www.elsevier.com/locate/ssi

Solid State Ionics 171 (2004) 173–181

Planar electrochemical sensors based on tape-cast YSZ

layers and oxide electrodes

Elisabetta Di Bartolomeo*, Narin Kaabbuathong, Maria Luisa Grilli, Enrico Traversa

Department of Chemical Science and Technology, University of Rome Tor Vergata, Via della Ricerca Scientifica, 00133 Rome, Italy

Received 20 January 2004; received in revised form 11 March 2004; accepted 14 May 2004

Abstract

Electrochemical planar sensors based on yttria-stabilized zirconia (YSZ) with semiconducting oxides (WO3 and LaFeO3) and mixed

conductors (La0.8Sr0.2FeO3) as sensing electrodes were investigated. The electromotive force (EMF) of the sensors was measured at fixed

temperature (450–700 jC) and different concentrations of NO2 and CO in air in the range 20–1000 ppm. The sensors were wholly exposed

to the same gas atmosphere, alternatively air and different gases in air, without using reference air. Fast and stable response was measured for

all the different sensors. The EMF of p- and n-type semiconducting-based sensors was measured in opposite directions upon the same gas

exposure. These results obtained for the response of planar sensors with different semiconductor sensing electrodes showed that the sensing

mechanism cannot be explained by mixed potential theory. The response of these sensors is better described by a differential electrode

equilibria sensing mechanism.

D 2004 Elsevier B.V. All rights reserved.

Keywords: Electrochemical gas sensors; Yttria-stabilized zirconia (YSZ); Semiconducting oxides; NO2 and CO gases; Automobile applications

1. Introduction

New emission regulations increase the need of reliable

and inexpensive sensors for monitoring and control of

automotive gas pollutants. The major gases to be detected

are CO, NO, NO2, and hydrocarbons (HCs), but also SO2,

CO2, O3, etc. To control and reduce these emissions, the On-

Board Diagnostic (OBD) system has been introduced in

most countries worldwide [1]. At present, given the lack of

reliable CO/hydrocarbons and NOx sensors, OBD is per-

formed by using oxygen sensors. One is placed upstream the

three-way catalytic converter (to control the air/fuel ratio),

another is located downstream in the exhaust to control,

through an electronic unit, the efficiency of the catalytic

converter. Solid oxide electrochemical O2 sensors are inex-

pensive and have been successfully shown to work in harsh

combustion exhaust environment. If modified to selectively

measure NOx or CO/hydrocarbon concentrations, this type

of sensors can be used to improve the combustion as feed-

back elements in engine control systems and thus comply

with the always much more strict automotive emission

regulations. This approach has been proposed by several

0167-2738/$ - see front matter D 2004 Elsevier B.V. All rights reserved.

doi:10.1016/j.ssi.2004.05.003

* Corresponding author. Tel.: +39-6-72594495; fax: +39-6-72594328.

E-mail address: dibartolomeo@uniroma2.it (E. Di Bartolomeo).

authors in relevant literature [2–15]. We have previously

studied bulk sensors based on yttria-stabilized zirconia

(YSZ) and semiconducting oxide electrodes [16–21].

In this work, planar sensors obtained by combining YSZ

(8 wt.% of Y2O3) tape-cast layers with WO3, LaFeO3, and

La0.8Sr0.2FeO3 semiconducting oxides as sensing electrode

were prepared. WO3 and LaFeO3 have been selected

because of their good performance in bulk and thick-film

form [2,3] as semiconductor NO2 sensors. La0.8Sr0.2FeO3

oxide was used to check if the mixed ionic and electronic

conductivity could improve the sensing response. The

electrochemical response of the sensors was studied in the

temperature range between 450 and 700 jC in the presence

of NO2 and CO in air (20–1000 ppm). The influence of the

grain size of La0.8Sr0.2FeO3 electrode on the gas response

was also investigated.

The main difference with previous work is that, in this

case, the metal and the oxide electrodes are on the same side

of the electrolyte but are closer to each other.

2. Experimental

YSZ (8 wt.% of Y2O3) tape-cast layers of 150-Amthickness were used for the sensor fabrication as solid

E. Di Bartolomeo et al. / Solid State Ionics 171 (2004) 173–181174

electrolyte material. Pt paste was used for the preparation

of metallic electrodes, which were deposited on one side

of the layers in parallel finger-like form at a distance of

about 5 mm. A scheme of the planar sensor is reported in

Fig. 1. Thin gold wires were connected for current

collection. The firing temperature of Pt paste was 750

jC for 10 min. For the fabrication of the sensing electrode,

commercial WO3 powders (99.995% purity) and LaFeO3

and La0.8Sr0.2FeO3 perovskite oxides, prepared in the

laboratory, were mixed with a screen-printing oil and the

slurry obtained was painted on one of the metallic elec-

trode and fired at 750 jC for 3 h.

Nano-sized LaFeO3 powders were prepared by the

thermal decomposition of a LaFe–hexacyanide complex

at 700 jC for 1 h [22]. La0.8Sr0.2FeO3 was prepared by a

sol–gel route. The precursors used were: La(NO3)2,

Sr(NO3)2, Fe(NO3)3, citric acid, ethylene glycol. The

nitrates and the citric acid were mixed in the following

molar ratio: La/Sr/Fe/citric acid = 0.8:0.2:1:2, while the

citric acid/ethylene glycol weight ratio was 40:60. Stoi-

chiometric amounts of salts were first dissolved into

ethylene glycol. When the precursors were completely

dissolved, controlled amount of citric acid was added. A

complete dissolution of precursors resulted in a clear red-

brown solution. The gel formation occurred at 120 jC, itstransparency giving an indication of a homogeneous

system. The samples were dried at 120–130 jC for few

hours and then heated at 700 jC for 3 h. To check the

influence of the electrode grain size on the sensing

response of the devices, the La0.8Sr0.2FeO3 oxide was also

heated at 900 jC for 4 h.

Fig. 1. Scheme of the planar sensor:

X-ray diffraction (XRD) analysis was preformed by

a Philips X’Pert 500 Powder Diffractometer on the

powders for phase identification, and microstructures

of the powders and films were observed by scanning

electron microscopy (SEM) using a Leica Cambridge

Model Stereoscan 360. From XRD patterns only the

peaks of orthorhombic perovskite-type LaFeO3 and

La0.8Sr0.2FeO3 structures were found, showing a good

crystallinity and purity of the powders as previously

published [22]. From SEM analysis, WO3 thick films

showed sub-micrometric grain size [20], LaFeO3 and

La0.8Sr0.2FeO3 films showed homogeneous nanometric

particles [17,18].

Sensing experiments were carried out in a conventional

gas-flow apparatus equipped with a controlled heating

facility. The sensor was alternatively exposed to air

and NO2 or CO (20–1000 ppm in air) at the total flow

rate of 100 ml/min in the temperature range between 550

and 700 jC. Electromotive force (EMF) measurements

were performed between the two electrodes of the sensors

using a digital electrometer.

3. Results and discussion

Figs. 2 and 3 show that the EMF response of WO3-

based sensors at fixed temperature (550–700 jC) was

in opposite direction upon exposure to NO2 and CO.

Positive EMF values were measured at different NO2

concentrations and negative EMF values at different CO

concentrations. The EMF changed quickly upon switching

(a) top view and (b) side view.

Fig. 2. EMF response of WO3-based sensors to different concentrations of NO2 in air at different operating temperatures.

E. Di Bartolomeo et al. / Solid State Ionics 171 (2004) 173–181 175

from air to different NO2 or CO concentrations in the

range 400–1000 ppm and steady-state values were

observed at all temperatures. At 600 jC, the EMF

magnitude was 25 and � 17 mV upon exposure to

1000 ppm of NO2 and CO, respectively. The response

(20 s) and recovery (40 s) times were quite fast for both

gas mixtures. It should be emphasized that a good

response was observed even at a temperature as high as

700 jC.A linear correlation was observed between the EMF

values and the NO2 and CO concentrations in logarithmic

scale. The best sensitivity to both NO2 and CO gases

was observed at 600 jC, as shown in Fig. 4. Unfortu-

nately, WO3-based sensors showed a lack of selectivity.

As shown in Fig. 4, the sensitivity to NO2 and CO at

600 jC were found to be quite close to each other

(about 17.8 and � 19.4 mV/decade, respectively). To

Fig. 3. EMF response of WO3-based sensors to different concen

improve the selectivity of the sensor to NO2, other

heterometallic oxide powders, LaFeO3 and La0.8Sr0.2FeO3, were investigated.

The LaFeO3-based sensors worked at lower temper-

atures as compared to WO3-based sensors. Fig. 5 shows

the EMF response of LaFeO3-based sensors to different

concentrations of NO2 in the range 400–1000 ppm at

different operating temperatures. The EMF responses of

the sensor to 1000 ppm of NO2 at 550, 600, and 650

jC were about � 30, � 15, and � 5 mV, respectively.

At 550 jC, the largest EMF values for all different NO2

concentrations together with the fastest response and

recovery times (40 s and 1 min, respectively) were

found. The EMF response of LaFeO3-based sensors to

NO2 gas was always negative at all measured temper-

atures, always in opposite direction to that of WO3-based

sensors.

trations of CO in air at different operating temperatures.

Fig. 4. EMF response of WO3-based sensors versus the concentrations of NO2 and CO in logarithmic scale at different operating temperatures.

E. Di Bartolomeo et al. / Solid State Ionics 171 (2004) 173–181176

Fig. 6 shows the sensing response of LaFeO3-based

sensors to NO2 in a low concentration range, from 40 to

100 ppm, at 550 and 600 jC. The sensing responses of the

sensors to 100 ppm of NO2 at 550 and 600 jC were about

� 10 and � 2 mV, respectively. EMF values decreased

with decreasing NO2 concentrations. However, no differ-

ence in the response and recovery times to 100 and 1000

ppm of NO2 was observed.

The CO cross-sensitivity of the LaFeO3-based sensors

was also studied. Fig. 7 shows the CO sensing response of

LaFeO3-based sensors at different CO concentrations. The

concentration of CO ranged from 400 to 1000 ppm. The EMF

response of the LaFeO3-based sensors to CO was always in

the positive direction. The EMF responses to 1000 ppm CO

at 550 and 660 jC were about 12 and 3 mV, respectively.

The NO2 and CO sensing characteristics of LaFeO3-

based sensors at three operating temperatures are com-

Fig. 5. EMF response of LaFeO3-based sensors to different concentrations

pared and shown in Fig. 8. The sensitivity to NO2 at 550,

600, and 650 jC were � 11.2, � 11.0, and � 3.7 mV/

decade, respectively. Thus, the best sensitivity to NO2

was observed at 550 jC. In the same figure, the sensi-

tivity to CO at 600 jC is also reported. The EMF values

are almost the same for all different CO concentrations,

so a very low sensitivity is obtained from the linear slope.

Making a comparison between the CO sensing property

of WO3-based and LaFeO3-based sensors at the same

temperature, e.g., at 600 jC, the sensing magnitude of

LaFeO3-based sensors was found to be smaller (almost

about 6 times) than WO3-based sensors. Thus, the CO

cross-sensitivity is obviously reduced by replacing WO3

with LaFeO3 semiconducting oxide powder as sensing

electrode. Though the improvement of NO2 selectivity of

LaFeO3-based sensor was successful, the further possibil-

ity to reduce CO cross-sensitivity by using another oxide

of NO2 in air (400–1000 ppm) at different operating temperatures.

Fig. 6. Response of LaFeO3-based sensors to different concentrations of NO2 in air (40–100 ppm) at different operating temperatures.

E. Di Bartolomeo et al. / Solid State Ionics 171 (2004) 173–181 177

such as La0.8Sr0.2FeO3, a mixed ionic-electronic conduc-

tor, was investigated further.

Fig. 9 shows the EMF response of La0.8Sr0.2FeO3-

based sensors (La0.8Sr0.2FeO3 powders decomposed at 700

jC) to different concentrations of NO2 in air at different

operating temperatures. The sensors were exposed to

various concentrations of NO2 gas (from 200 to 1000

ppm) at 550 and 600 jC. The EMF responses to 1000

ppm at 550 and 600 jC were measured to be � 30 and

� 15 mV, respectively. Similar values were obtained in

the same operating conditions for LaFeO3-based sensors.

Again, the response and recovery times of La0.8Sr0.2FeO3-

based sensors were found to be of the same magnitude as

those of previous WO3-based and LaFeO3-based sensors

(about 1 min).

Fig. 10 shows the sensing characteristic of the

La0.8Sr0.2FeO3-based sensors to different concentrations

Fig. 7. EMF response of LaFeO3-based sensors to different conc

of CO gas at two temperatures, 550 and 600 jC. Like

the LaFeO3-based sensors, the sensing direction to CO gas

of these sensors was always positive. The EMF responses

to 1000 ppm CO at 550 and 600 jC were reported to be

only 5 and 3 mV, respectively. It should be noted that for

LaFeO3-based and WO3-based sensors, the sensing

responses to 1000 ppm CO at 550 jC were found to be

about 12 and � 20 mV, respectively. The reduction of CO

cross-sensitivity at every measured temperature was sig-

nificantly achieved by using La0.8Sr0.2FeO3 powder as

sensing electrode. This is probably due to the mixed ionic

and electronic conduction mechanism: probably oxygen

ions absorbed on the surface can migrate inside the grains

and so a reduced number of adsorbed oxygen ions is

available for the reaction with CO. On the other hand, NO2

adsorption is competitive with O2 and thus NO2 sensitivity

is not reduced.

entrations of CO in air at different operating temperatures.

Fig. 8. EMF of LaFeO3-based sensors versus log NO2 and CO concentrations in air at different temperatures.

E. Di Bartolomeo et al. / Solid State Ionics 171 (2004) 173–181178

Fig. 11 shows the saturation EMF values to various

concentration of NO2 and CO of La0.8Sr0.2FeO3-based

sensors at 550 and 600 jC plotted versus the logarithm

of NO2 or CO concentration. The best sensitivity to NO2

of the sensor was observed at 600 jC, about � 20 mV/

decade, while the CO sensitivity at the same operating

temperature was found to be only 3 mV/decade. From

Figs. 10 and 11, we can confirm that the improvement of

reducing the cross-sensitivity to CO of the sensor was

achieved by using La0.8Sr0.2FeO3 semiconducting oxide

powder as sensing electrode as compared to WO3-

based and LaFeO3-based sensors at the same operating

temperature.

To investigate the sensing mechanism of these electro-

chemical sensors, a different geometry of the planar sensor

electrodes was tested. The reference and sensing electrodes

Fig. 9. EMF response of La0.8Sr0.2FeO3-based sensors to different co

were painted on the opposite sides of the YSZ layers. The

La0.8Sr0.2FeO3 powder that gave the best selectivity to

NO2 was selected for the sensing electrode. The Pt

electrode was exposed to reference air and the La0.8Sr0.2-FeO3 oxide electrode was exposed to different concentra-

tions of NO2 in air.

Fig. 12 shows the EMF response of the planar sensors at

450 jC and different concentrations of NO2 in air. The EMF

values were positive for all NO2 concentrations. These

results are just the opposite of what was observed in the

case of planar sensors with parallel finger-electrodes ex-

posed to the same atmosphere.

Fig. 13 shows the response at 450 jC of the same

sensor placed upside down—that is, exposing the oxide

electrode to reference air and the Pt electrode to 100 ppm

of NO2 in air. Also, in this case, a positive EMF response

ncentrations of NO2 in air at different operating temperatures.

Fig. 10. EMF response of La0.8Sr0.2FeO3-based sensors to different concentrations of CO in air at different operating temperatures.

E. Di Bartolomeo et al. / Solid State Ionics 171 (2004) 173–181 179

was observed, although the EMF values were smaller

and unstable in the same operating conditions. Moreover,

the EMF response in air showed a drift probably due

to the oxygen-ion conduction of the La0.8Sr0.2FeO3

oxide.

The same behavior to NO2 sensing characteristic was

confirmed by using a symmetric Pt/YSZ/Pt sensors. The

sensing response to 100 ppm of NO2 was small (few mV),

always positive but unstable.

3.1. Sensing mechanism of planar sensors

According to several authors [2–16], the sensing mech-

anism of solid-state electrochemical sensors with semicon-

ducting electrodes can be explained by mixed potential

Fig. 11. EMF of La0.8Sr0.2FeO3-based sensors versus log NO2

theory. The electrochemical reactions occurring at the

three-phase boundary between solid electrolyte, electrode,

and gas are: for NO2,

NO2 þ 2eZO2� þ NO ð1Þ

O2�Z1

2O2 þ 2e ð2Þ

for CO,

COþ O2�ZCO2 þ 2e ð3Þ

1

2O2 þ 2eZ2O2� ð4Þ

In the cases reported in the literature, these are the

reactions occurring at the sensing electrode with metal oxide

and CO concentrations in air at different temperatures.

Fig. 12. EMF response of La0.8Sr0.2FeO3-based sensors with reference air at metallic electrode to different concentrations of NO2 at 450 jC.

E. Di Bartolomeo et al. / Solid State Ionics 171 (2004) 173–181180

that give rise to the mixed potential. In the case of our

sensors where both electrodes are exposed to the same gas

environment, both reactions take place at both the electro-

des, though reactions (1) and (3) prevail at the sensing

electrode and reactions (2) and (4) at the metal electrode

side. These reactions can explain the results obtained in the

case of bulk sensors [16–20].

EMF measurements on planar sensors cannot be

explained by the occurrence of the electrochemical reactions

at the three-phase boundary, as obtained in the case of bulk

sensors [21]. In particular, the electrochemical reaction (1),

that should mainly happen at the oxide sensing electrode, as

in the case of bulk sensors, cannot account for the negative

EMF values of p-type semiconducting oxide-based sensors

in the presence of NO2 gas. This fact is probably due to the

electrodes design: two close finger-electrodes exposed to the

same gas atmosphere. Because of this geometry, the elec-

trochemical reactions take place with the same rate at both

Fig. 13. EMF response of La0.8Sr0.2FeO3-based sensors with refere

the electrodes. Thus, the adsorption mechanism character-

istic of semiconductors seems to become predominant. In

fact, the variation of EMF is compatible with the changes in

resistance induced by chemisorption of oxidizing or reduc-

ing gas on n- or p-type semiconductors.

In a different way, the response of the planar sensors with

reference air can be fully explained by using the mixed

potential theory. For NO2 detection, reactions (1) and (2)

take place at the electrode exposed to the pollutant gas,

giving rise to positive EMF values. These results suggest

that the oxide electrode plays an important role in promoting

the electrochemical reaction of NO2. In fact, the response of

the sensors with the oxide electrode exposed to the gas is

much larger and stable. Nonetheless, the geometry of the

electrodes is also very important. The mixed potential

theory, widely accepted by many authors [2–16], can fully

explain the behavior of sensors with reference air. The same

theory cannot explain the response of planar sensors with

nce air at the oxide electrode to 100 ppm of NO2 at 450 jC.

E. Di Bartolomeo et al. / Solid State Ionics 171 (2004) 173–181 181

parallel finger-electrodes. In this case, differential electrode

equilibria theory must be claimed to explain the opposite

responses obtained for p- and n-type semiconducting oxide-

based sensors [23,24], based on different electrocatalytic

activity and, most likely in this case, to resistance changes

by chemical sorption–desorption behavior of the electrodes.

4. Conclusions

Planar electrochemical sensors obtained by combining

YSZ with WO3, LaFeO3, and La0.8Sr0.2FeO3 semiconduct-

ing oxides as sensing electrodes are promising candidates

for pollutant detection at high temperatures. The use of an

oxide as electrode improves the thermal and chemical

stability of the electrochemical cell, which is extremely

important in automotive applications.

The sensing mechanism of oxide-based planar sensors

cannot be explained by mixed potential theory. The EMF

response can be also due to the different electrocatalytic

activity and/or chemical sorption–desorption behavior of

the electrodes (differential electrode equilibria).

Our findings on planar type sensors suggest that, due to

the electrode design (close finger-electrodes exposed to the

same gas atmosphere), the electrochemical reactions take

place with the same rate at both electrodes and thus the

chemical adsorption mechanism of gases on semiconducting

oxide is predominant.

Acknowledgements

This work was partly supported by the Ministry of

Education, University and Scientific and Research (MIUR)

of Italy.

References

[1] K Oishi, in: M. Butler, A. Ricco, N. Yamazoe (Eds.), Proceedings of

the Symposium on Chemical Sensors II, vol. 93-7, The Electrochem-

ical Society, Pennington, NJ, 1993, p. 443.

[2] N. Miura, H. Kurosawa, M. Hasei, G. Lu, N. Yamazoe, Solid State

Ionics 86–88 (1996) 1069.

[3] G. Lu, N. Miura, N. Yamazoe, J. Mater. Chem. 7 (1997) 1445.

[4] G. Lu, N. Miura, N. Yamazoe, Ionics 4 (1998) 16.

[5] T. Hibino, S. Wang, S. Kakimoto, M. Sano, Sens. Actuators, B,

Chem. 50 (1998) 149.

[6] N. Miura, G. Lu, M. Ono, N. Yamazoe, Solid State Ionics 117 (1999)

283.

[7] T. Hibino, S. Tanimoto, S. Kakimoto, M. Sano, Electrochem. Solid-

State Lett. 2 (12) (1999) 651.

[8] G. Lu, N. Miura, N. Yamazoe, Sens. Actuators, B, Chem. 65 (2000)

125.

[9] N. Miura, T. Shirahishi, K. Shimanoe, N. Yamazoe, Electrochem.

Commun. 2 (2000) 77.

[10] N. Miura, N. Yamazoe, in: H. Baltes, W. Gopel, J. Hesse (Eds.), Sen-

sors Update, vol. 6, Wiley-VCH, Weinheim, Germany, 2000, p. 191.

[11] E.L. Brosha, R. Mukundan, D.R. Brown, F.H. Garzon, J.H. Visser, M.

Zanini, Z. Zhou, E.M. Logothetis, Sens. Actuators, B, Chem. 69

(2000) 171.

[12] T. Hibino, A. Hashimoto, S. Kakimoto, M. Sano, J. Electrochem. Soc.

148 (2001) H1.

[13] N. Miura, S. Zhuiykov, T. Ono, M. Hasei, N. Yamazoe, Sens. Actua-

tors, B, Chem. 83 (2002) 222.

[14] E.L. Brosha, R. Mukundan, D.R. Brown, F.H. Garzon, Sens. Actua-

tors, B, Chem. 87 (2002) 47.

[15] R. Mukundan, E.L. Brosha, F.H. Garzon, J. Electrochem. Soc. 150

(12) (2003) H279.

[16] E. Di Bartolomeo, E. Traversa, M. Baroncini, V. Kotzeva, R.V.

Kumar, J. Eur. Ceram. Soc. 20 (2000) 2699.

[17] J.W. Yoon, M.L. Grilli, E. Di Bartolomeo, R. Polini, E. Traversa,

Sens. Actuators, B, Chem. 76 (2001) 483.

[18] M.L. Grilli, E. Di Bartolomeo, E. Traversa, J. Electrochem. Soc. 148

(2001) H98.

[19] M.L. Grilli, N. Kaabbuathong, A. Dutta, E. Di Bartolomeo, J. Ceram.

Soc. Jpn. 110 (2002) 159.

[20] A. Dutta, N. Kaabbuathong, M.L. Grilli, E. Di Bartolomeo, E. Tra-

versa, J. Electrochem. Soc. 150 (2003) H33.

[21] E. Di Bartolomeo, M.L. Grilli, N. Kaabbuathong, E. Traversa, MRS

Symp. Proc. 756 (2003) 127.

[22] E. Traversa, M. Sakamoto, Y. Sadaoka, J. Am. Ceram. Soc. 79 (1996)

1401.

[23] E.D. Wachsman, P. Jayaweera, in: E.D. Wachsman, W. Weppner, E.

Traversa, M. Liu, P. Vanysek, N. Yamazoe (Eds.), Proceedings of the

Electrochemical Society Symposium, Solid State Ionic Devices II—

Ceramic Sensors, vol. 2000-32, The Electrochemical Society, Pen-

nington, NJ, 2001, p. 298.

[24] E. Di Bartolomeo, M.L. Grilli, E. Traversa, J. Electrochem. Soc. 151

(5) (2004) 133.