Numbers make me cry

My job this week has been working on a delivery system for the Ar-36 gas that will be "plasmatized" (as of now, that is official jargon). I've got to come up with a way to get the gas from a tank into the ion source. It's not as simple as it sounds, unfortunately, and has been quite taxing on my sanity. The difficulty is being able to have this gas come into the ion source at a specific rate of flow, 0.02 std-cc/s specifically.

My first thought was to go to this company called Vacuum Technology Inc. (VTI), who operates in Oak Ridge, and have them design a specific reservoir and capillary leak (that's technical jargon for "tank that spews out gas") that fits our needs. My first calculation turned out to be super wrong when I estimated that given their maximum capacity tank we would only be able to run an experiment for around 14 hours before needing to refill. Since our experiments could run for weeks at a time, this isn't feasible.

I immediately threw that idea out and started looking into their open style capillary leaks, which differ in that you hook them up to your own source. The problem here started with the specs I have to provide to them: pressure coming into the device (known), the rate of flow (known), and the pressure coming out of the device (kind of known). Now, eventually the capillary goes into the ion source which is at vacuum. However, since we are transporting this gas a relatively long distance, there will be a difference in pressure from the exit of the leak to the entrance of the ion source. This pressure drop will affect the flow rate, which means if we want to have a final specific flow rate I need to figure out exactly how much pressure to have them calibrate this device to output.

It's easy, right! Just apply Poiseuille's equation, which relates pressure difference to rate of flow! Well, as I've come to find out, there are two different measures of flow rate: volumetric flow rate (measured in volume/second) and some other flow rate (measured in pressure*volume/second, which actually comes out to Joules/second). Poiseuille's equation uses volumetric flow rate and for most of the day I had been using the other kind. I also ran into numerous problems with determining the exact viscosity of Argon at the conditions we'd have it at as well as a cluster$%@# of unit conversions. Aghhh!

After my problems with the open capillary leak, my boss told me how to actually calculate how long my original plan would work for and it came out to 70 days or so. That's assuming a 1800 psi tank, which is approximately 122 atm, based on some specs I got from VTI. Thinking about it later, 122 atm is retardedly high for pressure so I'm going to have to call back and get those numbers again. Even if we can only squeeze 20 days out of a tank, that would work. Getting the capillary with reservoir calibrated is much easier than the open one so I'm hoping this works out.

I also examined getting one of the huge Argon gas tanks, getting a pressure regulator, and then attaching a flowmeter and adjusting the flow that way. But that wouldn't work because the flow rates possible with their tanks are way higher than what we need! Later I came to find out that I read their rates in std-cc/MINUTE and I was comparing that to std-cc/SECOND. So now that might actually be a viable option! Arghhhh!

The day wasn't wasted... I've learned a bit about how to be more careful with calculations and what not. I'm also that JJ (my boss) didn't get frustrated with me and wa shappy to help me out. I'm going to crunch the numbers again and give him a better report in the morning. Wish me luck? Please?