- Email: [email protected]
Oxide nanoparticle arrays for sensors of CO and NO2 gases
Vacuum 86 (2012) 590e593
Contents lists available at ScienceDirect
Vacuum journal homepage: www.elsevier.com/locate/vacuum
Oxide nanoparticle arrays for sensors of CO and NO2 gases S. Luby a, *, L. Chitu a, M. Jergel a, E. Majkova a, P. Siffalovic a, A.P. Caricato b, A. Luches b, M. Martino b, R. Rella c, M.G. Manera c a
Institute of Physics, Slovak Academy of Sciences, Dubravska cesta 9, 845 11 Bratislava, Slovakia Department of Physics, University of Salento, Via Arnesano, 73100 Lecce, Italy c Institute of Microelectronics and Microsystems, IMM-CNR, Unit of Lecce, 73100 Lecce, Italy b
a b s t r a c t Keywords: Nanoparticles LangmuireBlodgett layers Gas sensing
Metal oxide sensors with active films from Fe2O3 and CoFe2O4 hexagonally assembled nanoparticle (NP) arrays were studied. NPs were synthesized by high-temperature solution phase reaction. Sensing NP layers were deposited by LangmuireBlodgett (LB) technique. LB layers were characterized by XRD, SEM and magnetic measurements. NPs are monodomain and superparamagnetic at RT. Sensor active films are formed from 1 or 7 LB monolayers deposited on the alumina substrates equipped with heating meander and interdigitated contacts. LB layers were heat-treated or UV irradiated to remove the insulating surfactant. Sensing properties were studied in a test chamber containing reducing (CO) or oxidizing (NO2) gas in concentrations between 5 and 100 ppm in the mixture with dry air. Response current signal (Igas/Iair) vs. temperature and response vs. gas concentration calibration curves were measured. Best response values were obtained with CoFe2O4 devices between 300 and 400 C, being 3 and 10 for 100 ppm of CO and 5 ppm of NO2, respectively. The response and recovery times of sensors are between 3 and 30 min. The Fe2O3 and CoFe2O4 sensors with response of 8 or 10 to 5 ppm of NO2 might be of practical applications in the detection of explosives. Ó 2011 Elsevier Ltd. All rights reserved.
1. Introduction Bardeen and Brattain discovered that metal oxide semiconductor resistance could be modified by ambient gas [1]. Seiyama et al. proposed to use these materials in gas sensors [2]. Since that time many devices were developed. From the multitude of oxides for sensing of gases we mention at least SnO2 [3,4], WO3 [3,5], CoFe2O4 [6], Fe2O3, Fe3O4, In2O3 [7] and TiO2 [8]. Among the monitored gases attention is paid mainly to air pollutants like SO2, CO, CO2, CH4, NO, NO2, O2, O3, Cl2, C2H5OH, H2O2 [9,10]. Moreover, NO2 is of primary importance for trace amount detection of explosives like EGDN, TNT, PETN, RDX, etc. [11,12]. NO2 sensors must be extremely sensitive because some explosives have very low vapor pressure. At RT and atmospheric pressure of air it decreases from 50 ppm for EGDN to 30 ppb for RDX and PETN, the pressure of TNT and nitroglycerin being 1 ppm and 1 ppb, respectively [11]. In the detection of explosives a variety of techniques using fluorescent polymers, conducting polymer composites, surface acoustic waves, etc., are used [13]. Gas sensors and electronic noses attract a lot of interest because the use of dogs is expensive [11,13]. * Corresponding author. Tel.: þ421 903 713 696; fax: þ421 2 5477 6085. E-mail address: [email protected] (S. Luby). 0042-207X/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.vacuum.2011.07.018
Innovative sensors are built from nanoparticle (NP) arrays [9,14]. In comparison with continuous films the devices have higher surface/volume ratio. Consequently, the surface scattering is better influenced by adsorbed species and the change of sensor conductivity is increased. In this work we experiment with sensors from Fe2O3 and CoFe2O4 NPs fabricated in the colloid solution by a chemical route. They are non-toxic, moreover, the size dispersion of NPs is low (10%). Therefore, the conductivity of the array could be quite simply computed. 2. Experimental Monodisperse Fe2O3 (maghemite) and CoFe2O4 NPs were synthesized by high-temperature solution phase reaction of metal acetylacetonates (Fe(acac)3, Co(acac)2) with 1,2-hexadecanediol, oleic acid and oleyl amine in phenyl ether at 200 C/30 min plus 265 C/30 min heating [15]. Solvent was toluene. Structure of NPs was studied by grazing incidence (GI) XRD-D8 Bruker. Here NP solution was dropped onto Si substrate and dried. The thickness of the deposit was 10e20 NP monolayers (MLs). Size of NPs was estimated using HR SEM Leo 1550. Magnetic properties were studied between 4.2 and 278 K using vibrating sample magnetometer and diluted NP solution in a capillary.
S. Luby et al. / Vacuum 86 (2012) 590e593
591
Before measurements samples were held 12 h in dry air flux at 400 C to remove the surfactant coverage. Then NP arrays agglomerate. The 1 ML deposit transforms into labyrinth-like structure with certain level of percolation and 7 ML deposit aggregates into a compact structure. Alternatively, some samples were treated by UV radiation during 5 min at the flux 2 mW/cm2 in the ambient of ozone in the dry air (Fig. 1b). Here the transformed structures are less compact and they preserve NP nature [16]. Measurements were performed at different operating temperatures using the heating element. The sensor response to reducing (CO) or oxidizing (NO2) gas was monitored by an electrometer at a constant voltage of 5e10 V between electrodes. The desired concentrations of the gases in dry air were obtained using certified gas bottles and mass flow meters. The temperature of the bottles was about 22 C for both air and air containing CO and NO2. Flow of 100 sccm was fixed during the tests. At first conditioning of sensor in the test chamber in a flow of dry air was performed until stabilization of the sensor response signal. After it a mixture containing CO or NO2 was sent into a chamber. 3. Results and discussion
Fig. 1. Fe2O3 NP array before (a) and after (b) UV irradiation in the ambient of ozone in the dry air. The horizontal line corresponds to 100 nm.
NP LangmuireBlodgett (LB) films were prepared from the water subphase in the 612 D NIMA trough at the surface pressure of 20 mN/m. 1 or 7 ordered LB MLs were deposited by modified (patent pending) procedure. Gas sensors were fabricated on 3 mm 3 mm alumina substrates. They were equipped with 20 nm Ti/200 nm Pt interdigitated electrodes on the front side to read the measuring current and with 20 nm Ti/500 nm Pt heating meander on the back side. The temperature of sensors was obtained from poweretemperature curve. Calibration was done in an oven using a thermocouple and measuring the resistance of the heating element in dynamic conditions under the gas flow. After deposition of the sensing layers the devices were soldered onto a commercial TO-8 socket and they were fixed in a test chamber.
From GI XRD spectra it follows that the diffraction peaks of FeeO system correspond to either Fe3O4 (magnetite, cubic, fcc) or Fe2O3 (maghemite, cubic, primitive cell). From the peak 400 at larger 2q angles Fe2O3 composition is preferred. (Moreover, under the ageing/heating in the air ambient or UV irradiation Fe3O4 is oxidized to Fe2O3 [17].) In combination with Co, ternary compound CoFe2O4 is formed. It is isomorph with Fe3O4. The diameter of Fe2O3/Fe3O4 and CoFe2O4 NPs is 6.4 0.6 and 7.6 0.6, respectively, surfactant is 1.8 nm thick in both cases and blocking temperatures are 22 K and 204 K, respectively. NPs are monodomain and superparamagnetic at RT [15]. An example of hexagonal-like NP arrays before and after UV processing is shown in Fig. 1a,b, respectively. In Table 1 sensor characteristics, i.e. structure of the devices and working conditions for two monitored gases are summarized. The interval of working temperatures is 150e475 C. These temperatures are typical for many metal oxide semiconductor sensors [3,4]. The maximum in the sensor response vs. working temperature for CoFe2O4 with 7 MLs is shown in Fig. 2. Here any value of the response was obtained having the sensor at a specific working temperature in the test chamber. Data for other devices are in Table 1. The maxima of R vs. Tw are the consequence of reaching the best dynamic equilibrium between adsorption and desorption of the monitored gas [4,18]. In Table 1 measuring conditions and sensor responses are shown as well. Both Igas and Iair were obtained in the stabilized conditions described above. The base current in air is of the order 109e1011 A with 1 ML deposits, which corresponds
Table 1 Structure and operating conditions of sensors with 1 or 7 monolayers (MLs) of NPs (first column). Cg is the gas concentration, Tw is the interval of working temperatures and Tm e maximum response working temperature measured at the concentration Cm. Sensor base current in dry air Iair was obtained at Tm or in the middle of Tw interval. The sensor response is R ¼ Igas/Iair. NPs/No. of MLs
Gas
Cg [ppm]
Tw [ C]
Tm/Cm [ C]/[ppm]
Iair [A]
R
Fe2O3/1 Fe2O3/1 Fe2O3/7 Fe2O3/7 CoFe2O4/1 CoFe2O4/1 CoFe2O4/7 CoFe2O4/7 CoFe2O4/7 UV CoFe2O4/7 UV
CO NO2 CO NO2 CO NO2 CO NO2 CO NO2
20e100 5 100 5 20e100 100 2e100 100 100 5
325e475 375 200e475 300e475 175e450 350 300e475 375 250e475 300e475
450/100 Not evaluated 350/100 300/5 340/100 Not evaluated 425/100 Not evaluated 425/100 375/5
3 109 2.4 109 6 107 8 107 2 1011 4 1011 6 107 3 107 7 108 8 108
1.3 (400 C, 100 ppm) (1.8)1 (375 C, 5 ppm) 2.8 (350 C, 100 ppm) (8)1 (350 C, 5 ppm) 1.9 (350 C, 100 ppm) y1 1.6 (400 C, 100 ppm) y1 3 (350 C, 10 ppm) (10)1 (350 C, 5 ppm)
592
S. Luby et al. / Vacuum 86 (2012) 590e593
1.6
1.5
Response (Igas/Iair)
Response (Igas/I0)
1.3
1.4
1.3
1.2
1.1
1.0
1.2 300
350
400
0
450
20
40
60
80
100
CO concentration (ppm)
Working Temperature (°C) Fig. 2. Response of CoFe2O4 sensor with 7 MLs of NPs vs. working temperature at 100 ppm of CO gas.
to not fully continuous labyrinth-like structure. Therefore, we created 7 MLs deposits. The increase of the base current by 2e4 orders of magnitude is not a simple consequence of thicker conducting film but predominantly of the formation of continuous structure. This can be justified using Fig. 1b, where approximately 37% of the area is striped of NPs. Then the average thickness of the layer is 63% of the non-irradiated one and the thickness of 7 monolayers is 7 6.4 nm ¼ 4.5 nm, or even less, assuming the hexagonal closed packing in the vertical direction. (From calculation using the diameter of NPs and thickness of surfactant we obtain even smaller relative coverage of the surface by irradiated NP monolayer, being 41%. This, however, corresponds to densely packed irradiated NPs, which obviously is not the case.) The increase of the base current to 107e108 A (Table 1) might on one side screen the sensor response, but on the other side these values are important for easy electronic integration solution. The response at the chemisorption of oxidizing NO2 gas is shown as (X)1, because in this case capturing of electrons from sensing layers in the n-type semiconductor diminishes the conductivity. With reducing CO gas we have opposite effect [3,4]. From the comparison of the heat-treated and UV irradiated CoFe2O4 sensors with 7 MLs of NPs we can see, that in the second case we have much higher responses for both monitored gases (Table 1). This is attributed to the preservation of NP structure which is important for increasing the adsorption/absorption capacity of the sensor. In Fig. 3 are the time dependences of current in heat-treated Fe2O3 and UV irradiated CoFe2O4 sensors in oxidizing NO2 ambient. In both cases the response is good (Table 1) and devices are of some practical use.
Fig. 4. Calibration curve of CO sensor with 1 ML of Fe2O3 sensing film at a working temperature 400 C.
The published response and recovery times of metal oxide sensors are between tens of seconds and tens of minutes, depending especially on the structure and thickness of the sensing film. Short times are observed for thin nanocrystalline films where electron processes take place at the surface [3,4,7,10]. Our results are similar, the times are between 3 and 30 min, being lower for 1 ML sensing films. The recovery of sensors is satisfactory (Fig. 3). This is in accord with results of [7], where almost 100% recovery for FeeO system was reported. The calibration curves of sensors, i.e. response vs. gas concentration, are linear for both 1 ML (e.g. Fig. 4) and 7 ML sensing films at Tw approaching to 400 C. At temperatures around 300 C 1 ML sensors sometimes show saturation at CO concentration of 100 ppm. 4. Conclusions We have shown that gas sensors with sensing films composed from 1 or 7 monolayers of Fe2O3 or CoFe2O4 NPs are satisfactory for monitoring reducing CO or oxidizing NO2 gases. With 1 ML films the base current is low because after removing the surfactant the sensing film becomes partly discontinuous. On the other hand films from 7 LangmuireBlodgett monolayers are probably too thick and therefore they have higher response and recovery times. Improvement of devices is expected with sensing films composed from 2 to 4 monolayers of nanoparticles. The Fe2O3 and CoFe2O4 sensors with response of 8 or 10 to 5 ppm of NO2 might be of some practical applications in detection of explosives. Acknowledgment The work was supported by the following grants: CANR, ITMS code 26240220011 ERDF, APVV SK-IT-0027-08 and by the Scientific Grant Agency VEGA, 2/0126/09. The support of EU HASYLAB grant II-20080036 EC is also acknowledged.
Current (A)
10
10
References 10
10
10
0
1
2
3
4
5
6
7
8
time (h) Fig. 3. Dynamic responses to 5 ppm of NO2 at 350 C with Fe2O3, 7 ML heat-treated sensor (full line) and CoFe2O4, 7 ML UV irradiated sensor (broken line).
[1] Bardeen J, Brattain WH. Bell Syst Tech J 1953;32:1. [2] Seiyama T, Kato A, Fujishi K, Nagatani M. Anal Chem 1962;34:1502e7. [3] Heusler TP, Lorke A, Ifeacho P, Wiggers H, Schulz C. J Appl Phys 2007;102: 124305. [4] Gurlo A, Barsan N, Ivanovskaya M, Weimar U, Göpel W. Sens Actuat B 1998; 47:92e7. [5] LeGore LJ, Lad RJ, Moulzolf SC, Vetelino JF, Frederick BG, Kenik EA. Thin Solid Films 2002;406:79e86. [6] Chu X, Jiang D, Guo Y, Zheng ZC. Sens Actuat 2006;120:177e81. [7] Lin MS, Leu HJ. Electroanalysis 2005;17:2068e73. [8] Rothschild A, Edelman F, Komem Y, Cosandey F. Sens Actuat B 2000;67: 282e9.
S. Luby et al. / Vacuum 86 (2012) 590e593 [9] Baraton MI, Marhari L. J Nanopart Res 2004;6:107e17. [10] Vaseahsta A, Vaclavikova M, Vaseashta S, Gallios G, Roy P, Pummakarnchana O. Sci Technol Adv Mater 2007;8:47e59. [11] Fainberg A. Science 1992;255:1531e7. [12] Dorozhkin LM, Nefedov VA, Sabelnikov AG, Sevastjanov VG. Sens Actuat B 2004;99:568e70. [13] Yinon J. Anal Chem; 2003:99Ae105A.
593
[14] Riu J, Maroto A, Rius FX. Talanta 2006;69:288e301. [15] Chitu L, Jergel M, Majkova E, Luby S, Capek I, Satka A, et al. Mater Sci Eng 2007;27:1415e7. [16] Siffalovic P, Chitu L, Majkova E, Vegso K, Jergel M, Luby S, et al. Langmuir 2010;26:5451e5. [17] Kiyomura T, Maruo Y, Gomi M. J Appl Phys 2000;88:4768e71. [18] Göpel W. Sens Actuat B 1994;18e19:1e21.
Contents lists available at ScienceDirect
Vacuum journal homepage: www.elsevier.com/locate/vacuum
Oxide nanoparticle arrays for sensors of CO and NO2 gases S. Luby a, *, L. Chitu a, M. Jergel a, E. Majkova a, P. Siffalovic a, A.P. Caricato b, A. Luches b, M. Martino b, R. Rella c, M.G. Manera c a
Institute of Physics, Slovak Academy of Sciences, Dubravska cesta 9, 845 11 Bratislava, Slovakia Department of Physics, University of Salento, Via Arnesano, 73100 Lecce, Italy c Institute of Microelectronics and Microsystems, IMM-CNR, Unit of Lecce, 73100 Lecce, Italy b
a b s t r a c t Keywords: Nanoparticles LangmuireBlodgett layers Gas sensing
Metal oxide sensors with active films from Fe2O3 and CoFe2O4 hexagonally assembled nanoparticle (NP) arrays were studied. NPs were synthesized by high-temperature solution phase reaction. Sensing NP layers were deposited by LangmuireBlodgett (LB) technique. LB layers were characterized by XRD, SEM and magnetic measurements. NPs are monodomain and superparamagnetic at RT. Sensor active films are formed from 1 or 7 LB monolayers deposited on the alumina substrates equipped with heating meander and interdigitated contacts. LB layers were heat-treated or UV irradiated to remove the insulating surfactant. Sensing properties were studied in a test chamber containing reducing (CO) or oxidizing (NO2) gas in concentrations between 5 and 100 ppm in the mixture with dry air. Response current signal (Igas/Iair) vs. temperature and response vs. gas concentration calibration curves were measured. Best response values were obtained with CoFe2O4 devices between 300 and 400 C, being 3 and 10 for 100 ppm of CO and 5 ppm of NO2, respectively. The response and recovery times of sensors are between 3 and 30 min. The Fe2O3 and CoFe2O4 sensors with response of 8 or 10 to 5 ppm of NO2 might be of practical applications in the detection of explosives. Ó 2011 Elsevier Ltd. All rights reserved.
1. Introduction Bardeen and Brattain discovered that metal oxide semiconductor resistance could be modified by ambient gas [1]. Seiyama et al. proposed to use these materials in gas sensors [2]. Since that time many devices were developed. From the multitude of oxides for sensing of gases we mention at least SnO2 [3,4], WO3 [3,5], CoFe2O4 [6], Fe2O3, Fe3O4, In2O3 [7] and TiO2 [8]. Among the monitored gases attention is paid mainly to air pollutants like SO2, CO, CO2, CH4, NO, NO2, O2, O3, Cl2, C2H5OH, H2O2 [9,10]. Moreover, NO2 is of primary importance for trace amount detection of explosives like EGDN, TNT, PETN, RDX, etc. [11,12]. NO2 sensors must be extremely sensitive because some explosives have very low vapor pressure. At RT and atmospheric pressure of air it decreases from 50 ppm for EGDN to 30 ppb for RDX and PETN, the pressure of TNT and nitroglycerin being 1 ppm and 1 ppb, respectively [11]. In the detection of explosives a variety of techniques using fluorescent polymers, conducting polymer composites, surface acoustic waves, etc., are used [13]. Gas sensors and electronic noses attract a lot of interest because the use of dogs is expensive [11,13]. * Corresponding author. Tel.: þ421 903 713 696; fax: þ421 2 5477 6085. E-mail address: [email protected] (S. Luby). 0042-207X/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.vacuum.2011.07.018
Innovative sensors are built from nanoparticle (NP) arrays [9,14]. In comparison with continuous films the devices have higher surface/volume ratio. Consequently, the surface scattering is better influenced by adsorbed species and the change of sensor conductivity is increased. In this work we experiment with sensors from Fe2O3 and CoFe2O4 NPs fabricated in the colloid solution by a chemical route. They are non-toxic, moreover, the size dispersion of NPs is low (10%). Therefore, the conductivity of the array could be quite simply computed. 2. Experimental Monodisperse Fe2O3 (maghemite) and CoFe2O4 NPs were synthesized by high-temperature solution phase reaction of metal acetylacetonates (Fe(acac)3, Co(acac)2) with 1,2-hexadecanediol, oleic acid and oleyl amine in phenyl ether at 200 C/30 min plus 265 C/30 min heating [15]. Solvent was toluene. Structure of NPs was studied by grazing incidence (GI) XRD-D8 Bruker. Here NP solution was dropped onto Si substrate and dried. The thickness of the deposit was 10e20 NP monolayers (MLs). Size of NPs was estimated using HR SEM Leo 1550. Magnetic properties were studied between 4.2 and 278 K using vibrating sample magnetometer and diluted NP solution in a capillary.
S. Luby et al. / Vacuum 86 (2012) 590e593
591
Before measurements samples were held 12 h in dry air flux at 400 C to remove the surfactant coverage. Then NP arrays agglomerate. The 1 ML deposit transforms into labyrinth-like structure with certain level of percolation and 7 ML deposit aggregates into a compact structure. Alternatively, some samples were treated by UV radiation during 5 min at the flux 2 mW/cm2 in the ambient of ozone in the dry air (Fig. 1b). Here the transformed structures are less compact and they preserve NP nature [16]. Measurements were performed at different operating temperatures using the heating element. The sensor response to reducing (CO) or oxidizing (NO2) gas was monitored by an electrometer at a constant voltage of 5e10 V between electrodes. The desired concentrations of the gases in dry air were obtained using certified gas bottles and mass flow meters. The temperature of the bottles was about 22 C for both air and air containing CO and NO2. Flow of 100 sccm was fixed during the tests. At first conditioning of sensor in the test chamber in a flow of dry air was performed until stabilization of the sensor response signal. After it a mixture containing CO or NO2 was sent into a chamber. 3. Results and discussion
Fig. 1. Fe2O3 NP array before (a) and after (b) UV irradiation in the ambient of ozone in the dry air. The horizontal line corresponds to 100 nm.
NP LangmuireBlodgett (LB) films were prepared from the water subphase in the 612 D NIMA trough at the surface pressure of 20 mN/m. 1 or 7 ordered LB MLs were deposited by modified (patent pending) procedure. Gas sensors were fabricated on 3 mm 3 mm alumina substrates. They were equipped with 20 nm Ti/200 nm Pt interdigitated electrodes on the front side to read the measuring current and with 20 nm Ti/500 nm Pt heating meander on the back side. The temperature of sensors was obtained from poweretemperature curve. Calibration was done in an oven using a thermocouple and measuring the resistance of the heating element in dynamic conditions under the gas flow. After deposition of the sensing layers the devices were soldered onto a commercial TO-8 socket and they were fixed in a test chamber.
From GI XRD spectra it follows that the diffraction peaks of FeeO system correspond to either Fe3O4 (magnetite, cubic, fcc) or Fe2O3 (maghemite, cubic, primitive cell). From the peak 400 at larger 2q angles Fe2O3 composition is preferred. (Moreover, under the ageing/heating in the air ambient or UV irradiation Fe3O4 is oxidized to Fe2O3 [17].) In combination with Co, ternary compound CoFe2O4 is formed. It is isomorph with Fe3O4. The diameter of Fe2O3/Fe3O4 and CoFe2O4 NPs is 6.4 0.6 and 7.6 0.6, respectively, surfactant is 1.8 nm thick in both cases and blocking temperatures are 22 K and 204 K, respectively. NPs are monodomain and superparamagnetic at RT [15]. An example of hexagonal-like NP arrays before and after UV processing is shown in Fig. 1a,b, respectively. In Table 1 sensor characteristics, i.e. structure of the devices and working conditions for two monitored gases are summarized. The interval of working temperatures is 150e475 C. These temperatures are typical for many metal oxide semiconductor sensors [3,4]. The maximum in the sensor response vs. working temperature for CoFe2O4 with 7 MLs is shown in Fig. 2. Here any value of the response was obtained having the sensor at a specific working temperature in the test chamber. Data for other devices are in Table 1. The maxima of R vs. Tw are the consequence of reaching the best dynamic equilibrium between adsorption and desorption of the monitored gas [4,18]. In Table 1 measuring conditions and sensor responses are shown as well. Both Igas and Iair were obtained in the stabilized conditions described above. The base current in air is of the order 109e1011 A with 1 ML deposits, which corresponds
Table 1 Structure and operating conditions of sensors with 1 or 7 monolayers (MLs) of NPs (first column). Cg is the gas concentration, Tw is the interval of working temperatures and Tm e maximum response working temperature measured at the concentration Cm. Sensor base current in dry air Iair was obtained at Tm or in the middle of Tw interval. The sensor response is R ¼ Igas/Iair. NPs/No. of MLs
Gas
Cg [ppm]
Tw [ C]
Tm/Cm [ C]/[ppm]
Iair [A]
R
Fe2O3/1 Fe2O3/1 Fe2O3/7 Fe2O3/7 CoFe2O4/1 CoFe2O4/1 CoFe2O4/7 CoFe2O4/7 CoFe2O4/7 UV CoFe2O4/7 UV
CO NO2 CO NO2 CO NO2 CO NO2 CO NO2
20e100 5 100 5 20e100 100 2e100 100 100 5
325e475 375 200e475 300e475 175e450 350 300e475 375 250e475 300e475
450/100 Not evaluated 350/100 300/5 340/100 Not evaluated 425/100 Not evaluated 425/100 375/5
3 109 2.4 109 6 107 8 107 2 1011 4 1011 6 107 3 107 7 108 8 108
1.3 (400 C, 100 ppm) (1.8)1 (375 C, 5 ppm) 2.8 (350 C, 100 ppm) (8)1 (350 C, 5 ppm) 1.9 (350 C, 100 ppm) y1 1.6 (400 C, 100 ppm) y1 3 (350 C, 10 ppm) (10)1 (350 C, 5 ppm)
592
S. Luby et al. / Vacuum 86 (2012) 590e593
1.6
1.5
Response (Igas/Iair)
Response (Igas/I0)
1.3
1.4
1.3
1.2
1.1
1.0
1.2 300
350
400
0
450
20
40
60
80
100
CO concentration (ppm)
Working Temperature (°C) Fig. 2. Response of CoFe2O4 sensor with 7 MLs of NPs vs. working temperature at 100 ppm of CO gas.
to not fully continuous labyrinth-like structure. Therefore, we created 7 MLs deposits. The increase of the base current by 2e4 orders of magnitude is not a simple consequence of thicker conducting film but predominantly of the formation of continuous structure. This can be justified using Fig. 1b, where approximately 37% of the area is striped of NPs. Then the average thickness of the layer is 63% of the non-irradiated one and the thickness of 7 monolayers is 7 6.4 nm ¼ 4.5 nm, or even less, assuming the hexagonal closed packing in the vertical direction. (From calculation using the diameter of NPs and thickness of surfactant we obtain even smaller relative coverage of the surface by irradiated NP monolayer, being 41%. This, however, corresponds to densely packed irradiated NPs, which obviously is not the case.) The increase of the base current to 107e108 A (Table 1) might on one side screen the sensor response, but on the other side these values are important for easy electronic integration solution. The response at the chemisorption of oxidizing NO2 gas is shown as (X)1, because in this case capturing of electrons from sensing layers in the n-type semiconductor diminishes the conductivity. With reducing CO gas we have opposite effect [3,4]. From the comparison of the heat-treated and UV irradiated CoFe2O4 sensors with 7 MLs of NPs we can see, that in the second case we have much higher responses for both monitored gases (Table 1). This is attributed to the preservation of NP structure which is important for increasing the adsorption/absorption capacity of the sensor. In Fig. 3 are the time dependences of current in heat-treated Fe2O3 and UV irradiated CoFe2O4 sensors in oxidizing NO2 ambient. In both cases the response is good (Table 1) and devices are of some practical use.
Fig. 4. Calibration curve of CO sensor with 1 ML of Fe2O3 sensing film at a working temperature 400 C.
The published response and recovery times of metal oxide sensors are between tens of seconds and tens of minutes, depending especially on the structure and thickness of the sensing film. Short times are observed for thin nanocrystalline films where electron processes take place at the surface [3,4,7,10]. Our results are similar, the times are between 3 and 30 min, being lower for 1 ML sensing films. The recovery of sensors is satisfactory (Fig. 3). This is in accord with results of [7], where almost 100% recovery for FeeO system was reported. The calibration curves of sensors, i.e. response vs. gas concentration, are linear for both 1 ML (e.g. Fig. 4) and 7 ML sensing films at Tw approaching to 400 C. At temperatures around 300 C 1 ML sensors sometimes show saturation at CO concentration of 100 ppm. 4. Conclusions We have shown that gas sensors with sensing films composed from 1 or 7 monolayers of Fe2O3 or CoFe2O4 NPs are satisfactory for monitoring reducing CO or oxidizing NO2 gases. With 1 ML films the base current is low because after removing the surfactant the sensing film becomes partly discontinuous. On the other hand films from 7 LangmuireBlodgett monolayers are probably too thick and therefore they have higher response and recovery times. Improvement of devices is expected with sensing films composed from 2 to 4 monolayers of nanoparticles. The Fe2O3 and CoFe2O4 sensors with response of 8 or 10 to 5 ppm of NO2 might be of some practical applications in detection of explosives. Acknowledgment The work was supported by the following grants: CANR, ITMS code 26240220011 ERDF, APVV SK-IT-0027-08 and by the Scientific Grant Agency VEGA, 2/0126/09. The support of EU HASYLAB grant II-20080036 EC is also acknowledged.
Current (A)
10
10
References 10
10
10
0
1
2
3
4
5
6
7
8
time (h) Fig. 3. Dynamic responses to 5 ppm of NO2 at 350 C with Fe2O3, 7 ML heat-treated sensor (full line) and CoFe2O4, 7 ML UV irradiated sensor (broken line).
[1] Bardeen J, Brattain WH. Bell Syst Tech J 1953;32:1. [2] Seiyama T, Kato A, Fujishi K, Nagatani M. Anal Chem 1962;34:1502e7. [3] Heusler TP, Lorke A, Ifeacho P, Wiggers H, Schulz C. J Appl Phys 2007;102: 124305. [4] Gurlo A, Barsan N, Ivanovskaya M, Weimar U, Göpel W. Sens Actuat B 1998; 47:92e7. [5] LeGore LJ, Lad RJ, Moulzolf SC, Vetelino JF, Frederick BG, Kenik EA. Thin Solid Films 2002;406:79e86. [6] Chu X, Jiang D, Guo Y, Zheng ZC. Sens Actuat 2006;120:177e81. [7] Lin MS, Leu HJ. Electroanalysis 2005;17:2068e73. [8] Rothschild A, Edelman F, Komem Y, Cosandey F. Sens Actuat B 2000;67: 282e9.
S. Luby et al. / Vacuum 86 (2012) 590e593 [9] Baraton MI, Marhari L. J Nanopart Res 2004;6:107e17. [10] Vaseahsta A, Vaclavikova M, Vaseashta S, Gallios G, Roy P, Pummakarnchana O. Sci Technol Adv Mater 2007;8:47e59. [11] Fainberg A. Science 1992;255:1531e7. [12] Dorozhkin LM, Nefedov VA, Sabelnikov AG, Sevastjanov VG. Sens Actuat B 2004;99:568e70. [13] Yinon J. Anal Chem; 2003:99Ae105A.
593
[14] Riu J, Maroto A, Rius FX. Talanta 2006;69:288e301. [15] Chitu L, Jergel M, Majkova E, Luby S, Capek I, Satka A, et al. Mater Sci Eng 2007;27:1415e7. [16] Siffalovic P, Chitu L, Majkova E, Vegso K, Jergel M, Luby S, et al. Langmuir 2010;26:5451e5. [17] Kiyomura T, Maruo Y, Gomi M. J Appl Phys 2000;88:4768e71. [18] Göpel W. Sens Actuat B 1994;18e19:1e21.