Experimental study of filtration efficiency of nanoparticles below 20 nm at elevated temperatures

详细信息	查看全文 | 推荐本文 |

• 作者：W.G. Shin ; G.W. Mulholl ; S.C. Kim ; D.Y.H. Pui
• 关键词：Filtration efficiency ; Elevated temperature ; Nanoparticle ; Thermal rebound
• 刊名：Journal of Aerosol Science
• 出版年：2008
• 期刊代码：45_00218502
• 类别：ce
• 出版时间：June 2008
• 卷：39
• 期：6
• 页码：488-499
• 文件大小：502 K

摘要

For a number of reasons, a diffusion battery is not entirely a convenient instrument to study the interaction of nanometer particles with solid surfaces, although a theory was developed which included the thermal rebound of nanometer particles (Wang, 1993). Hence, some researchers prefer to use a single wire screen for experimental work (Ichitsubo et al., 1996).

The experimental data of nanometer size particle penetration through a series of screens SS40 (Tetko Inc., Elmsford, NY) at three different temperatures (295 K, 316 K and 337 K) are presented in this work. Aerosols of tungsten oxide and molybdenum oxide in the size range from 3.1 to 15.4 nm were studied. The data are compared with the theoretical model developed for a screen-type diffusion battery (Cheng and Yeh, 1980).

EXPERIMENTAL

Penetration of nanometer particles in the size range from 3.1 to 15.4 nm through a series of wire screens has been studied experimentally. A stainless steel screen SS40 (Tetko Inc., Elmsford, NY) was tested. The experimental system used in the study consisted of an aerosol generator, an electrostatic classifier with a short column of 11.11 cm (EC TSI Model 3071A), a thermal chamber with two identical filter holders inside and an ultrafine condensation particle counter (UCPC 3025A, TSI Inc.). Monodisperse aerosols of molybdenum oxide and tungsten oxide were obtained in the following way. Polydisperse aerosols were first produced by a photochemical aerosol generator (Baklanov et al., 1995). The aerosols were introduced into the EC at a volume flow rate of 2 l/min. Aerosol particles were separated according to their mobility, creating a monodisperse aerosol at the outlet of the classifier. (The most probable state of the oxides obtained by oxidation of carbonyls are MoO₃ and WO₃ but the composition of aerosol particles has not been checked).

We measured the penetration of monodisperse aerosols through 8 stages of wire screen which were placed in a filter holder with a diameter of 3.81 cm. An identical holder was mounted in parallel to the first one and could be switched into the line by means of a three way valve. Both holders, in turn, were housed in a thermal chamber in which we could stabilize and control the temperature of the aerosol flow. The temperature was measured by two thermocouples that were located in the rear part of the holders. During experiments the temperature inside the thermal chamber was varied from 295 to 337 K. The volume flow rate through the holders was kept at a constant value of 2 l/min, which corresponds a mean flow velocity 2.92 cm/s at temperature 21°C. The penetration was calculated as the ratio of particle fluxes after and before the holder. The concentration of aerosol particles was measured by the ultrafine condensation particle counter.

RESULTS AND DISCUSSION

The results of the measured penetration versus particle size for the studied substances at different temperatures of the thermal chamber are presented in Figs. 1 and 2. The dotted linesrender="y">mg src="http://www.sciencedirect.com/cache/MiamiImageURL/B6V6B-46204K8-5T-1/0?wchp=dGLbVtz-zSkWb" class="charImg" alt="image" title="image" height="16" width="95">Figure omittedhere are theoretical curves for maximum and minimum temperatures. These curves were calculated from the classical theory (Cheng and Yeh, 1980). Figure 1 shows agreement between the experimental data and the theory for tungsten oxide of particle size 3.1 nm and larger. Also, the data slightly depend on the temperature. A somewhat different picture is seen in Fig. 2 corresponding to molybdenum oxide. The theory and our experimental data agree reasonably well for particle sizes down to 5 nm. For smaller particles, the predictions differ from the observed data. Moreover, the discrepancy increases with decreasing particle size and the experimentally determined penetration is twice that of the theoretical value for a particle diameter of 3.1 nm. Another interesting result is that for particles less than 5 nm, the penetration depends on the temperature. This is clearlyrender="y">mg src="http://www.sciencedirect.com/cache/MiamiImageURL/B6V6B-46204K8-5T-2/0?wchp=dGLbVtz-zSkWb" class="charImg" alt="image" title="image" height="16" width="95">Figure omittedseen in Fig. 2. Analysis of the formulas shows that the penetration is a function of temperature and this function depends on the thermal rebound parameter R. If R>2 the dependence is very strong (Wang and Kasper, 1991). For 0.4<R<2 the dependence is a complex function of T, but weaker than the previous case. The observed behavior of the experimental data could be explained by the thermal rebound effect of particles against surfaces of wire screens.