Detection of anodising layer on aluminium 5

[gvh]

I have a little coated plate (50x50x0.6mm). After removing the paint, I have been asked to proove that there also is a anodising layer. I was thinking about dipping in nitric acid and watching how much material seem to come off. But I only have a small plate. If this does not work, I don't have a second chance. Anyone?

 

[CoryPad]

I see two ways to do this:

1) make a metallurgical cross-section, then polish and etch it. An anodized layer will be detectable.

2) measure the indentation hardness (either surface or cross-section). The anodized layer will be harder than the base metal. You will need to choose a proper indenter and force based upon the materials and geometry.

 

[kenvlach]

Two easy methods:
1) Check for electrical continuity -- anodize is a good insulator, whereas a probe will easily break through the oxide film on bare aluminum.
2) Try scratching with a piece of steel -- aluminum will scratch, anodize won't.

 

[davefitz]

Anodized aluminum has an infrared emmissivity index of greater than 0.85, while non anaodized aluminum has an infrared emmissivty less than 0.15. Try using a infrared radiation meter, adjust the settings to read emmissivity, and voila!

 

[TEV]

Scanning electron microscopy with energy dispersive x-ray spectroscopy will tell you if you have an anodized layer. You may need comparison samples known to be anodized and not anodized.

 

[kenvlach]

TEV,
For SEM to detect anodic coating, you don't need an EDS analyzer or standards. The anodic coating, being a nonconductor, will immediately charge up and deflect the electron beam all over the place. When studying anodic coatings via SEM, it is necessary to first prepare the sample by evaporating a conductive film of carbon (or gold).

gvh,
For a good answer by non-destructive means, find a local anodizing company and ask them to measure the anodize thickness. It is quick & easy using an eddy current tester, and if you know the alloy, it can be accurate to 0.00001 inch.

 

[AlanD]

Using a cross sectional view of the sample, the anodized layer should be clearly visible under a microscope after a HF etch or clearly visible in an SEM, no coating required.

 

[TEV]

Kenvlach - electron beam charging depends on anodization (oxide) thickness, primary beam voltage, and imaging mode. It is actually possible to image areas of differing oxide thickness in an SEM by varying the accelerating voltage. As you decrease the voltage, the secondary electron emission coefficient of oxides increases much more rapidly than that of the metal, and beam penetration also descreases. So, as you lower the accelerating voltage, thinner and thinner oxide areas become brighter in the secondary electron imaging mode. I have used this to plot oxide distribution on aluminum.

 

[kenvlach]

TEV,
I think my comment is generally true, to study a non-conductive surface by SEM requires a conductive coating.
If you studied the initial stages of oxide formation on aluminum, no C-coating would be needed since the aluminum would suffice. And of course for edge view as mentioned by Alan, you have adjoining conductive layer.
Re SE image, I recall that rastering the beam over a large area also helps reduce charging, but I had to use sufficient voltage (10 KeV?) to do Al, Zn & Mg mapping of the surface oxide for the alloy in question. I would like to see your reference for studying anodize w/o coating.

But, we're getting sidetracked. gvh has this small plate & he wants an NDT for determining anodize presence.
He clearly doesn't want cross-sectioning or need a method that costs more than replacing his plate. Other than the IR method mentioned by davefitz, I feel my suggestions are the only practical ones.

gvh, please print out this thread, take it with your plate to an anodizing shop and beg a free thickness test (better than bothering with billing) and ask their opinion of the various proposed methods.

 

[gvh]

Thanks you all.
I have tried to measure electrical resistance, but no resistance could be measured with a normal multimeter. So if there is an anodised layer, it will be very small. Since the component has very high corrosion resistance in Salt spary test (this was done on an other part from the the same plate) we know there must be some layer. But too small for us to detect, even by light microscopy. I do have an old SEM (no edx)so my next step will be to try this, with a very low accelerating voltage. I will keep you all informed here.

Extra remark for this site: I really did not think at first that so many people would respond. Now I have tried it, am very ethousisatic about this site. Keep up the good work!

 

[TVP]

kenvlach writes:

gvh has this small plate & he wants an NDT for determining anodize presence. He clearly doesn't want cross-sectioning or need a method that costs more than replacing his plate. Other than the IR method mentioned by davefitz, I feel my suggestions are the only practical ones.

To which I have the following comments:

1. gvh made no such claim that only NDT methods could be used. Obviously he/she does not want to destroy the entire sample, but that does not rule out some destructive testing.

2. The second sentence is complete conjecture. It is quite common to perform analyses on a component(s) for a variety of reasons (premature failure, quality control, customer demand, etc.) that cost more than the price of replacing an individual component.

3. Placing value judgements on the quality of the other replies based on one's limited perspective is borderline offensive. All of the suggested methods have merit, and it is up to gvh to make a decision on which method to pursue, if any.

 

[kenvlach]

TEV,
OK, you've made your point.
I admit, I don't know what's inside gvh's head. I interpreted his concern about damaging the plate as a cost issue, whereas it could be a forensic examination.

Your remark "one's limited perspective is borderline offensive" is rather offensive, as I have done plenty of slice & dice, mounting & polishing, SEM & EPMA work, so I have used all the methods mentioned except the one by IR.

I was speaking from years of experience in the aluminum anodizing business regarding what was practical. The vast majority of customers that I have dealt with were concerned with cost.
And compared to the SEM/TEM/EPMA methods my research group used to study oxide films, alumina-aluminum alloy composites, etc., plus those I later used to study anodic coatings, I can definitely say that there are faster, easier and less expensive means to test for anodize presence and thickness.
We routinely use $2 pen-size continuity testers having a battery & light to verify that racks are free of anodize, or whether a casting has been completely descaled, etc, & sometimes a VOM to check whether a scratch has penetrated the anodize layer on a part. Eddy current testers are used for QC, i.e., for fast, accurate thickness measurements.

 

[kenvlach]

I apologize, the above post was in response to TVP' post.
And, I suggest that as gvh didn't find any electrical resistance, all of this is moot.

 

[CoryPad]

kenvlach,

I believe this thread is not moot and that some of your comments are speculative, not factual. gvh never stated how the paint was removed - if it were sand blasted, maybe some aluminium oxide coating was removed. gvh didn't state details regarding the multimeter measurements - if the electrodes were sharp-pointed, perhaps they could deform/fracture a thin oxide layer. If "proof" of a coating is necessary, then a combination of the suggested techniques may be necessary. I don't think you should be offended by TVP's comments - take them as constructive advice since we all have limited perspectives that influence our discourse.

 

[nbucska]

Myself when youth... tried to measure the resistance
of not-anodized Al with a drop of Hg as probe -- it took
more than 300 V to break through the oxide layer.
<nbucska@pcperipherals.com>

 

[kenvlach]

CoryPad,
you are correct re 'moot.' I intended something like, &quot;all further discussion is moot until we have more info from gvh,&quot; but suddenly found I was late for an appointment.
I also admit to thinking of gvh's case as if he were a customer who walked in the door with a part. I would have his answer in ~no time and for no charge.

gvh, since CoreyPad raised the issue, how did you remove the paint?

I'll risk sticking my neck out again and suggest that sandblasting, as commonly practised, is not recommended for removing paint from aluminum unless you know ab initio that a layer of anodize exists. We sometimes use glass bead.

Further, it would be highly unusual for a very thin anodize layer (what we in the business call &quot;flash anodize&quot to be present beneath paint since that is normally used for appearance sake, to give some protection on bright polished parts. When we have a call-out for paint over anodize, it normally is a significant thickness (Type II or III) for added protection from corrosion and deep scratches. A conversion coating is much more commonly used pre-painting. Again, this is a only a probability based upon years of experience, but we could be dealing with the exceptional case.

For aluminum parts that are free of paint, another referee method for anodize thickness (weight) is ASTM B137, which I like since I have the acids and analytical balance required, but I considered it impractical, so again I could be wrong if gvh has a lab.

nbucska, your method has been used for studying the oxide film that develops naturally on cleaned aluminum as a function of time [for this, as CorePad pointed out, a sharp probe would be unsuitable]. A thin layer forms ~instantly in air; then slowly, at a parabollically decaying rate, reaches about 300 Angstroms (?) after 30 days. Perhaps TEV has studied this by high-tech means and could explain better.