Alicona Mex

[Probeman]

We have used the Alicona Mex software for quite a few years:

http://www.alicona.com/produkte/mex/

The product works well for transforming SEM images taken from two or three different sample tilt angles into a DEM (digital elevation model) to allow users to perform quantitative cross sections (profiles) and also to calculate volumes and areas. One needs to document the FOV, tilt angles and the working distance to obtain quantitative results. Here's an example of some work we've done in the past using their free viewer:



Recently we updated to their latest software and the new version now wants an additional parameter to be added to the working distance, which they call epsilon. The diagram in their manual defines it as this:



The working distance is easy, but we've hit a roadblock on trying to learn the epsilon values for our FEI Helios and Quanta instruments.  Calls to FEI/Thermo support have not been helpful.  Does anyone have any insight into what the epsilon values are for these FEI SEM/FIB instruments?

[UofO EPMA Lab]

We have used the Alicona Mex software for quite a few years:

http://www.alicona.com/produkte/mex/

The product works well for transforming SEM images taken from two or three different sample tilt angles into a DEM (digital elevation model) to allow users to perform quantitative cross sections (profiles) and also to calculate volumes and areas. One needs to document the FOV, tilt angles and the working distance to obtain quantitative results. Here's an example of some work we've done in the past using their free viewer:



Recently we updated to their latest software and the new version now wants an additional parameter to be added to the working distance, which they call epsilon. The diagram in their manual defines it as this:



The working distance is easy, but we've hit a roadblock on trying to learn the epsilon values for our FEI Helios and Quanta instruments.  Calls to FEI/Thermo support have not been helpful.  Does anyone have any insight into what the epsilon values are for these FEI SEM/FIB instruments?

I'm bumping this topic because I'm hoping that the lack of responses simply means that it has just been overlooked! My student and I can get the Alicona Mex software to produce reasonable results, but only after entering an epsilon value of 1000mm!   

Does anyone else use the Alicona Mex software for 3D reconstruction of tilted SEM images?

[Mike Matthews]

Hi John,

We use MeX but no, we haven't worked out how to determine epsilon either but I think the electron image working distance (as opposed to the z-axis height) 'should' be closer to 'pd' than 'wd' in the diagram. So far we've used relatively long working distances so epsilon is relatively small and we assume it's zero.

If you have a topographic standard it's obviously easy to test the settings. I haven't sketched it out or tested it to confirm but I think if the settings are wrong a flat surface would be modelled as curved.

[Probeman]

We use MeX but no, we haven't worked out how to determine epsilon either but I think the electron image working distance (as opposed to the z-axis height) 'should' be closer to 'pd' than 'wd' in the diagram. So far we've used relatively long working distances so epsilon is relatively small and we assume it's zero.

Hi Mike,
What is the "electron image working distance" and how is that different from the normal working distance?  Just to be sure we are on the same page, I call the distance from the bottom of the objective lens to the sample surface, the working distance. 

If you have a topographic standard it's obviously easy to test the settings. I haven't sketched it out or tested it to confirm but I think if the settings are wrong a flat surface would be modelled as curved.

Yes!  That's exactly what I suggested the student to try. That is, put in an object of known size and shape and try to "back out" this so called epsilon distance from the instrument. 

What gets me is that he's tried to ask Alicona what this epsilon distance should be (or how to measure it), but they don't respond.  Also their manual (in the version 6.2 that we have), has several critical typos, so it's making it rather difficult to troubleshoot this problem.

Finally, why do we only get halfway reasonable results when specifying an absurd value of 1000mm (1 meter!) for the epsilon distance?

[BenjaminWade]

Hi John
Its probably zero help as I cant give you the origin of the number, but we have MeX and use it occasionally and our epsilon value is entered as 6mm. I cannot say if that is something FEI told us for our XL30 FEG (which the 3D images are taken on), or if that was entered in by whomever setup our MeX in the past. Unfortunately no one left working here had anything to do with installing and setting up the MeX software so I can't enquire. Having said all that I have many years ago put a step standard in an it did a pretty decent job using those settings.

Cheers

[Probeman]

We use MeX but no, we haven't worked out how to determine epsilon either but I think the electron image working distance (as opposed to the z-axis height) 'should' be closer to 'pd' than 'wd' in the diagram. So far we've used relatively long working distances so epsilon is relatively small and we assume it's zero.

Hi Mike,
What is the "electron image working distance" and how is that different from the normal working distance?  Just to be sure we are on the same page, I call the distance from the bottom of the objective lens to the sample surface, the working distance. 

If you have a topographic standard it's obviously easy to test the settings. I haven't sketched it out or tested it to confirm but I think if the settings are wrong a flat surface would be modelled as curved.

Yes!  That's exactly what I suggested the student to try. That is, put in an object of known size and shape and try to "back out" this so called epsilon distance from the instrument. 

What gets me is that he's tried to ask Alicona what this epsilon distance should be (or how to measure it), but they don't respond.  Also their manual (in the version 6.2 that we have), has several critical typos, so it's making it rather difficult to troubleshoot this problem.

Finally, why do we only get halfway reasonable results when specifying an absurd value of 1000mm (1 meter!) for the epsilon distance?

This is somewhat embarrassing, but the problem was caused by me not realizing that with the new version of Mex (6.1), they calculate the horizontal field width differently. That is, one used to just enter the HFOV in whatever units specified, but with the new version of Mex they want you to enter in the distance by drawing a line on the image that corresponds to an object with a known length. They then use this number to calculate the distances needed...

It all works fine now!   

[Ben Buse]

Hi John,

That allows you to determine the horizontal sampling distance, but doesn't allow you to determine the epsilon value

Asking because want to know what the epsilon value should be - or is it - just use 6?

Ben

[Probeman]

Hi John,

That allows you to determine the horizontal sampling distance, but doesn't allow you to determine the epsilon value

Asking because want to know what the epsilon value should be - or is it - just use 6?

Ben

Well it would depend on the instrument in question, right?  Is this an SEM?  The optimum working distance for our old FEI was 10 mm (for optimized EDS), but of course the distance could vary depending on the stage height.

For a Cameca EPMA instrument that would be 9 mm. Not sure about a JEOL instrument..

[Ben Buse]

The epsilon value is not the working distance, it's the additional amount that the crossover is above the pole piece

The trouble is how do you know it

And does Mex default to epsilon of 6

[Probeman]

The epsilon value is not the working distance, it's the additional amount that the crossover is above the pole piece

The trouble is how do you know it

And does Mex default to epsilon of 6

Oh, that's right!  It's been years (since 2018!) since I played around with the Mex software.

I would contact the manufacturer of your instrument, whatever it is...

[Ben Buse]

Manual Projection distance: Note the projection distance during the capturing phase. Both images need the same projection distance. Please note that the projection distance(pd) is required for the stereo and auto calibration measurement consists of the EM working distance(wd) and the epsilon value (this value is EM dependent). In general an epsilon value around 10 mm leads into good measurement results. If you enter wrong or bad values for the projection distance (for example you have not entered the value) your result height model may look curved.

Just running some tests changing e from 4 to 10, but got to run for bus

[Probeman]

Manual Projection distance: Note the projection distance during the capturing phase. Both images need the same projection distance. Please note that the projection distance(pd) is required for the stereo and auto calibration measurement consists of the EM working distance(wd) and the epsilon value (this value is EM dependent). In general an epsilon value around 10 mm leads into good measurement results. If you enter wrong or bad values for the projection distance (for example you have not entered the value) your result height model may look curved.

Just running some tests changing e from 4 to 10, but got to run for bus