Very interesting! I would suggest that you may be centrifuging the pigments out. Do you see a gradient of pigment in the cross section? Look at the data sheets and see what the specific gravity of each component of the PDMS is, including the pigment. If they differ greatly, maybe you can reduce the RPM on the centrifuge to reduce the separation.

Hello Ron, thank you for your answer.

I had a closer look on the PDMS with a microscope. There are cave-like holes just below the surface of the PDMS. They are mainly closed with a very thin membrane. So depending on the angle of incoming light, the surface seems to be closed or the shadow in the caves are visible, forming the "marbeling-effect".

The caves are about 200 µm deep.

Right now I run a PDMS process with much higher speed in the centrifuge. I'm looking forward to see, if the caves disappear.

The attached images show a cut ("Cross-section_7x.jpg"), the surface on the right, the cross section on the left.

The other two images are a close ups from top-view of a cage. One image is focused to the surface, the other one to the bottom of the cave.

I don't want to cast aspersions on your centrifuging systems, but I don't think your PDMS is completely mixed. I've noticed similar phenomena when the PDMS and curing agent are not mixed completely, though I usually mix my PDMS by hand (and get rid of subsequent air bubbles in a vacuum chamber).

Yes, I also thought about the mixing first.

But the phenomena did not change after we switched to a professional mixing system (http://www.speedmixer.ch/e/home.asp). The mixture is perfect and reproducible. The phenomena disappears when I directly pour the PDMS on the mould. Also it now seems to disappear, when I use much higher centrifugal forces. But I only had one run with the higher rotation speed until now. So it's not reproducibly vanished by now. Also the density of the PDMS has changed and we have to see if that is a problem.

I think new air bubbles develop, when I pour the PDMS mix into the centrifugal system (http://www.3t-analytik.de/de/produkte/zentriform). If the rotation speed is too slow, the air is not pushed out fast enough.

Still I think the phenomena is strange. PDMS has a much higher density than air. In a centrifugal system, the air should never be trapped at plane of highest centrifugal force. For me it would be absolutely logical, if the phenomena would only appear at the back side. 

That's why I thought it might be a color-pigment problem.

In your first image, it looks like the PDMS is "cracked". Is there any difference in the PDMS film thickness between the moulded case and the centrifuged case?

Yes, it looks like it's cracked, but the surface is mainly closed. There is a cavity-system below the surface, as you can see in the last image i added ("cross-section_7x.jpg"). But it's possible, that this cavity-system develops from cracks during the polymerization process.

The direct moulded case has a thickness of 0.9 mm. In the cross section I can see air bubbles. Also on the surface are small round holes generated by air bubbles. But there is not the descirbed "marbeling" phenomenon.

When I have a look on the cast with a much higher rotation speed, the phenomenon disapears on the structured front side. Also there are no more cavities in the cross-section. But the back side has still a rough surface: (first image)

That's the top side: (second image)

Which pigment are you using? I would like to try this with some Silc Pig from Smooth-On and see if it happens with that. How many Gs on the centrifuge? The first image in your last post reminds me of the patterns generated by turbulence in settling silt in the mud pits of a drilling rig. (I could watch that stuff for hours when I was a kid.) I believe that, in the case of the silt, this results from the water moving upward as the silt moves downward creating a counter flow similarly to what happens with centrifugal elutriation of a dense suspension. When particles in the column separate to allow the fluid to pass, they have a tendency to stay separated since it requires more energy to move back together against the established flow in the gap, but once they collide, they have a tendency to stay together since this also requires additional energy to overcome adhesive forces. This causes what looks like aggregation of the sediments into tightly packed clusters with low flow between particles separated by less densely populated "channels" where there is a higher velocity of the liquid phase. Since the viscosity of PDMS is changing during the cure time, this could add some additional dynamic to the motion of elutriation which should be quite interesting to some of the folks in fluid dynamics. I certainly find it interesting. Maybe I'm just shooting the moon, but check out this paper. http://me.ucsb.edu/~meiburg/pubs/Bosse_et_al_2006.pdf Particularly, figure 6.

Hello Ron,

The datasheet by Wacker says:

Elastosil RT 607 A/B:

Component A: "reddish brown", density: 1.5g/cm^3, viscosity 15000mPa s

Component B "transparent", density: 0.97g/cm^3, viscosity 200mPa s

Wacker says that the pigment is iron oxide. I'm still waiting for the specifications from the supplier (DRAWIN) but I found a pigment-Datasheet from another supplier (http://kemper-co.de/pdf/aeznatron.pdf)  listing the density for brown iron-oxide: 4.7g/cm^3.

I am not sure if I understand your mud pit thesis right, but I think, that the pigment is not the problem meanwhile. I think the air-cavities build up during the polymerization if the rotation speed is not high enough.

The G-Value for 500rpm is 32.4 on the centrifuge (r=116mm). With the higher rotation speed (1500rpm) I get 291.8.

I have more pictures I can send you.

I agree with David. It is not well mixed. It is most likely that the PDMS monomers have a normal molecular weight distribution, though it is not detailed in its data sheet.  The centrifuging after mixing separates the monomer-catalyst mixture layer by layer, causing complications in curing as explained by Ron.    

Okay, I will check the mixing part. Probably the mixing speed is too high.

I called with the Wacker-PDMS supplier yesterday. The Elastosil 607 RT PDMS we use consists of about 50% filling material like quartz crystal and other components. They never saw a phenomena like mine, but a similar patterns