Instant adhesives based on ethyl 2-cyanoacrylate produce the strongest bonds with acetal (co)polymers. Henkel offers a dual adhesive system I was very successful with, that is Loctite 401 Prism with Loctite 770 Prism Primer. It worked perfectly with bonding acetals to many plastic substrates. I haven't tried it out to bond POM to rubbers, though. Surface roughening with fine sandpaper will also increase slightly the bond strength.
POM is one of difficulty bonded materials due to its low surface energy. That is why you should first roughen the substrate by sand paper. To ensure better result flame (or low-temperature) plasma treatment is recommended too. RELTEC LLC could offer adhesives (“BONDIT” type) which that will be suitable for your particular case.
Thank you Mr Sobeci for your instant answer. The rubber is DC-170 (polydimethylsiloxane). Unfortunately, now I can use only DC-170 and araldite.
Roughening I tried but there was no improvement – it was the same as before the use of sandpaper. I tried to warm the samples at c.a. 80ᵒ-85ᵒC, different times of warming and additionally pressed with something, but there was no help. Before plasma treatment I cleaned POM-C samples with water, benzene, benzene and alcohol in ultrasound bath but there was no improvement. After using sandpaper I clean the sample with compressed air and then I put the sample in the ultrasound bath. I tried also samples without washing in US bath. But up to know I did not move from the beginning.
If your bonding agent is not suitable for the system, any surface treatment techniques will not improve significantly the joint strength. Which type of Araldite adhesive have you used: cyanocrylate, acrylic, epoxy, polyurethane, or so? Is your PDMS used in neat form or filled with silica?
Mr. Sobeci, Araldite is epoxy, PDMS is not filled with silica and only quartz and some pigment were added.
Mr. Dikov after sample cleaning (on several ways described in my previous message) I treated already roughened and not roughened samples with Ar, Ar+O2 plasma and APP. But there was no success with bonding.