Welcome to the XRF Wiki. This Wiki is a repository of information contributed by members of [email protected]. Members can view and edit the pages. The pages are currently not viewable by the public. The pages below represent a rough overview of the technology and techniques associated with X-Ray Florescence Spectroscopy as well as serving as a place for members to organize the spectra that they have contributed to the forum.?
As a starting point, the pages will be populated by information taken from member posts from the past few years - with references/attribution. Perhaps in time, these pages can be edited by members to make them more complete and less choppy. Members can also edit this page, making the structure of the content more organized.?
Pages can contain information about a particular topic, links to relevant resources (such as manuals, research papers, etc.), links to relevant forum discussions on the topic, book titles, or anything that might be useful.
What follows is very much a work in progress.
?
Adding to the Wiki
For those unfamiliar with the Wiki phenomenon, it is basically a user editable encyclopedia. The idea is that there are pages with different articles, which the user can read and also choose to edit if desired. If a piece of information is incorrect, a link outdated, or the presentation choppy or unclear, the user can just click "Edit Page" at the bottom and then fix the mistake. There is also a "Page History," so if you a user were to make a mistake or maliciously destroy a page (which wouldn't happen here) an editor can come along and restore the page to a previous version.?
If you have used MediaWiki (the engine that runs behind the scene of Wikipedia.org) you will be familiar with a certain style of creating an article. There are tags that one uses to change the formatting of the parts of the article. The Wiki on Groups.io uses a different background engine and so the standard MediaWiki syntax does not apply here. Rather, the page is edited using the same tools as one would use to compose a message on the forums. If you are a programmer or someone who likes using the tagged syntax directly, you can access the underlying page source code, be activating the advanced editing toolbar (the icon with three lines on the far right of the basic toolbar) and then clicking on the source code icon <> on the far right of the Advance Editing Toolbar. You will immediately notice that the underlying source code is HTML. Most things can be done just by using the various icons, but advanced features like table can only be implemented by manually writing the source code. (Although there are website available that will generate the source code for you through a graphical interface.)
This section describes the hardware used for XRF. We will describe the basic theory of operation of each stage in the processing tool chain and then present different commercial systems that are available as these systems are often integrated together.
[Editor note: General theory needs to be in an independent page separate from specific manufacturer information.
[Editor note - question: Should this be arranged by category, e.g detectors, preamps, etc or by manufacturer with the manufacturers product line on one page. Perhaps it's best to put a list of products with theory but the details of the products on a manufacture page...]
Detectors?- what's available, theory of operation, tradeoffs
Detector Cooling?- keeping detectors cool to avoid thermal noise and keeping the TEC from overheating
Preamps?- theory of operation, brief summary of what out there (and what not to do - such as trying to use a PMT preamp for a SiPIN diode...)
Pulse Processing Theory - high level overview of the stages of going from detector pulse to channel peaks on the computer
Commercial Systems - many commercial systems are integrated so it seems to make sense to present, for example all Amptek products together. [question: are their other affordable integrated systems besides Amptek?]
On the longer PT time issue, I agree, but when I tried to go to 32, it was not offered, only up to 25.6. Maybe there's another setting that when selected limits the PT range? I'm sure I saw 32 and maybe longer? offered before.......
Sorry the cap stays on for protection, it is one of my modified caps with a Kapton window.
25.6 like last run, but auto threshold adjust applied. 8500s update .mca below. Didn't help?
"noise" at low end, did noticeably improve definition of peaks above 3 keV.
3.3 yes it is Ka1 of Potassium. But let's think about element K. for a moment, a portion of it is K-40 and that has something like 10% decay path to Ar gas. K-40 is loaded with 3.19 2.98 and 3.19 keVs. Then add the stable Ar in the air which is really easily ionized, you might be getting a huge dose of Ar XRF because of? the radioactive source.
Also the "noise" at the low end- I'm wondering if it has more to do with the X-Ray photon's photopeak being scattered? by O2, and other air molecules.?
Geo
? K at 3.3. noise at 4.5?? Cr at 5.42 and U Ll at 11.64
Hi Charles, close but no Am involved. The exciter is natural U-238, we are seeing the decay of the daughter Pa-234m (also known as UX2). We are essentially seeing only the birth of the?U-234daughter from? Pa-234m, just that part of the natural U-238 decays series.
When depleted U has been chemically separated from its daughters you wind up with only U-238, the quickly the ingrowth of the first 3 daughters Th-234>Pa-234m>U234.? U-234 has a 240,000 year half-life, so lower daughters will eventually grow back in, but in geological time scales.?
The disc is actually marked "UX2 beta, and is source of Pa-234m beta particles, but we're looking deeper than that.
A picture of that decay series and its STOP sign is in an earlier post.
----- Original Message ----- From: Charles David Young <charlesdavidyoung@...> To: XRF <[email protected]> Sent: Fri, 30 Oct 2020 22:27:34 -0400 (EDT) Subject: Re: [XRF] 59.5keV spectrum cleanup
I am not sure exactly what I am looking at here but I like the sharpness of the peaks and the low noise.? Is this a scan with an Am241 exciter?? See attached.
Charles
----- Original Message ----- From: Charles David Young <charlesdavidyoung@...> To: XRF <[email protected]> Sent: Fri, 30 Oct 2020 22:27:34 -0400 (EDT) Subject: Re: [XRF] 59.5keV spectrum cleanup
I am not sure exactly what I am looking at here but I like the sharpness of the peaks and the low noise.? Is this a scan with an Am241 exciter?? See attached.
25.6 like last run, but auto threshold adjust applied. 8500s update .mca below. Didn't help?
"noise" at low end, did noticeably improve definition of peaks above 3 keV.
3.3 yes it is Ka1 of Potassium. But let's think about element K. for a moment, a portion of it is K-40 and that has something like 10% decay path to Ar gas. K-40 is loaded with 3.19 2.98 and 3.19 keVs. Then add the stable Ar in the air which is really easily ionized, you might be getting a huge dose of Ar XRF because of? the radioactive source.
Also the "noise" at the low end- I'm wondering if it has more to do with the X-Ray photon's photopeak being scattered? by O2, and other air molecules.?
Geo
? K at 3.3. noise at 4.5?? Cr at 5.42 and U Ll at 11.64
25.6 like last run, but auto threshold adjust applied. 8500s update .mca
below. Didn't help?
"noise" at low end, did noticeably improve definition of peaks
above 3 keV.
3.3 yes it is Ka1 of Potassium. But let's think about element K. for a moment,
a portion of it is K-40 and that has something like 10% decay path to Ar gas.
K-40 is loaded with 3.19 2.98 and 3.19 keVs. Then add the stable Ar in the air
which is really easily ionized, you might be getting a huge dose of Ar XRF
because of? the radioactive source.
Also the "noise" at the low end- I'm wondering if it has more to do
with the X-Ray photon's photopeak being scattered? by O2, and other air
molecules.?
Geo
? K at 3.3. noise at 4.5?? Cr at 5.42 and U Ll at 11.64
I am not sure exactly what I am looking at here but I like the sharpness of the peaks and the low noise.? Is this a scan with an Am241 exciter?? See attached.
25.6 like last run, but auto threshold adjust applied. 8500s update .mca below. Didn't help?
"noise" at low end, did noticeably improve definition of peaks above 3 keV.
3.3 yes it is Ka1 of Potassium. But let's think about element K. for a moment, a portion of it is K-40 and that has something like 10% decay path to Ar gas. K-40 is loaded with 3.19 2.98 and 3.19 keVs. Then add the stable Ar in the air which is really easily ionized, you might be getting a huge dose of Ar XRF because of? the radioactive source.
Also the "noise" at the low end- I'm wondering if it has more to do with the X-Ray photon's photopeak being scattered? by O2, and other air molecules.?
Geo
? K at 3.3. noise at 4.5?? Cr at 5.42 and U Ll at 11.64
25.6 like last run, but auto threshold adjust applied. 8500s update .mca below. Didn't help?
"noise" at low end, did noticeably improve definition of peaks above 3 keV.
3.3 yes it is Ka1 of Potassium. But let's think about element K. for a moment, a portion of it is K-40 and that has something like 10% decay path to Ar gas. K-40 is loaded with 3.19 2.98 and 3.19 keVs. Then add the stable Ar in the air which is really easily ionized, you might be getting a huge dose of Ar XRF because of? the radioactive source.
Also the "noise" at the low end- I'm wondering if it has more to do with the X-Ray photon's photopeak being scattered? by O2, and other air molecules.?
Geo
? K at 3.3. noise at 4.5?? Cr at 5.42 and U Ll at 11.64
From: GEOelectronics@... To: [email protected] Sent: Friday, October 30, 2020 4:58:38 PM Subject: Re: [XRF] 59.5keV spectrum cleanup
Click on LINEAR view, looks pretty good to me, the L X-rays are the subject, but you are right I didn't do the threshold routine. Do you think 25.6 is too slow? I'm afraid if something with high count rates comes along it will not be the best, maybe 19 was a better compromise ?
Geo
----- Original Message ----- From: Dude <dfemer@...> To: [email protected] Sent: Fri, 30 Oct 2020 17:18:57 -0400 (EDT) Subject: Re: [XRF] 59.5keV spectrum cleanup
It looks like the low energy electronic noise is a problem below 1.8 keV. You¡¯ll need to try resetting the threshold filters or try a higher PT as noise counts are 10 times your signal
Click on LINEAR view, looks pretty good to me, the L X-rays are the subject, but you are right I didn't do the threshold routine. Do you think 25.6 is too slow? I'm afraid if something with high count rates comes along it will not be the best, maybe 19 was a better compromise ?
----- Original Message ----- From: Dude <dfemer@...> To: [email protected] Sent: Fri, 30 Oct 2020 17:18:57 -0400 (EDT) Subject: Re: [XRF] 59.5keV spectrum cleanup
It looks like the low energy electronic noise is a problem below 1.8 keV. You¡¯ll need to try resetting the threshold filters or try a higher PT as noise counts are 10 times your signal
It
looks like the low energy electronic noise is a problem below 1.8 keV. You¡¯ll
need to try resetting the threshold filters or try a higher PT as noise counts
are 10 times your signal
Looks
like Al at 1.4 keV and P at 2.01 but these need a MUCH longer count. K at 3.3. noise
at 4.5? ?Cr at 5.42 and U Ll at 11.64 ?and Zn 8.64 and 9.57 keV
Needs
way more counts before I put my signature on it
From: "DFEMER" <dfemer@...> To: [email protected] Sent: Friday, October 30, 2020 1:02:37 PM Subject: Re: [XRF] 59.5keV spectrum cleanup
It
would come off of the ROI. The FWHM will change with energy so its specific to
a peak.
In
DPPmca set an ROI on the peaks using the ^ peak like thing on the tool bar. Use
the cursor and left click the low side then left click the right side. Be careful
of the start and stop energies in the window they sometimes get out of wack, Add
the ROI and do the next one and when done click OK. Then go back to the analyze
tab and select ROI details, they¡¯ll all be listed there with the other useful
info. I always set my ROI¡¯s manually as the search takes too long to set up
correctly.
'
You¡¯ll note that the 19.2 peaking time has a
much better 13.6 keV FWHM resolution than the 4.8 (0.192 vs .396 FWHM)."
Dud,THANKS for that update and analysis>
is there a report log to determine FWHM for a particular
peak or do you have to set ROI, and if so how do you invoke FWHM?
Geo
?
From:
"DFEMER"
<dfemer@...> To: [email protected] Sent: Friday, October 30, 2020 11:40:36 AM Subject: Re: FW: [XRF] 59.5keV spectrum cleanup
?
Geo,
The
4.8 usec doesn¡¯t have the filter fast ?thresholds set correctly as a
result it has you seeing 3135 valid counts with a noise count of 128,980.While
The 19.2 PT has 4534 valid counts with a noise count of 56627 counts and both
with comparable count times. (Note that the 19.2 still has a lot of low energy
noise that needs a better threshold setting but its way better than the 4.8 PT
data set ). So the 19.2 has better count statistics 4534 vs 3135 counts for roughly
the same count time counts and is getting better peak shape. You¡¯ll note that
the 19.2 peaking time has a much better 13.6 keV FWHM resolution than the 4.8
(0.192 vs .396 FWHM).
Thanks
for that report Charles. Of the two, can you say what is the lowest peak that
can be resolved and is there a difference between them?? My just finished
overnight run using the Si-PIN on a weakly radioactive sample for Gamma Spec
(not excited XRF) is resolving individual peaks down to less than 1.5 keV. In
the Peaking Time topic thread I'll post that test @19.2us PT along with a
nearly similar test done last February at 4.8us. The main difference between
those seems to be both in the clarity and narrowness of the big peaks, and
better performance at the lowest end going to the 19.2.
?
Today
I'll try even longer peaking times, up to 32us just for the historical
record.?
I
think you've been using 32us and I can't fault your scans, that's for sure.
?
Geo
?
-----
Original Message -----
From: Charles David Young <charlesdavidyoung@...>
To: XRF <[email protected]>
Sent: Fri, 30 Oct 2020 08:26:25 -0400 (EDT)
Subject: Re: FW: [XRF] 59.5keV spectrum cleanup
?
I
did 2 runs on a brannerite specimen.? One like normal (red) and the other
with a paper adhesive label covering the buttons.? The proportions of the peaks
pretty much look the same to me.
?
Charles
?
On
Wed, Oct 28, 2020 at 11:30 AM Dude <dfemer@...> wrote:
Charles,
My
mistake, I now see the problem you¡¯re using a log energy axis. That should
always
be linear, counts can be log, lin or sqr as appropriate.
Looking
at the png you sent there is something seriously wrong on the low energy below
4 keV. Send the mca file for a look as it may just be a Theremino problem. For
these low energies you¡¯ll want to keep the Np in there, but to pull these
elements out you really need a high flux rate which the Am source isn¡¯t doing.
Even with a tube the efficiencies are very poor and the counts are low making
it hard to quantify. However, the Al Si, P, S, Ca and K ?are absolutely
essential to determining the basic mineralogy types and every effort should be
made to verify their presence.
To
get the count times up readjust the gain to look only at the low energy range
say from 10 on down and see ff that helps pull them out. Using a tight focused
beam and close up to the target will also help concentrate the flux.
Dud
?
From:[email protected]
[mailto:[email protected]] On
Behalf Of Charles David Young Sent: Wednesday, October 28, 2020 8:12 AM To: XRF Subject: Re: FW: [XRF] 59.5keV spectrum cleanup
?
Geo,
?
I agree and I did not mean to dis
this exercise, even if it
turns out to be primarily academic.
?
When I get a chance I'll try the
paper filter with a mineral
like Brannerite (UTi2O6) to see how it affects the Ti peak at 4.5keV.?
That is typically the lowest element that is essential to my application.?
Other peaks like P, K, Ca would be nice but I have learned to live without
them.? If we could somehow dig those out by an adjustment it would be
great.
"The filter comes at a
cost.? Am I reading it
correctly that it reduced the input count to %25 of the non-filtered?? Do
we need to wait 4x longer for equal counts when this is used as an exciter?
"
?
All that is to be determined
Charles. It well may be much
worse than that, if the low energies in the beam are what's helping to dig out
the low energy XRF signals we are looking for.
?
I agree about our perspective
points of view too.?
?
You system is as ideal as can be
for your main task, and
your display program is an advancement in the area of display and
identification, especially for the audience of that group. Now the photography
meets the same standard too. No one has added all those dimensions in one place
to the amateur mineral group before.? All for science' sake, and I applaud
that.
?
For you and others I would
recommend trying a paper filter
over the exciter source, just to see.
?
The rest is for the X-Ray Physics
and X-Ray Optics crowd. We
now have an inexpensive but powerful tool in our toolbox that wasn't there 2
days ago.
The filter comes at a cost.?
Am I reading it correctly
that it reduced the input count to %25 of the non-filtered?? Do we need to
wait 4x longer for equal counts when this is used as an exciter?? Was all
that low energy activity that is filtered out detrimental or beneficial when
the button is used to stimulate XRF in the target?
?
I come at this from a different
perspective.? I use the
Si-PIN with an Am241 exciter to identify minerals.? In my experience the
XRF coming from the target is much stronger than the noise introduced by the
Am241, which is mounted on a shield and facing away from the detector.?
Attached is a scan of xenotime, which shows a large Y component as well as some
U and Pb.? I turned on the Am241 reference lines to show that the Np and
Au peaks are very small in comparison.? One quickly learns to recognize
and ignore them when identifying the important peaks that emanate from the
target: Ti, Mn, Fe, Ta, Th, U, Y, Zr, Nb, REEs.
After trying many slight thickness
variations of Source
Filter components, the best result is show in the second picture.
First up is the spectra coming out of the Am exciter source with just a paper
alpha blocking filter. That alone made significant improvements at the low end
vs no alpha filter. The goal of this project is to apply a filter to the source
that will effectively remove most of the low energy component that is caused by
Np- daughter and self-XRF of elements in the button, such as Au, Ag, Fe.
Error! Filename not specified.
Now a filter applied to the source and another Gamma Spectrum ran.
Error! Filename not specified.
Thanks to member Dudley Emer for the design of the filter elements and
thicknesses,
based on selective filtering using the K-edge X-Ray absorption effect.
It
would come off of the ROI. The FWHM will change with energy so its specific to
a peak.
In
DPPmca set an ROI on the peaks using the ^ peak like thing on the tool bar. Use
the cursor and left click the low side then left click the right side. Be careful
of the start and stop energies in the window they sometimes get out of wack, Add
the ROI and do the next one and when done click OK. Then go back to the analyze
tab and select ROI details, they¡¯ll all be listed there with the other useful
info. I always set my ROI¡¯s manually as the search takes too long to set up
correctly.
'
You¡¯ll note that the 19.2 peaking time has a
much better 13.6 keV FWHM resolution than the 4.8 (0.192 vs .396 FWHM)."
Dud,THANKS for that update and analysis>
is there a report log to determine FWHM for a particular
peak or do you have to set ROI, and if so how do you invoke FWHM?
Geo
?
From:
"DFEMER"
<dfemer@...> To: [email protected] Sent: Friday, October 30, 2020 11:40:36 AM Subject: Re: FW: [XRF] 59.5keV spectrum cleanup
?
Geo,
The
4.8 usec doesn¡¯t have the filter fast ?thresholds set correctly as a
result it has you seeing 3135 valid counts with a noise count of 128,980.While
The 19.2 PT has 4534 valid counts with a noise count of 56627 counts and both
with comparable count times. (Note that the 19.2 still has a lot of low energy
noise that needs a better threshold setting but its way better than the 4.8 PT
data set ). So the 19.2 has better count statistics 4534 vs 3135 counts for roughly
the same count time counts and is getting better peak shape. You¡¯ll note that
the 19.2 peaking time has a much better 13.6 keV FWHM resolution than the 4.8
(0.192 vs .396 FWHM).
Thanks
for that report Charles. Of the two, can you say what is the lowest peak that
can be resolved and is there a difference between them?? My just finished
overnight run using the Si-PIN on a weakly radioactive sample for Gamma Spec
(not excited XRF) is resolving individual peaks down to less than 1.5 keV. In
the Peaking Time topic thread I'll post that test @19.2us PT along with a
nearly similar test done last February at 4.8us. The main difference between
those seems to be both in the clarity and narrowness of the big peaks, and
better performance at the lowest end going to the 19.2.
?
Today
I'll try even longer peaking times, up to 32us just for the historical
record.?
I
think you've been using 32us and I can't fault your scans, that's for sure.
?
Geo
?
-----
Original Message -----
From: Charles David Young <charlesdavidyoung@...>
To: XRF <[email protected]>
Sent: Fri, 30 Oct 2020 08:26:25 -0400 (EDT)
Subject: Re: FW: [XRF] 59.5keV spectrum cleanup
?
I
did 2 runs on a brannerite specimen.? One like normal (red) and the other
with a paper adhesive label covering the buttons.? The proportions of the peaks
pretty much look the same to me.
?
Charles
?
On
Wed, Oct 28, 2020 at 11:30 AM Dude <dfemer@...> wrote:
Charles,
My
mistake, I now see the problem you¡¯re using a log energy axis. That should
always
be linear, counts can be log, lin or sqr as appropriate.
Looking
at the png you sent there is something seriously wrong on the low energy below
4 keV. Send the mca file for a look as it may just be a Theremino problem. For
these low energies you¡¯ll want to keep the Np in there, but to pull these
elements out you really need a high flux rate which the Am source isn¡¯t doing.
Even with a tube the efficiencies are very poor and the counts are low making
it hard to quantify. However, the Al Si, P, S, Ca and K ?are absolutely
essential to determining the basic mineralogy types and every effort should be
made to verify their presence.
To
get the count times up readjust the gain to look only at the low energy range
say from 10 on down and see ff that helps pull them out. Using a tight focused
beam and close up to the target will also help concentrate the flux.
Dud
?
From:[email protected]
[mailto:[email protected]] On
Behalf Of Charles David Young Sent: Wednesday, October 28, 2020 8:12 AM To: XRF Subject: Re: FW: [XRF] 59.5keV spectrum cleanup
?
Geo,
?
I agree and I did not mean to dis
this exercise, even if it
turns out to be primarily academic.
?
When I get a chance I'll try the
paper filter with a mineral
like Brannerite (UTi2O6) to see how it affects the Ti peak at 4.5keV.?
That is typically the lowest element that is essential to my application.?
Other peaks like P, K, Ca would be nice but I have learned to live without
them.? If we could somehow dig those out by an adjustment it would be
great.
"The filter comes at a
cost.? Am I reading it
correctly that it reduced the input count to %25 of the non-filtered?? Do
we need to wait 4x longer for equal counts when this is used as an exciter?
"
?
All that is to be determined
Charles. It well may be much
worse than that, if the low energies in the beam are what's helping to dig out
the low energy XRF signals we are looking for.
?
I agree about our perspective
points of view too.?
?
You system is as ideal as can be
for your main task, and
your display program is an advancement in the area of display and
identification, especially for the audience of that group. Now the photography
meets the same standard too. No one has added all those dimensions in one place
to the amateur mineral group before.? All for science' sake, and I applaud
that.
?
For you and others I would
recommend trying a paper filter
over the exciter source, just to see.
?
The rest is for the X-Ray Physics
and X-Ray Optics crowd. We
now have an inexpensive but powerful tool in our toolbox that wasn't there 2
days ago.
The filter comes at a cost.?
Am I reading it correctly
that it reduced the input count to %25 of the non-filtered?? Do we need to
wait 4x longer for equal counts when this is used as an exciter?? Was all
that low energy activity that is filtered out detrimental or beneficial when
the button is used to stimulate XRF in the target?
?
I come at this from a different
perspective.? I use the
Si-PIN with an Am241 exciter to identify minerals.? In my experience the
XRF coming from the target is much stronger than the noise introduced by the
Am241, which is mounted on a shield and facing away from the detector.?
Attached is a scan of xenotime, which shows a large Y component as well as some
U and Pb.? I turned on the Am241 reference lines to show that the Np and
Au peaks are very small in comparison.? One quickly learns to recognize
and ignore them when identifying the important peaks that emanate from the
target: Ti, Mn, Fe, Ta, Th, U, Y, Zr, Nb, REEs.
After trying many slight thickness
variations of Source
Filter components, the best result is show in the second picture.
First up is the spectra coming out of the Am exciter source with just a paper
alpha blocking filter. That alone made significant improvements at the low end
vs no alpha filter. The goal of this project is to apply a filter to the source
that will effectively remove most of the low energy component that is caused by
Np- daughter and self-XRF of elements in the button, such as Au, Ag, Fe.
Error! Filename not specified.
Now a filter applied to the source and another Gamma Spectrum ran.
Error! Filename not specified.
Thanks to member Dudley Emer for the design of the filter elements and
thicknesses,
based on selective filtering using the K-edge X-Ray absorption effect.
?I
don¡¯t see much difference either, Maybe just a very slightly lower count ratio
of the 2 low energy peaks of the covered run.
Your
MCA files have no calibrates in them, that really has to be in there for us to
make sense of the spectra. You can also greatly improve these really long count
times by running 2048 channels, setting the thresholds and using a slower
peaking time. ?Short peaking times are used for high count rates and have more
noise and lower resolution. Low energy noise ties up the counting time, in this
case you¡¯re only getting 1 good count out of 6 ro 7 noise counts. Why are you
using a log energy x- axis in the Theremino plots? ??
From:[email protected]
[mailto:[email protected]] On Behalf Of Charles David Young Sent: Friday, October 30, 2020 5:26 AM To: XRF Subject: Re: FW: [XRF] 59.5keV spectrum cleanup
?
I did 2 runs on a brannerite specimen.? One like normal
(red) and the other with a paper adhesive label covering the buttons.? The
proportions of the peaks pretty much look the same to me.
?
Charles
?
On Wed, Oct 28, 2020 at 11:30 AM Dude <dfemer@...> wrote:
Charles,
My mistake, I now see the
problem you¡¯re using a log energy axis. That should always be linear, counts
can be log, lin or sqr as appropriate.
Looking at the png you
sent there is something seriously wrong on the low energy below 4 keV. Send the
mca file for a look as it may just be a Theremino problem. For these low
energies you¡¯ll want to keep the Np in there, but to pull these elements out
you really need a high flux rate which the Am source isn¡¯t doing. Even with a
tube the efficiencies are very poor and the counts are low making it hard to
quantify. However, the Al Si, P, S, Ca and K ?are absolutely essential to
determining the basic mineralogy types and every effort should be made to
verify their presence.
To get the count times up
readjust the gain to look only at the low energy range say from 10 on down and
see ff that helps pull them out. Using a tight focused beam and close up to the
target will also help concentrate the flux.
Dud
?
From:[email protected] [mailto:[email protected]] On Behalf Of Charles
David Young Sent: Wednesday, October 28, 2020 8:12 AM To: XRF Subject: Re: FW: [XRF] 59.5keV spectrum cleanup
?
Geo,
?
I
agree and I did not mean to dis this exercise, even if it turns out to be
primarily academic.
?
When
I get a chance I'll try the paper filter with a mineral like Brannerite
(UTi2O6) to see how it affects the Ti peak at 4.5keV.? That is typically
the lowest element that is essential to my application.? Other peaks like
P, K, Ca would be nice but I have learned to live without them.? If we could
somehow dig those out by an adjustment it would be great.
"The
filter comes at a cost.? Am I reading it correctly that it reduced the
input count to %25 of the non-filtered?? Do we need to wait 4x longer for
equal counts when this is used as an exciter? "
?
All
that is to be determined Charles. It well may be much worse than that, if the
low energies in the beam are what's helping to dig out the low energy XRF
signals we are looking for.
?
I
agree about our perspective points of view too.?
?
You
system is as ideal as can be for your main task, and your display program is an
advancement in the area of display and identification, especially for the
audience of that group. Now the photography meets the same standard too. No one
has added all those dimensions in one place to the amateur mineral group
before.? All for science' sake, and I applaud that.
?
For
you and others I would recommend trying a paper filter over the exciter source,
just to see.
?
The
rest is for the X-Ray Physics and X-Ray Optics crowd. We now have an
inexpensive but powerful tool in our toolbox that wasn't there 2 days ago.
?
Geo
?
?
?
?
?
-----
Original Message -----
From: Charles David Young <charlesdavidyoung@...>
To: XRF <[email protected]>
Sent: Wed, 28 Oct 2020 08:33:01 -0400 (EDT)
Subject: Re: FW: [XRF] 59.5keV spectrum cleanup
?
The
filter comes at a cost.? Am I reading it correctly that it reduced the
input count to %25 of the non-filtered?? Do we need to wait 4x longer for
equal counts when this is used as an exciter?? Was all that low energy
activity that is filtered out detrimental or beneficial when the button is used
to stimulate XRF in the target?
?
I
come at this from a different perspective.? I use the Si-PIN with an Am241
exciter to identify minerals.? In my experience the XRF coming from the
target is much stronger than the noise introduced by the Am241, which is
mounted on a shield and facing away from the detector.? Attached is a scan
of xenotime, which shows a large Y component as well as some U and Pb.? I
turned on the Am241 reference lines to show that the Np and Au peaks are very
small in comparison.? One quickly learns to recognize and ignore them when
identifying the important peaks that emanate from the target: Ti, Mn, Fe, Ta,
Th, U, Y, Zr, Nb, REEs.
After
trying many slight thickness variations of Source Filter components, the best
result is show in the second picture.
First up is the spectra coming out of the Am exciter source with just a paper
alpha blocking filter. That alone made significant improvements at the low end
vs no alpha filter. The goal of this project is to apply a filter to the source
that will effectively remove most of the low energy component that is caused by
Np- daughter and self-XRF of elements in the button, such as Au, Ag, Fe.
Error! Filename not specified.
Now a filter applied to the source and another Gamma Spectrum ran.
Error! Filename not specified.
Thanks to member Dudley Emer for the design of the filter elements and
thicknesses, based on selective filtering using the K-edge X-Ray absorption
effect.
From: "DFEMER" <dfemer@...> To: [email protected] Sent: Friday, October 30, 2020 11:40:36 AM Subject: Re: FW: [XRF] 59.5keV spectrum cleanup
Geo,
The
4.8 usec doesn¡¯t have the filter fast ?thresholds set correctly as a result it
has you seeing 3135 valid counts with a noise count of 128,980.While The 19.2 PT
has 4534 valid counts with a noise count of 56627 counts and both with
comparable count times. (Note that the 19.2 still has a lot of low energy noise
that needs a better threshold setting but its way better than the 4.8 PT data
set ). So the 19.2 has better count statistics 4534 vs 3135 counts for roughly
the same count time counts and is getting better peak shape. You¡¯ll note that the
19.2 peaking time has a much better 13.6 keV FWHM resolution than the 4.8 (0.192
vs .396 FWHM).
Thanks
for that report Charles. Of the two, can you say what is the lowest peak that
can be resolved and is there a difference between them?? My just finished
overnight run using the Si-PIN on a weakly radioactive sample for Gamma Spec
(not excited XRF) is resolving individual peaks down to less than 1.5 keV. In
the Peaking Time topic thread I'll post that test @19.2us PT along with a
nearly similar test done last February at 4.8us. The main difference between
those seems to be both in the clarity and narrowness of the big peaks, and
better performance at the lowest end going to the 19.2.
?
Today
I'll try even longer peaking times, up to 32us just for the historical
record.?
I
think you've been using 32us and I can't fault your scans, that's for sure.
?
Geo
?
-----
Original Message -----
From: Charles David Young <charlesdavidyoung@...>
To: XRF <[email protected]>
Sent: Fri, 30 Oct 2020 08:26:25 -0400 (EDT)
Subject: Re: FW: [XRF] 59.5keV spectrum cleanup
?
I
did 2 runs on a brannerite specimen.? One like normal (red) and the other
with a paper adhesive label covering the buttons.? The proportions of the
peaks pretty much look the same to me.
?
Charles
?
On
Wed, Oct 28, 2020 at 11:30 AM Dude <dfemer@...> wrote:
Charles,
My
mistake, I now see the problem you¡¯re using a log energy axis. That should
always
be linear, counts can be log, lin or sqr as appropriate.
Looking
at the png you sent there is something seriously wrong on the low energy below
4 keV. Send the mca file for a look as it may just be a Theremino problem. For
these low energies you¡¯ll want to keep the Np in there, but to pull these
elements out you really need a high flux rate which the Am source isn¡¯t doing.
Even with a tube the efficiencies are very poor and the counts are low making
it hard to quantify. However, the Al Si, P, S, Ca and K ?are absolutely
essential to determining the basic mineralogy types and every effort should be
made to verify their presence.
To
get the count times up readjust the gain to look only at the low energy range
say from 10 on down and see ff that helps pull them out. Using a tight focused
beam and close up to the target will also help concentrate the flux.
Dud
?
From:[email protected]
[mailto:[email protected]] On
Behalf Of Charles David Young Sent: Wednesday, October 28, 2020 8:12 AM To: XRF Subject: Re: FW: [XRF] 59.5keV spectrum cleanup
?
Geo,
?
I agree and I did not mean to dis this exercise, even if it
turns out to be primarily academic.
?
When I get a chance I'll try the paper filter with a
mineral
like Brannerite (UTi2O6) to see how it affects the Ti peak at 4.5keV.?
That is typically the lowest element that is essential to my application.?
Other peaks like P, K, Ca would be nice but I have learned to live without
them.? If we could somehow dig those out by an adjustment it would be
great.
"The filter comes at a cost.? Am I reading it
correctly that it reduced the input count to %25 of the non-filtered?? Do
we need to wait 4x longer for equal counts when this is used as an exciter?
"
?
All that is to be determined Charles. It well may be much
worse than that, if the low energies in the beam are what's helping to dig out
the low energy XRF signals we are looking for.
?
I agree about our perspective points of view too.?
?
You system is as ideal as can be for your main task, and
your display program is an advancement in the area of display and
identification, especially for the audience of that group. Now the photography
meets the same standard too. No one has added all those dimensions in one place
to the amateur mineral group before.? All for science' sake, and I applaud
that.
?
For you and others I would recommend trying a paper filter
over the exciter source, just to see.
?
The rest is for the X-Ray Physics and X-Ray Optics crowd.
We
now have an inexpensive but powerful tool in our toolbox that wasn't there 2
days ago.
The filter comes at a cost.? Am I reading it correctly
that it reduced the input count to %25 of the non-filtered?? Do we need to
wait 4x longer for equal counts when this is used as an exciter?? Was all
that low energy activity that is filtered out detrimental or beneficial when
the button is used to stimulate XRF in the target?
?
I come at this from a different perspective.? I use
the
Si-PIN with an Am241 exciter to identify minerals.? In my experience the
XRF coming from the target is much stronger than the noise introduced by the
Am241, which is mounted on a shield and facing away from the detector.?
Attached is a scan of xenotime, which shows a large Y component as well as some
U and Pb.? I turned on the Am241 reference lines to show that the Np and
Au peaks are very small in comparison.? One quickly learns to recognize
and ignore them when identifying the important peaks that emanate from the
target: Ti, Mn, Fe, Ta, Th, U, Y, Zr, Nb, REEs.
After trying many slight thickness variations of Source
Filter components, the best result is show in the second picture.
First up is the spectra coming out of the Am exciter source with just a paper
alpha blocking filter. That alone made significant improvements at the low end
vs no alpha filter. The goal of this project is to apply a filter to the source
that will effectively remove most of the low energy component that is caused by
Np- daughter and self-XRF of elements in the button, such as Au, Ag, Fe.
Error! Filename not specified.
Now a filter applied to the source and another Gamma Spectrum ran.
Error! Filename not specified.
Thanks to member Dudley Emer for the design of the filter elements and
thicknesses,
based on selective filtering using the K-edge X-Ray absorption effect.
The
4.8 usec doesn¡¯t have the filter fast ?thresholds set correctly as a result it
has you seeing 3135 valid counts with a noise count of 128,980.While The 19.2 PT
has 4534 valid counts with a noise count of 56627 counts and both with
comparable count times. (Note that the 19.2 still has a lot of low energy noise
that needs a better threshold setting but its way better than the 4.8 PT data
set ). So the 19.2 has better count statistics 4534 vs 3135 counts for roughly
the same count time counts and is getting better peak shape. You¡¯ll note that the
19.2 peaking time has a much better 13.6 keV FWHM resolution than the 4.8 (0.192
vs .396 FWHM).
Thanks
for that report Charles. Of the two, can you say what is the lowest peak that
can be resolved and is there a difference between them?? My just finished
overnight run using the Si-PIN on a weakly radioactive sample for Gamma Spec
(not excited XRF) is resolving individual peaks down to less than 1.5 keV. In
the Peaking Time topic thread I'll post that test @19.2us PT along with a
nearly similar test done last February at 4.8us. The main difference between
those seems to be both in the clarity and narrowness of the big peaks, and
better performance at the lowest end going to the 19.2.
?
Today
I'll try even longer peaking times, up to 32us just for the historical
record.?
I
think you've been using 32us and I can't fault your scans, that's for sure.
?
Geo
?
-----
Original Message -----
From: Charles David Young <charlesdavidyoung@...>
To: XRF <[email protected]>
Sent: Fri, 30 Oct 2020 08:26:25 -0400 (EDT)
Subject: Re: FW: [XRF] 59.5keV spectrum cleanup
?
I
did 2 runs on a brannerite specimen.? One like normal (red) and the other
with a paper adhesive label covering the buttons.? The proportions of the
peaks pretty much look the same to me.
?
Charles
?
On
Wed, Oct 28, 2020 at 11:30 AM Dude <dfemer@...> wrote:
Charles,
My
mistake, I now see the problem you¡¯re using a log energy axis. That should
always
be linear, counts can be log, lin or sqr as appropriate.
Looking
at the png you sent there is something seriously wrong on the low energy below
4 keV. Send the mca file for a look as it may just be a Theremino problem. For
these low energies you¡¯ll want to keep the Np in there, but to pull these
elements out you really need a high flux rate which the Am source isn¡¯t doing.
Even with a tube the efficiencies are very poor and the counts are low making
it hard to quantify. However, the Al Si, P, S, Ca and K ?are absolutely
essential to determining the basic mineralogy types and every effort should be
made to verify their presence.
To
get the count times up readjust the gain to look only at the low energy range
say from 10 on down and see ff that helps pull them out. Using a tight focused
beam and close up to the target will also help concentrate the flux.
Dud
?
From:[email protected]
[mailto:[email protected]] On
Behalf Of Charles David Young Sent: Wednesday, October 28, 2020 8:12 AM To: XRF Subject: Re: FW: [XRF] 59.5keV spectrum cleanup
?
Geo,
?
I agree and I did not mean to dis this exercise, even if it
turns out to be primarily academic.
?
When I get a chance I'll try the paper filter with a
mineral
like Brannerite (UTi2O6) to see how it affects the Ti peak at 4.5keV.?
That is typically the lowest element that is essential to my application.?
Other peaks like P, K, Ca would be nice but I have learned to live without
them.? If we could somehow dig those out by an adjustment it would be
great.
"The filter comes at a cost.? Am I reading it
correctly that it reduced the input count to %25 of the non-filtered?? Do
we need to wait 4x longer for equal counts when this is used as an exciter?
"
?
All that is to be determined Charles. It well may be much
worse than that, if the low energies in the beam are what's helping to dig out
the low energy XRF signals we are looking for.
?
I agree about our perspective points of view too.?
?
You system is as ideal as can be for your main task, and
your display program is an advancement in the area of display and
identification, especially for the audience of that group. Now the photography
meets the same standard too. No one has added all those dimensions in one place
to the amateur mineral group before.? All for science' sake, and I applaud
that.
?
For you and others I would recommend trying a paper filter
over the exciter source, just to see.
?
The rest is for the X-Ray Physics and X-Ray Optics crowd.
We
now have an inexpensive but powerful tool in our toolbox that wasn't there 2
days ago.
The filter comes at a cost.? Am I reading it correctly
that it reduced the input count to %25 of the non-filtered?? Do we need to
wait 4x longer for equal counts when this is used as an exciter?? Was all
that low energy activity that is filtered out detrimental or beneficial when
the button is used to stimulate XRF in the target?
?
I come at this from a different perspective.? I use
the
Si-PIN with an Am241 exciter to identify minerals.? In my experience the
XRF coming from the target is much stronger than the noise introduced by the
Am241, which is mounted on a shield and facing away from the detector.?
Attached is a scan of xenotime, which shows a large Y component as well as some
U and Pb.? I turned on the Am241 reference lines to show that the Np and
Au peaks are very small in comparison.? One quickly learns to recognize
and ignore them when identifying the important peaks that emanate from the
target: Ti, Mn, Fe, Ta, Th, U, Y, Zr, Nb, REEs.
After trying many slight thickness variations of Source
Filter components, the best result is show in the second picture.
First up is the spectra coming out of the Am exciter source with just a paper
alpha blocking filter. That alone made significant improvements at the low end
vs no alpha filter. The goal of this project is to apply a filter to the source
that will effectively remove most of the low energy component that is caused by
Np- daughter and self-XRF of elements in the button, such as Au, Ag, Fe.
Error! Filename not specified.
Now a filter applied to the source and another Gamma Spectrum ran.
Error! Filename not specified.
Thanks to member Dudley Emer for the design of the filter elements and
thicknesses,
based on selective filtering using the K-edge X-Ray absorption effect.
From what I know about XRD it requires careful preparation of the sample.? I know grinding is involved.? Interpreting the results is also not as straightforward as XRF.
For now, the XRF is working exceptionally well.? My hobby consists of a narrow niche of mineral collecting.? Between the XRF, the appearance of the mineral, and the minerals reported from the area I can determine with some confidence the id of a specimen.
I agree, and we can't expect anything below ~2keV in air, but what this shows me is with a helium atmosphere, we could have a shot at down to 1.
?I tried propane (once) and it changed things, bit of course is not practical, just as hydrogen is not practical. Vacuum is best, but at this time it's not in the cards for me.
Charles have you considered XRD for your minerals? If so check out the Tel-X-Ometer unit. I have several of those waiting a turn in the counting room. The way XRD works baffles me but it seems to be a very popular mode for determining crystal structure and maybe oxidation states (?). No immediate plans to do so but have used the internal copper-target X-Ray source in it many times for other purposes. Switch selectable 20 / 30 kV at 80uA. These have an integral goniometer, and the ones I get usually have that part damaged by corrosion due to leaving salt crystals in it when stored (they are a college teaching tool).
Geo
----- Original Message ----- From: Charles David Young <charlesdavidyoung@...> To: XRF <[email protected]> Sent: Fri, 30 Oct 2020 10:53:09 -0400 (EDT) Subject: Re: FW: [XRF] 59.5keV spectrum cleanup
It just looks like noise to me.? I would have to see a lot more counts and reproducible runs to believe it.
On Friday, October 30, 2020, 11:05 PM, GEOelectronics@... wrote:
Great idea on the reverse-osmosis water discharge!
Test everything!
Here's another great idea for collecting samples from water, a paper by?
"A Sensitive XRF Screening Method for Lead in Drinking Water"
Abstract
A novel method for quickly and quantitatively measuring aqueous lead in drinking water has been developed. A commercially available activated carbon felt has been found to effectively capture lead from tap water, and partnered with X-Ray fluorescence (XRF) spectrometry, it provides quantitative measurement of aqueous lead in drinking water. Specifically, for a 2-liter volume of tap water, the linear range of detection was found to be from 1 ¨C 150 ppb, encompassing the current EPA limit for lead in drinking water (15 ppb). To make a reproducible and easy to use method for filtering, a 2L bottle cap with a 1.25 cm diameter hole was used for filtering. Utilizing this filtration method, 75 solutions from 0 ppb to 150 ppb lead gave a 91% sensitivity, 97% specificity, and 93% accuracy and all the misclassified samples fell between 10 and 15 ppb. This method has also proved reliable for detecting calcium as well as several other divalent metals in drinking water including copper, zinc, iron, and manganese.
Note- She has tested all sorts of premade activated charcoal felt sheeting to find one that is inexpensive and yet performs very well for those heavy metals. I can send the paper by email but can't post it online.
Geo
----- Original Message ----- From: taray singh via groups.io <sukhjez@...> To: [email protected] Sent: Fri, 30 Oct 2020 10:14:27 -0400 (EDT) Subject: [XRF] Chinese water filter analysis
Hi guys I am interested in drinking water quality testing and most drinking water test kits are not accurate. They give positive readings on concentrated samples like sewage and not for trace elements in drinking water. So I wanted to try out this interesting Chinese? gadget. Search you tube for..?Water Electrolyzer (Quality Tester). The Most Dangerous Appliance Ever? Negative comments being dangerous? and a scam. I checked it out There are 2 pairs of electrodes A shiny aluminium and a dull? iron pair. Repeated usage caused the? iron to rust. Aluminium was confirmed? by exclusion being shiny,light,cheap and nondescript. My drinking water? sample is RO? waste water . This discharged RO waste water is about 4 times more concentrated than my tap water depending on machine cycling. Concentrated means more residue better than plain tap water for analysis. My 14 in 1test kit was basically negative? for heavy metals. I tested this RO concentrated with this Chinese toy. It produced a brownish residue as expected. Collected it and allowed it dry on cap and a tissue. This cap/tissue was analyzed with Xrf with Am241/Si pin detector. True enough like the you tubers described there is Fe from galvanic reactions. There appears to be Pb peak as well I need to repeat this test with a thicker sample? to confirm . Taray
