Printed RFID in 2010
2010年1月7日 Dr Peter Harrop
In 2009, the RFID market grew to $5.56 billion, having almost tripled in five years. Being largely based on government backed military, identity card, financial card, passport and other projects and legally mandated requirements such as animal tagging, this market is recession proof, continuing to grow rapidly. Of the total, $2.18 billion consisted of passive tags, the prime candidate for being printed in future. First to market with entirely printed transistor-based RFID is Kovio with the tag shown below.
The actual circuit printed by inkjet on stainless steel foil is shown below. It consists of about 1000 transistors. This is hugely significant for the future of RFID. See the IDTechEx report Printed and Chipless RFID.
IDTechEx projects that the market for passive RFID tags will grow to over $10 billion in sales by 2019. Much of it will continue to use silicon chips: for example, a passport chip has over one million transistors and no one will print that anytime soon. To be more precise, the backplane driver of the Plastic Logic QUE e-reader has one million organic transistors but that is A4 area - far too big and expensive for an RFID tag in a passport. The much smaller feature size of the Kovio printed nanosilicon transistors means that they could get around 250,000 on a typical credit card sized HF RFID tag - the main value market for RFID - but achieving even the 72,000 or more transistors of ISO 18000 memory chip RFID would surely be uneconomic. The same would be true of microprocessor smart cards, for the foreseable future, in the view of IDTechEx. 72,000 Plastic Logic organic transistors might be squeezed onto a credit card area but again the economics would be in question. Leaving space for the antenna around the edge is a further constraint unless the tag is to work only one way round.
Countries that dominate
The largest database of RFID projects in the world is the IDTechEx RFID Knowledgebase, which currently stands at 3900 projects in 111 countries including a huge number of new projects that make the above forecasts credible. China and the USA spend by far the most money on RFID and China will soon overtake the UK in number of projects. (Japan and China tend to have a smaller number of projects but they are much larger in value). Detailed market analysis and forecasts can be seen in the IDTechEx report RFID Forecasts, Players and Opportunities.
Number of profiled RFID projects by country
The antennas on some RFID tags have been printed for some time. Indeed, in 2007, Hyan Label in China printed HF RFID antennas directly onto paper adhesive labels reel to reel. These gave university students a discount on the Chinese railway system. See RFID in China for more details.
Checkpoint prints UHF antennas for its tags today. However, all this involves the precious metal silver, which is subject to price hikes and, on the occasion of the last hike, Avery Dennison reverted to non-printed aluminium for its UHF RFID antennas. It is hoped that nanosilver inks and dissolved silver inks can result in lower antenna costs and less exposure to silver prices because they use much less silver per antenna. However, these inks are more expensive, partially offsetting the benefit, and adoption in mass production is still awaited. The new printable copper ink may be another breakthrough here.
Printed HF RFID in 2010
The main cost of the tag is the silicon chip of course. So far, no passive RFID tags have been fully printed beyond experiments and trials because scale up and performance has proved troublesome. However, Kovio is well advanced with ink jet printed nanosilicon on stainless steel foil, having provided HF tags for secret trials to the world's favourite RFID specification ISO 14443. PolyIC promises demonstrator kits of printed transistor HF RFID on its website in 2010. These tags will be made by printing organic semiconductors reel to reel. This does not involve ink jet printing. The number of transistors and therefore the performance of these tags may be less but they may be lower cost in the end and they employ polyester substrate. Because of conductivity and cost problems with organic conductors, we expect PolyIC to print or plate metal antennas.
In 2009, Motorola, ORFID and OrganicID dropped out of the race for printed organic RFID but Holst Research Centre in the Netherlands with high bit rates and Sunchon University in Korea and Paru in Korea, using simple conventional printing technology, are among those making impressive demonstrations of organic printed RFID. All that is at HF (13.56 MHz), where most RFID money is spent. That market is in the form of cards, tickets, laundry tags, logistics, passport pages and library labels in the main.
IDTechEx believes that nanosilicon transistors have sufficient charge carrier mobility to achieve UHF RFID (around 915 MHz in the Americas and 868 MHz in Europe) with acceptable voltages. Not far behind in development are single layer transistors of Nano ePrint and carbon nanotube and graphene transistors, some even capable of terahertz performance. For more information see the IDTechEx report Carbon Nanotubes and Graphene for Electronics Applications.
UHF RFID is standard for aircraft parts, pallets and cases and air baggage and it is gaining inroads elsewhere. However, there has been heavy discounting by the traditional chip based UHF RFID tag suppliers for many years. Jointly, they have made well over $100 million loss on a suicidal attempt to reach viable volumes.
It turns out that paybacks are insufficient to justify profitable sale of UHF tags in the tens of billions yearly at tag prices of more than two cents or so (perhaps 3-4 cents for billions in mass transit ticketing). This is beyond the capability of chip technology. Suppliers of chip based UHF RFID selling at a loss at the 8 cent level against their manufacturing costs of around 10 cents achieved nothing at all. Because retailers mandated their consumer goods suppliers to fit these labels, those suppliers would have had to fit them at 10 cents. The result would be the same - a partly imploding program where no one much benefits in the value chain beyond some retailers exerting extreme market power.
After all these years, only a few hundred million UHF RFID tags are fitted to their pallet loads, cases and items every year. By contrast, printed RFID proponents project selling prices of only a few cents in the billions of RFID labels and some foresee sub one cent in much higher volumes. That will open up item level tagging in a big way and there are already enhancements being provided by printing such as the VRF Holdings UHF RFID swing tags for retailing that include radio controlled electrophoretic price indications and promotional messages. For more information see the IDTechEx report Brand Enhancement by Electronics in Packaging and the IDTechEx displays report.
Huge HF RFID potential
IDTechEx expects that the first major incursion of fully printed RFID will be at HF where there is huge untapped potential. For example, item level tagging in medicine and retailing may be best done with the controllable range of HF RFID, potential being trillions yearly. For anti-counterfeiting, the cigarette manufacturers, dominated by the China Tobacco Monopoly Authority, need over 50 billion tags on cigarette packets yearly, when affordable. For security and speed of transaction, the Chinese railway system needs three billion HF RFID tickets yearly if they can get them at the right price. It is already using hundreds of millions of relatively expensive silicon chip-based disposable tickets every year. Billions of price sensitive 70 cent financial cards are sold yearly based on memory silicon chips. Tens of millions of HF library tags with conventional silicon memory chips are fitted yearly and so on. Little wonder that all printed electronics companies targeting RFID choose HF first even where they are already capable of UHF as well.
The bottom line
Bottom line is that there is good news and bad news for the printed transistor developers targeting RFID. The bad news is that, in contrast to other applications of printed electronics, highest volume RFID is all about tag cost, not multiple benefits from printing such as tightly rollable, invisible, stretchable or edible. Certainly, none of the increasing number of developers of zinc oxide printed transistors is targeting RFID as first priority as they see their uniques of high current and transparency as more significant in photovoltaics and displays. Tag cost is 50-60% of total system cost in highest volume RFID so cost is the key parameter overall, particularly for highest volume.
The good news is that fully printed RFID can be one tenth or less of the cost of the conventional silicon chip based tag. In the long term it will be printed directly onto things like 85% of barcodes today, saving the label cost and getting RFID down to maybe one hundredth of the cost of the conventional RFID tag. IDTechEx believes that the printed RFID labels entering volume production in 2010 have a market potential of at least tens of billions yearly. Ones eventually printed directly onto things, no label being needed, have potential of 100 billion to one trillion yearly in consumer packaged goods, healthcare, postal and other applications. For more on printed transistors see the IDTechEx report Printed and Thin Film Transistors and Memory.
Elephant in the room
However, the elephant in the room is the ISO18000 EPC specification series designed with appalling complexity calling for 72,000 to 130,000 transistors in an RFID tag. Until these are simplified to what is really necessary, the printed electronics industry cannot take the market because of complexity. Even in the laboratory, printed RFID currently involves no more than 1000 transistors per tag: indeed some organic versions struggle to rise beyond hundreds of transistors. Tagging a gum ball or a letter does not need more than a few hundred transistors but will the specifications and custom and practice (eg obsession with read write and other complexity) let this shine through in new specifications? Will those specifications be written in something better than the usual geological timescales? Wild card would be a very large user, perhaps in China, where most of the world's manufactured goods are made, adopting the simpler printed option because of price and creating a de facto standard, which ISO eventually has to adopt. The Electronic Product Code of EPCglobal can subsume such a thing. Of its nature, it is not profligate.
Conventional silicon chip dead end
The silicon chip people behind the bloated ISO specifications have a Pyrrhic victory because they will never get down to the necessary low price. Even the read write capability of some printed RFID may be have to be abandoned to achieve the very lowest costs and highest volumes. It is taking time for people to learn from the "make it simple" history of barcodes and "those who do not study history are doomed to repeat it."
For more information attend the World's largest printed electronics event - the IDTechEx Printed Electronics Europe event on Dresden, Germany on April 13-14.
In addition, attend RFID Europe for the latest RFID implementations and technologies, held in Cambridge, UK in September 28-29 2010.
IDTechEx reports encompass all the latest market and technology trends, see Printed, Organic & Flexible Electronics Forecasts, Players & Opportunities and RFID Forecasts, Players and Opportunities.