Hello,
I have a bit of a problem with a NIRIM style cvd diamond deposition system
and am looking for advice. Before I go any further please let me state that
I'm a laser jock and DONT have any formal experience with this hardware.
It's a project my company picked up on the cheap, and were just in the
'fooling around' stages...

For the sake of reference, a NIRIM cvd reactor is essentially little more
than a quartz tube that has been 'stabed' through a waveguide connected to a
2.45 gig microwave soucre and tuner. the tube is pumped down to 10 to 100
torr and process gasses are maintained by mass flow controllers. The MW
radiation sustains an arc and allows for the deposition of diamond while at
the same time free hydrogen etches any graphite phase carbon that may be
deposited onto a heated substrate holder in the plasma column.

Now for my problem. I have noticed silicon inpurities in diamond films
produced by this reactor. As I understand it these types of reactors are
known to have impurity problems due to etching of the reactor wall tube by
the MW plasma. In an effort to reduce this problem, why cant the tube be
centered in a solinoid that squeezes the plasma into the central portions of
the process tube (if etching of the process tube isnt stopped, it should at
least be lessened I would think). I checked into the literture, and the only
such arraingments I could find related to ECR style systems where very low,
large area dischages (in other reactor designs, NOT an NIRIM reactor) lead
to a non-isothermal plasma and the magnetic feild only serves to aid in
energy transfer from hot electrons to the heavier gas species.

So.. What am i missing here? Is there some reason I'm overlooking that would
inhibit me from making use of such a magnetic feild to reduce this
contamination issue?

Thanks in advance for any help!

(crossposted to sci.physics.fusion.. i'd post to physics.plasma, but it
seems dead in there...)

bob wrote:
Hello,
I have a bit of a problem with a NIRIM style cvd diamond deposition system
and am looking for advice. Before I go any further please let me state that
I'm a laser jock and DONT have any formal experience with this hardware.
It's a project my company picked up on the cheap, and were just in the
'fooling around' stages...
....
Now for my problem. I have noticed silicon inpurities in diamond films
produced by this reactor. As I understand it these types of reactors are
known to have impurity problems due to etching of the reactor wall tube by
the MW plasma. In an effort to reduce this problem, why cant the tube be
centered in a solinoid that squeezes the plasma into the central portions of
the process tube (if etching of the process tube isnt stopped, it should at
least be lessened I would think). I checked into the literture, and the only
such arraingments I could find related to ECR style systems where very low,
large area dischages (in other reactor designs, NOT an NIRIM reactor) lead
to a non-isothermal plasma and the magnetic feild only serves to aid in
energy transfer from hot electrons to the heavier gas species.

I think you may be right in you accessment. You may want to visits the
chapter on plasmas in Jackson to calculate the field you need. Another
solution would be to replace the SiO2 tube with another material like
Al2O3; though, I think SiO2 was chosen because C and Si bonds similiar.

You may also want to determine the amount of hydrogen in the films. As I
remember doing diamond-like films in the 80's the ratio of H to C can be
quite high while giving sp3 bonding. You can determine the amount of H
in the film by doing nuclear resonance measurement. For this you'll need
to partner with someone that can accelerate N15 to ~ 7 MeV. That person
should be able to give you the H to C ratio as a function of depth. My
experience in diamond-like films that the H to C ratio decreased
dramatically as the deposition temperature increased. For my case the
ratio changed by 50% as the deposition temperature went from 50 C to
~300 C. Also, the ratio of sp2 to sp3 bonds and the deposition rate
decreased with increasing temperature.

If you're reporting the result, try to include the sp2 to sp3 bonding
ratio. This can be gotten from IR spectroscopy. I like to use the dual
beam ratio method for this measurement as it gives the baseline
transmission as opposed to FTIR measurements where you need standards to
get the ratio.





So.. What am i missing here? Is there some reason I'm overlooking that would
inhibit me from making use of such a magnetic feild to reduce this
contamination issue?

Thanks in advance for any help!

(crossposted to sci.physics.fusion.. i'd post to physics.plasma, but it
seems dead in there...)

Thganks for your input. BTW, if you are suggesting I check for H in the
film, I'm sure the issues I'm seeing arent due to H inclusion, as I'm basing
the culpret (si) on a particular raman line observed in the spectrum.

If I switch to a saphire or alumina tube, wouldnt contamination still be an
issue? I realize there is an electronic similarity between Si and C, but I
assume aluminum will find it's way into the latice if the plasma is
contaminated with it.

also, fyi this work is NOT being done for academic reasons (dont want to
mislead anyone). I'm trying to build up a reactor that scavenges waste
process material, as I would eventually like to grow high grade single
crystal cvd diamond out of isotopically enriched feedstock to produce a
heatsink of as high as possible thermal conductivity. All of the cvd diamond
suppliers wanted a very large sum of money for a dedicated growth run, and
their machines all waste unused carbon, an unacceptable expensive scenario
with highly enriched feedstock. I know astex manufactured a system that was
designed to reuse the feedstock, but it's designed for large area deposition
that would not be acceptable for my particular aplication (i eventually need
to grow very thick films, and with the kind of run times I'll need diamond
would eventually nucleate on anything that was at temp and waste feedstock
material by depositing a film in places where i didnt want it to.

so in other words, i'm trying to re-invent the wheel just because i dont
like the shape of al lthe other wheels on the market today!



I think you may be right in you accessment. You may want to visits the
chapter on plasmas in Jackson to calculate the field you need. Another
solution would be to replace the SiO2 tube with another material like
Al2O3; though, I think SiO2 was chosen because C and Si bonds similiar.

You may also want to determine the amount of hydrogen in the films. As I
remember doing diamond-like films in the 80's the ratio of H to C can be
quite high while giving sp3 bonding. You can determine the amount of H
in the film by doing nuclear resonance measurement. For this you'll need
to partner with someone that can accelerate N15 to ~ 7 MeV. That person
should be able to give you the H to C ratio as a function of depth. My
experience in diamond-like films that the H to C ratio decreased
dramatically as the deposition temperature increased. For my case the
ratio changed by 50% as the deposition temperature went from 50 C to
~300 C. Also, the ratio of sp2 to sp3 bonds and the deposition rate
decreased with increasing temperature.

If you're reporting the result, try to include the sp2 to sp3 bonding
ratio. This can be gotten from IR spectroscopy. I like to use the dual
beam ratio method for this measurement as it gives the baseline
transmission as opposed to FTIR measurements where you need standards to
get the ratio.





So.. What am i missing here? Is there some reason I'm overlooking that
would
inhibit me from making use of such a magnetic feild to reduce this
contamination issue?

Thanks in advance for any help!

(crossposted to sci.physics.fusion.. i'd post to physics.plasma, but it
seems dead in there...)

res1aah7 wrote:
Thganks for your input. BTW, if you are suggesting I check for H in the
film, I'm sure the issues I'm seeing arent due to H inclusion, as I'm basing
the culpret (si) on a particular raman line observed in the spectrum.

If I switch to a saphire or alumina tube, wouldnt contamination still be an
issue? I realize there is an electronic similarity between Si and C, but I
assume aluminum will find it's way into the latice if the plasma is
contaminated with it.

I am not saying there might not be a problem with aluminum
contamination. I am saying I don't know but al2o3 is usually less
reactive than sio2. It may decrease your problem. But if you do get Al
into the films then you should have holes generated similiarly as
putting B into natural diamond or into Si. Should make the diamond a bit
more conductive. The product coming off of the SiO2 should be SiHx +H20
while if you used an al203 tube it may be AlH3 +H20. I am not sure which
one would be more volatile. Have you at least had an Auger Electron
Spectrum preformed on the films? I hope you don't see Si or O inside the
film or the contamination is very bad.

But just given another thought is that you may want to modify the
surface of your reactor tube to cover up the SiO2 or anyother non-carbon
material. I can't see your microwave sourse arrangement clearly with
your tube. So this suggestion may be awful. Can you line the tube with
high purity graphic tube? If so you may be able to get rid of the Si
problem.


also, fyi this work is NOT being done for academic reasons (dont want to
mislead anyone). I'm trying to build up a reactor that scavenges waste
process material, as I would eventually like to grow high grade single
crystal cvd diamond out of isotopically

In pratical terms what is the difference you suspect in the heat
conductivity for an isotropically rich film versus one made with
ordinary carbon; though, I suspect you are presentlly a long way from
the isotopic scattering dominating the heat conductivity.

enriched feedstock to produce a
heatsink of as high as possible thermal conductivity.

We all would.

All of the cvd diamond
suppliers wanted a very large sum of money for a dedicated growth run, and
their machines all waste unused carbon, an unacceptable expensive scenario
with highly enriched feedstock.

My suggestions on the analysis of the films were not just for academic
reasons. If you don't know what the film is you can't improve the
process. You need feedback of film composition, latttice parameters, and
quality to be able to optimize your process for you application. I don't
think you said what your substrate was but if you want single crystal
diamond of any macrosize then you'll have to go with diamond substrate.
True if you grow diamond on silicon the anti-phase boundaries will begin
to grow together and from larger crystallites. But the difference is
going from a crystallite size of a 10's of nm to a few microns. Also,
the stress in the film will prevent you from growing crack free films of
thickness of more than a few micros without removing the silicon substrate.

But if you just want to have a very large thermal conductivity, you may
not need a single crystal film especially if the heat is to be
transmitted through the thickness of the film as opposed to across its
surface. The grain boundaries then should not significantly impede the
heat flow.

I know astex manufactured a system that was
designed to reuse the feedstock, but it's designed for large area deposition
that would not be acceptable for my particular aplication (i eventually need
to grow very thick films, and with the kind of run times I'll need diamond
would eventually nucleate on anything that was at temp and waste feedstock
material by depositing a film in places where i didnt want it to.

. Have you at least had an Auger Electron
Spectrum preformed on the films? I hope you don't see Si or O inside the
film or the contamination is very bad.
not as of yet, thusfar all i have done is spectroscopic analysis (ftir and
raman)

But just given another thought is that you may want to modify the
surface of your reactor tube to cover up the SiO2 or anyother non-carbon
material. I can't see your microwave sourse arrangement clearly with
your tube. So this suggestion may be awful. Can you line the tube with
high purity graphic tube? If so you may be able to get rid of the Si
problem.

http://www.mksinst.com/pdf/ASTEXax7610DS.pdf
thats the reactor I'm using... i dont think carbon walls would be realistic
as they would absorb too much of the microwave energy..



In pratical terms what is the difference you suspect in the heat
conductivity for an isotropically rich film versus one made with
ordinary carbon; though, I suspect you are presentlly a long way from
the isotopic scattering dominating the heat conductivity.

in terms of what I have been able to make inhouse, or in terms of what
others can made in practice? 99.9% C12 is about a 60% improvment at room
temp. Thats signifigant enough for my aplication to be worth great trouble.

My suggestions on the analysis of the films were not just for academic
reasons. If you don't know what the film is you can't improve the
process. You need feedback of film composition, latttice parameters, and
quality to be able to optimize your process for you application. I don't
think you said what your substrate was but if you want single crystal
diamond of any macrosize then you'll have to go with diamond substrate.
right now I'm using thin plates of natural diamond. I have been in contact
with a group who has had descent results with growing cm+ sized of single
crystal heteroepitaxy (on thin films of Ir iirc the latice mismatch is only
a few percent) later on down the road.

bob wrote:
. Have you at least had an Auger Electron
Spectrum preformed on the films? I hope you don't see Si or O inside the
film or the contamination is very bad.

not as of yet, thusfar all i have done is spectroscopic analysis (ftir and
raman)

Hmm, since you are growing on diamond substrates the charging may
prevent you from having AES or XPS done unless the film grown is
conductive. If you can't get decent AES analysis you can also try a
laser ablation time-of-flight mass spectrometer. I know those are fairly
rare.

...




in terms of what I have been able to make inhouse, or in terms of what
others can made in practice? 99.9% C12 is about a 60% improvment at room
temp. Thats signifigant enough for my aplication to be worth great trouble.



That improvement is very good. Almost negates the need for research
money into trying to make N4C3.

Well, good luck with your work. I hope you find the extent, if any, of
the impurities in your material and a solution to it. The magnetic field
may help your problem, but you may want to do a calculation on the free
mean path length of the ions and compare it to the geometry of your
system to see if the magnetic field will be of any benefit. Of course
the magneit field will have no impact on the neutral species.