Hi Genya and all
You are certainly right about the linear correlation between gas flow,
yield and ligit over large scales. Here are the results from recent unpol
runs:
flow dialed | sccm dialed | flow from yield | ligit
4.48 e16 a/s 0.05 2.83 e16 a/s 1.510 e-7
7.97 e16 a/s 0.08 3.69 e16 a/s 2.861 e-7
8.96 e16 a/s 0.10 7.82 e16 a/s 4.108 e-7
if one plot ligit over dial in sccm value, it is very much linear linear.
You are also cerntainly right when you say "RELATIVE".
Yield at 0.05sccm is 36% of the yield at 0.1sccm, ligit at 0.05sccm is
also 36% of the ligit at 0.1sccm. instead of 50%.
we had only 5 runs at 0.08sccm, statistics is too poor to say anything
with good confidence.
yield in 0.1sccm runs seems too high: 7.82/8.96 = 87%. reconstruction
software efficiency alone is lower than that.
Given these results, it seem to indicate the following:
1st. 0.1sccm dialing on MFC gives flow higher than 0.1sccm(1.3sccm
assuming detector efficiency is over all 70%).
2nd. 0.05sccm dialing on MFC gives 36% of whatever gas flow it is at
0.1sccm dial.
These are well with the design performance of the MFC though. We did the
calculation two years ago, that a 1sccm full scale MFC with 1% full scale
error calibrated by N2 gas, gives >= 25% measurement of 0.1sccm D2 gas
flow and >= 50% measurement of 0.05sccm D2 gas flow.
Given no absolute measurement of target flow, even with unpol gas flow
system, I am then facing the problem that I can not determine detection
efficiency or target density independently. Only thing I can say and so
far have been saying was: target density or flow is such and such
assuming detection efficiency is 100% (or an arbitrary plausible value)
like what I did above.
It seems to me that this situation must be improved.
For the sake of ABS diagnosing, it would be much better if we had
confidence in overall detector efficiency.
For detector diagnosing, it would be much better if we had confidence in
ABSOLUTE target density.
For the sake of myself, I need the ABSOLUTE target density and detector
efficiency to write into my thesis. Bill Bertozzi would give me an
extremely hard time if I write something like: target density is so and so
ASSUMING detector efficieny is so and so.
Since we have a simple alternative, why not?
Chi
This archive was generated by hypermail 2.1.2 : Mon Feb 24 2014 - 14:07:29 EST