Geo,
Well this sucks. I think
you¡¯re right and they are hardening the beam by dumping the low energy components
to reduce skin dose. What is your x-ray tube manufacturer and model number? Is
this a micro focus tube?. I thought that¡¯s what we were using up in Las Vegas. What¡¯s
the max operating voltage of the tube? I¡¯m sure you¡¯re right and is there is an
Al filter in there knocking down the <15 keV portion. That may standard on
medical x-ray tubes. Well that sucks.
The skimming set up as I understand
it is used for thin film analysis not general XRF sample analysis and isn¡¯t as efficient
as a standard set up. It¡¯d be interesting to see the difference between the
two.
?
Geo, I don¡¯t understand the
logic on continuing to use a wrong instrument setup with the 93 keV energy
span. Explain to me how it¡¯s not directly comparable. It has half the energy
resolution where an optimal 64 keV span will double that, both will show the
XRF but one will have trouble resolving overlapping peaks. You¡¯re needlessly degrading
the instrument¡¯s response and limiting the ability to see small important
features. Just to be compatible with a bad set up? That doesn¡¯t make sense.
Correct the problem now you¡¯ll be glad you did. If you look at the these
spectra they are very noisy for a 30 sec x-ray tube shot and going to 4096
channels is going to make it worse for no gain due to the detectors resolution limit.
A detector with a 0.2 keV FWHM is not going to resolve any peak separations close
to that no matter how many channels you are using. You have a perfect peaking
time / filter set up here but the system needs to be optimized for the energy range
of interest or you¡¯re just throwing out half your capability.
Fix the problem now and
move on.
Dud
?
?
?
That's correct. Same as
the ratio chart I showed.
Nothing at all at 15 keV
or lower.?
Looking at the beam
directly with a sensor tells why. There must be a compound filter in or right
at the output centered on tungsten L-XRF lines (~8.40/8.67keV). Like a high
pass filter in electronics. I can't identify the small peaks from the filter,
which is why it is probably a graded stack.
The two L line are there
but weaker than the bremsstrahlung peak to the right .I often wondered where is
the lower portion of the Brems continuum, besides the L peaks, it goes down to
the floor at the low end. Turned al, the way open, there appears another
continuum at the low end but still that gap between the two humps.
Makes sense, these tubes
are manufactured to drive X-Ray image intensifiers, where the lowest energies
are of no use and should be eliminated. They did a good job of doing so.
Like you mentioned, the
most energy is in the mid point or "top" of the hump. Merely setting
the cutoff point to a little above K-Edge does not give much output there as
that is the right hand bottom of the hump, so that's why there is a lag between
the K-edge and the HVpeak setting. So the top of the hump is where the K-edge
is set and that does work as most of the impressive peaks are on the left hand
slope of the hump.
This tube I've been using
for this since the first is a unitized module, the X-Ray tube, HV generator,
heat sink and X- shielding are all one sealed piece. Not much could be done
even if I wanted to.
I don't do any background
subtracting or software smoothing, so what my .mca's show is untampered with.
Right now the whole setup
is turned on it's side instead of facing straight up like before. Doing so
allowed an almost perfect 45 degrees between the beam and the detector, and the
results are good, better than the 90 degrees I use with the chamber installed.
But the convenience and safety features are not as good that way, so I'll stick
with the chamber for most routine stuff, and the skimming for those Ca vs K
determinations.
Meantime I'll stick with
93 keV top end and when finished messing around, switch back to 4096 channels.
If I change the top end, all the serious testing I've done that way won't be
directly comparable to future tests.
----- Original Message
-----
From: Dude <dfemer@...>
To: [email protected]
Sent: Sat, 23 Jan 2021 13:36:26 -0500 (EST)
Subject: Re: [XRF] Obsidian
margin: 0.0in;
font-size: 12.0pt;
font-family: "Times New Roman" , serif;
}
a:link, span.MsoHyperlink {
color: blue;
text-decoration: underline;
}
a:visited, span.MsoHyperlinkFollowed {
color: purple;
text-decoration: underline;
}
p {
margin-right: 0.0in;
margin-left: 0.0in;
font-size: 12.0pt;
font-family: "Times New Roman" , serif;
}
p.MsoNoSpacing, li.MsoNoSpacing, div.MsoNoSpacing {
margin: 0.0in;
font-size: 16.0pt;
font-family: Arial , sans-serif;
color: black;
}
span.EmailStyle19 {
font-family: Arial , sans-serif;
color: black;
top: 0.0pt;
letter-spacing: 0.0pt;
font-weight: normal;
font-style: normal;
}
*.MsoChpDefault {
font-size: 10.0pt;
}
div.Section1 {
page: Section1;
}
/*]]>*/
Geo,
Moist of these MCA files
have nothing in them
50kV, 40, 30 seem ok but
20 have very few counts and 15 has no counts same with all. You are getting no
signal below 30
The data are very noisy as
I think is due to the skimming and the collimator.
Rerun these using the
standard 45 degree shot and use a beam current around 40 or 50 and no narrow
collimator. Try for a spot size of about 6mm or larger
Dud
?
?
?
?
Theory says Fe and Ca should XRF with quite low energy excitement
=
"just above the K-edge".
. Fortunately the test was left as-is last night, so it was easy to redo and
save example .mca's. so here's a fresh set of .mca that show the reality of
this skimming arrangement.?
Note the Ca, Fe and Ba peaks as exemplars. Also remember the possibility of
stray "W" L-lines due to the Si-PIN collimator..?
The .mca marked as "ALLkVp" should contain the others too, if not
I've included each individual .mca's.
Once I'm able to strike this test, I'll take a repeat this test wiht the
standard 45 degree arrangement of exciter and detector angle.
Then a direct Gamma Scan of the beam itself to see how the internal tungsten
target characteristic L lines look. If they are being filtered internally that
would explain a lot.
Geo
![Skimming-1mm-Beam-Corizozo-Shield-Lava-ALLkVp.30s.png]()
