Determining carbon coating thickness

[Probeman]

OK, I have to admit that I'm a little surprised that more people are using the color on brass method than a quartz monitor.  I would have guessed the other way around.



But the data is still quite sparse (only 11 votes so far), so please, if you haven't already, log in and vote on what method your lab utilizes for monitoring the carbon coating thickness during deposition.

[Jarrod Crum]

I use the quartz thickness monitor to control thickness of my coatings.  The thickness monitor was calibrated by making coatings on a glass slide partially masked with Kapton tape. Then the tap was removed and the thickness was measured with an optical profilometer with a wide field of view to see the step height change from glass to coating.

I am curious.  How much impact does surface finish of the brass have on the color of the carbon coating?

[Probeman]

I use the quartz thickness monitor to control thickness of my coatings.  The thickness monitor was calibrated by making coatings on a glass slide partially masked with Kapton tape. Then the tap was removed and the thickness was measured with an optical profilometer with a wide field of view to see the step height change from glass to coating.

I am curious.  How much impact does surface finish of the brass have on the color of the carbon coating?

Hi Jarrod,
Do you mean the smoothness of the surface or how oxidized it might be?

We normally re-polish our brass disks using colliodial alumina (0.5 um), so they are very smooth and bright after a minute on the soft lap, but they tend to get a bit "wavy" after repeated polishings.  It doesn't seem to affect our ability to see Newton's colors. 
john

[Karsten Goemann]

Hi Jarrod,

We have a quartz thickness monitor on our new coater as well and an optical profiler in the building that we may be able to use for thickness calibration. So it would be good for us to establish this as an additional/alternate method of determining thickness.

Can I ask a few more questions regarding your setup and procedure? What type of coater do you use and do you see an effect of heat and/or light during evaporation on the quartz monitor thickness reading like we do? This seems to be one of the reasons for Leica to use pulsed deposition in their carbon thread coaters, although I haven't used a Leica coater myself, so I don't really know.

Have you got the quartz sensor mounted at the same distance to the carbon source as the samples? We use 100 mm as source-sample distance and there the amount of mass deposited on the sensor seems a bit low for our normal coatings (8-20 nm depending on application) to get a precise enough thickness reading, so I'm considering mounting it closer than the samples and using a tooling factor. That will probably affect heat/light sensitivity more...

Cheers,
Karsten

[BenjaminWade]

Hi all
Here are some measurements I did with our Quorum coater. There is no crystal, the carbon rods are shaped to what we hope is the same size every time then evaporate it under a specific current. Sometimes the whole rod evaporates...sometimes it leaves a little bit left...
We were assuming that each coat was nominally 10nm, so blank slides were masked with Kapton and coated up to 4 times. The stage only holds 4 slides and I should have made a note about which slides were on the outside vs the inside, but i didnt...so will probably have to repeat this again with perhaps moving them around after each coat to get some kind of average.




Of note is the thickness measurements were done on a laser profilometer and to measure the 1 and 2 coat thicknesses its probably getting to its resolution limit. I am going to analyse the same samples on the AFM in a couple of weeks to hopefully either confirm or change the thickness measurements of the 1 and 2 coat samples. If you just took the thickness of the 4x sample you do get nominally 10nm a coat.

Also when you pull the Kapton off it does lift the edge of the carbon coat, so you have to do your step height measurements a ways from the edge of the coat, at least 100um or more. Will repeat it trying a different masking technique using 4xTEM grids on a single slide in the centre of the chamber and report back.

Cheers

[BenjaminWade]

Hi all
Here are the results from the same measurements but using TEM grids as a mask. This time up to 5 repeat coats, all grids mounted on the same single thin section in the middle of the chamber.






So the average coat thickness is ~9nm +/-1nm, matching what I was told the nominal coat thickness of 10nm. I am surprised its so tight clustered really given that a different rod was used for every coat, and the errors in involved in measuring it all.
Here is the combined Resistance vs Thickness of all measurements. Might be able to use it as a very rough estimate for our users samples to determine an average thickness on their samples.



Cheers

[Karsten Goemann]

Hi Ben,

Awesome stuff. Good you're getting consistent results. Looks like with the TEM grid masking the optical profiler measurement works better? (with the Kapton it was looking more like the thickness doubled with every additional coating.)

Cheers,
Karsten

[BenjaminWade]

Hi Karsten
Yeah the TEM masking seemed more reproducible. I think the process of pulling the kapton off lifts the edge of the coat. To do these measurements I have to use the highest mag OBJ to get the resolution in Z I need, but this obviously limits my FOV. However I cant move far into the edge to measure thickness (where perhaps its more accurate), as I need a baseline/bare slide in the same FOV.

The masking still has/had its problems though. Below is a screen capture of a measurement from the 5x Coat



You can see the edges are obviously not sharp but a gradient, which I think is probably due to carbon bleeding/getting underneath the grid bars. Not a huge issue with the 5x coat as its so thick it reaches a plateau, but when you look at the 1x coat its not as clear cut and is much noisier. I held my grids down with strips of carbon tape, not sure how you could get around this as no matter how you hold them down they still wont lay perfectly flat and carbon will get underneath.

As you can probably tell from the image, the measurement shown above is from one of the edges of the grid, which I had to do so I could get baseline on both sides of the coat. With my 50x OBJ I cannot fit an entire "square" with baseline gridbars in the same FOV so had to resort to these edges. If I did it again I would probably use a 100 mesh grid so I could fit entire squares into the same FOV.

Also its not apparent in the 5x image above, but on the 1x/2x coat there are scanning artifacts created by our profilometer which utilises a scanning laser diode. This appears as horizontal striping at a z-scale big enough to make it a pain. I think your profilometer is an optical profilometer which should be much better at measuring thin coatings than ours.

Cheers

[sem-geologist]

I will give my two cents: We have both Quorum coater and Leica Coater. Quarum is using rods and Leica threads. The quality of it is like night and day, As I don't work in <15kV I am not so much caring about the real thikness of the carbon coating, however as I work a lot in high currents I am really interested in the quality. The same 20 nm on Leica is much lighter looking that 20nm on Quarum, they both use Quartz  thickness monitors. Now the stability under the beam is really where Leica couting shines, I have no problem with 200nA and 15 minutes!, and Quorum rhode coated (does not depend if with pulse method or with continues flash) coating is a joke - it disintegrates even in 50nA in 10 minutes despite looking much thiker than that with Leica. The quality of Leica coating is really worth the extra price tag in my opinion.

[Probeman]

Just so it's available to all here is the color change on brass chart that we have tacked over our carbon coater:



Also as an attachment below.
john

[Dan MacDonald]

Hi John,

We normally just use Wenol to polish off the coating on the slide, but something like diamond paste should work as well. The Ag paint wears off over time so we just repaint it every now and then.

Good morning, Karsten - I have done much he same thing in the past by measuring conductivity across the glass slide.  Instead of Ag-paint, I use copper tape with a conductive adhesive and it works very well, you just have to replace the tape every run.  Just an alternative for Ag-paint is all.

Best regards

Dan MacDonald
Dalhousie University