EIPC Technical Snapshot: Market Analysis and Advanced Manufacturing Tech
Only John Ling could have composed the invitation: “An EIPC webinar is just like one of their conferences, except that you do not have to travel, you do not have a beer with colleagues, you do not enjoy excellent food, and you do not enjoy convivial company. But we do not live in normal times, and some things are not the same. Manufacturing PCBs, however, remains comfortingly complex, so on November 18, 2020, at 1500 hours CEST, EIPC will hold a webinar at which three very reassuring gentlemen will help matters become a little clearer.”
In a well-attended Zoom meeting, EIPC president Alun Morgan introduced an outstanding webinar programme combining knowledgeable market analysis and advanced manufacturing technology. Thanks to Zoom, we had the benefit of a remarkably detailed inside-line view from Asia, as Prismark Managing Partner Dr. Shiuh-Kao Chiang reflected on a market that contrasted radically with original predictions and discussed key factors affecting its future direction and momentum.
2020 would certainly be a year to remember! The global electronics and PCB industries were acutely traumatised by the pandemic: PCB production patterns and demand expectations were substantially distorted. Originally a very strong year had been anticipated, when 5G was expected to create tremendous demand in smartphones and base stations, with a consequential increase in demand for high-end circuit boards. The pandemic forced a revision of the forecasts, and at the beginning of the year, a drop of 5–6% in the global economy had been predicted, with historical experience suggesting that the PCB industry would suffer as badly as it had in 2008, when a double-digit decline was followed by a global recession with a huge impact on the electronics and PCB industries.
But it did not turn out as badly as feared. The effects of work-from-home and the change in behaviour patterns effectively saved the electronics industry, with people continuing to purchase durable goods. And there remained a robust demand for semiconductors, resulting in the PCB industry seeing substantial growth in the demand for packaging substrates. Even though there had been a disruption of production and logistics and an evaporation of the demand for many consumer products, the effect on the PCB industry was limited in terms of the total value, but it was not uniformly distributed. Some sectors saw tremendous impact while others had been relatively undamaged.
Because of work-from-home, the computer market had been very strong, but the lock-in caused automotive demand to drop sharply. The markets for conventional consumer products and mobile phones were weak, but advanced consumer products—like wireless headphones, new gaming consoles, and smart wearables—were resilient. The analysts were predicting that the PCB market would see an overall growth in value of about 4.4% in 2020, with the majority coming from packaging substrates. The HDI market was also relatively good, although no growth was expected in conventional multilayers or rigid boards.
2020 could actually be seen as a positive, with the computer side as the strongest driver, which came as a surprise since the market had been in decline for the previous ten years, both desktop and notebook. As a consequence of work-from-home, it was no longer a case of one computer per home but one per person, and similar for tablets. A significant demand had been created for 4-, 6- and 8-layer boards, together with substrates for CPUs, GPUs, and AI devices. The server and data-centre market also continued to increase.
On the negative side was the huge drop in automotive. The medical sector had been steady, and some very strong growth had been seen in military, particularly in the U.S. The mobile phone market was interesting. Although the volume was forecast to decline by about 8%, the HDI, flex, and substrate value within each phone was increasing because 5G and high frequency required completely new RF modules and increased interconnection density.
For 2021 and beyond, the industry would continue in a transforming, recovering, and growing mode, with positive expectations for the PCB market. Substrates would continue to become more sophisticated, and leading substrate manufacturers had made billion-dollar investments. Demand for HDI would continue to grow, not just for the 5G smartphone but also because it was becoming the popular technology for many other applications, including consumer products, and local Chinese PCB shops were putting a lot of investment into HDI. High-speed, high-frequency applications were driving the demand for rigid multilayers in low-loss materials.
The pandemic had renewed the debate about the benefits of localisation versus globalisation for the security of supply and provided a stark reminder that we will never know what’s going to happen five years in the future!
Mike Vinson, chief operations officer at Averatek, described a novel semi-additive process based on a proprietary palladium surface activator for electroless metallisation and capable of producing extremely fine lines and spaces with minimal loss of conductor cross-section on flash-etching, offering significant benefits in signal integrity and impedance control. The process was currently capable of producing lines and spaces of less than fifteen microns, and sub-five-micron capability was in prospect. Ultimate resolution was limited only by photolithographic considerations.
The key feature was a non-aqueous surface coating treatment yielding a dense, coherent deposit of palladium, only a few nanometres thick. These characteristics offered significant advantages over traditional palladium catalysts in terms of coverage and adhesion, and enabled continuous metallisation to be achieved at minimal deposit thickness. The primary function was in providing an ultra-thin base layer of electroless copper on laminate for semi-additive manufacture of printed circuits on rigid or flexible substrates, although the treatment could also be used to metallise a variety of cavities, channels, and three-dimensional surfaces, as well as to facilitate the plating of copper on aluminium, ceramics, and textile materials.
The metallised coating could be imaged using photo-lithography and other printing techniques. Alternatively, it could be directly patterned using a laser or by selective blocking. After pattern plating and resist stripping, there was very little copper to remove at the flash-etching stage; therefore, the conductor cross-section was accurately maintained, enabling close control of impedance and allowing precise inductive and capacitive coupling across narrow spaces.
It was claimed that the process could dramatically reduce the area, layer count, and weight of electronics systems, as well as provide significant RF benefits. It had been designed to be easily integrated with traditional PCB manufacturing equipment and materials, and the coating retained its adhesion and activity through multiple processing steps. It was well suited to improving microvia reliability, and high-density copper traces could be used to reduce or eliminate microvia issues. It was successful in plated through-holes and had been used to metallise silicon vias with diameters as small as 40 nanometres.
Averatek was working closely with major laminate suppliers to determine compatibility with a wide range of rigid and flexible materials: standard, mid-loss, and low-loss. Vinson showed composite coupons for high-speed passive circuit testing and preliminary RF analysis of microstrips, differential pairs, and resonance structures. Averatek’s process had been licensed to multiple commercial manufacturers.
The rapid implementation of HDI technology has focused increased attention on the thermo-mechanical reliability of stacked microvias. Factors like base material properties, lamination and drilling parameters, de-smearing, and electroless copper activation had been extensively studied. Research by Atotech indicated that the detailed formulation of the copper electroplating process significantly influenced the nature of the metal deposited on and near the surface and that subtle differences in the co-deposition of additives played an important role in determining the integrity of the interface between stacked micro-vias. Roland Herold, Atotech senior specialist in electroplating, gave an overview of the latest findings and developments.
How could the bond be improved? His Ishikawa diagram was indeed “a fish with many bones,” but he chose to focus on the functions and purposes of essential additives: halides, brighteners, levellers, and carriers, and the extent to which these could be preferentially co-deposited at or near the interface. Clearly, the surface of the electroless copper at the target pad needed to be perfectly clean to enable the best metallic bond to be achieved by undisturbed epitaxial copper crystallization, but even when the prior processes were perfectly under control, it was possible for components of the electroplating process itself to effectively contaminate the target pad surface and thus weaken the metallic bond.
Herold described the chemistry of organic additives, which typically contained sulphur and nitrogen in combination with carbon and oxygen and were normally associated with chlorine, and discussed methods by which they could be detected and measured. Dynamic secondary ion mass spectrometry (dSIMS) had been successfully used as an analytical technique for investigating failed interfaces. He gave a detailed account of electrodeposition mechanisms to explain the interaction of brighteners, carriers, and levellers and their influence on the initiation and growth of copper crystal structure. A fascinating lesson in theoretical electrochemistry!
Investigational work at Atotech had demonstrated that leveller compounds could be co-deposited onto interfaces and into the electrodeposited copper deposit, synergistically with the brightener, and the rate of co-deposition of additives was highly dependent on their chemistry. The co-deposition of additives could be controlled to a low level by careful choice and balance of chemical composition, together with optimisation of operating temperature, current density, and reverse pulse frequency, as well as the configuration and maintenance of the anode system.
After a question and answer session, Alun Morgan thanked the presenters, the organisers, and the participants, and brought EIPC’s very successful and extremely informative second technical snapshot to a close. The next one in the series is planned for December 16.