- Email: [email protected]
On the electrochromic properties of antimony–tin oxide thin films deposited by pulsed laser deposition
Solid State Ionics 156 (2003) 463 – 474 www.elsevier.com/locate/ssi
On the electrochromic properties of antimony–tin oxide thin films deposited by pulsed laser deposition N. Naghavi, C. Marcel, L. Dupont, J-B. Leriche, J-M. Tarascon* Laboratoire de Re´activite´ et Chimie des Solides, Universite´ de Picardie Jules Verne, 33 rue Saint Leu, F-80039 Amiens Cedex, France Received 5 April 2002; received in revised form 9 July 2002; accepted 19 August 2002
Abstract Antimony – tin oxide (ATO) thin films were deposited using pulsed laser deposition. From a wide survey of deposition conditions (deposition atmosphere and substrate temperature), we deduced that the optimum experimental parameters to obtain electrochromic active films were a deposition temperature of 200 jC at a 10 2 mbar oxygen pressure followed by an annealing at 550 jC, the latter being critical. Through potentiostatic tests, we showed that depending on the composition, which influences film morphology, the Sn – Sb – O films could present either a faradic or a capacitive-like behavior, associated to a color or a neutral switching over a wide range of potentials, respectively. In addition, a relationship between the film annealing temperature and its structure as well as its electrochromic properties was evidenced by means of complementary electrochemical, X-ray diffraction and electron microscopy techniques. Finally, the electrical and optical properties of the ATO films were discussed in terms of the energy band theory. D 2003 Elsevier Science B.V. All rights reserved. Keywords: Antimony – tin oxide; Pulsed laser deposition; Electrochromic properties
1. Introduction In recent years, electrochromic devices have received considerable attention as they are able to reversibly change their optical properties when an electric current flows through the device [1,2]. Due to this feature, they find applications in rearview mirrors [3], «smart windows» [2], display panels [4], etc. The most extensively studied solid-state electrochromic systems are based on tungsten oxide [5– 8] *
Corresponding author. Fax: +33-32-2827590. E-mail address: [email protected] (J.-M. Tarascon).
working electrode thin films, namely a proton or lithium insertion material well known for its blue electrochromism. With respect to the counter electrodes, two options that consist in using either optically neutral (e.g., CeVO4 [9]) or complementary colored electrodes such as IrO2 [10] or hydrous Ni oxide [11] are being pursued. However, the latter, exclusively used in proton-functioning devices, present non negligible drawbacks such as a high cost for the first one and a poor stability in acidic electrolytes for the second one. Thus, alternative materials are required for proton-conducting counter electrodes in electrochromic windows. Recently, a large number of studies have been devoted to low Sb-doping in ATO compounds for their
0167-2738/03/$ - see front matter D 2003 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 7 - 2 7 3 8 ( 0 2 ) 0 0 7 4 9 - X
464
N. Naghavi et al. / Solid State Ionics 156 (2003) 463–474
transparent conducting properties [12]. Such materials, which can be grown in thin film forms either by spray pyrolysis [13], dip coating [14,15] or sputtering [16], present a great interest with respect to antistatic coatings [17], gas sensors [18] and energy storage applications [19]. However, the electrochromic properties of ATO thin films containing high antimony content received very little attention with a few exceptions. Coleman et al. [20] demonstrated for the first time the feasibility of fast switching devices for display application made with ATO fine grains embedded in a fluoro-elastomer resin. In light of these promising results, Marcel et al. [21] studied the electrochromic properties of Sn – Sb– O powders and thin films grown by pulsed laser deposition. They pointed out that depending on the deposition conditions, ATO films could either exhibit a grey reversible coloration upon reduction in relation with a faradic behavior, or present a neutral switching over a wide range of potentials associated to a capacitive-like behavior. The present paper aims at providing more information about the influence of deposition conditions (especially postannealing temperature and Sn – Sb – O composition changes) on the electrochemical, structural and optoelectronic properties of the resulting films in the hope of identifying the origin of their competing faradic/ capacitive-like behavior.
2. Experimental Thin oxide films were prepared by pulsed laser deposition using a KrF excimer laser beam (Lambda Physic, Compex 102, k = 248 nm) with a laser fluence of 1 –2 J/cm2. The targets were pellets of commercial SnO2 (Aldrich 99.9%) and Sb2O3 (Aldrich 99.9%) mixtures in stoichiometric proportions annealed for 20 h at a relatively low temperature (700 jC) in order to avoid any antimony loss while achieving a pellet density of 60 – 70%. Films of 1 1-cm2 area were deposited either on (SnO2:F)-coated glass for electrochemical tests or simple glass for structural and electronic characterizations. Deposition time was fixed at 15 min with a repetition rate of 3 Hz. The thickness, determined by profilometry using a Dektak St instrument, was estimated as being in the range of 200– 250 nm. Film crystallinity was examined by X-ray diffraction (XRD) with a Philips diffractometer model PW
1710 (kCuKa = 0.15418 nm). The surface morphology and film composition were investigated by scanning electron microscopy (SEM) with a Philips XL 30 field emission gun (FEG) coupled to an Oxford Link instrument for energy-dispersive X-ray spectroscopy (EDS). High-resolution transmission electron microscopy (HRTEM) was carried out using a JEOL 2010 microscope equipped with an EDS analyzer. Optical transmission spectra in the UV – visible and near infrared regions (250 – 2500 nm) were obtained using a Varian double beam UV – Vis – NIR spectrometer ‘‘CARY-5E’’. The electronic conductivity of the films was taken as the inverse function of the resistivity, which was determined by the four-point probe method using the formula q=(V/ I )C(a/d, d/s), as reported by Smith [22]. Voltages and DC currents were measured with a multimeter (Keitley 175 A). Within the above formula, V is the voltage and I is the current, a and d are the sample lengths (in our case a = d ), and C(a/d, d/s) is a correction factor accounting for the sample geometry. The resistivities were obtained by assuming that our samples were of the same thickness and dimensions. Therefore, such values should not be taken as quantitative but as qualitative values enabling to compare our variety of samples having different compositions or anneals. The electrochemical properties of the films were characterized by cyclic voltammetry performed with an Autolab PGSTAT 30 system. The cell consists of the ATO film as the working electrode and of a platinum wire as the counter electrode, both immersed in a 0.1 M aqueous H3PO4 electrolyte. The potentials were given versus the saturated calomel electrode (SCE) reference. For all these experiments, cyclic voltammograms (CV) were scanned at a rate of 10 mV/s.
3. Results 3.1. Influence of the deposition conditions Our previous work dealing with the ATO system [21] showed that optimum electrochromic properties were achieved for films made from removing a (70% Sn – 30% Sb or 60% Sn – 40% Sb) target composition under an oxygen pressure of 10 1 and 10 2 mbar for
N. Naghavi et al. / Solid State Ionics 156 (2003) 463–474
faradic and capacitive-like behavior, respectively. First, we assumed that the difference between these electrochemical processes was nested in film morphology (namely, the porosity). To test this hypothesis, we studied the influence of temperature and oxygen pressure on film morphology during deposition and annealing steps, and thereby on the film electrochemical/electrochromic properties. While maintaining both film composition (60% Sn –40% Sb) and pressure (10 2 mbar), we noted that a slight grey cathodic coloration could only be generated if films were first grown at 200 jC, and then post-annealed for 30 min at 550 jC. Indeed, attempts to directly grow films at 500 jC resulted in polycrystalline films, which were black and electrochemically inactive. In contrast, amorphous films, which were colorless with no electrochemical activity in H3PO4 liquid electrolyte, were obtained when the growth temperature was limited at 200 jC. It should be noted that the annealing step resulting in film crystallization was accompanied by a slight darkening. While maintaining the same target composition and growth protocol (deposition at 200 jC followed by a 550 jC annealing step), we varied the oxygen pressure between 10 4 and 10 1 mbar. From pressures smaller than 10 2 mbar, all the deposits were electrochemically active. In contrast, the films grown at pressures greater than 10 2 mbar were electrochemically inactive, while being both less crystalline and more porous, as deduced from XRD and SEM measurements, respectively. Moreover, EDS analysis revealed that increasing the deposition pressure resulted in ATO films having an antimony content greater than the target nominal composition. In light of such findings, we checked the importance of the Sn/Sb ratio on film morphology and electrochemical properties. 3.2. Influence of the Sb content Films of various antimony – tin oxide compositions were grown from targets whose nominal composition ranged between 0% and 80% at Sb. More specifically, the films were deposited at 200 jC under 10 2 mbar, annealed for 30 min at 550 jC under the same oxygen pressure, and characterized for their electrochromic properties. Under such deposition conditions, as deduced from EDS analysis, we noted that the Sn/Sb
465
atomic ratio was relatively well preserved between the target and grown films. 3.2.1. Structural characterization SEM micrographs of Sn-Sb-O films deposited on FTO-glass substrate are shown in Fig. 1 for various antimony contents. While films containing less than 50% Sb appear dense and homogenous, those having higher Sb contents reveal the presence of cracks underlining a porosity, thereby implying a decrease in film density. For Sb percentages greater than 80%, the deposits were poorly sticking to the substrate, making their characterization meaningless. Note that the apparent porosity observed for dense films with low Sb contents is a visual effect due to FTO texture. Fig. 2 shows the X-ray diffraction patterns of ATO thin films deposited on glass substrate, which are typical of SnO2 cassiterite phase. Increasing Sb amount yields a loss of crystallinity; this trend being also observed for ATO thin films deposited on FTO substrate. The absence of peaks pertaining to antimony oxide phases (whatever the oxidation states) lets suppose the complete dissolution of antimony in the nanostructured SnO2. Beyond 70% Sb, films become amorphous. Purely coincidental or not, ATO nanoparticles with a similar Sb content and prepared by our group via a solution route became amorphous [21]. Such results could first indicate that the ‘‘apparent solubility limit’’ of antimony in nanocrystalline tin oxide is reached near the 70% Sb. At this point, however, a legitimate question was whether these Sb-enriched ATO specimens presented large amounts of amorphous antimony-containing material as previously reported [23] or were single phases. To unambiguously answer such a question, HRTEM measurements were carried out on our heavily substituted ATO films. Fig. 3a,b display HRTEM micrographs of films deposited from targets composed of 60% Sb and 40% Sb, respectively. The high-resolution images and the electron diffraction patterns recorded from these materials gave no evidence for the presence of any antimony oxide phase. For both compositions, the diffraction pattern contains well-resolved diffraction rings characteristic of an SnO2 cassiterite phase. Indeed, films consist of small disordered crystals of a cassiterite-type phase and no trace of amorphous
466
N. Naghavi et al. / Solid State Ionics 156 (2003) 463–474
Fig. 1. SEM micrographs of ATO thin films with antimony contents ranging from 10% to 80% Sb: (a) 20%; (b) 40%; (c) 50%; (d) 60%; (e) 70%; and (f) 80%.
materials is observed independently of the antimony amount. However, changes in film texture with composition are visible. While the film with the lowest antimony content (40% Sb, Fig. 3b) is very dense and exhibits a poor definition of particle shape, the film containing 60% Sb (Fig. 3a) appears less dense with ˚ . In short, the better defined grains of about 50 A TEM study confirmed the formation of Sn –Sb – O solid solution with a SnO2 cassiterite-type structure up to a concentration of 60% Sb. 3.2.2. Electrical and optical properties The effect of the Sb-content on film conductivity is shown in Fig. 4. The conductivity initially increases with an antimony percentage up to 10% Sb, and then continuously decreases for higher Sb contents. The initial conductivity increase was ascribed [12] to an increase in the carrier concentration as a result of Sb doping. Indeed, as commonly reported for low Sb amounts, Sb5+ ions provide an n-type doping (electrons) within the conduction band [12 –16]. Note that in our case, Sn0.9Sb0.1O2 films were not as conductive as a classical transparent conducting oxide (TCO) whose conductivity averages 103 – 104 S/cm. This is most likely due to the fact that the two-step heat treatment
Fig. 2. X-ray diffraction patterns of ATO thin films with 10 V %Sb V 60 having a SnO2-cassiterite structure.
N. Naghavi et al. / Solid State Ionics 156 (2003) 463–474
Fig. 3. TEM micrographs of ATO thin films with composition of (a) 60% Sb and (b) 40% Sb.
467
468
N. Naghavi et al. / Solid State Ionics 156 (2003) 463–474 Table 1 Evolution vs. Sb percentage of the threshold temperature for which the resistivity suddenly drops during postannealing in air for ATO thin films deposited at low temperature (200 jC)
Fig. 4. Variation of the conductivity (r) as a function of antimony content of ATO films deposited at 200 jC, then annealed at 550 jC.
process that we used to induce electrochemical properties slightly differs from the best established growth protocol to achieve optimum conductivity. For higher Sb content, the observed conductivity downfall may have different explanations. It may be due either to the existence of traps created by crystal defects increasing with Sb content thus dominating the electron scattering mechanism or to the deep Sb3+ levels that act as an electron trap (as proposed by Kojima et al. [24] from XPS measurements results). In fact, a realistic scenario involves the simultaneous presence of both Sb5+ and Sb3+ ions, which can be described as: 5þ 3þ Sn4þ 1x Sbxy Sby O2ðx=2yÞ
Thus, concomitant with the Sb substitution for Sn4+ ions would be the appearance of Sb5+ and
Fig. 5. Variation of resistivity (R) of a film containing 10 – 40% Sb as a function of the temperature (T).
Composition
10% Sb
20% Sb
30% Sb
40% Sb
Threshold temperature (jC)
300
450
500
510
Sb3+ species that would lead to a noticeable lowering of the electron mobility consistent with the observed resistivity drop. The electrical properties of films containing 10– 40% Sb deposited at 200 jC as a function of the temperature are shown in Fig. 5. Whatever their antimony content the as-made films were all insulating. Nevertheless, upon annealing them in air, their resistivity sharply dropped over a temperature range of 300 –510 jC with increasing antimony contents (Table 1). This temperature-driven drastic conductivity onset could be explained as a consequence of the film crystallization that leads to a reorganization of the band structure, and thereby the creation of Sb+5 donor levels below the bottom of the conduction band. Fig. 6 shows the transmission spectra in the UV – visible and near infrared regions for films having Sb content ranging from 10% to 60% Sb. While films grown at 200 jC were amorphous and highly transparent independent of composition, the same films annealed at 550 jC exhibited a transmittance greatly influenced by the Sb amount. The lowest transmittance pertains to (90% Sn –10% Sb) sample and the highest to the most Sb-charged composition (i.e., 40% Sn–
Fig. 6. Optical transmission spectra of a film deposited at 200 jC (30% Sb) and post-annealed ATO thin films with 0 V %Sb V 60.
N. Naghavi et al. / Solid State Ionics 156 (2003) 463–474
Fig. 7. Variation of the optical band gap (Eg) estimated from optical spectra as a function of %Sb for polycrystalline films annealed at 550 jC. Band gaps of 4.05 and 3.2 – 3.5 eV are reported for SnO2 and Sb2O3, respectively (see oxide Handbook).
60% Sb). Thus, film coloration varies from light grey to colorless with raising antimony content. Finally, note that the transmittance drop observed in the infrared part is more pronounced for low Sb rates. This is totally consistent with our conductivity measurements since, according to Drude’s model based on free carrier absorption, the transmittance should be the lowest for the highest conducting films. Regarding the UV region, the transmittance falls very sharply due to the onset of fundamental absorption associated to band-gap energy (Eg). For each composition, we deduced the Eg values (Fig. 7) by extrapolation of the linear portion of a2 versus hm for crystallized annealed films, where a is the absorption
469
coefficient. After post-annealing, we observed an overall widening of the band gap presenting a minimum at about 30% of Sb. Beyond 30%, the band gap widening is a natural result of the observed transmittance enhancement with increasing the antimony content. The observed heat treatment-driven optical film behavior is still open to discussions. Some authors suggested that antimony-segregated atoms form clusters in the lattice resulting in a darkening of the films [13]. Others could evidence an antimony-rich amorphous phase, interpreted as Sb2O4, yielding a color lightening [23] of the ATO sample. However, none of these assumptions could be corroborated by our HRTEM study. Therefore, based on our results in conjunction with Kojima et al.’s [24] studies, we propose the following scenario. According to these authors, antimony firstly contributes in low temperature deposited ATO thin films to the widening of the conduction and valence bands, but not to the formation of light absorbing levels in the energy gap consistent with the transparency of the Sn– Sb –O films deposited at 200 jC. After a heat treatment, the sharp decrease in resistivity depicted in the previous section is due to the formation of Sb5+ energetic levels overlapping the bottom of the conduction band (Fig. 8a). However, since the measured resistivities were inferior to TCOs in low Sb content samples because of the poor crystalline nature of the films, the darkening coloration was possibly due to the presence of Sb5+ impurity levels lying close to the bottom of the conduction band. Beyond 10% Sb
Fig. 8. Schematic representation of the band structure for ATO films annealed at 550 jC (polycrystalline) with different antimony amounts: (a) 10% Sb, (b,c) 20 – 40% Sb, and (d) 60% Sb.
470
N. Naghavi et al. / Solid State Ionics 156 (2003) 463–474
content range, the presence of Sb3+ reinforces light absorption until 30% Sb by creating supplementary defects causing a band-gap reduction (Fig. 8b,c). With further increasing Sb content, we enhance the number of Sb3+ ions that may act as traps for conductive electrons, and thereby form localized states within the SnO2 band gap. These levels may become deeper and deeper with increasing antimony amounts, and generate less and less light absorption, leading to a progressive bleaching of the films (Fig. 8d). 3.2.3. Electrochemical characterization Cyclic voltammograms (CVs) of our ATO films were obtained by sweeping the potential from 0.7 to 1.5 V versus SCE. Depending on the film Sb content, two types of CV traces were observed (Fig. 9). Within the 10 – 40% Sb range, CV curves are featureless adopting a pseudo-capacitive-like shape as previously shown by Marcel et al. [21], and
furthermore, the transmittance remains the same as shown for the (60% Sn – 40% Sb) sample in Fig. 10a. From 60% to 70% Sb content, a pronounced faradic behavior appears, possibly corresponding to (Sb5+ X Sb3+ ) charge transfer with simultaneous H+ (de)intercalation. Indeed, a redox couple, visible at ca. 0 V in oxidation and 0.7 V in reduction, produces a dark-grey coloration with a cathodic sweep, followed by a bleaching process with an anodic one (see Fig. 9e). Fig. 10b shows that the optical contrast for the composition (40% Sn – 60% Sb) is reinforced with increasing Sb amount. It is interesting to note that upon further decreasing the cathodic voltage limit to 0.8 V, the CV of films having a pseudo-capacitive shape moved toward a faradic form (Fig. 9g). As previously shown [21], the films have a pseudo-capacitive behavior over the 0.3– 0.3 V range as deduced from the linear variation of the current as a function of the scan rate
Fig. 9. Voltammograms of ATO half-cells cycling against Pt in 0.1 M H3PO4 electrolyte with various antimony rates: (a) 10%, (b) 30%, (c) 40%, (d) 50%, (e) 60%, and (f) 70%. CVs were recorded between 0.7 and 1.5 V versus SCE with a sweep rate of 10 mV/s, except for a film containing 40% Sb (g), whose CV was enlarged between 0.8 and 1.5 V.
N. Naghavi et al. / Solid State Ionics 156 (2003) 463–474
Fig. 10. Optical transmittance spectra of thin films of compositions (a) 60% Sn – 40% Sb and (b) 40% Sn – 60% Sb, bleached at 1.5 V, and cathodically colored at 0.7 V.
increase from 5 to 50 mV/s, while for other voltage domains, the variation of the current density j as a function of the scan rate r yielded a linear law j=(r)1/2 [21]. In fact, beyond 0.3 to 0.3 V, faradic currents could be due either to decomposition of the solution and H electrodeposition or to changes of oxidation state of the electrode material. This behavior is confirmed by light coloration changes due to the reduction of Sb5+ ! Sb3+ at a potential lower than 0.7 V and the presence of the peak at 0 V, corresponding to the partial oxidation of Sb3+ ! Sb5+ . The explanation of these two types of curve could go as follows. For specimens having a capacitive-like behavior (typically those containing 40% Sb), we experienced that the charge density increased with cycling. The origin of this cycling-driven activation is most likely nested in the nongranular film nature that limits the
471
proton diffusion, as neither the surface morphology nor the thickness were modified after cycling [20]. In fact, TEM measurement on films containing 40% Sb showed no real change in their morphology after about 100 cycles (Fig. 11a). These films remained very dense leading to a slow proton diffusion that can explain the presence of two broad peaks located at about 1 V in oxidation and at 0.3 V in reduction for films with Sb content < 50%. With respect to applications, it is worth noting that such films are cycled for more than 8 months. On the other hand, films with composition z 60% Sb lost their capacity after cycling. After about 100 cycles, TEM micrographs indicated a loss of density associated to a better grain definition (Fig. 11b), which may facilitate the proton diffusion leading to a well-defined faradic behavior. Suggesting that these films tend to slightly dissolve in H3PO4 electrolyte, the resulting cycling lifetime is apparently short. To circumvent this issue, we modified the electrolyte by changing either the concentration or its nature. Another interesting result is the evolution of the capacity as a function of antimony contents. For each composition, we reported in Fig. 12 the ‘volumic’ capacity Q (i.e., the charge density divided by film thickness) as a function of antimony content. Q+ symbolizes the capacity upon oxidation, whereas Q is written for the capacity calculated over the cathodic sweep. An increase in the antimony content is correlated to an increase in capacity, up to 50% (69
Fig. 11. Evolution of the ‘volume’ charge Q (mC/cm2/Am) versus the antimony amount ( Q + is written for anodic scan, and Q for cathodic scan).
472
N. Naghavi et al. / Solid State Ionics 156 (2003) 463–474
Fig. 12. TEM micrographs of ATO thin films of compositions of (a) 60% Sn – 40% Sb and (b) 40% Sn – 60% Sb after about 100 cycles.
mC/cm2/Am), but goes back to decreasing again. It should be noted that Q is always superior to Q + due to the unavoidable hydrogen production at the end of the reduction process. The decrease in capacity for films with a composition Sb>50% is surely linked to
the decrease in density of the films leading to a decrease in the electrochemically active materials. At last, Fig. 13 summarizes the overall electrochromic behavior for ATO thin films when cycled between 0.7 and + 1.5 V. While SnO2 stays color-
N. Naghavi et al. / Solid State Ionics 156 (2003) 463–474
473
Fig. 13. Overall representation of the ATO thin film electrochromic properties. Schematic bubbles are drawn to indicate the presence of hygroscopic antimony oxide.
less upon reduction, films containing 10 – 30% Sb are too dark to exhibit any optical contrast. From 40% to 70% Sb range, the optical contrast is visible with a maximum at 70% Sb, but from this last composition, films no longer stick to the substrate due to hygroscopic antimony oxide traces. In the 80 –100% Sb domain, films are not electrochromic.
capacity retention. Those features enable them to be very good candidates in their use as cheap counter electrodes in proton-working devices. The feasibility of an all-solid electrochromic window constituted by WO3 blue-switching electrode exchanging H+ ions with an ATO film containing 40% Sb is being studied.
4. Conclusion
Acknowledgements
In summary, through a systematic study of Sn – Sb – O thin films, we gave evidence for a direct relationship between the ATO thin film electrochromic behavior and morphology via changes in Sb content. Optical study emphasized that post-annealing was responsible for a partial lowering of an antimony oxidation state leading to film darkening. Indeed, the presence of concomitant Sb5+ and Sb3+ levels causes an intense light absorption and is also responsible for the cathodic coloration revealed in a proton electrolyte over the 40– 70% Sb range. Based on a charge transfer between Sb3+ and Sb5+, the optical contrast between bleached/colored states is thus dictated by the antimony concentration in tin oxide matrix. It increases to the detriment of electrochemical stability, which yields the most antimony-rich compositions (60 – 70% Sb) presenting a typical faradic behavior to be unsuitable proton-conducting switching electrodes. On the other hand, films containing 40% Sb present a neutral coloration over a wide potential scale associated to a capacitive-like behavior and good
The authors would like to give special thanks to A. Rougier and C. Gue´ry for helpful discussions.
References [1] P.M.S. Monk, R.J. Mortimer, D.R. Rosseinsky, Electrochromism Fundamentals and Applications, VCH, New York, 1995. [2] C.G. Granqvist, Handbook of Inorganic Electrochromic Materials, Elsevier, Amsterdam, 1995. [3] F.G. Baucke, J.A. Duffy, Chem. Br. 21 (1985) 643. [4] B. Buffat, F. Lerbet, Recherche 231 (21) (1991) 434. [5] S.K. Deb, Philos. Mag. 29 (1973) 801. [6] S.K. Deb, Appl. Opt., Suppl. 3 (1969) 192. [7] K. Machida, M. Enyo, J. Electrochem. Soc. 137 (1990) 1169. [8] C.G. Granqvist, Solid State Ionics 70/71 (1994) 678. [9] F. Artuso, G. Picardi, F. Bonino, F. Decker, S. Bencic, A. Surca Vuk, U. Opara Krasovec, B. Orel, Electrochim. Acta 46 (2001) 2077. [10] M.A. Petit, V. Plichon, J. Electroanal. Chem. 444 (1998) 247. [11] A. Azens, L. Kullman, G. Vairars, H. Nordborg, C.G. Granqvist, Solid State Ionics 449 (1998) 113. [12] H.L. Hartnagel, A.L. Dawar, A.K. Jain, C. Jagadish, Semiconducting Transparent Thin Films, Institute of Physics Publishing, VCH, New York, 1995.
474
N. Naghavi et al. / Solid State Ionics 156 (2003) 463–474
[13] S. Shanti, C. Subramanian, R. Ramasamy, J. Cryst. Growth 197 (1999) 858. [14] D.U. Ju, J.H. Chung, D.Y. You, S.G. Kim, Ind. Eng. Chem. Res. 37 (1998) 1827. [15] P. Olivi, E.C. Periera, E. Longo, J.A. Varella, J. Electrochem. Soc. 140 (1993) L81. [16] B. Sterjna, E. Olsson, C.G. Ganqvist, J. Appl. Phys. 76 (1994) 6. [17] M. Yoshizumi, K. Wakabayashi, Plast. Eng. 43 (1987) 61. [18] D.S. Vlachos, C.A. Papadopoulos, J.N. Avaritsiotis, Sens. Actuators B24 (1995) 883.
[19] J. Yang, Y. Takeda, N. Imanishi, O. Yamamoto, J. Electrochem. Soc. 146 (1999) 4009. [20] J.P. Coleman, A.T. Lynch, P. Madhukar, J.H. Wagenknecht, Sol. Energy Mater. Sol. Cells 56 (1999) 375. [21] C. Marcel, M.S. Hegde, A. Rougier, C. Maugy, C. Gue´ry, J.M. Tarascon, Electrochim. Acta 46 (2001) 2097. [22] F.M. Smith, Bell Syst. Tech. J., (1958) 37711. [23] F.J. Berry, D.J. Smith, J. Catal. 88 (1984) 107. [24] M. Kojima, H. Kato, M. Gatto, J. Non-Cryst. Solids 218 (1997) 230.
On the electrochromic properties of antimony–tin oxide thin films deposited by pulsed laser deposition N. Naghavi, C. Marcel, L. Dupont, J-B. Leriche, J-M. Tarascon* Laboratoire de Re´activite´ et Chimie des Solides, Universite´ de Picardie Jules Verne, 33 rue Saint Leu, F-80039 Amiens Cedex, France Received 5 April 2002; received in revised form 9 July 2002; accepted 19 August 2002
Abstract Antimony – tin oxide (ATO) thin films were deposited using pulsed laser deposition. From a wide survey of deposition conditions (deposition atmosphere and substrate temperature), we deduced that the optimum experimental parameters to obtain electrochromic active films were a deposition temperature of 200 jC at a 10 2 mbar oxygen pressure followed by an annealing at 550 jC, the latter being critical. Through potentiostatic tests, we showed that depending on the composition, which influences film morphology, the Sn – Sb – O films could present either a faradic or a capacitive-like behavior, associated to a color or a neutral switching over a wide range of potentials, respectively. In addition, a relationship between the film annealing temperature and its structure as well as its electrochromic properties was evidenced by means of complementary electrochemical, X-ray diffraction and electron microscopy techniques. Finally, the electrical and optical properties of the ATO films were discussed in terms of the energy band theory. D 2003 Elsevier Science B.V. All rights reserved. Keywords: Antimony – tin oxide; Pulsed laser deposition; Electrochromic properties
1. Introduction In recent years, electrochromic devices have received considerable attention as they are able to reversibly change their optical properties when an electric current flows through the device [1,2]. Due to this feature, they find applications in rearview mirrors [3], «smart windows» [2], display panels [4], etc. The most extensively studied solid-state electrochromic systems are based on tungsten oxide [5– 8] *
Corresponding author. Fax: +33-32-2827590. E-mail address: [email protected] (J.-M. Tarascon).
working electrode thin films, namely a proton or lithium insertion material well known for its blue electrochromism. With respect to the counter electrodes, two options that consist in using either optically neutral (e.g., CeVO4 [9]) or complementary colored electrodes such as IrO2 [10] or hydrous Ni oxide [11] are being pursued. However, the latter, exclusively used in proton-functioning devices, present non negligible drawbacks such as a high cost for the first one and a poor stability in acidic electrolytes for the second one. Thus, alternative materials are required for proton-conducting counter electrodes in electrochromic windows. Recently, a large number of studies have been devoted to low Sb-doping in ATO compounds for their
0167-2738/03/$ - see front matter D 2003 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 7 - 2 7 3 8 ( 0 2 ) 0 0 7 4 9 - X
464
N. Naghavi et al. / Solid State Ionics 156 (2003) 463–474
transparent conducting properties [12]. Such materials, which can be grown in thin film forms either by spray pyrolysis [13], dip coating [14,15] or sputtering [16], present a great interest with respect to antistatic coatings [17], gas sensors [18] and energy storage applications [19]. However, the electrochromic properties of ATO thin films containing high antimony content received very little attention with a few exceptions. Coleman et al. [20] demonstrated for the first time the feasibility of fast switching devices for display application made with ATO fine grains embedded in a fluoro-elastomer resin. In light of these promising results, Marcel et al. [21] studied the electrochromic properties of Sn – Sb– O powders and thin films grown by pulsed laser deposition. They pointed out that depending on the deposition conditions, ATO films could either exhibit a grey reversible coloration upon reduction in relation with a faradic behavior, or present a neutral switching over a wide range of potentials associated to a capacitive-like behavior. The present paper aims at providing more information about the influence of deposition conditions (especially postannealing temperature and Sn – Sb – O composition changes) on the electrochemical, structural and optoelectronic properties of the resulting films in the hope of identifying the origin of their competing faradic/ capacitive-like behavior.
2. Experimental Thin oxide films were prepared by pulsed laser deposition using a KrF excimer laser beam (Lambda Physic, Compex 102, k = 248 nm) with a laser fluence of 1 –2 J/cm2. The targets were pellets of commercial SnO2 (Aldrich 99.9%) and Sb2O3 (Aldrich 99.9%) mixtures in stoichiometric proportions annealed for 20 h at a relatively low temperature (700 jC) in order to avoid any antimony loss while achieving a pellet density of 60 – 70%. Films of 1 1-cm2 area were deposited either on (SnO2:F)-coated glass for electrochemical tests or simple glass for structural and electronic characterizations. Deposition time was fixed at 15 min with a repetition rate of 3 Hz. The thickness, determined by profilometry using a Dektak St instrument, was estimated as being in the range of 200– 250 nm. Film crystallinity was examined by X-ray diffraction (XRD) with a Philips diffractometer model PW
1710 (kCuKa = 0.15418 nm). The surface morphology and film composition were investigated by scanning electron microscopy (SEM) with a Philips XL 30 field emission gun (FEG) coupled to an Oxford Link instrument for energy-dispersive X-ray spectroscopy (EDS). High-resolution transmission electron microscopy (HRTEM) was carried out using a JEOL 2010 microscope equipped with an EDS analyzer. Optical transmission spectra in the UV – visible and near infrared regions (250 – 2500 nm) were obtained using a Varian double beam UV – Vis – NIR spectrometer ‘‘CARY-5E’’. The electronic conductivity of the films was taken as the inverse function of the resistivity, which was determined by the four-point probe method using the formula q=(V/ I )C(a/d, d/s), as reported by Smith [22]. Voltages and DC currents were measured with a multimeter (Keitley 175 A). Within the above formula, V is the voltage and I is the current, a and d are the sample lengths (in our case a = d ), and C(a/d, d/s) is a correction factor accounting for the sample geometry. The resistivities were obtained by assuming that our samples were of the same thickness and dimensions. Therefore, such values should not be taken as quantitative but as qualitative values enabling to compare our variety of samples having different compositions or anneals. The electrochemical properties of the films were characterized by cyclic voltammetry performed with an Autolab PGSTAT 30 system. The cell consists of the ATO film as the working electrode and of a platinum wire as the counter electrode, both immersed in a 0.1 M aqueous H3PO4 electrolyte. The potentials were given versus the saturated calomel electrode (SCE) reference. For all these experiments, cyclic voltammograms (CV) were scanned at a rate of 10 mV/s.
3. Results 3.1. Influence of the deposition conditions Our previous work dealing with the ATO system [21] showed that optimum electrochromic properties were achieved for films made from removing a (70% Sn – 30% Sb or 60% Sn – 40% Sb) target composition under an oxygen pressure of 10 1 and 10 2 mbar for
N. Naghavi et al. / Solid State Ionics 156 (2003) 463–474
faradic and capacitive-like behavior, respectively. First, we assumed that the difference between these electrochemical processes was nested in film morphology (namely, the porosity). To test this hypothesis, we studied the influence of temperature and oxygen pressure on film morphology during deposition and annealing steps, and thereby on the film electrochemical/electrochromic properties. While maintaining both film composition (60% Sn –40% Sb) and pressure (10 2 mbar), we noted that a slight grey cathodic coloration could only be generated if films were first grown at 200 jC, and then post-annealed for 30 min at 550 jC. Indeed, attempts to directly grow films at 500 jC resulted in polycrystalline films, which were black and electrochemically inactive. In contrast, amorphous films, which were colorless with no electrochemical activity in H3PO4 liquid electrolyte, were obtained when the growth temperature was limited at 200 jC. It should be noted that the annealing step resulting in film crystallization was accompanied by a slight darkening. While maintaining the same target composition and growth protocol (deposition at 200 jC followed by a 550 jC annealing step), we varied the oxygen pressure between 10 4 and 10 1 mbar. From pressures smaller than 10 2 mbar, all the deposits were electrochemically active. In contrast, the films grown at pressures greater than 10 2 mbar were electrochemically inactive, while being both less crystalline and more porous, as deduced from XRD and SEM measurements, respectively. Moreover, EDS analysis revealed that increasing the deposition pressure resulted in ATO films having an antimony content greater than the target nominal composition. In light of such findings, we checked the importance of the Sn/Sb ratio on film morphology and electrochemical properties. 3.2. Influence of the Sb content Films of various antimony – tin oxide compositions were grown from targets whose nominal composition ranged between 0% and 80% at Sb. More specifically, the films were deposited at 200 jC under 10 2 mbar, annealed for 30 min at 550 jC under the same oxygen pressure, and characterized for their electrochromic properties. Under such deposition conditions, as deduced from EDS analysis, we noted that the Sn/Sb
465
atomic ratio was relatively well preserved between the target and grown films. 3.2.1. Structural characterization SEM micrographs of Sn-Sb-O films deposited on FTO-glass substrate are shown in Fig. 1 for various antimony contents. While films containing less than 50% Sb appear dense and homogenous, those having higher Sb contents reveal the presence of cracks underlining a porosity, thereby implying a decrease in film density. For Sb percentages greater than 80%, the deposits were poorly sticking to the substrate, making their characterization meaningless. Note that the apparent porosity observed for dense films with low Sb contents is a visual effect due to FTO texture. Fig. 2 shows the X-ray diffraction patterns of ATO thin films deposited on glass substrate, which are typical of SnO2 cassiterite phase. Increasing Sb amount yields a loss of crystallinity; this trend being also observed for ATO thin films deposited on FTO substrate. The absence of peaks pertaining to antimony oxide phases (whatever the oxidation states) lets suppose the complete dissolution of antimony in the nanostructured SnO2. Beyond 70% Sb, films become amorphous. Purely coincidental or not, ATO nanoparticles with a similar Sb content and prepared by our group via a solution route became amorphous [21]. Such results could first indicate that the ‘‘apparent solubility limit’’ of antimony in nanocrystalline tin oxide is reached near the 70% Sb. At this point, however, a legitimate question was whether these Sb-enriched ATO specimens presented large amounts of amorphous antimony-containing material as previously reported [23] or were single phases. To unambiguously answer such a question, HRTEM measurements were carried out on our heavily substituted ATO films. Fig. 3a,b display HRTEM micrographs of films deposited from targets composed of 60% Sb and 40% Sb, respectively. The high-resolution images and the electron diffraction patterns recorded from these materials gave no evidence for the presence of any antimony oxide phase. For both compositions, the diffraction pattern contains well-resolved diffraction rings characteristic of an SnO2 cassiterite phase. Indeed, films consist of small disordered crystals of a cassiterite-type phase and no trace of amorphous
466
N. Naghavi et al. / Solid State Ionics 156 (2003) 463–474
Fig. 1. SEM micrographs of ATO thin films with antimony contents ranging from 10% to 80% Sb: (a) 20%; (b) 40%; (c) 50%; (d) 60%; (e) 70%; and (f) 80%.
materials is observed independently of the antimony amount. However, changes in film texture with composition are visible. While the film with the lowest antimony content (40% Sb, Fig. 3b) is very dense and exhibits a poor definition of particle shape, the film containing 60% Sb (Fig. 3a) appears less dense with ˚ . In short, the better defined grains of about 50 A TEM study confirmed the formation of Sn –Sb – O solid solution with a SnO2 cassiterite-type structure up to a concentration of 60% Sb. 3.2.2. Electrical and optical properties The effect of the Sb-content on film conductivity is shown in Fig. 4. The conductivity initially increases with an antimony percentage up to 10% Sb, and then continuously decreases for higher Sb contents. The initial conductivity increase was ascribed [12] to an increase in the carrier concentration as a result of Sb doping. Indeed, as commonly reported for low Sb amounts, Sb5+ ions provide an n-type doping (electrons) within the conduction band [12 –16]. Note that in our case, Sn0.9Sb0.1O2 films were not as conductive as a classical transparent conducting oxide (TCO) whose conductivity averages 103 – 104 S/cm. This is most likely due to the fact that the two-step heat treatment
Fig. 2. X-ray diffraction patterns of ATO thin films with 10 V %Sb V 60 having a SnO2-cassiterite structure.
N. Naghavi et al. / Solid State Ionics 156 (2003) 463–474
Fig. 3. TEM micrographs of ATO thin films with composition of (a) 60% Sb and (b) 40% Sb.
467
468
N. Naghavi et al. / Solid State Ionics 156 (2003) 463–474 Table 1 Evolution vs. Sb percentage of the threshold temperature for which the resistivity suddenly drops during postannealing in air for ATO thin films deposited at low temperature (200 jC)
Fig. 4. Variation of the conductivity (r) as a function of antimony content of ATO films deposited at 200 jC, then annealed at 550 jC.
process that we used to induce electrochemical properties slightly differs from the best established growth protocol to achieve optimum conductivity. For higher Sb content, the observed conductivity downfall may have different explanations. It may be due either to the existence of traps created by crystal defects increasing with Sb content thus dominating the electron scattering mechanism or to the deep Sb3+ levels that act as an electron trap (as proposed by Kojima et al. [24] from XPS measurements results). In fact, a realistic scenario involves the simultaneous presence of both Sb5+ and Sb3+ ions, which can be described as: 5þ 3þ Sn4þ 1x Sbxy Sby O2ðx=2yÞ
Thus, concomitant with the Sb substitution for Sn4+ ions would be the appearance of Sb5+ and
Fig. 5. Variation of resistivity (R) of a film containing 10 – 40% Sb as a function of the temperature (T).
Composition
10% Sb
20% Sb
30% Sb
40% Sb
Threshold temperature (jC)
300
450
500
510
Sb3+ species that would lead to a noticeable lowering of the electron mobility consistent with the observed resistivity drop. The electrical properties of films containing 10– 40% Sb deposited at 200 jC as a function of the temperature are shown in Fig. 5. Whatever their antimony content the as-made films were all insulating. Nevertheless, upon annealing them in air, their resistivity sharply dropped over a temperature range of 300 –510 jC with increasing antimony contents (Table 1). This temperature-driven drastic conductivity onset could be explained as a consequence of the film crystallization that leads to a reorganization of the band structure, and thereby the creation of Sb+5 donor levels below the bottom of the conduction band. Fig. 6 shows the transmission spectra in the UV – visible and near infrared regions for films having Sb content ranging from 10% to 60% Sb. While films grown at 200 jC were amorphous and highly transparent independent of composition, the same films annealed at 550 jC exhibited a transmittance greatly influenced by the Sb amount. The lowest transmittance pertains to (90% Sn –10% Sb) sample and the highest to the most Sb-charged composition (i.e., 40% Sn–
Fig. 6. Optical transmission spectra of a film deposited at 200 jC (30% Sb) and post-annealed ATO thin films with 0 V %Sb V 60.
N. Naghavi et al. / Solid State Ionics 156 (2003) 463–474
Fig. 7. Variation of the optical band gap (Eg) estimated from optical spectra as a function of %Sb for polycrystalline films annealed at 550 jC. Band gaps of 4.05 and 3.2 – 3.5 eV are reported for SnO2 and Sb2O3, respectively (see oxide Handbook).
60% Sb). Thus, film coloration varies from light grey to colorless with raising antimony content. Finally, note that the transmittance drop observed in the infrared part is more pronounced for low Sb rates. This is totally consistent with our conductivity measurements since, according to Drude’s model based on free carrier absorption, the transmittance should be the lowest for the highest conducting films. Regarding the UV region, the transmittance falls very sharply due to the onset of fundamental absorption associated to band-gap energy (Eg). For each composition, we deduced the Eg values (Fig. 7) by extrapolation of the linear portion of a2 versus hm for crystallized annealed films, where a is the absorption
469
coefficient. After post-annealing, we observed an overall widening of the band gap presenting a minimum at about 30% of Sb. Beyond 30%, the band gap widening is a natural result of the observed transmittance enhancement with increasing the antimony content. The observed heat treatment-driven optical film behavior is still open to discussions. Some authors suggested that antimony-segregated atoms form clusters in the lattice resulting in a darkening of the films [13]. Others could evidence an antimony-rich amorphous phase, interpreted as Sb2O4, yielding a color lightening [23] of the ATO sample. However, none of these assumptions could be corroborated by our HRTEM study. Therefore, based on our results in conjunction with Kojima et al.’s [24] studies, we propose the following scenario. According to these authors, antimony firstly contributes in low temperature deposited ATO thin films to the widening of the conduction and valence bands, but not to the formation of light absorbing levels in the energy gap consistent with the transparency of the Sn– Sb –O films deposited at 200 jC. After a heat treatment, the sharp decrease in resistivity depicted in the previous section is due to the formation of Sb5+ energetic levels overlapping the bottom of the conduction band (Fig. 8a). However, since the measured resistivities were inferior to TCOs in low Sb content samples because of the poor crystalline nature of the films, the darkening coloration was possibly due to the presence of Sb5+ impurity levels lying close to the bottom of the conduction band. Beyond 10% Sb
Fig. 8. Schematic representation of the band structure for ATO films annealed at 550 jC (polycrystalline) with different antimony amounts: (a) 10% Sb, (b,c) 20 – 40% Sb, and (d) 60% Sb.
470
N. Naghavi et al. / Solid State Ionics 156 (2003) 463–474
content range, the presence of Sb3+ reinforces light absorption until 30% Sb by creating supplementary defects causing a band-gap reduction (Fig. 8b,c). With further increasing Sb content, we enhance the number of Sb3+ ions that may act as traps for conductive electrons, and thereby form localized states within the SnO2 band gap. These levels may become deeper and deeper with increasing antimony amounts, and generate less and less light absorption, leading to a progressive bleaching of the films (Fig. 8d). 3.2.3. Electrochemical characterization Cyclic voltammograms (CVs) of our ATO films were obtained by sweeping the potential from 0.7 to 1.5 V versus SCE. Depending on the film Sb content, two types of CV traces were observed (Fig. 9). Within the 10 – 40% Sb range, CV curves are featureless adopting a pseudo-capacitive-like shape as previously shown by Marcel et al. [21], and
furthermore, the transmittance remains the same as shown for the (60% Sn – 40% Sb) sample in Fig. 10a. From 60% to 70% Sb content, a pronounced faradic behavior appears, possibly corresponding to (Sb5+ X Sb3+ ) charge transfer with simultaneous H+ (de)intercalation. Indeed, a redox couple, visible at ca. 0 V in oxidation and 0.7 V in reduction, produces a dark-grey coloration with a cathodic sweep, followed by a bleaching process with an anodic one (see Fig. 9e). Fig. 10b shows that the optical contrast for the composition (40% Sn – 60% Sb) is reinforced with increasing Sb amount. It is interesting to note that upon further decreasing the cathodic voltage limit to 0.8 V, the CV of films having a pseudo-capacitive shape moved toward a faradic form (Fig. 9g). As previously shown [21], the films have a pseudo-capacitive behavior over the 0.3– 0.3 V range as deduced from the linear variation of the current as a function of the scan rate
Fig. 9. Voltammograms of ATO half-cells cycling against Pt in 0.1 M H3PO4 electrolyte with various antimony rates: (a) 10%, (b) 30%, (c) 40%, (d) 50%, (e) 60%, and (f) 70%. CVs were recorded between 0.7 and 1.5 V versus SCE with a sweep rate of 10 mV/s, except for a film containing 40% Sb (g), whose CV was enlarged between 0.8 and 1.5 V.
N. Naghavi et al. / Solid State Ionics 156 (2003) 463–474
Fig. 10. Optical transmittance spectra of thin films of compositions (a) 60% Sn – 40% Sb and (b) 40% Sn – 60% Sb, bleached at 1.5 V, and cathodically colored at 0.7 V.
increase from 5 to 50 mV/s, while for other voltage domains, the variation of the current density j as a function of the scan rate r yielded a linear law j=(r)1/2 [21]. In fact, beyond 0.3 to 0.3 V, faradic currents could be due either to decomposition of the solution and H electrodeposition or to changes of oxidation state of the electrode material. This behavior is confirmed by light coloration changes due to the reduction of Sb5+ ! Sb3+ at a potential lower than 0.7 V and the presence of the peak at 0 V, corresponding to the partial oxidation of Sb3+ ! Sb5+ . The explanation of these two types of curve could go as follows. For specimens having a capacitive-like behavior (typically those containing 40% Sb), we experienced that the charge density increased with cycling. The origin of this cycling-driven activation is most likely nested in the nongranular film nature that limits the
471
proton diffusion, as neither the surface morphology nor the thickness were modified after cycling [20]. In fact, TEM measurement on films containing 40% Sb showed no real change in their morphology after about 100 cycles (Fig. 11a). These films remained very dense leading to a slow proton diffusion that can explain the presence of two broad peaks located at about 1 V in oxidation and at 0.3 V in reduction for films with Sb content < 50%. With respect to applications, it is worth noting that such films are cycled for more than 8 months. On the other hand, films with composition z 60% Sb lost their capacity after cycling. After about 100 cycles, TEM micrographs indicated a loss of density associated to a better grain definition (Fig. 11b), which may facilitate the proton diffusion leading to a well-defined faradic behavior. Suggesting that these films tend to slightly dissolve in H3PO4 electrolyte, the resulting cycling lifetime is apparently short. To circumvent this issue, we modified the electrolyte by changing either the concentration or its nature. Another interesting result is the evolution of the capacity as a function of antimony contents. For each composition, we reported in Fig. 12 the ‘volumic’ capacity Q (i.e., the charge density divided by film thickness) as a function of antimony content. Q+ symbolizes the capacity upon oxidation, whereas Q is written for the capacity calculated over the cathodic sweep. An increase in the antimony content is correlated to an increase in capacity, up to 50% (69
Fig. 11. Evolution of the ‘volume’ charge Q (mC/cm2/Am) versus the antimony amount ( Q + is written for anodic scan, and Q for cathodic scan).
472
N. Naghavi et al. / Solid State Ionics 156 (2003) 463–474
Fig. 12. TEM micrographs of ATO thin films of compositions of (a) 60% Sn – 40% Sb and (b) 40% Sn – 60% Sb after about 100 cycles.
mC/cm2/Am), but goes back to decreasing again. It should be noted that Q is always superior to Q + due to the unavoidable hydrogen production at the end of the reduction process. The decrease in capacity for films with a composition Sb>50% is surely linked to
the decrease in density of the films leading to a decrease in the electrochemically active materials. At last, Fig. 13 summarizes the overall electrochromic behavior for ATO thin films when cycled between 0.7 and + 1.5 V. While SnO2 stays color-
N. Naghavi et al. / Solid State Ionics 156 (2003) 463–474
473
Fig. 13. Overall representation of the ATO thin film electrochromic properties. Schematic bubbles are drawn to indicate the presence of hygroscopic antimony oxide.
less upon reduction, films containing 10 – 30% Sb are too dark to exhibit any optical contrast. From 40% to 70% Sb range, the optical contrast is visible with a maximum at 70% Sb, but from this last composition, films no longer stick to the substrate due to hygroscopic antimony oxide traces. In the 80 –100% Sb domain, films are not electrochromic.
capacity retention. Those features enable them to be very good candidates in their use as cheap counter electrodes in proton-working devices. The feasibility of an all-solid electrochromic window constituted by WO3 blue-switching electrode exchanging H+ ions with an ATO film containing 40% Sb is being studied.
4. Conclusion
Acknowledgements
In summary, through a systematic study of Sn – Sb – O thin films, we gave evidence for a direct relationship between the ATO thin film electrochromic behavior and morphology via changes in Sb content. Optical study emphasized that post-annealing was responsible for a partial lowering of an antimony oxidation state leading to film darkening. Indeed, the presence of concomitant Sb5+ and Sb3+ levels causes an intense light absorption and is also responsible for the cathodic coloration revealed in a proton electrolyte over the 40– 70% Sb range. Based on a charge transfer between Sb3+ and Sb5+, the optical contrast between bleached/colored states is thus dictated by the antimony concentration in tin oxide matrix. It increases to the detriment of electrochemical stability, which yields the most antimony-rich compositions (60 – 70% Sb) presenting a typical faradic behavior to be unsuitable proton-conducting switching electrodes. On the other hand, films containing 40% Sb present a neutral coloration over a wide potential scale associated to a capacitive-like behavior and good
The authors would like to give special thanks to A. Rougier and C. Gue´ry for helpful discussions.
References [1] P.M.S. Monk, R.J. Mortimer, D.R. Rosseinsky, Electrochromism Fundamentals and Applications, VCH, New York, 1995. [2] C.G. Granqvist, Handbook of Inorganic Electrochromic Materials, Elsevier, Amsterdam, 1995. [3] F.G. Baucke, J.A. Duffy, Chem. Br. 21 (1985) 643. [4] B. Buffat, F. Lerbet, Recherche 231 (21) (1991) 434. [5] S.K. Deb, Philos. Mag. 29 (1973) 801. [6] S.K. Deb, Appl. Opt., Suppl. 3 (1969) 192. [7] K. Machida, M. Enyo, J. Electrochem. Soc. 137 (1990) 1169. [8] C.G. Granqvist, Solid State Ionics 70/71 (1994) 678. [9] F. Artuso, G. Picardi, F. Bonino, F. Decker, S. Bencic, A. Surca Vuk, U. Opara Krasovec, B. Orel, Electrochim. Acta 46 (2001) 2077. [10] M.A. Petit, V. Plichon, J. Electroanal. Chem. 444 (1998) 247. [11] A. Azens, L. Kullman, G. Vairars, H. Nordborg, C.G. Granqvist, Solid State Ionics 449 (1998) 113. [12] H.L. Hartnagel, A.L. Dawar, A.K. Jain, C. Jagadish, Semiconducting Transparent Thin Films, Institute of Physics Publishing, VCH, New York, 1995.
474
N. Naghavi et al. / Solid State Ionics 156 (2003) 463–474
[13] S. Shanti, C. Subramanian, R. Ramasamy, J. Cryst. Growth 197 (1999) 858. [14] D.U. Ju, J.H. Chung, D.Y. You, S.G. Kim, Ind. Eng. Chem. Res. 37 (1998) 1827. [15] P. Olivi, E.C. Periera, E. Longo, J.A. Varella, J. Electrochem. Soc. 140 (1993) L81. [16] B. Sterjna, E. Olsson, C.G. Ganqvist, J. Appl. Phys. 76 (1994) 6. [17] M. Yoshizumi, K. Wakabayashi, Plast. Eng. 43 (1987) 61. [18] D.S. Vlachos, C.A. Papadopoulos, J.N. Avaritsiotis, Sens. Actuators B24 (1995) 883.
[19] J. Yang, Y. Takeda, N. Imanishi, O. Yamamoto, J. Electrochem. Soc. 146 (1999) 4009. [20] J.P. Coleman, A.T. Lynch, P. Madhukar, J.H. Wagenknecht, Sol. Energy Mater. Sol. Cells 56 (1999) 375. [21] C. Marcel, M.S. Hegde, A. Rougier, C. Maugy, C. Gue´ry, J.M. Tarascon, Electrochim. Acta 46 (2001) 2097. [22] F.M. Smith, Bell Syst. Tech. J., (1958) 37711. [23] F.J. Berry, D.J. Smith, J. Catal. 88 (1984) 107. [24] M. Kojima, H. Kato, M. Gatto, J. Non-Cryst. Solids 218 (1997) 230.