Strange spectra

[Probeman]

Something really weird happened yesterday on our Quanta SEM.

I had a user on the instrument running on the instrument who was switching between imaging and acquiring EDS spectra, and I was starting to hear hearing him cursing a bit, and so went over to see what the matter was. He said that the EDS system was randomly not working and so I looked at it and sure enough, it was only displaying a very ragged peak roughly around the 0 keV region.  The spot size looked good, keV looked good, so I then turned on the chamber light to see what the geometry looked like and all of a sudden the full spectrum started being acquired, though still a little ragged looking.

The chamber light is of course the usual LED infrared system on most SEMs these days, but I was curious as to why turning on the chamber light would cause the spectrum to go from just a small noisy peak near 0 keV to a full spectrum. I tried several times and each time the light was off, the spectrum would be just a small noisy peak near 0 keV, and each time I turned on the chamber light it would go full spectrum to 15 keV.

Finally I noticed that the user had the sample at around a 7mm working distance and so since our EDS system is aligned for a 10 mm working distance, I lowered the stage a few mm, and lo and behold, the spectra came right back with nice sharp peaks!

Now I understand that if the backscatter detector is partially blocking the EDS detector line of sight to the beam, the spectrum will look pretty weird. I've also seen similar spectra on the EPMA when using samples that are top referenced, leaving a small lip around the sample edge, when we attempt an analysis within a mm or so of that lip. But here's my question: why the heck did turning on the chamber light make such a difference?  Particularly an infrared chamber light?  I'd expect a worse spectrum when the chamber light was on, but in this case it got much better...

Any ideas?

[SteveSeddio]

Am I understanding correctly that when the IR light was on, a spectrum with "normal," expected peaks was acquired?

What was the deadtime when this happened?

[Probeman]

Am I understanding correctly that when the IR light was on, a spectrum with "normal," expected peaks was acquired?

What was the deadtime when this happened?

Hi Steve,
Yes, when the IR light was off, no peaks, but when the light was on the spectrum looked essentially normal. The deadtime was quite low I think.

I suspect it was some sort of a grazing effect from the BSE detector, due to the working distance being so short.

[SteveSeddio]

Does something physically change position in your chamber when the IR comes on?
That sounds pretty weird.

[Probeman]

Does something physically change position in your chamber when the IR comes on?

Not that I know of.

That sounds pretty weird.

Exactly my words in my initial post.