Sunpower IBC cell: process flow

Sunpower IBC cell: process flow


In the last video, we were talking about this SunPower cell design, which uses an
interdigitated back contact. And we talked about all the nice things
which come along if you use this interdigitated
back contact design. First of all, your your solar cell, it
looks very aesthetically appealing. It looks all pitch black from the front –
there is no metal fingers or bus bars at the front.
So it minimizes the shading. And also some of the unique features that
we discussed, such as the use of n-type
silicon, these, localized p+ and n+ diffuision, which are which are all which are all very
native to this SunPower cell design. Now in this video, what I want to do is
discuss the process flow which is needed to manufacture such a cell. And, you’ll realize that the devil is
really in the details, when we, so we talked about all the advantages that you get by having these contacts at the
back. But the real problem or you know the real
challenge comes along when you think about how to manufacture these these these cells, especially how to make these
contacts at the back, such that they’re localized and
they’re interdigitated. And that’s where I think SunPower deserves
a lot of credit for commercializing this technology, and
making the price point come down at a point at which it becomes competitive. So, you know to give you an idea of how this process flow can very quickly
become complex, I want to show you a sample, an example of the process flow which can
be used to realize this realise this IBC kind
of a cell which SunPower makes. So we start with, you know, similar to how we start the processing for our
traditional silicon solar cells. So you make this make these textured, texturing to you.
Make this make these textured texturing on the top surface, very similar to how
you do a, for a, for a conventional cell. Over here you start with an n-type wafer
instead of a p-type wafer. Then you will do this you do this p-type
diffusion on the top to essentially form these
surface field at the top. So this is, you know, also similar to how you do it for a
conventional cell, only you’re, you’re making a p-type
diffusion instead of a. doping it n+ type, at you would do in a
conventional cell. So the next process is essentially forming
these localized diffusion. So, what is you know, one of the simplest
way you can think of is to, you know, form a
lithographic pattern, so do lithography and open up these windows.
And then essentially you know, then essentially dope them with
n-type dope glass and then diffuse these n+ dopants into your, into your device.
So there is one lithotography step which would be later required over here but now you need to form the p+ diffusion as well and that needs to be formed in these spaces which are in, or which is interdigitated between these n+ diffusion regions. So you’ll need to do another lithography step, so you’ll need to cover your n+ diffusion and open up this space between uh, between these, the n+ diffusion. So you’ll need a second lithography step which is aligned with this, which is aligned with this first lithography step. And now you can form these p+ diffusion
over here. So you can form, these are my p+
diffusion, these are my n+ diffusion. And then what I can do is I can
essentially you know, passivate the whole back surface with say
silicon oxide or some other dielectric. And then I can screen print these
contacts, so, I can screen print these contacts and then I can fire them in such
that they connect to these local p+ and p+ diffusion. But you know, you can see that I required, I’ve used two lithography steps
in this scheme. And you know, anybody who works in a solar
factory, they hate this, they hate this L word, that it
stands for lithography. So every time you add a lithography step,
it adds substantially to the cost of your to the overall dollar
per watt cost because each of these lithography steps, you know
you need to deposit a resist, you need to clean the resist off and, you
know, if you need to align the first lithography step with the first, you
need, again you know, it becomes even more
complex. And it slows down your, your manufacturing
line substantially if you need, if you include
any lithography steps in your process flow. So, you can see that your inclusion of
these two lithography steps makes this process flow very slow,
and also makes it expensive. Sometimes to, you know, to even improve,
further improve their cell characteristics, what is desired is
that these back contacts, Instead of placing them throughout the back
surface, these are you know, the whole back surface is and instead
passivated with this silicon oxide. And you open up these contact holes where
you will make contact between your metal and your semiconductors instead of
making these contacts over the whole line you make them now localized in these
just these holes. So this again requires a third lithography
step. So it requires a third lithography steps, which
again, adds to your process flow cost as well as to its
complexity. So this is why it’s, you know, it’s very expensive to make these
interdigitated back contact cells. What SunPower did and it should receive a
lot of credit for that is that it essentially simplified this
process flow such that it minimizes this amount of lithography steps. So one of the, you know, one of the papers that they published on this, back in the
late, late 80s, was to demonstrate this self-aligned
process flow and that reduce the number of lithography steps to
to one. And let me illustrate that, so how that is
done. So you start with the n-type wafer and then you texture if at the top, just like you would do in a conventional
process flow. And then you form this p-type diffusion
from the top. So you form, that gives your surface field which helps in minimizing this
surface recombination. And then you start working on making these
localized, localized diffusions at the back, so now
this is where it becomes really interesting, so
the way this the SunPower paper described is that you coat all of the back surface, you
coat it with a layer that is essentially, it constitutes of, of n-type dopants, so you coat it with a PSG or a layer like that
which contains these n-type dopants. And now you do one lithography step, where
you open up these open up this layer which had
n-type doping. And then you also recess etch, so you etch your
underlying silicon as well. So you you etch this underlying silicon so that it it forms these trenches inside
your, inside your entire, inside your silicon as
well. And now you coat this so you know, we did, let’s keep count of
the lithography steps. So we did one lithography step to
essentially etch this n-dopant layer and recess, recess into
this silicon. Now you coat the whole thing again, you
coat the whole thing again with another layer which has which is now
a source of p-type dopant. So you coat it with your glass-like layer which is a
source of p-type dopants. So, all over here, you have this this surface, especially that recessed
surface over here. It is now contacted with a source of
p-type doping. And now essentially, you fire this
together, so you. you essentially, you know, apply some
thermal budget such that you form these n+
diffusion inside your silicon. So you form this n+ diffusion in the
region which is facing which is in contact with this layer containing n
diffu, n-type dopants. And some of you found this inside this
recess you found this p+ localized diffusion because that surface is in
contact with this layer which contains p-type
dopants. And now you remove these layers, these
layers which were used to you know, source the dopant
atoms. So you remove these glass-like layers
which were which were essentially supplying
these dopants. So you are left with these junctions which
have now been formed inside your silicon. And you are left with this recess at the
back. And this recess is is very important.
as we’ll see shortly. So the next thing you do is now
essentially you know, sputter the whole thing or, the whole thing or. you know, you essentially deposit aluminum on, throughout this, throughout this back
surface. And when you deposit you know, aluminum
usually when you deposit over a step feature, so let’s say you have a step
feature step feature like this. And you, … let me use the same color.
So you have a step feature like this. So this is assumed to be a view of what’s happening over here. And now when you start depositing
aluminium it would deposit something like this, where layer by layer layers of aluminium will deposit here, and
this, this, this surface will expand in this
direction. But that will prevent aluminum from
depositing on the edges of this bottom surface. So if you look at the bottom surface and this aluminium is depositing from this
side. So essentially it will automatically
self-align these contacts such that they don’t short with each other. So when you deposit aluminum, it deposits in such a way such
that these metals, they hopefully don’t short
with each other. And then you deposit, just to be double sure, you deposit a layer of titanium over
here. So your deposit a layer of titanium on the top.
And then you etch back this, on the top.
And then you etched back this etch back this aluminum.
So this titanium layer, this aluminum.
So this titanium layer with it acts as a mask, so even if there was some tita-, some
aluminum which got deposited over here, once you use this
titanium layer you can etch this back and you can make
sure that this these metal contacts which are connecting my n-type diffusion
and my p-type diffusions are not shorted with
each other. So now this would form my p bus bar, and
this would form my n bus bar and using this self-aligned process flow I can
make sure that you know, they’re not shorted with
each other. Also I’ve used only one lithography step
in this whole process flow. And this is illustrated further, in this
in this publication from SunPower, and it shows how they have formed this
interdigitated n+ and p+ contacts, where this p+ contact is located, is entrenched inside this trench which was
created. And this n+ contact, it’s located at a slightly higher feature as compared to
this p+ contact, and you can see this this this
recess etch going on over here as well. So the key, over here, to separate these
two layers is this titanium layer which is deposited
on the top, and using that and a [INAUDIBLE] mask you can etch into this aluminum. So you can etch into this aluminum, and
make sure that these p bus bar, and these n bus bars are not
shorted with each other. So here I’ve illustrated a simpler process
flow, where you were able to achieve this using one
lithography step. More recently we know that Sunpower has
eliminated all lithography steps. So it uses zero lithography steps.
And it relies on screen printing of these dopant layers. So these instead of forming these doping
by diffusion, you essentially deposit or screen
print these dopant [UNKNOWN] layers and then diffuse them in. And more recently they also used this
copper plating at the back, at the back of these, to form
these n+ and p+ interdigitated contacts, they use
electroplating of copper, which also helps in providing improved mechanical strength, improved mechanical
strength to this solar cells. So more details of this are not you know,
this detail of this process flow is not
available in public literature. But you get some hints of what’s happening by reading the patents which are being
published. But anyway, I hope that, you know, I have
given you some idea of how this process flow of forming
these interdigitated back contact works. And hopefully left with you with some food
for thought or some ideas where you can
simplify it further.

2 COMMENTS

Leave a Reply

Your email address will not be published. Required fields are marked *