Active programs are evaluating the use of nano-sized tin, silver and copper particles in Pb-free solders that can be processed below 200°C.

Following completion of the 2004 International Electronics Manufacturing Initiative Roadmap, a conference on innovation was convened at which international speakers and participants looked at how the industry could extricate itself from what some commentators consider a slump toward total commoditization. Some fret that no new huge killer app such as video games, PCs or cellphones is on the horizon. Some believe the three-cylinder engine that has so successfully driven new products - the trio of revenue, profitability growth and investment in R&D - is becoming unbalanced. And worries abound over our ability to support consumer expectations based on Moore's Law because electronics has become a consumer-driven business. The laws of physics, unfortunately, do not respect that. Further confounding the issues is the stretching of supply chains and the uncertainty over just who will bring innovations to market.

What are the takeways from the iNEMI roadmap and subsequent meeting?

• Despite a lack of killer apps, customers are looking for "killer experiences" that exceed expectations such as better displays, longer battery life and more useful cellphones with different features, form factors and finishes.

• Research by Prismark Partners concluded that the biggest R&D spenders are OEMs and IC manufacturers, at 64% and 23%, respectively. The remaining (and surprisingly puny) 13% is split between IC packaging services, EMS, passive components and materials - all critical to the size and economic issues that will determine which new products we buy in the future.

• It takes time for inventions to become innovations and then to appear in products - think seven to eight years for an incremental improvement, such as the adoption of Pb-free SAC alloy, and 15-plus years for a disruptive one that requires serious infrastructure changes, such as the adoption of the transistor. Although product cycles are becoming much faster, process and materials cycles are not.

• Companies need to learn how to become more open in working with third parties. Many developments in nanotechnology are coming from companies in other disciplines, or even from customers, suppliers and competitors. The 2000 or more startups scrambling to find a foothold in the emerging nanotechnology market will inevitably be positioned as competitors to the established supply chain in many industries, including electronics, in the absence of a way to cooperate with established players. A collaborative environment coupled with a more open IP would accelerate progress in a number of areas.

Almost all the speakers at the iNEMI forum mentioned nanotechnology as a key factor in future electronics development. The Semiconductor Research Council, many of whose members are active in the ITRS semiconductor roadmap, recently formed the Nanoelectronics Research Corporation to support and encourage university work in this area, in coordination with the National Science Foundation and the U.S. National Nanotechnology Initiative. The European Nanoelectronics Initiative has similar goals.

Let's explore how to use nanotechnology to overcome technical hurdles and re-energize the industry. In reality, nanotechnology is not really one technology, it is a grouping of techniques - vapor phase, liquid phase, solid state, self-assembly - that permit the manipulation of materials and structures at the nano scale - less than 100 nm (0.1 µm). It is a toolkit for the electronics industry, giving us the gear to make nanomaterials and nanostructures with special properties modified by ultra-fine particle size, crystallinity, structure or surfaces. These are interesting on a scientific level but, no matter how clever the technology, it will become commercially important only when it gives a clear cost and performance advantage over existing products or allows us to create new products.

Often there is a clear size effect with nanomaterials - a "tipping point," below which the surface energy of particles and features or quantum effects starts to take effect (Figure 1). In the case of silver powder, for example, the sintering temperature starts to decline rapidly below 100 nm with a dramatic reduction to below 200°C when the particle size is below 50 nm - this for a metal with a melting temperature of 961°C. This is the basis for the widely accepted definition of nanotechnologies and nanostructures as having a key dimension below 100 nm. A new ISO standards committee, TC229, has been tasked to develop a consistent nomenclature for nanotechnologies.

FIGURE 1: The "tipping point" in particle and feature properties.

Nanotechnologies - leaning on techniques borrowed from chemistry, physics and biology - can offer:

• Uniform particles - metal, oxide, ceramics, composite.

• Reactive particles - as above.

• Unusual optical, thermal and electronic properties - phosphors, analogs of semiconductor devices, heat pipes, percolation-based conductors.

• Nano-structured materials - tubes, balls, hooks, surfaces.

• Directed-assembly - liquid-based, vapor based or even by diffusion in the solid state.

In most cases, the use of a nanotechnology will be invisible to the consumer who only notices a non-scratch surface, a brighter display or longer battery life.

Long-Term Issues

Once CMOS technology dips below about 20 nm resolution, quantum effects such as electron tunneling start to result in phenomena like unacceptable leakage; the only way to move below that size is to use these and other quantum effects in new types of minute structures, be they pure electronic or bio-electronic (remember, the most effective and energy efficient computer available sits on your shoulders).

Both production issues and performance issues abound. As the semiconductor industry moves below 20 nm features, the need for different structures is becoming apparent. Once the industry moves to ultraviolet and then x-ray lithography, it seems there is nowhere to go (in a practical and economic sense) to process ultra-small features using conventional techniques.

Nanotechnology approaches to producing a logic device can be novel and diverse. Imagine making a semiconducting carbon nanotube, then coating it with differently doped materials and assembling it (preferably self-assembling it) in an array. Imagine creating quantum dots that can store a single electron charge or spin. Imagine trapping atoms inside a nanotube and using the electron spin to create a quantum computing device. There are a large number of potential routes to new computing, storage and optical devices. The devices we are making now are clumsy compared with established semiconductor technology. But they will surely improve.

It is fairly clear that the substrate of choice will continue to be silicon. Challenges will be to connect nano materials to silicon in order to detect tiny transitions and to overlay a smaller logic circuit over a larger one (not unlike the redistribution we have to do with silicon to connect 90 nm circuits with 0.1 mm pitch circuit boards). Tiny devices will need to interface with the outside world and a circuit board is still probably the most effective means (Figure 2). Based on this premise three issues arise:

• How to manage the architecture for a regular array that is mismatched to a larger array.

• How to manage a fault-tolerant architecture that can tolerate upwards of 25% defective connections (this will also be needed for silicon as feature sizes decline and devices become more susceptible to thermal or other damage). Note: This is exactly how nervous systems in many organisms have developed, with redundant structures and repair mechanisms to aid survivability in case of injury. Perhaps the most extreme example is the "rewiring" of brain functions that occurs when a person is recovering from a stroke.

• How to develop non-CMOS based logic structures based on spin transitions and other effects such as those used in Nantero's or HP's clever memory devices based on carbon nanotubes.

FIGURE 2: An additional hierarchy of interconnect.

A huge amount of work is being done to commercialize semiconducting carbon nanotubes for electronics. Issues include:

• Producing them cost effectively.

• Making them straight.

• Making them a uniform length.

• Sorting semiconductive from conductive nanotubes (or alternatively vaporizing the conductive tubes or post-treating them to make them semiconducting).

Similarly, work is being carried out to scale up spintronic molecules containing two atoms (typically metal atoms) in an organic system such that spin can be transferred from one to the other and sensed.

Significant progress has been made in all of these areas over the past year and we are starting to see memory and other devices reach the market in developmental quantities.

Mid-Term Opportunities

In many areas of technology, once an area of concern is reached, we can develop a workaround. Hence clock speed, which many followed as the measure of processing capability, has been replaced in some devices by distributed processing with two processors placed on the same chip. This reduces the heat penalty and gives some breathing room - many upper-end processors generate between 100W and 200W - but the heat issue has not gone away. Several unusual properties of nanoscale materials - enhanced thermal conductivity of carbon nanotubes, diamond-like films, nano-metal dispersions - have the promise of aiding heat removal.

Nanowires and other structures using atomic cluster deposition show promise for interconnects, ESD protection structures and sensors whose small size and ability to integrate onto silicon logic circuits using lithography or other imaging techniques coupled with low-temperature assembly promise rapid response and low cost (Figure 3).

FIGURE 3: Hydrogen sensor performance and a high surface area nanowire structure assembled by cluster deposition. The wire diameter is 1 µm. (Nano Cluster Devices Ltd.).

Nanomaterials and nanostructures also increase the efficiency of many types of energy conversion devices (photovoltaic, thermoelectric, battery and fuel cell). This area will get increased attention as the energy supply and demand equation becomes more complex. In fuel cells, nanomaterials can control the microstructure to channel gas, ion and electron flow as needed and can create thin impermeable electrolytes with higher efficiencies - improving to an order of magnitude higher power output per cell than five years ago.

Immediate Opportunities

Enhanced shielding materials, solders, conductive adhesives, underfills, etc. are now possible as nano-sized materials become available and economic. An iNEMI program is starting to evaluate use of nano-sized tin, silver and copper to explore the development of SAC Pb-free solders that will form reliable solder contacts at temperatures below 200°C.

Other opportunities exist in composite conductors. A project at University of Binghamton and supported by a New York State SPIR initiative, for example, is looking at metal powders such as silver and copper as well as carbon nanotubes in composite materials to explore their properties as conductors and shielding materials.

One area receiving a great deal of attention is printable electronics. The concept of printing circuit traces is not new; the technique been used in ceramic hybrid circuits and in flexible circuits used in membrane switches and keypads for many years. The printed electronics market is difficult to quantify because definitions differ, but many experts agree that it is poised to grow dramatically over the next five years.

What is driving this change? It is a combination of new materials, circuit structures and market opportunities. Many of the markets are nascent, the structures are not optimized and the materials still require further development but all areas are receiving worldwide attention with a potential of at least $10 billion by 2010.¹

Printing techniques are of interest for a number of reasons:

• Environmental. Printing processes are additive. Many circuit-forming processes are additive-subtractive, and it can take up to 8 kg of material to produce a 1-kg circuit board, which builds in cost and environmental constraints.

• Flexibility. It can be digitally driven serial deposition or it can be massively parallel deposition using flexographic or lithographic printing. Materials can be deposited on 3-D surfaces such as casings using inkjet or transfer printing. Digital offers flexibility, parallel provides low cost.

• Cost. It can be adapted to low-cost processes; e.g., reel-to-reel on flexible substrates. For the past several years the fastest growing substrate has been in flex, traditionally polyimide for solderability but polyester is used widely as a low-cost substrate in keyboards and membrane switches.

• Low-temperature processing. Non-fired composite or low-temperature (below 250°C) silver systems can be used to create functional circuit elements.

Materials that can be printed include:

• Conductors: To use low-cost substrates such as polyester or paper (instead of epoxy, polyimide or ceramic), process temperatures must be reduced below 200°C.

• Semiconductors: polymers or polymer composites can be printed as components of structures such as solar cells (Graetzl cells), LEDs for displays or transistors.

• Dielectrics: high-K for example for embedded capacitors or low K for insulation.

• Phosphors and other functional materials.

Competing processes include:

• Plating processes: well established but use aggressive chemical baths.

• Etching laminated planar copper on glass epoxy to develop traces and pads: another well-established, low-cost technology.

• Semiconductor processes: e.g. spin-on, lithographic, CVD, ALD, etc.

Printable materials include metal powders and nanotubes - silver for conductors, nickel for MLCC electrodes, copper for component terminations, and carbon nanotubes for thermal and electrically conductive structures (Figure 4).

FIGURE 4: Nanomaterials developed for applications in electronics (clockwise from top left: 200 nm nickel, 20 nm silver, 500nm silver platelets, 50 nm diameter multiwall carbon nanotubes) (NanoDynamics Inc.).

Assembly Not Going Away

How will this affect electronics assembly business? In the near term, higher performance process materials - better conductivity, lower temperature processing, etc. - will be available. They will not be obviously nano in the same way that the tires on your automobile do not look different because they contain nano silica and carbon black, they just grip better and last longer.

Further down the road, expect to see structures that will need special low-temperature assembly techniques - small structures can be destroyed by thermal diffusion in the same way that optoelectronic devices can be. But as with optoelectronics, we can address this by using daughterboards or modules. Also expect to see more pressure to move to printed electronics and flexible circuits (the most rapidly growing board market segment).

Even further down the road, you will still be assembling circuit boards! The components may not use the same logic systems or materials, but they will still have to be interconnected to the "real world" to be useful.

References

1. Nanomarkets, "Printable Electronics: Roadmaps, Markets, Opportunities," September 2005.

Resources

• U.S. National Nanotechnology Initiative (nano.gov)

• Prismark Partners LLC (prismark.com)

• Nanoelectronics Research Corp. (NERC) (src.org/nri/)

• European Nanoelectronics Initiative (ENIAC) (cordis.lu/ist/eniac/)

• Nantero (nantero.com)

• Nano Cluster Devices Ltd. (nanoclusterdevices.com)

 

Alan Rae is vice president of market and business development at NanoDynamics Inc. (nanodynamics.com), and director of research for iNEMI (inemi.org); area@nanodynamics.com. He was formerly vice president of technology at Cookson Electronics (cooksonelectronics.com).

• Prev
• Next