Electro-optic displays with reduced remnant voltage

Abstract

The invention provides materials and methods (including driving methods) for reducing the effects of remnant voltages in electro-optic displays.

Claims

1. An electro-optic display comprising a layer of electro-optic material and voltage supply means arranged to apply a voltage not greater than a predetermined value in either direction across the layer of electro-optic material, wherein the electro-optic material has a threshold voltage which is greater than zero but less than about one third of the predetermined value.

2. An electro-optic display according to claim 1 wherein the electro-optic material has a threshold voltage which is not less than about one-fiftieth but less than about one third of the predetermined value.

3. An electro-optic display according to claim 2 wherein the electro-optic material comprises a particle-based electrophoretic material comprising a suspending fluid and a plurality of electrically charged particles held in the suspending fluid and capable of moving therethrough on application of the voltage across the layer of electro-optic material.

4. An electro-optic display according to claim 3 wherein the electrophoretic material is an encapsulated electrophoretic material, a polymer-dispersed electrophoretic material or a microcell electrophoretic material.

5. An electro-optic display according to claim 3 wherein the suspending fluid is gaseous.

Description

BRIEF DESCRIPTION OF DRAWINGS

(1) As already mentioned,

(2) FIG. 1 of the accompanying drawings is a graph showing a typical curve of the decay of remnant voltage with time in an electro-optic display.

(3) FIG. 2 is a schematic side elevation showing the contact circle between capsules and surrounding liquid during the coating of an encapsulated electrophoretic medium.

(4) FIG. 3A is a schematic side elevation illustrating the forces acting on sparsely coated capsules during the coating of an encapsulated electrophoretic medium.

(5) FIG. 3B is a schematic side elevation, similar to that of FIG. 3A but showing the form of the capsules in the final dried capsule layer as a result of the forces illustrated in FIG. 3A.

(6) FIGS. 4A and 4B are schematic side elevations, similar to those of FIGS. 3A and 3B respectively, showing the forces acting on closely packed coated capsules and the form of the capsules in the final dried capsule layers.

DETAILED DESCRIPTION

(7) As already mentioned, the present invention provides several different improvements in electro-optic displays and media, and in waveforms and controllers for driving such displays. The various aspects of the present invention will be described separately (or in related groups) below, although it should be understood that a single display or medium may make use of more than one aspect of the present invention. For example, a single display may contain a volume resistivity balanced electrophoretic medium of the present invention and use a waveform selection method of the invention to drive this medium.

(8) Methods for Determining Remnant Voltages, and Addressing Methods and Controllers for Electro-Optic Displays which Exhibit Remnant Voltages

(9) As already indicated, in view of the deleterious effects of remnant voltages on the optical performance of electro-optic displays, when a display is subject to such remnant voltages it is typically necessary or desirable to use an addressing method that minimizes the impact of remnant voltage.

(10) For a given pixel of an electro-optic display, the state of the remnant voltage is greatly affected by the image history, i.e., the electric fields that have been applied previously, and is thus affected by parameters such as the waveforms used, the electric field intensity, and the elapsed times between successive image updates.

(11) One helpful class of addressing methods described in the aforementioned 2003/0137521 and Ser. No. 10/249,973 employs knowledge of the previous image data. A look-up table is employed in which, for example, the waveform for a black pixel being switched to white may be different, depending on whether the black pixel had previously been white, or had previously been gray (the transition from gray presumably being a different waveform that would have created a different amount of remnant voltage). Practically it has been found that such prior n-state lookup tables do tend to reduce ghosting attributable to remnant voltage.

(12) There are, however, several disadvantages to this approach. Firstly, while the previous optical states are tracked, in some cases the algorithm used does not take into account the elapsed time between each image transition (change of optical state), and as a result the values chosen for the look-up table must be selected with some usage model in mind, for example, an update on average once per second. Secondly, this method requires additional memory, and to achieve higher accuracy, the size of the look-up table must be increased and the amount of memory required goes up further, especially as n increases past 2 or 3. As discussed in the aforementioned applications, in some cases the very large look-up tables required may be difficult to accommodate in portable devices.

(13) In accordance with the waveform selection method of the present invention, an alternative approach is now proposed in which the remnant voltage of each pixel is first determined (or estimated by use of various parameters known to be related to remnant voltage), and thereafter one of two or more waveforms is selected based at least in part on the determined or estimated remnant voltage. Such a waveform selection method may make use of several possible approaches to estimating or predicting the remnant voltage based on known or measured display characteristics. A waveform selection method may also involve direct measurement of remnant voltage.

(14) In an exhaustive method, the complete update history of each pixel may be recorded including both voltage applied and the elapsed times between image updates. A decay model is used to forecast the remnant voltage remaining from each previous update. Updates that occur a sufficiently long time (typically about 10 minutes) before the transition being considered may be ignored and their history erased because their contribution to remnant voltage level has been reduced essentially to zero. The remnant voltage of the pixel may then be modeled as the aggregate of remnant voltages from each previous relevant update.

(15) In practice, a preferred approach requiring less memory is to track a single remnant voltage value and time stamp for each pixel. Prior to each image update, the remnant value for each pixel is reduced by an amount determined by the decay function of the display and the time stamp for the pixel is updated. After each update, the remnant voltage value is increased or decreased by an amount based on the actual waveform used, and the time stamp is updated. In this way, remnant voltage is tracked at all times but only two data values have to be stored per pixel.

(16) The decay functions and change functions may be calculated in any suitable manner, such as by logical computation based on a formula and data parameters, through analog logic device, or by a look-up table with adequate gradations for the display application. The actual updating of the stored remnant voltage and time stamp values may occur in any suitable manner, such as a single step combining the results of both calculations. If a waveform used for an image update comprises a series of pulses spread over a long period (e.g. 300-1000 msec), it may be advantageous to update remnant voltage and/or time stamp values at intervals during the image update itself.

(17) The decay function for a given display is highly sensitive to many factors, such as the materials, manufacturing methods, and system design features used in the display. Hence, it is necessary to vary the decay function or function parameter values for different electro-optic displays. In practice, due to the complexity of electro-optic media and displays, it has been found most useful to measure the display system experimentally for remnant voltage response and decay against a series of applied voltages and to thereby create a look-up table or to fit a function to the data. It may be helpful to repeat this measuring step periodically in the manufacturing process, for example when switching to new material set or making a batch change. It may also be helpful to individually characterize the decay function for each display after it is assembled, and to record the resulting parameters in a display controller.

(18) The remnant voltage decay and response function for a given display may be affected by environmental factors such as temperature and humidity level. A sensor or user-selectable input value may be added to the display (or device containing the display) to track such factors. Thus, it may be advantageous to use a remnant response and decay function or look-up table that allows for these environmental parameters. It may also be advantageous to update the remnant voltage value regularly (e.g. every 30-300 seconds) regardless of whether the display has been updated so that the stored values allow for environmental changes such as temperature and humidity, and remain accurate.

(19) If the pixels in a display are small, the overall remnant voltage on a pixel may be significantly affected by the remnant voltages of its neighbors. Accordingly, remnant voltage updating functions may be used that allow for lateral field effects, or a pre- or post-processing algorithm may be introduced to allow for such effects. Again, it may be useful to update the remnant voltage value periodically for each pixel in response to the remnant voltage values of its immediate neighbors, and to thereby achieve an adequately accurate estimate of actual remnant voltage values.

(20) The foregoing discussion has focused on methods for estimating the remnant voltage based on system inputs and characteristics. An alternative approach is to measure the remnant voltage directly. Techniques for sensing the state of an electrophoretic display are described in U.S. Pat. No. 6,512,354. Similar techniques may be used for sensing the remnant voltage in other electro-optic displays. The aforementioned 2003/0137521 specifically describes the use of comparators to measure remnant voltage. Direct measurement of remnant voltage may be performed prior to each image update or periodically to update and correct the data values described above.

(21) Estimating and direct measuring methods may be used together. For example, an estimating method may be used at each image update, but the remnant voltage value may be updated periodically based on actual measurement. Because the remnant voltage response and decay rate may change over the working lifetime of a display, which may be a period of years, it can be advantageous for display controller software to track such changes and to use an adaptive algorithm, for example a Bayesian algorithm, that updates its predictive parameters based on actual data.

(22) Similarly, an estimating method may be used for each pixel, the remnant voltage value may be directly sensed at one or more test pixels, and the remnant voltage value for the remaining pixels adjusted, at least in part, based on the difference between the estimated and measure remnant voltage for the test pixels. The test pixels may or may not be pixels visible to an observer of the display.

(23) In at least some cases, the remnant voltage characteristics of a display may be highly sensitive to the electrical properties of one or more specific layers of the display, for example, an adhesive layer. Accordingly, the foregoing approaches to estimating or measuring remnant voltage may be modified to estimate or measure the electrical characteristics of the specific layer (such as the adhesive layer) having a major effect on remnant voltage characteristics, and to modify the algorithms for the remnant values of the pixels appropriately. A sensor may be used to probe the specific layer of the display and may or may not probe material associated with visible pixels. Furthermore, a physical sample of the material of the relevant layer may be provided outside the display as part of a sensor that incorporates the material directly to measure its responses and changes over time.

(24) Remnant Voltage Aware Waveforms and Addressing Methods

(25) Having estimated or measured the remnant voltage (or a proxy variable) by the above or any other suitable methods, in accordance with the waveform selection method of the present invention, an addressing method is selected based at least in part on the estimated or measured current remnant voltage or proxy variable. The addressing method may be chosen based on the remnant voltage of a specific pixel, the remnant voltage of the pixel and the surrounding pixels, or on the overall remnant voltage across all or a portion of the display larger than one pixel and its immediate neighbors.

(26) Various methods may be used to modify a standard waveform (i.e., a waveform which is not remnant voltage aware) to allow for remnant voltage of a specific pixel or group of pixels. For example, the remnant voltage may be subtracted from the desired waveform and a reduced voltage applied, so that the effective waveform experienced by the pixel is the original desired waveform. Alternatively, a scaling factor or other transformation may be applied to the waveform. Alternatively, the voltage levels in the waveform may be held unchanged, but their durations may be adjusted. For example, if a standard waveform requires a 10 V 50 ms pulse, but the pixel has a 2 V remnant voltage, the pulse may instead be 40 ms at 10 V, 50 ms at 8 V, 44.7 ms at 8.94 V, or even two pulses of 20 ms at 10 V with an intervening pause of 10 ms at 0 V (for simplicity these examples do not take account of the decay rate of the 2 V remnant voltage and could be adjusted more precisely to match the expected remnant impulse decline from an initial value of 2 V). These waveforms may also be adjusted to allow for the fact that it is not necessarily ideal for the net impulse to be exactly constant, since the electro-optic medium may have a slight threshold or be otherwise asymmetric in its optical response to the voltage or duration of a pulse in a waveform.

(27) Direct calculation of waveform adjustments in this manner may impose a significant overhead on the display controller. To reduce such overhead, the controller may instead select an addressing waveform, algorithm, formula or look-up table from a series of options, each associated with a range of remnant voltage values. Thus, the waveform selection method of the present invention extends to the selection of from among two or more essentially equivalent waveforms (waveforms that do not substantially differ in the final optical state of the pixel after the waveform has been completed) so as to minimize the change in aggregate remnant voltage within a pixel (i.e., to yield a very low remnant voltage waveform). The optimum waveform may be determined by modeling the decay rates of the electro-optic medium or by direct experimentation and a process of tuning and optimization for the waveforms. The waveform selection method of the present invention also extends to selecting among essentially equivalent waveforms that generate equivalent or non-minimal remnant voltages, choosing the waveform that brings the net remnant voltage of a given pixel closer to zero; such a waveform may be called an off-setting remnant voltage waveform.

(28) As described in the aforementioned Ser. No. 10/814,205 and Ser. No. 10/879,335, it is possible, and often desirable, to use drive schemes in which the waveforms used for individual transitions are DC-balanced (as opposed to the overall drive scheme being DC-balanced). In other cases, specific parts of a waveform may be DC-balanced even if the entire waveform is not DC-balanced; examples are shake-up pulses, blanking pulses (see below), and many rail-stabilized addressing methods. During such DC-balanced waveform sequences, as part of a sequence which involves a pixel being driven to both its extreme optical states (hereinafter assumed for convenience to be black and white) the controller may select in which direction to switch first, towards white or towards black. When a switch to one extreme optical state is followed by a switch to the other extreme optical state, the second switch will normally have a greater impact on remnant voltage simply because it occurs later in time and the effects of remnant voltage decay with time. Thus, in a DC-balanced waveform sequence, selecting whether to switch towards black or white first can determine whether the remnant voltage for a given pixel is slightly increased or decreased. This is another example of an offsetting remnant voltage waveform of the invention.

(29) Some drive schemes require a periodic (typically every 10 minutes or so, or every image update) blanking pulse which drives a pixel to both extreme optical states; see, for example, the aforementioned 2003/0137521. For example, the blanking pulse can switch the display to all-white then all-black, or to all-black then to all-white. In accordance with the waveform selection method of the present invention, a choice between these alternatives can be made to reduce remnant voltage and thus to reduce perceived ghosting. Alternatively, by determining whether the pixels of the display have, overall, positive or negative remnant voltages, the total remnant voltage on the display may be reduced by choosing the appropriate blanking sequence (black/white or white/black) without increasing the image update time. In a variation, the decision as to which optical rail (extreme optical state) to hit first is not based on the aggregate remnant voltage but is based on the number of pixels that have high remnant voltage in either direction. More generally, any suitable algorithm may be used to determine to which rail the medium will be driven first in order to minimize outliers or other distracting visual artifacts of the display caused by remnant voltages, given the user preferences for the targeted application.

(30) If desired, the algorithm may also provide for introducing an additional blanking sequence (white-black-white or white-black-white-black) when the remnant voltages are extreme in one or both directions. It will be apparent that the voltage level of the blanking pulse on each pixel could be modified instead of its duration.

(31) The waveform selection method of this invention also extends to extending the period of a voltage pulse during the time when the electro-optic medium is already in an extreme optical state (i.e., is at an optical rail), thereby increasing or decreasing remnant voltage without a distracting optical change. An opportunity for such voltage pulse extension exists every time a pixel is brought to an extreme optical state. The blanking pulses mentioned above are one example. The waveform selection method therefore provides for the blanking pulse duration (or voltage) to be varied on a pixel-by-pixel basis. By lengthening the pulse in either direction on a calculated basis for each pixel, a net remnant voltage component may be applied, and the total remnant voltage for that pixel thereby reduced or eliminated. Thus, a blanking pulse could be used to reduce remnant voltage across all pixels of the display without apparent optical impact. As a practical matter, the degree to which the pulse is lengthened could be quantized, i.e., the pixels could be grouped into categories based on remnant voltage ranges and the same adjustment applied to all the pixels in each category.

(32) So-called rail stabilized drive schemes are known (see, for example, the aforementioned 2003/0137521, Ser. No. 10/814,205 and Ser. No. 10/879,335), which allow any given pixel to undergo only a limited number of transitions without touching an optical rail, and thus provide for each pixel to be switched to one of its extreme optical states on a frequent basis. For example, to transition a pixel from one gray level to another gray level, the pixel may be switched first to either a dark or a white state (possibly for an extended period) and then a subsequent pulse applied to reach a desired gray level. Such a transition tends to create positive or negative remnant voltage by virtue of the long period in which the pixel is addressed toward the extreme optical state. According to the waveform selection method of the present invention, the remnant voltage of the pixel may be minimized by causing the transition to use the optical rail for which the remnant voltage created by the direction of the switch will tend be opposite in sign to the remnant voltage carried by the pixel just prior to the transition.

(33) One reason for using a rail-stabilized drive scheme is to mitigate the optical effects of remnant voltage. The estimate of measurement of remnant voltage on a pixel-by-pixel level, as described above, can reduce the need for the use of such rail-stabilized drive schemes. A hybrid approach is to use rail-stabilized methods for pixels in which remnant voltage is fairly high, but to switch directly to the desired state (a direct impulse method) when remnant voltage is low and would not affect the image.

(34) Another approach to reducing remnant voltage is to identify pixels for which the remnant voltage is extreme (i.e., has a magnitude greater than some predetermined value), and, prior to a general image update, to apply an off-setting voltage to such pixels to reduce their remnant voltages, or otherwise pre-condition the remnant voltage levels across the display. Such pre-conditioning may enable a reduced period of rail stabilization and achieve a faster perceived image update time. If the off-setting voltage is small and applied over a sustained time, or if it tends to lengthen a period of rail stabilization rather than pull the particles back from the rail, the reduction of remnant voltages may be accomplished without distracting visual impact.

(35) The above discussion has focused on tracking net remnant voltages and selecting appropriate algorithms for reducing remnant voltages. Another parameter of image history of a pixel or display is net DC imbalance. It will be apparent to one skilled in the imaging art that most of the methods described above can be modified to track and correct for net DC imbalance, either in combination with, or independently from, any adjustments for remnant voltage. For example, DC imbalance can be used when determining which optical rail to select first and what pre-conditioning is appropriate in the above approaches. Also, for example, even when remnant voltage may be reduced by modifying a waveform, the drive scheme could omit this correction if the pixel is already DC-imbalanced and would become more imbalanced after the adjustment for remnant voltage. Similarly, any conditioning of the display or adjustment of waveforms to achieve a net DC balance would be allowed for when estimating the remnant voltage across each pixel.

(36) Thus, the waveform selection method of this invention may be generalized as an addressing method for an electro-optic display capable of exhibiting a remnant voltage, wherein a data value corresponding to remnant voltage is determined and an addressing waveform is selected at least in part based on the remnant voltage value. In such a method, time and remnant voltage values, or data representing each, are typically explicitly tracked. However, it should be recognized that addressing waveforms for electrophoretic and other electro-optic displays may account for time and remnant voltage values implicitly or approximately. For example, the so-called prior n-state addressing methodologies described the aforementioned 2003/0137521, Ser. No. 10/814,205 and Ser. No. 10/879,335 and above may not track time, but they do track a history of prior pixel optical states, and this can be a proxy for time if the drive scheme designer has some knowledge of typical usage models and common elapsed times between image updates. Hence, it is now recognized that such methods tend to reduce remnant voltage and thus show improved ghosting behavior.

(37) One practical reason why such methods have previously been used is that the display controller in many electro-optic displays does not have access to clock information to track elapsed time between image updates, perhaps because such elapsed time data is most useful for bistable displays and few bistable displays have hitherto been commercialized. In a preferred form of the waveform selection method of the present invention, the controller does comprise a clock or equivalent timing mechanism. Alternatively, the controller may be in logical communication with an external information source (such as the device which uses the display as its output device) that generates an elapsed time value and provides this information to the controller. For example, the device may provide time information along with a function call to the display controller or along with each new set of image data. Such time information may be quantized (e.g. immediate, 0.5 second, 1 second, 2 seconds, 10 seconds, 30 seconds, 60 seconds, more than 60 seconds) thereby reducing data bandwidth and yet still providing useful information, especially if the quantized time bands are chosen to correspond to the substantially exponential decay of the remnant voltage.

(38) In general, it is most useful for the controller to receive elapsed time data for each pixel, since some pixels may not change during an update. However, it is still useful for the controller to receive data corresponding to the elapsed time since the most recent image update, most recent blanking pulse, or most recent update for a set of pixels. Additionally, the controller may receive data indicating the likely update frequency of the display, for example, a flag indicating whether the user is currently entering text, which may require many updates in rapid succession to the whole display or to a defined region thereof.

(39) Another form of approximate correction of remnant voltage is used in the dwell time waveform selection method of the present invention, which provides for choice among multiple waveforms to effect an image transition, where the selection among the multiple waveforms is based at least in part on the dwell time of the relevant pixel in its initial gray state, or some proxy for this dwell time. Such time-sensitive selection among multiple waveforms implicitly accounts for the decay of remnant voltage with time, even though remnant voltage is not explicitly tracked, estimated or measured.

(40) For example, a specific dwell time waveform selection method of the present invention might be applied to a controller for a display with four gray levels using a drive scheme based on a logical transition table with 16 entries, each entry corresponding to the transition from one gray level (0,1,2,3) to another (0,1,2,3). Selection of the entry is based on knowledge of the initial and final gray levels of the desired transition. Within each entry, there are three possible waveforms. The controller selects the first waveform when the image transition occurs within 1 second after the prior image update, the second waveform when the image transition occurs between 1 and 5 seconds after the prior image update, and the third waveform when the image transition occurs more than 5 seconds after the prior image update.

(41) In the dwell time waveform selection method, the waveforms may be represented by look-up tables (as described above), may be modified (or split into sub-tables) to allow for variation in environmental conditions, and may be set in whole or in part during manufacture of the display to include specific parameters of an individual display. In short, the waveforms used in this method may include any of the optional components and variations described in the aforementioned 2003/0137521, Ser. No. 10/814,205 and Ser. No. 10/879,335.

(42) From the foregoing it will be seen that, although in the dwell time waveform selection method of the present invention remnant voltage is not explicitly tracked, and although the elapsed time may be based on elapsed time since the display was updated and not on elapsed time since a specific pixel was updated, the dwell time waveform selection method does implicitly approximate both remnant voltage and elapsed pixel update time and therefore exhibits improved ghosting behavior over prior art drive schemes.

(43) Materials Selection

(44) As already indicated, the selection of materials for use in electro-optic displays can have a major influence on the remnant voltages which exist in such displays during their operation, and hence upon the electro-optic performance of such displays.

(45) Also as discussed above, when used in an electro-optic display certain materials exhibit Type I polarization which contributes to remnant voltage. It is believed (although the invention is in no way limited by this belief) that this polarization is frequently due to the mobility and concentration of ions moving through at least one of the component materials.

(46) The speed of decay of remnant voltage may be measured in any specific material by preparing a test cell in which the material is in contact with the same interfaces as in the proposed display. For example, test cells have been prepared consisting of a controlled thickness of laminating adhesive coated onto an ITO substrate, and an electric field applied across the laminating adhesive/ITO interface. Remnant voltage peak values and decay were then measured by opening a charging circuit, and monitoring the voltage across the pixel with a high impedance voltmeter.

(47) It has been found that laminating adhesives with higher ionic mobility show faster remnant voltage decay rate. A preferred lamination adhesive has a volume resistivity of less than about 10.sup.11 ohm cm.

(48) Previous E Ink patent applications, for example the aforementioned U.S. Pat. No. 6,657,772 and Patent Publication No. 2003/0025855, and application Ser. No. 10/708,121, filed Feb. 10, 2004 (Publication No. 2004/0252360, now U.S. Pat. No. 7,110,163), describe lamination adhesives with controlled resistivities, or which are heterogeneously or anisotropically conductive, for example Z-axis adhesives. Such adhesives may provide a further benefit of reducing remnant voltage.

(49) A lamination adhesive may also exhibit Type II polarization. In test cells, increased adhesive thickness has been found to be associated with higher remnant voltage. Since polarization at the interfaces should be independent of the film thickness, this result suggests the existence of internal charge polarization sites, characteristic of Type II polarization effects. Consequently, care must be taken in the selection of the adhesive thickness and, in the case of encapsulated electrophoretic displays, its morphology around the capsules. The same test lamination adhesive was heated to drive out suspected impurities and crystalline regions. Thereafter it exhibited reduced remnant voltage.

(50) Type I polarization may occur anywhere in the display where a material interface exists. It has been found that, by using the same material for lamination adhesive and binder (i.e., the material used to surround the capsules and form them into a cohesive layer, as described in many of the aforementioned E Ink and MIT patents and applications), an interface is eliminated and the remnant voltage is reduced. Therefore, the present invention provides an electrophoretic display comprising a microcavity binder and a laminating adhesive in which the materials are either identical or similar in composition or electrically equivalent in conductivity or ionic mobility. In some cases where the materials are different in composition, it may be desirable to dope the less conductive material to achieve substantially equal ionic mobility on both sides of the interface.

(51) Type I polarization at some interfaces can be affected by surface roughness. It may be advantageous either to planarize or to introduce a texture to some interfaces, thereby providing a degree of interpenetration of the materials on either side of the interface. These techniques may result in either increased polarization at the specific interface or in decreased polarization, either of which could be beneficial depending upon the specific display being considered. For example, increased polarization at one location that offsets polarization elsewhere in the display may cause a reduced remnant voltage across an electro-optic medium. Typically if the interface results in a remnant voltage that is strongly coupled to the electro-optic medium, then reducing the degree of polarization at the interface and its decay rate is desirable.

(52) Type I polarization at some surfaces may also be affected by surface cleanliness. Cleaning of substrates prior to coating and lamination is desirable in order to achieve consistent electrical behavior.

(53) Conductive Paths in Electrophoretic Layer

(54) In microcavity electrophoretic displays, a cell wall (a term which is used herein to include the capsule wall of an encapsulated display) exists that is electrically in parallel with the electrophoretic internal phase (the suspending fluid and the electrically charged particles). Current, in the form of electrically charged ions, can flow through the internal phase or through the cell wall. The cell wall can be a polymer, such as a gelatin, or any other suitable material. The cell wall is typically further surrounded by a binder, as mentioned above. Therefore, some current may flow between the electrodes of the display via the binder or cell wall without flowing through the electrophoretic internal phase, and thus without contributing to changes in the electro-optic state of the display or a pixel thereof.

(55) In the preferred electrophoretic displays described in the aforementioned E Ink and MIT patents and applications, the conductivities of the cell wall and binder are typically slightly higher than those of the internal phase. Relaxation of remnant voltage may thus occur in part through the binder and cell wall.

(56) During the application of an electric field to the electrophoretic medium, charged particles move toward the two electrodes of the display. If charged particles cluster near the front electrode (the electrode through which an observer normally views the display) for a period of time, corresponding electrons or oppositely charged ions may flow through the cell wall and/or binder in response. The charged regions thus created may create a remnant voltage that affects a subsequent image update. Consequently, the conductivities and ionic mobilities of the cell wall and binder are of importance, as are their morphologies.

(57) The remnant voltage of a specific cell/binder morphology may be measured by methods similar to those described above for a lamination adhesive. In accordance with the volume resistivity balanced electrophoretic medium aspect of the present invention, it is preferred that the binder and cell walls have a volume resistivity at least two times less than the volume resistivity of the electrophoretic internal phase and that both have a volume resistivity of less than about 10.sup.11 ohm cm. More generally, in an electrophoretic medium comprising a plurality of discrete droplets of a suspending fluid dispersed in a continuous phase (which may have the form of a single continuous phase in a polymer-dispersed medium, a combination of cell walls and binder in an encapsulated electrophoretic medium, or cell walls only in a microcell electrophoretic medium), the droplets comprising a plurality of electrically-charged particles held in a suspending fluid and capable of moving therethrough on application of an electric field to the electrophoretic medium, it is preferred that the continuous phase have a volume resistivity not greater than about one-half of the volume resistivity of the droplets, and that both the continuous phase and the droplets have a volume resistivity of less than about 10.sup.11 ohm cm. In a preferred embodiment, the binder and cell walls occupy between about 5 and about 20% by volume of the electrophoretic layer (with the remainder being the electrophoretic internal phase), and the binder is evenly distributed among the capsule walls.

(58) Zeta Potential Considerations, and Charge Balanced Dual Particle Electrophoretic Medium

(59) A preferred type of electrophoretic medium described in many of the aforementioned E Ink and MIT patents and applications is a so-called opposite charge dual particle medium, in which the electrophoretic internal phase contains two different types of particles bearing charges of opposite polarity (see, for example, the discussion of the different types of electrophoretic media in the aforementioned 2002/0171910). The amount of charge on each particle may be controlled, for example by surface modification as described in U.S. Patent Publication No. 2002/0185378 (now U.S. Pat. No. 6,870,661), and in copending application Ser. No. 10/711,829, filed Oct. 7, 2004 (now U.S. Pat. No. 7,230,750). The number of particles in each microcavity may also be controlled in a predictable way by selecting the total amount of particles provided in the electrophoretic internal phase prior to encapsulation or filling of microcells. By multiplying the average charge per particle by the average number of particles per microcavity, it is possible to estimate the total charge of each type of particle in the microcavity.

(60) It has been found that if the total charges of the oppositely charged types particle are not approximately balanced, a particularly large polarization is produced in the polarized microcavity, which induces a corresponding large and slowly decaying polarization in the continuous phase material(s). It has further been found that by varying the net total charge of the particle types, it is possible to vary an encapsulated electrophoretic display between a regime in which an electric field leaves a remnant voltage of the same sign (so that a subsequent update in the opposite direction is retarded), a regime in which very little remnant voltage occurs, and a regime in which an electric field leaves a remnant voltage of the opposite sign (so that a subsequent update in the opposite direction is promoted).

(61) In accordance with the charge balanced dual particle aspect of the present invention, it is preferred that neither type of electrophoretic particle have more than about twice the total charge of the other. It is also preferred, in accordance with the low remnant voltage electrophoretic medium aspect of the present invention, that, in an opposite charge dual particle electrophoretic display, the particle charge, particle mass, and particle mobility be selected so that the display exhibits Low Remnant Voltage Behavior, herein defined as having a remnant voltage measuring less than about 1 V (and desirably less than about 0.2 V) exactly 1 second after the application thereto of a square wave DC addressing pulse of 15 V for 300 milliseconds.

(62) To assess charge balance in an opposite charge dual particle electrophoretic internal phase, it is helpful to analyze the charge on each particle relative to its mass (since mass can be easily measured at the time of manufacture). It is believed, although the invention is in no way restricted by this belief, that the charge to mass ratio may be estimated using the following relationship:
q/M is proportional to /d.sup.2(1)
where: q=particle charge M=mass =zeta potential (mV) d=particle diameter.

(63) The total net charge of the electrophoretic internal phase should desirably be controlled by careful co-optimization of particle charge, particle mass, particle diameter, and zeta potential.

(64) In a substantially charge-balanced electrophoretic medium exhibiting Low Remnant Voltage Behavior (as defined above), such behavior may typically cease if any of the following occurs: (a) the average charge on either type of particles is changed by about 20% to 100%; (b) the relative mass of one type of particle is changed by about 50% to 300%; (c) the average diameter of one type of particle is changed by about 30% to 200%; and (d) the average mobility of one type of particle is changed by about 20% to 100%.

(65) Suspending Fluid Additives

(66) It has been found that the addition of surfactants to the suspending fluid of the electrophoretic medium may reduce remnant voltage. For example, when single pixel displays were prepared using otherwise identical dual particle opposite charge electrophoretic media but in which sorbitan trioleate (sold commercially as Span 80) was added to one of the suspensions, the display containing the sorbitan trioleate demonstrated reduced remnant voltage.

(67) It is believed, although the invention is in no way restricted by this belief, that the surfactant alters the relative charge balance of the two types of electrophoretic particles. It is further believed that the surfactant reduces Type III polarization by modifying charge relaxation rates in the electrophoretic internal phase so that they more closely balance the corresponding relaxation rates in the external phase.

(68) Thus, this invention provides an electrophoretic display exhibiting Low Remnant Voltage Behavior (as defined above), which behavior ceases if the concentration of a surfactant or charge control agent in the electrophoretic internal phase is changed by about 30% to 200%.

(69) Materials for External Phases of Microcavity Electrophoretic Displays

(70) It is possible to select external phase materials for use in microcavity electrophoretic displays, or to mix, dope or condition such materials, to achieve desired remnant voltage relaxation rates. As described above, the relaxation rate of an internal phase may be affected by numerous factors, including the choice of electrophoretic particle(s) and the concentration of surfactants and charge control agents. One aspect of the present invention provides that the external phase materials and the internal phase materials be balanced (within a factor of 2) in relaxation rates.

(71) This aspect of the invention provides an electrophoretic display exhibiting Low Remnant Voltage Behavior (as defined above), which behavior ceases if the conductivity of the external phase materials is changed by about 30% to 200%.

(72) In a typical encapsulated electrophoretic display, a critical external phase material is the gelatin capsule wall. The conductivity of the wall is significantly affected by moisture. In a preferred embodiment, the electrophoretic display comprises moisture and is resistant to changes in the relative humidity (RH) of the operating environment. In a further preferred embodiment, the display is conditioned (by placing it in a controlled humidity environment until it has come to equilibrium and/or by manufacturing the display in a controlled humidity environment) so as to achieve between 20% RH and 55% RH, and preferably 35% RH, for the electrophoretic layer within the final display.

(73) Thus the invention provides a method of manufacturing an electrophoretic display that comprises RH conditioning the display material. The electrophoretic display may also comprise moisture barriers or substrates that are impermeable to water.

(74) Low Threshold Electro-Optic Displays

(75) A small threshold in an electrophoretic or other electro-optic display may be produced in many ways. The threshold can result from attractions between particles and walls, or among particles. The attractions can be electrical, such as from oppositely charged particles; physical, such as from surface tension; or magnetic. A threshold can also result from the nature of a suspending fluid, which may be strongly shear-thinning, or have an apparent yield stress (such as for a Bingham fluid), or electro-rheological properties. An additional electrical field, for instance a field created by in-plane electrodes or a control grid, can substitute for a threshold.

(76) For purposes of this application, a threshold is considered present at a particular voltage level when a square wave DC pulse of 1 second duration applied to the display at that voltage level results in an optical change of less than 2L*.

(77) It is known in the display art that a threshold in an electro-optic medium can serve as a basis for a passive addressing scheme. Typically such a scheme relies on a threshold equal to half of the switching voltage (V/2); in some drive schemes, passive addressing can be achieved with a minimum threshold of one-third of the switching voltage (V/3).

(78) In contrast, as described above, a threshold of as little as 1 V, as compared with a switching voltage of 15 V, can be useful in reducing the impact of remnant voltages on electro-optic performance. Accordingly, the low threshold display aspect of the present invention provides an electro-optic display operating at a voltage not greater than V, wherein the electro-optic material has a threshold voltage which is greater than zero but less than about V/3.

(79) Manufacturing Electrophoretic Displays with Reduced Remnant Voltage

(80) A final aspect of the present invention relates to various improvements in the manufacture of electrophoretic displays to reduce the remnant voltages exhibited by the displays thus manufactured.

(81) During the manufacture of encapsulated electrophoretic displays, capsules are typically suspended in a slurry, which comprises the capsules and a polymeric binder, and may also comprise various additives, for example water, plasticizers, pH adjusters, biocides, and surfactants or charge control agents. For present purposes, such a slurry may be regarded as containing a binder consisting of the nonvolatile components of the slurry, excluding the capsule. In some cases, the binder materials may separate during preparation of the slurry, or during shipment and storage, and may not always be mixed adequately prior to coating. As a result, regions of area-to-area heterogeneity may exist that can cause Type II polarization problems in the final display. To reduce such problems, it is desirable to thoroughly mix such binder materials through appropriate means such as mixing by propeller blade or on a roll mill for extended periods.

(82) The dried binder material should desirably have uniform electrical characteristics such that, following a 15 V voltage pulse applied for 300 ms and a 1 second pause, the measured remnant voltage of the binder material itself should be less than about 1 V, and preferably less than 0.2 V.

(83) As mentioned above, it is desirable to control the amount of space between capsules that is occupied by binder because this space can contribute to Type III polarization. Electrodeposition may be used to control capsule spacing directly, as described in copending application Ser. No. 10/807,594, filed Mar. 24, 2004 (Publication No. 2004/0226820; see also the corresponding International Application PCT/US2004/009421, Publication No. WO 2004/088002). In microcell or photo-patterned electrophoretic displays, microcavity spacing can be controlled directly.

(84) In coated encapsulated electrophoretic displays, dried capsule spacing and morphology are the result of many controllable factors, as discussed in several of the aforementioned E Ink and MIT patents and applications. To summarize, capsule morphology can be adjusted by varying capsule wall thickness and elasticity, the formulation of the coating slurry, the surface energy of the coating substrate, the height of the coating die off the substrate, the amount of coating slurry passing through or pumped through the die onto the substrate, the speed of a substrate web, and the drying conditions of the wet coated film such as temperature, duration and air flow. Useful principles for control of capsule spacing and morphology are described below.

(85) A: Capsule Wall Property Effects on Dried Capsule Shape

(86) Capsule wall properties vary with materials and process variables of encapsulation, especially mixing speed. The capsule wall should desirably be elastic enough to allow an overall capsule height/diameter ratio between 0.33 and 0.5. However the capsule wall should also ideally permit local variations enabling nearly a 90-degree bend radius on sharp corners for hexagonal close packing of the capsules on the substrate on to which they are coated, as described for example in U.S. Pat. Nos. 6,067,185 and 6,392,785.

(87) It is believed (although the invention is in no way limited by this belief) that capsule wall elasticity can be affected by the degree of cross-linking of the capsule wall material (less cross-linking typically giving a more flexible capsule wall) and by the thickness of the wall. Wall thickness is affected by internal phase formulation, gelatin/acacia levels and process parameters. For a given capsule packing pattern, reducing the wall thickness can improve the aperture ratio (i.e., the fraction of the area of the electrophoretic medium which undergoes change of optical state; the areas occupied by the capsule walls cannot undergo such change) of the medium; however, walls that are too thin may burst easily.

(88) Certain process parameters that have been found important in affecting wall thickness are set out in the Table below. The results shown in the Table were generated through encapsulation experiments done at a 4 L scale. Also shown in the Table are the relative qualitative rankings for the wall thickness compared to a standard operating procedure for encapsulation. The standard process conditions for the 4 L encapsulation are acacia level (index at 100% of standard level), pH (4.95), emulsification temperature (40 C.), cooling rate (3 hours), and rate of internal phase addition. In the Table, a Rank of 3 denotes walls of standard thickness, with 1 denoting a very thin wall and 5 a thick wall.

(89) TABLE-US-00001 TABLE Effect on wall thickness Rank Rank Parameter Low High Acacia: 25% variation in mass 3 1.5 pH: 3% variation on pH scale 3 2 Emulsification Temp: 10% variation 4 3 Cooling Rate: 2 hours 2 4 Rate of IP addition: Spray/Dropwise 2.5 3.5

(90) pH is a critical parameter for the wall properties, not just in terms of the wall thickness but also because the solid content and viscosity of the coacervate are quite different at different pH levels. Finally, the type of gelatin and acacia used may have a dramatic impact on the wall properties.

(91) B: Binder Evaporation as a Mechanism for Changing Dried Capsule Shape

(92) The effect of binder evaporation varies depending on how closely packed the capsules were while the coated slurry was still wet. The same binder ratio, with the same capsule diameters, may result in either flattened (oblate ellipsoidal) or tall (substantially prismatic) capsules according to wet capsule proximity.

(93) FIG. 2 shows the situation where a capsule/binder slurry has been coated on a substrate 110 so that the capsules 112 are sparsely coated, i.e., are separated from one another by gaps comparable to the diameter of the capsules 112. The capsules 112 comprise electrically charged particles 116 in a suspending fluid 118. As shown in FIG. 2, in these circumstances the capsules 112 are only partially immersed in the uncured binder 114, so that the portions of the capsules 112 remote from the substrate 110 protrude from the layer of binder 114, and the boundary between the binder and each capsules is a circle of radius r.sub.c around each substantially spherical capsule. It can be shown that the downward force exerted on a capsule by surface tension forces during drying is given by:
F=2r.sub.c sin .sub.c
where: F is the downward force on the capsule; is the surface tension of the liquid surrounding the capsule; r.sub.c is the radius of the contact circle which the liquid makes with the capsule, as illustrated in FIG. 2; and .sub.c is the complement to the contact angle of the liquid surrounding the capsule (i.e., 90the contact angle).

(94) From FIG. 2, it will be seen that r.sub.c will increase as the capsules increase in size and as the level of the surrounding liquid is lowered.

(95) Two extreme cases of the effects of this downward force are shown in FIGS. 3A-3B and 4A-4B respectively of the accompanying drawings. In FIGS. 3A and 4A, arrows A denote evaporation of water from the wet binder, while arrows B denote the forces on the capsules exerted by surface tension. In FIGS. 3B and 4B, C denotes dried binder. (Note that throughout FIGS. 3A to 4B, the presence of binder outside the two illustrated capsules is ignored.) FIGS. 3A and 3B illustrate the effects of the downward force on sparsely coated wet capsules, i.e., capsules coated so as to leave gaps between adjacent capsules which are a substantial fraction of a capsule diameter. From FIGS. 3A and 3B, it will be seen that the effect of the downward force is to flatten the original spherical capsules into oblate ellipsoids, which typically will touch each other in the final dried layer, but that little or no distortion of these ellipsoids occurs by contact between adjacent capsules. In contrast, FIGS. 4A and 4B illustrate the effects of the downward force on closely packed coated wet capsules, in which the wet capsules as coated are in contact with one another. From FIGS. 4A and 4B, it will be seen that the effect of the downward force is to force the capsules to contact each other over progressively larger areas, so that in the final dried layer the capsules have substantially the form of polygonal prisms having a height substantially greater than their width; if the wet capsules are hexagonally close packed, as is ideally the case, the dried capsules will have substantially the form of hexagonal prisms. It should be noted that if capsules are too sparsely packed, voids may be left between the capsules as they dry and tend to associate into clusters.

(96) FIG. 4B illustrates the capsules in an electro-optic display formed by placing the capsules between electrodes 120 and 122, a voltage supply means V being connected between the electrodes 120 and 122 for applying a voltage not greater than a predetermined value in either direction across the layer of capsules. Particles 124 bearing charges of one polarity are attracted to electrode 120 while particles 126 bearing charges of the opposite polarity are attracted to electrode 122.

(97) C: Slurry Preparation Effects on Dried Capsule Shape

(98) The pH level of the capsules may affect dried capsule shape. As pH is increased, the charge on gelatin changes, affecting the attraction of the gelatin to the substrate (typically an ITO surface) on which the capsules are coated and making it harder or easier for capsules to shift location. Choice of substrate surface energy by changing substrates may affect this relationship.

(99) Surfactant level affects the adhesion (stickiness) of capsules to each other and possibly to the binder. Amore surface active surfactant weakens surface tension and should reduce the surface tension forces acting on the capsules during drying. A less surface active formulation may help flatten capsules.

(100) Binder ratio is a critical factor affecting dry capsule shape. Lower binder ratios result in rounder capsules. A binder ratio of 2:1 (i.e., two parts by weight of capsules to one part by weight of binder) is sufficient to fully surround each capsule as a perfect sphere when dried and hence results in the least flattened capsules. Lower binder ratios allow the binder, when dry, to fill the interstices between capsules. A binder ratio of 8:1 is adequate to achieve flat capsules or heightened (polygonal) capsules depending on coating conditions.

(101) D: Coating Parameter Effects on Dried Capsule Shape

(102) As described above, the coating process should optimally deposit wet capsules a predetermined distance apart. Critical parameters in achieving such desired spacing include coating speed, die type, die height, and slurry flow rate.

(103) Experimentally, it has been found that increasing the slurry flow rate in a coater, while holding all other parameters constant, tends to increase coating weight, with the result that the wet capsules are placed closer together. This could result in some capsules being too closely spaced, resulting in less flattened/more heightened dried capsules.

(104) In related coating experiments, decreasing the gap in the die to a low value (e.g., 40-50 m) brought the die height to a value comparable to the size of the wet capsules being used. At this die height, a monolayer of capsules was virtually assured, but the packing was usually very tight. Coatings at lower die heights tended to result in heightened dried capsules. Ideally, the capsules would be coated almost touching together but not packed together when wet.

(105) E: Drying Parameter Effects on Dried Capsule Shape

(106) Experimentally, it has been found that a conveyor oven drying at 60 C. for 2 minutes is able to create capsule-containing films with flattened, heightened and spherical dry capsules. Attempting to dry the capsules too quickly may cause a skin to form on the top of the binder; this skin traps moisture within the binder and causes the film to dry very slowly.

(107) The rate of air flow while drying affects the rate of evaporation and affects whether evaporative gases are trapped among the capsules. It is difficult to achieve good success in drying without adequate ventilation; air flow across the binder is helpful.

(108) The foregoing description has emphasized the application of the invention to electrophoretic displays. Such electrophoretic displays may be of any type and still benefit from at least some aspects of this invention. Thus, the displays may include microcavity electrophoretic displays, such as encapsulated, microcell, microcup, and polymer-dispersed displays; electrophoretic displays using one or more species of particles (except of course for those aspects of the invention specific to dual particle electrophoretic displays); electrophoretic displays using clear or dyed suspending fluids; electrophoretic displays comprising oil-based and gaseous suspending media; flexible and rigid electrophoretic displays; electrophoretic displays addressed by non-linear devices (such as thin film transistors), by passive means (such as a control grid) and by direct drive; electrophoretic displays operating by lateral or in-plane motion of the electrophoretic particles, by vertical or electrode-to-electrode motion or any combination thereof; and full-color, spot-color and monochrome electrophoretic displays.

(109) Finally, it is again emphasized that although this invention has been principally described as applied to electrophoretic displays, many aspects thereof are applicable to any electro-optic display or medium capable of a remnant voltage, with particular importance for bistable electro-optic displays.

(110) It will be apparent to those skilled in the art that numerous changes and modifications can be made in the specific embodiments of the present invention described above without departing from the scope of the invention. Accordingly, the whole of the foregoing description is to be construed in an illustrative and not in a limitative sense.