 --> 00:37:31.690
Pyrophoric means that it can
combust at room temperature.

00:37:31.690 --> 00:37:34.650
It can catch on fire, meaning
there are more stable compounds

00:37:34.650 --> 00:37:38.370
than this that can form when
you react this gas with air

00:37:38.370 --> 00:37:42.170
and, during the early days
of silane development,

00:37:42.170 --> 00:37:44.145
folks really didn't
know much about it

00:37:44.145 --> 00:37:45.520
and there's some
early research--

00:37:45.520 --> 00:37:47.853
some of the earliest research
done here at MIT, in fact.

00:37:47.853 --> 00:37:51.050
They would fill up an evacuated
chamber with silane gas

00:37:51.050 --> 00:37:53.350
and spark and
nothing would happen.

00:37:53.350 --> 00:37:55.100
Spark a second time,
nothing would happen.

00:37:55.100 --> 00:37:56.500
Spark a third time, boom.

00:37:56.500 --> 00:37:57.110
OK.

00:37:57.110 --> 00:37:58.330
That's critical limit.

00:37:58.330 --> 00:37:59.220
Such and such amount.

00:37:59.220 --> 00:38:00.890
You know, they'd keep
increasing the amount

00:38:00.890 --> 00:38:02.000
and finally it would go boom.

00:38:02.000 --> 00:38:03.150
Tell you what, lets
repeat the experiment

00:38:03.150 --> 00:38:04.310
since we're good scientists.

00:38:04.310 --> 00:38:06.434
They'd repeat it and, at
low concentrations, click.

00:38:06.434 --> 00:38:07.305
Boom.

00:38:07.305 --> 00:38:07.930
That's strange.

00:38:07.930 --> 00:38:09.260
That was much lower this time.

00:38:09.260 --> 00:38:10.880
Let's repeat the
experiment one more time.

00:38:10.880 --> 00:38:11.210
Click.

00:38:11.210 --> 00:38:11.550
Click.

00:38:11.550 --> 00:38:11.880
Click.

00:38:11.880 --> 00:38:12.180
Click.

00:38:12.180 --> 00:38:12.470
Click.

00:38:12.470 --> 00:38:12.730
Click.

00:38:12.730 --> 00:38:12.960
Click.

00:38:12.960 --> 00:38:13.180
Click.

00:38:13.180 --> 00:38:13.390
Click.

00:38:13.390 --> 00:38:13.921
Boom.

00:38:13.921 --> 00:38:14.420
All right.

00:38:14.420 --> 00:38:16.090
I don't really
understand this gas,

00:38:16.090 --> 00:38:18.830
but I'm going to say it's
really dangerous so I'm

00:38:18.830 --> 00:38:21.200
going to have little
warning bells that

00:38:21.200 --> 00:38:23.170
will detect the silane
gas if it's leaking

00:38:23.170 --> 00:38:25.560
and tell people to get the
heck out of the building

00:38:25.560 --> 00:38:29.180
if it starts being leaked.

00:38:29.180 --> 00:38:31.540
It's also toxic for
humans, by the way.

00:38:31.540 --> 00:38:34.940
Very small dilute
concentrations can kill you

00:38:34.940 --> 00:38:37.920
and so three buildings
on campus, only three

00:38:37.920 --> 00:38:40.770
to my knowledge, are set up
with the proper safety equipment

00:38:40.770 --> 00:38:42.900
to use silane gas
in the laboratory.

00:38:42.900 --> 00:38:45.590
Building 13, which is the
material science building,

00:38:45.590 --> 00:38:48.340
and then-- MTL and related.

00:38:48.340 --> 00:38:52.189
So we have this gas right here.

00:38:52.189 --> 00:38:52.980
Extremely powerful.

00:38:52.980 --> 00:38:55.980
There are variants thereof.

00:38:55.980 --> 00:39:00.020
You can replace some
of the hydrogens

00:39:00.020 --> 00:39:10.280
with chlorine, like this and
now you have trichlorosilane.

00:39:10.280 --> 00:39:11.930
It's all one word.

00:39:11.930 --> 00:39:15.330
So tricholorsilane, I've just
replaced three of my silanes--

00:39:15.330 --> 00:39:17.310
my hydrogens with
chlorine and now I

00:39:17.310 --> 00:39:19.850
have a different molecule,
still silicon bearing,

00:39:19.850 --> 00:39:22.480
still very reactive,
but now reactive

00:39:22.480 --> 00:39:25.820
at different temperatures
and I can modify my process

00:39:25.820 --> 00:39:29.310
by substituting out some of
the hydrogens for chlorines.

00:39:29.310 --> 00:39:33.000
So we have the silane gas or
trichlorosilane or the variants

00:39:33.000 --> 00:39:36.370
thereof, loaded into some
transportation vehicle that

00:39:36.370 --> 00:39:40.030
is very safe, leak-proof
and preventing accidents

00:39:40.030 --> 00:39:45.020
on the road, to deliver
it to where it is going

00:39:45.020 --> 00:39:50.510
to be consumed, which are
these so-called polysilicon,

00:39:50.510 --> 00:39:54.300
or Siemens reactor
as shown here.

00:39:54.300 --> 00:39:56.350
Excuse me.

00:39:56.350 --> 00:39:59.240
What happens, or how the
process actually flows,

00:39:59.240 --> 00:40:00.362
let me go back one step.

00:40:00.362 --> 00:40:02.820
We're going to start from up
at the very top of the process

00:40:02.820 --> 00:40:04.830
and move all the way
down, showing you

00:40:04.830 --> 00:40:07.520
what the manufacturing equipment
looks like at each step.

00:40:07.520 --> 00:40:09.730
So, this is the
distillation process

00:40:09.730 --> 00:40:11.960
used to create the
silane and when

00:40:11.960 --> 00:40:15.740
you see one of these factories
just think of a refinery.

00:40:15.740 --> 00:40:20.770
In fact, the people who don't
like this particular process

00:40:20.770 --> 00:40:24.170
who aren't a fan of the
silane refining process

00:40:24.170 --> 00:40:26.330
and opt for other ways of
purifying their silicon,

00:40:26.330 --> 00:40:28.830
liken this to an oil refinery.

00:40:28.830 --> 00:40:32.190
The imagery is very stark there.

00:40:32.190 --> 00:40:36.040
The polysilicon production,
this is the Siemens reactor,

00:40:36.040 --> 00:40:38.540
it's much smaller in comparison
to the metallurgical grade

00:40:38.540 --> 00:40:39.620
silicon furnace.

00:40:39.620 --> 00:40:41.970
Much smaller than the
carbothermic reduction furnace.

00:40:41.970 --> 00:40:44.850
Here, we have a small
human or human next

00:40:44.850 --> 00:40:47.170
to the small contraption.

00:40:47.170 --> 00:40:49.500
Here are a series of
them lined, almost

00:40:49.500 --> 00:40:52.720
like little pods and,
out of this material,

00:40:52.720 --> 00:40:54.750
actually inside of
the furnace, you

00:40:54.750 --> 00:41:00.940
have these rods that are
passing current and heating up

00:41:00.940 --> 00:41:04.920
and the silicon is
cracking onto the rods.

00:41:04.920 --> 00:41:07.220
So we wind up with
six nines, usually

00:41:07.220 --> 00:41:11.390
called 6N solar grade silicon
as a result of this process.

00:41:11.390 --> 00:41:16.180
We could also go up to, even,
nine nines using the Siemens

00:41:16.180 --> 00:41:16.750
process.

00:41:16.750 --> 00:41:18.510
It could be very,
very pure depending

00:41:18.510 --> 00:41:21.011
on how fast you grow, what the
purity of your silane gas is.

00:41:21.011 --> 00:41:21.635
AUDIENCE: Yeah.

00:41:21.635 --> 00:41:22.540
What is cracking.

00:41:22.540 --> 00:41:23.654
What does that mean?

00:41:23.654 --> 00:41:24.320
PROFESSOR: Sure.

00:41:24.320 --> 00:41:27.330
So what it means is
this gas molecule comes

00:41:27.330 --> 00:41:31.970
in, sees a solid surface,
the central atom right here,

00:41:31.970 --> 00:41:34.740
the silicon atom, gets
deposited onto the surface,

00:41:34.740 --> 00:41:37.550
becomes an adatom, which
means it's a surface atom,

00:41:37.550 --> 00:41:41.370
it's scuttling around and
the remaining elements

00:41:41.370 --> 00:41:44.215
within this molecule are then
free to move away as a gas.

00:41:44.215 --> 00:41:45.840
AUDIENCE: So you've
broken those bonds.

00:41:45.840 --> 00:41:46.464
PROFESSOR: Yes.

00:41:46.464 --> 00:41:49.600
Effectively, you've added
the core constituent

00:41:49.600 --> 00:41:53.740
of this molecule
onto the surface.

00:41:53.740 --> 00:41:56.580
It's joined the collective if
you will and, in this matter,

00:41:56.580 --> 00:41:59.530
the diameter of those
rods grows with time.

00:41:59.530 --> 00:42:01.480
So what I'm going
to do is pass around

00:42:01.480 --> 00:42:06.320
an example of a chunk coming
from this Siemens rod.

00:42:06.320 --> 00:42:07.960
Be very gentle with it please.

00:42:07.960 --> 00:42:10.400
On the outside you can
see a corrugated, rough,

00:42:10.400 --> 00:42:12.520
cauliflower-like structure.

00:42:12.520 --> 00:42:15.570
That's because you're
optimizing for deposition speed,

00:42:15.570 --> 00:42:17.120
not for beauty of the surface.

00:42:17.120 --> 00:42:18.770
You don't really
care how flat it

00:42:18.770 --> 00:42:20.990
is, unless you're trying to
grow a very specific type

00:42:20.990 --> 00:42:23.210
of material called flotsam,
which we get to the second,

00:42:23.210 --> 00:42:25.180
but in general, if you're trying
to crack it up and break it

00:42:25.180 --> 00:42:27.138
into a smaller piece and
into a chunk like this

00:42:27.138 --> 00:42:28.970
and throw it into a
big ingot furnace,

00:42:28.970 --> 00:42:31.178
it doesn't really matter
what the surface looks like.

00:42:31.178 --> 00:42:33.590
On the inside, it's
pretty dense silicon

00:42:33.590 --> 00:42:36.310
and, if look very carefully,
right in the middle there

00:42:36.310 --> 00:42:37.410
you can see the rod.

00:42:37.410 --> 00:42:38.910
The initial seeding rod.

00:42:38.910 --> 00:42:40.812
It's a slightly different color.

00:42:40.812 --> 00:42:42.770
So I'll pass these around
and please be gentle.

00:42:42.770 --> 00:42:45.270
AUDIENCE: Is the seeding
rod just silicon?

00:42:45.270 --> 00:42:46.900
PROFESSOR: It's
actually doped silicon,

00:42:46.900 --> 00:42:49.030
so it's lower
resistivity so you can

00:42:49.030 --> 00:42:52.330
pass more current through it.

00:42:52.330 --> 00:42:57.600
This here is chunks, or smaller
chunks of the polysilicon

00:42:57.600 --> 00:42:59.840
so, essentially, just
crushed polysilicon

00:42:59.840 --> 00:43:02.610
and if you're trying to load
a crucible with big chunks

00:43:02.610 --> 00:43:04.730
like this you'll leave
a lot of empty space

00:43:04.730 --> 00:43:07.660
unless you crush some of this up
and make finer grains out of it

00:43:07.660 --> 00:43:08.930
and fill in the gaps.

00:43:08.930 --> 00:43:11.120
So I'll pass these
around right here

00:43:11.120 --> 00:43:12.590
so you can have a look at them.

00:43:12.590 --> 00:43:16.500
Those are examples of the
Siemens grade polysilicon.

00:43:16.500 --> 00:43:19.940
This is a bigger rod.

00:43:19.940 --> 00:43:22.150
Here is the seed coming
right through the middle.

00:43:22.150 --> 00:43:23.650
Here's the surface
where you can see

00:43:23.650 --> 00:43:29.060
it's kind of rough
and corrugated

00:43:29.060 --> 00:43:32.570
and one of the biggest issues
with this feedstock refining

00:43:32.570 --> 00:43:34.580
process is that there
are very large plants

00:43:34.580 --> 00:43:36.070
and long lead times.

00:43:36.070 --> 00:43:38.460
This is a plant construction
going on right now,

00:43:38.460 --> 00:43:39.720
you can see.

00:43:39.720 --> 00:43:44.070
Typical lead times are
between 18 and 24 months.

00:43:44.070 --> 00:43:46.680
That's a long time between
when the board says yes, we

00:43:46.680 --> 00:43:50.360
will create new silicon
refining capacity and product

00:43:50.360 --> 00:43:52.040
starts to roll off
the production line

00:43:52.040 --> 00:43:53.360
and into customers' hands.

00:43:53.360 --> 00:43:55.660
It's a long time and
what this results in

00:43:55.660 --> 00:43:58.090
are drastic oversupply
and undersupply

00:43:58.090 --> 00:43:59.500
conditions in the market.

00:43:59.500 --> 00:44:01.310
So the silicon
feedstock price goes

00:44:01.310 --> 00:44:04.110
very high during periods
of undersupply and very low

00:44:04.110 --> 00:44:06.370
in periods of oversupply
and we're in an oversupply

00:44:06.370 --> 00:44:08.290
condition right now.

00:44:08.290 --> 00:44:11.530
Five years ago, let
me quantify this.

00:44:11.530 --> 00:44:13.830
Five years ago if you
went to the spot market--

00:44:13.830 --> 00:44:16.080
maybe four years ago-- if
you went to the spot market,

00:44:16.080 --> 00:44:20.480
you could pay $100 to $500
per kilogram of silicon.

00:44:20.480 --> 00:44:23.070
That material that was just
right there I bet one you

00:44:23.070 --> 00:44:25.900
would put it into your bag and
run away out the door right now

00:44:25.900 --> 00:44:28.590
and be able to go to Mexico.

00:44:28.590 --> 00:44:32.170
Now the polysilicon
prices are much,

00:44:32.170 --> 00:44:33.520
much lower on the spot market.

00:44:33.520 --> 00:44:36.830
Somewhere in the order of
$30 to $50 per kilogram.

00:44:36.830 --> 00:44:39.680
About an order of
magnitude lower.

00:44:39.680 --> 00:44:43.241
AUDIENCE: Isn't lower cost
silicon better for the PV

00:44:43.241 --> 00:44:43.950
industry, though?

00:44:43.950 --> 00:44:45.782
PROFESSOR: Is it better
for the PV industry?

00:44:45.782 --> 00:44:47.920
As a customer most definitely,
it is good for you.

00:44:47.920 --> 00:44:50.380
As an installer, it is most
definitely good for you.

00:44:50.380 --> 00:44:52.440
As a polysilicon
producer who wants

00:44:52.440 --> 00:44:54.440
to be a sustained
industry presence,

00:44:54.440 --> 00:44:55.530
it's not good for you.

00:44:55.530 --> 00:45:00.710
So this wide oscillation
between fat cat and scrawny

00:45:00.710 --> 00:45:03.220
is not very good
for any industry.

00:45:03.220 --> 00:45:06.450
It's unpredictable and it
causes some players to drop out.

00:45:06.450 --> 00:45:06.992
AUDIENCE: OK.

00:45:06.992 --> 00:45:09.241
PROFESSOR: And the investments
are very large as well.

00:45:09.241 --> 00:45:11.570
As you go from the early
stage portions of the value

00:45:11.570 --> 00:45:15.870
chain toward the module, the
investments generally decrease

00:45:15.870 --> 00:45:19.650
and so this is an outlook
coming from last year--

00:45:19.650 --> 00:45:22.040
the numbers are still a little
bit outdated-- polysilicon

00:45:22.040 --> 00:45:25.500
production is buttressing
up against 200,000

00:45:25.500 --> 00:45:28.900
metric tons per year at
this point in about 3/4

00:45:28.900 --> 00:45:30.340
to the PV industry.

00:45:30.340 --> 00:45:33.960
The cost of manufacturing
is between $20 and $25

00:45:33.960 --> 00:45:38.450
per kilogram and 2010 prices
were around $50 to $70.

00:45:38.450 --> 00:45:42.560
Now they're on
$30 to $50 in 2011

00:45:42.560 --> 00:45:46.150
and the 2008 prices were around
$500 per kilogram in the spot

00:45:46.150 --> 00:45:49.670
market and it really boils
down to the inability

00:45:49.670 --> 00:45:51.200
to adapt to demand.

00:45:51.200 --> 00:45:53.550
If you have a very
large contraption that

00:45:53.550 --> 00:45:55.932
produces the feedstock materials
and it takes a long time

00:45:55.932 --> 00:45:57.390
to build the
factories, you're just

00:45:57.390 --> 00:46:00.870
not going to be able
to adjust fast enough.

00:46:00.870 --> 00:46:03.700
Here's supply and demand,
demand being the red

00:46:03.700 --> 00:46:05.210
and supply being the blue.

00:46:05.210 --> 00:46:08.380
You can see how the oversupply--
the undersupply condition

00:46:08.380 --> 00:46:13.740
of the mid 2000s really led
to our current condition.

00:46:13.740 --> 00:46:17.300
So, alternatives to solar grade
silicon feedstock refining.

00:46:17.300 --> 00:46:20.270
What are some people thinking
in terms of other processes

00:46:20.270 --> 00:46:21.690
that they can use?

00:46:21.690 --> 00:46:26.300
These are two processes
right here and, mind you,

00:46:26.300 --> 00:46:30.180
when we were in this
situation with this price

00:46:30.180 --> 00:46:32.215
for the silicon,
everybody and anybody

00:46:32.215 --> 00:46:34.840
was coming up with new ideas of
how to manufacture the silicon.

00:46:34.840 --> 00:46:39.630
Now that we're barely selling
at cost and in an oversupply

00:46:39.630 --> 00:46:42.150
condition, many of these
ideas are having a struggle--

00:46:42.150 --> 00:46:43.320
a hard time in the market.

00:46:43.320 --> 00:46:44.936
They're struggling right now.

00:46:44.936 --> 00:46:47.310
So fluidized bed reactor and
upgraded metallurgical grade

00:46:47.310 --> 00:46:49.260
silicon.

00:46:49.260 --> 00:46:52.050
Let's talk about each
of those in turn.

00:46:52.050 --> 00:46:54.040
So what the fluidized
bed reactor folks

00:46:54.040 --> 00:46:57.550
realized was, gee, if
we're depositing on a rod,

00:46:57.550 --> 00:47:01.940
our surface area to volume
ratio is really large-- sorry,

00:47:01.940 --> 00:47:03.120
is really small.

00:47:03.120 --> 00:47:06.480
Our surface area to volume
ratio is going to be very small.

00:47:06.480 --> 00:47:07.650
So think of it this way.

00:47:07.650 --> 00:47:10.080
If we have a sphere,
a sphere would

00:47:10.080 --> 00:47:11.910
be the quintessential
example where

00:47:11.910 --> 00:47:15.880
we'd have a very large
surface area to volume ratio.

00:47:15.880 --> 00:47:18.270
If we had a plate, we
would have, as well,

00:47:18.270 --> 00:47:21.260
a very large surface
area to volume ratio

00:47:21.260 --> 00:47:26.130
and in the case of the condition
prior, where you have this rod,

00:47:26.130 --> 00:47:28.370
you really can't deposit
that quickly and so what

00:47:28.370 --> 00:47:30.390
these folks decided was,
what we're going to do

00:47:30.390 --> 00:47:35.290
is introduce small silicon
granules into this vessel,

00:47:35.290 --> 00:47:38.114
into this evacuated
chamber, and-- here's

00:47:38.114 --> 00:47:39.530
the evacuated
chamber right here--

00:47:39.530 --> 00:47:45.482
and we're going to flow silane
gas into the system right here

00:47:45.482 --> 00:47:46.940
and the smaller
particles are going

00:47:46.940 --> 00:47:49.550
to go higher up because
of this flow of gas

00:47:49.550 --> 00:47:52.820
coming in the bottom and
those will grow and eventually

00:47:52.820 --> 00:47:56.230
settle down down here where
we can extract the bottom.

00:47:56.230 --> 00:47:58.800
So we'll wind up with these
beautiful little silicon

00:47:58.800 --> 00:47:59.585
granules.

00:47:59.585 --> 00:48:01.710
These ones shown right
here, which I'll pass around

00:48:01.710 --> 00:48:04.480
as well, those are coming
from a fluidized bed reactor,

00:48:04.480 --> 00:48:08.020
and they're nice beautiful,
spherical granules that

00:48:08.020 --> 00:48:10.540
are grown a lot faster, I
mean, a lot more silicon

00:48:10.540 --> 00:48:14.240
is deposited per unit time than
through the Siemens process

00:48:14.240 --> 00:48:15.640
as shown there in the back.

00:48:15.640 --> 00:48:17.920
As a result, the energy
intensity is lower,

00:48:17.920 --> 00:48:20.610
the cost is lower, there's
a very tricky process

00:48:20.610 --> 00:48:22.586
to nail to get just
right, because you

00:48:22.586 --> 00:48:24.085
have to get the gas
flows right, you

00:48:24.085 --> 00:48:27.820
have to design the chamber
well, redo some purity contents.

00:48:27.820 --> 00:48:29.520
It's a tricky
process, and so this

00:48:29.520 --> 00:48:31.670
is being produced
right now, I believe,

00:48:31.670 --> 00:48:32.980
by only a few companies.

00:48:32.980 --> 00:48:35.510
REC has a capability
of doing it.

00:48:35.510 --> 00:48:37.930
MEMC, as well,
has the capability

00:48:37.930 --> 00:48:39.680
of doing this process.

00:48:39.680 --> 00:48:41.370
By and large, most
silicon is coming

00:48:41.370 --> 00:48:44.770
from the Siemens process.

00:48:44.770 --> 00:48:45.270
Yup.

00:48:45.270 --> 00:48:46.936
AUDIENCE: Sorry, both
of those companies

00:48:46.936 --> 00:48:48.590
have the normal
refining process?

00:48:48.590 --> 00:48:49.760
PROFESSOR: They have the
normal refining process.

00:48:49.760 --> 00:48:50.801
AUDIENCE: The [INAUDIBLE]

00:48:50.801 --> 00:48:53.640
PROFESSOR: Yup, and that's why
they developed this new one.

00:48:53.640 --> 00:48:55.600
They had these smaller,
internal projects

00:48:55.600 --> 00:48:57.220
that we're able to develop.

00:49:00.750 --> 00:49:01.810
So, yeah.

00:49:01.810 --> 00:49:03.734
I was just mentioning
the energy intensity.

00:49:03.734 --> 00:49:05.400
This is the kilowatt
hours per kilogram,

00:49:05.400 --> 00:49:07.690
going back to your question
about energy intensity.

00:49:07.690 --> 00:49:10.660
This is trichlorosilane
based Siemens process,

00:49:10.660 --> 00:49:12.600
silane based Siemens process.

00:49:12.600 --> 00:49:15.370
They're more ore less comparable
in terms of energy intensity.

00:49:15.370 --> 00:49:18.400
And the silance based
fluidized bed reactor process.

00:49:18.400 --> 00:49:20.180
According to
internal REC numbers,

00:49:20.180 --> 00:49:21.945
which are little rosy,
but never the less,

00:49:21.945 --> 00:49:23.420
the trend is correct here.

00:49:23.420 --> 00:49:25.560
It is lower somewhere
in the order

00:49:25.560 --> 00:49:29.080
of an order of magnitude
energy intensity,

00:49:29.080 --> 00:49:31.620
and cost is lower as well.

00:49:31.620 --> 00:49:36.330
So let's move away from
the silicon refining

00:49:36.330 --> 00:49:38.270
by distillation
process entirely.

00:49:38.270 --> 00:49:41.817
Let's leave gaseous
distillation aside and say,

00:49:41.817 --> 00:49:44.150
what if we were to take this
metallurgical-grade silicon

00:49:44.150 --> 00:49:47.330
and, through liquid
purification routes,

00:49:47.330 --> 00:49:49.450
result in high purity silicon.

00:49:49.450 --> 00:49:50.440
How would we do that?

00:49:50.440 --> 00:49:52.660
Well, if we turn to other
industries, the ones

00:49:52.660 --> 00:49:56.770
that smelter aluminum or
refine manganese and so forth,

00:49:56.770 --> 00:49:59.910
we would see a multitude
of different options

00:49:59.910 --> 00:50:01.070
that we could borrow.

00:50:01.070 --> 00:50:03.940
Slag refining, bleaching,
leaching solidification.

00:50:03.940 --> 00:50:06.780
Let me walk through
them one by one.

00:50:06.780 --> 00:50:09.430
Leaching-- that's
fairly straightforward.

00:50:09.430 --> 00:50:14.370
So if we put in some
acid, for instance,

00:50:14.370 --> 00:50:17.430
that dissolves the metals but
doesn't dissolve the silicon

00:50:17.430 --> 00:50:19.827
we could leach the metals
out of the material,

00:50:19.827 --> 00:50:21.410
and so that's the
essence of leaching.

00:50:21.410 --> 00:50:23.210
You might crush
up your material,

00:50:23.210 --> 00:50:27.270
in other ways other ways expose
the metals, or impurities,

00:50:27.270 --> 00:50:29.830
to the acids inside
of your system.

00:50:29.830 --> 00:50:31.820
Slag refining says,
gee, what if we

00:50:31.820 --> 00:50:34.210
were to introduce some
material that could

00:50:34.210 --> 00:50:36.240
absorb the metals into it?

00:50:36.240 --> 00:50:37.920
The solubility of
the metals would

00:50:37.920 --> 00:50:39.930
be higher inside
of the slag agent

00:50:39.930 --> 00:50:41.490
than inside of the liquid.

00:50:41.490 --> 00:50:45.210
Maybe we throw in calcium
oxide or yttrium oxide or some,

00:50:45.210 --> 00:50:47.810
usually it's a metal oxide
that has a very high melting

00:50:47.810 --> 00:50:51.910
temperature that remains a solid
or, at least a glassy solid,

00:50:51.910 --> 00:50:54.300
and we pour it on
top of our silicon

00:50:54.300 --> 00:50:57.730
and it's able to absorb, say,
the phosphorus or the boron

00:50:57.730 --> 00:51:00.920
that's inside of our silicon so
that we reduce impurity content

00:51:00.920 --> 00:51:02.480
and then we can add the
phosphorous and boron later

00:51:02.480 --> 00:51:04.146
intentionally, but
to the concentrations

00:51:04.146 --> 00:51:07.150
we want not to exuberantly
high concentrations that

00:51:07.150 --> 00:51:08.800
might be found in nature.

00:51:08.800 --> 00:51:11.580
Solidification-- during
this solidification process

00:51:11.580 --> 00:51:13.455
you're taking your
molten silicon

00:51:13.455 --> 00:51:15.080
and you're solidifying
it directionally

00:51:15.080 --> 00:51:18.356
from the bottom up and, because
the solubility of impurities

00:51:18.356 --> 00:51:20.730
tends to be larger in the
liquid than it is in the solid,

00:51:20.730 --> 00:51:23.140
it's like dragging a comb
through the entire material

00:51:23.140 --> 00:51:24.400
dragging out the impurities.

00:51:24.400 --> 00:51:28.610
Concentrating them in the liquid
and leaving a more pure silicon

00:51:28.610 --> 00:51:29.481
behind.

00:51:29.481 --> 00:51:30.980
Obviously, at the
very, very end you

00:51:30.980 --> 00:51:32.980
have this highly
concentrated region

00:51:32.980 --> 00:51:35.680
of impurities which then you
have to slice off and remove,

00:51:35.680 --> 00:51:37.350
so the solicitation
process doesn't

00:51:37.350 --> 00:51:39.089
come without a yield penalty.

00:51:39.089 --> 00:51:40.880
You still throw away
some of your material.

00:51:40.880 --> 00:51:43.244
So you can't repeat the
solidification over and over

00:51:43.244 --> 00:51:44.660
and over again, I
guess you could,

00:51:44.660 --> 00:51:47.140
but you'd be losing
material every step.

00:51:47.140 --> 00:51:49.720
So some combination of these
processes here, and others.

00:51:49.720 --> 00:51:51.450
Other trickery.

00:51:51.450 --> 00:51:55.910
Low temperature eutectic
formation with other elements,

00:51:55.910 --> 00:51:57.230
for example.

00:51:57.230 --> 00:52:00.260
Some combination of this is
used to refine the silicon

00:52:00.260 --> 00:52:02.000
without creating
a gas out of it.

00:52:02.000 --> 00:52:03.830
So wafer fabrication.

00:52:03.830 --> 00:52:05.407
We're now going from
feedstock, we're

00:52:05.407 --> 00:52:06.990
leaving feedstocking
behind, and we're

00:52:06.990 --> 00:52:08.760
going to be talking about how
do you go from the feedstock

00:52:08.760 --> 00:52:10.718
materials that are being
passed around the room

00:52:10.718 --> 00:52:14.010
right now into a wafer that
you can then manufacture

00:52:14.010 --> 00:52:15.330
a solar cell device out of?

00:52:15.330 --> 00:52:16.890
One of these for instance.

00:52:16.890 --> 00:52:20.280
So let's talk about wafer
fabrication right here.

00:52:20.280 --> 00:52:21.980
So again, just to
situate ourselves,

00:52:21.980 --> 00:52:24.730
we've gone from raw materials
to silicon feedstock

00:52:24.730 --> 00:52:27.250
and now we're going to
feedstocks to wafers.

00:52:27.250 --> 00:52:30.240
Any questions right now
before we dive into that?

00:52:30.240 --> 00:52:30.740
Yeah.

00:52:30.740 --> 00:52:32.470
AUDIENCE: Question about supply.

00:52:32.470 --> 00:52:35.552
So silicon is very abundant
but the high purity

00:52:35.552 --> 00:52:38.900
silica deposits-- are
they really abundant too?

00:52:38.900 --> 00:52:40.180
PROFESSOR: Great question.

00:52:40.180 --> 00:52:43.010
So the question was are the
high purity silica deposits

00:52:43.010 --> 00:52:45.400
as abundant as, say, silicon.

00:52:45.400 --> 00:52:45.900
Certainly.

00:52:45.900 --> 00:52:49.120
If you bend over and rub your
fingers against the ground

00:52:49.120 --> 00:52:52.080
you're probably going to
come up with, probably,

00:52:52.080 --> 00:52:55.930
millions of trillions of silicon
atoms in your fingernails.

00:52:55.930 --> 00:52:57.620
Those are not very purity.

00:52:57.620 --> 00:53:01.260
So the highest security
quartz deposits are more rare

00:53:01.260 --> 00:53:03.620
and they are sought after,
and so they're are known.

00:53:03.620 --> 00:53:04.950
Their locations are known.

00:53:04.950 --> 00:53:07.170
There's one specific
one in Norway,

00:53:07.170 --> 00:53:09.310
one specific one
in North Carolina,

00:53:09.310 --> 00:53:13.150
and so forth around the
world and there-- in a sense,

00:53:13.150 --> 00:53:14.140
they go to places.

00:53:14.140 --> 00:53:16.830
People have adjusted their
metallurgical-grade silicon

00:53:16.830 --> 00:53:18.750
refineries and their
subsequent down process

00:53:18.750 --> 00:53:19.922
for that particular ore.

00:53:19.922 --> 00:53:22.130
Once you run out of it, it's
not that the world ends,

00:53:22.130 --> 00:53:26.580
we just have to adjust for
the next feedstock source.

00:53:26.580 --> 00:53:29.640
So, in principle, there
are people looking

00:53:29.640 --> 00:53:31.080
at a variety of silicon inputs.

00:53:31.080 --> 00:53:35.580
Anything from the dirtier,
compressed, metamorphic quartz

00:53:35.580 --> 00:53:36.600
that I mentioned.

00:53:36.600 --> 00:53:39.390
Some people looking
at rice husks,

00:53:39.390 --> 00:53:41.206
which are silica rich as well.

00:53:41.206 --> 00:53:42.580
Other people
looking at seashells

00:53:42.580 --> 00:53:44.996
which, mostly calcium carbonate,
but other things as well.

00:53:44.996 --> 00:53:47.860
I mean, there was a wide range.

00:53:47.860 --> 00:53:50.690
When the price of silicon
was $500 per kilogram,

00:53:50.690 --> 00:53:52.520
you got a multitude of ideas.

00:53:52.520 --> 00:53:54.850
When the price comes
back down, people

00:53:54.850 --> 00:53:57.744
tend to be more conservative.

00:53:57.744 --> 00:54:02.100
AUDIENCE: Is silicon considered
a renewable resource?

00:54:02.100 --> 00:54:04.389
PROFESSOR: Is silicon
considered a renewable resource.

00:54:04.389 --> 00:54:06.180
It is not a renewable
resource in the sense

00:54:06.180 --> 00:54:09.150
that, once you mine
it from the ground,

00:54:09.150 --> 00:54:11.650
you've mined it from the ground
and you used in some way.

00:54:11.650 --> 00:54:14.740
The reason it's
considered not an issue

00:54:14.740 --> 00:54:16.920
is because there's
so much of it.

00:54:16.920 --> 00:54:20.610
Not all of it, though, is
in the easy to access form.

00:54:20.610 --> 00:54:21.110
Right?

00:54:21.110 --> 00:54:22.651
Some of the silicon
might be bound up

00:54:22.651 --> 00:54:26.200
within heavily
contaminated sources

00:54:26.200 --> 00:54:30.800
and that's where the refining
ingenuity comes into play.

00:54:30.800 --> 00:54:34.210
As long as prices remain low,
there's not too much interest,

00:54:34.210 --> 00:54:36.100
say, for example,
that mine in Peru

00:54:36.100 --> 00:54:42.780
that has titanium oxide needles
throughout their silicon

00:54:42.780 --> 00:54:44.960
because why would
you want heavily

00:54:44.960 --> 00:54:47.070
titanium contaminated silicon?

00:54:47.070 --> 00:54:50.030
But as the price of
silicon, it probably will,

00:54:50.030 --> 00:54:52.800
rise again then people might
take another look at that mine

00:54:52.800 --> 00:54:55.380
and say gee, how can we
phase separate the rutile

00:54:55.380 --> 00:54:57.240
and anatase from
the quartz early

00:54:57.240 --> 00:54:59.110
on in the process by
crushing and etching

00:54:59.110 --> 00:55:04.050
or something so we can access
this feedstock material.

00:55:04.050 --> 00:55:06.090
We'll see.

00:55:06.090 --> 00:55:08.950
It really depends on how the
market evolves, where people

00:55:08.950 --> 00:55:11.385
go looking for their silicon,
but there's a lot of it

00:55:11.385 --> 00:55:12.260
in the earth's crust.

00:55:12.260 --> 00:55:14.936
AUDIENCE: You're not
concerned about silicon?

00:55:14.936 --> 00:55:15.830
PROFESSOR: No.

00:55:15.830 --> 00:55:16.850
Nope.

00:55:16.850 --> 00:55:20.280
What is a bigger bottleneck
are are the refining steps

00:55:20.280 --> 00:55:21.090
in between.

00:55:21.090 --> 00:55:23.854
First it was the
reactors and soon it's

00:55:23.854 --> 00:55:26.020
probably going to be the
metallurgical-grade silicon

00:55:26.020 --> 00:55:28.950
reactors as well.

00:55:28.950 --> 00:55:29.600
All right.

00:55:29.600 --> 00:55:30.840
Wafers.

00:55:30.840 --> 00:55:33.020
How do we get to these
from the raw feedstock

00:55:33.020 --> 00:55:36.260
materials that are being passed
around the room right now?

00:55:36.260 --> 00:55:38.741
So single crystalline
silicon ingot growth.

00:55:38.741 --> 00:55:39.990
Let's walk through that first.

00:55:39.990 --> 00:55:42.160
How do we get these
beautiful ingots?

00:55:42.160 --> 00:55:45.910
They're about half of all
silicon market right now.

00:55:45.910 --> 00:55:48.250
The biggest growth
method, by far,

00:55:48.250 --> 00:55:50.530
is called Czochralski
growth and, named

00:55:50.530 --> 00:55:54.920
after the Polish physicist
there Jan Czochralski.

00:55:54.920 --> 00:55:58.320
What you do is you have
a bath of molten silicon.

00:55:58.320 --> 00:55:59.840
A crucible, if you will.

00:55:59.840 --> 00:56:01.480
This tends to be a
circular crucible,

00:56:01.480 --> 00:56:03.150
rounded at the
bottom, usually made

00:56:03.150 --> 00:56:05.195
of quartz with
heaters on the outside

00:56:05.195 --> 00:56:07.070
to heat up the molten silicon.

00:56:07.070 --> 00:56:09.180
To heat up the silicon
chunks in here.

00:56:09.180 --> 00:56:10.640
Once everything
is molten, looking

00:56:10.640 --> 00:56:12.370
like a big bathtub
of silicon, you

00:56:12.370 --> 00:56:15.890
introduce a small
crystalline silicon seed

00:56:15.890 --> 00:56:18.450
into that molten
silicon and then

00:56:18.450 --> 00:56:20.880
you begin pulling while
rotating that seed.

00:56:20.880 --> 00:56:22.992
So the seed is a
single crystal material

00:56:22.992 --> 00:56:24.950
and what ends up happening
is, as you introduce

00:56:24.950 --> 00:56:27.020
the seed into the material
and begin pulling,

00:56:27.020 --> 00:56:29.000
you start pulling
out this crystal.

00:56:29.000 --> 00:56:30.900
Single crystalline crystal.

00:56:30.900 --> 00:56:33.850
It's a thing of beauty
and this seed is actually

00:56:33.850 --> 00:56:35.790
very, very narrow in diameter.

00:56:35.790 --> 00:56:38.370
It might be about that big
around so pretty narrow

00:56:38.370 --> 00:56:40.830
in diameter and it's
being able to support

00:56:40.830 --> 00:56:43.870
this ingot of a
few, usually a few,

00:56:43.870 --> 00:56:46.310
tens to hundreds of kilograms
of mass underneath it

00:56:46.310 --> 00:56:49.790
and that's because silicon is
very strong even though it's

00:56:49.790 --> 00:56:51.260
brittle.

00:56:51.260 --> 00:56:55.100
So if you weren't to apply,
say, for example, a shear

00:56:55.100 --> 00:56:57.850
force on your silicon but
just to apply an axial load,

00:56:57.850 --> 00:57:00.570
you could support a very,
large weight underneath it.

00:57:00.570 --> 00:57:04.140
So the [INAUDIBLE] of silicon
is grown from the bottom

00:57:04.140 --> 00:57:06.640
and eventually you wind
up with this nice ingot,

00:57:06.640 --> 00:57:08.290
as shown right there.

00:57:08.290 --> 00:57:12.690
The art that goes into growing
this properly is amazing.

00:57:12.690 --> 00:57:15.770
I'll highlight it with one
small little example just

00:57:15.770 --> 00:57:18.630
to illustrate the bigger
picture that a lot of effort

00:57:18.630 --> 00:57:21.340
goes into making these
defect-free, quote unquote,

00:57:21.340 --> 00:57:22.487
defect-free crystals.

00:57:22.487 --> 00:57:24.820
They're called defect-free
because they contain no grain

00:57:24.820 --> 00:57:27.080
boundaries and no dislocations.

00:57:27.080 --> 00:57:30.490
They have impurities, they
have intrinsic point defects,

00:57:30.490 --> 00:57:32.394
meaning vacancies or
interstitial atoms,

00:57:32.394 --> 00:57:34.560
but they don't have grain
boundaries or dislocations

00:57:34.560 --> 00:57:37.930
and so they're called
defect-free silicon.

00:57:37.930 --> 00:57:40.380
You introduce that seed
down into the liquid melt.

00:57:40.380 --> 00:57:41.370
Thermal stress happens.

00:57:41.370 --> 00:57:41.870
Right?

00:57:41.870 --> 00:57:47.690
Because you have the shock
between the solid silicon seed

00:57:47.690 --> 00:57:49.550
encountering the liquid
for the first time.

00:57:49.550 --> 00:57:51.040
So this locations [? form ?].

00:57:51.040 --> 00:57:54.040
And you have to pull the
seed out in such a way,

00:57:54.040 --> 00:57:57.040
you slowly rotate and
make this shoulder.

00:57:57.040 --> 00:57:59.130
The shoulder has to be
as quick as possible

00:57:59.130 --> 00:58:01.960
because you don't want
to waste material.

00:58:01.960 --> 00:58:05.140
Everything inside this shoulder
right here gets thrown away.

00:58:05.140 --> 00:58:07.920
So that little piece of material
right there gets tossed out.

00:58:07.920 --> 00:58:10.800
So you want to make the
shoulders as narrow and as

00:58:10.800 --> 00:58:12.450
quick as possible
so you can utilize

00:58:12.450 --> 00:58:14.140
the majority of your ingot
but, at the same time,

00:58:14.140 --> 00:58:15.690
you have to make
it thick enough so

00:58:15.690 --> 00:58:17.850
that the dislocations
can move all the way

00:58:17.850 --> 00:58:19.920
and propagate all the
way to the outside

00:58:19.920 --> 00:58:21.960
and end and terminate
in the shoulder

00:58:21.960 --> 00:58:24.420
before propagating
into the crystal.

00:58:24.420 --> 00:58:26.849
So that's just one
example of the technology

00:58:26.849 --> 00:58:28.140
that goes in the growing these.

00:58:28.140 --> 00:58:30.660
Another might be, gee,
we're PV industry,

00:58:30.660 --> 00:58:34.220
we want to make the stuff fast
whereas, in the IC industry

00:58:34.220 --> 00:58:38.060
you can invest up to a few of
dollars per gram of silicon

00:58:38.060 --> 00:58:40.110
and still make a profit
because you're selling

00:58:40.110 --> 00:58:41.660
a computer at 1,000 bucks.

00:58:41.660 --> 00:58:43.500
In the PV industry,
we can invest,

00:58:43.500 --> 00:58:46.122
at most, a few tens of
cents per gram of silicon.

00:58:46.122 --> 00:58:47.580
So we have to make
this stuff fast.

00:58:47.580 --> 00:58:48.617
We can't dilly dally.

00:58:48.617 --> 00:58:50.450
You might want to crank
up the growth speed,

00:58:50.450 --> 00:58:52.480
then you run into issues
with defect concentrations,

00:58:52.480 --> 00:58:54.080
intrinsic point
defect concentrations,

00:58:54.080 --> 00:58:55.410
during the growth.

00:58:55.410 --> 00:58:57.370
I'm illustrating this
just to highlight

00:58:57.370 --> 00:59:01.820
the complexity of the growth
process of making these ingots,

00:59:01.820 --> 00:59:04.560
and the latter example was
one that the PV industry

00:59:04.560 --> 00:59:05.460
is facing today.

00:59:05.460 --> 00:59:06.925
It's actually a
hot research topic.

00:59:06.925 --> 00:59:07.425
Yeah.

00:59:07.425 --> 00:59:08.091
And then Ashley.

00:59:08.091 --> 00:59:11.074
AUDIENCE: The rotation speed
does that just affect time,

00:59:11.074 --> 00:59:12.490
or does it affect other things?

00:59:12.490 --> 00:59:15.036
PROFESSOR: So it affects
a multitude of things.

00:59:15.036 --> 00:59:16.410
One of the things
that it affects

00:59:16.410 --> 00:59:19.660
is the flow of, the
convective flow, of the melt.

00:59:19.660 --> 00:59:21.540
So the liquid flow
inside of the melt

00:59:21.540 --> 00:59:24.470
is, in part, determining
how much oxygen

00:59:24.470 --> 00:59:26.430
gets transported
from this crucible

00:59:26.430 --> 00:59:28.360
here into the growing crystal.

00:59:28.360 --> 00:59:31.290
If you manage to suppress that
convective flow in the melt,

00:59:31.290 --> 00:59:33.797
you will also suppress
oxygen transport

00:59:33.797 --> 00:59:35.630
since the fusion is
going to be a lot slower

00:59:35.630 --> 00:59:40.920
than turbulent transport
or [INAUDIBLE] transport

00:59:40.920 --> 00:59:42.811
or convective transport.

00:59:42.811 --> 00:59:43.310
Yeah.

00:59:43.310 --> 00:59:43.600
Question?

00:59:43.600 --> 00:59:44.860
AUDIENCE: I just
have two questions

00:59:44.860 --> 00:59:46.510
so one is how fast
do you rotate it

00:59:46.510 --> 00:59:49.430
and the other is what that
does control-- the diameter

00:59:49.430 --> 00:59:51.286
because I've heard
of 12 inch wafers

00:59:51.286 --> 00:59:53.150
versus like 18 inch wafers.

00:59:53.150 --> 00:59:53.980
PROFESSOR: Sure.

00:59:53.980 --> 00:59:56.210
So one of the things
that controls diameter

00:59:56.210 --> 01:00:00.750
is the balance of
heat extraction.

01:00:00.750 --> 01:00:03.160
So if you cool something down,
especially molten silicon,

01:00:03.160 --> 01:00:04.560
it will freeze, it will grow.

01:00:04.560 --> 01:00:06.330
If you heat it up,
it will shrink.

01:00:06.330 --> 01:00:08.930
So that's one of the components
that controls the diameter.

01:00:08.930 --> 01:00:12.850
The pull speed and how
you grow that shoulder,

01:00:12.850 --> 01:00:14.700
essentially how you
heat up the material

01:00:14.700 --> 01:00:17.020
and how fast you pull
at those initial stages,

01:00:17.020 --> 01:00:18.440
also dictates the
diameter and you

01:00:18.440 --> 01:00:20.110
can see in the
ingots themselves,

01:00:20.110 --> 01:00:21.140
they're not perfect.

01:00:21.140 --> 01:00:23.260
They have a little
bit of corregation

01:00:23.260 --> 01:00:27.010
and that's the fluctuations of
the temperature of the melt,

01:00:27.010 --> 01:00:28.710
fluctuations of
the heater output,

01:00:28.710 --> 01:00:32.540
fluctuations of pull
speed, maybe what's pulling

01:00:32.540 --> 01:00:35.760
this entire contraption
is kind of a stepper motor

01:00:35.760 --> 01:00:38.500
that has a certain
granularity to it.

01:00:38.500 --> 01:00:40.770
Results in corregated edges.

01:00:40.770 --> 01:00:42.530
It's not perfect
and so there will

01:00:42.530 --> 01:00:47.640
be some adjustment made to
the form factor of the edge

01:00:47.640 --> 01:00:50.910
to get this nice round
wafer at the end of the day.

01:00:50.910 --> 01:00:52.910
AUDIENCE: Does the
seed rod [INAUDIBLE]

01:00:52.910 --> 01:00:54.965
all the way to the ingot
or just near the top?

01:00:54.965 --> 01:00:55.840
PROFESSOR: All right.

01:00:55.840 --> 01:00:57.540
So the entire ingot
becomes pattern

01:00:57.540 --> 01:00:59.200
or templated by the seed rod.

01:00:59.200 --> 01:01:01.750
So this entire ingot has the
same crystalline orientation

01:01:01.750 --> 01:01:02.673
as a seed.

01:01:02.673 --> 01:01:07.650
AUDIENCE: And is the seed doped
differently than the silicon?

01:01:07.650 --> 01:01:09.490
PROFESSOR: It might
be but I'm not aware

01:01:09.490 --> 01:01:11.600
that that affects
the overall process.

01:01:11.600 --> 01:01:13.930
It could be that it's one
of the critical pieces

01:01:13.930 --> 01:01:16.263
of the magic sauce that makes
it work but I'm not aware.

01:01:19.730 --> 01:01:20.530
Rotation speed.

01:01:20.530 --> 01:01:23.620
It's not rotating like
this it's a slow rotation

01:01:23.620 --> 01:01:26.652
so I would-- let's see.

01:01:26.652 --> 01:01:27.860
How many radians per second--

01:01:27.860 --> 01:01:28.910
AUDIENCE: Can you see it?

01:01:28.910 --> 01:01:30.185
PROFESSOR: You can
visually see it

01:01:30.185 --> 01:01:31.518
if you looked at it long enough.

01:01:31.518 --> 01:01:32.100
Yeah.

01:01:32.100 --> 01:01:34.330
Yeah.

01:01:34.330 --> 01:01:36.540
So one modification,
one variant,

01:01:36.540 --> 01:01:38.760
of the single
crystalline growth method

01:01:38.760 --> 01:01:40.100
is called float-zone growth.

01:01:40.100 --> 01:01:44.100
You take a rod of poly, much
like that right over there

01:01:44.100 --> 01:01:49.720
that's inside of here,
and you pass an RF coil,

01:01:49.720 --> 01:01:52.220
radio frequency
coil, next to the rod

01:01:52.220 --> 01:01:54.140
and what that does is,
essentially, heats up

01:01:54.140 --> 01:01:56.870
the silicon, if it's
doped highly enough.

01:01:56.870 --> 01:01:59.600
It will melt the
silicon locally.

01:01:59.600 --> 01:02:03.230
Folks have probably heard
of fancy high-end stoves

01:02:03.230 --> 01:02:07.130
that we can only probably hope
to afford in 10 or 15 years,

01:02:07.130 --> 01:02:11.560
but these stoves that
are inductive heaters.

01:02:11.560 --> 01:02:12.060
Right?

01:02:12.060 --> 01:02:15.334
They're not resistive
heating elements,

01:02:15.334 --> 01:02:17.250
they're inductive heating
elements and the way

01:02:17.250 --> 01:02:21.860
that works is you have a radio
frequency source that then

01:02:21.860 --> 01:02:25.870
is absorbed by, in the
case of the RF heater,

01:02:25.870 --> 01:02:28.630
I believe it's a
specific type of iron

01:02:28.630 --> 01:02:34.580
that the inductive
heating ovens need.

01:02:34.580 --> 01:02:38.160
And so this RF coil here
is emitting energy, which

01:02:38.160 --> 01:02:40.440
is absorbed by the
silicon and melting it,

01:02:40.440 --> 01:02:44.100
and you start with the
polycrystalline rod coming

01:02:44.100 --> 01:02:47.210
from the Siemens process
and in that case,

01:02:47.210 --> 01:02:51.540
this rough, corrugated
material right here won't do.

01:02:51.540 --> 01:02:52.070
Right?

01:02:52.070 --> 01:02:54.360
This is too rough
for that RF coil

01:02:54.360 --> 01:02:57.620
to pass over and be a
consistent distance away.

01:02:57.620 --> 01:02:59.070
In the case of
float-zone growth,

01:02:59.070 --> 01:03:02.687
you actually have to modify your
polysilicon production process.

01:03:02.687 --> 01:03:04.270
You have to modify
the Siemens process

01:03:04.270 --> 01:03:08.030
so that you get a
nice smooth rod, which

01:03:08.030 --> 01:03:11.080
you can then pass the
RF coil next to and melt

01:03:11.080 --> 01:03:13.360
and you again start with
a seed at the bottom,

01:03:13.360 --> 01:03:16.600
your RF coil starts down here
and then the RF coil moves

01:03:16.600 --> 01:03:18.950
through the material, almost
like a comb from the bottom

01:03:18.950 --> 01:03:21.700
to the top, converting
the polysilicon

01:03:21.700 --> 01:03:27.240
into nice single crystalline
material and, in the process,

01:03:27.240 --> 01:03:29.560
it concentrates impurities
in this liquid region.

01:03:29.560 --> 01:03:31.893
Since the liquids have a
higher solubility in the liquid

01:03:31.893 --> 01:03:34.440
than they do in the solid,
the impurities are then

01:03:34.440 --> 01:03:37.550
aggregated inside of the liquid
region and, again, like a comb,

01:03:37.550 --> 01:03:40.050
they just get swept
out of the material.

01:03:40.050 --> 01:03:41.810
Not all of them, but
a large percentage

01:03:41.810 --> 01:03:43.510
of them, and so you
can make multiple

01:03:43.510 --> 01:03:46.605
passes with this RF coil
to further concentrate

01:03:46.605 --> 01:03:48.490
the impurities and
the extremities

01:03:48.490 --> 01:03:50.990
and remove them
from the material.

01:03:50.990 --> 01:03:52.240
So that's a float-zone method.

01:03:52.240 --> 01:03:55.609
Very expensive material,
very high purity.

01:03:55.609 --> 01:03:57.150
One of the reasons
it has high purity

01:03:57.150 --> 01:04:00.310
is because you don't have
this quartz crucible nearby,

01:04:00.310 --> 01:04:01.760
you don't have
this molten silicon

01:04:01.760 --> 01:04:03.990
that's absorbing or
dissolving the quartz

01:04:03.990 --> 01:04:06.920
and transporting the
oxygen into your crystal.

01:04:06.920 --> 01:04:09.100
You have much lower carbon
and oxygen concentrations

01:04:09.100 --> 01:04:10.590
to [INAUDIBLE]
float-zone material.

01:04:10.590 --> 01:04:13.360
So if anybody is doing
experiments with silicon,

01:04:13.360 --> 01:04:16.090
for whatever reason, using
it as a substrate material,

01:04:16.090 --> 01:04:19.190
you want to think carefully
about what type of silicon you

01:04:19.190 --> 01:04:21.490
source and from
where you source it.

01:04:21.490 --> 01:04:23.792
You can find some very
poor quality silicon

01:04:23.792 --> 01:04:25.250
out there in the
market, especially

01:04:25.250 --> 01:04:26.930
if you going into
the aftersale market,

01:04:26.930 --> 01:04:28.764
and we know this from some--

01:04:28.764 --> 01:04:29.680
AUDIENCE: [INAUDIBLE].

01:04:29.680 --> 01:04:32.600
PROFESSOR: --very
painful experiences.

01:04:32.600 --> 01:04:35.055
And so there are
some better sources

01:04:35.055 --> 01:04:36.680
from which to get
your wafers and we're

01:04:36.680 --> 01:04:39.670
happy to talk
about that offline.

01:04:39.670 --> 01:04:42.019
So, again, single
crystalline silicon.

01:04:42.019 --> 01:04:43.560
We're going to
venture into the world

01:04:43.560 --> 01:04:46.240
of multicrystalline silicon
ever so briefly here.

01:04:46.240 --> 01:04:48.050
First, we'll start
about cast material

01:04:48.050 --> 01:04:50.200
and, just to emphasize
here, we have

01:04:50.200 --> 01:04:51.820
regions of crystalline
material that

01:04:51.820 --> 01:04:54.737
have grain boundaries
separating the adjacent grains

01:04:54.737 --> 01:04:56.820
and the reason we go into
multicrystalline silicon

01:04:56.820 --> 01:05:00.110
is really oftentimes, it
is a lower cost method

01:05:00.110 --> 01:05:03.221
of producing a silicon wafer
although you have the grain

01:05:03.221 --> 01:05:03.720
boundaries.

01:05:03.720 --> 01:05:06.980
So, again, single crystalline,
Czochralski and float-zone, you

01:05:06.980 --> 01:05:10.030
wind up with round
wafers, typically

01:05:10.030 --> 01:05:12.880
single crystalline variety, and
multicrystalline silicon wafers

01:05:12.880 --> 01:05:19.680
tend to be more square-like
and more visibly multi-grained,

01:05:19.680 --> 01:05:21.180
if you will.

01:05:21.180 --> 01:05:22.860
So let's talk about
those for a minute.

01:05:22.860 --> 01:05:26.880
How do you make a
multicrystalline silicon wafer?

01:05:26.880 --> 01:05:31.000
Again, you would start with
the solar-grade silicon that

01:05:31.000 --> 01:05:33.150
could either be coming
from the Siemens process,

01:05:33.150 --> 01:05:35.233
it could be coming from
the fluidized bed reactor,

01:05:35.233 --> 01:05:36.710
it could be coming
from an upgraded

01:05:36.710 --> 01:05:38.876
metallurgical-grade silicon,
the liquid purification

01:05:38.876 --> 01:05:41.870
route but, somehow, some way,
you get chunks of silicon,

01:05:41.870 --> 01:05:44.292
or granules of silicon, that
have a high enough purity

01:05:44.292 --> 01:05:45.750
for you to make
solar cells out of,

01:05:45.750 --> 01:05:47.270
and high enough
purity is typically

01:05:47.270 --> 01:05:50.510
in the order of one part per
million impurity content.

01:05:50.510 --> 01:05:53.500
So you put your solar-grade
silicon into a crucible

01:05:53.500 --> 01:05:55.750
and then you melt the
silicon inside of it.

01:05:55.750 --> 01:05:58.522
Silicon melts at
1,414 degrees Celsius.

01:05:58.522 --> 01:05:59.730
It's a very high temperature.

01:05:59.730 --> 01:06:04.600
So 1,414 degrees Celsius is the
melting temperature of silicon.

01:06:04.600 --> 01:06:05.940
And then it's cooled.

01:06:05.940 --> 01:06:09.140
Not just randomly, but from
the bottom up and the reason

01:06:09.140 --> 01:06:11.790
it's cooled from the
bottom up is because,

01:06:11.790 --> 01:06:15.130
and here I guess you'll actually
have to come up and see this

01:06:15.130 --> 01:06:18.200
after class, it's rather
difficult to see from here,

01:06:18.200 --> 01:06:20.447
but this is a cross
section of a small ingot.

01:06:20.447 --> 01:06:22.030
This is the outside
of the ingot where

01:06:22.030 --> 01:06:25.210
it was contacting the
wall, these little pieces

01:06:25.210 --> 01:06:27.050
of white stuff that
are flaking off,

01:06:27.050 --> 01:06:32.290
this is the fused quartz silica
that forms the crucible wall,

01:06:32.290 --> 01:06:34.770
and the silicon
nitride coating that

01:06:34.770 --> 01:06:37.000
form the anti-stick
coating that prevented

01:06:37.000 --> 01:06:39.484
the silicon from
sticking to the crucible,

01:06:39.484 --> 01:06:41.150
and so it's kind of
rough and corrugated

01:06:41.150 --> 01:06:44.210
but, if we were to rotate this
around and look at the inside,

01:06:44.210 --> 01:06:47.269
this here is a cross
section of the actual ingot

01:06:47.269 --> 01:06:49.060
from the inside and,
if you look carefully,

01:06:49.060 --> 01:06:51.170
you'll see grains growing
from the bottom to the top.

01:06:51.170 --> 01:06:52.760
You probably can't
see them from here,

01:06:52.760 --> 01:06:54.450
you'll have to come up
after class and take a look,

01:06:54.450 --> 01:06:57.070
but the grains are growing
from the bottom to the top

01:06:57.070 --> 01:06:59.880
and that is called directional
solidification, or the result

01:06:59.880 --> 01:07:01.130
of directional solidification.

01:07:01.130 --> 01:07:02.130
Directional
solidification is when

01:07:02.130 --> 01:07:03.870
you solidify from
the bottom to the top

01:07:03.870 --> 01:07:05.230
and, typically, your
grain boundaries

01:07:05.230 --> 01:07:06.771
are going to be
running perpendicular

01:07:06.771 --> 01:07:09.300
to the solid-liquid interface,
so your grain boundaries

01:07:09.300 --> 01:07:12.010
will be running up like this
as you grow your material

01:07:12.010 --> 01:07:13.660
from the bottom to the top.

01:07:13.660 --> 01:07:16.660
If you were to do
uncontrolled solidification

01:07:16.660 --> 01:07:18.457
and all walls would
freeze the same time,

01:07:18.457 --> 01:07:20.290
you'd have grains growing
in from the sides,

01:07:20.290 --> 01:07:22.248
you'd have grains growing
in through the bottom

01:07:22.248 --> 01:07:24.740
and then, when you slice
your wafer out horizontally,

01:07:24.740 --> 01:07:26.990
the grain boundaries wouldn't
be running perpendicular

01:07:26.990 --> 01:07:27.710
to the surface.

01:07:27.710 --> 01:07:30.260
They might be running parallel
to the surface, in which case

01:07:30.260 --> 01:07:32.610
they could wreck havoc on your
minority care diffusion length.

01:07:32.610 --> 01:07:34.155
Imagine you being
an electron having

01:07:34.155 --> 01:07:37.600
to travel across that grain
boundary that's between you

01:07:37.600 --> 01:07:39.270
and the P-N junction.

01:07:39.270 --> 01:07:41.720
Whereas, if the grain boundaries
are running perpendicular

01:07:41.720 --> 01:07:43.390
to the surfaces,
now they're only

01:07:43.390 --> 01:07:47.570
affecting very small areas of
the entire solar cell wafer.

01:07:47.570 --> 01:07:49.170
So when I pick up
a wafer like this,

01:07:49.170 --> 01:07:53.200
this wafer was chopped
from the ingot this way

01:07:53.200 --> 01:07:55.090
or, to put it into
perspective here,

01:07:55.090 --> 01:07:58.240
this wafer was sliced
out like that from this.

01:07:58.240 --> 01:08:00.400
So the grain boundaries
were running perpendicular

01:08:00.400 --> 01:08:04.270
to the surfaces and that way
they don't impede as much

01:08:04.270 --> 01:08:06.690
with electron transport.

01:08:06.690 --> 01:08:08.790
So the multicrystalline
silicon ingot is formed.

01:08:08.790 --> 01:08:11.230
The ingot is then chopped
into these blocks,

01:08:11.230 --> 01:08:17.330
usually between 16 and 24, that
means four bricks to an edge

01:08:17.330 --> 01:08:19.270
or five bricks to an edge.

01:08:19.270 --> 01:08:23.010
Some folks are exploring
six by six, so 36 bricks,

01:08:23.010 --> 01:08:26.300
and then the bricks are
rotated on their side

01:08:26.300 --> 01:08:30.020
and then sliced into wafers
and individual wafers come out.

01:08:30.020 --> 01:08:34.689
So you can see the wafers
I've sliced from the bricks

01:08:34.689 --> 01:08:35.960
as I showed you right here.

01:08:35.960 --> 01:08:38.700
Is this diagram clear to folks?

01:08:38.700 --> 01:08:40.300
In general since--
any confusions?

01:08:40.300 --> 01:08:41.859
Any questions?

01:08:41.859 --> 01:08:42.359
No.

01:08:42.359 --> 01:08:43.760
AUDIENCE: How do
they cut the wafers?

01:08:43.760 --> 01:08:45.176
PROFESSOR: How do
they cut wafers!

01:08:45.176 --> 01:08:48.460
So this is a process called
wire sawing sign and this is

01:08:48.460 --> 01:08:50.229
one of the most
beautiful technologies

01:08:50.229 --> 01:08:52.939
because it was invented
in the PV industry

01:08:52.939 --> 01:08:56.646
and transported back,
adopted by the IC industry.

01:08:56.646 --> 01:08:58.520
So it's one of the few
examples of technology

01:08:58.520 --> 01:08:59.760
that went the other way.

01:08:59.760 --> 01:09:01.330
Let me get to that point.

01:09:01.330 --> 01:09:02.890
AUDIENCE: How was
it done before?

01:09:02.890 --> 01:09:04.939
PROFESSOR: It was
done by ID saws,

01:09:04.939 --> 01:09:06.689
for instance, inner
diameter saws,

01:09:06.689 --> 01:09:10.974
that would slice off wafers
like a wafer off of a salami.

01:09:10.974 --> 01:09:13.604
AUDIENCE: So not a
wire but like a disk?

01:09:13.604 --> 01:09:14.729
PROFESSOR: Like a disk saw.

01:09:14.729 --> 01:09:15.229
Yeah.

01:09:15.229 --> 01:09:16.229
Exactly.

01:09:16.229 --> 01:09:19.450
Like the inner diameter meaning
your saw is like a rotating

01:09:19.450 --> 01:09:22.351
blade and you're just
using, you know-- Yeah.

01:09:22.351 --> 01:09:22.850
OK.

01:09:22.850 --> 01:09:25.880
So directional solidification
of multicrystalline silicon.

01:09:25.880 --> 01:09:29.569
This is a cross section of a
furnace that is solidifying

01:09:29.569 --> 01:09:30.470
an ingot right here.

01:09:30.470 --> 01:09:31.470
Here's your ingot.

01:09:31.470 --> 01:09:34.109
This is a liquid silicon
and, essentially, it's

01:09:34.109 --> 01:09:36.960
solidifying from the
bottom to the top

01:09:36.960 --> 01:09:39.540
and, hopefully, we'll have
a tour of one of the world's

01:09:39.540 --> 01:09:45.012
largest ingot solidification
furnace manufacturing

01:09:45.012 --> 01:09:45.970
companies in the world.

01:09:45.970 --> 01:09:47.594
So, they don't
manufacture the silicon,

01:09:47.594 --> 01:09:50.619
they manufacture the furnace
that manufactures the silicon.

01:09:50.619 --> 01:09:51.655
If that makes sense.

01:09:51.655 --> 01:09:52.810
AUDIENCE: Do they also
make the crucible?

01:09:52.810 --> 01:09:53.130
PROFESSOR: No.

01:09:53.130 --> 01:09:54.180
That would be Vesuvius.

01:09:54.180 --> 01:09:54.680
Yeah.

01:09:54.680 --> 01:09:58.130
It would be other companies
that make the crucibles.

01:09:58.130 --> 01:10:01.090
And these are some of
the furnaces right here.

01:10:01.090 --> 01:10:02.230
The keyboard and monitor.

01:10:02.230 --> 01:10:03.750
For size comparison, stairs.

01:10:03.750 --> 01:10:05.630
So they're about
two stories tall.

01:10:05.630 --> 01:10:07.890
You can go up here to the
top and look down into them.

01:10:07.890 --> 01:10:09.077
It's pretty cool.

01:10:09.077 --> 01:10:11.160
Using a little infrared
lens to block out the heat

01:10:11.160 --> 01:10:14.460
so you don't get blinded
and the furnace itself--

01:10:14.460 --> 01:10:16.310
all the action happens
inside of here.

01:10:16.310 --> 01:10:19.910
The top can lift-- typically,
they're the bottom loaded.

01:10:19.910 --> 01:10:22.370
You'll see this little
seal right here.

01:10:22.370 --> 01:10:24.110
So this bottom part
typically comes down

01:10:24.110 --> 01:10:26.026
because you want to trap
the heat inside of it

01:10:26.026 --> 01:10:29.690
so you're not losing all that
and the bottom is removed,

01:10:29.690 --> 01:10:33.540
the forklift comes in, picks
up this ingot and crucible

01:10:33.540 --> 01:10:36.280
which could be a few of
kilograms in mass-- up

01:10:36.280 --> 01:10:39.210
to about 600, maybe even
a ton-- and removes it

01:10:39.210 --> 01:10:41.415
and places in the
proper location.

01:10:41.415 --> 01:10:43.310
It's a pretty dirty environment.

01:10:43.310 --> 01:10:46.030
The operator will typically
take a garden hose

01:10:46.030 --> 01:10:48.240
and hose it down
inside afterward.

01:10:48.240 --> 01:10:51.150
It's really an
antithesis of an IC

01:10:51.150 --> 01:10:53.420
fab at this stage right here.

01:10:53.420 --> 01:10:56.830
These are graphite
insulation materials

01:10:56.830 --> 01:11:01.310
on the sides of the crucible
and this yellowish dust

01:11:01.310 --> 01:11:03.830
that you see everywhere
is silica, again.

01:11:03.830 --> 01:11:09.270
That nice fine grained dust
that's bad for you lungs.

01:11:09.270 --> 01:11:12.350
The directional
solidification process

01:11:12.350 --> 01:11:15.450
can be, to some degrees,
used interchangeably

01:11:15.450 --> 01:11:18.030
with the so-called
Bridgeman process.

01:11:18.030 --> 01:11:20.000
It's also a name
for a specific type

01:11:20.000 --> 01:11:21.900
of directional solidification.

01:11:21.900 --> 01:11:27.120
This is your ingot, this is
the ingot chopped into bricks,

01:11:27.120 --> 01:11:30.539
and then the bricks-- here's an
ingot coming out of a furnace.

01:11:30.539 --> 01:11:31.830
Those are the bricks over here.

01:11:31.830 --> 01:11:32.970
This is a really tiny one.

01:11:32.970 --> 01:11:34.330
It's like lab scale.

01:11:34.330 --> 01:11:40.930
The big ones are about over
a meter along the long edge.

01:11:40.930 --> 01:11:43.250
And then, to saw
them into wafers,

01:11:43.250 --> 01:11:46.561
we use what's
called wire sawing.

01:11:46.561 --> 01:11:49.610
These are several kilometers
of wires-- of wire.

01:11:49.610 --> 01:11:51.270
One continuous wire,
several kilometers

01:11:51.270 --> 01:11:56.880
long, typically of a
steel-based composite.

01:11:56.880 --> 01:12:00.790
Running in these
bricks right here,

01:12:00.790 --> 01:12:03.330
in the presence of a
glycol-based slurry, typically,

01:12:03.330 --> 01:12:06.200
and silicon carbide
or diamond grit,

01:12:06.200 --> 01:12:08.540
and the grit is being
pressured by the wire

01:12:08.540 --> 01:12:09.850
against the silicon.

01:12:09.850 --> 01:12:12.210
The grit is very small
in size-- micron size--

01:12:12.210 --> 01:12:14.970
and it's, essentially, chipping
out small pieces of silicon

01:12:14.970 --> 01:12:16.720
as this wire is
progressing through

01:12:16.720 --> 01:12:19.080
and, over a period of
around 6 to 8 hours,

01:12:19.080 --> 01:12:22.690
you saw through the entire
brick and you use, maybe,

01:12:22.690 --> 01:12:24.650
four or eight of them at a time.

01:12:24.650 --> 01:12:26.400
So if that wire were
to snap about halfway

01:12:26.400 --> 01:12:29.440
through the process, all
those bricks are gone.

01:12:29.440 --> 01:12:32.040
So it's very important
that the wire be

01:12:32.040 --> 01:12:35.890
very robust and able to
support the sawing process

01:12:35.890 --> 01:12:38.910
and, as I said, it's
several kilometers long

01:12:38.910 --> 01:12:42.650
and moving at a speed of
a few meters per second.

01:12:42.650 --> 01:12:44.850
So this is zinging along
through your material

01:12:44.850 --> 01:12:47.915
in the presence of very
small grit and slurry,

01:12:47.915 --> 01:12:50.540
and so the consumables that are
used in the wire sawing process

01:12:50.540 --> 01:12:53.150
are enormous, and you lose
about half of your silicon

01:12:53.150 --> 01:12:55.980
due to sawdust in this
process right here.

01:12:55.980 --> 01:12:59.220
So this is a prime
candidate for replacement

01:12:59.220 --> 01:13:01.910
of the manufacturing
process, even though it's

01:13:01.910 --> 01:13:04.050
so commonly used today.

01:13:04.050 --> 01:13:06.460
What I'm going to do is give
a quick pause right here

01:13:06.460 --> 01:13:07.870
until our next
class, where we'll

01:13:07.870 --> 01:13:10.200
pick up and talk about
ribbon growth, which

01:13:10.200 --> 01:13:12.750
seeks to get around
all the complexities

01:13:12.750 --> 01:13:14.770
of multicrystalline
silicon ingot growth

01:13:14.770 --> 01:13:16.910
while still keeping
the cost advantage.

01:13:16.910 --> 01:13:19.640
So with that, thank you.