Electrochimica Acta 46 (2001) 2015 2022 www.elsevier.

nl/locate/electacta

Electrochromics and polymers

Harlan J. Byker *
Pleotint, L.L.C., 7705 West Oli6e Road, West Oli6e, MI 49460, USA Received 16 September 2000

Abstract Polymers are used in commercial and proposed electrochromic devices as everything from the electrochromic materials themselves to solution thickeners, the matrix or solid phase of gels, solid polymer electrolytes and sealants. Each type of application of polymers in electrochromics is discussed and several denitions are compared with regard to gels and solution-based electrochromic redox systems. A review of electrochromics and polymer gels, especially from the perspective of the extensive patent literature on the subject, is given. 2001 Elsevier Science Ltd. All rights reserved.
Keywords: Electrochromic; Polymer; Gel; Solution phase; Redox

1. Introduction Polymers and electrochromic (EC) technology and EC devices are intimately linked, with a few exceptions. Polymers can be EC materials themselves such as in the well-known systems: polyaniline and polythiophene in which case the EC redox material is the polymer forming monomer. Also, the formation of the EC redox materials can be a polymer forming reaction such as with certain polyviologens. An EC redox material can be reacted with another monomer to form a polymer or can be reacted with functional groups on an already formed polymer such as with certain other polyviologens. If the polymer in these systems is immobilized, say for example by being surface conned on an electrode surface, then the EC redox materials are immobilized by the covalent bonds that form the polymer or attach the redox materials to the polymer. A polymer lm can contain physically immobilized redox materials in a fairly high concentration (in some cases also containing a little solvent) and can function as an EC material when the lm is in contact with an
* Fax: + 1-616-7387661. E-mail address: harlanb@pleotint.com (H.J. Byker).

electrode. The solvent in this case can be referred to as a plasticizer, although this term is sometimes misused in descriptions of EC devices in an attempt to mask the fact that mobile, solution phase materials are present. These polymers might or might not be crosslinked. A polymer, either organic, inorganic or a hybrid, can form a gel with a solvent or solution and this type of EC layer can contain one or more mobile EC redox materials. As will be discussed below the gel can be a physical gel without chemical bond formation between the polymer chains or the gel can be a chemical gel in which the chemical bonds (usually covalent bonds) are formed between the polymer chains. The chemical bonds can be formed directly by reaction of functional groups on the polymer or by added crosslinking agents that tie polymer chains together by reacting with functional groups on more than one polymer chain. In a chemical gel, a solution phase EC system may be interspersed in a matrix phase formed by the polymer. Often, an EC redox material is free to diffuse and/or migrate in the solution phase of a gel almost as if it was contained in a free solution. In a more simplistic case, a polymer can be added to a liquid solution simply to provide thickening or increased viscosity of the liquid solution. Thickening can

0013-4686/01/$ - see front matter 2001 Elsevier Science Ltd. All rights reserved. PII: S 0 0 1 3 - 4 6 8 6 ( 0 1 ) 0 0 4 1 8 - 2

2016

H.J. Byker / Electrochimica Acta 46 (2001) 20152022

be valuable if the device leaks or is broken and can help counteract the effects of natural convection in solution phase EC devices. Obviously, as the amount of polymer thickener is increased in a liquid solution, the viscosity increases. As the addition of polymer is continued, the viscosity increases to the point eventually where the solution is self-supporting or free standing and can be considered a physical gel. When EC redox systems, like metal oxide lms, are conned on electrode surfaces, a polymer containing electrolyte can provide for mobile ions to charge compensate the conned redox systems when they are electrochemically oxidized or reduced. These polymer containing electrolytes range from pure polymersalt combinations to thickened salt solutions. A basic polymer electrolyte is a polymersalt combination in which one or more ions from the salt is mobile and can be transported through the electrolyte like for example in poly(ethylene oxide) containing a lithium salt. The addition of some solvent/plasticizer often increases ionic conductivity in polymer electrolytes. Salts dissolved in the solution phase of physical and chemical gels can achieve ionic conductivities approaching those of liquid electrolytes. Polymeric sealant materials are often critical to the life of an EC device and may represent as big a R&D challenge as the EC system itself in a device commercialization effort. Even an all solid state EC system consisting of multiple thin lm ceramic coatings on glass that has been proposed for use in a double pane, insulated glass unit conguration, may have to rely on polymeric sealant materials for a portion of its life expectancy [1,2].

2. Solution phase electrochromics The vast majority of EC devices manufactured today involve at least one and in most cases more than one solution-based EC redox material. These materials are free to diffuse and/or migrate to the surface of the electrodes and undergo oxidation or reduction. In the preferred systems, both materials change from colorless to colored when the device is activated. In self-erasing devices with solution-based anodic and cathodic EC redox materials, the colored redox materials diffuse from the electrode surface to the bulk of the solution and meet each other. Upon meeting, they undergo homogenous electron transfer to reform the colorless form of the redox materials and they then can be transported back to the electrode to again participate in the color forming electrochemical reaction. This process is illustrated in Fig. 1. The self-erasing process requires continuous passage of current to maintain the colored state of the device. However, for a large EC window with a slow switching time, the current required to maintain deep coloration can be less than 5 mA/cm2. In addition, self-erasing helps to make certain EC devices, like automatic dimming rearview mirror, fail safe since they spontaneously go to the clear condition in the case of power failure. Some examples of cathodic solution-based EC redox materials are shown in Fig. 2 and a number of specic materials are given in Refs. [3 5]. These materials generally undergo two reductions of one electron each. The reductions take place at different redox potentials and the rst reduction is used in EC devices to provide color formation. The fact that the product of the second electrochemical step is chemically stable (at least to a reasonable extent) means that the product of the rst reaction can be highly stable, chemically. The author has observed no perceptible change in coloration of solutions for electrochemically formed viologen cation radicals of some viologens in sealed glass tubes at room temperature over a period of more than 12 years. Some examples of anodic solution-based EC redox materials are shown in Fig. 3 and a number of specic materials are again given in Refs. [3 5]. These materials generally undergo two oxidations of one electron each. The oxidations take place at different redox potentials and the rst oxidation is used in EC devices to provide color formation. Again, the fact that the product of the second electrochemical step is chemically stable (at least to a reasonable extent) means that the product of the rst reaction can be highly stable, chemically. With the proper choice of cathodic and anodic EC redox materials, devices have been constructed that cycle through full charge/discharge cycles at 70C for many millions of cycles.

Fig. 1. Single compartment, self-erasing EC device with two solution-phase EC redox materials.

H.J. Byker / Electrochimica Acta 46 (2001) 20152022

2017

Fig. 2. Examples of cathodic solution-phase EC redox materials: (1) viologens (color of the one electron reduced species depends on the R groups); (2) 7,7,8,8-tetracyanoquinodimethane; (3) substituted anthraquinones.

During EC device operation solution-based EC redox materials are subjected to natural convection or a process known as aggregation [6] or segregation [7]. This process is illustrated in Fig. 4. In this process the oxidation of the anodic materials causes the solution near the anode to increase in density relative to the bulk of the solution and the density gradient causes this more dense solution to sink in the gravitational eld of the Earth. The increase in density can be understood in terms of the tendency of polar organic solvents like those used in EC devices to solvate cations and neutral materials better than anions. Thus as the anodic material is oxidized it becomes better solvated or the solvation sphere tends to move in toward the newly formed oxidized materials and hence the solution becomes more dense. The reverse happens at the cathode for the same reason and the solution becomes less dense and the solution rises in the gravitational eld of the Earth. During electrolysis of solutions containing these types of anodic and cathodic materials the colored solutions can be readily observed to stream up at the cathode and down at the anode. Also, an EC device with an unthickened solution left in the colored state for 20 min or more, will exhibit colored bands of the cathodic material near the top and the anodic material near the bottom when an attempt is made to clear the device. Thickeners can be added to help suppress natural convection and can also be added to thicken the solution in case of leakage or breakage. One of the earliest

suggestions of adding thickeners is from Manos [8]. Acrylic ber, poly(vinylbutyral) and a three-dimensional matrix forming material, colloidal silica, are described for use as thickeners in EC devices containing solution phase EC redox materials. The segregation phenomena is signicantly reduced by adding polymer thickeners to the solution. Unfortunately the non-uniform potential distribution in an EC device with transparent electrodes results in its own segregation of materials and color band formation. Interestingly, with the proper choice of solution viscosity and the judicious choice of placing the positive lead at the top of the device and the negative lead at the bottom the two phenomena can be played off against each other to give almost no color band formation in some device congurations [7]. While thickening the solution provides some benet when a device leaks or cracks, it is often desirable to completely retain the liquid portion of the device. An even greater motivation for the development of EC devices with self-supporting or free standing EC systems comes from the need to eliminate hydrostatic pressure in windows that are even reasonably large. A simple liquid sealed between two pieces of glass over a surface area of say 1 m2 will exert forces large enough to bow out the glass on the bottom half of the window, collapse the glass pieces together in the top half of the window and in some cases rupture the seals holding the assembly together.

2018

H.J. Byker / Electrochimica Acta 46 (2001) 20152022

3. Polymer based gels Gelling an electrolyte layer including an EC layer has been suggested as a means of increasing the structural integrity of layer and as a means of eliminating hydrostatic pressure concerns. However, not everything that is termed a gel is self-supporting, free standing and capable of eliminating hydrostatic pressure. A simple experiment was conducted that illustrates the concern over assuming that a polymer-thickened solution is a gel. Three solutions were prepared of lithium perchlorate and PMMA, average Mw :996,000 from Aldrich Chemical Company, in acetone and propylene carbonate. The solutions were prepared in round bottom asks. The acetone was removed at reduced pressure and elevated temperature with a rotary evaporator and the remaining solutions, each contained 0.5 M lithium perchlorate and at least 20, 30, and 40 wt% PMMA in propylene carbonate. All three asks were stoppered and placed on their sides in an oven at 80C. After the viscous solutions had moved to the side of the asks, the asks were set up and the times for the

solutions to move to the bottom of the asks at 80C were observed. For the 20 wt% PMMA in propylene carbonate solution the time was only 15 min. For the 30 wt% PMMA in propylene carbonate solution the time was about 5 h. For the 40 wt% PMMA in propylene carbonate solution the time was about 36 h. Albeit a crude experiment, it serves to illustrate that care should be used in picking a polymer system for use in an EC window that must survive times at elevated temperatures during sunlight exposure, under the forces of gravity, for many years. However, as the weight percentage of the polymer thickener increases, one can imagine a case in which the viscosity increases to the point where the electrolyte solution will not ow over the time scale of interest for the use of an EC device. At this point one may consider the polymer containing EC layer to be a physical gel. As an alternative to physical gels, a relatively small amount of polymer thickener can be used as a means of making a free-standing, self-supporting gel system by providing a modest amount of crosslinking between the

Fig. 3. Examples of anodic solution-phase EC redox materials: (1) substituted phenazines; (2) substituted para-phenylenediamines; (3) substituted benzothiazole azines.

H.J. Byker / Electrochimica Acta 46 (2001) 20152022

2019

Fig. 4. Illustration of segregation phenomenon in solutionphase EC devices. In the absence of polymer thickeners, the solution of oxidized anodic material sinks while the solution of reduced cathodic material rises during electrolysis.

Fig. 5. Illustration of polymer thickeners, polymer gels, and polymer electrolytes as a function of viscosity and polymer crosslinking.

polymer chains of the thickener. If the crosslinking is provided by chemical bond formation, the resulting gel can be termed a chemical gel. In some cases, forming a crosslinked polymer matrix with as little as about 3 or 4 wt% total polymer content in the presence of as much as 96 97 wt% solvent can give rise to a non-light scattering gel that will not ow even at elevated temperatures. Fig. 5 provides a qualitative comparison of polymer thickeners, polymer gels, pure polymer electrolytes and their variation in viscosity. The viscosity as a function of polymer content is highly dependent on crosslink density. In this illustration, one starts with a low viscosity liquid and adds some soluble uncrosslinked polymer thickener to increase viscosity. High molecular weight linear polymers are signicantly more effective thickeners than low molecular weight polymers. However, in either case, as the polymer content is increased the viscosity increases and as discussed above and as illustrated in the left arm of Fig. 5 either sooner or later a highly thickened system becomes a physical gel in which chain entanglement, van der Waals forces between the polymer chains and solution effects are sufcient to make the system self-supporting and free standing. Having even higher polymer content (i.e. lower relative solvent content) can lead to a system in which the remaining solvent is considered merely as a plasticizer for the polymer. Finally if the correct polymer and electrolyte salt are chosen, the solvent can be eliminated and a pure polymer electrolyte can be formed which still allows at least certain ionic species to have mobility within the polymer. The right arm of Fig. 5 is intended to illustrate that the viscosity can increase dramatically when the polymer chains are crosslinked and once a three-dimensional network of even a low polymer content is present in the electrolyte solution the system can support itself and be used to eliminate hydrostatic pressure issues. While it is possible to form a crosslinked polymer and then swell that polymer with an electrolyte solution, for ease of device fabrication it is often desirable to provide the precursors for polymer matrix formation in a relatively low viscosity solution, ll the device with this solution and form the gel in situ once the device is lled. Having large amounts of crosslinked polymer is possible if there is good compatibility between the electrolyte solution and the matrix. However, as the polymer content increases issues like light scattering, shrinkage, cracking, crazing, and syneresis (having free liquid expelled or exuded from the gel) become possible. Very high levels of polymer content (probably greater than 80 wt% in most systems) appear to be necessary to prevent mobility of redox species in the gel and thus even devices containing large amounts of polymer and relatively low concentrations of EC redox materials continue to function via mass transport to

2020

H.J. Byker / Electrochimica Acta 46 (2001) 20152022

and from the electrode, although with slower speeds of coloring and clearing. As illustrated in Fig. 5, if the solvent content is taken to zero the crosslinked matrix can also result in a pure polymer electrolyte that happens to have some crosslinking in the polymer. It is interesting to look at Fig. 5 in light of the history of polymer electrolyte development for lithium batteries [9], which of course is also of interest for lithium ion-based EC devices. The history given in Table 1 of Ref. [9] starts with pure polymer electrolytes and progresses through networks and crosslinked polymers to the addition of solvent type plasticizers to improve mobility and progresses further to gel-polymer electrolytes with ionic conductivities near to those of free liquid electrolyte systems.

4. EC redox in gel type devices For EC devices in which there is a self-supporting, free-standing gel, in which the electrolyte contains one or more of the active EC redox materials, the redox materials can participate in electron transfer with the electrode in at least three ways: (1) by mass transport to the electrode; (2) via redox hopping; and/or (3) via mediated electron transfer utilizing a redox materials capable of mass transport to the electrode surface. In the case of redox hopping the active EC redox materials must be present at very high concentration in order to be close enough to allow for electrons to hop from one site to another. EC devices are described by Tsutsumi et al. [1012] which contain organic EC redox materials, about 20 wt% solvent, some supporting electrolyte and a high proportion of crosslinked polymer. The crosslinked polymer is formed by in-situ polymerization once the device is lled with a solution containing all of the EC layer materials including the polymer forming materials. The EC layer is almost certainly self-supporting at this level of polymer content. The organic redox materials are reported to color when a voltage is applied to the device in the same manner as they color on electrochemical reduction in a simple solution with the solvent used in the device. While the organic EC redox materials chosen are not particularly stable in their colored states, the devices exhibit measurable changes in absorption in 320 s after a voltage is applied. The concentration of EC redox material in these devices is less than 50 mM and as will be discussed below this concentration is too low to provide signicant electrochemical electron transfer via redox hopping. Thus even with about 20 wt% solvent there appears to be some mobility for redox materials in the electrolyte layer. Hirai et al. [13,14] describe EC devices that have organic EC redox materials physically contained in a

polymer lm formed by solvent casting a solution of polymer, electrolyte, and redox material followed by an attempt to remove most, if not all, of the solvent. The EC layer is solvent cast on a transparent electrode and in most cases a counter electrode is provided by vacuum depositing a metal counter electrode on the EC layer. While EC redox materials are considered by these authors to be immobile in the EC layers and redox hopping is invoked as the means for electron transport, the concentrations in the solution used for solvent casting are ambiguous as they are given in mole ratios or moles per liter and the molecular weights of the polymers are not specied. EC systems that are termed polychromic [15,16] are formed by reaction of an EC monomer composition to give what is described as an EC layer in which the EC compounds are not free to migrate. These systems are generally based on lling an EC device with an EC monomer composition and in situ reaction of monomers, at least some of which are multifunctional. A large number of examples are given which describe either polyacrylate and/or polyurethane formation after a solution with the proper precursors has been placed in an EC device. In most cases anodic and cathodic EC materials are present that do not react with the polymer forming monomers and the solvent content of the nal EC layers typically ranges from 40 to 60 wt% for the polyacrylate-based layers to nearly 90 wt% for the polyurethane-based layers. The solvent is referred to as a plasticizer even when it makes up nearly 90 wt% of the layer. In some of the examples of polychromic devices, one of the cathodic EC materials has a functional group that is reactive with the polymer forming monomer. The concentrations of the EC materials are similar to those in solution phase EC devices and the performance for speed of coloration and clearing and depth of coloration are also similar. If the EC materials are not truly free to migrate in these EC layers, then these compositions represent a remarkable example of fast redox hopping over long distances (given the low concentrations) or some new form of electron transfer. Ferrocene is one of the most used EC materials in the examples of polychromic devices. It is interesting to note that a diffusion coefcient for ferrocene in an electrolyte layer contained in a crosslinked polyacrylate matrix has been reported that is only a factor of 2.5 less than that in a free liquid solution [17]. In contrast to polychromic systems, EC systems with similar compositions are described as free-standing gels by another group [18 20]. A free-standing gel is described as a system of polymer matrix with interspersed EC solution. Once again the EC layers are formed by lling devices with a relatively low viscosity

H.J. Byker / Electrochimica Acta 46 (2001) 20152022

2021

solution of monomers and/or uncrosslinked polymers in a solution, followed by in situ reaction to give a crosslinked polymer matrix. Preferred polymer or crosslink forming reactions are addition reactions such as the reaction of isocyanates with alcohols since no by-products are formed. These gels can be thought of as a two-phase system in which a continuous solution phase is interspersed in a three-dimensional matrix formed by a polymer phase. The solution portion of the gel is open enough to allow relatively large molecules and ions to diffuse and/or migrate through out the phase, including to and from the electrodes. This description of a gelled, solution phase EC system is in stark contrast to that for polychromics for which the inventors claim that the EC materials are not free to migrate. In the two-phase concept of a gel, the two phases are intimately related in that the solvent of the solution phase is believed to solvate the polymer chains and keep them from collapsing or agglomerating to form light scattering regions and the polymer matrix holds the solution in place so that there is no leakage or hydrostatic pressure exerted by the EC layer composition once it is reacted or cured. In essence the polymer matrix is a single macromolecule dissolved in the solution remaining after the macromolecule is formed. The above denition is in keeping with a number of literature denitions of gels quoted below: 1. A gel is a crosslinked polymer network swollen in a liquid medium. Its properties depend strongly on the interaction of these two components. The liquid prevents the polymer network from collapsing into a compact mass, and the network, in turn, retains the liquid [21]. 2. We interpret a gel to consist of continuous solid and uid phases of colloidal dimensions. Continuity means that one could travel through the solid phase from one side of the sample to the other without having to enter the liquid; conversely, one could make the same trip entirely within the liquid phase [22]. 3. Gel: a two-phase colloidal system consisting of a solid and a liquid in more solid form than a sol [23]. 4. A gel may contain very considerable quantities of the dispersion medium (over 99% in some cases). If the concentration of colloidal substance in the gel is not too high, the gel may behave almost exactly the same as the liquid sol, apart from mechanical properties. Thus the rate of diffusion of crystalloids in the gel is practically the same as in the sol, the electrical conductivity is unchanged, and chemical reactions take place in the gel just as they do in the ordinary solution [24].

5. Immobilized redox materials EC and other electrochemical systems are known in which the redox materials are believed to be immobilized. One example involves conducting polymers like polythiophenes and polyanilines. If one were to estimate the density of a polyaniline lm formed on a transparent electrode by electrochemical oxidation of aniline as 1 g/cm3, then the concentration of phenylenediamine redox groups can be calculated to be on the order of 10 M. In a similar fashion a lm of polythiophene, which was assumed to have a density of 1 g/cm3, would have a concentration of thiophene groups on the order of 12 M. In these cases the electron transfer can proceed by redox hopping or via the electronic conductivity of the polymer itself. And even in these cases, in order for the layers to function in EC devices, they must allow for mobility of charge compensating ions within the polymer layers. An extensively studied EC system with viologen molecules immobilized by providing reactive pendants that form a highly crosslinked polymer matrix is described in several patents [25 27]. Surface-conned EC layers are formed with a high effective concentration of redox sites by reaction of viologen monomers that have trialkoxysilane groups attached to the bipyridinium moiety. If the density of these lms is assumed to be 2 g/cm3 then the concentration of viologen redox sites can be of the order of 4 M. While there is expected to be little or no mobility for the viologen redox sites in these lms the concentration is believed to be high enough to support redox hopping and the surfaceconned, immobilized viologen lms can be used to provide signicant EC activity. Viologen molecules are also believed to be immobilized on an electrode surface by coating an electrode with a solution of a polyviologen/polyanion complex, followed by drying to remove solvent [28]. Placing the coated electrode in an electrolyte solution in which the coating is insoluble and then polarizing the electrode is reported to give rise to EC activity. Here again, the concentration of redox sites is high enough to support redox hopping. Support for the requirement of having high molar concentrations of immobilized redox sites to support redox hopping is given in many places including Fig. 2 of Ref. [29] which shows that the charge-transport diffusion coefcient for electrochemical reaction of immobilized organometallic species does not reach any kind of signicant values until the concentration reaches greater than about 0.8 M. Similarly, Fig. 8 of Ref. [30] shows a rapid increase to appreciable values for the apparent diffusion coefcient for an electrochemical oxidation of immobilized redox species as the concentration increases to values above about 1 M.

2022

H.J. Byker / Electrochimica Acta 46 (2001) 20152022 [2] S.C. Schulz et al., United States Patent 6,055,089, April 25, 2000. [3] R. Srinivasa et al., United States Patent 6,037,471, March 14, 2000. [4] R. Srinivasa et al., United States Patent 5,998,617, December 7, 1999. [5] K.L. Baumann et al., United States Patent 6,020,987, February 1, 2000. [6] I.V. Shelepin, et al., Elektrokhimia 13 (1977) 346 (English translation). [7] H.J. Byker, United States Patent 4,902,108, February 20, 1990. [8] P. Manos, United States Patent 3,451,741, June 24, 1969. [9] D. Fauteux, et al., Electrochim. Acta 40 (13) (1995) 2185. [10] H. Tsutsumi, et al., Sol. Energy Mater. Sol. Cells 39 (1995) 341. [11] H. Tsutsumi, et al., J. Polym. Sci. 30 (1992) 1725. [12] H. Tsutsumi, et al., Electrochim. Acta 37 (1992) 369. [13] Y. Hirai, et al., Appl. Phys. Lett. 43 (7) (1983) 704. [14] Y. Hirai United States Patent 4,550,982, November 5, 1985. [15] D.V. Varaprasad et al., United States Patent 5,910,854, June 8, 1999. [16] D.V. Varaprasad et al., United States Patent 6,002,511, December 14, 1999. [17] K.M. Abraham, M. Alamgir, J. Electrochem. Soc. 137 (1990) 1657. [18] W.L. Tonar et al., United States Patent 5,679,283, October 21, 1997. [19] W.L. Tonar et al., United States Patent 5,888,431, March 30,1999. [20] W.L. Tonar et al., United States Patent 5,928,572, July 27, 1999. [21] T. Tanaka, in: J.I. Kroschwitz (Ed.), Concise Encyclopedia of Polymer Science and Engineering, Wiley, New York, 1990, p. 432. [22] C.J. Brinker, G.W. Scherer, Sol Gel Science, The Physics and Chemistry of Sol Gel Processing, Academic Press, New York, 1990 (p. 8). [23] S.P. Parker (Ed.), McGraw-Hill Dictionary of Chemical Terms, McGraw-Hill, New York, 1984. [24] H. Remy, Treatise on Inorganic Chemistry, vol. vol. II, Elsevier, Amsterdam, 1956, p. 674 (translated by J.S. Anderson). [25] N. Leventis et al., United States Patent 5,818,636, October 6, 1998. [26] N. Leventis et al., United States Patent 5,457,564, October 10, 1995. [27] M.S. Wrighton et al., United States Patent 4,473,695, September 25, 1984. [28] H. Akahoshi, et al., J. Phys. Chem. 85 (1981) 818. [29] J.S. Facci, et al., J. Am. Chem. Soc. 104 (1982) 4959. [30] M. Kitagawa, et al., Polym. Adv. Technol. 6 (1994) 204.

6. Seals and sealants Practical EC devices for many applications are required to have useful lives of from 10 to 30 years. In most proposed EC systems the useful life of the devices is limited by the stability of one or more of the EC materials in either its colorless or colored state or by the process of cycling between these states. If an EC system is developed which has adequate stability of its own accord, the next challenge is to prevent internal materials form escaping from the device and external materials from entering the device. EC devices that involve electrochemical reduction and oxidation to form color have a good chance of being sensitive to the oxidizing power of oxygen from air. Also, many EC systems are sensitive to water from atmospheric moisture or other environmental contaminates like ammonia, chlorine, hydrogen sulde, etc. While the reactivity of the materials may be low, it must be considered that the reaction times of interest for degradation processes are on the order of the 1030 years, corresponding to the desired life of the devices. Since the seal of a device may be as critical a factor in the life of a device as the stability of the EC system itself, the challenge in nding and consistently providing a satisfactory seal should not be underestimated. In principle true hermetic seals can be provided between glass substrates by glass to metal seals and solder glass or low melting frit glass seals. However, these present difcult technical issues for high volume production. Thus efforts usually focus on polymeric seals like acrylics, epoxies, and rubbers. Exact formulations are closely guarded secrets in most manufacturing situations. Accelerated testing is often a key to the development of satisfactory seals and tests like boiling water immersion, exposure to water vapor in an autoclave environment at 120C with 100% relative humidity, thermal shock, exposure to high pressure oxygen at high temperatures, and measurement of physical strength of seals are helpful in seal development.

References
[1] D.K. Benson et al., United States Patent 5,384,653, January 24, 1995.