Developments in Printed RFID Vivek Subramanian Published by Pira International Ltd Cleeve Road, Leatherhead Surrey kt22 7ru UK T +44 (0) 1372 802080 F +44 (0) 1372 802079 E [email protected] W www.piranet.com The facts set out in this Pira International Ltd acknowledges product, service and company names referred to in this publication are obtained from report, many of which are trade names, service marks, trademarks or registered trademarks. sources which we believe to be reliable. However, we accept no legal liability of any kind for the publication contents, nor any information contained therein nor conclusions drawn by any party from it. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise without the prior permission of the Copyright owner. © Copyright Pira International Ltd 2005 ISBN 1 85802 964 3 Head of publications and events Philip Swinden [email protected] Customer services manager Denise Davidson [email protected] T +44 (0)1372 802080 Typeset in the UK by Jeff Porter, Deeping St James, Peterborough,Lincs [email protected] Contents List of figures iii All-printed RFID 30 1 Architectures and standards 30 Interplay between device and circuit 34 Printable materials 38 Printable semiconductors 38 Introduction 1 Printable dielectrics 42 RFID classification 2 Conductors 45 135kHz RFID 3 Process integration 46 13.56MHz RFID 4 Printed transistors 47 900MHz and 2.4GHz 5 Gate architectures 47 UWB RFID 6 Top-gate architecture 48 2 Bottom-gate architecture 49 Critical process parameters 49 Opportunities 50 Outlook 51 4 Printing silicon-based RFID devices 7 HF versus UHF 7 Tag construction 9 Printed antennas 10 Choosing HF antennas 10 Value-added technology 53 Choosing UHF antennas 11 Batteries 53 Antenna fabrication 11 Performance parameters 53 Printing methods 11 Battery chemistries 55 Screen printing 11 Sensors 55 Gravure, flexo, offset 13 Time–temperature integration 56 Inkjet printing 15 Temperature readout 56 Materials 17 Peak temperature excursion 56 Metallic pastes 17 Vapour sensing 57 Particle-based inks 18 Displays 57 Fusible nanoparticle inks 19 Bistable displays 57 Organometallic precursors 22 Non-bistable displays 59 5 Materials influence design 23 Fabrication opportunities 23 Summary and outlook 24 3 Major players 61 Printed materials 61 Conductors 61 Printed circuitry 27 Semiconductors and dielectrics 61 Scaling costs 27 Devices and circuits 62 Page iii © Copyright Pira International Ltd 2005 List of figures 1.1 Predicted market size versus RFID tag 3.2 PICA attachment process 29 cost 1 3.3 FSA attachment process 29 1.2 Broadcast power limits 3 3.4 Effect of Q-boosting on voltage 1.3 Some typical 135kHz tags 4 coupled to a tag 31 1.4 Some typical 13.56MHz tags 5 3.5 Printed HF RFID tag 33 1.5 UHF antenna configurations 6 3.6 Printed UHF RFID tag 34 2.1 RFID tag 7 3.7 Archetypal printed transistor 35 2.2 HF RFID antenna stage 10 3.8 CMOS gates have power 2.3 Archetypal UHF antenna 11 advantages 36 2.4 Screen printing 12 3.9 Rectification schemes 37 2.5 Gravure printing 13 3.10 Some well-known printable polymer 2.6 Flexo printing 14 semiconductors 39 2.7 Offset printing 14 3.11 Some printable oligomer 2.8 Thermal inkjet system 15 semiconductors 40 2.9 Piezo inkjet system 16 3.12 Performance trends in soluble 2.10 Antenna production 17 organic semiconductors 40 2.11 Metallic paste ink 18 3.13 Some n-type organic 2.12 Methods to produce nanoparticulate semiconductors 41 inks 19 3.14 Thermally convertible precursors for 2.13 Melting point is related to some printable organic the diameter of metallic semiconductors 41 nanoparticles 20 3.15 Some commonly reported polymer 2.14 Typical nanoparticle synthesis 21 conductors 46 2.15 Nanoparticle sintering process 21 3.16 Archetypal printed transistor 47 2.16 Chemical transformation of an 3.17 Fabricating a top-gate printed organometallic precursor into a metal film 22 transistor 48 2.17 Metallisation using organic metallic 3.18 Fabricating a bottom-gate printed precursors followed by plating 23 transistor 49 3.1 Archetypal RFID attachment 4.1 Electrophoretic display 58 hierarchy 28 4.2 Rolling ball display 58 Page iv © Copyright Pira International Ltd 2005 1 Introduction Over the past decade there has been a dramatic surge of activity in radio frequency identification (RFID). It has gained attention for increasing the efficiency of inventory control, stock management, factory workflow and product shipment or tracking. In an RFID system, items to be tracked are tagged with a small electronic circuit, which can then communicate with an external reader, providing it with a unique identifier that allows it to determine the nature of the tagged item. RFID systems can track pallets in a warehouse, manufactured goods in a factory, or individual items in a department store, and could give a huge productivity boost in any environment requiring careful stock control. The potential applications of RFID are tremendous. The number of items that could be tagged is staggering. It could ultimately replace optically scanned bar codes, currently used all over the world. The biggest drawback is cost. For RFID to achieve widespread use, the cost of tagging must be negligible, so the economics of tagging, reading, etc., are outweighed by the benefits of increased tracking efficiency. There are various estimates of the relationship between tag cost and tag sales volume. Figure 1.1shows a strong inverse relationship: volume goes up as cost goes down. For RFID tags to become ubiquitous, the cost will have to reach a few cents per tag or even less. FIGURE 1.1 Predicted market size versus RFID tag cost 15 14 13 12 11 s) g a 10 umber of t 89 ntication n e og (10 7 e auth L u 6 al v h- 54 Hig are card a arcode 3 allet ase, f harm em b 2 P C P It 10 1 0.3 0.1 0.01 Tag cost ($) Source: Pira International Ltd Figure 1.1shows the cost of tags. Since tags are the repeating cost in an RFID system, volume–cost analyses are usually based on tag costs, but reader costs need to be considered in some applications. For example, in department store inventory control Page 1 © Copyright Pira International Ltd 2005 Developments in Printed RFID Introduction applications, many readers will be used throughout the store and their total cost could be substantial, strongly influencing RFID deployment. Unfortunately, it is difficult to estimate reader costs based on a tag configuration. Reader costs are not covered in this report, but always investigate an RFID application to see whether reader costs are important. RFID classification In general, RFID tags may be classified by how they obtain their power and how they communicate with the reader. One classification scheme divides tags into active tags, which have an internal power source, i.e. a battery, and passive tags that get their power from the reader. Active tags Active tags cost much more than passive tags, but active tags are already being used in several inventory management and tracking applications, since the battery typically gives them enough power to operate over several metres or more. Yet the cost of active tags, based on integrating and packaging a battery into the tag, means they are rarely considered promising candidates for ubiquitous RFID applications such as department store inventory tracking, warehouse pallet tracking, etc. Battery technology is considered briefly on page 53. Passive tags Passive tags do not contain a battery. Instead power is supplied to them by the reader. The reader broadcasts large amounts of power (the specific power limits are imposed by the relevant government licensing agencies) from its antenna. Tags in turn have antennas to capture or harvest this power and charge up an internal capacitor. Tags that are sufficiently close to the reader can collect enough power to become energised. Since this power transfer is extremely inefficient, the range of passive tags is usually substantially smaller than for active tags. The typical operating range of passive tags is a few centimetres to a few metres. The inefficiency of power transfer is because the reader doesn’t know the tag’s location, so it radiates power isotropically, or equally in all directions. The tag can obtain power only from the electromagnetic field that actually interacts with its small antenna, and this wastes large amounts of the reader’s power output. Figure 1.2shows the limits of power output imposed by regulations. Many of these figures are being revised upwards as RFID gains wider acceptance. Page 2 © Copyright Pira International Ltd 2005 1 Developments in Printed RFID Introduction FIGURE 1.2 Broadcast power limits 7 6 1) –m / V µ 5 / h gt n e str eld 4 (fi10 g o L 3 2 0.125 13.56 900 Frequency (MHz) Source: Pira International Ltd The broadcast power limits, the antenna cross-section on the tag and the power requirements of the tag circuitry determine the range of a given combination of tag and reader. Active tags have internal batteries, so they are constrained only by the sensitivity of communication between the reader and the tag, not by the need to provide power to the tag, so active tags tend to work over a much longer range than passive tags. RFID tags may also be classified by how they communicate with the reader. This is usually based on the frequency at which the reader broadcasts information to the tag; also, the power to passive tags is radiated at this frequency. Several frequency ranges are commonly used for RFID applications. The availability of ranges for RFID is determined by government agencies, as the readers must not interfere with the ranges licensed to broadcasters such as television stations. In general, the frequency bands already used for RFID around the world or under consideration are 135kHz, 13.56MHz, 900MHz, 2.4GHz and 5GHz. 135kHz RFID RFID at135kHz has been widely deployed. Two major types of RFID tag exist at this frequency: tags for very short range (near-contact) applications and tags that operate over longer ranges (several centimetres to several metres). The tags operate in the near field of the reader, i.e. they interact primarily with the magnetic component of the electromagnetic signal broadcast by the reader. In this regime the tag antenna consists of an inductor, and the reader antenna also includes an inductor. When the tag is within a usable operating range of the reader, the two inductors are coupled, creating a mutual Page3 © Copyright Pira International Ltd 2005 Developments in Printed RFID Introduction inductance between them. This allows the reader to provide power to the tag and allows communication between the tag and the reader. RFID at 135kHz is attractive as this frequency doesn’t significantly interact with water and other fluids and also works fairly well in the presence of metals. Therefore tags that operate at 135kHz or lower work well in environments containing these materials; indeed they are widely inserted into livestock for inventory control; an external reader can read the tag inside the animal The main disadvantages of 135kHz RFID are the size of the inductor, the configuration of the tag and the range of operation. In general, the lower the operating frequency, the bigger the inductor. At 135kHz the inductor is extremely large; if it were implemented as a planar spiral on a plastic sheet, its radius would be several centimetres, creating a very large tag. Large planar spiral inductors typically give very low efficiency in the power coupling between the reader and the tag. The parameter Qfor an inductor gives an indication of how lossy it is. Large spiral inductors are essentially very long wires, so they have a very large series resistance, hence a low Q(often with an upper limit in the range of 1). This limits the range of 135kHz RFID tags with planar spiral inductors; typical ranges are a few centimetres at most. Range can be extended substantially by using another configuration instead of a planar spiral. The value of Qis dramatically higher for a wound coil on a ferrite core, extending the range of the tag to several metres. Unfortunately, the wound inductor and ferrite costs tens of cents, so there is a price penalty. Therefore 135kHz RFID is not considered promising for ubiquitous tagging applications. If someone found a way to reduce the cost of ferrite-cored antennas, it could revolutionise the RFID industry, since 135kHz has great advantages in a world full of liquids and metals. Figure 1.3shows some typical 135kHz tags. FIGURE 1.3 Some typical 135kHz tags Coil wound Circuit board with Wound inductor on ferrite rod silicon chip, etc. on ferrite Chip Source: Pira International Ltd 13.56MHz RFID One of the most important RFID frequencies is 13.56MHz. At this frequency, power is still inductively coupled. However, compared to 135kHz, the antenna is substantially smaller, with a typical tag inductor of radius 1–2cm. This results in a smaller, hence cheaper tag. And since the inductor is smaller, it is possible to achieve higher Q; planar spiral configurations have achieved Q= 5–20without too much difficulty. Unlike 135kHz RFID, Page 4 © Copyright Pira International Ltd 2005 1 Developments in Printed RFID Introduction there is no need to resort to expensive inductor configurations. At this frequency, operating ranges of several tens of centimetres have been achieved, consequently 13.56MHz RFID has been considered for item-level tracking applications; it is already widely used in library inventory control and will probably soon appear in pharmaceutical tracking applications. 13.56MHz RFID works well in the presence of liquids, but is fairly susceptible to interference from nearby metals. The metal interaction problem may be mitigated by using an insulating spacer between the metal surface and the tag, but it could still limit the usefulness of 13.56MHz RFID in some applications. Figure 1.4shows some typical 13.56MHz tags. FIGURE 1.4 Some typical 13.56MHz tags: (a) Texas Instruments, (b) Rafsec (a) (b) Source: Pira International Ltd 900MHz and 2.4GHz RFID tags operating at 900MHz and 2.4GHz are typically called ultra-high frequency (UHF) tags; tags operating at 13.56MHz are typically called high-frequency (HF) tags. UHF tags operate very differently from 135kHz and 13.56MHz tags. They typically operate in the far- field region of the reader’s electromagnetic field, so they primarily interact with the electric component of the field. This means that inductive coupling is seldom used for UHF tags; the tag antenna isn’t an inductor. UHF tags use backscatter systems for communication, and the antenna typically has a dipole configuration. This is easily formed in a planar geometry and may be made long and thin, producing tags that may be several centimetres long but less than a centimetre wide. The antenna is crucial to UHF RFID and antenna designs are valuable intellectual property. Figure 1.5shows some examples. Page5 © Copyright Pira International Ltd 2005 Developments in Printed RFID Introduction FIGURE 1.5 Some common UHF antenna configurations: (a) Rafsec, (b) Texas Instruments, (c, d) Alien Technologies (a) (b) (c) (d) Source: Pira International Ltd UHF RFID systems can typically achieve long-range operation; some systems have ranges of several metres. This makes them extremely attractive for applications requiring longer ranges than offered by HF tags. As a result, UHF tags are being rapidly deployed in pallet-level and case-level tracking applications. Unfortunately, they are extremely sensitive to liquids – liquids absorb radiation very strongly at 2.4GHz – and also interact strongly with metals. They are prone to null points, regions where the signal strength is extremely low but still within the system specification. As a result, UHF tags may be problematic for use in item-level tagging where there are liquids or metal containers, and the readers have to be very carefully positioned to reduce the impact of null points on tag detection. But given their long range, UHF tagging remains arguably the largest growth area in RFID. UWB RFID In the past few years there has been an upsurge in RFID implementations at frequencies of about 5GHz, often called ultra wide band (UWB). Here are some of the advantages. First, the antenna may be even smaller than UHF antennas; the operating frequency is higher, so the wavelength is smaller, hence the antenna can be smaller. Second, the reader is potentially much simpler than a UHF reader, reducing overall deployment costs. The third advantage relates to anti-collision. Anti-collision is how the reader discriminates between the many tags that potentially exist within its reading range at any moment. Using a predefined protocol, anti-collision allows the reader to identify each tag within its reading range, not a jumble of noise from all the tags talking at once. Anti-collision is potentially easier in a UWB implementation, reducing the overall system cost and possibly increasing the read data rates. Page 6 © Copyright Pira International Ltd 2005