Continue to Site

Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

  • Congratulations cowski on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!

Bonding A380 Die-Cast Aluminum and 6061 Aluminum 1

Status
Not open for further replies.

wagner89

Mechanical
May 12, 2011
5
Hello adhesive experts,

I am currently working on the next generation of a solar panel microinverter and am tasked with finding an adhesive that will act as both a sealant and a permanent bond. We require the sealant aspect in order to keep out moisture and the permanent bond feature with the aim of reducing or eliminating the number of mechanical fasteners used. The adhesive needs to bond two aluminum surfaces (A380 Die-Cast and 6061) and be able to withstand a temperature range of -40 to 90 degrees Celsius, an outdoor environment, potential exposure to UV light, and have a life of 25 years. The adhesive will need to be applied in a "channel" that runs the perimeter of the product. The total volume of the “channel” is approximately 0.08 cubic inches. What would be the optimal adhesive for these requirements?

I would appreciate any information/advice/insight into the proper adhesive for this application.

Thanks in advance!
 
Replies continue below

Recommended for you

It's unfortunate that you seem to have put the cart before the horse in designing the mechanical configuration of the items to be bonded before giving any consideration to the adhesive.

Your conditions appear challenging. A more holistic design approach would have been better.

It would be quite useful if you could better describe the configuration, and even more useful if you could define the loads that the bond needs to support.

 
Thanks for your response. The conditions are indeed challenging.

I forgot to mention in my initial post that the sealing action of the adhesive is the primary concern and the bonding aspect would be nice but not necessary. The microinverter won't undergo any substantial loads other than maintaining a seal between the two aluminum surfaces. There may be some vibration due to wind or forces due to thermal expansion.

As for configuration, the microinverter will consist of the A380 die-cast aluminum that will hold the electronics and the 6061 aluminum cover. I have attached a pdf file showing a cross section of the design with an arrow indicating the location of the "channel" where the adhesive will be applied. The "channel" has a depth of 0.04 inches, a width of 0.06 inches, and a total volume of approximately 0.07335 cubic inches.
 
 http://files.engineering.com/getfile.aspx?folder=42e893d7-110a-4bfb-8641-04f6e2150cd6&file=Adhesive_Cross_Section.pdf
With a temperature range of 130K you can be certain that there will be forces associated with thermal expansion.

I'm too lazy to look up what the difference in thermal expansion between A380 and 6061 is, but you should not be.

I'm assuming that you've shown a section. What will the pressure differential from inside to outside be?

I think either a two-part silicone or polyurethane could work in the environment, but I don't think that you have sufficient gap to keep the shear stress within capability.

The shape suggests that you'll get some bending of the parts, and thus pealing of the adhesive. You might not have enough surface area to resist those loads.

You're going to have to do your homework to get a much better handle on the expected loading. Then give Dow or Master Bond a call.
 
Polyurethane seems like a good choice due to bonding, flexibility, and UV resistance.
 
The selection of the adhesive is one issue, but the surface preparation of the aluminium is just as important as the type of adhesive. It is relatively easy to generate a short-term strong bond with minimal treatment (e.g. just a solvent clean) BUT clearly for your customer's sake and your warranty validity, you need long-term bond durability, and that is not as easy to establish.

Firstly, recognise that adhesive bonds require chemical reactions at the interface, and for those reactions to occur the surface must be chemically active. This can be achieved by etching or anodising or even grit blasting to remove surface oxides. A solvent degrease is not adequate, and hand abrasion is marginal at best.

However, even that is not sufficient to assure long term bond durability. Aluminium surfaces have an affinity for forming hydrated oxides. Over time, Al2O3 forms Al2O3.2H2O and in the process, any short-term chemical bonds originally formed with the adhesive will dissociate because it takes less energy to form the hydrate than it takes to maintain the chemical bonds at the interface. Hence, the original chemical bonds dissociate so that the hydrate can form, and you have a disbond.

Now don't kid yourself that the "sealant" action will keep water out. Moisture will gradually penetrate almost any bond to aluminium. The key to success is to treat the surface to prevent hydration. The best method is phosphoric acid anodising, but a simpler method may be to use an organo-functional coupling agent which is a polymer in which one end of the chain attaches to the metal surface by a covalent bond. The other end of the polymer must be compatible with the adhesive type you select. We have used such a product for epoxy bonds for aircraft repairs and achieved a reduction of repeat-repair rates at one major airbase from 43% in 1992 to less than 0.07% since then, and all of those failures occurred because of technician malfunctions. Where the correct procedures were followed no bond failures have been recorded. The USAF experience with the same process over fifteen years is identical.

The best accelerated test to validate bond durability is the wedge test ASTM D3762, BUT the acceptance criteria stated in that standard are hopelessly inadequate. Better acceptance criteria are available in the attached link.

Now if you get the adhesive selection right and combine that with a valid surface preparation approach, and the bond overlap length is acceptable (generous), you may be able to eliminate the sealant groove, and that may save you the additional thickness of the material to allow for the groove and the cost of machining/casting the groove. And you will have a product that stays bonded for a very long time.

Regards

Max Davis

PS For our US friends, please replace the "s" with "z" in such words as anodiZe, and drop the extra "u" from "Aluminium". Never did understand why the US is not consistent; if they want Aluminum, why not have Lithum, Berilum, Chromum, Uranum, Plutonum, Thorum etc. ;-)} The } is my beard.
 
 http://www.tc.faa.gov/its/worldpac/techrpt/artn0657.pdf
The English language allows such variations and many are for
historical reasons.

I have not read Emerson but he is reported to have said:

'A foolish consistency is the hobgoblin of little minds.'

Apply as required.

 
@MintJulep: I appreciate your advice and the suggestion of potential adhesives. I will definitely be doing my homework. I first wanted to get advice on the forum on where to start. I believe it will be sealed at atmospheric pressure so any pressure differential would occur in changes in the outside pressure. We may redesign to eliminate the "channel" and have two flat surfaces being bonded together. Are there equations I can use to calculate the gap and surface area? I'm not quite sure about the bending forces; I may be able to post a better image of the entire design so you can get a better look.

@CoryPad: Thanks for the confirmation of polyurethane as a type of adhesive to research.

@blakmax: I appreciate your post and the attachment. I hadn't considered surface treatment and it sounds like a major aspect for choosing the proper adhesive. I believe that the substrates are being anodized at the moment but I will definitely research the proper surface treatment for our product. To clarify, I need to decontaminate the surfaces, then anodize in order to chemically activate the surfaces, and then it must be made resistant to hydration. As I mentioned to MintJulep, we are considering a redesign to eliminate the groove. How can I calculate the minimum bond overlap length? Thanks for the advice on testing for the durability of the bond. Also, part of your post makes it sound like it's not possible to make a durable bond between two aluminum substrates but the other part makes it sound possible, could you clarify this?

@jedward: I understand the quote but I am not quite sure how it is relevant.
 
There is a very good adhesive used to hold washing machine motor frames togather. I was going to sample it to see what it was but my scrap motor got scrapped.

I'm trying to find a hot glue/sealer that is used for Aluminum, my internal tickler file isn't working.

Here are two people I've dealt with for adhesives and sealants. Make sure you put your life expectancy requirement in any conversation.



 
Wags
You are probably using a hard anodise, which may or may not not be suitable for bonding. You would need to undertake a wedge test. Don't rely on strength tests because they only measure short term strength. Phosphoric acid anodising is very effective for bonding surfaces, but it would probably not be suitable for geberal environmental protection anodising.

It is certainly possible to generate a durable bond to aluminium. As I said in my post, we have not had bond failures for nineteen years.

The procedure should be to

1. degrease (don't use detergents because they will wet the surface and act as a contaminant themselves)
2. chemically activate the surface. We usually grit blast, even for anodising.
3. Anodise or treat with a couplig agent.
4. You may wish to use a corrosion inhibiting bonding primer especially if you anodise. Most of our work used the couplig agent and we never used a primer, and never found corrosion on our bonds because they didn't fail.

The overlap length will depend on loads, the thickness of the materials you are joining and the properties of the adhesive at the maximum service temperature. The procedure is in the reference I attached last posting. I suspect that the loads are relatively low so you could probablyy get away with Hart-SMith's rule of thumb and make the overlap 30 times the thickness of the material being joined.

Regards

Max

PS Glad you understood Jedwards quote.
 
The word microinverter suggests a high switching frequency.
... which in turn suggests that electrical bonding may also be necessary in order to keep the FCC off your back.

Ask your sparkys about that.





Mike Halloran
Pembroke Pines, FL, USA
 
A friend of mind that has done considerable work in the adhesive/sealer word stated that you should be looking at sealer first and adhesion second if you are using mechanical fasteners..
He didn't have a specific recommendation as he stated he has been out the loop too many years. He did mention Henkel, Devcon, and Lord.

One industrial spy is worth 3 PHDs. From a local shop that bonds and seals all manner of materials the following companies products stood out.

The "Metal Welder" from Devcon. Devcon also carries the Permatex line of gaskets sealers. The metal/glass sealers from Ramapro. last but not least is Henkel, especially the Hysol Brand materials




SE_HTML.htm
 
@unclesyd: I appreciate the advice and the potential companies. I'll definitely contact them.

@blakmax: Thanks again for your informative post. I spoke with a sales representative today who informed me that anodizing the "channel" would not be necessary and that I should use an adhesive primer instead. We will already be anodizing the parts so it seems unnecessary to cover the channel while anodizing and then using a primer. What are your thoughts on this? Also, for clarification, what exactly is the bond overlap length?

@MikeHalloran: What does FCC stand for?
 
Wags

The same thing applies to a bonding primer! The primer requires a chemically active surface to form chemical bonds, just the same as the adhesive requires. I have seen a number of failures where the primer forms a neat layer on the surface of the adhesive, but disbonds from the metal.

A primer will only work on a metal without pretreatment if it produces a strong acid-base reaction. I would still recommend a light grit blast after solvent degreasing, even if you use the primer. That way, if the primer requires a chemcially active surface, you have it. If it has a strong acid-base reaction, then that will also occur. Don't solvent degrease after grit blasting because that will spread contamination over the chemcially active surface. Blow the dust off with dry gas. We use bottled nitrogen as the grit blast propellant and also to blow the dust away.

As I last posted, the overlap length will depend on the thickness of the materials being joined and the loads expected to be carried. It also depends on the design philosophy you employ for the joint. For reasonable thin materials (less than about 0.14 inches) it is actually possible to design the joint such that it will be stronger than the metal, and the adhesive will never be the locus of failure for the joint. The 30 times the thickness I previously mentioned would achieve that.

However, I guess that you are really not dealing with loads that will break the metals, so you probably could use the 500 psi average stress approach. (A large number of manufacturers in the US use this to design aircraft, but it is totally incorrect because the shear stresses in adhesive bonds are not uniformly distributed. The only way the aviation industry gets away with it is that the design is supported by a huge test program.) One outcome is that the adhesive may be the locus of failure.

You would need to estimate (or guesstimate)the loads to be transferred (in lb/in) and then divide that load by 500. The answer will be the overlap in inches. Now if the structure was to carry significant loads, I would strongly advise against using this method.

Hope this helps

Regards

Blakmax
 
Okay, I've got a few updates updates.

First, we will most likely not be eliminating mechanical fasteners so my concentration is now on finding a cost effective sealant against liquid water.

Second, the location where the adhesive will be applied has changed along with the dimensions. The dimensions are 0.09 inches wide, 0.01 inches high, and about 30 inches around the perimeter. I have attached a pdf showing the new location/dimensions. The thickness of the lid is 0.06 inches for the prototype and will be 0.04 inches when it goes into production.

@blakmax: I've been doing a lot of reading and I keep coming across the terms overlap length and bond width but no one defines these terms. Could you clarify what exactly these two terms refer to when talking about adhesives?

 
 http://files.engineering.com/getfile.aspx?folder=84d5981c-265d-4727-aedc-0691356d6ccf&file=Adhesive_Cross_Section_1.pdf
An RTV silicone sealant is probably a good choice for your application.
 
Wags 89
You ask:

@blakmax: I've been doing a lot of reading and I keep coming across the terms overlap length and bond width but no one defines these terms. Could you clarify what exactly these two terms refer to when talking about adhesives?

Overlap length is the length of the bond in the load direction. The bond width is transverse to that direction. If you define the loads in lb/in width, the bond width is meaningless. The overlap length is the critical issue, You need to have enough adhesive in the major load direction to support the applied loads.

In the example you state, the lengths are totally inadequate for any significant structural loads. Bond overlap lengths are typically more than 0.5 inches for thin materials and high performing adhesives, and increase with adherend thickness and maximum service temperature. You are talking about lengths less than 0.1 in. that would not support even moderate loads.

With regard to the use of fasteners, lets be clear. If you have a structurally effective bond and your surface preparation is effective and if your overlap length is adequate, any fasteners will carry absolutely minimal load. They are totally superfluous.

I really think you are talking about a sealant to keep moisture out, rather than an adehsive to form a structural joint. In that case design your joint as if there is no contribution from the bond. Under no circumstances try to design for the load capacity of the fasteners and then think you can add onto that the contribution from the adhesive. These strengths are not additive.

Regards

Blakmax
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor