CHAPTER 4 HIGH TEMPERATURE BEHAVIOUR OF ELECTROCHROMICS

							                                  CHAPTER 4




 HIGH TEMPERATURE BEHAVIOUR OF ELECTROCHROMICS

                      J.P. Matthews, J.M. Bell and I.L.Skryabin




Published: Renewables: The Energy for the 21st Century
Proceedings of the World Renewable Energy Congress VI, Brighton, UK (A.A.M.
Sayigh Ed.), 230-235 (2000).
Contributions of Authors



This paper presents the results of experimental work carried out by J.P. Matthews, under

the supervision of J.M. Bell and I.L. Skryabin. The paper was written by J.P. Matthews

and revised by J.M. Bell before final submission of the manuscript. This paper was

presented by John Bell as an invited paper at the World Renewable Energy Congress VI,

Brighton, UK, 2000.




                                          84
        HIGH TEMPERATURE BEHAVIOUR OF ELECTROCHROMICS

                        J.P. Matthews1, J.M. Bell1 * and I.L.Skryabin2
                   1
                       Research Concentration in Materials Technology,

             School of Mechanical, Manufacturing and Medical Engineering,

                       Queensland University of Technology, Australia
                           2
                               Sustainable Technologies, Australia,

                        11 Aurora Ave, Queanbeyan, NSW, Australia



*
    Author to whom correspondence should be addressed.




Abstract

Sol-gel deposited electrochromic films have been cycled at elevated temperatures under

various environmental conditions. Significant irreversibility was observed during

cycling of films when moisture was present in the electrolyte, especially at high

temperatures. A proportion of the injected charge did not cause colouration under these

conditions, which caused an apparent decrease in colouration efficiency at high

temperatures. An experiment was carried out which enabled the observation of the slow

bleaching of these films in an electrolyte solution, even though the working electrode

was electrically isolated from the external circuit. This self-bleaching was associated

with irreversible charge injection under conditions where moisture was present. Films

cycled under very dry conditions exhibited very reversible behaviour, and the

colouration efficiency was found to be independent of temperature.



                                               85
Keywords: Electrochromic thin films; switchable glazing; colouration efficiency; self-

bleaching; temperature effects




4.1 Introduction

Previous experiments investigating the effects of temperature on electrochromic device

switching voltages [1] have shown that the magnitude of the voltages that are required to

colour and bleach electrochromic (EC) films decreases with increasing temperature.

When the films were cycled at temperatures exceeding approximately 30ºC some

irreversibility was observed in the charge injection/extraction process. Some of the

charge injected during colouration was unable to be extracted during the bleaching

process however the optical density of the bleached film was consistent with cycles that

were totally reversible. The amount of charge unable to be extracted each cycle

increased with temperature and this trapped charge apparently did not contribute to the

colouration of the film. Although the amount of charge trapped per cycle was relatively

small, the cumulative effect over many cycles is very significant. The amount of charge

available for transfer between the working and counter electrodes of an EC device is

limited by the amount of charge incorporated during device fabrication. After device

fabrication no more charge can be introduced so a reduction in the reversibility of the

EC process limits the maximum possible change in optical density and ultimately the

device lifetime.



In order to better understand the effect of this charge trapping phenomenon, a series of

self-bleaching experiments were carried out. These experiments involved colouring of

                                            86
EC films to a specific charge density, and disconnecting the counter electrode thereby

electrically isolating the working electrode from the external circuit. These coloured

films were observed to undergo a slow self-bleaching process and this change was

monitored by continual measurement of the electromotive force (emf) and optical

density of the working electrode.



After the self-bleaching experiment the WO3 substrate was dissolved off the glass/FTO

substrate with an alkaline solution. Chemical analysis of this solution revealed that

there was a large amount of lithium still present in the film, confirming that there

actually was lithium still inside the film, and that some of this lithium did not contribute

to colouration.



The reversibility problems encountered at high temperatures hindered simulation of the

experimental data because the amount of charge extracted varied for each cycle. In

order to try and establish reversible cycling at elevated temperatures, experiments were

repeated in a nitrogen filled dry-box, in which very low levels of humidity were

stringently maintained. It was found that the EC reaction was reversible over the entire

experimental temperature range and that the colouration efficiency was linear and

independent of temperature.




                                             87
4.2 Experimental


4.2.1 Electrode preparation

Mixed tungsten-titanium oxide electrochromic films were deposited using sol-gel

processing from organic precursors onto 10cm × 10cm substrates of LOF TEC8/3 glass

using the sol-gel dip coating method [2]. The alkoxide precursor solutions used in the

sol-gel dipping have been described previously [3].




4.2.2 Electrochemical testing

Electrochemical measurements were made using a three electrode cell. The counter

electrode was a platinum foil (Area=64cm2) and the reference electrode was a Ag/AgCl

cell filled with an ethanolic solution of KCl, saturated with AgCl. The electrolyte used

was 1M LiClO4 in propylene carbonate, which was stored over molecular sieves after

preparation. Experiments were carried out in a glass tank (filled with electrolyte

solution) partly submersed in a larger heating tank filled with mineral oil. An electrical

heater/stirrer unit was used to control the temperature of the oil bath, and hence the

electrolyte solution. The electrolyte solution was also stirred during the experiments to

minimise the temperature difference between the heating oil and the electrolyte solution.



The WO3 films were cycled using a voltage-limited constant current technique,

described previously [4,5]. Films were cycled to 15mC/cm2 (Current density =

0.1mA/cm2, film area = 100cm2). Optical measurements were made by directing a

1mW, 670nm laser beam through the electrodes, and onto a silicon photodiode. The

                                            88
photocell voltages reported in the results are the output voltages of the silicon

photodiode. As the film is coloured, the intensity of the laser beam reaching the

photodiode is reduced, hence photocell voltages decrease with increasing optical density

of the film.



Experiments were carried out either in the ambient environment of the laboratory, or

inside a dry-box. The experiments carried out in the ambient laboratory environment

were carried out while slowly bubbling dried nitrogen through the electrolyte solution

before and during testing to maintain a slight positive pressure and minimise the mount

of moisture in the electrolyte. During the dry-box experiments the complete

experimental apparatus including electrolyte tank, oil bath, heater unit and optical bench

remained in a nitrogen-filled dry glovebox. The atmosphere inside the glovebox was

kept dry by exposing it to P2O5 desiccant, and recirculating the nitrogen through a

column of dried molecular sieves. The humidity level inside the glovebox was

monitored with a HMP235 humidity and temperature transmitter, manufactured by

Vaisala. The humidity during the dry-box experiment was maintained at 1.05ppm

absolute humidity (Relative humidity = 5.3% and temperature = 22.4ºC immediately

prior to experiment).



The self-bleaching experiment was performed under the ambient laboratory conditions

described above. Films were coloured to 20mC/cm2 (Current density = 0.1mA /cm2,

film area = 100cm2), and the counter electrode was disconnected from the electrical

circuit immediately after completion of the coloration. The counter electrode was



                                             89
disconnected for 30 minutes and the electrolyte solution was maintained at the

appropriate temperature during this time. At the end of the 30 minute self-bleaching

period, the counter electrode was reconnected, the film was bleached and the

temperature was increased for the next data set.


4.2.3 Chemical analysis

The amount of lithium and tungsten in a bleached film was determined, after the self-

bleaching experiment, by inductively coupled plasma-atomic emission spectroscopy

(ICP-AES). The film was washed off the glass/FTO substrate with aqueous sodium

hydroxide, and the solution was diluted to 50mL. Tungsten and lithium stock standard

solutions were used prepared a series dilution of calibration standards. Calibration

graphs were prepared and used to determine the concentrations of lithium and tungsten

in the sample solution. The ICP-AES measurements were performed on a Spectroflame

spectrometer, manufactured by Spectro Analytical Instruments, West Germany.




                                           90
4.3 Results and Discussion


4.3.1 Effect of temperature on coloration efficiency

Figure 4.1 shows the changing optical properties of two sol-gel deposited WO3 films

during colouration and bleaching at various temperatures. The experiment shown in

Figure 4.1(a) was carried out in ambient laboratory conditions, and nitrogen was

bubbled through the electrolyte solution in an attempt to keep water (from the air) out of

the system. The results shown in Figure 4.1(b) are for an experiment carried out in an

extremely dry environment, inside a nitrogen filled glove box. Plots of change in optical

density versus injected charge are expected to be linear for EC films and devices, and

the slope is defined as the colouration efficiency (CE) [6]. It is evident that this

behaviour is observed only in the very dry case (Figure 4.1(b)) where the plots at

different temperatures are virtually the same with an average CE of 38.6cm2/C

(R2=0.99).




Figure 4.1 Change in optical density versus injected charge for WO3 films cycled to

15mC/cm2 at elevated temperatures. The results shown in (a) are for an experiment

carried out in the ambient environment, while the results shown for (b) are for an


                                             91
experiment carried out in a dry-box.

The colouration efficiency for the first experiment (Figure 4.1(a)) is approximately

linear for the room temperature cycle (20.6ºC), and the colouration efficiency is

determined from a linear regression to be 43.9cm2/C (R2=0.93). As temperature

increases the optical density, for each level of injected charge density, decreases

indicating that some of the injected charge is not contributing to colouration. The

reversibility of the charge injection process in this experiment was also affected by

temperature, with significant irreversibility noticeable for temperatures above 30ºC.

The film was cycled using a constant current charge injection/extraction technique (as

described in the experimental section, above), and the bleaching process was terminated

when the voltage reached some safe limit, predetermined to prevent damage to the film.

Any charge remaining in the film after bleaching was therefore unable to be extracted

without applying larger voltages which would have damaged the film. Figure 4.2 shows

the relative amount of charge unable to be extracted from the films each cycle, during

cycling at elevated temperatures, for both of the experiments discussed above.




                                            92
Figure 4.2 Reversibility of cycling at elevated temperatures, represented as the

percentage of the injected charge density trapped per cycle.




It is evident that the ion injection process for the film cycled in ambient conditions was

much less reversible than for the film cycled in the dry-box. The relative amount of

charge not extracted each cycle in the ambient case increases with temperature, and at

50ºC approximately 3.5% of the charge injected was unable to be extracted during the

bleaching process. The reversible limit of x = 0.4 (in LixWO3) [7] was not exceeded

during these cycles so the irreversibility must be accounted to some other reaction

involving the lithium ions.



At room temperature the EC reaction of the film cycled in the dry-box was very

reversible, and the percentage of charge trapped is very close to zero, within the limits

of experimental error. As temperature increases, the amount of charge extracted

actually exceeds the level of charge injected for that cycle, which would suggest some

                                            93
experimental errors. This may be due to the combination of the switching regime used,

and the reduction in switching voltages which occurs at elevated temperatures. Before

the experiment was carried out, the film was ‘pre-loaded’ with charge. The films do not

cycle reversibly for the first few cycles, so 20 cycles were performed where it is

common for a significant proportion of the injected charge to remain in the films after

bleaching, even though the films appear to be bleached as normal.



As temperature rises, the magnitude of the voltage required to achieve a given charge

density decreases and so applying a set voltage limit for the bleaching cycle, we are

driving the bleaching process further at high temperature. In this experiment the same

voltage limit was applied to the bleaching process at all temperatures, so it is possible

that some of the pre-loaded lithium was removed at higher temperatures. The fact that

the amount of charge extracted increases with temperature for the dry-box experiment

supports this proposition. It is also possible that there is a small experimental error

associated with the measurement of the currents, such as a bias towards the

measurement of the bleaching current. An electrical calibration error of this kind would

be expected to be independent of temperature, and this would also mean that the amount

of charge trapped in the ambient conditions was even greater than that shown in Figure

4.2. The fact that there are some negative results for the charge trapped during the

drybox experiments therefore does not affect the conclusions made about the experiment

carried out in the ambient environment.



The large differences between the results observed from films cycled in ambient and



                                             94
very dry conditions suggests that the problems associated with irreversibility and charge

not causing colouration may be ascribed to water present in the system. Although an

attempt was made to keep the ‘ambient’ experiment dry by bubbling dry nitrogen

through the electrolyte, the high humidity of the experimental location (Brisbane,

Ausralia) combined with the highly hygroscopic nature of the propylene carbonate

electrolyte means that it is unlikely that there was no water present in the electrolyte.

Inside the drybox, it is relatively easy to ensure that there is very little water present and

so the presence of water is thought to be the major difference in the conditions of the

two experiments described above. In order to further investigate the cycling

irreversibility and the proportion of injected lithium not causing colouration at high

temperature, a self bleaching experiment was carried out using the ambient environment

conditions described in the experimental.




4.3.2 Observation of self-bleaching



Figure 4.3(a) shows the change in photocell voltage of the WO3 film during self-

bleaching at elevated temperatures and Figure 4.3(b) shows emf measurements taken

over the same time period. The WO3 electrode was electrically isolated at 150seconds

(after the end of colouration) and then reconnected after a 30 minute period, just prior to

the bleaching half-cycle. The increase in photocell voltage observed in Figure 4.3(a)

indicates a reduction in the optical density and suggests that the concentration of lithium

in the film is decreasing or that some of the lithium is being converted to an optically

non-active form. The drift in emf observed in Figure 4.3 indicates a changing chemical

                                              95
potential of the film, and the drift towards more positive potentials is consistent with a

decreasing lithium concentration over time and at higher temperature. These results

suggest that the film is bleaching as per the normal EC reaction but with the counter

electrode disconnected there is no path for electron flow from the back of the working

electrode. Any lithium reaction must therefore be with some species already present in

the system which also supports the theory that water in the system is responsible for

some of the injected ions not causing colouration and for the irreversibility observed.




Figure 4.3 Change in (a) photocell voltage and (b) emf of WO3 electrode during self-

bleaching experiment.




The currents measured during colouration and bleaching were integrated with respect to

time to determine the amount of charge injected and extracted respectively. These

values were used to calculate the measured amount of charge which was trapped per

cycle. Calibration curves of photocell voltage and emf versus injected charge density

were constructed for each temperature, by interpolation of measurements made at the

highest and lowest temperatures. These calibration curves were used in conjunction

with photocell and emf values at the start and end of the self-bleaching period (ie. from

                                             96
(a) and (b)), in order to estimate the amount of charge apparently lost during the 30

minute self-bleaching period.



These estimated values of charge loss were then correlated with the measured values

obtained from the difference between the integration of the colouration and bleaching

currents.

Figure 4.4 shows a plot of the estimates of charge lost versus the measured charge loss,

during the self-bleaching period. If the lithium ion concentration in the film was

decreasing (eg. lithium was reacting at the electrode surface to form a new species

outside the film) we would expect the plots of estimated versus measured charge loss to

be linear with a slope of one and intercept of zero.




Figure 4.4 Correlation of estimated and measured quantities of charge lost during self-

                                        bleaching.




                                            97
These plots are indeed linear however the slope is not one and the intercept is not zero.

The photocell measurements are an indication of the number of lithium ions contributing

to colouration, and the charge loss estimated from photocell measurements is therefore

an indicator of the amount of lithium no longer causing colouration at the end of the

self-bleaching period. The measured amount of charge remaining in the film after

bleaching at each temperature is smaller than the estimates, which suggests that some of

the charge that was extracted was not contributing to colouration of the film. For

example, at the end of the self-bleaching period at 50.5ºC, the amount of injected charge

no longer causing colouration is estimated from the photocell voltage (immediately prior

to bleaching) to be 4.7mC/cm2. The amount of charge remaining in the film after the

subsequent bleaching cycle was 4.3mC/cm2. Approximately 4.7mC/cm2 of lithium ions

therefore were not causing colouration after the half hour self bleaching period, and

0.4mC/cm2 of this was later electrochemically extracted from the film. The remaining

4.3mC/cm2 of lithium ions either stayed in the film, was lost into the electrolyte solution

to a side reaction or a combination of both.




4.3.3 Determination of trapped lithium in WO3 film by ICP-AES



In order to answer some of the questions regarding the location of lithium ions which

could not be extracted from the WO3 film by the bleaching process, a portion of a film

used in another self-bleaching experiment was subjected to further chemical analysis.

The WO3 film was washed off the glass/FTO substrate with a sodium hydroxide

solution, and then diluted to 50mL. The film area used was 55.6cm2, and inductively

                                               98
coupled plasma-atomic emission spectroscopy (ICP-AES) was used to determine the

lithium and tungsten concentrations of this solution. The ICP-AES analysis revealed

that there was approximately 100µg of lithium and 8mg of tungsten in the 50mL

solution, which corresponds to x=0.33 in LixWO3 or an injected charge density of

approximately 15mC/cm2 (for a 200nm thick film, with molar volume of 42cm2/mol).

The total measured charge lost during this experiment was approximately 130mC/cm2, a

value clearly very much larger than the amount of ions recovered from the film.

The fact that a large proportion of the measured injected charge was not recovered

suggests that a large amount of the injected ions were either lost to side reactions or that

the ion injection process was not 100% efficient.



Considering that the charge measurement was made by integration of the electron

current with respect to time, any side reactions occurring simultaneously along with the

normal ion intercalation would contribute to the measured current and hence the

measured charge. If the measured current resulted solely from ion injection, the

remaining charge which was not recovered was presumably lost to side-reaction(s) to

form a new species, which then dissolved into the electrolyte. Another possibility is that

side reactions such as gas evolution occurred during ion injection and made a significant

contribution to the measured current, however no evidence of gas evolution was

observed during the experiment.




                                             99
4.4 Conclusions

Sol-gel deposited EC films were cycled under various conditions and a range of

temperatures. Significant irreversibility was observed for films cycled with moisture

present, especially at high temperature. This irreversibility was associated with a

proportion of the injected charge not causing colouration, and consequently there was an

apparent reduction in colouration efficiency at high temperatures. In very dry

conditions, films cycled very reversibly and colouration efficiency was independent of

temperature.



Coloured films were observed to slowly self-bleach in an electrolyte which was not

completely dry, even though the counter electrode was disconnected. After leaving the

film in this electrolyte for 30 minutes, some of the charge remained in the film even

after the bleaching process, and this charge did not give rise to colouration.

Measurements of the photocell voltage and emf of the film during the 30 minute period

were used to estimate the amount of charge trapped during the self-bleaching period.

These estimates were compared to the measured values of charge remaining in the film

after bleaching, and a reasonable correlation was attained.

ICP-AES was used to confirm that there was actually lithium trapped in the film after

self-bleaching, but only a small proportion of the expected amount of lithium was found.

This implied that some charge was trapped in the film while a much larger proportion

was lost to a side reaction, probably reaction of lithium ions with water present in the

electrolyte.




                                            100
Acknowledgements



This work is supported by an Australian Postgraduate Award (Industry) scholarship

from the Australian Research Council, and Sustainable Technologies Australia (STA).

The work described in this paper has been supported by the Australian Cooperative

Research Centre for Renewable Energy (ACRE). ACRE’s activities are funded by the

Commonwealth’s Cooperative Research Centres Program. We would also like to thank

Pat Stevens for his advice and assistance with the ICP-AES measurements.


REFERENCES

[1] J.P. Matthews, J.M. Bell and I.L. Skryabin, Electrochimica Acta 44 (1999) 3245.

[2] J. M. Bell, G. B. Smith, I. L. Skryabin, B. G. Monsma, N. C. Ruck and T. Dinh,

“Sol-gel Deposited Electrochromic Devices”, in Proceedings of Windows Innovations

Conference WIC ‘95, 383-391, Minister of Supply and Services, Canada.

[3] A. Koplik, Australian Patent Application, PP0274 (1997).

[4] J. M. Bell and I. L. Skryabin, Solar Energy Materials and Solar Cells, 56 (1999)

437.

[5] I. L. Skryabin and J. M. Bell, Control of Electrochromic Devices, International

Patent Application PCT/AU97/00697.

[6] C.M. Lampert, V-V. Truong, J. Nagai and M.G. Hutchins, in Characterization

Parameters and Test Methods for Electrochromic Devices in Glazing Applications,

International Energy Agency Task X-C Final Report, University of California (1994).

[7] J.-G. Zhang, D.K. Benson, C. Edwin Tracy and S.K. Deb, J. Mater. Res., 8, 2657-

2667 (1993).

                                          101
102