increase in carbon/epoxy modulus with hot/wet? 2

[RPstress]

We have some data for twill weave HS carbon in epoxy which shows an increase in compression compression modulus for hot/wet tests.

Over RT/dry in the 2-direction this is about 15% for 90°C/wet (13% for 70°C/wet). In the 1-direction it is an increase of 9.5% at 90°C/wet (9% for 70°C/wet). There is also a small increase of 1.1% for 70°C/dry in the 1-direction.

It is an average of several tests, probably between six and ten (not sure but this was qualification testing).

I would expect compression modulus to go down for hot/wet. There are small reductions in the tensile modulus figures (except for a 5% increase for 70°C/dry).

Does anyone have any idea what could be going on?

Tabular summary:
[tt]
Test Condition E GPa % change
Tensile 0° RT/dry 61.37
70°C/dry 64.37 4.9
70°C/wet 61.63 0.4
90°C/wet 59.29 -3.4
Tensile 90° RT/dry 59.5
70°C/wet 56.07 -5.8
90°C/wet 58.4 -1.8

Compression 0° RT/dry 52.3
70°C/dry 52.93 1.2
70°C/wet 56.85 8.7
90°C/wet 57.29 9.5
Compression 90° RT/dry 48.46
70°C/wet 56.85 17.3
90°C/wet 57.29 18.2
[/tt]

 

[SWComposites]

RP -

how is strain being measured? strain gages? extensometers? gage length?

what test method / specimen is being used?

how is modulus being calculated? secant between two defined strain values? best fit?

what is the cure temp?

SW

 

[RPstress]

The testing was carried out by a reasonably reputable resin manufacturer for an important third party, so I have limited info their test practices but they are quite likely to be good. It is just possible that a fourth party (a qualified test house) actually did the testing.

This was three-batch testing so the total number of specimens was almost certainly between 18 and 30.

They used EN2597 for tension and EN2850B for compression. Those specs don't specify the strain measurement technique but give a minimum accuracy and require strains from both side of the specimens, and they suggest strain gauges as the device to be used. Strains used for modulus are suggested as between 0.1 and 0.5%. Specimens are 10 mm wide.

Cure temp was 180°C in autoclave, probably 100 psi pressure but not certain.

 

[SWComposites]

Hmm, very strange. Makes no sense for compression modulus to increase significantly at hot/wet.

Where the compression modulus values from separate untabbed specimens? not the same ones as for compression strength (which need tabs)?

What is the cell size for the fabric weave? need to be sure the strain gage or extensometer length spans several weave cells.

Can you get the load-strain data? check for misalignment, buckling, etc. What is the compression strain to failure at each environment? is it close to the 0.5% strain value for the upper end of the modulus value?

I have seen many "reputable" labs bugger up modulus measurements, so I would treat the data as guilty of being wrong until absolutely proven to be correct.

SW

 

[RPstress]

Tabs optional in test spec., can't tell if they used 'em.

Re cloth cell-size, one hopes they're not daft. Here're the UD material results:
[tt]
Test Condition E GPa % change
Tensile 0° RT/dry 140.58
70°C/dry 148.87 5.9
70°C/wet 131.56 -6.4
90°C/wet 145.43 3.4
Tensile 90° RT/dry 7.76
70°C/wet 7.53 -3.0
90°C/wet 6.79 -12.5
Compression 0° RT/dry 111.10
70°C/dry 109.09 -1.8
70°C/wet 117.11 5.4
90°C/wet 117.00 5.3
Compression 90° RT/dry 7.89
70°C/wet 8.18 3.7
90°C/wet 8.21 4.1
[/tt]
So less tendency to show increased stiffnesses but still the possibility of an effect, although low percentages aren't very convincing.

The test work was mostly done in late 2006 and no other results (like the deflections or strains) are available to us.

The test results included strength and those generally show a sensible reduction with hot/wet. The strengths were for a number of loadings including bearing, interlaminar shear, open and filled hole, etc., and included a variety of laminates. (A 50/40/10 layup had a 2% increase for 70°C/dry in tension, but that's the only slightly anomolous increase.)

The moduli were only found for the basic material properties. The in-plane shear modulus was also found from a ±45° tension test and show a sensible drop with hot and/or wet (except RT/wet which was just 4% up on RT/dry).

Someone had an idea that maybe there was some sort of 'prestressing' by the fibre-resin stresses increasing but it sounded a very unlikely longshot.

Thanks for your thoughts SW.

 

[Moving2012]

Hi,

I think it maybe due to the resin itself. The wet condition doese weaked the resin/fiber interface that's why you get lower stress and IPS number. But it may not soft your resin matrix so you didn't have a drop on modulus. Or the Wet Tg of the resin is much higher than your test temperature so the resin doesn't soft. I would like suggest you to look neat resin after condition. Try to get more understanding from your resin supplier.

 

[RPstress]

Thanks for the throught Moving2012. The mechanism you suggest does explain why the modulus might not go down. The environment affecting the fibre/resin interface more than the bulk resin properties is not something we considered. However, that doesn't appear to explain why the modulus actually went up!

 

[mcclain]

Could there have been some post curing during the wet conditioning?

 

[RPstress]

In theory no, mcclain, though thanks for the thought. I believe that conditioning is always done at less than 100°C, even if it's a crude X-day water boil (well, I guess if the weather's high pressure it could go a little bit over). Although someone might know better? This material should have had a cure of at least 120° and a post cure of 180°C, or an initial cure of 180.

 

[berkshire]

Rp,
How old were the samples? As Mcclain says was there any post curing.
Post curing can be time dependent as well as temperature.
The resin might have been cured at 180* ,but if it was not held long enough, you might not have had full crosslinking.
B.E.

The good engineer does not need to memorize every formula; he just needs to know where he can find them when he needs them. Old professor

 

[RPstress]

berkshire: The testing was done by the resin's manufacturer for an important client (a very big end user of the material) and for an important purpose (material qualification testing) so it is very unlikely that they didn't cure it properly. I would expect the resin to be at least 98% cured if not 100%. However, I have no DSC results to prove the resin's condition. mcclain's suggestion is the first one that might account for the increase. However, the strengths did not go up with the conditioning. It is possible that some big one-off error was made in manufacture of the cloth modulus specimens. I will suggest this.

 

[TheBigM]

If they've used strain gauges for the determination of modulus the adhesive may have softened under heat and moisture allowing the gauge to creep more slowly than the actual sample itself.

There's always the possibility that they've used an extensometer at RT (to save money on gauges) and strain gauges hot/wet which may provide a difference.

M

 

[RPstress]

It would be a little odd for the resin supplier to select a strain gauge adhesive with inadequate resistance to a known test environment, but this would again explain the modulus discrepancy. It would also cause the strains/stresses for the strength part of the test to be too low, which while it is explicable is a possibility. As for effectively using a different standard between RT and HW to save money it's not unknown. Thanks for the suggestions.