polycarbonate tubes containing LED strips cracking, suspect solvent fumes

[gonkin]

Hi, I'm an Australian artist who is working on a fairly large scale light sculpture. We have had a problem that I would like to seek input on from this forum.

The main component of the sculpture consists of a 5x5 grid of 5m PC tubes, each of which contain two back to back LED strips. After we assembled this structure we left it running overnight, the next day most of the tubes were crazed and cracked. The tubes were warm but not hot.

(see images below)

After much research I came across this forum and found information that has led me to believe that there is a problem with solvent fumes. The LED strips are embedded in an epoxy resin that is encased in a silicone sleave. On removing them from their packaging the strips smelt strongly of an aromatic solvent. My attempts to identify the solvent with the Chinese manufacturer have been rather confused, but it smells like old school marker pens ie Toluene or Xylene, which I gather from other posts here is particularly bad for polycarbonate. I think that what happened is that the heat of running the LEDs caused a higher concentration of fumes for the first 24 hours.

my questions

1) Some of the tubes didn't crack in the first 24 hours (I suspect that some of the strips weren't sealed in their packaging as well as others - allowing the fumes to evaporate more before being placed in the tubes). These uncracked tubes have not subsequently cracked in spite of another 7 days of continual running. Perhaps the solvent has evaporated enough not to be a problem? The strips are still somewhat smelly. Because the resin is enclosed in silicone I think it will take quite sometime to fully evaporate. If we replace all the tubes is it likely that we will have problems with cracking in the long term given a projected lifespan of 15 years?

2) given Polycarbonate seems especially bad for Environmental Stress-Cracking Resistance, is there another plastic that we should consider for the tubes? They are packed into a stainless steel grid structure that will be suspended between two buildings. They don't need to be especially impact resistant. They will receive some direct sun for a max of 1 or 2 hours per day.

many thanks

John Tonkin

https://dl.dropboxusercontent.com/u/4955130/grid-01b.jpg

https://dl.dropboxusercontent.com/u/4955130/cracks.jpg

 

[rickfischer51]

Is it possible the tubes that didnt crack are from a different source? Maybe the local distributor you used gets tube from two different sources. Is it possible that the lights are form two different sources? Try taking new lights and tubes. Turn on one light and let it run long enough to bake off the solvent. Put this light and a virgin lite in tubes and turn them on and see if there is a difference.

There is a technique called solvent-wash that can be used to determine the presence of a solvent. You take a solvent different from what you suspect and run FT-IF on it to get a baseline of the solvent. Then you dunk your contaminated part in, let it soak a bit, remove the contaminated part, and run FT-IR on the solvent again. You then subtract the first trace from the second. The difference is whatever is on the sample. You need a good lab to do this for you. I think there is an ASTM spec covering this procedure. Also, I think PC needs a UV stabilizer to prevent degradation. Check with your supplier. A test would be take a samples from a crazed tube and one of the tubes that survived and have them melt flow tested. A higher melt flow value means lower viscosity due to chain scission. I believe that ESC does not result in chain scission, so if the melt flows are the same, it means solvents are not the issue. If you have chain scission, that could be UV degradation, the extruder used too much regrind, poor process control, etc.

As far as alternate materials goes, stay away from acrylic. It crazes as bad or worse than PC. There is some clear rigid PVC pipe available that is supposed to be chemically resistant.

Rick Fischer
Principal Engineer
Argonne National Laboratory

 

[IRstuff]

Can't see your dropbox photos; our router blocks the site. The cracking looks to be full thickness cracks; it's a wonder that the tubes didn't just crumble, given the amount of cracking.

One issue I have with the solvent hypothesis is that the silicone sleeve presumably is not that permeable to the solvents in question. Your statement implies that you could smell the solvent in the sleeve, but not in the PC tube.

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529


Of course I can. I can do anything. I can do absolutely anything. I'm an expert!
There is a homework forum hosted by engineering.com:

http://www.engineering.com/AskForum/aff/32.aspx

 

[btrueblood]

Agree with Rick Fisher on alternate materials and UV susceptibility. Acrylic could work if the LED strips were outgassed first (before assembly), but determining how to, and how much outgassing is necessary is a bit of a science project.

Silicone is ridiculously permeable to vapors.

What'd be wrong with glass tubing?

 

[Compositepro]

All of the rope lights I've seen are made from flexible PVC and not silicone. PVC is made flexible with plasticizers, which are basically high boiling solvents. A cleaning solvent may have been used on some, which would also get partly absorbed, or perhaps toluene was used to clean the extruder in the factory.

The plastic safety sleeves made for fluorescent light tubes are polycarbonate, so that seems a good choice (

http://www.thermoplasticprocesses.com/products-arm-a-lite.html).

You must not clamp both ends of the tubes. The plastic tubes must be allowed to expand and contract freely. It is also possible that

 

[gonkin]

many thanks for all of your replies!


rickfischer51

Because the tubes are >3m they were manufactured specifically for this project - so they are definitely from the same batch. We are communicating with the supplier (dotmar.com.au) about possible issues in the manufacturing process.

The LEDstrips are all from the same manufacturer. We removed them from their ziplock packaging to test them about a month before they were installed. When we repackaged them we weren't especially careful about resealing the packaging, so I suspect some strips had more opportunity to outgas - leading to some inconsistency in the strips in relation to fumes.

I've written some more about outgassing below in response to btrueblood

A quick search for UV resistant clear PVC tubing came up with

http://www.clearpvcpipe.com/uvresistantclearpvcpiping.aspx

it seems finding this in 5m lengths might be an issue, I'll follow this up with our supplier.


IRstuff

the dropbox links are to the same images. It is surprising the affected tubes didn't crumble, although the cracking wasn't consistent along the entire length of the tubes.

My impression is that the silicone tubing is not especially permeable to the solvent. I have some strips that have been unpackaged for 4 months while I have been writing and testing the software. While they smell much less than when first unwrapped, they still are noticeable when I first walk into my studio. I guess it could be that the cured outer layers of the epoxy resin is slowing down the evaporation of solvent more internal to the resin???


btrueblood

re Outgassing: We plan to investigate this by baking in the LED strips - running them at maximum brightness (they get quite warm) for several days. We can then retest for cracking. The tubes that did crack did not seem to get worse after the first 24 hours (so perhaps they were then sufficiently out gassed). My concern is whether any residual traces of the solvent are likely to cause cracking in the long term (eg 10 years) that is not apparent in shorter testing over a week or so?

re: glass tubing. Is it possible to source 5m lengths of 40mm diameter toughened glass?


Compositepro

The strips are covered in a silicone tube/sheath. (similar to

http://www.ebay.com.au/itm/200m-10m...80?pt=LH_DefaultDomain_0&hash=item487413a620)

The support structure has been designed to allow for thermal expansion and contraction.

 

[btrueblood]

IRstuff, is it because when stored (low temperature) the solvent is less volatile? Or is it because the "silicone" is actually PVC as CP suggests?

gonkin, at one time I was able to find 20' sticks of pyrex tubing in large diameters (2 or 3", can't remember), you could contact Dow Corning and see what they say. It's also possible to butt weld two sticks of pyrex. I'm not gonna say it would be cost effective...but haven't looked either.

Your concern about short term tests possibly not correlating to longer term results is exactly why I described it as a science experiment.

 

[IRstuff]

PVC is susceptible to toluene-like solvents, so it the tubes got crazed, the a PVC sleeve would have been even more damaged.

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529


Of course I can. I can do anything. I can do absolutely anything. I'm an expert!
There is a homework forum hosted by engineering.com:

http://www.engineering.com/AskForum/aff/32.aspx

 

[btrueblood]

Irstuff,

We have a sight gage on the side of a large stainless steel tank, which we fabricated using "clear" pvc pipe and "solvent weld" fittings. The pvc gets gooey/soft under the action of the solvent (a mix of toluene and ketones), but no crazing/cracking is visible in the clear tube, even at the weld joints. Not sure what the difference is, but I think pvc pipe is made intentionally with some added plasiticizers, which reduce the tendency to craze.

 

[IRstuff]

Fair enought, BT. I certainly won't claim expertise in this sort of thing.

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529


Of course I can. I can do anything. I can do absolutely anything. I'm an expert!
There is a homework forum hosted by engineering.com:

http://www.engineering.com/AskForum/aff/32.aspx

 

[Pud]

Has anybody noticed the patterning of the cracking? Predominantly circumferential top and bottom, axially half way round.

Odd?

No, I don't know either, unless its the stress due it's beam function. I wonder if there's less cracking at the ends/near the supports?

Could be processing - cold die? (Or anything else that leaves residual stress). I know extruded sheet is higher impact than moulded parts due it it being stress-free.

Otherwise it's an excellent example of ESC!

H

www.tynevalleyplastics.co.uk

It's ok to soar like an eagle, but weasels don't get sucked into jet engines.

 

[Compositepro]

I agree that stored stresses in the tubes from the manufacturing process is probably a major factor. Annealing the tubes by warming (200F?) would relieve stored stresses. This is something the manufacturer should do. It may be that the cracked tubes were from the start of the run, before everything warmed-up and stabilized.

 

[btrueblood]

The neat thing about transparent plastics is that you can use photoelasticity to see residual stresses, and the effect of any attempts to anneal them.

http://www.ptonline.com/articles/stress-diagnose-it-before-it-ruins-your-parts

 

[rickfischer51]

PUD: The cracks look more plentiful on the top. With gravity loading, that would be in compression. Cracks generally don't initiate in a compressive stress field. In fact a compressive stress field would have lower free volume than the tensile field on the bottom. Solvent action is function of free volume, so it makes sense they would start on the bottom. Also, when cracks propagate, the tend to branch out, not combine. This would also explain

It would fun to section a crack and look at it under a microscope. Gonkin, want to send me a sample?

Rick Fischer
Principal Engineer
Argonne National Laboratory

 

[Pud]

rickfischer: Indeed, more cracks on top. Hence my "dunno" comment.

Food for thought: The patterning is reminiscent of the way tempered glass breaks into almost regular shapes (old type car windscreens). That is made by chilling the heated glass either side with cold air to pre-stress the surfaces. Perhaps something similar with PC?

Cheers

Harry



www.tynevalleyplastics.co.uk

It's ok to soar like an eagle, but weasels don't get sucked into jet engines.

 

[Compositepro]

Yes, after exiting the extrusion die the tubes would be cooled from the outside. After full cooling, the outside surface would be in compression and the inside in tension. This explains why the cracks do not go through the full wall thickness. It seems more likely that the cracking would occur to due to the tubes getting cold (cold night?) rather than the warmth of the LED's.

 

[gonkin]

hi again

The sculpture is being assembled in Adelaide (South Australia) where it will be installed whereas I live in Sydney. So unfortunately it is not so easy for me to send samples (but perhaps eventually). I'm embedding some different pictures of the cracking below that confirm some of the points you have made.

As Pud points out, there is a very specific pattern to the cracking. While the tubes will eventually be vertical they are being tested while horizontal. As rickfischer points out, this gravity loading would cause compression at the top and expansion at the bottom. The cracking at the top consists of parallel lines running across the tube, whereas at the bottom it seems much more randomly orientated.

The tubes are being supported at either end and by a collar in the centre. The cracking tends to be more focussed near these support points. I thought that this might be because they are slightly higher (the fumes rise???), but it makes sense that the tensions are higher there, and in fact in the third image below you can see that the density of the parallel cracks increases towards the end of the tube.

As Compositepro points out the cracks do seem to be internal to the tube. You can't feel them on the outside surface of the tubes. Adelaide in late December (Summer) (when this happened) doesn't get especially cold at night, maybe 10 - 15 celsius. The tubes were warm after running overnight, so I guess there was something of a temperature differential across them, but I don't think it would be extreme. Compositepro's comments about variation in annealing across the batch because of the warming up of the manufacturing process sounds very interesting.

The cracking only really occurred in the first 24 hours. The tubes were run continuously for a further 7 days but they didn't seem to get noticably worse.

We left a tube in the sun for a few days (so it & the LED strips were being heated by the sun). It contained the LED strips but was only sealed at one end. It was horizontal, but much more evenly supported (ie much less tension due to gravity). It developed some more minor cracks towards the closed end of the tube but none towards the open end of the tube where presumably the fumes could escape. This would seem to imply that there is residual tension from the manufacturing process?

thanks again for your continuing ruminations

 

[chicopee]

One thing that I know about Xylene on transparent polycarbonate surface is that it will turn the polycarbonate opaque.

 

[chicopee]

Polycarbonate such as Lexan is a pretty tough plastic. Some public schools that I know of, have used polycarbonate as window pane replacement which students could not break, however, students could still damage these panes by burning holes in them with cigarette lighters. I would have a testing lab do research with different chemicals to try to duplicate and isolate the problem. UV should not be a cause. Also contact Lexan manufactuers such as GE for their input.

 

[D2S]

Looks like classic solvent exposure in PC. The silicone outgasses alcohols and possibly other solvents that could be the source or the adhesive tape under the light strips could be the culprit if there's adhesive tape in there.

Only amorphous thermoplastics are clear. That leaves you with PMMA, PS, PC, PEI PSU, PES, PPSU and Amorphous nylon (Grilamid TR90 for instance). Grilamid TR90 is very chemical resistant for an amorphous thermoplastic. It's used in the eyewear industry for sunglass frames (resistant to sunscreen, sebum, DEET, etc.) If you can't burn off the volatiles, try Grilamid TR (the TR90 is an injection molding grade so an extrusion grade would be more appropriate). Make sure it's UV protected and seek help on long term UV exposure issues as I don't have a good handle on those.