begin our discussion on epitaxial techniques for gallium arsenide which will be used for
high speed devices. Yesterday, we discussed in detail, how to realize semi-insulating
gallium arsenide material because when you want to make a device or integrated circuit
what we have discussed yesterday was you need an insulating substrate if not insulating
at least semi-insulating. For example, if you take SiO2 the resistivity is something
like 10 to the power of 14 ohm centimeters. Whereas, what you have particularly achieved
with gallium arsenide as a semi-insulating substrate is something like 10 to the power
of 9 to 10 to the power of 10. 10 to the power 9 is quite good, resistivity that you can
get. It is not a perfect insulator that is why it is called as semi-insulating material.
We have also seen that you can do it in two ways: one, take poly crystalline gallium arsenide
melt it and grow by sucrose technique, at that point of time when you do that what you
will get will be an unintentional doped gallium arsenide doped with silicon. Now, then you
compensate it by adding chromium which introduces a deep level at the mid gap. Another approach
which is now popular with availability of high pressure crystal growth system is, you
make the compound in c2 by taking gallium and arsenic. In fact, you can get 7 9’s
not 6 9’s, 7 9 impurity 99. 99999.
So, 7 9 purity gallium arsenide, gallium 7 9 purity arsenic, all those you can get. You
can get better purity gallium arsenide substrate. You grow that but when you grow you end up
with p type material, if nothing else happens due to the contamination from carbon due to
contamination from that liquid encapsulates, which you are using. These are things we saw
earlier. But the compensation for that acceptor level, shallow acceptor is done automatically
by excess arsenic that is present or the arsenic which is sitting on the gallium sites. In
fact, if you have some vacancies present when you grow the crystal, you adjust the composition
such that you have deliberately gallium vacancy then some of these excess arsenic atoms will
go and occupy the gallium site, that is called anti site defect. Whenever you have got an
additional level, this anti site defect is called e l 2 levels. I really do not have
information why it is called e l 2. It so happens, that it is called e l 2 level, instead
of saying chromium level you say e l 2 level. It is actually due to the anti site defect,
instead of adding impurity you have the arsenic sitting there. There is no extra impurity
that is added, this itself is acting as an impurity. Whenever you have loss of periodicity
or loss of structure, you get additional levels and this gives donor levels at the mid gap
and we have gone in detail;, if you have acceptor level shallow, you can compensate very easily
without worrying too much, whether I am having too much of donor level, put the donor level
at the mid gap, these are some of things that we discussed. You do not have to worry about
that e l 2 name, it is the name given to the defect which is arising out of arsenic sitting
on the gallium site that is all. Now having done that, now you see the growth that is
Having done that red color or brown color is slowly growing here. That is I put it back
once again. You can see gallium arsenide substrate then, on the top of that you have that layer
slowly growing that is the layer which is the active layer. In fact, I mentioned yesterday,
if this is semi-insulating gallium arsenide is chromium doped, before you put this active
layer, you introduce a buffer layer. The abuse is taken by the buffer layer. In the sense,
when you do high temperature processing all the chromium that is present may go into the
buffer layer and actual layer on the top does not get affected.
But today, with the ability to grow high purity gallium arsenide with compensation of e l
2, you do not have to worry. You can straight away grow the epitaxial layer on that active
layer on that. Now what we want to discuss today is, about this particular layer. What
is that? In fact, you will call this by different names in the sense, if you have leisure gives
an epitaxial layer if you are in a hurry it is epitaxy. When you say epitaxy, it is actually
growth of epitaxial layer. What is done here is, let us talk of various things. Supposing
you are growing silicon on silicon then that is silicon epitaxial layer. Supposing you
are growing gallium arsenide on gallium arsenide substrate can have any doping. It can be semi-insulating
p type, n type. But here, what we are talking of is semi-insulating gallium arsenide structure.
You can grow epitaxial layer of gallium arsenide on the top of that. The layer that you are
growing, if you are going by this, whatever technique you are talking of you can adjust
the thickness. We can say I want 0.1 micron or 0.2 micron. We can adjust the doping of
that level when you grow the crystal when you grow this layer that is the epitaxial
layer. When you say grow the epitaxial layer, the meaning is you are arranging the atoms
from the gallium arsenide on the semi-insulating gallium arsenide substrate. Arrangement, if
you give proper opportunity to these atoms, which are being laid out on the surface, it
will take the same crystal structure at the substrate. That is why that is called; in
fact, you can call it as auto epitaxy or homo epitaxy.
If the substrate is doped differently but still it is called homo epitaxy n type, p
type does not matter but if the substrate is totally different, for example when you
grow silicon epitaxial layer on sapphire that is hetero epitaxy. They are totally different
substrate. But still you can grow a 1 0 0 oriented silicon on sapphire because there
is a good lattice match between silicon with 1 0 0 orientation to that particular layer,
one particular orientation of sapphire. So that is hetero epitaxy. Now the process involved
in the growth of this epitaxy. What is that? Process is two things should happen. One you
must be able to transport these constituents, which form this layer, onto the surface. If
it is silicon the constituent is silicon alone. You must transport silicon onto the surface
of this silicon substrate. If it is gallium arsenide, you must be able to transport gallium
and arsenic, both should be transported onto substrate. How to transport it depends upon
the particular process that you are using and you call the particular process by a name
depending upon how you transport this impurity, this constituent elements atom.
We will see what the different processes which are involved. Now you transport these constituent
elements it is not enough if you transport them. You must give opportunity for these
constituent elements. Let us talk of silicon because only one you have to talk of only
one element. Silicon atom you have transported onto surface what does it do? If you transport
and just do not, if you have just kept this and transport the silicon atom, it will go
and rest there. That is not epitaxy; it is just dumping silicon there. You do not want
to dump silicon. You want to have silicon arranged with the arrangement same as that
of the subject. How can you do that? You must give opportunity for those atoms to move around.
It should be able to move around and attach onto some of the sites, which you call it
as nucleation sites. Nucleation site is actually a favored site, which will go and sit.
For example if I have a surface a slight kink in some portion. There may be a number of
kinks due to the atom itself substrate atom itself, on the atomic scale. There will be
small kink, it need not be a kink like that and small kink there this atom that you have
put will go and just move around. When it sees the kink it will just stop there, that
is stops there that is we can call it as nucleation site. It also involves proper wetting. Simple
minded thinking will tell you that, wherever kink sites are there, the atom which is moving
around because of the energy given to that. What is the source of energy? Thermal, that
is why in epitaxy you must heat the substrate sufficiently, so that the atoms have enough
energy to move around. It will go and attach it the kink sites and from that kink site
it will move laterally because if I have one atom sitting there, here next atom will come
near that, side by side. One layer atom should be arranged once that one layer is complete.
Next layer grows; it keeps on growing till you stop the reaction. That is the epitaxy
that is the arrangement of atom that is being done.
That is why this diagram, what we saw in this diagram, we can see, we can watch this. You
see that is the way it grows unfortunately no color there. It is actually from there
onwards, you can see at least the size coming up. You have put like that, you can see that
growth so that additional layer that is growing there is gradually growing like this. That
is layer by layer. Now if you do not provide additional temperature, additional energy
for the atoms to move around, what will happen? Room temperature we cannot have the growth
taking place. We will have lumps of silicon or lumps of gallium arsenide coming up. It
is no use for making devices. But if the temperature is slightly lower than what is required, the
atoms will not have enough energy to arrange themselves. You will end up with not single
crystal, poly crystalline material even if the substrate is single crystal. Suppose the
substrate is not single crystal, if you have amorphous material like oxide. On the top
of it if you grow a layer that will turn out to be poly crystalline. It can be amorphous
or poly crystalline depending again upon the temperature. Amorphous is random. Poly crystalline
is number of single crystal layer regions, which are joined together.
Each region has got you want to see that, at one region it may be 1 0 0 orientation,
you will have another region here which is 1 1 1 orientation. You will have another region
here 1 1 0 orientation. I am just arbitrarily putting it can be 2 1 1, 2 2 0 whatever. We
will have all these things coming up. That is poly crystalline; we are not looking for
poly crystalline material here. But in silicon technology you could hear about poly crystalline
material very often. Where do you hear that? It is part and parcel of the present day integrated
circuit technology because it forms the gate region of the MOS device. That gate region
is done not by epitaxy; the temperatures are low for silicon epitaxy. You need about 1000
degree centigrade at least. If you take single crystal silicon and raise the temperature
to about 1000 degree centigrade and pass silian gas through that. Silian is nothing but Si
H4. That is silian gas. If you pass that, at that temperature pyre opens, it will crack
into silicon and hydrogen and that silicon atom will be sitting on the surface. If it
is single crystal at 1000 centigrade substrate is single crystal 1000 degree centigrade,
you will get single crystal silicon.
Onto that you add arsine silian is there, this will give you silicon plus 2 H2. So that
is the one which is being deposited. You want to dope to the gas face, you should add arsenic
you will get arsenic plus, you will get As 2 plus 3 H2 or 2 As. That is chemistry. You
have arsenic doped silicon, you can add phosphine. You get fifth group phosphorus. You can have
silicon doped with arsenic or phosphorus depending upon the dopants added. What I am trying to
bring out is I took the example of silicon because it is simple to understand. You can
grow an epitaxial layer undoped the epitaxial layer. Undoped means not intentionally dope,
you may have contaminants coming in. You can intentionally make it n type, or p type, by
adding the suitable constituents to the way you bring in the atoms. Now here this in silicon
that poly you obtain there is due to two factors. One is temperature is about 625 degree centigrade
in the M O H technology. Two, you grow it on oxide, you deposit on oxide both put together
you get poly crystalline 625. You lower the temperature down to 55 centigrade you will
get amorphous. You will get amorphous silicon there. It all depends how much energy it has
got to move around. At least to some degree of orientation you get the temperature is
625. Now let us not spend our time too much on the silicon epitaxy. This is relevant silicon
epitaxy, if you are thinking of silicon on sapphire. That is the same process silian
broken into silicon and hydrogen. They get sapphire sufficient that is the temperature
1000 to the deposition. In fact, that is one of the materials which are used for high speed
MOS devices. That is why it is relevant to this particular presentation discussion. But
if you are talking of compound semi conductor only then, let us see what the things are,
that is the basic process that is involved in the epitaxy.
Now let us talk of epitaxy of gallium arsenide. As I mentioned one of the requirements is
to keep the substrate hot. Other requirement is transport these consequent elements to
the vicinity of the substrate. Depending upon the way you transport those consequent elements,
you call it vapor phase epitaxy. What it is we will see one by one liquid phase epitaxy
I will give a glimpse of liquid phase epitaxy when I discussed about p type, n type dopant
by adjusting temperature in the previous lecture, molecular beam epitaxy. That if you talk to
the people depending upon what they have at their disposal they will highlight on that.
If you talk to MBE person they will say that is the best. Objectively, each has its own
merits; each has its own drawbacks also. You have to compromise depending upon what you
Let us see vapor phase epitaxy VPE of gallium arsenide. Here this involves the transport
of species by some of the volatile compounds of the material. For example, if you talk
of silicon, you talk of vapor phase epitaxy. That is vapor phase epitaxy. You talk of the
volatile compounds like arsine As H3 or you talk of arsenic tetra chloride, not arsine
it is saline or silicon or silicon tetra chloride. Then release silicon from that, the reaction
will be taking from the surface of the substrate, that is the thing. In the case of gallium
arsenide, you have to transport a compound containing gallium you have to transport a
compound containing arsenic. Then forced reaction in that reaction locker you provide the actual
temperature etc, for the reaction to take place forced reaction of the substrate. You
do not want; you do not prefer the reaction to take place in the gas space.
Supposing I am having a substrate here like this, if the gas is moving through that you
do not prefer the reaction to take place there. You prefer the reaction to take place on substrate
because in a gas space it is the reaction which is taking place, you have these gallium
arsenide formed there it may be droplets of gallium arsenide coming out there that is
not what is required you must have the reaction taking place just near the surface on the
surface. So that atoms once they are released they move about and arrange themselves that
is called heterogeneous reaction. The reaction will be homogeneous if the reaction is taking
place in the gas phase, the reaction will be heterogeneous. If it is taking place on
the surface itself, you are the favored or the one that you like is heterogeneous reaction.
To put it in simple terms, you want the reaction to take place in some surface that is what
happens in a heated substrate right there.
Now let us see you can transport, see after all you have to transport arsenic and gallium.
Today I am giving you the whole gist of epitaxy. If you start talking of this, you can talk
for ten hours on this. Just the idea what is involved, what are the compounds involved
so that you get a flavor of what is going on around. The transport of arsenic is done
in the form of the vapor phase epitaxy is done in the form of arsenic tri chloride or
arsine AsH3. Arsenic tri chloride is liquid. Arsine is gas. Gallium is transported again
in the form of a volatile compound either as gallium chloride or tri methyl gallium.
Now depending upon the compound used for transport, you call it by different names you call it
for vapor phase epitaxy. First you have got three classifications. Let me just go back.
You have got three classifications depending upon the mode of transport here the transport
is through vapor phase compound, the liquid phase it is actually precipitation from the
liquid phase. You keep the substrate in contact with the liquid, from the liquid you precipitate
those constituent elements that is liquid phase; whereas, molecular beam epitaxy physical
transport of gallium physical transport of arsenic in the form of a beam that is why
molecular beam. That particular point is that, that is why that forms classification. Now
coming to the vapor phase epitaxy, depending upon what sort of compounds you use for transport
of the vapors, you call it halide process, hydride process, or MOCVD or OMCVD is one
of the most popular epitaxial techniques for gallium arsenide today. Metal organic, chemical
vapor deposition, or organo metallic. Both same it is only who is talking somebody will
say metal organic CVD or organo metallic CVD. You deal with the same compound trimethyl
gallium. That is called depending upon that particular thing, you call it MOCVD. Now let
us just take a look at some of the, each one of them in detail.
Not in detail just going through the name. Halide process, you can see here, this is
a furnace tube, watts tube which is laid out there and it is a 2 zone furnace. Fancy name
given to that is the reactor. An epitaxial reactor is nothing but that it frightens you
when you say reactor. But if I say it is a furnace that is all that is it. You have a
furnace here, for epitaxial reactor, which has got 2 zones. One zone where you can heat
it about 800 to 850 degree centigrade simultaneously you must be able to heat other zone to 750
degree centigrade for growth. The philosophy here is this or the approach is we have hydrogen
gas and arsenic tri chloride liquid. You have a bubbler, you pass hydrogen through that.
Like this I am just showing a bubbler it can be from flask then you have the arsenic tri
chloride. Really a lot of chemistry comes here. We will prove with that with this lecture
so, this arsenic tri chloride and a tube coming into that. You pass hydrogen through that,
because you can get it in a very pure form. You do not want to use any gas with contaminants
from the epitaxial layer. If you pass hydrogen through that if it is very pure, 6th 9 purity
and then you notice that is the liquid and this bubbles hydrogen through that. When it
comes out it picks up the vapors of arsenic tri chloride. What are transported now are
the vapors of arsenic tri chloride that is what is shown here. Arsenic tri chloride vapor
is occurring here. What was before that, I have not shown that is a secret. It is what
is present there. You control the temperature of the bubbler. We can control the vapor pressure.
Once you control the vapor pressure, extent of arsenic tri chloride comes in this control.
The controller is there that is the temperature of the bubbler. Now rest is just chemistry.
A few equations we will see. Arsenic tri chloride notice here, you have a crucible which contains
gallium. Gallium is melt even at 30 degree centigrade. It is the molten gallium. Past
this arsenic tri chloride vapor comes in this is our career gas additional to dilution,
if you want arsenic tri chloride plus hydrogen will react even at 425 degree centigrade to
give you arsenic in that form As4 and HCL.
Now this arsenic which is coming out here will slowly react with gallium where you have
kept the temperature at about 800 to 850 degree centigrade. It will form gallium arsenide
on the surface. You may have to keep for that reaction to take place, to form a crust of
gallium arsenide on the gallium surface; you may have to keep it for several hours. When
I say several hours it is even 24 hours so that a crust is formed. Now, you have a top
player of gallium, which is converted into gallium arsenide. Now notice the first reaction
that is taking place. You have arsenic which is reacting. You have also hydrochloric acid
HCL. This gallium arsenide with HCL will react with the crust, which is formed gallium arsenide
and this one here is again an arrow left hand side is the reaction taking place. These are
the reaction products. This reaction is either way it will take place. It will proceed from
this to that when the temperature is high. You keep the temperature high 800 to 850 back
of the temperature is 850, this reaction between gallium arsenide and HCL will take place to
release gallium chloride.
What is this whole effort about? Ultimately you want to bring in gallium chloride here
and arsenic here and force gallium chloride and arsenic to react, to release gallium arsenide.
This weapon that is used is this equation. This equation is a double edged sword. The
double edged sword in the sense, we can force the reaction to take place from left to right
that is gallium arsenide HCL reaction taking place releasing gallium chloride and hydrogen
and arsenic that is taking place here because temperature is about 850 degree centigrade.
What is transported across into this, this is this thing what is on the right hand side.
Now here you keep the temperature at 750 degree centigrade lower temperature. The reaction
proceeds in the upside direction, same species which have been produced at high temperature
recombine. This is the thermodynamics of the process that will release gallium arsenide
on the surface. This is the graphite substrate you keep it at an angle etc., so that to get
uniform deposition all through. To get all through uniform deposition, you keep it in
the tilted position so that even if the vapors get depleted, you get higher velocities so
that diffusion takes place more on this side. Result is you get uniform layer from this
end to that end. This is the graphite substrate which is heated that is the semi-insulating
gallium arsenide. On the top of that you will realize this layer that is this reaction taking
place releasing gallium arsenide. That is the thing you need, 2 zone furnace with which
you can use, gallium as the source to convert it to gallium chloride by the reaction of
This reaction HCL, gallium arsenide and this what you get when you bubble? The process
that I discussed here is summarized and that it is available for you to see the hydrogen
is of course, you can get 6th 9 purity that is why we use hydrogen as the carrier gas,
all that I have mentioned.
Now, the halide process called hot wall reaction process as I mentioned already, high quality
gallium arsenide epitaxial layer of career concentration. 10 to the power of 13 per centimeter
cube you do not get semi-insulating there because it is just a thing which vary growth
vary or transporting. Some of these compounds which will carry some of the impurities. You
get 10 to the power of 13 undoped n type. It turns out to be because of some silicon
contamination it comes from gallium source that has some silicon contamination. That
is enough to give the doping concentration and arsenic tri chloride of course is you
can get here. I just want to correct this, because this particular quantity is 1 9 added
there and this also 1 9 added there, 7 9 purity because of this purity you can get this. Such
load doping level concentrations if you do not have such purity it will be 10 to the
power of 15. Now that is the halide process. Quite popular because we get layers which
are quite pure.
Next process is slightly different. The way you transport these ultimately over here,
you get gallium chloride and arsenic. So, instead of arsenic tri chloride which you
use in a hydride process, have halide process. You use arsenic tri chloride which is used
to release arsenic and also gallium. Gallium is released through gallium chloride; arsenic
trichloride converts gallium to gallium chloride. I hope chemistry is right. Now here, same
gallium is there liquid form, all mechanism same thing and arsine is let inside deep inside.
In the previous case arsenic tri chloride was coming right up to this point only so
that reaction is forced. Now you straight away transport arsine there and you have HCL
also here. That is separately required, if you require you can use it because I just
mentioned in the previous thing, before you start growing the layer you must etch the
substrate in c 2. Once the temperature everything is reached or even higher temperature, you
remove a layer of gallium arsenide. In the case of silicon epitaxy also when you grow
silicon just do not go and dump silicon on that. You remove a layer of silicon so that
a fresh surface is exposed using HCL etching hydrochloric acid, etching HCL etching is
done. That is initially you pass HCL directly to that point which will actually react with
gallium arsenide and if the temperature is high initially you keep it at 900 degree centigrade.
If you recall in the previous equation HCL plus gallium arsenide will react and remove
that h gallium arsenide, if the temperature is high. But the products will be formed gallium
arsenide will be formed if the temperature is 750. You can see it here. Just quickly
go through that because you can just see it in looking into this arsine breaks up into
As4 and hydrogen then here HCL is passed also on to this previous reaction. You convert
gallium into gallium arsenide and allowed arsenic tri chloride to react with that to
release gallium chloride transport gallium here HCL is directly supplied to react with
gallium to release gallium chloride and gallium chloride with this gallium chloride and this
arsenic will react to give gallium arsenide deposit here at 750 degree centigrade. This
reaction would have been in opposite direction. The temperature is high. That is actually
the hydride process.
Whatever I have said is written here Arsine is used for arsenic transport purity of arsine
is lower than that of arsenic tri chloride arsenic tri chloride purity is about 7 9s
today about ten years back it was 6 9s now, 7 9s; whereas, arsenic purity is still about
5 9 purity. Result is the layer that we get you cannot go down to those doping concentrations
which are 10 power 13 undoped but you want undoped 10 to the power of 15. No problem
you can use this arsine. So the difference here is the ratio of arsenic and chlorine
can be varied very easily so they are different sources. If you are using arsenic tri chloride
ratio of arsenic to chlorine is 3.33. So you have bit more flexibility for deposition that
may not be a big advantage but then that is one of the merits hydride process course is
there. If you just go back to the previous thing, you will see in the hydride process
or the halide process, first process arsenic trichloride you have to convert gallium to
gallium arsenide that takes long time, 24 hours I mentioned; whereas, if you use the
hydride process there is no such conversion. Go back and see here straight away, gallium
chloride is released by reaction of HCL and gallium in the previous process. Go back further
down in the previous process you release gallium chloride required for gallium by reaction
of arsenic tri chloride arsenic gallium arsenide formation here. So that process is not there
that is what we mentioned here. That is, it does not take it does not wait for the gallium
arsenide crust form on gallium straight away pass HCL that will convert gallium to gallium
Coming to the third process that is M O C V D or metal organic chemical vapor deposition,
I just want to give a word of caution. We are talking of vapor phase epitaxy. Sometimes
it is also referred to as chemical vapor deposition. It is a generic term, general term that is
used when you say chemical vapor deposition; it may be epitaxy or may not be epitaxy. What
we are talking of is all these processes were chemical vapor deposition because the transport
is through the chemical vapors of the chemical arsenic tri chloride arsine they are all vapors
so chemical vapor deposition. Now, same thing that we talked of if you use philine in silicon
at lower temperature. It is not single crystal that is not epitaxy but if you take a single
crystal substrate, on that use proper temperature then use the chemical C V D that becomes vapor
phase epitaxy. The correct term is vapor phase epitaxy. In fact, even here they should have
used metal organic vapor phase epitaxy M O V P E. We will use it, but general term is
M O C V D. This is epitaxy. What we are telling is if you use M O C V D on a foreign substrate
like, molybdenum you get poly crystalline material. If you use M O C V D on the single
crystal substrate like gallium arsenide you will get single crystal.
Here it is a similar process, much simpler. All that you do is here, it is pyrolitic decomposition.
We can use a substrate graphite substrate heated by rf-radio frequency coupled to the
substrate alone. When you do that, the deposition will take place only on substrate in a hot
wall epitaxy deposition will take place not only on the substrate, it will deposit on
the tube also. You have to remove the tube every time and clean it. Here also some deposition
will be there. But neither deposition will be here. See here, the organo metallic compound
is used for transporting gallium and for transporting arsenic, arsine is used. Not so high purity
like arsenic tri chloride but you put up with that. CH3 thrice gallium is tri methyl gallium
that is bubbled hydrogen is bubbled through that bubbler is shown here, which is maintained
at room temperature usually 0 degrees easy to maintain. Put ice and maintain and then
that is passed through that thing and this tri methyl gallium and in another tube, you
have got arsine coming up. Over here, you maintain a substrate temperature at about
650 degree centigrade. That is sufficient for epitaxial gallium arsenide 650 to 750.
650 degree centigrade you maintain and then you will have this reaction taking place on
the surface. To release gallium arsenide and methane, you can see the simple chemistry
also is simple here one chemical reaction taking place there. But more than that you
actually have the merit of using this rf plus you can use this for growing gallium arsenide.
Use another bubbler which has tri methyl gallium and tri methyl aluminum, instead of CH3 Ga
CH3 Al. what will you get? I have one bubbler which takes tri methyl gallium vapors. One
more bubbler I attach to take tri methyl aluminum and arsine gas. What will you get? From tri
methyl gallium, you get gallium. Tri methyl aluminum, you get aluminum. Arsenic you get
arsine. So, you get aluminum gallium arsenide. Such a process is not possible with the other
two processes, which we have discussed. You have the aluminum compound, if you realize
you have to go to 1100 degree centigrade; whereas, here the aluminum compound is organo
metallic compound or metal organic compound. Aluminum tri methyl aluminum this is one of
the key merits of M O C V D. Tri methyl gallium tri methyl aluminum. We can tell names tri
methyl indium, if you want indium, tri methyl zinc if you want zinc. Dope it in p type put
another bubbler. Either you can call tri methyl zinc t m z, or di ethyl zinc d e z ooze out
anything names you want. Those are some of the names. The merit of this metallic organic
C V D is you can get binary compounds like gallium arsenide. We can get ternary compounds
like aluminum gallium arsenide, gallium indium arsenide. You can also dope them with zinc
with using diethyl zinc or tri methyl zinc. Of course we can add dopants like sulphur.
Use another bubbler that means you will have all bubblers hanging around connected to the
Sulphur you can use hydrogen sulphide or you can use sulphur mono chloride that is the
sulphur. After all, what you need is a compound of sulphur hydrogen sulphide or sulphur mono
chloride. You can add those things; bubbles sulphur mono chloride is liquid. In fact one
of the systems which we have been working at, way back was sulphur mono chloride this
system. That was used for growing gallium arsenide on molybdenum. Thin layer of molybdenum,
gallium arsenide doped with sulphur to make solar cells. There is a huge program on gallium
arsenide by solar cells and grow it on molybdenum substrate in terms of poly crystalline. Now
this is mo c v d in very nut shell that is the thing. All the advantages I have just
told you already. These are put down here.
The growth rate is about 0.1 to 0.5 microns per minute. If you want 10 microns there,
at least about half an hour or some thing like that you may have to put. 0.3 microns
used to grow about half an hour to grow 10 micron layer. That is why I am remembering
that number. This growth rate we can control by controlling the partial pressure of tri
methyl gallium. You have plenty of arsenic in all these cases. You must have plenty of
arsenic. Why? Arsenic over pressure must be maintained. Otherwise, you will lose arsenic
from the growing gallium arsenide. For all through you put excess arsenic there to maintain
over pressure. That is done here also that arsine is maintained at over pressure, but
growth rate is controlled. How much t m z is there for reaction to take place? Excess
arsenic is there, through arsine. Now more t m z you add more will be the growth rate.
You do not want to have excessive growth rate. The crystal will be better if the growth rate
is smaller because it has enough time to organize itself. You prepare further growth rates.
Advantage of M O C V D is ability to grow aluminum gallium arsenide etcetera which I
have said just now. You can use tri methyl aluminum for transporting aluminum; this is
not possible in halide.
This is just the system which is shown schematic MOCVD system, which you know that all I have
mentioned just now. You can have tri methyl aluminum bubbler, tri methyl gallium bubbler,
diethyl zinc bubbler, to dope p type if you want to and all of them we can bubble hydrogen
through that. These are all the valves which will be opened and closed by means of hydraulic,
that is air pressure. You press a button the valve will open or close to let in that. If
you do not want to dope it you close this wall. If you want to dope it open that both
of them. Then of course you have got arsine bubbled through that straight away hydrogen
sulphide coming in here. Those are for doping if you require tri methyl gallium, this is
The last one that we will have here is the molecular beam epitaxy, the molecular beam
epitaxy is different form the vapor phase epitaxy, in the sense the constituent element
are directly transporting them physically, a very good example of physical transport.
I am sure all of you are familiar, is vacuum deposition. What do you have in the vacuum
deposition system, you have chamber in which vacuum is created? You have a filament which
can be heated and you keep. For example, if you want to evaporate aluminum, that is you
want to transport aluminium on to silicon or gallium arsenide, what you do? That filament
is heated. You put aluminum there, heat it. It melts and evaporates and it goes all over
and on to substrate. Now same thing is done here. Physical transport, this is nothing
but a physical transport. Now you call it molecular beam. Not because just molecules,
you will have different effusion cells. These are the different cells crucibles which are
heated just like putting some evaporant. You have these crucibles, which you call it as
effusion cells. Sometimes they call it as k cells Knudsen cells. These are some of the
jargons which the M B E people throw at you. It is just a source, which is heated. Crucible
is heated; this heating coil is put around that. Now there are shutters all around here.
The entire thing is a stainless steel chamber which is evacuated, by means of complicated
pumping system oil free pumping system and each one is the cell which is used for let
me use that language evaporation. You may not like it, if I say evaporation. It is actually
physical transport. Let us say of gallium, pure gallium 7 9 purity gallium is put inside
the cell. You have to fill that in. Arsenic well arsenic is put poisonous deadly poison.
There is always a joke, which tells what is poisonous plus, is good for semi conductor.
You look at the gases; you look at the substance that you are using. It is only as a lighter
sense, do not take it literally. It turns out that arsenic and if you want aluminum,
put aluminum here. Indium you put in indium here, whatever if you want to put there the
sources. What you do is whatever and dopant you want to put, you can put it there. You
can get the required layer of the material gallium arsenide, indium gallium and arsenide
etcetera by putting using the proper source. Now, the substrate, the single crystal substrate
is put here. And it is attached on to a heated substrate, and the substrate is molybdenum.
Molybdenum is used because it is a good thermal conductor. That is chosen because of that
and how it is mounted on that. It is mounted on that using indium for mounting and of course
indium melts at 150 degree centigrade. Even gallium can be used but it is held there due
to surface tension. Even if it melts there, this is held on to the substrate by means
of surface tension. Now the key thing here is this actually goes as a beam and it will
go as a beam if it is not colliding. It will collide if any of the particles, if there
are particles. What are those particles? If it is atmospheric pressure there, what will
happen? When you evaporate it, it will not go there. It will just collide and collapse
near that. In fact, if you take a vacuum evaporation system and if the vacuum is not proper, aluminium
will not go to the substrate. It will just stay there. Spread out. This is one thing
which you have seen. You have to create a vacuum of at least 10 to the power of minus
5 so that the mean free path length is much larger than this gap. This will be about 30
centimeters. The distance between the source and the substrate will be about 30 centimeters.
It must be much larger than that. A calculation shows if I have about 10 to the power of minus
5 torque; 1 atmosphere is 760 torque this is 10 to the power of minus 5 torque. If you
use the mean free path is at least 10 meters, you can be happy with 10 to the power of minus
But if you grow gallium arsenide at that pressure 10 to the power of minus 5 torque and this
substrate heated to about 600 degree centigrade, why do you heat the substrate? The atom should
be able to move if you heat that. The vacuum here is 10 to the power of minus 5 torque.
You will get gallium arsenide. But if you use harsh word that will be junk which in
the sense it will have lot of impurities which are coming from this ambience. If you want
to reduce the contamination level by another 10 to the power of minus 5 factor you have
to go down in vacuum by another factor of 10 to the power of minus 5. The vacuum level
that you are talking of here is some thing like 10 to the power of minus 10 to 10 to
the power of minus 11 torque. Very high vacuum, ultra high vacuum that is U H V is ultra high
vacuum and so the constituents of this m b e system are these, k cells which are the
sources of the operand with shutters, which can be opened or closed which ever you want.
All of them are heated. You close it, all the things you close, if not only gallium
arsenide. Open only these two shutters that is the idea and these shutters can be computer
controlled. What we are telling is very high vacuum is created and that is done by using
special pumps. I will not read out that.
There are all sorts of forms which are available are oil free. Now, what you have here is most
important, that I want to point out is I just have drawn a circle there or ellipse there.
This chamber is made up of stainless steel, high quality stainless steel. Iif you have
to get down to those vacuum, and before deposition you feed that chamber to high temperature
to degas and pump down all that is done. Not only is the chamber stainless steel, you know
when you have this, it looks totally closed. There are ports somewhere in these things.
There are ports, stainless steel projecting out with. You know gaskets and valves doors
which can be closed, that is not shown here. You get inside to put all these things wafer,
chlorine etc, you need those things. When you put close the chamber like that, there
will be a gasket. You cannot use any gasket that you like. What you use will be metal
gasket. Metal gaskets, what metal? hHold your breath, gold. Gold gaskets are used between
those two. Those are some of the things one does to get high vacuum and the entire process
of course is simple. Control the flow of gallium and arsenic by adjusting the temperature.
The beams which are coming, you get the beams because of the very high vacuum that is why
we call molecular beam. It is coming like that from another one like that on to the
substrate. You can get gallium arsenide, very high purity gallium arsenide 10 to the power
of 13, 14 etc.
Now there is a pumping system and you can get the advantage here is, you can grow 1
micron per hour, very small growth rate you can do when you grow very small rate of growth.
Very high purity, excellent crystal power quality and thickness uniformity over a 5
centimeter wafer you can get within 1%. People who want to make quantum wells, very small
layers, they resort to this. Today M O C V D is competing with this to grow thin layers.
Let me just warn you and leave you on that. This is M B E is very good and M O C V D is
competing because mass production if you want MOCVD is better than this. High purity material
can be grown and any composition we can get;, Indium, gallium, arsenide and also any dopants
that you can use. Key thing is control of thickness, control of impurities, precisely
done to the extent of a mono layer.
Some of the other advantages are that we can just say. You can actually introduce some
of the diagnostic tools along with this thing. Now let me not go into this in detail on to
that right now. Let me just sum up at this stage, the M B E system had gained lot of
popularity and today people are accepting to use this for some commercial applications
because of the high quality that you will get. Particularly those who want very thin
layers;, for example, aluminum, gallium, arsenide, then gallium arsenide a very thin layer at
the Armstrong level then we use those type of approach is highly attractive. This is
one of the most popular techniques which are used along with MOCVD. Other two techniques
are hydride and halide. It is more of academics. But still a poor man’s quality epilayers
will be some of those high qualities if you want you can use that halide process. With
that I think I will stop the discussion today. Few of the things which are meritorious for
gallium arsenide growth with M B E, I will bring out later. Then we will go on to the
devices afterwards. We will see in the next lecture.