MicroTech 2007

Citation

Ling, J. (2007), "MicroTech 2007", Microelectronics International, Vol. 24 No. 2. https://doi.org/10.1108/mi.2007.21824bac.001

MicroTech 2007

MicroTech 2007

“Advanced Interconnection” was the theme of the iMAPS UK Annual Conference, held this year at the Helidon Lakes Hotel in Northamptonshire, on 6 and 7 March. The many delegates were welcomed by Peter Ongley, outgoing iMAPS UK Chairman.

As well as the conference there was a table-top and poster exhibition, at which companies such as Stevenage Circuits, PandA Europe, Polymer Systems, Taylor-Hobson, Accelonix, the University of Greenwich, Exception PCB, amongst others.

6 March

The Keynote Speaker was Graham Pink of CSR (Cambridge Silicon Radio) who wanted to look at changing the way the world connects. CSR was founded eight years ago, by nine people, now they have 1,000 employees, operate in ten countries, and had sales revenue last year of some $700 million. They produce blue tooth modules, some 250 million devices a year, and their chips are sold into companies manufacturing headsets, mobile “phones, play stations, satellite navigation systems; CSR say that their devices are making radio technology available to the masses.”

Package for a radio ought to be simple, but it is not. Customers always want something different, various host interfaces, then there is RF, analogue, digits and power supply – and then there is cost. Here, they make the silicone as small as possible, and they bridge the gap between size and functionality by using WLCSP products. They look like BGAs. Graham described how they were made by their assembler, UNISEM Europe (Atlantic) in South Wales and for the company packaging was at the time low on their list of “things to do”. But then along came Bluetooth, which brought in a whole new packaging technology, based on BGA architecture, a Chip On Board without underfill, which made placement easier, improved yields, reduced inspection, and cut rework costs. Quite why some companies have not moved on was a mystery to Graham, given that their own market in the mobile telephone industry will contract.

Technical programme

Advanced material technologies.Dr Andy Wright from Fibrelok UK described a new compliant bonding technology, and opened by referring to the application of gallium-arsenide to silicon; here there was a great mismatch, which lead to many defects – epidaxial lift-off and transfer, for example, dislocation filtering, so now it is gallium-nitride which needs to be put down onto silicon, but fatigue in solder joints is a problem here. So, why not create a compliant bond of joint which is lead-free, conductive, and low cost? Current possibilities include copper pillars, spring structures, with better compliance by going vertical. A mechanical “Velcro” seemed attractive, but too expensive, so a solid state adhesive would work, it was felt. The two shoe-brush phenomenon was illustrated, which is the basic principle behind FibreLok. Here, vertical metal fibres made of nickel are created, which can work as a plug connector, and sub- micron parts are possible – fibres can be 500nm and down; there are multilayer fibres from various metals, nickel, copper, gold, etc. Applications include chip to chip; chip to package, MEMs, chips on glass; solvent proof bonds; and integrated opto-electronics. It is generic, it is lead-free, and it works.

We thought that Ragu Das of IDTEchEX was going to give a paper on materials and ultra – low cost assembly techniques used in RFID technology, but he made it all much more interesting by saying that printed electronics will require the packaging of billions of chips in the future for RFID alone, so the microelectronics industry will be busy. RFID is digital information, pure and simple. Wal-Mart insists that all their suppliers use RFID, and Coca-Cola have estimated that with RFID a 300-day cycle can be reduced to 30. RFID tags will top one billion this year, and 20 billion in five years time. RFID is an enabling technology, can be used in toys as a functional system, for example. Library books can have an RFID tag and be checked without removing them from the shelf. Many other examples included air baggage, banknotes, credit cards, and supermarket barcodes brings demands into the trillions.

Supermarkets are forcing the pace, you can have networks within networks; you can have an RFID chip in your mobile phone that registers every time you have consume something in particular. Target price is less than 5 cents per tag.

Antenna type chips were described. Chip attach systems were also illustrated, the RFID road map showed that there will be a strap attach system as well, printed electronics replaces the silicon chip, and can be used in many different ways; it is a big topic that is on its way.

In a well-illustrated talk, Dr Chris Hunt of the NPL spoke about measuring true stress and relaxation properties of solder joints. They have worked with two SAC alloys, simulating thermal cycles, with temperature controlled cabinets ranging from 2110 to 1508C. NPL had carried out a highly detailed series of tests, using an experimental design, which showed true stress for hysterisis loops at different temperatures, and stress relaxation at various dwell times and temperatures can be assessed, and they can see the stress crack take place under these controlled conditions.

After lunch, iMAPS ran a session entitled “Market watch – is the supply chain broken?”

Aubrey Dunford is the man from Afdec, the Association of Franchised Electronic Components, whose members operate franchised component distribution, sitting between importer and end-user. We are no longer a mass-market, except possibly the automotive sector, and the key driver is investment in systems and in the infrastructure, all of which depends upon the state of the economy. Manufacturing has shown the biggest drop over the years, and DTAM is correspondingly low. There will be no great growth in 2007, with interest rates rising and industry slowing, and the outlook for 2007, said Aubrey, is that production will shift eastwards (Europe and Far East), whilst there will be limited growth in the Western economies. Diversification is the answer; many distributors are diversifying into embedded software, system level boards, component assemblies, and system design. Large suppliers are moving their support and design teams into Great Britain, a country now strong in the service industry sector. Is the supply chain broken? No, but it needs to change and adapt, and success will be bound to follow.

Paul Palmer is with KTN (formerly the Prime Faraday Partnership) and they have been going since 1998. Where there was once Celestica, GEC, Thorn- EMI, Plessey, Ferranti, Marconi, etc. all giants in their day, now companies evolve, and evolving companies such as Brindley Electronics do well, sub- contracting assembly, and the key to what they do is the clever use of information, using uncommitted software. Merlyn Electronics is another, they have reinvented themselves in the supply chain, they have moved into the area where they are installing electronics into bus shelters, – they have new ways of working, with interconnectivity with low cost of communication; virtual workspaces; they have a global perspective. All change here.

Professor Chris Bailey from the University of Greenwich valiantly fought off ill-health to come along and talk about education. The education pipeline – skilled people, are they available in the UK, or would you have to go abroad to find them? The pipeline could be in North Americas, the Far East, but remember that the UK is the second largest producer in Europe, and have the second largest market. What remains as drivers for technology are design, military, aerospace, automotive and medical industries. Packaging is changing. Universities such as Loughborough, Heriot-Watt, Lancaster, Bolton, Greenwich and Nottingham cut across the traditional barriers, multi- disciplinary sectors packaging designers of the future need to be multi-disciplinary. We have good industry bodies such as the Smart Group, NPL, Soldertec Global, and IeMRC, an initiative from the DTI, driven by industry, and funded by the EPSDRC. Knowledge transfer partnerships – supported by the DTI for the next two years; in the UK we punch above our weight. Education in the UK is much the same as in USA, but we do need to keep the pipe line here for as long as possible.

Indro Mukajee said that he was not a ghost or a has-been, he manufactures in the UK, and he wants to stay here. He runs a company called C-Mac Micro Technology – but the competition comes from all over the world, especially China. You need to know your market; know the buying trends; invest in innovation; and partnerships are an essential part, as well. What are the issues regarding graduate intake to university? Is electronic engineering too daunting for the young? We must have engineers going into schools and talking to the pupils, and we need a skilled set of them who are prepared to. We need a skill set that are prepared to move, and have the talent to drive it forward. Framework 7 asks for training skills which are not in university. Analogue needs maths, and maths is not popular with the young.

The panel fielded some really interesting points and questions from the delegates, mostly hinging around the link between schools and industry; how do you make industry attractive; how do you find or inspire the engineers of tomorrow? Tell them may be one way, tell them that CSR, for example, is a key player in the manufacture of Playstation3 and Apple mobile phones and you may get a flicker of interest. Undergraduates, where do they come from?

Technical programme

Future connections.Andy Longford of PandA, and iMAPS UK Chairman, gave us the results of road map exercises that ELFNET has taken. ELFNET was an EC Funded project (e2.3 million) with 19 partners across Europe, and was the European Lead Free NETwork. They have a very powerful web site, supported by 380 European expert members, 530 industry members, and have 120 lead-free projects running. Last year they had 300,000 visitors to the web site. Deliverables included solder alloy database with their properties, a reliability book on failures and testing, and roadmap. There are five major topic areas, and a good quote on lead-free says that “unless new technologies are developed, high reliability applications will never be able to comply with RoHS”

Roadmapping is a 79 pp. document available on the ELFNET web site. In Framework 7 electronics is the key player and affects every other subject from food to building. Interestingly, component packaging came out of the exercise as an area where more help is required. High reliability solders are a major driver, with CTE matching, process technologies, process optimisation, effective rework and repair, a need to decrease the number of process steps, more process control; testing and reliability.

Amir Wallrabenstein from the University of Paderborn wondered why optical system interconnects would be needed? Large-scale computers and switches they need play a role here, as the physical properties constrain bandwidths, and the best way to overcome this bottleneck problem is to use complex switching; and here the most promising solution is optical interconnects. Optical interconnects allow high bandwidth by use of high clock rates. EOCB – electro-optical circuit boards enable high bandwidth on board data transfer by board integrated optical interconnects. The system is compatible with existing technologies. The waveguides are made from polymeric material, the pitch is 250mm, incorporated by a standard lamination process. A 10Gbps-system with AXC-technology was also explained which can be incorporated into a standard board lamination process. The requirements for a design system were also described.

Jan Vanflerteren of IMEC in Belgium presented the case for stretchable conductor technologies for elastic electronic systems especially for portable applications. These offer not only a reduction in weight but also enable the interconnects to be folded, rolled and even wrapped so the circuit can be deformed in an arbitrary way to suit the curvatures of human body parts providing maximum comfort to the wearer.

This emerging technology will have a major impact on biomedical electronic applications and for wearing electronics in clothes possibly for military defence purposes.

The main challenge for elastic stretchable circuits is the development of low resistance conductors typically used for conventional rigid and flex circuits of today.

The development project of stretchable elastic conductors was funded by the E.C. STELLA project and from the BIOFLEX Contract supported by the Flemmish Community.

This Presentation was awarded with the Best Paper Award based on various criteria of: Originality and Innovation, Industrial/Commercial Impact, Published and Presentation Quality.

7 March

Process technologies

Dr Rolf Beckert from Hesse and Knipps talked about the current and future aspects of wedge bonding. Wedge bonding sits alongside wire bonding and ball bonding for the bonding of wires to chip carriers. There is wedge- wedge bonding, ribbon bonding and ball-wedge bonding. Rolf described flip- chip and compared this to other assembly technologies; it carries less weight, has better electronic properties, Wedge bonding versus ball bonding was illustrated, where wedge bonding worked better on small pad sizes, lower loop shapes, and had no requirement for part heating, even if it was a slightly slower process. Ball bonding is popular for the automotive industry where ceramic bonds are needed. Placement accuracy with wedge bonding is excellent – accuracy of 5m or less is offered. Ribbon bonding will become more and more popular, thought Rolf, replacing wire bonding, and there will be a renaissance of thin wire wedge bonding, mainly due to the rising cost of gold.

Josef Pupp from Datacon in Austria looked at the challenges in multi-chip die-bonding. These include accuracy, reliable dispensing, theta corrected stamping, stack dies, thin dies, flip chip, multi-chip, and a reduced cost of ownership. The 2200Evo is their new die attach machine, which is offering higher throughput, and which can assemble very complex packages. Josef described the new bond head, with infinitely controllable bond force, the new tool changer, eject system and flip chip systems of this new machine, in an impressive presentation. About 25mm placement accuracy at very high speed for SiPs, MEMS and stacked dies. A short video was shown, and a great deal of interest was shown in this impressive machine.

The mass imaging of electronics materials was described by Clive Ashmore from DEK Printing Machines Ltd He wondered if there would still be a market for assembly. Clive traced the path to high density gate arrays, with the market being driven by the giants of China, India and possibly Russia. Their consumers push the interconnect technology forward, smaller, cheaper, more functionality, always the three code words. Now it is thinner, too. Greater integration is needed between device, package, system designers and manufacturing companies. DEK have come a long way since their screen printing days with just conductive inks; nowadays a screen printer is a high accuracy mass imaging system, which can handle die (wafer) bumping, interposer bumping, no-flow flux underfill, ball placement down to 0.175mm.

CSPs 0.1 pitch might be a problem, DEK can print small, but once you mix UFP components with standard SMT, life gets complex. The key lies in the stencil, maybe in nano-film coatings. Printed electronics is a subject they are watching closely, with very considerable potential.

Mydata Automation AB from Sweden had sent along Dr William Holm to talk about the future of ink jet printing. Jet printing can be traced back to 1850, Lords Kelvin and Raleigh made a system even then. Adhesive jetting came along in 1993, and in 2005 MyData offered solder paste ink jet machines. In the future, batch sizes will shrink, lead times will be shorter, changeovers will be frequent, there will be fewer operators and there will be less cleaning. “SMT meets semicon” is the scenario where jet printing will apply; in the field of non-planar surfaces where components need to be placed, ink-jet could handle it, screen printing would not. High density electronics calls for very close placement, 1mm and not 10mm, for example. Ink jet can give you 1mm. Ink jet lends itself to more diversification, with less tooling, lower operator costs and is a production process entirely governed by software. About 0.3mm is the smallest solder paste pad that can be defined by inkjet.

System integration

The ebullient David Topham knows all about 3D circuitry, which is where his company – 3D Mintegration – designs in 3D, manufactures in 3D, and tests in 3D. The market drivers are automotive, aerospace telecom medical; and consumer products, and his company wants to break down the compromises of traditional manufacturing. It is a grand leap from the customer planer integration of PCBs, semiconductors and components. They make toolsets for industry, a virtual reality design environment in which designers can be immersed to enable their comprehension of the concept. Innovative thinking, fast modelling, looking at the risks, 32 processes are being developed by the project, and process capabilities are being extracted. Brunel University has been chosen as a partner in this project, micro manufacturing at low costs of entry. All brave steps into the unknown, but once they get there this should be a presentation worthy of a very large audience indeed.

Rob Dry of Filtronics looked at a low cost packaging approach for high frequency MMICs to meet the costn demands of the telecom industry, and reflected that industry wants smaller and smaller products at lower and lower prices. Traditional commercial HF product design showed high package costs, with a high number of assembly processes, so Filtronics went for bare MMICs die directly mounted and interconnected on to top level assembly. But here you had high product board costs, complex board metallurgy, and in-house chip and wire capability was necessary. So lower costs were seen in QFN surface mount package outlines, a plastic package. Filtronics found that with QFNs they could save 18 per cent with chip-and- wire assembly, no gold finish is required, and laser cut pockets would no longer be required. A QFN can be tuned to give a good signal performance, was one of their findings, and the overall picture is good.

Professor Marc Desmulliez from Heriot-Watt University knows about HUMMS (Health and Useage Monitoring MicroSystems). These can be used in larger electronics systems and used to monitor the health of cable wiring. Cable wiring is used in aircraft, for example there are 30km of it inside an F-111, for example (are they still flying?). Here, we are back to 3D, but with a simple sensor inside the cable itself, measuring temperature, pressure, humidity, strain and current spikes. Such sensors have to be integrated into the cable, have to be diagnostic, so here the Rogowski Coil works well. And they wanted to integrate this into the fabric of the cable, so they have now produced a 3D version of the coil. Direct metal writing onto polyimide is something else they are working on, so this is the system they are going for.

Substrate technologies

Andreas Ostmann from the Fraunhofer Institute in Germany had come over to talk about lamination techniques for SiPs. Embedded chips in PCBs already exist, such as the Imbera system in Finland, the Casio in WLP; GE have a thin film technology on top of an embedded chip. At Fraunhofer they have the IZM, using build-up layers, chip in polymer, which gives extremely thin packaging, has the capability for 3D component stacking, gives a very short interconnect between chips and substrates, and one can used established processes and materials. You can have CiP with conductive die-bond; a CioP with a non-conductive die bond, and an embedded flip-chip with very thin interconnects. HIDING DIES was described, an EC funded project for embedded chips in a 16£24in. board. Herr Ostmann explained precisely how the CiP was laminated and then processed for connectivity, using laser drilling and normal electroplating of the laser-drilled microvias. Chips are 50m thick, and vary from 2.5 to 10mm. The embedded chips pass all the usual and stringent reliability tests, temperature storage, thermal shock, humidity storage, and Jedec Level 1. They have also successfully laminated chips 400m thick, for a power MOSFET module. Fraunhofer now have a generic process for embedding thin chips into polymer.

Wim Christiaens from IMEC in Ghent, Belgium spoke on embedding active components in multilayer flex boards. The SHIFT Project, another EC funded project, uses thinned wafers 20-30m integrated into flex material. Thinning is best done by silicon dry- etching rather than sawing, which is damaging. Dicing by thinning can give you very small chips, using standard adhesive tape Wafer. They screen print the adhesive (acrylic) for the thinned chips, and this printed result is laminated, laser-drilled and electroplated in the usual fashion for microvias. Embedded die can be used as a package. And can be solder assembled on rigid or flex PCBs. Dicing by thinning makes it possible to get wafers down to 20-30m. Flip Chip in Flex can be mounted directly on or in flexible laminate.

David Woodley runs Technograph Microcircuits Ltd and was making the case that Thick film is still relevant. UK microelectronics thick film manufacturers include MCE, C-MAC, Extec, Welwyn, Hitech, and Pelikon, who use TFT for high volume applications. System In Package might be a better name but the applications are for needs such as size, mass, and where high speeds are demanded, for the medical aerospace and automotive applications. Thick film is medium cost, high performance, forms a valuable part of the armoury, can be used in conjunction with other technologies. Ceramics, or glass, or metal (for high power, integral heat sinks, etc.). Examples included laser detection devices for aircraft, an amplifier which operates best at -200¸C; an “event” detection system. The “Typhoon” Eurofighter has over 80 packages on board based on thick film technology. Myriad uses, thick film is as applicable to-day as it ever was.

Debbie Raynes from Ansoft UK was the last speaker of the day, and she spoke with huge knowledge about SiP and stacked package solutions. SiP contains multiple dies, combinations of wire-bond and flip-chip, and memory and logic is combined with passives, filters, antennae, etc. More functionality in a smaller and smaller space is demanded by the mobile phone industry where the equipment is now multi-purpose and multi-function. Ansoft SiP solution is to merge package layouts, using a merge utility software, merging.siw files. Their software is able to model the package, and look at electrical, thermal and mechanical effects; it can simulate optimal run times, and can print out the thermal footprint so that one can see the areas of stress.

iMAPS always manage to attract some really interesting speakers, this conference was no exception. March 2008 will see iMAPS celebrating their 40th Anniversary with MicroTech being held in Chelsea Village, London, with the precise dates to be advised.

John LingAssociate Editor, Microelectronics International