Issue link: https://iconnect007.uberflip.com/i/712650
August 2016 • The PCB Magazine 91 provide subsequent adhesion. A list of tested dispersant/solvent combinations is shown in Table 1. Five of these provide sufficient disper- sion stability to allow further coating and adhe- sion testing. These inks were Mayer bar coated, vacuum dried, and flash sintered into 6x9 inch samples in the lab and the resulting uniform copper samples were sent to a local printed cir- cuit house for preparation of peel strength cou- pons by masking, plating and etching. An as- sessment of the adhesion of these materials is shown in Table 2. The maximum peel strengths of 2−3 peel test coupons on the plated (to ~30 µm) and etched films are recorded and listed in Table 2. Two samples showed very poor adhesion of the final film; zero peel strength is assigned as the traces fell off. For the other three samples, the numbers are quite variable, indicating non-uni- form adhesion on the 6x9 inch coatings. The non-uniformity can be attributed to the sinter- ing by overlapping repetitive flashes from a sin- gle bulb, resulting in less particle necking in the overlap area, as mentioned previously. Another possible reason is the contamina- tion on the polyimide surface (dust, grease, etc.) which largely affects the surface energy and therefore the thin film formation during ink drying. Interestingly, the last two inks on the table, that both employed polymethacrylate copolymer as dispersant/binder, showed very different adhesion results as different solvents were used. This suggests that the configuration of the polymeric dispersant/binder in the ink is possibly important to the adhesion. Overall, it is believed that we can further improve the adhesion by testing more binders that have similar chemical structures and properties with polyvinylpyrrolidone, hyperdispersant A and polymethacrylate copolymer. It has been reported in the literature that the adhesion between copper and polyimide can be improved by applying a thin layer of chro- mium[3,4]. It has been attempted to sputter a thin (<100 nm) layer of metal like Cr, Ti, or Cr/ Cu as a tie layer on plasma-treated polyimide substrates (20% O 2 /Ar, 3 min at 15 mtorr), upon which to coat the ink. We found that there was a great deal of difficulty sintering these samples with a metal layer underneath, while no sig- nificant difference was observed on the plasma- treated (control) sample, as shown in Figure 8. The coated copper developed holes on sinter- ing showing the sputtered metal underneath. The unexpected loss of adhesion might come from the nature of the deposited copper on the metal interlayer. In the literature, the metal layer promotes the adhesion of electron-beam evaporated[4] or sputtered[3] pure copper, but the copper nano ink includes dispersants/bind- ers that may interfere with the interface. Since these sputtered samples had very poor adhe- sion at this point, none were masked, plated, and etched. FACTORS AFFECTING THE ADHESION OF THIN FILM COPPER ON POLYIMIDE Table 1: A list of tested dispersants and solvents. Table 2: Maximum peel strength of 1 mm wide peel test coupons on the plated and etched cop- per nano ink coatings (N=2-3).