Explain This If You Can!

[Probeman]

OK, I'm starting what I hope will be a fun topic where we can challenge each other to try and figure out what we are seeing in an image or graph.

Please post your "weird science" images or plots!

[Probeman]

Ok, I'll start with this recent image of mapping a glass melt inclusion in olivine about 120 um in size. Conditions were 15 keV, 30 nA and 2 sec (2000 ms) per pixel. 256 x 256 pixels.



It is quite reproducible and also seen to a much lesser degree in the other elements (e.g., K) as might be expected.

[Probeman]

OK, here's another weird image- but this one I can explain (see attachment for a good hint).

[qEd]

Polyphase saturation intergrowth following rind crystallation that occurred after melt entrappment.

[John Donovan]

Ed,
Which image does your "explanation" refer to? 

Meanwhile, George Morgan asks what was the maximum Na count rate on the x-ray map, so here is a map of the Na raw data intensities in cps (attached).

I also attach a BSE image of the same inclusion before the map acquisition. The circular feature in the inclusion is where the carbon coat was starting to lift off but it "subsided" before the map was acquired.

[Ben Buse]

rgd kileaua melt inclusion

Must be beam damage and recovery, or beam damage depleting one area and enriching an adjacent area (alhtough never seen it laterally). Is the map one really slow map or a stack of intergrated maps? Don't suppose you can do TDI on your maps can you, that's probably asking to much. I surpose the only way to make sense of it is to try different step sizes or speeds and see if the pattern changes.

[qEd]

I was referring to the electron images in the 3:01 post

[srmulcahy]

RE Kilauea Melt Inclusion

At first look I'd agree with Ben.  What's your spot size for this image?  My guess is that the green regions on the raw intensity map are the spot locations with Na-depletion and you're getting enrichment in the regions between the adjacent points.

[Probeman]

Polyphase saturation intergrowth following rind crystallation that occurred after melt entrappment.

Not that I understand anything about "polyphase saturation intergrowth..." but you'll notice that this is a secondary electron image (see the FEI databar at the bottom of the image where it says Sig SE). So that should be a good hint.

[Probeman]

RE Kilauea Melt Inclusion

At first look I'd agree with Ben.  What's your spot size for this image?  My guess is that the green regions on the raw intensity map are the spot locations with Na-depletion and you're getting enrichment in the regions between the adjacent points.

Ok, except that the Na x-ray map is a 256 x 256 pixel stage scan so the analysis positions are much, much smaller than the artifacts and the stage was moving perfectly smoothly!

You can see the individual pixels if you click on the raw data x-ray map:

http://probesoftware.com/smf/index.php?action=dlattach;topic=144.0;attach=296;image

[Probeman]

Don't suppose you can do TDI on your maps can you, that's probably asking to [sic] much...

Funny you should mention this, but I've got a very cool idea on how one *might* do this, that I hope to test soon.

[Zack Gainsforth]

Ok, I'll start with this recent image of mapping a glass melt inclusion in olivine about 120 um in size. Conditions were 15 keV, 30 nA and 2 sec (2000 ms) per pixel. 256 x 256 pixels.



It is quite reproducible and also seen to a much lesser degree in the other elements (e.g., K) as might be expected.

Can you compare closely to other elements?  If this is in fact diffusion, then you should expect a complimentary pattern.  Specifically, Na diffuses much more easily than many other elements, and you may have another cation diffusing oppositely to the Na to balance charge.  If so, then the patterns for all elements should not be the same, but they should be complimentary.

If they are all the same, then I'm going to put my money on an instrumental fluctuation: like (to make stuff up) the emission from your electron gun is varying, or your detector amp has some noise on it.  You maybe can check this by looking at the amplitude of the signal.  I notice the "Wt%" of the Na is varying by about 1/2 Wt%.  So, do other nearby elements also vary by about that?  Compute the variance of the concentration as a function of element Z.  Maybe you get an overvoltage curve....

[Probeman]

Can you compare closely to other elements?  If this is in fact diffusion, then you should expect a complimentary pattern.  Specifically, Na diffuses much more easily than many other elements, and you may have another cation diffusing oppositely to the Na to balance charge.  If so, then the patterns for all elements should not be the same, but they should be complimentary.

If they are all the same, then I'm going to put my money on an instrumental fluctuation: like (to make stuff up) the emission from your electron gun is varying, or your detector amp has some noise on it.  You maybe can check this by looking at the amplitude of the signal.  I notice the "Wt%" of the Na is varying by about 1/2 Wt%.  So, do other nearby elements also vary by about that?  Compute the variance of the concentration as a function of element Z.  Maybe you get an overvoltage curve....

At the risk of being wrong I'd have to say it's not an instrumental effect as the Na signal does not vary at all in the host phase (olivine)- though it is essentially zero.

Also I recall Si and Al do not seem to diffuse as much as Na and K under normal TDI measurements. Typically one will see TDI variations of Na and K several times larger than Si and Al. See here for examples:

http://probesoftware.com/smf/index.php?topic=116.msg461#msg461

On the other hand, maybe the images *are* consistent with that hypothesis- though wouldn't that mean that the ion migration is real, rather than an instrumental artifact? But then what causes the periodicity?  Is it related to the size of the inclusion for some reason?

Of course, I've always assumed that it is the thermal heating of a glass which mobilizes the alkali ions (which is why we see Kearns' so called "incubation time" as seen here):

http://probesoftware.com/smf/index.php?topic=116.msg454#msg454

And subsequently those alkali ions are then attracted to the sub-surface (dynamic) primary electron charge, hence the ions are drawn *deeper* into the sample, therefore exhibiting greater absorption losses, while the Si and Al atoms are less mobile and only show a minor increase in intensity due to the reduced degree of absorption by Na and K?

But, here's some of the other elements plotted so you can see them yourself:

[Karsten Goemann]

Don't suppose you can do TDI on your maps can you, that's probably asking to [sic] much...

Funny you should mention this, but I've got a very cool idea on how one *might* do this, that I hope to test soon.

Couldn't you do multiple passes over the same area similar to what many EDS softwares do? Would require very stable position though.

[John Donovan]

Don't suppose you can do TDI on your maps can you, that's probably asking to [sic] much...

Funny you should mention this, but I've got a very cool idea on how one *might* do this, that I hope to test soon.

Couldn't you do multiple passes over the same area similar to what many EDS softwares do? Would require very stable position though.

Maybe... stay tuned!