To read the first part of this article, click here.
Well-known at EIPC conferences for her fascinating bioelectronics work, Dr. Despina Moschou, from the University of Bath in the UK, gave an update on recent developments in lab-on-PCB technology for medical diagnostic applications. She began by reviewing examples of microfluidic systems designed to bring together microscopic volumes of liquids, transducers and microelectronic components to form biosensors, and the need to identify an integration platform that could be upscaled cost-effectively to commercialise the production of cheap disposable diagnostic devices. The lab-on-PCB approach had been proposed in the 1990s, but had been sidelined by easier microfluidic fabrication processes such as soft lithography and glass/polymer processing. But it was now re-emerging as a very strong candidate, particularly because of its inherent upscaling potential. The PCB industry was well established world-wide, with standardised fabrication facilities and processes, but currently commercially exploited only for electronics. It offered opportunities for low-cost upscaling of complex micro-total-analysis systems, integrating microfluidics, sensors and electronics on the same PCB platform. There were adequate microfabrication capabilities and the integration of electronics was intuitive.
Dr. Moschou showed examples of a micro polymerase chain reaction module for DNA analysis, integrated on a PCB, and PCB biosensors for glucose and lactate. A module for measuring gamma-interferon in human serum had been developed as part of the ELISA project, and a pre-diabetes diagnosis test for mass-population preventative screening of children was in development in the CHIRP project. And biosensors for next-generation sepsis point-of-care diagnosis had been successfully ink-jet printed.
Many European academic groups were working on PCB based prototypes, the microfluidics industry was keen to engage with the PCB industry, and a substantial commercial market was forecast.
A quote from an article I wrote in PCBDesign007 made a convenient preface to the presentation from Jan Pedersen, senior technical advisor at Elmatica in Norway:
"An enormous amount of information is needed to precisely and unambiguously define all of the fabrication details for a PCB and ensure that it is manufactured, tested, qualified and delivered exactly as the customer specified. The principal area of concern is not the image data that constitutes the board design—there are already well-established and well-accepted formats for conveying those instructions across the CAD/CAM interface—but all of the other bits and pieces of essential information needed to fulfil the order requirement that need to be communicated over the same interface. It costs the PCB industry substantial time and money interpreting this information when it is presented in different styles by different customers. A uniform language would save these costs and avoid misinterpretations."
Pedersen provided the solution with his presentation on CircuitData, an open source language for communicating PCB specification.
Leading PCB broker Elmatica, together with its software development partner NTTY, started this initiative to improve communication of PCB Article Specifications. CircuitData was released from Elmatica as an open source site and established as a free standing organisation. The primary objectives were to reduce engineering queries to a minimum while being able to prepare quotations without exposing intellectual property.
The basic CircuitData structure divided a specification into major groups: sections, layers (including stackup information), process functions, metrics, tolerances, logistical and configuration. Pedersen discussed each of these and indicated the web addresses where further details could be found.
A CircutData material database was being compiled, that would hold material data relevant for printed circuits, and present data from different sources in one generic structure compatible with the CircuitData language. And there was a CircuitData on-line forum where new ideas could be discussed for improvements, upgrades and new technology. Also, it was planned to integrate CircuitData into Ucamco Integr8tor and IPC 2581.
“Who can join? Everyone—a language will work only when spoken!” Anyone who was interested could contribute by joining the CircuitData Forum and helping to evolve the language by discussing their requirements, adding their special parameters and encouraging colleagues, customers and suppliers to get involved.
The final conference session was concerned with new materials and processes for PCB manufacturing, moderated by Oldrich Simek, owner of Pragoboard in the Czech Republic and vice president of EIPC.
The first presentation came from Rick Nichols, Atotech’s global product manager for surface finishing, who discussed the results of a study of the initiation speed of the palladium, within an electroless nickel, electroless palladium and immersion gold (ENEPIG) system, with regard to solder joint reliability.
He commented that the reliability of solder joints became more and more important with smaller features, and the objective of the investigations was to improve solder joint reliability using high-speed shear testing as the measurement tool. Multiple designs of experiment were carried out for this optimization, and it was observed that rinsing technology had a significant impact on the results.
An Intel test vehicle was used, with 400-micron SAC 405 solder balls on 380-micron pads assembled with customer-specified flux and reflow parameters. The assemblies were aged through one additional reflow cycle, then shear-tested at 1.8 metres per second with the shear blade set 10 microns from the surface of the PCB. 3 sets of 20 balls were tested in each run.
Initial results indicated a relationship between total shear energy and the palladium initiation stage of the ENEPIG finishing process, and this was studied further. Several methods were employed to study the palladium initiation reaction: x-ray fluorescence spectroscopy, scanning electron microscopy with energy dispersive x-ray, high-resolution transmission electron microscopy and time-of-flight secondary ion mass spectrometry. Electrochemical methods were used to measure initiation times. There was definite correlation between palladium initiation time and shear results: the later the initiation, the poorer the result.
Bill Bowerman, director of metallisation technologies with MacDermid Enthone Electronics Solutions, presented a review of the effect of metallisation interfaces on microvia reliability. “We plate metals on surfaces and join them together!” was his opening line as he discussed the consequences of miniaturisation and multi-functionalisation of electronics: complex board architecture, mixed materials, smaller features, stacking of features, greatly reduced contact areas, lower “absolute” bond strength and a higher potential for failure under “non-optimum conditions.” He remarked that control of all aspects of the manufacturing process became increasingly critical.
Microvia reliability encompassed the ability to survive reflow assembly and the ability to survive operation through useful life. Contributing factors were the number and size of vias, their aspect ratio and shape, whether they were stacked or staggered, and their position within the stack-up. Additional factors were the cleanliness of the target pad, the type and quality of the primary metallisation, and the quality of the secondary metallisation. And it was essential to understand each process and the underlying steps of each process.
Several interfaces were typically encountered in microvia processing, although it was preferred to minimise their number. Interfaces tended to be the weak links. They could occur at drilling, desmear, electroless copper or direct metallization, flash plating, via fill plating and conformal plating, and each presented an opportunity for reliability issues. Of the many factors influencing interfaces, features such as oxidation and residues, materials of construction, differences in thermal expansion coefficients and differences in grain structure were significant. Bowerman examined each one in detail and commented on how it could influence reliability. Reliability could be optimised minimising the number of interfaces and implementing, monitoring, and controlling best plating practices for each applied layer.
Andre Bodegom, managing director of Adeon, discussed the benefits of parameterised AOI in aid of traceability. He listed some of the challenges experienced by companies developing AOI systems to meet current and future industry demands. Designs often had different materials and technology levels per layer or materials of different brands with different contrast. Acceptance levels could vary, according to the customer specification or the IPC definition. There was also the need to maintain design-embedded intelligence from CAD/CAM to the AOI system, include key component and feature information, and to automatically attach specific inspection parameters.
Further desirable features were the ability to create a “black box” machine language and automatically select magnification levels, illumination and greyscale settings, together with an automatic and intuitive set up procedure for operators. To maximise equipment utilisation, it was necessary to have the flexibility to accept a high mix of different jobs, to combine applications such as IC substrates, PCB inner and outer layers, and laser vias, and to cope with a wide range of material variants, all on shorter cycle times. Traceability was an essential requirement, as was the ability to communicate with all CAM stations and industry formats, to inspect, measure, record, report and to communicate with the outside world.
Quite a shopping list! So how could all of these objectives be achieved? Bodegom maintained that the answer lay in parameterising optical inspection, with automatic set up and material recognition, a calibration library, black-box data preparation, and an intuitive graphical user interface. Parameterisation was about translating information into intelligence, automatically populating AOI parameters, identifying types of defect and their impact, and recording and reporting all of the data. Additional options were an enhanced metrology function to enable local height measurement and 3D profiling, finished board inspection capability and the ability to verify and classify defects off-line.
“Be flexible. Go digital.” Was the headline of the presentation from Rita Torfs, project manager at Agfa Materials, proposing inkjet as “more than the traditional way to apply solder mask.”
Agfa had an extensive background in inkjet in wide-format printing systems and industrial printing, and a substantial portfolio of inkjet products for the PCB, printed electronics and chemical milling industries. Although there remained a steady demand for phototool films, they saw great growth potential in PCBs for etch resist, solder mask and legend inks, and saw the change from analogue to digital imaging techniques as bringing benefits in cost, environmental friendliness, speed and flexibility.
But there were major challenges to be overcome in formulating inks for jetting compared with screen printing. In particular, inks for jetting were much lower in viscosity and needed pigments and filler particle sizes to be in the nanometre, rather than micron, range. Whereas aqueous-based inks were good for printing on paper and solvent-based inks were suitable for printing on PVC and other substrates, PCBs required solvent-free UV-curable formulations.
Torfs explained the complex interactions that were taken into account in optimising jetting performance: the design and geometry of the print-head, the viscosity, dynamic surface tension and high frequency elasticity of the ink, and the waveform of the print-head electronics. And all of this was before the ink arrived at the surface of the PCB, where a further series of complex interactions had to be considered: the surface energy of the substrate and the ink, the physics of wetting and flowing, the coalescence of successive drops, the photochemistry of UV pin-curing and the UV or thermal post-cure. Images were created in single or multiple passes, and could be selectively built up depending on the print strategy and the system electronics.
Principal components of a solder mask ink were monomers, photoinitiators, pigments and additives such as surfactants and adhesion promoters. Pigments had to be ground fine enough to give a stable dispersion and good jetting behaviour. And the resulting formulation had to meet the specification requirements of coverage, image quality and edge definition, with fast curing and excellent adhesion and hardness.
Torfs showed examples demonstrating the coverage and imaging capability of Agfa’s solder mask inks and listed test results confirming that they satisfied all of the requirements of IPC-SM-840.
Agfa having discussed inkjet from the formulator’s perspective, it was interesting to learn from the experiences of the equipment manufacturer. Carsten Schimansky, founder and CEO of Notion Systems, described new technologies for solder mask application. He began by referring to the standard process sequence for liquid photoimageable solder mask, and indicating the substantial savings that could be made in floor space, materials, maintenance costs, labour costs and energy costs by changing to an inkjet process. There were significant technical benefits as well, especially the ability to deposit material only where it was required, with no mask in holes or on pads, and to control deposit thickness.
The natural edge profile of an ink-jet solder mask pattern was an advantage where it came close to a component pad, because it avoided any entrapment of subsequent process chemistry. Alternatively, the mask could be brought in contact with the edge of the pad and form a seal. Schimansky showed details of the range of equipment his company could supply, and described some of the finer points of ink-jet printing technique, how keeping the distance between print-head and substrate to a minimum gave higher accuracy, less satellites and less drop deviation by airflow. The provision of an integrated drop-watcher enabled real-time monitoring of main drop volume, satellite volume, drop velocity and drop angle, and was of great assistance in fine-tuning the printer for different inks. An interesting feature was the ability to adjust the reflectance level of a solder mask simply by changing the print strategy, to suit customer preference.
With 25 presentations over two days, covering a huge spectrum of the state of the art in PCB technology, this had been a memorable 50th anniversary conference for EIPC. Indeed, I learned that the call for papers had been oversubscribed and the final selection had involved some difficult choices. But it was a great learning experience for all who attended—and more than that, a splendid networking opportunity for the leaders of the European PCB industry. There is always a wonderfully friendly and sociable atmosphere surrounding EIPC events, and it is always a great pleasure to have the opportunity to attend.
Alun Morgan wrapped-up the proceedings, thanking delegates for their attention, speakers for sharing their knowledge and sponsors for their generous support. Special thanks and good wishes to Michael Weinhold for his enormous contribution over the years, and to Kirsten Smit-Westenberg and Carol Pelzers for their calm and professional organisation and management of another superb event.
“We are 50 years old, and many things have changed. But some things remain the same, and we are still here to support the European PCB industry!”
My thanks to Alun Morgan for kindly sharing his photographs with us.