The Effect of Aerosol Loading on the Performance of Thermal Desorption Tubes and the Development and Evaluation of a Pre-filter for Adsorption Tubes
Date Issued
2016
Date
2016
Author(s)
Chang, Chien-Yu
Abstract
Thermal desorption tubes are commonly used to quantify trace amount of VOCs in the workplace. Despite a small piece of glass wool normally placed in front of the sorbent, it is unlikely an absolute filter, and aerosol penetration and deposition on the sorbent are inevitable. Therefore, this study aimed to characterize the effect of aerosol loading on the performance of thermal desorption tubes. The ultimate goal was to design a pre-filter for a better performance of thermal desorption tubes in practical dusty working environments. Homemade thermal desorption tubes loaded with 100 mg Carbopack X sorbents were used in the present study. The sampling flow rate was 0.2 L/min. Acetone vapor was generated to perform the adsorption tests and the nitrogen gas was used for desorption. A flame ionization detector (FID) was employed to measure the acetone concentrations upstream and downstream of the thermal desorption tubes. A constant output aerosol generator (TSI 3076) and an ultrasonic atomizing nozzle (Sono Tek) were used to generate sub-micrometer-sized and micrometer-sized aerosol particles, respectively. For aerosol penetration test, a scanning mobility particle sizer (TSI SMPS) and an aerodynamic particle sizer (TSI APS) were employed to measure the aerosol concentrations and size distributions upstream and downstream of the test filters. Both solid (NaCl) and liquid (DEHS) particles were generated and the pressure drop across the filter was simultaneously monitored. Glass wool plug, stainless steel mesh (#400, #1500), polyurethane foam (110 ppi) and fibrous filter disc cut from N95 filtering facepieces were tested in this work. Only the glass wool installed upstream of sorbent was tested for aerosol penetration. To verify the performance of the pre-filter, two thermal desorption tubes (one with the pre-filter) were simultaneously challenged with DEHS particles and then compared with the initial breakthrough curves. The experimental results showed that the sub-micrometer-sized and micrometer-sized solid particle loading (loaded mass 0.6 mg) decreased the adsorption capacity, 5% and 13% less, respectively. The liquid particles could significantly deteriorate the sorption performance, because the deposited liquid particles might form the film covering the activate sites. The most penetrating particle size (MPPS) of the glass wool was 0.3 ~ 0.5 μm and the aerosol penetration of MPPS was about 60 ~ 75% with the pressure drop of 21.2 ± 8.1 mmH2O under the sampling flow of 0.2 L/min. As for the 400 mesh stainless steel, the aerosol penetration of 40 pieces was comparable to that of the glass wool with the face velocity of 17 cm/s and pressure drop of 32 mmH2O. High aerosol collection efficiency (for example, 90%) can be achieved by increasing the mesh number and decreasing the face velocity. However, the use of stainless steel with high mesh number (1500 mesh in this case) was not cost effective. With the use of 110-ppi foam, the total length of the foam was estimated to be as long as 300 mm to attain the required collection efficiency (90%) at a face velocity of 0.68 cm/s. The aerosol collection efficiency can be enhanced by increasing the foam packing density. However, it was difficult to guarantee the foam packing quality to gain reliable performance. Moreover, the N95 filter disc (D = 6.4 mm, H = 0.8 mm) showed an excellent performance on aerosol collection with a fairly low pressure drop of 9.8 mmH2O. Among the filter materials tested, the N95 filter disc worked best, for low cost, low pressure drop and stable quality. Finally, the devastating effect of aerosol loading on the adsorption performance of thermal desorption tubes can be significantly leveraged with the use of a N95 pre-filter.
Subjects
Sampling devices
particles
particle loading
pre-filters
Type
thesis
File(s)![Thumbnail Image]()
Loading...
Name
ntu-105-R02841028-1.pdf
Size
23.32 KB
Format
Adobe PDF
Checksum
(MD5):6f43314e98ad0c727b13eb75e579bf80