Issue link: https://iconnect007.uberflip.com/i/1533339
MARCH 2025 I PCB007 MAGAZINE 73 conductors compared to bare Cu over 0–50 GHz range 3, 4, 5 . 5G cellular networks, which are currently in use worldwide, are using millime- ter frequency bands at the higher end of that range; for example, Verizon (U.S.) is using a 29 GHz band, and AT&T (U.S.) is using a 39 GHz band 6 . Higher GHz bands are already being dis- cussed for cell networks between 50–100 GHz due to the higher throughput of data avail- able at higher frequencies 7 , and automotive radar already uses a 76-81 GHz band 8 . Signal loss associated with the use of nickel in ENIG surface finish is already a concern at the signal speeds of existing cellular networks; that con- cern will only increase with ever-higher sig- nal speeds of future technologies. Already, the number of high-frequency PCBs as a percent- age of the overall electronics industry is 15%, and is expected to grow (Figure 2) 9 . With the growing need for circuit boards and designs for high-frequency applications, the Figure 2: High-frequency PCB market share in China 9 . Figure 3: Nickel-less surface finish that is tested in this article, which includes a nano-engineered barrier layer instead of the Ni-P layer in ENIG (layers not to scale). insertion loss due to ENIG surface finish is becoming unacceptable in the industry. In these applications, a new type of surface finish must be developed. e require- ments for a replacement for ENIG in this sphere should be: 1. No nickel, in order to remove the high insertion loss and ferro- magnetic element of ENIG surface finish. 2. Gold final finish in order to retain the high reliability and long shelf-life rating cur- rently associated with ENIG. A surface finish solution satisfying both of these criteria could reasonably replace ENIG in high-frequency applications without sacri- ficing the advantages ENIG provides in lower- frequency circuits. In this article, a proposed alternative to ENIG for high-frequency applications is discussed— a Ni-free surface finish solution in which a bar- rier layer is deposited on bare copper in place of nickel in ENIG, and gold is deposited on the barrier layer (Figure 3). ere is no nickel in this approach and therefore no adverse effects of insertion loss for high-frequency circuits, and the reliability aspect of the gold final finish is still upheld. Figure 4: ENIG surface finish, including the problematic nickel-phosphorous (Ni-P) layer (layers not to scale). Figure 5: ENIG surface finish, including the problematic nickel-phosphorous Layers (layers not to scale).