Calibrated leak as deuterium gas supply.

[Steven Sesselmann]

Hi guys,

My next project, which I have optimistically called F.I.C.S. fusion calls for a special gas supply, with extremely low flow rate. I intend to use a calibrated leak, so I picked up this one cheap on ebay, but before I take to it with a spanner, I was hoping that some of you might be able to give me a few clues...

Where does the leak actually sit, is it in the valve or inside the cylinder?

How would I go about filling this with Deuterium?

If his leak is calibrated for Helium at 1.1 x 10^-7, can I assume the leak would be similar for deuterium?

Any help would be appreciated

Picture:
The fitting on the left hand side does not seem to have any important function, I can see right through it and into the valve.

The nib sticking out on the right hand side is a protective sleeve, it unscrews, and there is a brass tube inside which has been squeezed flat and silver soldered.

Steven

[Frank Sanns]

The fill is on the right side which is the tube that is crimped and soldered. Not setup for an easy refill.

The valve is to shut the leak off but it is not very effective since the volume of gas between the valve and the leak is significant at the low flow rates.

The rate of a leak with another gas can be calculated from Grahms's Law: Rate1 / Rate2 = square root of (Mass2 / Mass 1)


Frank Sanns

[Edward Miller]

What is the size of the right hand side tube? You can probably put a Swagelok fitting on it.

[Alexi Hammond]

This leak will not work for deuterium. Helium leaks use He diffusion through a glass membrane. Only He has high enough diffusion rates in glass at room temperature.

Hydrogen diffuses very well through some metals at elevated temperature. We have made a hydrogen filter @work out of a thin-walled Ni tube heated directly by current (~6A) up to ~500C.

We have used this tubing:

http://www.alfa.com/en/GP100W.pgm?DSSTK=13465

Alex

[Chris Bradley]

What flow rate are you after, Steven, and what pressures are we talking in your chamber?

[Steven Sesselmann]

Hi Guys,

Thanks to all of you for informative replies. I took another picture of the nib, which is left quite short. The best option if I was to reuse this cylinder would be to drill out, or desolder the whole tube, and solder in a new tube.

However, if Ed Miller is correct, then this idea might be dead in the water.

To Chris, I do not know at this stage what the pressure in my chamber will be, though it might be possible to calculate this, as it would be a function of the voltage and temperature of my plasma.

It is critical to the design, that a continuous flow albeit small flows into the chamber. My starting point was going to be 1 sccm.

There are companies that will make up calibrated leaks specifically for deuterium. I have been quoted a price of $2000 for a leak designed to my specification.

Might be the way to go.

Steven

[Tyler Christensen]

Unless you need it to be precisely 1SCCM regardless of the pressure differential, why not just use a laser drilled orifice? They can be had for less than $100 on NPT fittings, and the flow math is easy to do for whatever gas you want (in fact I think there is a flow calculator on Lennox Laser's site)

Tyler

[Chris Bradley]

Steven Sesselmann wrote:
> It is critical to the design, that a continuous flow albeit small flows into the chamber. My starting point was going to be 1 sccm.
So are you looking for accuracy, or consistent & repeatable precision? It sounds like it might actually be the latter, if you do not yet know exactly what figure you want/need.

There are ways to achieve the latter, through flow measurement, with kit you probably have available already.

[George Schmermund]

Alexi is correct. What you have is a diffusion leak. When you get a helium leak detector you'll be glad that you didn't tamper with it. A helium leak detector is just about impossible to tune on the He peak without the cal leak. Also, the leak is designed to have the He reservoir diffuse into an atmosphere containing only the natural background partial pressure of He. If you keep the valve closed the partial pressure of the volume above the valve will increase to a level where the diffusion rate cannot maintain its calibration level. Some leaks can take quite a while to return to a steady state. The valve is just to help tune the mass spec with abrupt on/off switching. In steady state diffusion a typical leak of this sort will lose less the 0.5% of its calibration value per year. Therefore, keep the valve open.

[Edward Miller]

Alexi deserves the credit.

I love the Swagelok low flow metering valves. I will be there this week getting some parts so if you want me to send you two I'd be happy to. I've been able to get really steady vacuum pressures with two in serial. Here's a link to the parts

http://www.swagelok.com/search/find_pro ... ce&RPR=100

[Steven Sesselmann]

Thanks all,

Tyler, thanks for the link to Lennox Laser, this sounds like a good way to go. What would be the best fitting to have a hole drilled in? Having 2 or three fittings that stack on top of eachother would give some adjustment.

Chris, if you think back on my design, you will realize that anything like an MFC or pressure regulator is out of the question, it would be too complex to power it and control it. It needs to be a static leak.

Ed, I am familiar with the Swagelock needle valves, I used these in my MK1 STAR reactor. it was extremely hard to control the pressure in such a small chamber.

My approach with FICS is to have a steady flow of D2 into the chamber, and to control the pressure with a butterfly valve on the vacuum side only. Maybe a bit wasteful, but far simpler in my case.

Steven

[Chris Bradley]

Steven Sesselmann wrote:
> Chris, if you think back on my design, you will realize that anything like an MFC or pressure regulator is out of the question, it would be too complex to power it and control it. It needs to be a static leak.

That's why I am asking if you are actually seeking accuracy, or just precision, because I think there are straight-forward, hands-off mechanical solutions to obtain constant [but not necessarily accurate] flow.

If you obtain a flow orifice that will give you ~1sccm at ~1atm differential, then set up a regular gas regulator (pref a two stage, for more accurate regulation) and adjust it until you get the flow rate you want. The flow rate will remain as steady as the pressure regulator maintains a steady pressure on the orifice, which you can check and monitor by operating a pressure gauge directly above your turbo. The turbo will pump at a constant rate, so the pressure above it is a direct reading of your throughput, and using that reading you can check and ensure the flow is regular and repeatable.

I do this in my setup and in a static hands-off configuration the pressure gauge I fit above my turbo tells me I can run a steady 1 sccm with +/- 5% for a 20 minute run on the gas containted in the HP side of the regulator [alone]. The '1 sccm' may not be accurate, because I assume the turbo flow rate to be as nominally stated, but I can measure the consistency of flow easily enough.

[Steven Sesselmann]

Chris Bradley wrote:
>set up a regular gas regulator (pref a two stage, for more accurate regulation) and adjust it until you get the flow rate you want.

Yes, I think this is the way to go, a small sample cylinder with regulator, and a flow orifice.

Swagelock regulators aren't cheap, but I could get lucky on ebay..

Steven

[Tyler Christensen]

I got mine in a 1/8" NPT nipple. I'm not so sure about cascading multiple orifices, it seems to me that it would takes hours for the intermediary pressures to stabilize to whatever equilibrium they'd want to be at, but I could be wrong about that (I've just never really liked the idea of multiple restrictions in cascade for precision gas flow control). You should be able to get the flow you want with a single orifice anyways. Like you said, you will primarily control the tank pressure with your vacuum valve on top of the high-vac pump. You can also get some control by changing the back pressure on the orifice with your regulator pressure, since the pressure differential across the orifice controls flow rate. If I recall correctly (don't quote me on this) I had a 7 micron orifice with about 5PSIG behind the orifice (and of course ~0PSIA on the other side) and was getting about 1.5SCCM flow.

My orifice clogged up (something to be careful of, especially if you are continually bringing chamber up to air and working on the system) and now I use a swagelok needle valve and just leave it at a fixed position and control chamber pressure with my vacuum valve, personally I like that better.

[Steven Sesselmann]

Tyler,

Forgive my ignorance, but do you weld a plate in the nipple first?

If so, how thick does it need to be?



Steven

[Tyler Christensen]

They manufacture it fully, you get a nipple in the mail that has a plate in it with the hole (so small you can't even see light through it, assuming you're in the <10 micron range). No assembly required.

http://www.lenoxlaser.com/index.php?dis ... gory_id=13

No need for calibrated (unless you want to be able to calculate more precisely the actual flow)

Tyler

[Steven Sesselmann]

Thanks Tyler,

Too easy...

Steven