Flexible e-Book Displays Produced in Standard TFT and Module Factories

Thin, light, and robust flexible e-book displays are being prepared for mass production with most of the same equipment and processes used for glass displays.

by Ian French

THE FIRST e-BOOKS, such as the Rocket and Softbook reader in the 1990s, were not major commercial successes. This was probably due to a combination of limited reading material being available, poor content distribution, and the fact that these devices used active-matrix LCD screens that did not prove popular for immersive reading. In the last few years, there have been major developments that have led to wider acceptance of e-books. First, high-quality electrophoretic displays were created that closely mimicked the visual appearance of printed paper. Then, Amazon, Sony, and other companies developed the infrastructure to make large numbers of books, newspapers, and magazines quickly and easily available, either by wireless or via the Internet.

Electrophoretic displays had been under development by several groups since the 1970s, but by the early 1990s they had largely fallen out of favor because of the rise of LCDs and the many practical problems associated with making high-quality electrophoretic displays with acceptable performance and lifetime. In 1995, the Jacobson group at MIT revisited electrophoretic-display technology and developed a foil with microcapsules containing black-and-white particles with opposite electrical charges on them in a clear liquid. The particles could be moved by external electrical fields, so that either the black or white ones were toward the top, or they could have different degrees of mixing within the capsule. An observer would then see the capsule as black or white, depending on which type of particle was uppermost, or a shade of gray that depended on the distribution of mixed particles. Along with good control of the surface chemistry of the charged particle, this approach solved many of the pre-existing problems of electrophoretic displays.1The development of electrophoretic foils was spun out from MIT just 2 years later when the E Ink Company was formed in 1997.

To make high-resolution e-book displays that can show pages of text and monochrome pictures, E Ink foils are laminated onto amorphous-silicon (a-Si) TFT backplanes of the type used in most active-matrix LCDs. The first e-book to use E Ink foils was the Sony LIBRIé, which was launched in 2005. Initially, TFT-backplane manufacturing and module integration were carried out by Philips, but these activities were passed over to Prime View International (PVI) late in 2005 when Philips decided to stop manufacturing displays.

Currently, the e-paper market is one of the most dynamic in the display industry. Market-research-company DisplaySearch estimates that it will grow from a value of $100 million in 2009 to $9 billion in 2018. It is ironic that this dramatic growth is based on a technology that would seem like a backward step if we only consider the usual display metrics of brightness, contrast ratio, color gamut, and speed of response. Electro-phoretic displays have slow response, a lower contrast ratio than LCDs, and, at the moment, they are only monochrome. On the positive side, they benefit from long battery life because they only draw power when the page is rewritten, but their major attraction is their paper-like appearance.

Although e-books have captured much of the visual experience of the printed page, their feel is very different. This can be important for immersive reading, which people typically do while holding a book, magazine, or newspaper. At the moment, the e-book displays have a rigid, glass TFT backplane that must be enclosed in a metal case for protection. This makes them comparatively heavy, particularly for larger displays, such as 8 in. and above. They are well protected by the metal case, but they can still sometimes break if they are dropped or if an object presses hard against them, as can happen in a bag. The obvious solution to these issues is to replace the glass backplane with an array of TFTs on a plastic substrate to make the display thinner, lighter, and more robust. In fact, it was so obvious that E Ink first started working with Lucent Technologies to make flexible displays with organic TFTs on a plastic substrate as long ago as 1999. Since then, E Ink has demonstrated high-quality flexible displays, with more than 10 different companies and research institutes using a range of different TFT types, flexible substrates, and fabrication techniques. Surprisingly, given that so many companies have successfully made good demonstrators over the years, flexible displays have still not reached the market. One of the main reasons for this is that many of them were made using experimental techniques on small substrates that would have required large investments in new machines and methods of handling in order to scale up to mass production. Organic or plastic TFTs also require new materials systems that have not yet made it into commercial products, despite having been in development for use in TFTs for more than 10 years.

The EPLaR Process

The electronics on plastic by laser release process (EPLaR)2 is one method that can be used to produce flexible displays. (It is the method that has been adopted by the author's company, Prime View International.) This process employs more than 95% of the same steps and equipment that are used for making glass e-book displays, which minimizes invest-ment cost and development time because no new factory is needed and relatively few new process machines have to be purchased. EPLaR displays also directly benefit from years of experience of manufacturing and using glass displays in e-books.

The main steps in making active-matrix electrophoretic displays on glass involve a-Si TFT-array fabrication, followed by display-module assembly. A field-shielded pixel structure having a relatively thick polymer insulator over the TFT is used. This allows the ITO pixel to lie over the TFT, rows, and columns, maximizing the ITO area and therefore the area of the electrophoretic foil that is being controlled. This is the same TFT structure that is used in LCDs with large-optical-aperture pixels. The steps in module making are scribe and break to separate the displays on the motherglass; lamination of an electro-phoretic foil; attachment of row and column drivers by a chip-on-glass process; and then, finally, attachment of a flexible printed circuit board (PCB) to provide connections to external driver electronics.

The EPLaR process closely follows that for making glass electrophoretic displays, but with some additions. The first extra step is inserted before TFT-array fabrication. A 10-μm-thick polyimide layer is applied to a standard glass substrate. This will eventually become the plastic substrate. As long as the correct interface treatments are used in conjunction with the correct type of polyimide and curing schedule, then the polyimide adheres very strongly to the glass substrate. The polyimide can withstand all of the standard processes used for making TFTs on glass substrates.

To reduce the level of particles and contamination, all TFT factories minimize the number of workers inside them by using integrated automated handling systems. These use cassettes on automated guided vehicles for carrying the glass substrates between process machines. The substrates are transferred from the cassettes to processing machines, and between adjacent process stations, by robotic arms. The glass substrates are so large, even for Gen 2 factories, that they sag in the cassettes, and the spacing must be very accurately controlled to obtain the maximum number of displays in a cassette while still allowing the pick-up arm of the robots to get between the substrates. The automated handling is so precise and finely tuned that, in the past, factories have not been able to change glass-substrate thicknesses from 1.1 to 0.7 mm because all of the cassettes andhandling systems would have needed to be changed. This illustrates the kinds of difficulties that should be expected if major changes are made to the size and weight of substrates in TFT factories. In the EPLaR process, the polyimide increases the substrate thickness by only about 1.4% and the weight of the substrate by less than 1%. These small changes allow EPLaR substrates to be used with all of the existing automated handling and mass-production equipment in TFT factories.

During processing, the polyimide layer is strongly anchored to the glass substrate. This means that the design rules for EPLaR displays can be kept exactly the same as for TFT arrays on glass, allowing use of the same mask sets for making glass and EPLaR displays. Figure 1(a) shows a photograph of a pixel of a glass electrophoretic display and Fig. 1(b) a pixel from a flexible display made by the EPLaR process. It can be seen that they are identical, except for the color, which comes from the thin polyimide. In comparison, flexible displays made on relatively thick pre-formed plastic substrates must use design rules that allow for shrinkage and swelling during processing because the substrate absorbs moisture and loses it during heating. For this reason, relatively thick pre-formed plastic substrates are generally not so well-suited for making high-resolution displays.


Fig1a_online_tif (a)    Fig1b_online_tif (b)

Fig. 1: These photographs of electrophoretic-display pixels were taken through the substrate: (a) on glass and (b) taken through a laser-released EPLaR polyimide substrate.


Electrophoretic foils, driver chips, and electrical interconnects are laminated onto the substrates after TFT-array fabrication with the same equipment and process settings that are used for glass electrophoretic displays. At this stage, there are fully working electro-phoretic displays on glass, but with a thin polyimide layer between the glass and the bottom of the TFT array. To make them into flexible displays, the second additional EPLaR step is needed. The polyimide is released from the glass substrate using a laser process, and the polyimide becomes the plastic substrate for the flexible display. Figure 2 shows part of a batch of 9.7-in. EPLaR displays immediately after laser release at the PVI module factory in Yangzhou, China.

Electrical Characteristics

Electrical characteristics measured on test TFTs from a standard glass substrate and an EPLaR display after laser release are shown in Fig. 3. The TFT characteristics are effectively identical and no significant difference in ON-currents, threshold voltage, sub-threshold slope, or OFF-currents are detectable. All of these characteristics are important because together they determine the drive voltages and timing needed to drive the display. For instance, TFTs with a small sub-threshold slope have a wider voltage difference between their OFF and ON states, so that more expensive driver chips with a larger voltage range are needed. The other TFT characteristics all affect the driver voltages and timing in some manner. The same drive electronics and drive schemes can be used for glass and EPLaR e-book displays because of their identical characteristics, which is a significant advantage to a company that is manufacturing both.

DC-stability measurements at elevated temperatures and high gate fields were also made on glass and EPLaR TFTs. Again, the results were identical for laser-released EPLaR TFTs and test devices on glass. The results show that TFTs in EPLaR displays will have the same performance, lifetime, and stability as a-Si TFTs in LCDs and e-books.



Fig. 2: These 9.7-in. flexible displays have just gone through laser release. The ones on the table and in the carriers are face-down while the one being held has the front face of the display toward the camera.



Fig. 3: The graph on the left shows the transfer characteristics of test a-Si TFTs on glass and the graph on the right, on an EPLaR display.


EPLaR Displays and Future Plans

All active-matrix displays that are already in mass production use glass as the TFT substrate. This includes TFT-LCDs, active-matrix OLEDs, and e-books. EPLaR displays are called flexible to indicate that they are not rigid in the same way as glass displays are; instead they are thin, light, and robust. Initially, EPLaR displays will be used in applications where they are held flat or curved to a fixed shape, rather than ones in which they are constantly rolled and unrolled. In the past, E Ink has described displays that have the feel and thickness of a mouse pad to convey the idea of a robust display that is normally used flat, but that can withstand a certain amount of bending. There is no reason why EPLaR displays should not be made rollable in the future, but this would require a redesign of the module, especially the placing of the rigid row and column drivers, and extensive mechanical testing.

PVI has made EPLaR displays with 1.9-, 6-, and 9.7-in. diagonals, which are three of the standard sizes the company already uses for glass electrophoretic displays. Figure 4 is a photograph of an EPLaR 9.7-in. display without a case, showing the image of a newspaper page.

Although the EPLaR process has been developed primarily for use in e-books and e-newspapers, it will also allow other applications to benefit from having thin, light, and robust displays. One obvious possibility is incorporating a small display into a smart card to show customers account information and maybe alternate between the customer's photograph and signature during payment. Glass displays cannot be used in smart cards because they are too thick and rigid, but EPLaR displays can be less than 0.4 mm thick. Small flexible displays could also be used as electronic shelf labels that do not break when knocked by a can or bottle, or for secondary text displays on mobile phones. Figure 5 shows a photograph of a 1.9-in. EPLaR display having drive electronics on a flexible PCB that has been attached to a shaped holder to demonstrate that it can conform to a curved shape. The image on the EPLaR screen has 16 gray levels, and it is easy to see that it could act as an ID photograph in an electronic badge. The electronics and drive scheme for this portable demonstrator were made by MpicoSys.3

EPLaR displays are currently being readied for commercial release, with production split between the PVI TFT factory in Taiwan and the module factory in China. The electrical characteristics and stability of the a-Si TFTs on EPLaR displays are the same as TFTs made on glass for LCDs, which have been used in many millions of devices, ranging from mobile phones to LCD TVs. EPLaR displays can be used in a range of products, such as e-books, e-newspapers, smart cards, and electronic shelf labels. In the future, they will benefit from planned improvements in electrophoretic displays on glass, such as the development of full-color displays and video rates, which will also be applicable to EPLaR displays.


1J. Ritter, "The Promise of 'Electronic Ink,' " Information Display 12/1, 22–25 (2001).

2I. D. French, et al., "Thin Plastic Electro-phoretic Displays Fabricated by a Novel Process," SID Symposium Digest 36, 1634–1637 (2005).

3http://www.mpicosys.com. •



Fig. 4: This 9.7-in. EPLaR display is flexible, though not rollable.



Fig. 5: This 1.9-in. EPLaR display in a curved holder can hold this shape indefinitely.


Ian French is principal scientist at PVI, concentrating on the industrialization of EPLaR. He worked at Philips Research for many years on a range of display and sensor applications. He can be reached at ianfrench@pvi.com.tw.