Performance Evaluation of Four Commercial Optical Particle Counters ()
1. Introduction
Single particle light scattering is one of the most widely used techniques for measuring the particle number size distribution in the range from 0.2 µm up to several microns. Single Optical Particle Counters (OPCs) measure the light elastically scattered from a single particle illuminated by a well defined light source while it is passing through the sensing volume of the instrument. The scattered light intensity is utilized as a measure of the particle’s size. The Mie theory is used to predict the light scattering intensity of an electromagnetic wave by a homogenous spherical particle. Consequently, the relative signal response of an OPC can be determined by knowing the characteristics of the light source, the detector, and the physical configuration of the sensitive volume, and the properties (mainly size, refractive index and shape) of the particle [1]. Several efforts have been made to test experimental OPC’s response against Mie theory, using polystyrene latex [2-4].
In recent years, measurement techniques based on aerosol light scattering have received greater attention because of the possibility of deploying the aerosol size distribution as a proxy of the health related fractions like PM10, PM2.5 and PM1 [5]. The increasing importance of adverse health effects of air pollution has driven research towards the development of real time measurement techniques, which can couple meteorological parameters and particulate matter characteristics (mainly concentration and size distribution), thus providing an improved understanding of the relationship between sources and effects of mitigation actions that could be taken. In order to estimate the aerosol mass concentration from the OPC size distribution measurements, as well as the influence of particle shape and refractive index, particle density must also be taken into account. Since the latter is generally not known over the size distribution spectrum, only empirical conversion matrix factors can be used. Therefore, only sampling site-dependent correlations can be obtained [6].
In the present study, four commercial OPCs were tested in parallel with monodisperse generated aerosols, in order to compare their particle counting efficiencies and particle size classification capabilities.
The aim of the present study is to caution against careless use of the counters in monitoring situations. The study was promoted by the Italian Aerosol Association (IAS) in the context of the working group on “PMx aerosols”.
2. Experimental Set-Up
The experimental work was carried out at the Institute of Atmospheric Sciences and Climate (ISAC) of the Italian National Research Council (CNR) in Bologna during summer 2010. Monodisperse particles were generated by means of MAGE (Monodisperse Aerosol Generator, Lavoro & Ambiente, Bologna, Italy). MAGE is a condensation-type generator capable of producing monodisperse particles in a wide size range. A stream of nuclei (mainly NaCl nuclei) is exposed to the vapour of a low-volatile liquid at an elevated temperature, and the controlled heterogeneous condensation of the vapour onto the nuclei results in the generation of monodisperse aerosol [7,8]. MAGE was used with Carnauba Wax (CW) at 270˚C and 300˚C oven temperatures in order to obtain two different monodisperse aerosol distributions. Particles size was ascertained by aspirating the aerosol on a Nuclepore filter (Whatman, 0.20 mm pore size) at the outlet of the MAGE and observing the filter under SEM. The measured diameter were (1.00 ± 0.08) µm at the lower temperature and (1.40 ± 0.15) µm at the higher temperature value. Figure 1 shows a typical SEM image of the aerosol generated by the MAGE at 270˚C, sampled directly from the generator outlet. It can be seen that the particles are of spherical shape. Their density is supposed to be the same as the CW bulk density (about 990 Kg∙m−3), while the refractive index is 1.45˚C at 20˚C.
Since the particle number concentration was very high at the MAGE outlet (of the order of 1012 particles∙m−3), two diluters were positioned in series, after which all OPCs sampled from a mixing volume. The diluters were based on the Venturi effect, as shown in the paper of Yoon et al. [9].
A test was also carried out with Polystyrene Latex particles (PSL, Agar Scientific) with nominal diameter of
Figure 1. Monodisperse carnauba wax particles obtained with the MAGE at the working temperature of 270˚C.
1.0 mm and refractive index of 1.59. The PSL particles were nebulized with a six-jet TSI atomizer model 9306, aerosol generator, and were sampled without dilution.
The generated aerosols were not strictly monodisperse due to various causes. As a matter of fact some NaCl nuclei could not have been nucleated by the CW vapour. Based on the concentration of the NaCl solution and the size of the droplets generated by the Collison, the nuclei should have a diameter of about 60 nm [10]. In addition the generation of PSL particles involves even residue particles which are below 100 nm [11]. Therefore the OPC counters should not be able to detect both the NaCl nuclei and PSL residue particles. A possible coagulation of the aerosol at the exit of the MAGE due to the high particle concentration was avoided by diluting the aerosol.
The OPCs tested were the following: DustMonit (Con. Tec Engineering srl, Milan) with sampling flow rate of 1 lpm, size separation range of 0.3 - 10 mm and 8 channels; Grimm 1.107 (Grimm Aerosol Technik, GmbH) with sampling flow rate 1.2 lpm, size separation range of 0.2 - 32 mm and 31 channels; Grimm 1.108 (Grimm Aerosol Technik, GmbH) with sampling flow rate 1.2 lpm, size separation range of 0.3 - 20 mm and 15 channels; ParticleScan CR (ParticleScan, Advanced Particle Counters. IQAir) with sampling flow rate 2.8 lpm, size separation range of 0.3 - 5 mm and 6 channels. Each OPC name is shortened to the following initials: DM (DustMonit), EN (Grimm 1.107), GR (Grimm 1.108) and PS (ParticleScan). When comparing the OPCs, the particle number concentration of the EN counter was considered only for sizes larger than 0.3 mm (first two channels excluded). The tested OPCs have a laser wavelengths in the range from 675 nm (Grimm 1.107) to 780 nm (Grimm 1.108). In accordance with the manufacture’s recommendations, Grimm OPCs are calibrated with polystyrene latex, while for DustMonit and ParticleScan the calibration procedures are not available.
The OPC performances were compared by measuring the size distribution of monodisperse aerosol particles simultaneously sampled from a 1 liter mixing volume. Before each test session the background particle number concentration in the laboratory was measured and verified to be less than 1% of the particle number concentration generated in the comparison experiments. In order to compare the findings from each OPC, the particle number concentration in each size channel (dN) was normalized by considering the channel width (dN/dDp). During each run (for both particle diameters), an aerosol sample generated by MAGE was sampled on Nuclepore filter media for SEM observations. Only tests with a total particle number concentration less than 2 × 106 l−1 were considered, as this is the maximum particle number concentration recommended by the respective manufacturers to avoid coincidence errors.
In addition, in order to confirm the results of the comparison, the GR counter was subsequently sent to the manufacturer for the yearly revision (calibration checks and adjustments). On its return, a further test was performed against a Las-X (PMS, Inc) active cavity laser.
3. Results and Discussion
The results are presented below, separately for the total particle number concentration and the particle size classification.
3.1. Total Particle Number Concentration Measurements
1) Carnauba wax particles
Figure 2 shows the ratio of the total particle number concentration measured by each OPC and the average total particle number concentration (M), obtained by averaging the total counts of the single OPC, for the four tests performed with monodisperse particles at (1.00 ± 0.08) mm (MAGE at 270˚C). The bars show one standard deviation. The histogram shows the results considering the OPC channels from 0.3 mm and from 0.5 mm separately.