Shrinkage of flat RTM plaques...

[RPstress]

We have some flat HS carbon fibre/PR 520 epoxy RTM plaques which are coming out a bit thinner than we think they should. The tool cavity is 3.09 mm (0.122") deep at room temp and the cure is 180°C.

The final room temp plaque thickness is averaging about 3.026 mm (0.119"), so it's shrinking about 2.1% for a 160°C cool down.

The resin Tg is in advance of (higher than) the cure temp, so just from a CTE estimated at 40 microstrain/°C we'd expect to get a 0.64% shrinkage. (We have ~59%–62% fibre vol and pure resin CTE is about 53; laminate CTE might actually be nearer 25–30 if transverse fibre CTE is about 9.)

Could the additional reduction be due to resin shrinkage on cure? It seems a bit high, but there is rather limited data for this resin, so it may be possible that that's it.

Does anyone have any thoughts/experience of PR 520?

 

[Demon3]

Most resins shrink on cure so I wouldn't be surprised to find that is the cause. Isn't it going to be more pragmatic to increase the depth of your mold by a little than to spend energy speculating on the causes?


Chris DeArmitt

 

[Compositepro]

The shinkage you are measuring is what I would expect to see. Most epoxies shrink 1.5 to 2%. In a laminate that shows-up almost entirely in thickness skinkage. Most shrinkage test methods measure total shrinkage due to cross-linking and due to cooling from cure temperature. Thus high cure temperatures usually mean higher shrinkage.

In RTM surface pits and "wormholing" are common due to resin shrinkage in a closed cavity mold. Slowly ramping up cure temperature can improve part quality by compensating for resin shinkage with expansion due to increasing tempearature. High pressure helps by "inflating" the mold cavity so it can shink along with the part. This prevents cavitation or separation of the part from the mold surface.
Cavitation can cause voids and surface pits in parts.

 

[spudpeeler]

Like all other epoxy resins, PR520 does shrink during gel. With a very stiff tool this can cause cavitation but if the tool is flexible enough (most tools are) the tool will deform to allow the shrinkage to occur.

You are probably seeing more loss of thickness in the center of the panel as this is where the volume change can occur. If the tool, or a lining in the tool, can float with the part the maximum thickness change is less as it will be spread over the full tool surface.

I do not know of a cure cycle that will minimize this effect.

 

[RPstress]

Thanks all. It's annoying that it's shrinking as much as it is. Other RTM resins (e.g. Cycom 890) claim a good deal less cure shrinkage than this even for the neat resin, let alone when it's only 40% of the laminate.

The main reason to understand the cause of the shrinkage is to be able to allow for it in future molds with minimal re-cutting.

 

[CompoMan]

RPstress, as a matter of interent, you mentioned that the Tg of the cure laminate is much higher, then what is the % shrinkage estimated for from the TG to RT??

 

[RPstress]

Turns out I was talking boll*x...Tg (dry) of PR 520 is 161°C according to Cytec. Maybe this is a big factor in the shrinkage; CTE above Tg tends to be quite a bit bigger than below, though I've no data for PR 520 above its Tg. Thanks for making me check!

 

[allcomposites]

Hi RP Stress, I'm new on the forum and saw your query.
We have worked hard characterizing the PR520. It shows a chemical shrinkage of 2% during cure. The CTE is nearly linear from room temperature up to 140C then it increases 4 times (when approaching its Tg).
The 890 shows a nearly similar behavior on shrinkage and CTE.
Hope it's not too late

 

[Compositepro]

Allcomposites, I'm curious how you define (i.e., measure) "chemical shinkage". All the standard tests measure the combined shrinkage due to crosslinking after gellation and CTE in cooiling to room temp. This is about 2% for most epoxies.

 

[RPstress]

Allcomposites: it's too late for this particular mold, but it's still of great potential use for the future. Thanks.

The combination of high (and probably non-linear) CTE when heating up during the (sort of) liquid phase, gellation/chemical shrinkage during cure/cross-linking at more-or-less constant cure temperature (depends on exotherm, etc.), then lesser CTE-based shrinkage (when solid) back to RT can be a bit complicated to assess, especially when the above Tg/below Tg CTE varies and Tg is below cure T.

The manufacturers seem a bit cagey about shrinkage; I guess they think we can just cast a rectangle and measure it...however it seems a bit of a nuisance to have to do this at a cost of a couple of hundred $/€/£ when it's a pretty basic property.

The CTE of carbon across the fiber is also hard to come by; you wind up having to make a test panel or two. It all gets pretty complicated (and expensive) for simply predicting final what mold thickness you need if you have a required final laminate thickness, especially if you also have a required fibre volume range.

 

[allcomposites]

Hi Compositepro,
Since chemical shrinkage problems are quite known in industry, I have developed a laboratory instrument to measure it. I have combined a DMTA instrument with a volume controlled chamber that has thermocouples and a heat flux sensor on it (acting as a DSC). This allows the combined characterization of resin volume, pressure and degree of conversion in time. It is similar to a PVT instrument for thermoplastics.

 

[allcomposites]

RPStress,
I fully agree with your last post. Even if this issue seems to be simple, when you look deep into it, it is complex and very time consuming since you need to know the properties of each phase (liquid, gelled and cured resin). I did my PhD on this (some time ago) and expended most of the time characterizing the materials.

Note that in order to make a proper inverse analysis to define the final mold thickness, you must also consider the fibers reorientation (due to draping over a non-planar surface). Draping results in locally increased fiber volume content and thickness.

For the CTE of the laminate, you can characterize it on a TMA instrument. You measure CTE on each direction and then calculate the CTE tensor.