Hi, harikrishnan,

If you have time to do a small study about how the pdms surface cures over time, you could use an AFM and monitor insitu the visco elasticity of the surface of the PDMS being cured. Alongwith this simultaneously observe by physically tampering the pdms surface (In another sample) how much stickiness is present as curing proceeds, as cited in this paper (http://pubs.rsc.org/en/Content/ArticleLanding/2014/RA/C3RA45578E#!divAbstract) This way you can independently calibrate the curing process and use it for you application of bonding two pdms together. You can try this with some other non invasive/invasive simple method which will give you the same results. The entire process is pretty simple and takes the time for which the pdms is being cured.

Regards 

For creating my microfluidic PDMS chips I use a similar method. I have one channel already completely cured. The other chip half consists of ITO electrodes with a 3um layer of PDMS spun over them. This half we put in an oven for just 20 minutes at 65 C, which ensures the layer is partially cured, but still liquid/sticky on top. So, my thoughts would be for you to use lower temperature and curing time, and just check at which time the PDMS stops being sticky by pressing your finger on it at several time intervals.

Hi Harikrishnan,

The following things you might need to check:

1. What kind of PDMS did you use? It is better to use RTV615 for two layer devices.

2. For the thinner layer, just several-um thick, the curing time need reduce to 20 min to reduce the stiffening of the materials.

Thanks. I was using Sylgard 184. I believe the protocol is defined for RTV.

I too noticed it works better for RTV. With Sylgard 184 you could use other methods like stamp and stick or partial cure (assemble with one layer having a bit of tack, off ratio not needed) if you want it to withstand high pressures (like the valve actuation).

Hello,

I am also working with multi-layer PDMS bonding. I found this discussion very useful. I have 2 more questions. Any suggestion on these?. 

1) For Sylgard 184 what is a good ratio for the thick and thin layer - 10:1 and 20:1 respectively? 

I tried 5:1 (base:curing agent) @ 80 C for 1 hour. But the PDMS repeatedly got stuck to the wafer inspite of having a fully cross linked SU-8 base layer. Can the PDMS cross-liker chemically bind to SU-8 ? 

2) Are you able to punch neat holes on a partially cured 10:1 PDMS layer (thicker one- 2-4 mm thick) ? (Curing conditions - 80 C for 1-2 hours)

I have tried curing in a hotplate and oven. Though the PDMS is rigid enough to peel off the silicon wafer, the holes are not neat at all.

Thanks!

Multilayer based on off-ratio doesn't work well for me (and two other people here) in Sylgard so I have no idea. Sticking to a mold and your hole quality is a partially separate problem. But yes, messing with the suggested formula makes it stiffer/harder or stickier/softer and is making your process more sensitive to: feature density, cure temp/time, mold age, passivation method, hole punch size, hole punch age...

Thank you for the reply. 

For a multilayer (2 layers) PDMS device with the same base:cross-linker ratio of 10:1, do you get good bonding between the two layers using the partial curing method?

Yes it's very good:

http://iopscience.iop.org/article/10.1088/0960-1317/18/6/067001/meta

Thank you for the reply and the paper. Partial curing is the best method according to the paper

I have tried this recipe before. But I assumed that I din't get good bonding between the PDMS layers. 

-Thick layer- 2-4 mm - PDMS - 10:1 - fully cured overnight @ 60 C

-Thin layer- about 50 um - PDMS - 10:1 partially cured @ 60 C for 55 minutes

-2 layers aligned and bonded and left @ 60C overnight for full curing of the interface

-Device (with both the layers bonded) peeled off the wafer 

-Device bonded to glass using oxygen plasma and left at 100 C for 1 day

When I tried to peel off the PDMS from the glass slide to test the bonding, the bottom thin PDMS layer remained stuck to glass and the top thick PDMS layer came off pretty easily, indicating the poor bonding between PDMS layers, but good bonding between PDMS and glass.

Do you have any suggestion on improving the process? Thank you!

I have tried different temperatures with the same ratio, but I do not see any bond strength improvement. 

The partially cured layer should be tacky but not wet (a flat pair of tweezers touching the surface then removed will draw a string that snaps back without wetting the tweezers). It makes more sense to vary time and not temperature. You currently may not have a good sense of how the mechanical properties of PDMS changes as it cures over time. EDIT: It is likely that the curing time of thin layer is too long. It is closer to 20 minutes, and you will have to test tackiness around there.

I have uploaded the protocol which worked for us (based on RTV 615). We tried working with Sylgard but rather than reinventing the wheel settled for RTV. The protocol was originally developed for RTV and is being used commercially by Fludigm (C1).

Method Protocol for making multilayer PDMS devices (multi layer sof...

Thank you Philip Thomas for your help! 

Thank you Harikrishnan Jayamohan for the protocol. This would be a good starting point once we try RTV 615!

Thank you Harikrishnan. I only could get it to work with RTV615 too, but it is funny if your whole lab buys Sylgard 184, it is tough to order RTV just for one student! :)

 How do you test the bond strength and quality between thick and thin layer of PDMS? 

1) After bonding the thick to the thin layer on the wafer and peeling the assembly off the wafer, do you check if it comes out as one single piece instead of seeing the thick layer come off by itself without the thin layer?

OR

2) Do you apply pneumatic pressure in the channel and see if the two layers de-laminate at a certain pressure when they are not supposed to do so?

OR

3) Do you bond the 2-layer device to glass with oxygen plasma, try to peel off the device after a while and see if the device comes out as 1 piece and does not leave back the thin layer on to the glass?

You should be able to do 1 after curing. I've attached a video from Stanford microfluidics foundry where they do the same thing.

https://www.youtube.com/watch?v=xWdRczefirs