About this document
This Technical Note explains how acoustic excitation changes ion supply inside high-aspect-ratio through-glass vias during copper electroplating, and how the apparatus, operating conditions and evaluation method are set up. The full report is available in Korean. An English summary is provided on this page.
Download
An English edition is not available yet. The Korean document is below.
Summary
At the center of a high-aspect-ratio via, ion supply depends on diffusion, so the entrance closes first and voids remain inside. Sound launched along the via axis creates an axial response, and the time-averaged flow it drives in the wall boundary layer exchanges the solution inside the via. The note covers how the operating band is selected, an apparatus built around open anodes, the operating sequence and the evaluation items for comparison tests.
Conclusion
At the center of a high-aspect-ratio via, ion supply relies on diffusion. Surface agitation sweeps past the entrance without renewing the solution inside, so the entrance overgrows, the interior stays thin, and the entrance closes before the via is filled. The document treats prevention of this early entrance closure, rather than thickness variation itself, as the key effect, and uses the internal fill ratio at the moment of closure as the evaluation metric. With the transducer face parallel to the substrate, sound enters along the via axis and an axial response is established. In the first mode the via center is a pressure antinode and a velocity node, so high acoustic pressure is not read directly as the fastest flow or plating. Solution exchange comes from a time-averaged flow generated secondarily in the wall boundary layer, and whether the exchange front reaches the via center is the criterion for operating conditions. In the literature, the center plating rate in a PCB through-hole via about 175 μm in diameter with an aspect ratio of about 5.7 was about 3.1 μm/h with continuous excitation and about 0.4 μm/h without excitation (literature values). For the reference geometry (100 μm diameter, 1,000 μm length), a one-dimensional open-pipe calculation places the first axial response near 0.69 MHz, and a ±10% band around it serves as the initial search band. When the bore narrows uniformly from 100 μm to 15 μm during plating, the calculated frequency shift is only about 7%; large shifts occur when boundary conditions change, for example when a central bridge forms. Because the calculated curve rests on design assumptions, the operating band is fixed by measured acoustic-pressure response and cross-section evaluation of internal growth and entrance opening. Literature results in which the same 1 MHz either promoted or suppressed plating, depending on via conditions, support this procedure. The baseline apparatus holds a glass substrate vertically, keeps both faces in contact with the plating solution and excites from one side. Because the anodes sit between the transducer and the substrate, they are open so that sound can pass, with rib thickness set at half the acoustic wavelength in the electrode material. The baseline layout is judged by whether the acoustic, flow and current paths overlap at the via entrance. Double-sided excitation, multiple transducer modules, nanobubble coupling and large-panel configurations are added as extra components while the initial installation stays in place. The comparison test is designed to evaluate five conditions with the same items: no excitation, conventional agitation, out-of-band excitation, fixed in-band excitation and growth-linked operation. Any reduction in plating time is attributed to three causes: better internal supply, prevention of early entrance closure and a higher usable current. Fill times in the document are conversions that assume a constant wall growth rate, not predictions. The original marks Cools' values as design conditions and calculated values, not measurements.
Topics
- Plating
- Glass substrates
