Catalytic properties of oxide nanoparticles applied in ... - Springer Link

2 downloads 0 Views 1MB Size Report
Doina Lutica, Michael Stranda, Anita Lloyd-Spetzb, Kristina Buchholtb, Eliana Ievac, Per-Olof Ka¨llc, and Mehri Sanatia,*. aDepartment of Chemistry, Va¨xjo¨ ...
105

Topics in Catalysis Vol. 45, Nos. 1–4, August 2007 ( 2007) DOI: 10.1007/s11244-007-0248-1

Catalytic properties of oxide nanoparticles applied in gas sensors Doina Lutica, Michael Stranda, Anita Lloyd-Spetzb, Kristina Buchholtb, Eliana Ievac, Per-Olof Ka¨llc, and Mehri Sanatia,* a Department of Chemistry, Va¨xjo¨ University, SE-351 95 Vaxjo, Sweden S-SENCE and Divison of Applied Physics, Linko¨ping University, SE-581 83 Linkoping, Sweden c Division of Physical and Inorganic Chemistry, Linko¨ping University, SE-581 83 Linkoping, Sweden b

A series of gas sensing layers based on indium oxide doped with gold were prepared by using the aerosol technology for deposition as the active contact layer in a metal oxide semiconductor capacitive device. The interaction between the measured species and the insulator surface was quantified as the voltage changes at a constant capacitance of the device. The sensor properties were investigated in the presence of H2, CO, NH3, NO, NO2 and C3H6 at temperatures between 100–400 C. Significant differences in the morphology of the layer and its sensitivity were noted for different preparation methods and different gas environments. KEY WORDS: indium oxide; nanoparticles; gold; catalytic effect; capacitive sensors; MIS gas sensors; gas detection; NOx; CO; NH3; C3H6..

1. Introduction Gas monitoring is an important task in various household and industrial applications, especially in what concerns people and environment security and determination of product quality. For most of the chemical sensors for gases, the detection mechanism occurs by adsorption, chemi- or physisorption and eventually followed by chemical reactions, between the solid sensor surface and the species to be detected, resulting in the modification of the charge carrier density. The magnitude of the modifications is a measure of the number of the sensed gas molecules, provided that the interaction mechanism is known in detail. Nanosized oxides are a convenient class of materials for gas sensors, due to the diversity of their chemical composition, the crystalline or amorphous character, the porosity, size and shape of the particles. The peculiar properties of the layer influence to a high extent the interaction between the surface and the measured gas. The use of nanoparticles assures a high accessible surface for the gas molecules, allowing the use of very small amounts of substance in the sensitive layer and thus the miniaturizing of the devices. In2O3 is explored as a sensor material based on the resistivity changes to different gases by several authors [1–6]. In this work we are communicating a series of results obtained using indium oxide-based catalytic layers doped with Au, as capacitive-type devices for measuring gases such as hydrogen, carbon monoxide, hydrocarbons, nitrogen oxides, ammonia and mixtures thereof, in oxygen-containing atmosphere. According to our knowledge, there are no papers dealing with this * To whom correspondence should be addressed. E-mail: [email protected]

particular kind of measuring devices and sensing layers in the literature.

2. Experimental The catalytic gate material, indium oxide + gold, playing the role of the contact to the dielectric in a capacitor, was produced by aerosol technique and deposited on the substrate consisting of three layers: an insulator (based on silicon dioxide and silicon nitride densified to yield a top layer of SiOx (5 nm)), a semiconductor (silicon carbide n-doped) and a metal stack (Ni/Ti/Pt) playing the role of a back side ohmic contact. On the front side of the substrate there is a bonding pad of noble metal (Pt and Ti (an adhesion layer)), also assuring a good electrical contact to the catalytic nanoparticle layer within the electrical circuit. These kinds of substrates will be referred to as insulator/SiC. Details about this kind of devices can be found in one of our previous works [7]. The indium oxide nanoparticles were obtained by thermal decomposition of indium nitrate (Fluka) in air flow in a tubular horizontal oven acting as a flow reactor. The salt was delivered to the reaction zone as a fine sprayed solution (0.01–0.2 mole L)1) in ultra pure water, using an atomizer (AGK 2000, Palas GmBH). The particles were then deposited by inertial impact technique on the cleaned substrates (oxidation with hydrogen peroxide 30% in the presence of ammonia, at 85 C) under low pressure. A very similar system was described in detail in [8]. The gold nanoparticles were either obtained by the same aerosol technique, using tetrachloroauric acid (Fluka) (0.3 mole L)1), or by electrochemical synthesis.

1022-5528/07/0800-0105/0  2007 Springer Science+Business Media, LLC

106

D. Lutic et al./Catalytic properties of oxide nanoparticles in gas sensors

For the electrochemically method a three-electrode cell is used with a reference electrode, an anode electrode consisting of a thin Au sheet, and a cathode electrode of Pt [9]. The characteristics of the aerosol flow were analysed by measuring the particle size distribution and mass flow of the solid using an electric mobility spectrometer (SMPS 3080, TSI Inc.) incorporating a Differential Mobility Analyser (DMA 3081) and a Condensation Particle Counter (CPC 3010), after a 125 fold dilution of the aerosol flow. The particle morphology was investigated by a LEO 1550 VP field emission scanning electron microscopy (SEM). The capacitance measurements were performed on MISiC (metal insulator silicon carbide) capacitive samples using a Boonton 7200 bridge as a multi capacitance meter, MCM, (time shared measurements of up to 15 capacitors), at constant temperature values comprised between 100 and 400 C. The sensor chip and Pt 100 device for temperature control were glued onto a heater and mounted on 16 pin capsules (figure 1(a)). The capsules were introduced in aluminium cells (figure 1(b)) sealed by Teflon joints, within which controlled gas flows were introduced. The gas flows and compositions were controlled by a series of valves connected to a computer and use of specific software. A first series of measurements were performed at variable voltage, between )5 ‚ 10 V and concerned measurement of the capacitance values in the dynamic regime; these measurements are called C-V curves. A normal test cycle consisted in measurements of the CV characteristics performed at 200, 300, 400, 100 and then again 200 C, in different gas environments and synthetic air, between which purges with air for 5 min were effected. Several cycles were performed for each sample series in the same conditions, in order to investigate their stability, i.e. the reproducibility of the C-V data. A subsequent series of measurements were performed at constant capacitance, at the value where the variable voltage measurements reached the maximum change due to different gas ambients. The voltage changes due to different concentrations of the measured gases were

measured in air containing: H2: 250, 500 and 1000 ppm, CO: 100, 200 and 500 ppm, NH3: 250, 500 and 1000 ppm, NO: 50, 100 and 250 ppm, NO2: 50, 100 and 250 ppm and NO + NH3: 100 + 250 ppm, 200 + 500 and 250 + 625 ppm, respectively. All the gases were purchased from Air Liquide and have high purity.

3. Results and discussion 3.1. Characterization of the particles A narrow size distribution of solid particles produced by the aerosol technique is desirable in order to have an uniform sensing layer. Several parameters were changed in order to obtain a stable and uniform layer on the substrate: concentration of the precursor solution salt, pressure of the air used for spraying the solution, diluting air flow, oven temperature, reaction zone length in the oven, impact zone pressure and peculiar design of the impact region. An optimum was reached in the following conditions: 0.01 mole L)1 indium nitrate, air of 3 bar for spraying (equivalent to about 60 mL h)1 solution sprayed), 1 L min)1 diluting air, 800 C, 50 cm and 5 mbar respectively. The particle number size distribution in this series of conditions is displayed in figure 2. A series of Au-In2O3 samples were prepared on same type of substrates. The Au particles were deposited using two different methods on the surface, see table 1. The SEM images indicate a low degree of packing of almost spherical and quite uniform indium oxide particles (figure 3(a)), indicating the good accessibility to the grains for the gas. The gold deposited by aerosol technique on the indium oxide layer displays much bigger balls stacked on the indium oxide, preserving a good porosity of the layer (figure 3(b)). 3.2. Electrical measurements Our interest was to obtain sensitive sensors towards chemical species currently present in various car exhaust gases and flue gases (carbon monoxide, hydrocarbons

Figure 1. (a) Sensor chip on 16-pin capsule; (b) cells for gas sensitivity tests.

D. Lutic et al./Catalytic properties of oxide nanoparticles in gas sensors

107

Au elch/In2O3/insulator/SiC

The OH groups on the insulator exhibits large dipoles and thus forms a strongly polarized layer, which will introduce an electric field in the insulator and, for a positively biased n-type semiconductor, the depletion area will decrease; the OH groups acts as an extra applied positive voltage. Thus the polarization of the insulator surface causes a voltage shift of the capacitance versus voltage curve, proportional to the concentration of hydrogen. For the sensing of other gases, e.g. ammonia, the presence of triple phase boundaries, where gas, metal and insulator are all in contact, are necessary in order to obtain dissociation of the molecule in such a way that hydrogen atoms are released to the insulator to form the polarized layer in the same way as for hydrogen [12–17]. A voltage shift appears between the C-V curves in the presence of air (reference gas) and some ppm of test gas in air. The C-V curve is the starting point to determine the capacitance value where the magnitude of the modifications on the surface is maximal, i.e. the sensitivity of the layer is highest. Some examples of C-V curves are presented in figure 4.

and nitrogen oxides). The modifications occurring in the electron density in the semiconductor are the result of the electronic interactions between the gas and the sensing layer during the adsorption or chemical reaction. In the case of the very thin metallic layers involved in field effect type devices like the MIS capacitors used here, the catalytic interaction occurs between the metal and the gas molecule. The hydrogen containing molecules dissociate on the catalytic metal sites, hydrogen atoms spill over to the insulator and adsorbs predominantly as OH groups. There is evidence of these phenomena after investigation by DRIFT spectroscopy using model surfaces of Pt or Ir impregnated SiO2 powder. Clear evidences for OH groups and even eventually OH2+ groups were found on the SiO2 sites during hydrogen and ammonia exposure in air [10,11].

Figure 3. (a) SEM images of the sensitive surface of the indium oxide layer (200k x); (b) Au deposited by aerosol technique on indium oxide layer (50k x).

-3

dN/d lg dp [cm ]

8.0E+04

6.0E+04

4.0E+04

2.0E+04

0.0E+00 10

100

1000

Mobility Diameter dp [nm]

Figure 2. Particle size distribution for indium oxide (precursor: indium nitrate 0.01 mole L)1; temperature: 800 C; atomizer pressure: 3 bar; diluting air: 1 L min)1; reaction zone length: 45 cm; dilution degree of the aerosol; 125).

Table 1 Au-In2O3 samples used as sensing layers Sample

Details concerning preparation

Indium oxide layer deposited on freshly cleaned substrate, then Au nanoparticle suspension deposited as a droplet [7] Au aerosol/In2O3/insulator/SiC Indium oxide layer deposited on freshly cleaned substrate, then Au nanoparticles obtained by aerosol technique suspension deposited as a droplet Au elch/In2O3/Au Au nanoparticle suspension elch/insulator/SiC deposited as a droplet on freshly cleaned substrate, then indium oxide deposited by aerosol technique, then again Au nanoparticle suspension deposited as a droplet [7] (Au + In2O3) Nanoparticles of indium oxide and aerosol/insulator/SiC gold obtained by aerosol technique, by spraying a solution containing both precursors

108

D. Lutic et al./Catalytic properties of oxide nanoparticles in gas sensors

As observed from the C-V curves, the preparation technique has an essential effect on the capacitance value of the samples. Moreover, the capacitance differs to a big extent according to the nature of the detected gas. The samples obtained by simultaneous spraying of Au and In have very low capacitance values (less than 100 pF), making them unuseful as sensor devices. The deposition of gold by aerosol technique gives also quite low capacitance values in comparison with the other samples, obtained by droplet deposition of Au. The stability of Au deposited by droplet on indium oxide is however very low; the capacitance increases dramatically for each cycle, indicating a poor stability of the layer. From the point of view of the catalytic layer, these heating cycles seem to produce a reorganisation of the structure of the layer, which begins to accumulate charges in a manner not so easy to explain from this type of investigations only. Most likely gold particles, which have a higher conductivity as compared to the In2O3 particles, restructures into a network with contact paths between the gold grains. The constant capacitance measurements gave important indications about the catalytic potential of the layer with respect to the gas species to be detected. The expected changes of the capacitance values were effective only at high testing temperatures (350 and 400C), indicating a lower catalytic activity of the Au/ In2O3 sensing layer as compared to for example porous layers of Pt or Ir [13,14] and nanoparticles of Ru or RuO2 [18]. Apparently, rather high activation energy is necessary to catalyse the reactions involved in the sensing mechanism. Our experiments clearly show that the catalytic system Au/In2O3 reacts with the reducing species, i.e. hydrogen, carbon monoxide and propene, and that there is also a response, though in the opposite direction to NO (normally denoted NOx, since some

amount of NO2 will always be present in air), see figure 5. It should be noted that the direction of the response to reducing and oxidizing species also here follows the normal behaviour for MIS / FET (field effect transistor) devices [12–15] but due to experimental circumstances the polarity of the sensor signal is switched. Interesting results in figure 5 is the increased response to CO, which is quite poor for comparable devices in Refs. [14,15]. The response to NO at a temperature of 400 C is even more encouraging. Probably the higher response in figure 5 of the sensing layer with the electrochemically synthesized gold nanoparticles is due to the higher Au content [19], also reflected in the higher accumulation capacitance of the CV-curve in figure 4 of this sample. Sensors for detection of NOx are needed for example to control SCR [14], selective catalytic reduction, in diesel exhausts. Here an operation temperature > 300 C is required. The metal oxide sensor systems, which measure resistivity changes during gas exposure normally detects NOx only at temperatures around or below 200 C [1–6]. Thus through the use of a less catalytic active material such as In2O3, the operation temperature is increased. Another interesting thing to note in the results in figure 5 is the response to the combination of NO/NH3, which seems to be simply the addition of the single responses indicating an independent detection mechanism for NH3 and NOx, respectively. While the detection mechanism for NH3 is quite well understood [10,11,13] the mechanism for detection of NOx is still under investigation. Finally it can be noticed that for the structure of the sample containing the indium oxide layer between two Au droplet deposited layers the reproducibility of the signal after a number of test cycles was very poor, indicating a kind of capacitor properties breakdown, for example related to a leakage current in the insulator of the device.

1400 Au elch./In2O3/insulator/SiC Au aerosol/In2O3 aerosol/insulator/SiC 1200 Au aerosol/In2O3 aerosol/insulator/SiC (2)

C a p a c it a n c e , p F

Au elch/In2O3/Au elch/insulator/SiC 1000

800

600

400

200

0

-5

-4

-3

-2

-1

0

1

2

3

4

5

6

7

8

9

Voltage, V Figure 4. C-V curves for the series of samples in air containing100 ppm NO.

10

109

D. Lutic et al./Catalytic properties of oxide nanoparticles in gas sensors H2

_ CO

_

NH3

_

NO _ NO+NH3 _ C3H6

0 0

20

40

60

80

100

120

140

160

180

-0,5

Voltage shift, V

-1

-1,5

-2

-2,5

-3

Au elch./In2O3/insulator/SiC (up) Au aerosol/In2O3 aerosol/insulator/SiC (middle)

-3,5

Au aerosol/In2O3 aerosol/insulator/SiC (2) (down)

-4

Time, min Figure 5. Constant capacitance measurements for a series of gases at 400 C.

4. Conclusions The use of aerosol technology is a convenient way to obtain nanoparticles that could be deposited as sensitive layers in gas sensors for sensing species such as carbon monoxide, hydrogen, hydrocarbons, ammonia and NOx. In our work we obtained some promising results after having tested indium oxide doped with gold as the catalytic sensing layer involved in the fabrication of a capacitive type gas sensor. The preparation details have a big impact on the results concerning the stability of the samples, due to the requirements of a conducting sensing layer in the capacitive sensors. Different ratios of In2O3 to Au particles display interesting differences in sensor properties. The temperature range for a high gas response is between 350 and 400 C. We found devices with a high response to NOx at 400 C and other devices with an increased response to CO.

Acknowledgments D.L. acknowledges the financial support as a postdoctoral researcher by the Environmental Foundation of The Swedish Association of Graduate Engineers (Civilingenjo¨rsfo¨rbundet). A grant from the Swedish Research Council is also acknowledged.

References [1] J. Mizsei, Sensors and Actuators B 23 (1995) 173. [2] H. Yamaura, T. Jinkawa, J. Tamaki, K. Moriya, N. Miura and N. Yamazoe, Sensors and Actuators B 35–36 (1996) 325. [3] G. Korotcenkov, V. Brinzari, A. Cerneavschi, M. Ivanov, A. Cornet, J. Morante, A. Cabot and J. Arbiol, Sensors and Actuators B 98 (2004) 122.

[4] V. Golovanonv, M.A. Ma¨ski-Jaskari, T.T. Rantala, G. Korotcenkov, V. Brinzari, A. Cornet and J. Morante, Sensors and Actuators B 106 (2005) 563. [5] A. Gurlo, N. Barsan, M. Ivanovskaya, U. Weimar and W. Go¨pel, Sensors and Actuators B 47 (1998) 92. [6] Z. Zhan, D. Jiang and J. Xu, Materials Chemistry and Physics 90 (2005) 250. [7] A. Salomonsson, S. Roy, C. Aulin, J. Cerda, P.-O. Ka¨ll, L. Ojama¨e, M. Strand, M. Sanati and A. Lloyd-Spetz, Sensors and Actuators B 107(2) (2005) 831. [8] M. Strand, A. Salomonsson, J. Einvall, C. Aulin, L. Ojama¨e, P-O. Ka¨ll, A. Lloyd Spetz and M. Sanati, Proceedings of ‘‘European Aerosol Conference 2005’’, Editor M. Maenhaut, Ghent, Belgium, 28 Aug–2 Sep, pp. 735. [9] M.T. Reetz and W. Helbig, J. Am. Chem. Soc. 116 (1994) 7401. [10] M. Wallin, H. Gro¨nbeck, A. Lloyd Spetz and M. Skoglundh, Applied Surface Science 235(4) (2004) 487. [11] M. Wallin, H. Gro¨nbeck, A. Lloyd Spetz, M. Eriksson and M. Skoglundh, J. Phys. Chem. B 109 (2005) 9581. [12] L.-G. Ekedahl, M. Eriksson and I. Lundstro¨m, Acc. Chem. Res. 31(5) (1998) 249. [13] M. Lo¨fdahl, C. Utaiwasin, A. Carlsson, I. Lundstro¨m and M. Eriksson, Sensors and Actuators B 80 (2001) 183. [14] H. Wingbrant, H. Svenningstorp, P. Salomonsson, D. Kubinski, J. Visser, M. Lo¨fdahl and A. Lloyd Spetz, IEEE Sensors Journal 5(5) (2005) 1099. [15] M. Andersson, P. Ljung, M. Mattsson, M. Lo¨fdahl and A. Lloyd Spetz, Topics in Catalysis 30–31 (2004) 365. [16] M. Eriksson, A. Salomonsson and I. Lundstro¨m, J. Appl. Physics 98 (2005) 034903–1. [17] A Lloyd Spetz, H. Wingbrant, M. Andersson, A. Salomonsson, S. Roy, G. Wingqvist, I. Katardjiev, M. Eickhoff and S. Nakagomi, Proc. International Symposium on Advanced Materials and Processing (ISAMAP2K4), Kharagpur, India, 6–8 December, (2004) 755. [18] A. Salomonsson, R.M. Petoral Jr., K. Uvdal, C. Aulin, P.-O. Ka¨ll, L. Ojama¨e, M. Strand, M. Sanati and A. Lloyd Spetz, J. Nanoparticle Research, DOI 10.1007/s11051-005-9058-1 (2006). [19] C.K. Buchholt, E. Ieva, B. Johansson, P. Ka¨ll, L. Torsi and A. Lloyd Spetz, Proc. IMCS11, Brescia, Italy, July 17–19 (2006).