首页   按字顺浏览 期刊浏览 卷期浏览 Nitrogen dioxide and volatile sulfide sensing properties of copper, zinc and nickel chr...
Nitrogen dioxide and volatile sulfide sensing properties of copper, zinc and nickel chromite

 

作者: Colin L. Honeybourne,  

 

期刊: Journal of Materials Chemistry  (RSC Available online 1996)
卷期: Volume 6, issue 3  

页码: 277-283

 

ISSN:0959-9428

 

年代: 1996

 

DOI:10.1039/JM9960600277

 

出版商: RSC

 

数据来源: RSC

 

摘要:

Nitrogen dioxide and volatile sulfide sensing properties of copper, zinc and nickel chromite? Colin L. Honeybourne and Raymond K. Rasheed Faculty of Applied Sciences, University of the West of England, Bristol, UK BSl6 1QY Polycrystalline powders of CuCr20,, ZnCr204 and NiCr204 were prepared from the decomposition of metal nitrates with glycine, and their structures confirmed by X-ray diffraction. The three spinels were mixed with ethyl cellulose dissolved in a suitable solvent (80 :20 toluene-ethanol) to form pastes which were applied to substrates and dried. The gas sensing properties of the resulting thick films were investigated with ppm levels of NO2, CH3SH and C,H5SH at 220 "C. It was found that the spinels can be used to detect the presence of these gases.Further experiments showed that by substituting some of the A or B cations with Li and Mg, respectively, the gas sensing properties of the spinels can be altered significantly. The monitoring of volatile sulfides and nitrogen dioxide is increasing in importance in industrial and domestic areas, because even in dilute concentrations they can be harmful, and in the case of sulfides their smell can be offensive. The majority of metal oxide sensors for volatile sulfides and nitrogen dioxide B (0.04 moles) were dissolved in 100ml of distilled water. Using valency values of -2, + 1 and 0 for 0, H and N, respectively, the oxidising valency for one mole of each metal nitrate was calculated. These values were then corrected for each nitrate to correspond to the number of moles used in the are based on tin dioxide containing additives or d0~ants.l~~ reaction and then summed to give a The aim of the research described here is to demonstrate the potential of chromite spinels as sensitive gas sensors.To this end it was decided to focus on metal oxides containing transition-metal ions, and target gases that may bring about a change in their oxidation state. The resulting changes in oxidation state would then alter the conductivity of the oxide, and give a measurable response to the target gas. The oxides chosen for this study were CuCr,O,, ZnCr,O, and NiCr20,. These oxides belong to a class of inorganic compounds known as spinels which have the general formula AB20,., The oxide ions form a close-packed array and the A and B cations occupy tetrahedral and octahedral sites, respectively.Spinels are ther- mally stable and because of their high activity they are used in a variety of important industrial reactions, such as the dehydrogenation of hydrocarbons.' Chromites are p-type semi- conductors and conduction proceeds via hopping of charge carriers between Cr3+ and Cr4+.6,7 The catalytic and gas sensing properties of CuCr,O,, ZnCr20, and NiCr,O,, among other chromites, have been studied.6*8 In most catalytic reactions involving spinels the reaction follows the Mars- van Krevelen mechani~m.~ Lattice oxygen is used and appears in the oxidation products. The resulting oxygen vacancies in the spinel are then replenished by gaseous oxygen from the atmosphere.It therefore seemed logical to investigate the gas sensing properties of CuCr204, ZnCr20, and NiCr,O,. The effects of substituting some of the A or B cations with other cations of different valencies were also studied. Experimental Preparation of spinels The spinels (AB204) were synthesized from nitrates using a wet chemical method." The basis of the reaction is the exothermic redox decomposition of metal nitrates and a suit- able organic compound. The starting materials used were Cu(NO,), (99.999%), Cr(NO,), 9H20 (99.99%), LiNO, (99.99%), Mg(N0,)- 6H20 (99.995%), Zn(N0,) (99.999%), Ni( 6H20 (99.999%) and glycine (99 +YO),all obtained from Aldrich. Metal nitrate A (0.02 moles) and metal nitrate i Presented at the Second International Conference on Materials Chemistry, MC', University of Kent at Canterbury, 17-21 July 1995.combined oxidising valency. Using the same elemental valency values for 0,H, N and +4 for C, the reducing valency for one mole of glycine was calculated. This value was used to calculate the number of moles of glycine required to balance the combined oxidising valency of the nitrates in solution. The glycine was then added to the solution and dissolved. The solution was heated at 100°C to remove most of the water. During this stage of the reaction a fine powder appeared. As the solution became more concentrated the viscosity increased rapidly and a clear resin was obtained. The resin was then heated at 300°C on a hotplate, causing it to ignite and yield a fine precursor powder.In the final step of the preparation, the precursor powder was calcined at 800 "C for 24 h in air to produce the spinel product. The spinels obtained were identified by X-ray diffraction (XRD) with Cr-Ka radiation, and the particles were observed under the transmission electron microscope (TEM). Electron spectroscopy for chemical analysis (ESCA) was used to deter- mine the cationic species present and their relative amounts. The analyses were carried out using a VG ESCALAB Mark I1 instrument with in-house software. Fabrication and testing of sensor elements Silver electrodes, approximately 1 mm apart, were painted onto glass substrates (25 mmx 8 mm) and dried at room temperature for 1 h.The powdered spinel was mixed into a paste with ethyl cellulose dissolved in 80 :20 toluene-ethanol. The ethyl cellulose was used as a binder and its loading was fixed at 5% by mass. The paste was then applied to the substrate to produce a thick film that partially covered the electrodes. In all preparations the thick film was dried for 3 h at 200 "C. The prepared sensors were mounted on heaters and tested at 220 "C to ppm levels of methanethiol, ethanethiol and NO2. A gas-flow rig was constructed so that the responses could be measured accurately. Together with a gas chamber, the major components of the flow rig are a Keithley 617 programmable electrometer, a Keithley 705 scanner, a Keithley 7056 voltage card and 7158 current card, thermal mass flow controllers, and a Viglen IV/33 personal computer.Electrical measurements were made using the Keithley 617 programm- able electrometer coupled with the Keithley 705 scanner. The electrical responses of the gas sensors to different atmospheres were recorded as current changes. This was achieved by using J. Muter. Chem., 1996, 6(3), 277-283 277 the electrometer to apply a potential difference across the sensors and measure the resulting current In expenments using NO2 and CH,SH, model 5850TR thermal mass flow controllers were used to administer a specified gas together with a diluting gas (air) from separate gas cylinders to the gas chamber Measurements of sensor responses were made under continuous flow conditions In experments with C2H5SH the chamber was sealed with rubber septums and C2H5SH vapour was injected into the chamber through a septum with a gas- tight syringe Results and Discussion XRD The XRD patterns for all the samples prepared confirm the formation of a spinel phase Many of the peaks are very broad owing to the fact that the average crystal size in the powder samples is rather small In XRD broadening of diffracted X-ray beams may occur if the average particle size is below 200 nm l1 Gas sensing properties of CuCr,04, ZnCr204 and NiCr204 Fig 1 and Table 1show that thick films of CuCr204 can detect NOz and that the sensitivity can be altered by substituting some of the Cu or Cr for Li and Mg respectively (see Fig 2 and 3 and Table 1) Exposure to NO2 causes the conductivity of CuCr204 to increase steadily, and in some cases the time taken to reach a steady-state response may be as long as 1 h Upon the removal of NO2 the sensors recovered quickly The 51 f.0 tlmin 80 Fig. 1 Response of a CuCr,O, sensor at 220°C to 20 ppm NO, (sensitivity 6 65) a. 0 t/min 90 Fig. 2 Response of a Cuo 8Lio ZCrzO4 sensor at 220 "C to 20 ppm NO, (sensitivity 41 76) 0 tlmin Fig. 3 Response of a CuCr, 9Mgo sensor at 220 "C to 20 ppm NO, (sensitivity 1102) rise in conductivity produced by NO2 is the result of processes that occur on the surface whereby the density and/or mobility of charge carriers is increased It is probable that NO2 adsorbed on the surface behaves in the same way as oxygen adsorbed on tin dioxide, in that electrons in the conduction band or acceptor states become trapped at the surface12 To see how this would lead to an increase in conductivity it is necessary to examine bnefly the processes involved in charge-carrier formation As previously stated, the charge carriers in chro- mites are mobile holes that hop from one Cr3+ site to another Their formation in CuCr204 is balanced by the reduction of Cu2+ to Cu+, and by the adsorption of oxygen to a lesser extent Note that Cr3+ and Cu2+ lose a significant amount of crystal-field stabilisation energy through oxidation and reduction, respectively This may limit the number of charge carriers produced in that Cu+ may revert to Cu2+ and Cr4+ to Cr3+, z e the holes are annihilated If it is assumed that charge-carrier formation is in equilibrium with opposing fac- tors, then the trapping of electrons at the surface would be expected to shift the equilibnum in favour of charge-carrier formation Since NO2is an electron-seeking gas, its adsorption in the immediate vicinity of Cu+ would be expected to trap electrons at that locale, and thus prevent the annihilation of holes (charge carriers) An additional factor leading to the observed increase in conductivity may be the existence of oxygen vacancies These may react with oxygen from the decomposition of NO2 to produce charge carriers, this is shown in eqn (1) using Kroger-Vink notation Vo ++O+OOx+2h (1) Introduction of Li for Cu and Mg for Cr both enhance the sensitivity of CuCr204 to NO2 (see Fig 2 and 3) The com- pounds that showed the most significant improvements in sensitivity were Cu, *Lie 2Cr204 and CuCr, ,Mgo XRD studies of these compounds revealed a spinel phase only, and TEM images (Fig 4) show that there are very minor differences in their microstructures It is proposed that sensitivity is enhanced by the formation of oxygen vacancies in the spinel lattice The presence of Li+ for Cu2+, and Mg2+ for Cr3+, would be expected to produce ionised oxygen vacancies as Table 1 Synopsis of observed sensitivity values for copper chromites after activation of the sensor surface by an initial influx of the gas to be detected (values for some repeated observations are given) ~~~~ PPm ~ sensitivity of gas used C u C r 0, cuO gL1O ZCrZo4 cuO gL10 ICrZ04 CuCrl 9Mg0 lo, NO, 20 6 65 36 65 - 13 09 50 5 25 37 10 - 13 79 50 4 99 42 61 - 14 78 CH,SH 10 0 90 124 6 33 - CzH,SH 67 6 71 7 11 12 05 - HZS 10 - - 6 21 - 278 J Muter Chem , 1996, 6(3), 277-283 Fig.4 TEM images of (a)CuCrz04,(b)CuCr, 9Mg0,104 and (c) Cuo 9Lio ,CrZO4. The particles of each sample are very similar. shown in eqn. (2): Li,O 2cuo 42Li,,' +0,"+V,' . (2) and likewise for Mg2+ [eqn. (3)]: 2Mg0 92Mg,,' +20," +V,' * (3) Note that in SnO, oxygen vacancies act as chemisorption sites, and this would also be true for spinels. The extra oxygen vacancies created by the presence of Li' and Mg2+ in the compounds Ch.8Li, 2Cr204 and CUC~~.~M~,-,~O~ would there- fore lead to a greater interaction with NO, and hence promote sensitivty. Exposure toppm levels of volatile sulfides causes a sharp conductivity decrease in CuCr,04 (see Fig.5). Some typical sensitivity values are reported in Table 1. It is not known exactly how volatile sulfides interact with CuCr,O,, but two a C 0 t Imin 30 Fig. 5 Response of a CuCrz04 sensor at 220 "C to 67 ppm CH,CH,SH (sensitivity 4.36) processes are believed to occur. In the first, volatile sulfides incident on the surface of CuCr204 react with chemisorbed oxygen to give C02, H20 and SO2. The loss of oxygen ions from the surface releases electrons back into the lattice and these may combine with Cr4+, thus annihilating the charge J.Mater. Chem., 1996, 6(3), 277-283 279 carriers and lowering the conductivity. It is also possible that some of the electrons could combine with Cu2+ ions. This would have no effect on the conductivity because the charge carriers would remain intact. In the second process it is assumed that Cu+ exists on the surface, and that lattice oxygen is used in the decomposition of volatile sulfides. Again, oxygen vacancies act as chemisorption sites. It is then proposed that the adsorbed sulfides react with lattice oxygen in accordance with the Mars-van Krevelen mechanism, leave the surface as C02 and HzO, and deposit sulfur as S2-ions. The S2-ions would then be expected to interact with Cu+ cations since Cu+ is a soft metal and forms strong complexes with ligands such as alkyl sulfides.The interaction between S2-and Cu+ to form Cu-S bonds requires that Cu+ loses an electron. The free electrons combine with Cr4+and thus lower conductivity. The long recovery times of the sensors due to the difficulty in breaking these bonds supports this idea. It was found that introduction of Li for Cu to give Cuo.9Lio.1Cr204improved sensitivity towards volatile sulfides (see Fig. 6 and Table 1). As well as containing Cr4+and Cu+,Cuo.9Lio.lCrz04will contain additional oxygen vacancies [see eqn. (2)] and will thus allow greater interaction between volatile sulfides and the spinel nature of the surface of spinels has not been determined, there is evidence to suggest that the cations in tetrahedral sites are not exposed to the atm~sphere.~If this were the case, then only through oxygen vacancies would the Cu+ ions become exposed and accessible to chemisorbed species.Furthermore, the increased sensitivity of Cuo~9Lio~lCr,0,would then be due to the greater number of exposed Cu' ions. It was found that the introduction of Mg had no significant effect on the sensing properties of CuCr204towards volatile sulfides. This could be due to the fact that oxygen vacancies produced by the presence of Mg2+ are not ideally situated in that they do not expose Cu+ ions to the atmosphere. The results of ESCA studies on the compounds CuCr204, CUC~~.~M~.,O, are shown in Fig.7. Forand Cuo.9Lio.1Cr204 illustrative purposes the three spectra are placed on the same set of axes. The region between 930 and 935 eV is of particular interest. Close examination of this region reveals that there are two peaks for each sample. The larger peak at ca. 935 eV can be attributed to tetrahedrally coordinated Cu2+ions, and the smaller peak at ca. 932eV to Cu' in the same environ-ment.' For CuCr20, the two peaks are very close together, the Cu' peak appearing as a shoulder on the Cu2+ peak. The lattice. Further substitution of Cu to give CU,,~L~~.~C~~O,shoulder becomes more distinguished for CuCrl~9Mgo~104,had and a less significant effect on sensitivity; sensitivity values of this for Cb.9Lio.lCr204the two peaks are almost completely compound are only slightly greater than those of CuCr,O,.resolved. In this sequence it can be seen that the binding Before proposing the second mechanism, it was assumed that energy corresponding to Cu+ decreases slightly. This is due to Cu+ ions exist on the surface. CuCr204 is a normal spinel the reduced influence of oxygen ions on the Cu+ cation and with Cu2+located in tetrahedral sites, and although the exact indicates the presence of oxygen vacancies. The fact that these shifts can be detected by ESCA proves that the process of oxygen vacancy formation is not limited to the bulk. From the results shown in Fig. 8 and Table 2 it can be seen that greater sensitivity to volatile sulfides is achieved with263m-1521a1 I ZnCr,O,.Like CuCr,O,, ZnCr,O, is a p-type semiconductor and the charge carriers are expected to be formed in the same way.' In CuCr,O, most of the valence electrons released by Cr3+ are accepted by Cu2+ to give Cu+. A similar process in ZnCr20, is unlikely because in spinels Zn2+ is generally con- 0 t/min . 40 sidered to be inert. The valence electrons can then only be accepted by gaseous oxygen through chemisorption, and Fig. 6 Response of a Cu0,,Li,,,Cr2O4 sensor at 220°C to CH,SH (sensitivity 6.33) 10ppm acceptor states lying just above the valence band. It is then reasonable to assume that at the surface of ZnCr20, oxygen al"bY 04 I I I I1 I I I I I 1 925 930 935 940 945 950 955 960 965 970 975 binding energylev Fig.7 Results of ESCA studies of (a) CuCr,O,, (b) CuCr,~,Mgo,,04and (c) Cu,.,Lio.,Cr,04 280 J. Mater. Chew., 1996, 6(3), 277-283 Table 2 Synopsis of observed sensitivity values for zinc chromites after activation of the sensor surface by an initial influx of the gas to be detected (values for some repeated observations are gven) Termmatron of 691 06 GR CH-,SFI~trcaxn~ 12 3 0 t/min 50 Fig. 8 Responses of ZnCr,O, to (a) 10 ppm CH3SH (sensitivity 299 23) and (b)1 ppm CH3SH (sensitivity 3.20) has the same function as Cu2+ in CuCr,O,. As with CuCr,O,, it is proposed that two processes occur on the surface of ZnCr,O, when volatile sulfides are adsorbed. The first process is a simple reaction between volatile sulfides and chemisorbed oxygen to give CO,, H,O and SO,.All the electrons released by the loss of chemisorbed oxygen would then recombine with Cr4+ and hence cause a fall in conductivity. This may be one reason why ZnCr,O, is more sensitive than CuCr,O,. In CuCr,O, some of the released electrons may be accepted by Cu2+ instead of Cr4+ and thus have no affect on conductivity. The second process thought to occur follows the Mars-van Krevelen mechanism in which lattice oxygen reacts with gases NO2 20 299 287 20 290 285 CH3SH 1 3 20 2 88 10 297 292 10 299 29 1 C,H,SH 15 278 262 from the air and the oxidised products leave the surface. This produces oxygen vacancies in the ZnCr,O, lattice which, when ionised, releases electrons that could annihilate charge carriers and reduce conductivity. Substituting some of the Cr for Mg to give ZnCr, ,Mg, had little effect on the microstructure (see Fig.9) or the sensitivity; the sensitivity values remained fairly high (see Table 2). It is therefore assumed that the interactions between volatile sulfides and the two samples are the same. The two samples gave similar responses to NO, and also showed greater sensitivity than CuCr,O, (see Fig. 10 and Table 2). The rise in conductivity would be due to NO, abstracting electrons from the valence band during chemisorption and thus produc- ing hole charge carriers. Again, the presence of Cu+ may explain why CuCr,O, is less sensitive than ZnCr,04 to NO,. As well as the valence band, the Cu+ ions may act as a source of electrons for the chemisorption of NO, by being oxidised to Cu2+.As a result the conductivity of CuCr,O, changes by a smaller amount than that of ZnCr,O,.Fig. 9 TEM images of (a)ZnCr,O, and (b)ZnCr, ,Mg, The particles of each sample are very similar and cuboid in shape J. Mater. Chem., 1996,6(3), 277-283 281 Fig. 10 Response of a ZnCr204 sensor at 220 "C to 20 ppm NO,(sensitivity 290 36) From the results shown in Fig 11 and Table 3 it can be seen that reasonable responses to CH3CH2SHcan be achieved with NiCr204 Although the NiCr,04 sensors are less sensitive than the ZnCr,04 sensors, they are more sensitive than the CuCr204 sensors In companng ZnCr204and CuCr204sensors, it was proposed that the ease with which copper ions change oxi-dation state is detrimental to sensor performance This means that processes occurnng on the surface of CuCr204involving the exchange of electrons are not limited to the valence band where charge are formed Or destroyed The Of copper ions to change oxidation state means that they can interact with adsorbed species and as a result the electrons involved do not affect conductivity In ZnCr,04 sensors, electronic processes are restricted to the valence band, which explains their greater sensitivity Since NiCr204shows greater sensitivity to than CuCr204, It Is --279MQ Q: 13 1 -< 400 tlmn Fig.ll Response of an NiCr204 sensor at 220°C to 67ppm CH3CH2SH(sensitivity 26 33) 8000 -.I h 7000 5000 a u) g!r 4000 3000 2000 Table 3 Synopsis of observed sensitivity values for nickel chromites after activation of the sensor surface by an initial influx of the gas to be detected (values for some repeated observations are given) PPm sensitivities of gas gas used NiCr,O, NIO SLIO 2cr204 NO2 20 0 43 188 50 0 59 170 CH3SH 10 3 33 C2H5SH 67 16 64 67 26 33 For the detection of NO,, sensitivity S is defined as S=(R,,, -RNo~)/RNo~~ are the resistances in air and NO2,where Rair and RNO~ respectlvel~For the detection of volatile sulfides (VS) sensitivity is defined as S =(Rvs -R,,,)/R,,,, where Rvs and R,,, are the resistances in and to assume that nickel ions do not change oxidation state as readily as copper ions With respect to gas sensing, this means that more of the electrons released by the reactions betweenadsorbed oxygen and are accepted by Cr4+ ions Since fewer electrons are accepted by Ni2+ to give Ni', conductivity in NiCr204 changes to a greater extent ESCA studies on this compound (Fig 12) revealed two peaks for Ni at binding energy values of 856 50 and 861 95 eV It is likely that these peaks correspond to Ni2+,the presence of Ni+ could not be confirmed The ESCA spectra for the compounds NiCr, ,Mg, ,04and Ni, &lo 2Cr2O4 (Fig 12) also had Ni peaks at the same binding energy values Thus it was concluded that the introduction of Mg for Cr and Li for Ni had no effect on the environment of the Ni2+ions However, the behaviours of the two compounds were quite different NiCr, ,Mg, gave reasonable responses to CH3CH2SH with slightly lower sensitivity values than NiCr204 Ni, &io 2Cr2O4, on the other hand, showed very little sensitivity to volatile sulfides Based on the assumption that Li has no affect on the environment of Ni2+,it is proposed that its presence in NiOsLio2Cr2O4 is balanced by the oxidation of 25% of the remaining Ni2+ions or 10% of the Cr3+ ions TEM images also show that the microstructures of the compounds are very similar (Fig 13) o! I1 I II I1 I I 1 I 1 I I1 i 845 e50 855 860 865 a70 875 880 885 890 895 binding energy/eV Fig.12 Results of ESCA studies of (a) Nio *Lie &r204, (b) NiCr2O4 and (c)NiCr, ,Mg, 282 J Muter Chem , 1996,6(3), 277-283 Fig.13 TEM images of (a)NiCr,O,, (b) NiCr, ,Mg, The NO, sensing properties of NiCr,O, were found to be relatively poor in comparison with ZnCr204. Table 3 shows the sensitivity values obtained for NiCr,04 to various concen- trations of NO2. NiCrl.,Mgo.lO, was found to be even less sensitive, but the sensitivity of Ni0.8Li0.2Cr204 was surprisingly high (Table 3). TEM images of the compounds NiCr204 and Nio.8Lio.2Cr204showed no difference in their microstructures; therefore, it is reasonable to assume that the enhancement in sensitivity is due to the presence of Li. We propose that the electropositive nature of Li modifies the surface of Nio~,Li0.,Cr,O4 and makes it more reactive to oxidising species such as NOz.Conclusion We have shown that the chromites of copper, zinc and nickel have potential use as solid-state sensors of both oxidising and reducing species. We have observed that the sensing character- istics of chromites can be modified by replacing some of the A or B cations with cations of different valencies. and (c)Ni, *Lie ,Cr,O,. The particles of each sample are very similar. References 1 G. Sberveglieri and S. Groppelli, Sensors Actuators B, 1991,4,457. 2 J. Mizsei and V. Lantto, Sensors Actuators B, 1991,521. 3 J. Tamaki, Chem. Lett., 1991, 575. 4 A. F. Wells, Structural Inorganic Chemistry, Oxford University Press, London, 1975, p. 489. 5 N. J. Jebarathinam and M. Eswaramoorthy, Bull. Chem. SOC.Jpn., 1994,67,3334. 6 D. Basak and J. Ghose, J.Solid State Chem., 1994,112,222. 7 K. S. R. C. Murthy and J. Ghose, J. Catal., 1994,147,171. 8 Y. Shimizu and S. Kusano, J.Am. Ceram. SOC., 1990,73,818. 9 J. Jacobs and A. Maltha, J. Catal., 1994,147,294. 10 J. J. Kingsley and L. A. Chick, Better Ceramics Through Chemistry V,Mater. Res. SOC.Symp. Proc., 1992,271, 113. 11 A. R. West, Solid State Chemistry and its Applications, Wiley, Chichester, 1984,p. 173. 12 P. T. Mosely, A. M. Stoneham and D. E. Williams, Techniques and Mechanisms in Gas Sensing, Adam Hilger, Bristol, 1991, pp. 108-138. Paper 5/04826E;Received 21st July, 1995 J. Mater. Chem., 1996, 6(3), 277-283 283

 

点击下载:  PDF (1094KB)



返 回