Rh

SD

a

ARRAA

KCMHNXT

1

dosrglamhrsth[

saaw[T

0h

Sensors and Actuators B 194 (2014) 260– 268

Contents lists available at ScienceDirect

Sensors and Actuators B: Chemical

journa l h om epage: www.elsev ier .com/ locate /snb

apid synthesis of cerium oxide nanoparticles with superiorumidity-sensing performance

harada Thakur, Pradip Patil ∗

epartment of Physics, North Maharashtra University, Jalgaon 425001, India

r t i c l e i n f o

rticle history:eceived 22 October 2013eceived in revised form 6 December 2013ccepted 16 December 2013vailable online 24 December 2013

a b s t r a c t

This paper reports a simple and rapid microwave-assisted method for synthesizing cerium oxide (CeO2)nanoparticles for the fabrication of high-performance humidity sensors. The humidity-sensing investi-gation reveals that the sensor based on CeO2 nanoparticles exhibits a high and linear response withinthe entire relative humidity (RH) range of 11–97% at an operating frequency of 60 Hz. The correspondingimpedance changes by approximately three orders of magnitude within the entire humidity range from11% to 97%. The response and recovery times are approximately 3 and 16 s, respectively. Additionally, the

eywords:erium oxide nanoparticlesicrowave synthesisumidity sensoryquist impedance plots-ray diffraction

sensor exhibits a rapid and reversible response characterized by a very small hysteresis (∼1%RH), excel-lent repeatability, long term stability and a broad range of operation (11–97%RH). The Nyquist impedanceplots of the sensor at different RHs were used to elucidate the sensor’s humidity-sensing mechanism viaan electrical equivalent circuit. The experimental results provide a possible method for the rapid synthesisand fabrication of high-performance humidity sensors based on CeO2 nanoparticles.

ransmission electron microscopy

. Introduction

Over the last decade, considerable efforts have been made toevelop suitable humidity-sensitive materials for the fabricationf high-performance humidity sensors [1–4]. To fabricate a sen-or that is suitable for practical use, a high response value, a linearesponse, a fast response and recovery behavior, a small hysteresis,ood reproducibility, a wide range of relative humidity detection,ong-term stability and low cost are required. The performance of

humidity sensor mainly depends on the properties of the sensingaterials. Metal oxides [5–7], polymers [8,9] and inorganic/organic

ybrids [10–12] have been widely investigated as sensing mate-ials for the fabrication of humidity sensors. Among the variousensing materials, metal oxides have been widely investigated dueo their unique properties such as chemical and physical stability,igh mechanical strength and a wide operating temperature range5–7].

Recently, the nanostructures of metal oxides have received con-iderable interest in the fabrication of sensors to detect humiditynd gases due to the large surface-to-volume ratio, high surfacectivity and effective electron transport of these oxides [13–19]. A

ide range of nanostructured metal oxides such as SnO2 nanowires

13], ZrO2 nanorods [14], ZnO nanorods [15], Al2O3 nanowires [16],iO2 nanotubes [17], BaTiO3 nanofibers [18] and ZnSnO3 nanocubes

∗ Corresponding author. Tel.: +91 2572257474.E-mail address: pnmu@yahoo.co.in (P. Patil).

925-4005/$ – see front matter © 2013 Elsevier B.V. All rights reserved.ttp://dx.doi.org/10.1016/j.snb.2013.12.067

© 2013 Elsevier B.V. All rights reserved.

[19] have been investigated as humidity-sensing materials duringthe past few years.

In the context of humidity sensors, CeO2 is less studied but is apotential ceramic material for sensing applications due to the mate-rial’s good response to humidity [20]. Recently, hydrothermallysynthesized CeO2 nanowires were used for humidity sensing byFu et al. [20]. The CeO2 nanowires indicate an exponential decreaseof the wire’s resistance with an increase in the RH, and both theresponse and the recovery times were found to be ∼3 s. However,the synthesis method of CeO2 nanowires is time-consuming. Thehumidity-sensing properties of the Ba-doped CeO2 nanowires [21]and Mn-doped CeO2 nanorods [22] synthesized by using a facilecomposite-hydroxide-mediated (CHM) route have been recentlystudied. However, the other humidity-sensing characteristics suchas the humidity response, the response and recovery times, thehysteresis and the repeatability have not been investigated. Anumber of methods have been used to synthesize nanostructuredCeO2 including sol–gel [23], hydrothermal [24], homogeneous pre-cipitation [25], reverse micelles [26], flame spray pyrolysis [27]and sonochemical methods [28]. However, some of these meth-ods require long processing times, and usually, the ceria precursorformed during precipitation is calcined to obtain the product.

In this paper, we report the synthesis of CeO2 nanoparticles by asimple, cost-effective and rapid microwave-assisted method and a

study of the humidity-sensing characteristics of a humidity sensorbased on CeO2 nanoparticles. We have determined the responsetime, the recovery time and the humidity response to understandthese sensor’s true potential as alternative humidity sensors. The

d Actuators B 194 (2014) 260– 268 261

sTts

2

2

eatasdaocd(tayditwtf

2

fAc(s(i

2

es2op(tgtidTMessTF

S. Thakur, P. Patil / Sensors an

tability and repeatability of the response have also been studied.he obtained experimental results provide a possible method forhe rapid synthesis and fabrication of high performance humidityensors based on CeO2 nanoparticles.

. Experimental

.1. Synthesis of CeO2 nanoparticles

The synthesis of CeO2 nanoparticles (NPs) was performed bymploying a microwave-assisted method using analytical-grademmonium cerium(IV) nitrate and propylene glycol without fur-her purification. The ammonium cerium(IV) nitrate was useds the source of cerium, and propylene glycol was used as atabilizing agent. Microwave irradiation was performed with aomestic microwave oven (CQ138S, Samsung, Japan) operatedt a microwave frequency of 2.45 GHz with a maximum powerf 800 W. In a typical experiment, 0.1 M each of ammoniumerium(IV) nitrate and propylene glycol were dissolved in double-istilled water and stirred continuously for 1 h at room temperature25 ◦C). The appropriate amount of ammonia was added drop-wiseo this solution with continuous stirring until the final pH value ofpproximately 10 was achieved in the solution. The resulting pale-ellow-colored precipitate was filtered, washed several times withouble-distilled water and alcohol, and subjected to microwave

rradiation at 600 W microwave power for 10 min. After irradiation,he yellow-colored precipitate was harvested by centrifugation,ashed several times using double-distilled water and ethanol, and

hen dried in an oven at 90 ◦C overnight to obtain the end-productor further characterization.

.2. Characterization

The structural analysis of the as-synthesized CeO2 NPs was per-ormed using an X-ray diffractometer (XRD, D8 Advance, BrukerXS) with Cu K� radiation (� = 1.5418 A). The surface morphologi-al study was performed using a transmission electron microscopeTEM, 1200 EX, JEOL, Japan). The Fourier transform infrared (FTIR)pectroscopy analysis was performed using a FTIR spectrometerFTIR, IMPACT 420 DSP, Nicolet) by the conventional KBr methodn the spectral range 4000–400 cm−1.

.3. Sensor fabrication

In the present work, the sensor consisting of an interdigitatedlectrode (IDE) and a layer of CeO2 NPs coated on top as a humidity-ensing material was fabricated. The size of the entire device is3 mm × 15 mm and the typical dimensions of the sensing areaf the CeO2 NPs are 20 mm × 15 mm. The IDE consists of fiveairs of Cu tracks screen-printed onto an epoxy glass substrate25 mm × 20 mm). The width of a Cu track and the gap betweenwo successive tracks are each 1 mm. Prior to use, the IDE-epoxylass substrates were cleaned by an ultrasonic treatment in ace-one, then rinsed thoroughly with double-distilled water and driedn vacuum. The as-synthesized CeO2 NPs powder was mixed withouble-distilled water in a weight ratio of 100:25 to form a paste.he paste was then spin-coated using the Spin Coater (SPN 2000,ilman Thin Film Systems, Pvt. Ltd., Pune, India) onto the IDE-

poxy glass substrates at 2000 rpm for 12 s to form CeO2 films. After

pin coating, the CeO2 films were dried at 80 ◦C for 12 h and used asensing elements to evaluate the humidity-sensing characteristics.he schematic diagram of a fabricated humidity sensor is shown inig. 1.

Fig. 1. Schematic diagram of a fabricated humidity sensor.

2.4. Humidity-sensing measurements

The different RH levels were generated by the different sat-urated salt solutions in closed bottles at room temperature. Thebottles were made of glass with heights of 19 cm and diameters of6 cm. The six different standard saturated aqueous salt solutions ofLiCl (11 ± 0.30%RH), MgCl2 (33 ± 0.14%RH), K2CO3 (43 ± 0.20%RH),NaCl (75 ± 0.15%RH), KCl (85 ± 0.24%RH) and K2SO4 (97 ± 0.16%RH)were used to act as humidity sources. The saturated salt solu-tions were placed in the bottles for 12 h to ensure that the air inthe bottles reached equilibrium states. The CeO2 film was placedsuccessively into the bottles with different RH levels at room tem-perature, and the impedance of the film was measured as a functionof RH. A humidity probe (Model 6517-RH Humidity probe, Keith-ley Instruments, USA) was also placed into the bottles along withthe CeO2 film to monitor the RH during the measurement. Theimpedance of the CeO2 film was measured as a function of RH usinga simple two-probe configuration with a LCR Meter 4300 (WayneKerr Electronics, USA) controlled by the test software supplied byBiotronic systems, Mumbai, India. The voltage applied was AC 1 V,and the frequency was varied from 60 Hz to 1 kHz.

The Nyquist impedance plots were also recorded using aSolartron impedance gain phase analyzer (SI 1260, Solartron, U.K.)in a two-point configuration in the relative humidity range 11%to 97% and in the frequency range 40 Hz to 100 kHz with analternating-current excitation potential of 10 mV. The analysis ofthe impedance plots was performed by the fitting the experimentalresults to an electrical equivalent circuit with the Z-View softwarefrom Scribner Associates [29]. The quality of the fit to the equivalentcircuit was judged first by the x2 value and second by the limitationof the relative error in the value of each element in the equivalentcircuit to 5%.

3. Results and discussion

3.1. XRD analysis

The XRD pattern of the as-synthesized product shown in Fig. 2(a)is in good agreement with that of the pure CeO2 crystalline phase(JCPDS No.: 81-0792). All of the diffraction peaks can be indexed toCeO2 with a face-centered cubic structure. No other peaks wereobserved, indicating that no impurities were present and con-

firming that the adopted synthesis method gives pure CeO2 NPs.The average crystallite size was calculated by fitting the [1 1 1]diffraction peak (2� = 27.76◦) width using a Gaussian function and

262 S. Thakur, P. Patil / Sensors and Actuators B 194 (2014) 260– 268

(d) H

im

D

giTr

3

ai1ttC7stbb

3

eam(m

Fig. 2. (a) XRD pattern, (b) FTIR spectrum (c) TEM image,

nserting the diffraction angle and peak full line width at half ofaximum (FWHM) in the Debye–Scherrer formula—

= 0.94�

B cos �(1)

where D is the average size of the crystallite, assuming that therains are spherical, � is the wavelength of the X-ray radiation, Bs the peak FWHM in radian and � is the diffraction peak position.he average crystallite size of the CeO2 NPs was found to be in theange of 5–10 nm.

.2. FTIR analysis

The FTIR spectrum of the CeO2 NPs (Fig. 2(b)) shows the bandst approximately 540 and 851 cm−1 corresponding to a character-stic of Ce–O stretching vibrations [30]. The bands at approximately061 and 1385 cm−1 may be attributed to the characteristic vibra-ions of CeO2 [31]. A broad band at ∼3320 cm−1 may be assignedo the stretching vibrational mode of an O–H group bonded to thee atom, i.e., Ce–OH. The bands at approximately 1522, 996 and32 cm−1 are similar to those of commercial CeO2 powders (nothown here) and CeO2 NPs [32]. Compared with the FTIR spec-rum of commercial CeO2 powder, there is no apparent distinctionetween them, which reveals that the nanosized CeO2 synthesizedy the microwave-assisted method is pure.

.3. Morphological study

The TEM image of the as-synthesized product, shown in Fig. 2(c),xhibits the presence of CeO2 NPs with clear boundaries. The

verage grain size of the CeO2 NPs is estimated to be approxi-ately 5–10 nm, which supports the XRD result. The HRTEM image

Fig. 2(d)) shows well-developed lattice fringes, which are in agree-ent with the XRD result. The corresponding selected area electron

RTEM image and (e) SAED pattern of CeO2 nanoparticles.

diffraction (SAED) pattern (as shown in Fig. 2(e)) further confirmsthe random orientations of the CeO2 NPs and that no secondaryphase exists.

3.4. Humidity-sensing performance

To evaluate the humidity-sensing performance of the CeO2NP-based sensor, the impedance of the sensor as a function ofRH has been measured at different frequencies, as shown inFig. 3. The measuring frequency has a noticeable effect on thelog(impedance)–RH characteristic plots. The sensor’s impedancedecreases with increasing RH at each frequency. Moreover, thevalue of the impedance is lower at higher frequencies and becomesalmost independent of the frequency at high RH. The high humid-ity response and linearity in the entire RH range are observedat relatively low measuring frequencies. For example, at 60 Hz,the sensor’s impedance changes by three orders of magnitudefrom 3.91 × 108 to 3.58 × 105 � as the RH increases from 11% to97%. The broken line in the inset of Fig. 3 indicates the linear fit(logZ(˝) = 9.08 − 0.35RH, R2 = 0.9993 where R2 represents the cor-relation coefficient) to the experimental data, illustrating the goodquality of the fit. The change in the sensor’s impedance with RH isdue to the adsorption of water molecules onto CeO2 NPs. The NPs ofCeO2 can increase the active adsorption sites for water moleculesand promote the dissociation of water molecules adsorbed on theirsurface, resulting in a large decrease in the sensor’s impedance withincreasing RH. Hence, the sensor exhibits a high humidity responseand a wide range of humidity detection. At high measuring frequen-cies (i.e., 1 kHz), the impedance plot (cf. Fig. 3) becomes flat becausethe adsorbed water molecules are not polarized, due to the rapid

change in the direction of the electric field at higher frequencies.

A number of experiments have been performed to measurethe impedance of the CeO2 NP-based sensor at 60 Hz by exposingthe sensor to humidity atmospheres ranging from low (11%RH) to

S. Thakur, P. Patil / Sensors and Actuators B 194 (2014) 260– 268 263

F red at various frequencies and 1 V. Inset depicts a linear fit to the impedance-RH curvem

hsas

lTu

S

wahlFiwt6fst

aiimer(wtrsrTtta

100806040200

106

107

108

109

(a)

Impe

danc

e (ΩΩ

)

Relative Humidity (%)

100806040200101

102

103

104

105

106

(b) 60 Hz 10 0 Hz 1 kHz

Hum

idity

Res

pons

e (%

)

ig. 3. Variations in the impedance of CeO2 NPs film with changing RH (%) measueasured at 60 Hz and 1 V.

igh (97%RH), and the corresponding log(impedance)–RH plots arehown in Fig. 4(a). The impedance of the sensor element maintainspproximately constant values at all times, which suggests that theensor has relatively good repeatability.

The humidity response for detecting the humidity was calcu-ated to reveal the characteristics of the CeO2 NPs toward moisture.he humidity response (S) for detecting the humidity is calculatedsing the following expression [33]:

= Zd

Z× 100 (2)

here Zd and Z are the values of the impedance taken at 11%RH andt a particular RH, respectively. Fig. 4(b) shows the variations of theumidity response with the RH and frequency calculated from the

og(impedance)–RH characteristic plots (cf. Fig. 3). As seen fromig. 4(b), the overall response is higher at the lower frequencies,.e., 60 Hz. Additionally, the humidity response increases linearly

ith RH. Because the best linearity of the impedance-RH charac-eristic plot and the best high humidity response were observed at0 Hz, the operating conditions of 1 V and 60 Hz were employed inurther studies to investigate other characteristics of the humidityensor such as the response and recovery behavior, the hysteresis,he stability and the reproducibility.

The response and recovery behavior, corresponding to thedsorption and desorption processes of water molecules, is anmportant characteristics for evaluating the performance of humid-ty sensors [19]. The impedance of the CeO2 NP-based sensor was

onitored at 60 Hz by alternately exposing the sensor to twoxtreme humid atmospheres, 11% and 97%RH. The response andecovery characteristics of the sensor for the low (11%RH)–high97%RH)–low (11%RH) cycle are shown in Fig. 5(a). When the sensoras exposed to a 97%RH atmosphere from an 11%RH atmosphere,

he impedance of the sensor rapidly decreased and then graduallyeached a relatively stable value. Subsequently, when the sen-or was switched to an 11%RH atmosphere again, the impedanceapidly increased and gradually reached a relatively stable value.

he response time and recovery time are defined as the timeaken by a sensor to achieve 90% of the total impedance change inhe case of adsorption and desorption, respectively. The responsend recovery times of the sensor are approximately 3 and 16 s,

Relative Humidity (%)

Fig. 4. (a) Replicates of the measurements for the impedance of CeO2 NPs film as afunction of the RH and (b) dependence of the humidity response on the RH and thefrequency for CeO2 NPs film.

264 S. Thakur, P. Patil / Sensors and Actuators B 194 (2014) 260– 268

50403020100

0

100

200

300

400(a)

Responselow (11 % RH ) to high (97 % RH )

Recoveryhigh (97% RH ) to low (11% RH )

Impe

danc

e (M

ΩΩ)

Time (s)

200150100500

0

50

100

150

200

250

300

350

(b)

97 %R H

11 %R H

Impe

danc

e (M

ΩΩ)

Time (s)

F6(

rr9tbcAioiaaardtoaer

s(ra

ig. 5. (a) Response and recovery characteristics of CeO2 NPs film measured at 1 V,0 Hz and (b) repetitive response of CeO2 NPs film when exposed to three high97%RH)–low (11%RH)–high (97%RH) cycles.

espectively, indicating that the sensor exhibits quick response-ecovery characteristics to an increase in humidity from 11% to7%RH. Although, the sensing element was inevitably exposed tohe laboratory atmosphere during the switching process from oneottle to another, the humidity atmosphere change process wasompleted as soon as possible (which could be done in less than 1 s).s can be seen from Fig. 5(b), the sensor showed good reproducibil-

ty in the continuous measurements and therefore, the correctnessf this experiment is acceptable. The adsorption of water moleculess an exothermic process, and the desorption of these molecules isn endothermic process. Biju et al. [34] reported that the responsend recovery times of the humidity sensor depend on the heat ofdsorption/desorption and the diffusion of a gaseous molecule. Theesponse time should be equal to the recovery time for an idealiffusion-controlled adsorption–desorption process. The observa-ion of the rapid response to humidity (∼3 s) indicates that the heatf adsorption is higher, which implies that the CeO2 NPs have morective adsorption sites for water molecules. Consequently, a higherxternal energy is required for the desorption process, which iseflected in the relatively slow observed recovery.

The impedance of the CeO2 NP-based sensor was mea-

ured by repeatedly exposing the sensor to a low (11%RH)–high97%RH)–low (11%RH) cycle to examine the reproducibility andeversibility. The measurements were repeated for three cycles,nd the resulting response and recovery characteristics are shown

Fig. 6. Humidity hysteresis for CeO2 NPs film.

in Fig. 5(b). The impedance of the sensor always reverts to theoriginal value when the RH is restored to the former state, indicat-ing that the humidity-sensing process is extremely reversible. Theresponse and recovery times do not change during the three cyclesof measurements, indicating a good reproducibility of the humid-ity response. Thus, the CeO2 NP-based humidity sensor exhibitsa super-rapid response (a response time of 3 s) with very goodrepeatability and stability.

Hysteresis is one of the most important characteristics of ahumidity sensor [19] and is defined as the maximum differencebetween the adsorption and desorption curves. The humidity hys-teresis of the CeO2 NP-based humidity sensor was determined bymeasuring the impedance of the sensor as a function of the RH forthe low (11%RH)–high (97%RH)–low (11%RH) cycle at 60 Hz and 1 V,as shown in Fig. 6. Note that the sensor exhibits highly reversiblesensing properties and that the sensing curves for the humidifica-tion and desiccation processes almost overlap, showing very smallhysteresis. The humidity hysteresis error (�H) was calculated usingthe following expression [35]:

�H = ±�Hmax

2FFS(3)

where �Hmax is the difference in the output of the forward andbackward operations and FFS is the full scale output. The maximumabsolute value of the humidity hysteresis error �H is found to be∼1%RH in the range of 11–97%RH, indicating a good reliability ofthe sensor.

The stability behavior is a crucial characteristic for evaluatingthe performance of humidity sensors in practice [19]. To inves-tigate the stability of the CeO2 NP-based humidity sensors, thesensors were exposed in air for 30 days, and a measurement ofthe impedances was performed once in every five days at 11%, 43%and 97%RH. The impedance variations with time are shown in Fig. 7.As seen from Fig. 7, after 30 days of testing, it was found that theimpedance variation is less than 2% in each humidity region, whichdirectly confirms the excellent stability of the sensor.

To understand the humidity-sensing mechanism, the Nyquistimpedance plots for the CeO2 NP-based humidity sensor wererecorded in the frequency range from 40 Hz to 100 kHz within theRH range of 11–97%. The typical Nyquist impedance plots of theCeO2 NP-based humidity sensor are shown in Fig. 8. At low RH lev-

els (11%, 33% and 43%RH), the impedance plots describe a part of thesemicircle, and the radius of curvature decreases with increasingRH. At an RH level of 75%, a complete semicircle becomes visi-ble inside the measurable impedance range, and a spur in the low

S. Thakur, P. Patil / Sensors and Actu

30252015105105

106

107

108

109

(c)

(b)

(a)

Impe

danc

e (ΩΩ

)

Time (days)

fsl

lNatddtri

tcwNn[tctsbbiafsctll

fwtcaobai

decreases linearly with an increase in the RH. For low humidity

Fig. 7. Stability of CeO2 NPs film measured at (a) 11%, (b) 43% and (c) 97%RH.

requency region begins to appear along with the semicircle. Theize of the semicircle in the high frequency decreases, and the spurengthens with further increasing RH.

Several researchers have attributed the observed semicircle atow RH to the intrinsic impedance of the sensing film, i.e., the CeO2Ps film [8,33,36]. In this condition, only a few water molecules aredsorbed by the chemisorption. Because the coverage of water onhe surface of CeO2 NPs is not continuous, the ionic conduction isifficult. The interaction of water with CeO2 NPs is also very weakue to the adsorption of a few water molecules. Therefore, whenhe RH is low, the conduction process is mainly due to the CeO2 NPsather than the adsorbed water molecules and thus, the complexmpedance plot is a semicircle (cf. Fig. 9(a)).

With increasing RH, more water molecules are physisorbed onhe surface of CeO2 NPs. Additionally, there exists a high localharge density and a strong electric field in the tips and defects,hich promotes the ionization of water physisorbed on the CeO2Ps surface [20]. The ionization produces a large number of hydro-ium ions (H3O+) as charge carriers by a chain of exchange reactions37]: H3O+ + H2O ↔ H2O + H3O+. Simultaneously, H+ transfer alsoakes place between adjacent H2O molecules in clusters. The con-entrations of H3O+ and H+ become higher as the RH increases andhe diffusion phenomenon takes place. For the RH level of 75%, theemicircle becomes smaller, and a spur in the low frequency regionegins to appear along with the semicircle. The spur looks like theeginning of another semicircle in the low frequency region, which

s a consequence of the diffusion of the electroactive species H3O+

nd H+ in the adsorbed water layer toward the electrode. This dif-usion of ions results into the accumulation of ions at the electrodeurface [8,33,36]. Thus, the spur represents the mass transfer pro-ess, which is related to the amount of water vapor adsorbed onhe sensing layer. As the relaxation time for diffusion of ions isonger than that for charge transfer, the spur is observed only atow frequencies.

With a further increase in the RH, the spur observed in the lowrequency region, which reflects the contribution of the adsorbedater to the conduction, becomes larger and larger. At the same

ime, the semicircle in the high frequency region, which reflects theontribution of the CeO2 NPs to the conduction, becomes smallernd smaller as the RH increases. In this condition, the coveragef physisorbed water molecules on the surface of the CeO2 NPs

ecomes continuous. Moreover, more and more water moleculesre trapped between the electrodes and the CeO2 NPs, resultingn the stronger diffusion behavior. The physisorbed water layers

ators B 194 (2014) 260– 268 265

gradually exhibit a liquid-like behavior, and the hopping of the pro-tons (H+) between adjacent water molecules occurs with a chargetransport by a Grotthuss chain reaction [38]. The free movement ofH+ ions along with water layer causes a decrease in the impedanceof the film. Thus, these results reveal that three conduction mech-anisms, viz., ion hopping, ion diffusion and electrolytic conduction,depend on the amount of the adsorbed water on the surface of theCeO2 NPs. At low RH, the intrinsic charge carriers of the CeO2 NPsare the dominant charge carriers for conductivity. At medium RH,H3O+ ions mainly contribute to the conductivity, whereas H3O+ ionsand H+ hopping become dominant in the transport process at highRH.

The equivalent circuit depicted in Fig. 9(a), as suggested by Linet al. [8], was used to model the Nyquist impedance plots (cf. Fig. 8)of the humidity sensor based on the CeO2 NPs. The analysis of theimpedance plots was performed by fitting the experimental resultsto an equivalent circuit with the Z-View software from ScribnerAssociates [29]. The equivalent circuit consists of series and par-allel combinations of resistances (R) and CPEs. The experimentaldata points are represented as dots, and the fits are represented ascontinuous lines. Because the impedance plots exhibit depressedsemicircles, the CPEs have been used instead of the pure capacitorsin the equivalent circuit.

The series resistance Rs is related to the region near the elec-trode, which is weakly dependent on the RH. The resistance R1characterizes the bulk resistance of the CeO2 NPs film and repre-sents the contributions of all conduction mechanisms. The CPE1element characterizes the contributions of electronic and ionictransduction among CeO2 NPs and the capacitance effect betweenthe NPs. The CPE2 element represents the contact impedancebetween the electrode/sensing film interface and includes the con-tributions of both the capacitance and resistance. The CPE has acomplex impedance given by [39]

ZCPE = 1Q (jω)n (4)

where j = √−1, Q is the capacitance, ω is the angular frequencyand n is a constant that lies between 0 and 1. At n = 0 and n = 1,the CPE element behaves like a resistor (ZCPE = 1/Q, n = 0) and acapacitor (ZCPE = −j 1

Qω , n = 1), respectively. Based on the equiv-alent circuit, the overall impedance of the CeO2 NPs sensor can beexpressed as

Z = 11

ZCPE1+ 1

(R1+ZCPE2) + RS

(5)

Thus, the humidity response of the CeO2 NP-based sensordepends on variations of all the parameters in the components ofthe equivalent circuit with RH. The impedance of the CPE elementwas calculated by using Eq. (4) with ω = 60 Hz. The variation of theresistance R1 as a function of the RH is shown in Fig. 9(b). The valueof R1 decreases linearly with an increase in the RH and is attributedto the improved electrical conductance of the CeO2 NPs film dueto the film’s enhanced interaction with an increasing number ofadsorbed water molecules. A change in R1 of three orders of magni-tude, from 1.55 × 108 to 1.28 × 105 �, has been observed as the RHincreases from 11% to 97%. The relationship between the resistanceR1 and the RH can be expressed as

log R1(˝) = −0.035RH + 8.54 (6)

The variations of the index (n1) and the impedance of CPE1 as afunction of the RH are shown in Fig. 9(c). The impedance of CPE1

levels, n1 is approximately 1, and consequently, the CPE1 elementbehaves more like a capacitor, thereby exhibiting a high impedance(ZCPE1 = 2.54 × 109 ˝). With an increase in the RH, although a

266 S. Thakur, P. Patil / Sensors and Actuators B 194 (2014) 260– 268

2001501005000

50

100

150

200(a) Experime ntal data

Fitted cu rveZ

(im) /

MΩΩ

Z (re)/ M ΩΩ6040200

0

20

40

60(b) Exper iment al data

Fitt ed curve

Z (im

) /M

Ω

Z (re)/ MΩ

141210864200

2

4

6

8

10

12

14 (c) Exper imen tal data Fitted curve

Z (im

) /M

Ω

Z (re)/ MΩ

1.51.20.90.60.30.00.0

0.3

0.6

0.9

1.2

1.5(d) Experimental da ta

Fitt ed cur ve

Z (im

) /M

Ω

Z (re)/ MΩ

75060045030015000

150

300

450

600

750(e) Experime ntal d ata

Fitted curve

Z (im

) / k

Ω

Z (re)/ k Ω28024020016012080

0

40

80

120

160

200

240

280(f) Exper imental da ta

Fitted curve

Z (im

) /k

Ω

Z (re)/ kΩ

F H, (c)

p .

stThra

l

Fig. 9(d) shows the variations of the index (n2) and theimpedance of CPE2 as a function of the RH. The n2 fluctuates approx-imately 0.4, and the impedance of CPE2 decreases linearly with anincrease in the RH, which can be expressed as

ig. 8. Nyquist impedance plots for CeO2 NPs film recorded at (a) 11%RH, (b) 33%Roints, and continuous lines represent simulated curve using the equivalent circuit

light decrease in the value of n1 is observed, the impedance ofhe CPE1 element still exhibits a high value (ZCPE1 = 3.97 × 108 ˝).hus, only one order of magnitude change in the impedance of CPE1as been observed over the entire RH range (11–97%RH), and the

elationship between the impedance and the RH can be expresseds

og ZCPE1 (˝) = −0.0093RH + 9.56 (7)

43%RH, (d) 75%RH, (e) 85%RH and (f) 97%RH. Dots represent the experimental data

log ZCPE2 (˝) = −0.035RH + 8.78 (8)

S. Thakur, P. Patil / Sensors and Actuators B 194 (2014) 260– 268 267

F f the v(

CRhoctwitr

4

srnfmstnoaraohT

ig. 9. (a) Equivalent circuit used to model the impedance plots and dependence on2) and impedance of the CPE2 element (d) on the RH.

A change of three orders of magnitude in the impedance ofPE2 from 2.97 × 108 to 3.76 × 105 � has been observed as theH increases from 11% to 97%. Thus, these results reveal that theumidity response of the CeO2 NP-based humidity sensor mainlyriginates from the variation of the bulk resistance (R1) and theontact impedance (ZCPE2 ). The linear decrease of the magnitudes ofhree components of the equivalent circuit (i.e., R1, ZCPE1 and ZCPE2 )ith an increase in the RH results in the significant decrease in the

mpedance of the CeO2 NP-based humidity sensor with the adsorp-ion of water, which is in good agreement with the experimentalesults shown in Fig. 3.

. Conclusions

In conclusion, CeO2 nanoparticles were successfully synthe-ized at a low-cost using a microwave-assisted method. The XRDesult reveals the formation of a face-centered cubic phase of CeO2anoparticles with good crystallinity. The TEM study showed the

ormation of nanoparticles with average particle sizes of approxi-ately 5–10 nm. The experimental results demonstrate a low-cost,

imple, rapid, and suitable microwave-assisted method for the syn-hesis of CeO2 nanoparticles. A humidity sensor based on CeO2anoparticles exhibits a high response with an impedance variationf three orders of magnitude to humidity change, a linear response,

wide humidity detection range, a small hysteresis, excellentepeatability and a rapid response and recovery behavior (3 s for

dsorption and 16 s for desorption). The Nyquist impedance plotsf the sensor at different RHs were used to elucidate the sensor’sumidity-sensing mechanism via an electrical equivalent circuit.his study demonstrates that the CeO2 nanoparticles synthesized

alue of R1 (b), the index (n1) and impedance of the CPE1 element (c) and the index

by the microwave-assisted method can be used as the humidity-sensing material for the fabrication of humidity sensors.

Acknowledgments

The authors acknowledge the financial support from the Boardof Research in Nuclear Sciences (BRNS), Department of AtomicEnergy (DAE), Mumbai (Project No. 2008/37/44/BRNS/2468) andfrom the University Grants Commission (UGC), New Delhi (ProjectNo. F. 530/2/DRS/2010—SAP-DRS, Phase-II). Sharada Thakur isthankful to UGC, New Delhi, for awarding the Rajiv Gandhi NationalFellowship.

References

[1] Z. Chen, C. Lu, Humidity sensors: a review of materials and mechanisms, SensorLett. 3 (2005) 274–295.

[2] B. Cheng, B. Tian, C. Xie, Y. Xiao, S. Lei, Highly sensitive humidity sensor basedon amorphous Al2O3 nanotubes, J. Mater. Chem. 21 (2011) 1907–1912.

[3] A. Zainelabdin, G. Amin, S. Zaman, O. Nur, J. Lu, L. Hultman, M. Willander,CuO/ZnO nanocorals synthesis via hydrothermal technique: growth mech-anism and their application as humidity sensor, J. Mater. Chem. 22 (2012)11583–11590.

[4] Z. Wang, Y. Lu, S. Yuan, L. Shi, Y. Zhao, M. Zhang, W. Deng, Hydrothermal synthe-sis and humidity sensing properties of size-controlled zirconium oxide (ZrO2)nanorods, J. Colloid Interface Sci. 396 (2013) 9–15.

[5] P. Biswas, S. Kundu, P. Banerji, S. Bhunia, Super rapid response of humiditysensor based on MOCVD grown ZnO nanotips array, Sens. Actuators, B 178(2013) 331–338.

[6] Y. He, X. Liu, R. Wang, T. Zhang, An excellent humidity sensor with rapidresponse based on BaTiO3 nanofiber via electrospinning, Sens. Lett. 9 (2011)262–265.

[7] L. Li, K. Yu, J. Wu, Y. Wang, Z. Zhu, Structure and humidity sensing propertiesof SnO2 zigzag belts, Cryst. Res. Technol. 45 (2010) 539–544.

2 d Actu

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[[

Maharashtra University, Jalgaon, India. He received his master’s degree (M.Sc.) in

68 S. Thakur, P. Patil / Sensors an

[8] Q. Lin, Y. Li, M. Yang, Investigations on the sensing mechanism of humiditysensors based on electrospun polymer nanofibers, Sens. Actuators, B 171 (2012)309–314.

[9] T. Fei, H. Zhao, K. Jiang, X. Zhou, T. Zhang, Polymeric humidity sensors withnonlinear response: properties and mechanism investigation, J. Appl. Polym.Sci. 130 (2013) 2056–2061.

10] K. Suri, S. Annapoorni, A.K. Sarkar, R.P. Tandon, Gas and humidity sensorsbased on iron oxide–polypyrrole nanocomposites, Sens. Actuators, B 81 (2002)277–282.

11] D. Patil, Y.K. Seo, Y.K. Hwang, P. Patil, Poly(o-anisidine)—tin oxide nanocom-posite: synthesis, characterization and application to humidity sensing, Sens.Actuators, B 148 (2010) 41–48.

12] P.G. Su, L.N. Huang, Humidity sensors based on TiO2 nanoparticles/polypyrrolecomposite thin films, Sens. Actuators, B 123 (2007) 501–507.

13] Q. Kuang, C. Lao, Z.L. Wang, Z. Xie, L.J. Zheng, Zn-doped SnO2 with 3D cubicstructure for humidity sensor, Am. Chem. Soc. 129 (2007) 6070–6071.

14] Y. Lu, Z. Wang, S. Yuan, L. Shi, Y. Zhao, W. Deng, Microwave-hydrothermalsynthesis and humidity sensing behavior of ZrO2 nanorods, RSC Adv. 3 (2013)11707–11714.

15] F. Fang, J. Futter, A. Markwitz, J. Kennedy, UV and humidity sensing propertiesof ZnO nanorods prepared by the arc discharge method, Nanotechnology 20(2009) 245502–245509.

16] Z. Feng, X.J. Chen, J. Chen, J. Hu, A novel humidity sensor based on aluminananowire films, J. Phys. D: Appl. Phys. 45 (2012) 225305–225311.

17] Q. Wang, Y.Z. Pan, S.S. Huang, S.T. Ren, P. Li, Resistive and capacitive responseof nitrogen-doped TiO2 nanotubes film humidity sensor, Nanotechnology 22(2011) 025501–025522.

18] Y. He, T. Zhang, W. Zheng, R. Wang, X. Liu, Y. Xia, J. Zhao, Humidity sensingproperties of BaTiO3 nanofiber prepared via electrospinning, Sens. Actuators,B 146 (2010) 98–102.

19] D. Bauskar, B.B. Kale, P. Patil, Synthesis and humidity sensing properties ofZnSnO3 cubic crystallites, Sens. Actuators, B 161 (2012) 396–400.

20] X.Q. Fu, C. Wang, H.C. Yu, Y.G. Wang, T.H. Wang, Fast humidity sensors basedon CeO2 nanowires, Nanotechnology 18 (2007) 145503–145507.

21] Z. Zhang, C. Hu, Y. Xiong, R. Yang, Z.L. Wang, Synthesis of Ba-doped CeO2

nanowires and their application as humidity sensors, Nanotechnology 18(2007) 465504–465509.

22] C.H. Hu, C.H. Xia, F. Wang, M. Zhou, P.F. Yin, X.Y. Han, Synthesis of Mn-dopedCeO2 nanorods and their application as humidity sensors, Bull. Mater. Sci. 34(2011) 1033–1037.

23] M. Alifanti, B. Baps, N. Blangenois, J. Naud, P. Grange, B. Delmon, Character-ization of CeO2–ZrO2 mixed oxides: comparison of the citrate and sol–gelpreparation methods, Chem. Mater. 15 (2003) 395–398.

24] Y.C. Zhou, M.N. Rahaman, Hydrothermal synthesis and sintering of ultrafineCeO2 powders, J. Mater. Res. 8 (1993) 1680–1686.

25] J. Shih, Y.J. Chen, M.H. Hon, Synthesis and crystal kinetics of cerium oxidenanocrystallites prepared by co-precipitation process, Mater. Chem. Phys. 121(2010) 99–102.

26] S. Tsunekawa, T. Fududa, A. Kasuya, Blue shift in ultraviolet absorption spectraof monodisperse CeO2−x nanoparticles, J. Appl. Phys. 87 (2000) 1318–1321.

ators B 194 (2014) 260– 268

27] S. Masson, P. Holliman, M. Kalajia, P.J. Kluson, The production of nanoparticu-late ceria using reverse micelle sol gel techniques, J. Mater. Chem. 19 (2009)3517–3522.

28] J.C. Yu, L. Zhang, J.J. Lin, Direct sonochemical preparation of high-surface-areananoporous ceria and ceria–zirconia solid solutions, J. Colloid Interface Sci. 260(2003) 240–243.

29] Electrochemical Corrosion Software, CorrWare and CorrView, Scribner Asso-ciates, Southern Pines, NC, 2006.

30] J. Liu, F. Wang, R. Dewil, CeO2 nanocrystalline-supported palladium chloride:an effective catalyst for selective oxidation of alcohols by oxygen, Catal. Lett.130 (2009) 448–454.

31] L. Yue, X.M. Zhang, Structural characterization and photocatalytic behaviors ofdoped CeO2 nanoparticles, J. Alloys Compd. 475 (2009) 702–705.

32] D.S. Zhang, H.X. Fu, L.Y. Shi, C.S. Pan, Q. Li, Y.L. Chu, W.Y. Yu, Synthesis of CeO2

nanorods via ultrasonication assisted by polyethylene glycol, Inorg. Chem. 46(2007) 2446–2451.

33] D. Patil, Y.K. Seo, Y.K. Hwang, J.S. Chang, P. Patil, Humidity sensitive poly(2,5-dimethoxyaniline)/WO3 composites, Sens. Actuators, B 132 (2008) 116–124.

34] K.P. Biju, M.K. Jain, Effect of polyethylene glycol additive in sol on the humiditysensing properties of a TiO2 thin film, Meas. Sci. Technol. 18 (2007) 2991–2996.

35] X. Wang, M. Ye, Hysteresis and nonlinearity compensation of relative humiditysensor using support vector machines, Sens. Actuators, B 129 (2008) 274–284.

36] S. Hu, G. Fu, Humidity-sensitive properties based on liquid state LiZnVO4-dopedSnO2, Sens. Actuators, A 163 (2010) 481–485.

37] J.J. Fripiat, A. Jelli, G. Poncelet, J. Andre, Thermodynamic properties of adsorbedwater molecules and electrical conduction in montmorillonites and silicas, J.Phys. Chem. 69 (1965) 2185–2197.

38] F.M. Ernsberger, The nonconformist ion, J. Am. Chem. Soc. 66 (1983) 747–750.39] P.M. Faia, C.S. Furtado, A.J. Ferreira, AC impedance spectroscopy: a new equiva-

lent circuit for titania thick film humidity sensors, Sens. Actuators, B 107 (2005)353–359.

Biographies

Sharada Thakur is a research scholar in Department of Physics, North MaharashtraUniversity, Jalgaon, India, doing research work in the field of conducting poly-mers/metal oxide nanocomposites. She received master’s degree (M.Sc.) in physicsin 2011 from North Maharashtra University, Jalgaon, India. She is currently pursuinga Ph.D. in physics. Her research interests include conducting polymers/metal oxidenanocomposites for chemical and humidity sensors.

Pradip Patil is working as professor in physics at Department of Physics, North

physics in 1983 and Ph.D. degree in 1988 from the University of Pune, Pune, India. Hismain interests include development of conducting polymers and conducting poly-mer nanocomposites for their applications as corrosive protective coatings, chemicaland biological sensors.