Electrochromic properties of vacuum-evaporated organic thin films

197

J. Electroanal. Chem., 283 (1990) 197-204 Elsevier Sequoia S.A., Lausanne - Printed

Electrochromic films

in The Netherlands

properties of vacuum-evaporated

Part 3. The case of 4,4’-dicyanophenyl

organic thin

viologen

Akio Yasuda and Jun’etsu Seto Sony Corporation Research Center, I74 Fujitsuka-rho, (Received

15 August

Hodogaya-ku,

1989; in revised form 28 November

Yokohama 240 (Japan)

1989)

ABSTRACT Electrochemical studies on the vacuum-evaporated film of a viologen (4,4’-dicyanophenyl-bipyridinium dichloride, abbreviated as Cyv) are described. CyV was found to form a charge-transfer complex with Ru(CN)z(abbreviated as Ru) in the solid and in the solution state. Vacuum-evaporated thin films of CyV and CyV-Ru could be formed and they were found to retain their electrochromic properties through vacuum evaporation. During the vacuum evaporation, CyV and CyV-Ru were considered to be reduced, because the colours of the films coincided with those of their reduced states. Thermogravimetry-differential thermal analysis (TG-DTA) suggested a chemical reaction in CyV during the evaporation. The electrochromism of the CyV film was found to depend on the size of the counter-cation for charge compensation. Moreover, the counter-cations were considered to be injected without their hydration sheaths.

INTRODUCTION

Studies on the electrochromic properties of viologens in the solution state have been reported previously [l-4]. One of the most important reasons for studying electrochromics is the realization of electrochromic displays (ECDs). Considering the applications of ECDs, all-solid-state type ECDs are desirable. To this end, an electrochromic thin film should be formed by a dry process. Therefore, the formation of a viologen film on the electrode using vacuum evaporation was investigated, because the vacuum evaporation technique has the advantages that the process is completely dry, and that patterning can be made easily. It was necessary to select a viologen which is durable for vacuum evaporation. Following the screening of viologens for vacuum evaporation, N, N ‘-dicyanophenyl-4,4’-bipyridinium dichloride (Cyv) was selected. 0022-0728/90/$03.50

0 1990 - Elsevier Sequoia

S.A.

198

CYV Fig. 1. Redox reaction of CyV.

CyV shows good electrochromism with a colour change from transparent to green in an aqueous electrolyte solution by one-electron reduction (Fig. 1). and a colour change from green to red-violet by two-electron reduction. It was expected that the electrochromic properties of CyV would remain even in a thin film state. It was found that a vacuum-evaporated film of CyV in an as-deposited state was coloured green or red-violet. This may indicate that CyV was reduced during the vacuum evaporation (heat resistance method). Moreover, the evaporated CyV film was found to show a redox response and electrochromism in contact with an electrolyte solution. On the formation of a charge-transfer (CT) complex between a viologen and ferrocyanide [4], CyV was known to form a CT complex with ferrocyanide [5]. CyV was found to form a CT complex also with Ru(CN);f-. Furthermore, it was possible to evaporate the CyV-Ru(CN)icomplex (CyV-Ru) and CyV-Fe(CN)z(CyV-Fe) to form electrochromic thin films in their reduced state. In this paper, the spectroelectrochemical and electrochemical properties of the newly-developed vacuum-evaporated thin films of CyV and CyV-Ru are discussed. EXPERIMENTAL

Reagents CyV was supplied from the Japanese Research Institute Dyes Co. Ltd. The white powder of CyV was used without Other reagents were all commercially available.

for Photosensitizing further purification.

Instrumentation For cyclic voltammetry, a Hokuto Denko HA-301 potentio/galvanostat combined with a HB-105 function generator was used. The working electrode was IT0 (indium tin oxide). The reference electrode was a Ag/AgCl electrode and the counter-electrode was a Pt plate. 1 M RbCl aqueous solution was the electrolyte solution. For investigating the evaporation process of CyV, thermogravimetry-differential thermal analysis (TG-DTA, Rigaku Denki 807861) was applied. During monitoring of the TG-DTA, a nitrogen atmosphere was maintained. The rate of the temperature increase was 150°C h-i and the reference was Al,Os. In-situ absorption changes and the absorption spectra of the CyV film and CyV-Ru were obtained by using a Hitachi 220A double-beam monochromator combined with a potentiostat and a function generator.

199

$! 0.5

0 500 Wavelength

700

900

h / nm

Fig. 2. Absorption spectra of charge-transfer complexes. The solid line represents line represents CyV-Fe. CyV: 0.01 M; Fe(CN)zand Ru(CN)i-: 0.01 M.

CyV-Ru;

the dotted

Vacuum evaporation technique Vacuum evaporation was performed by a Tokuda CT-6ST vacuum evaporator with a 40 cm diameter glass belljar. The distance between the evaporation boat and the substrate was 25 cm and a heat resistance method was used. The current range was lo-30 A at 10 V. Film thicknesses were monitored and were controlled by a quartz oscillator thickness monitor (Sloan DTM-200) during the evaporation.

RESULTS

AND DISCUSSION

The formation Fe(CN)i _

of a charge-transfer

(CT) complex

between

CyV and Ru(CN)i-

or

Absorption spectra of a mixture of CyV and Ru(CN)z(Ru), and a mixture of CyV and Fe(CN);f(Fe) aqueous solution are shown in Fig. 2. CyV, Ru or Fe themselves have no absorption in the visible region. It was suggested that the appearance of absorption in the visible region indicated the formation of a CT complex between CyV and Ru or Fe. In the case of heptylviologen and ferrocyanide CT complex, the absorption maximum appeared at 560 nm [I]. As shown in Fig. 2, the absorption maximum appeared at 530 nm for the CyV-Fe complex and at 430 nm for CyV-Ru. We reported that for the 4,4’-dibenzyl bipyridinium(dibenzylviologen) and ferrocyanide complex, the ratio of viologen and ferrocyanide was 2 to 1 [4]. For CyV and ferrocyanide, it was reported that the ratio of CyV to ferrocyanide was 2 to 1 [5]. The shorter wavelength of the absorption maximum of CyV-Ru suggests that CyV-Ru forms a stronger CT bond than CyV-Fe.

200

500

Fig. 3. Absorption about 25 nm.

Absorption

900

700

Wavelength spectra

h / nm

of CyV and CyV-Ru

in the as-deposited

state.

The film thicknesses

were

spectra of as-deposited films of CyV and CyV-Ru

Absorption spectra of as-deposited films of CyV and CyV-Ru are shown in Fig. 3. The film thicknesses of both films were about 25 nm. For the CyV film, absorption maxima were observed at 600 and 660 nm and the film was green, while for the CyV-Ru film a red-violet colour was observed. These colours of the films

60-l 22

I -4.0 50

100

150

200 Temperature

250

300

350

400

/ “C

Fig. 4. Thermogravimetry-differential thermal analysis of CyVzt 2 Cl- in a nitrogen atmosphere. weight of CyV 2+ 2 Cl- was 2.78 mg. The sample pan was Al. The rate of the temperature increase 150°C h-‘. The reference was A1203.

The was

201

may indicate the reduced state of CyV: green for the one-electron reduction state and red-violet for the two-electron reduction state. This suggests that during the vacuum evaporation, CyV was reduced. In order to investigate the process of vacuum evaporation, TG-DTA of CyV was performed. As shown in Fig. 4, in TG, CyV started to decrease in weight at around 300” C. Judging from the DTA, an exothermic reaction could be considered to occur, corresponding to the decrease in weight of CyV. If the polycrystalline sample dissolves first and then evaporation occurs, the evaporation process is such that an endothermic reaction should predominate. The exothermic reaction may indicate that a chemical reaction of reducing CyV occurred thermally in the CyV’+ 2 Cl- salt itself. Although one of the possible candidates for the reductant is considered to be Cl-, which exists as a counter-anion, there is no concrete evidence for the participation of Cl- at present. Cyclic voltammetry

and absorption spectra of Cy V and Cy V-Ru films

Cyclic voltammograms of vacuum-evaporated films of CyV and CyV-Ru in 1 M RbCl are shown in Figs. 5a and 5b, respectively. In Fig. 5a, the as-deposited film of CyV, which was light green, changed to red-violet with the reduction. Only in the first scan in the negative direction, was an extraordinary large peak observed at - 630 mV. With repeated scans, the height of the reduction peak around - 700 mV increased, though it decreased in the second and third 3rd scans, while the height of the reduction peak at around - 300 mV decreased with repeated redox cycling. Cyclic voltammograms of an as-deposited film of CyV-Ru are shown in Fig. 5b. The colour of the CyV-Ru film was red-violet. and in the first scan the reduction

r 0

0 N

7 E ”

k

”

-100

u

Q I b h .S : XI z E 5

-1 oc 3 2

1 . ._

-200

-300

h .= z % E

-200

-300

-400

2 5 ”

-400

v

-800

-600

Potential

-400

-200

E /mV

0 (vs. Ag/AgCI)

-800

-600

Potential

-400

-200

E /mV

(vs. Ag/AgCI)

Fig. 5. (a) Cyclic voltammograms of CyV in 1 M RbCl aqueous solution. Scan rate = 10 mV/s voltammograms of CyV-Ru in 1 M RbCl aqueous solution. Scan rate = 10 mV/s.

0

(b) Cyclic

202 TABLE

1

Relation

between

the peak potential

and electrolyte

Electrolyte

E, /mv

&/mV

RbCl NaF LiCl

- 370 -415 - 375

-700 - 680 -615

(Ag/AgCU

peaks were at -370 and -690 mV. With the oxidation, the colour changed from red-violet to green. With repeated scans, the current of the reduction reaction around - 700 mV and at - 350 mV decreased gradually. Finally, the cyclic voltammogram of CyV-Ru followed the same curve. Although the reason for the extraordinary behaviour in the first scan cannot be explained clearly at present, the process is considered to occur as follows. When the reduction of the CyV film occurs, counter-ions should be injected into the film in order to compensate the charge change. Since the electronic conductivity of the film was considered to be low, the reduction reaction could expand from the electrode side through the electrolyte solution side. Therefore, counter-ions had to penetrate the film in the first reduction reaction. If this is so, the injection of a counter-ion may be affected by the radius of the ion. If the shift of the reduction potential depends on the size of the cation, the reduction could be proved to occur with the cation injection for charge compensation. If the cation was larger, the peak potential of the reduction would shift in the negative direction. The reduction potentials in the first scan with various aqueous electrolyte solutions are shown in Table 1. The radii of the bare cations and those of the hydrated cations are listed in Table 2 [6]. The peak potentials E2 in Table 1 and the radii of the cations in Table 2 (either bare or hydrated) are plotted in Fig. 6. It was considered that the peak potential should shift towards negative potentials with increasing radius of the cation. The clear tendency of the reduction peak depending on the radius of the bare cations (without hydration) was confirmed. That is, the reduction potential shifted towards negative potentials with an increase of the bare radius. Therefore, it was considered that the counter-cation injection could be confirmed , and that the cations lose their hydration sheaths and are then injected into the film. However, no relationship was discovered between the reduction potential and the radius of the ion on the first reduction peak of the CyV-Ru film.

TABLE

2

Radii of the ions with and without

hydration r/nm

With hydration Without hydration

Li+

Na+

Rb+

0.340 0.060

0.276 0.095

0.228 0.148

203

- 600 Reduction

-650 potential

Fig. 6. Relation circles represent hydration.

peak

-700 E / mV(vs.

Ag/AgCI)

between the second reduction the radii of the ions without

potential of CyV-Ru and the radius hydration and the crosses the radii

of the ions. The of the ions with

Cyclic voltammetric confirmation of the redox state of the CyV film in the as-deposited state is shown in Fig. 7. In the positive scan, the height of the anodic current decreased greatly with repeated scans and finally no redox response was observed, because the CyV film dissolved in the aqueous solution. Since CyV2+ dissolves well in water, the as-deposited film was considered to be in the reduction state by one electron. The reason why the re-deposition of CyV was not observed is that the concentration of CyV dissolved in the electrolyte solution was too low to detect a faradaic current. The concentration of CyV was estimated at 1 X 1OP8

(3

-100’ -200 Potent ial

Fig. 7. Cyclic voltammograms mV/s.

0

200

400

E / mV Ivs. Ag / AgCl

600

)

of CyV in the positive

scan in 1 M RbCl aqueous

solution.

Scan rate = 30

I

m

Time Fig. 8. Colouring-erasing response The potential changes corresponded

/s

of the CyV film. The absorption changes were monitored to the cyclic voltammogram of Fig. 5a.

at 500 nm.

mol/l (film thickness = 50 nm, film density = 1 g/cm3, volume of the electrolyte = 10 ml and electrode area = 1 cm2). In both the CyV and CyV-Ru films, the colour changed reversibly, corresponding to the cyclic voltammetry. The colouring-erasing properties of the CyV film corresponding to the cyclic voltammetry in Fig. 5a are shown in Fig. 8. It can be seen that the absorption at 500 nm changed reversibly corresponding to the change of the electrode potential. CONCLUSION

It was first found that CyV and CyV-Ru were reduced during the vacuum evaporation. The reason why the vacuum evaporation technique is suitable for these materials is the formation of water-insoluble electrochromic films, because the reduced state of CyV is insoluble in water. Moreover, since the electrochemical reversibility of viologens is known to be very high, it is a great advantage that the viologen has been found to form a vacuumevaporated film in which the the electrochromic properties are retained. On the reduction of the CyV-Ru film, counter-cations were found to be injected. The peak potential of the reduction was found to shift in the negative direction with increasing radius of the cation. This may indicate that 4he counter-ion is injected into the CyV-Ru film without hydration. REFERENCES 1 2 3 4 5 6

A. Yasuda, H. Mori, Y. Takehana, A. Ohkoshi and N. Kamiya, J. Appl. Electrochem., 14 (1984) 3‘23. A. Yasuda, H. Kondo, M. Itabashi and J. Seto, J. Electroanal. Chem., 210 (1986) 197. A. Yasuda, H. Mori and J. Seto, J. Appl. Electrochem., 17 (1987) 512. A. Yasuda and J. Seto, J. Appl. Electrochem., 18 (1988) 333. H. Mori and J. Mizuguchi, Jpn. J. Appl. Phys., 26 (1987) 1356. F.A. Cotton and G. Wilkinson, Advanced Inorganic Chemistry, 2nd ed., Wiley, New York, 1966.