Crystal Alignment or Intensity Issues

[Probeman]

I need some help.

We have Edgar Chavez here to re-align some of our analyzing crystals since our CAMCOR engineer replaced the crystal flip motors on some of them.  During the process of attempting to align the TAP crystals, Edgar noticed that the intensities were quite low.  At first he thought we might be using detector windows that were too thick, but I don't think that can be right because even though it is a TAP crystal, we are aligning it using the Fe Ka *8th order* reflection so the x-ray energy is still that of Fe Ka.  So the window thickness should not be an issue...  Just FYI the Moxtek part# is LWN00014/LW100402004 and Edgar says it looks a "little bit thick" but OK.

Anyway, we are wondering why even with 4 micro-amps(!!!) of beam current we are getting such a small Fe Ka (8th order) peak on our TAP crystals. 

[Mike Jercinovic]

We did lose about 25 to 30% of counts on UMb when we switched to the Moxtek detector window compared to Cameca's original polypropylene detector window.  Fe Ka is, of course, much more energetic...

[gmorgan@ou.edu]

We also just had Edgar replace the Moxtek windows on all three of our low-pressure spectros with stretched polyporylene due to poor light element intensities. The support grid Moxtek uses really absorbs the light element x-rays. Something like at least 20-25% loss at F Ka (I couldn't really peak or do a PHA adjustment on F Ka using F-topaz at 200 nA current...), and on the order of 75-80% loss at B Ka.

[Probeman]

Hi Mike and George,
I get what you guys are saying, but these are very high energy lines and detectors with Mylar windows, so the detector window is a red herring in this case I think.

In fact, we still don't know why the intensity dropped so much after the alignment, as described here:

http://probesoftware.com/smf/index.php?topic=567.msg3362#msg3362

[Mike Jercinovic]

John,
Your original post only mentioned the TAP crystals, so this seemed to point to detector windows (probably not column windows).  Now I see that this effect is across all spectrometers, all the Ti counts are down by almost exactly the same amount across the board (including with different crystals, different counter tube pressures, and different column and detector windows), but not Mn.  Can't explain that with detector gas as intensities would vary with tube pressure, and, as you say, can't be windows.  Seems like this has to be something really simple that effects x-ray counts to all spectrometers...the Ti itself?  I just don't see a scenario that makes any sense at the moment.

[Probeman]

John,
Your original post only mentioned the TAP crystals, so this seemed to point to detector windows (probably not column windows).  Now I see that this effect is across all spectrometers, all the Ti counts are down by almost exactly the same amount across the board (including with different crystals, different counter tube pressures, and different column and detector windows), but not Mn.  Can't explain that with detector gas as intensities would vary with tube pressure, and, as you say, can't be windows.  Seems like this has to be something really simple that effects x-ray counts to all spectrometers...the Ti itself?  I just don't see a scenario that makes any sense at the moment.

Hi Mike,
Yup, TAP crystals were mentioned but you will note that I said we were using the Fe Ka *8th order* reflection so the x-ray energy is still that of Fe Ka...  so detector window thickness is not a concern, obviously, as you said so yourself.

Dang, I sure hope we can figure this out.  Do you have any similar measurements on your own instrument (just LIF, LLIF, PET and LPET, not your super awesome VLPET crystals!), for Mn and Ti Ka that we can compare?

I would like to know, what are "normal" intensities for Mn and Ti Ka at 15 keV, say 30 nA... on LIF and PET... and/or LLIF and LPET...
john

[Probeman]

We have Edgar Chavez here to re-align some of our analyzing crystals since our CAMCOR engineer replaced the crystal flip motors on some of them.  During the process of attempting to align the TAP crystals, Edgar noticed that the intensities were quite low.  At first he thought we might be using detector windows that were too thick, but I don't think that can be right because even though it is a TAP crystal, we are aligning it using the Fe Ka *8th order* reflection so the x-ray energy is still that of Fe Ka.  So the window thickness should not be an issue...  Just FYI the Moxtek part# is LWN00014/LW100402004 and Edgar says it looks a "little bit thick" but OK.

Anyway, we are wondering why even with 4 micro-amps(!!!) of beam current we are getting such a small Fe Ka (8th order) peak on our TAP crystals.

Update: we still don't understand why the TAP crystals seemed to have such a low intensity, at least according to Edgar Chavez our Cameca engineer who commented on this, soon after he arrived on site. He now suspects a problem with the Moxtek windows, but I note that Cameca now provides Moxtek windows, so not sure what the differences might be?  Also, Fe Ka is pretty energetic so also not sure how a detector window designed for low energy work could be too thick for Fe Ka?

But as far as the general crystal alignment procedure goes, I checked today on a number spectrometers looking at various standard intensities before and after the alignment, and honestly, there doesn't seem to be much change in the intensities.  The biggest change is that the difference between the measured peak offsets and the theoretical spectrometer positions, for our most problematic spectrometers, seem to be closer to zero, as would be expected from an alignment.

So as to why I now see 3x less intensity on Ti Ka (but not Mn Ka) on the LiF and PET crystals is a bit of a mystery. I can only assume that I mis-measured the Ti Ka intensities somehow. So except for the TAP Fe Ka issue, I think our spectrometer alignment procedure with Edgar went quite well.
john

[edgrchvz]

OK, Hello. I would like to add a short comment about the “loss” of intensities at  Donovan crystals.
 In French and Spanish we say: We are always wrong if we are absent. For this reason I want to add this comment.
I compared the intensities of the crystals after my intervention. I took the actual intensities and those from the time the instrument was installed. They are more or  less the same. Only some decrease on those for the low pressure spectrometers This is very probably due to presence of Mostek windows. These windows are a  thicker and have a grid. That grid is a bit wide on the window surface. The grid will affect the intensity for all elements. The window thickness will affect the light elements like F, Na, etc.
I found  as well a additional problem. The values on the gain and bias were wrong. Once we adjusted it to the Cameca PHA way, all became better.

This is the results counts for Ti ka :
At Cameca 06/03/2005   At U of Oregon 10/13/15   Decrease or increase
Sp1= 20585                18912                                 -8%
Sp2=58019                   56661                                  -2%
SP3=64623                   66552                                 +3%
Sp4=19468                   17568                                  -9.6%
SP5=25217                   27744                                  +10%

I'm glad to clarify my intervention didn't produce any kind of loss on crystals intensities or any kind of misalignment.
      
Edgar Chavez
      

[Probeman]

This is the results counts for Ti ka :
At Cameca 06/03/2005   At U of Oregon 10/13/15   Decrease or increase
Sp1= 20585                18912                                 -8%
Sp2=58019                   56661                                  -2%
SP3=64623                   66552                                 +3%
Sp4=19468                   17568                                  -9.6%
SP5=25217                   27744                                  +10%

Hi Edgar,
As we say in English- don't worry so much!   No one thought you caused any problem!  We're just trying to understand the instrument's performance over time.

You say "all became better" but the numbers above in your post still record an average loss in intensity over time.  Not surprising for an 8 year old instrument I should add!  But we'd still like to understand why the SX50 produces slightly *better* intensities on a 28 year old instrument...

That is, if there is still a problem causing a decrease in x-ray intensities on our Sx100, we would like to get to the bottom of it with your advice and assistance if possible.

Thanks.
john

[edgrchvz]

I wouldn’t say the SX50 keep better performances. We have to know and compare the crystals performances for the SX50 when it was installed. Possible, the SX50 came with very good crystals. Normally, in good conditions they should remain always good.  The physics on both instruments is very similar.
On the SX100, the loss on intensities is only on the low pressure spectrometers (1, 2 and 4). You see the high pressure (3 and 5) are with the same intensity or better.   The Mostek windows are for something.
I assure you, a good periodic maintenance will help a lot. Probably, many of the users prefer to use the SX100 over the SX50.

[Probeman]

I wouldn’t say the SX50 keep better performances. We have to know and compare the crystals performances for the SX50 when it was installed. Possible, the SX50 came with very good crystals. Normally, in good conditions they should remain always good.  The physics on both instruments is very similar.
On the SX100, the loss on intensities is only on the low pressure spectrometers (1, 2 and 4). You see the high pressure (3 and 5) are with the same intensity or better.   The Mostek windows are for something.
I assure you, a good periodic maintenance will help a lot. Probably, many of the users prefer to use the SX100 over the SX50.

Hi Edgar,
I guess it is possible the SX50 started off life with better crystals than the SX100...  but time (and thermal cycling) does take a toll on 28 year old crystals, I also assume...

When you say "I assure you, a good periodic maintenance will help a lot", exactly what maintenance do you think will help the crystal intensities?  At one point you mentioned a column maintenance, and although I can understand that a column maintenance would help the beam alignment, given the same faraday beam current, how could a cleaner column improve the crystal intensities?
john

[Probeman]

Here is data from the Drift.exe application that demonstrates that the crystal alignment performed by Edgar Chavez (Cameca) generally improved the SX100 spectrometer intensities (see attached images below).  Basically spectrometers 1 and 2 showed the most improvement. Which is interesting because spectrometer 3 was the one that was most out of alignment based on the size of the offset from the theoretical peak positions.

The Drift application is a useful app for monitoring your spectrometer intensities over time.  See the Drift menu in the Start button, Probe Software | Utilities menu. Basically one simply selects a folder tree to search for just an element and x-ray, but one can also constrain it to specific spectrometers and/or standards which is what I did below. Here is a screen shot of the search for Y La intensities on my YPO4 std (1016):



And when it asks "Do you want to export the data to Excel?" click Yes and here is what you do to plot it in Excel:



What we'd like to figure out now is why Edgar said after he first arrived, that the TAP and LTAP crystals showed such low intensities for Fe ka (high order) reflections...  is there something we can do to find out what is causing these low intensities?

I'm going to try some "in the hole" tests next using a W or Mo 100 um aperture.

[BenjaminWade]

Hi all
Ressurecting an old post...
I have finally had some time to get around to looking at some of my spectrometer intensities over time on various standards/xtals using the Drift program. I am after an opinion on some of the trends I am seeing, and whether or not they are indicative of the need for some xtal alignments, or if there is some other effect/problem.

Looking at the first attachment, it is my Ti Ka intensity on LPET(sp3) on one of the rutile standards collected over the course of ~3 years. Over that period the intensity (blue line) has dropped, perhaps more drastically in a couple of bigger steps, but no denying is lower now than what it was three years ago. The offset of peak from theoretical (orange line) was never bang on, but you could argue is also definitely further away from theoretical than a few years ago.

The second attachment is my Si Ka intensity on TAP(sp4). In this plot it is more a steady decline in peak intensity over a few years. The peak offset from theoretical, although noisy, is hovering around zero, but this is a function of Si being the element Cameca users "verify" on for their TAP xtals, forcing it to its correct sintheta value.

I know there are probably many variables (how well I peaked, standard degradation? etc), but on the whole perhaps they are in need of some alignment? How often are people getting their xtals aligned?

While on this, perhaps I am understanding the Cameca verification routine wrong, and/or not completely understanding what it does. Take the LLIF xtal where we verify on Fe Ka forcing it to the theoretical position, then applying a linear offset/correction to the entire spectrometer. Given that verification is done often, shouldn't this be correcting for the "Peak offset" seen in the Ti Ka plot I have attached? Or is this the kind of error involved in that offset/correction and its not linear?

Not sure if any of that made sense..

Cheers

[Probeman]

This is an old topic but Philipp Poeml asked me to post my old lab notes for aligning the Bragg crystals on a Cameca instrument and this seems to be the most related existing topic.

The scanned pages (attached below, please login to see attachments) are from notes I took more than 25 years ago when my Cameca SX51 was installed at UC Berkeley. Of course some of the computer specific procedures will be different, but the notes on the mechanical adjustments should still be relevant.

[Probeman]

This is an old topic but Philipp Poeml asked me to post my old lab notes for aligning the Bragg crystals on a Cameca instrument and this seems to be the most related existing topic.

The scanned pages (attached below, please login to see attachments) are from notes I took more than 25 years ago when my Cameca SX51 was installed at UC Berkeley. Of course some of the computer specific procedures will be different, but the notes on the mechanical adjustments should still be relevant.

In case anyone is wondering, ADP is a Bragg crystal composition no longer seen very often:

Ammonium Dihydrogen Phosphate, known as ADP crystal (NH4H2PO4)

https://www.unitedcrystals.com/KDPProp.html

It was sometimes utilized to get a high spectral resolution for Al Ka (or Br La):