Charles,
These spectra are wrong as well. It appears you have a
bad energy cal that you were using.
?
This is the Ag cal on 2/24/202
START_TIME -
02/24/2020 04:58:50
SERIAL_NUMBER - 0
<<CALIBRATION>>
LABEL - Channel
312 3.605
1275 14.96
I¡¯d surely like to know how you got that cal Sb and Y?
?
Here¡¯s the Ce cal Its using Ag from 3/2/2020
<<CALIBRATION>>
LABEL
- Channel
293
3.151
2031
22.163
?
Here¡¯s the Nb cal its using Ag as well
<<CALIBRATION>>
LABEL
- Channel
293
3.151
2031
22.163
?
Using the Ag cal, Ka is at ch 1881 with a 22.1 keV which
is correct. Then looking at ?channel 2953 (the Ce peak) it shows 34.75 keV
which is also correct
However using the Nb file the Ka peak at channel 2953 is
32.25? keV
The Ce file Ka at channel 2953 also shows ?a 32.5
keV which is the Ka of Ba . Ce should be at 34.7 keV ?
So while the Ag cal file is correct the Ce and Nb files
are energy shifted implying a gain change or a different cal file was used.
Looking at the 2 point data that¡¯s pretty clear.
Looking at the Ag file at channel 2953 gives 34.75 keV
Ch 2953 for the Nb file is? 32.25 - Nb and Ce use
the same cal. The Ag file is different
Ch 2953 for Ce is 32.25? - Nb and Ce use the same
cal. The Ag file is different
The energy offset is 32.25? ¨C 34.75 = -2.5 keV error
Your Ce should have come in at 34.77 keV but the cal is
-2.5 keV to low due to using an old Ag erroneous cal which ID¡¯s it as Ba at
32.25 keV.? The same effect occurs with the Ce and Nb La¡¯s producing an
offset making Ce look and ID like Ba.
?
This is a classic example of what you are doing with Theremino is
wrong. Look at the attached Ag Cu Ce ?.png file. Ag and Ce are lined up
correctly but nothing could be further from the truth.? Look at the raw
.mca ¨C files nothing lines up correctly. The problem is you have taken
?erroneous data without even examining it, quickly moved it over to
theremino and manually fudged it with ¡°sliders¡± to fit a preconceived notion
making it look ¡°pretty¡±. ?
The fault here ?lies in your comment - ¡°These were
scanned on different days.? It takes hours to create each scan so there is
no point in trying to recalibrate each day.? If this Si-PIN has any drift
I have not been able to detect it¡±.
I understand the issue of a really weak source and the
time it takes, but skipping absolutely essential steps in spectroscopy is just
not done and this is a perfect example of why.? If you had looked at and
interpreted the raw data spectra in DppMCA it might have been apparent ( it
wasn¡¯t for me I thought it was Ba and it fit Ba , even DppMCA thought it was
Ba.) but looking at the known Np peaks should have said something was wrong as
they were shifted too. Yeah, I didn¡¯t do that either.
The
essential steps in spectroscopy that always need to be done are pretty straight
forward.
1.????? Determine what energy range you
want to look at and set up the gain to do that. In your case your source is Am,
so a good energy range would include 59.5 keV. For the low end you can get an
source that covers the low end like Fe55 and use the Am Np peaks. Use a 3pt or
better cal as anything fits 2 pts and if one is off you¡¯ll never know it.
2.????? For THAT energy range and Gain run
a hot source calibrate. Check other peaks that you didn¡¯t use to make sure it¡¯s
a good cal fit. Save that cal and use it for that energy/gain range the next
time. But ¨C see #3
3.????? At the beginning of the day or
before the first spectra, do a Cal Check. Use the hot cal source and verify
that all channels and their energy line up and are correct. This doesn¡¯t take
any time as it¡¯s a hot cal source you are using. At the end of the days run do
another Cal Check ¨C QA/QC you know. If it doesn¡¯t match throw out the data and
start over. That¡¯s pretty rare though as these things shouldn¡¯t drift.
4.????? Using DppMCA start ID¡¯ing the big
peaks working down using your eyeballs and an energy table looking for close
interferences. Put a manual ROI on it and verify Ka,b- La,b and when you have a
positive ID leave the ROI on it. A positive ID needs both a and b lines. A not
determined has only one line so don¡¯t bet on that horse. Do not use computer
picks as that is only there to mislead and embarrass you.
?If you must use the Theremino pretty picture thing then Cal
it correctly. The very fact that it doesn¡¯t have a linear or quadratic
regression in it is a real clue it¡¯s not ready for prime time.? Being that
you¡¯re an accomplished
software geek with this program use the linear or quadratic equation from
DppMCA under the Calibrate tab ?(Energy =A+Bx ?or for a quadratic E= A+B*x + C*x^2
where x is channel number) to calculate the channel vs energy table then import
that directly into the Thermino and avoid the problematic slider thing.?
Or ?program it in to Theremino and have it calc it.
The slider thing is probably needed for the sound cards non
linear response or missing codes. It¡¯s, at best, an approximation that works
for NaI but shouldn¡¯t be used ?for high res detectors.
?
Dud
?
?
?
?
Dud,
here are the .mca files along with another one that I did of Nb last
night.? I would be interested in seeing what calibration points you can
come up with.? These were scanned on different days.? It takes hours
to create each scan so there is no point in trying to recalibrate each
day.? If this Si-PIN has any drift I have not been able to detect it.
?
"Dud, I made a special decluttered plot
just for you.? And in the process I discovered that the SmKa1 peak that I
had previously identified is actually CeKb2 so this Ce metal is actually a
little purer than I thought!
?
Charles"
?
Be sure to pay attention to L-gamma and M- shell lines
too.?
?
Geo
?
?
Ok,
so the deal is that using AgKa1 as one of the points on the E cal line is not
very accurate by the time it gets up to CeKa1 so the peak finder in DppMCA
assumed
that peak was BaKa1.? Use these cal points instead in DppMCA:
You
may ask why I never noticed this discrepancy?? The reason is that I use
DppMCA strictly for collecting data.? The E cal in DppMCA is not
critical.? All analysis is done in my highly customized Theremino where I
can label things nicely.? The table format of
DppMCA just doesn't cut it for me.
Dud,
I made a special decluttered plot just for you.? And in the process I
discovered that the SmKa1 peak that I had previously identified is actually
CeKb2 so this Ce metal is actually a little purer than I thought!
On
Wed, Mar 4, 2020 at 2:44 PM Dude <dfemer@...> wrote:
Charles,
Check
your ¡°slider¡± calibration. This is Ba not Ce. Go back to the
.mca file and look at it with the original cal using the DppMCA program. It
looks like you used Ag for a cal but when was that cal done? Before you ran the
¡°Ce¡± or was it an old one?
Please
clean up the theremino plot, you have so many lines and
labels on their it¡¯s hard to tell what is what. Get rid of anything that¡¯s not
ID¡¯d or an interference anything else is clutter.? One shouldn¡¯t ?do
an interp
based on pre-established labels. Treat each peak separately starting at the low
energy and work your way up ruling out close interferences. Then put an ROI and
?ID label on it and use a real MCA program.? Thermino is great for
NaI but it¡¯ll
get you in trouble with these high res detectors.
There
is no Sm , La or Ce in here and it looks like you were at
83% dead time which is too high which can shift energy and FWHM.
Think
scientifically. The hypothesis was its Ce. You need to prove
that.? I¡¯ll guess you used your pre-established Ce labels and did the
¡°slider¡±
thing to line it all up to a preconceived notion. ??Wrong approach
and guaranteed
to get you in trouble- Don¡¯t use a secondary Cal - use the data acquisition
program to ID peaks.? Look at the periodic table ¨C Ba, La, and Ce line up
so
they are going to have the same spectral structure close energies and look
alike if the cal isn¡¯t right.? An interp starts by looking at each
individual peak
energy and any surrounding interferences with no pre-conceived notions, then
decide
based on Ka,b and La,b and move on to the next. If it doesn¡¯t take you an hour
or more you¡¯re not doing it right. Computer picks suck, stay a long way away ¨C
use your eyes, brains and an energy table.
Dud
?
?
This
is from Gunnar Faerber who did his own analysis.?
He is also the one that identified the Betafo speciman as brannerite.
Meanwhile,
this is an interesting scan of Ce metal.? It
is evidently hard to purify it of the other lanthanides because it shows plenty
of Sm and even detectable La.? The La1 and Lb1 peaks echo the Ka1 and Kb1
peaks nicely.? Taken together this set of peaks is quite useful for
calibration from low to high.
?
On
Tue, Mar 3, 2020 at 9:03 PM WILLIAM S Dubyk <sdubyk@...> wrote:
Charles,
who did you get your specimen of ishikawaite from,
and do you have a microprobe analysis for it? The reason I ask is that out of
curiosity I looked at some of the samarskite analyses from the Petaca district,
and several of the Queen mine specimens may be that mineral. I will ask Al
Falster, who is somewhat of an expert on the samarskite minerals, about what
analyses to use to determine this - oxides versus cation ratios. From my
understanding, this mineral has U+Th greater than Y+REE, and that is what I see
in the Queen mine specimens.
Steve
?
?
