Lab Problem - Carbon Coating not Conductive

[Andrew Mott]

Our lab has developed an issue with our carbon coating process over the last month. Samples are coated as normal (same protocol as the last two years that worked successfully), but any sample that is mounted in epoxy or geologic thin sections (non-conductive samples essentially) that are subsequently placed into our instrument are heavily charged - resulting in charging not only on the sample, but on the sample holder.

(1) Samples that are mounted in conductive mounting or that are carbon taped to a metal cylinder for analysis still appear to be fine and can be imaged and analyzed with good totals after coating.
(2) Non-conductive samples that were coated > 2 months ago still work (for imaging and analysis) - both epoxy disks and thin sections

I have successfully imaged and analyzed epoxy disks (today) that were coated 3 months ago, then polished and re-coated them. Subsequently they have the charging issue and cannot be imaged or analyzed.

I've been in touch with the carbon coater company and with our instrument engineers and we're a bit stumped on what the problem could possibly be.

Does anyone have any ideas? I have done a complete clean out of the coating chamber and tried different sets of carbon rods. All that I have left to check is the hosing for some sort of contamination.

Thanks for your time.

[Probeman]

Hi Andrew,
This has happened to me in the past also.

In my case it was due to a small air leak in the chamber that was perhaps oxidizing the carbon vapor? I don't know, but once we fixed the leak everything was good again.

Also I have noticed on our Edwards 306A coater that if the heat (current) to the rods is low, the coating is much less conductive.  I generally mention to students that they should look for sparks about 2 to 4 cm long when evaporating carbon. No more, but no less in general.

On a slightly related topic we started seeing about 200 PPM of chlorine in all our standards after one of our maintenance people replaced a backing pump line on our carbon coater with a PVC hose. We had to clean the entire coater vacuum system and replace the hoses with high vacuum compatible hoses and all was normal again.

http://probesoftware.com/smf/index.php?topic=423.msg2288#msg2288

john

[Andrew Mott]

John,

Going off your ideas I increased the length of time I took between the coating and when I opened the air inlet valve to the bell jar (in an effort to combat potential oxidation of the surface while the apparatus was still heated). I also tried to maximize the current I was using for the coating. Some combination of these two changes made the difference and my samples are now conductive.

Thanks for the help!

Andrew

[Probeman]

John,

Going off your ideas I increased the length of time I took between the coating and when I opened the air inlet valve to the bell jar (in an effort to combat potential oxidation of the surface while the apparatus was still heated). I also tried to maximize the current I was using for the coating. Some combination of these two changes made the difference and my samples are now conductive.

Thanks for the help!

Andrew

Hi Andrew,
Glad to hear that worked for you too.

But isn't it strange that simply reducing the rod heating current not only reduces the conductivity of the carbon coat, but apparently even makes it insulating!   Someone should look into this.  I'll bet there's some interesting chemistry going on in our carbon coaters...
john

[Les Moore]

A little trick that has been shown to me that I use all the time is to make the point "trumpet shaped".  By this I mean that the point is an almost parallel section that curves to the full thickness.  This deposits more C earlier and you don't need to bake your sample with radiant heat to get a thick coating.

I use 300grit papre around a thick felt tip marker body - messy but effective. 

[Probeman]

A little trick that has been shown to me that I use all the time is to make the point "trumpet shaped".  By this I mean that the point is an almost parallel section that curves to the full thickness.  This deposits more C earlier and you don't need to bake your sample with radiant heat to get a thick coating.

I use 300grit papre around a thick felt tip marker body - messy but effective.

Hi Les,
I think I know what you mean (neat idea), but please post a couple of photos, especially showing how you create this "trumpet shaped" tip.
john

[Probeman]

There are several more sophisticated methods for determining carbon coating thickness while carbon coating, but the color on polished brass seems to me to be a fairly direct (what about density variations?) method, considering that it relies on Newton's colors. 

By the way, I suspect the main problem for some labs is the reliance on sputter carbon coaters (as opposed to actual carbon evaporation) for quantitative analysis of non-conducting samples. If you do use a carbon sputter coater for quantitative analysis, you might want to test the uniformity of your carbon coat thickness by placing a large (5 cm diameter or more), polished brass plate in your sputter coater and see how much variation you see from the center to the edge.

This is a chart that Julie Chouinard produced when she started in our lab, and it's currently laminated in plastic and posted over our Edwards 306A carbon coater:



An observation: for reasonable conductivity one really only needs 10 or 15 nm of carbon (depending on the smoothness of the surface), but the transition from red to blue (at 20-25 nm) is quite striking and therefore we use this color change for thickness reproducibility.

Note that for analysis of low energy x-rays (e.g., oxygen and nitrogen Ka), one should carbon coat their standards and unknowns together at the same time.

I also attach the PowerPoint presentation of the carbon coat color document in case any one wants to modify it for their own purposes (remember to login to see attachments).