- Email: [email protected]
Design and performance of a single-pass bubbling bioaerosol generator
ARTICLE IN PRESS
Atmospheric Environment 39 (2005) 3521–3533 www.elsevier.com/locate/atmosenv
Design and performance of a single-pass bubbling bioaerosol generator Gediminas Mainelisa,, David Berrya, Hey Reoun Ana, Maosheng Yaoa, Kevin DeVoeb, Donna E. Fennella, Rudolph Jaegerc a
Rutgers University, Department of Environmental Sciences, 14 College Farm Rd., New Brunswick, NJ 08901, USA b BGI Inc., 58 Guinan Street, Waltham, MA 02451, USA c CH Technologies, Inc., 263 Center Avenue, Westwood, NJ 07675, USA Received 31 August 2004; accepted 8 February 2005
Abstract We describe and analyze a new particle generator that utilizes a bursting bubble principle and eliminates carrier fluid reuse. In this Liquid Sparging Aerosolizer (LSA), a suspension of particles or microorganisms is pumped at a flow rate of 0.2–2 mL min1 to the top surface of a porous stainless-steel disk where it forms a thin suspension film. Filtered air is then sparged through the disk into the film causing it to break into bubbles that subsequently burst, releasing particles into the air. The released particles are then captured by the sparging air stream and are carried away. Particles that impinge the glass vessel and liquid droplets not captured by the air stream drain to the bottom of the vessel and play no further role in the aerosolization process. We tested the LSA with disks of different pore sizes (0.2, 0.5, 2.0 and 10.0 mm) and different air flows (2–30 L min1) through the porous disks while generating polydisperse and monodisperse particles. Our tests showed that the use of 0.5 and 2.0 mm porosity disks resulted in the highest output of PSL particles in the desired size range, i.e., comparable to bacterial size. Each pore size seemed to have an optimal air flow rate; the produced aerosol concentration increased with increasing suspension delivery rate. The LSA also demonstrated stability of output concentration when aerosolizing particles over extended periods of time. In addition, the size distribution of injury-sensitive Pseudomonas fluorescens bacteria virtually did not change during 90 min of continuous aerosolization by the LSA. In fact, there was no (0%) viability loss, whereas the bacterial spectrum produced by a Collison nebulizer changed significantly over 90 min and there was a 50% loss in viability. The results indicate that the new instrument could be used to generate particles for the evaluation of pathogen collection methods, inhalation and other studies where extended delivery of stable and undamaged biological aerosols is required. r 2005 Elsevier Ltd. All rights reserved. Keywords: Microorganism aerosolization; Microorganism viability; Bursting bubble
1. Introduction
Corresponding author. Tel.: +1 732 932 7166;
fax: +1 732 932 8644. E-mail address: [email protected] (G. Mainelis).
Inhalation studies, instrument calibration, evaluation of pathogen collection methods and other aerosol investigations require stable and reliable aerosol and bioaerosol generators. Reliable generation of biological
1352-2310/$ - see front matter r 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.atmosenv.2005.02.043
ARTICLE IN PRESS 3522
G. Mainelis et al. / Atmospheric Environment 39 (2005) 3521–3533
aerosols is also needed to study transport and deposition of indoor bacteria and mold. Among several methods used to aerosolize microorganisms, pneumatic nebulization is probably the most commonly used tool (Chen et al., 1994; Jensen et al., 1992); however, nebulizing generators produce high shear forces that may affect microorganism viability (Griffiths and DeCosemo, 1994). Commercially available nebulizers, such as the Collison nebulizer, are able to produce high concentrations of aerosol, but have been suspected of injuring and fragmenting microorganisms (Reponen et al., 1997). Griffiths et al. (1996) used a glass atomizer to aerosolize Escherichia coli cells and Penicillium expansum spores and found that aerosolization reduced the culturable fraction of E. coli to near zero. Israeli (1973) also indicated that bacteria submitted to aerosolization lose their ability to form colonies on nutrient media. In addition, in a standard Collison nebulizer with 20 mL of water in the flask, most of the liquid solution is recirculated about every 6 s (May, 1973). Thus, because the suspension is ‘‘recycled’’ the bacteria are repeatedly subjected to shear force stress. It has also been shown that nebulizers produce large concentrations of aerosol in the 0.1–0.5 mm size range, an unimportant range for bio-aerosol generation (Dennis et al., 1990). When aerosolized microorganisms are used to evaluate the efficiency of various bioaerosol collection and control methods, preservation of the organisms’ structural and biological integrity is highly desired. Thus, there is a need for a bioaerosol generator that can not only efficiently produce a stable bioaerosol output, but also minimize injury and death to sensitive microorganisms. Ulevicius et al. (1997) proposed a generator in which particles, suspended in a liquid, are produced from a bubbling liquid. Injection of tangentially oriented drying jets dries the airborne droplets and carries the particles away from the bubbling source and out. The bursting bubble action has also been shown to produce a higher aerosol concentration than nebulization for the production of short-chain fatty acid mists (White and Lucke, 2003). The bubbling mechanism has been studied as a naturally occurring phenomenon and has been recognized as a significant factor in aerosolization of seawater and suspended contaminants from breaking waves (Saint-Louis and Pelletier, 2004). The generator proposed by Ulevicius et al. (1997) was shown to effectively aerosolize polystyrene latex (PSL) particles of 0.71–5.1 mm in size. In addition, it introduced a lower metabolic injury to Pseudomonas fluorescens bacteria compared to Collison nebulizer (Reponen et al., 1997). The microorganisms aerosolized using the generator were found to carry fewer electrical charges compared to those generated by Collison nebulizer (Mainelis et al., 2001), which indicates that liquid disruption is not as violent as in pneumatic nebulizers. However, even in this generator a certain fraction of the suspension was still
being recycled. It was also shown that the particle output depends on the distance between the drying jets and the liquid surface. Thus, as the particles and the liquid are being dispersed at a rate of approximately 8.7 g h1, the level of the suspension inside the generator is lowered, a process that increases the distance between the drying jets and the liquid surface resulting in a lower aerosol concentration. Also, the described two-flow generator requires two separate air flows—one for bubbling the liquid and one for drying the droplets—a requirement that may complicate the experimental setup. To overcome this complication, Ulevicius et al. (1997) also suggested a single-flow generator in which the bubbling air flow and the drying air flow are combined into one. However, the presented single-flow design was expected to have a lower utility than the twoflow design and was only minimally investigated. In the research described in this paper, we present and investigate a new aerosol and bioaerosol generator that not only utilizes the bubbling liquid principle to induce low shear stress to microorganisms, but also features no suspension ‘‘recycling’’ and a simple design made possible by the use of the same air flow for both bubbling and drying air. Use of the single-pass design was expected to minimize the output of fragments and non-viable microorganisms when producing biological aerosols. In our proposed Liquid Sparging Aerosolizer (LSA), a suspension of particles or microorganisms is pumped at a certain flow rate to the top surface of a porous stainless-steel disk where it forms a liquid film. The air is then sparged into the film through pores in the disk causing the liquid film to break into bubbles that subsequently burst, releasing particles into the air. The released particles are then captured by the same air stream and are carried out of the device. Particles and liquid not captured by the air stream coalesce, drain to the bottom of the vessel and play no further role in the aerosolization process. The primary goal of this investigation was to analyze the performance of the LSA as a function of the disk pore size, the suspension delivery rate, and the sparging (aerosolization) air flow rate. A separate set of experiments was performed to investigate the generation efficiency and stability as well as the possible effect of the proposed aerosolization device on sensitive microorganisms.
2. Materials and methods 2.1. Design of the Liquid Sparging Aerosolizer The LSA features a bursting bubble design that minimizes shear forces on particles being aerosolized (Fig. 1). The particle suspension used to generate an aerosol is delivered to the device via a peristaltic pump operating at liquid delivery rate, QL , from 0.17 to
ARTICLE IN PRESS G. Mainelis et al. / Atmospheric Environment 39 (2005) 3521–3533
3523
Fig. 1. Schematic representation of the Liquid Sparging Aerosolizer (LSA).
2 mL min1 (a syringe can be used instead of a peristaltic pump). The suspension travels downward through a narrow vertical needle and is dripped onto a 1 in (2.54 cm) porous stainless-steel disk (Mott Corporation, Farmington, CT) where it forms a thin liquid film. The disk is approximately 1 cm below the needle outlet. The aerosolization air flow, QLSA , is delivered under pressure to the underside of the porous disk via a separate airtight channel. The air is then forced by pressure upwards through the disk’s pores into the liquid (air sparges into the liquid). The multiple air streams emanating from the porous disk break the suspension film into bubbles within which particles are contained. An expanding pressure gradient between the inside and outside of the bubbles causes them to burst, aerosolizing the particles.
Visually, the process resembles boiling of fine bubbles. The same multiple air streams ðQLSA Þ then contribute to the drying of the airborne particles and liquid filaments, carrying the resulting particles and liquid residues out the exit port. Particles and liquid not captured by the QLSA air stream (as shown later they constitute a small fraction) are collected at the bottom of the glass vessel and play no further role in the aerosolization process. Thus, sensitive particles, such as microorganisms, participate in the generation process only once and there are no strong shear forces involved. Therefore, this generator should impart less damage to fragile elements compared to traditional pneumatic nebulization. The dimensions of the glass vessel are the same as those of the Collison nebulizer (Model MRE CN24/25, BGI Inc.,
ARTICLE IN PRESS 3524
G. Mainelis et al. / Atmospheric Environment 39 (2005) 3521–3533
Waltham, MA). The 1-in diameter porous disk facilitating the bubbling action can easily and quickly be replaced with other 1-in disks featuring different pore sizes. The liquid delivery flow rate, QL , and the aerosolization flow rate, QLSA , can also be easily adjusted for optimal performance. The effect of these parameters on the LSA performance was investigated in our research described below. 2.2. Test particles The performance of the LSA was investigated with monodisperse PSL particles (Bangs Laboratories, Fishers, IN) of three different sizes: 0.99, 2.93, and 5.09 mm in diameter. PSL suspensions were prepared by suspending aliquots of PSL particle stock using filtered (Milli-Q system, Millipore, Billerica, MA) water. Generation stability experiments were performed with polydisperse NaCl particles prepared by dissolving 1 g of reagent quality NaCl (Mallinckrodt Baker, Inc., Phillipsburg, NJ) into 500 mL of purified water. The effect of the LSA on viability and size distribution of sensitive bacteria was investigated using P. fluorescens vegetative cells. The rod-shaped gram-negative P. fluorescens bacteria are 0.7–0.8 mm in diameter and 1.5–3.0 mm in length (Palleroni, 1984) and are commonly found in ambient air (Go´rny and Dutkiewicz, 1998). P. fluorescens vegetative cells were grown in tryptic soy agar (Becton Dickinson Microbiological System, Sparks, MD) at 26 1C for 18 h. The bacteria were separated from the broth by centrifuging three times at 7000 rpm (BR-4, Jouan, Winchester, VA) for 7 min each time. The bacterial suspension was prepared by suspending the pellet in freshly filtered (Milli-Q system,
Millipore, Billerica, MA) water which had been sterilized. 2.3. Test system and test parameters The test system used in this research (Fig. 2) was designed to test the performance of the LSA under different experimental conditions. For certain experimental conditions, the LSA performance was compared to that of the Collison nebulizer (BGI Inc., Waltham, MA). Particles produced by the LSA and the Collison nebulizer at aerosolization flowrates of QLSA and QCOLL , respectively, entered a HEPA-filtered dry air flow, QDRY ¼ 50 L min1 , and were carried into the measurement chamber. The concentration, C AIR , and size of the test particles in the measurement chamber were isokinetically measured using an optical particle counter (Model 1.108, Grimm Technologies Inc., Douglasville, GA). This device measures particles from 0.3 to 20 mm in 15 size channels and operates at a flow rate QGRIMM ¼ 1:2 L min1 . Since both the LSA and the Collison nebulizer were operated at several different aerosolization air flows, which changes the total air flow amount at the measurement point, the number of particles produced by each device under different conditions was evaluated in terms of total particle output, N TOT , defined as N TOT ¼ C AIR ðQA þ QDRY Þ,
(1)
where QA is aerosolization air flow for the generator used: QLSA for the LSA or QCOLL for the Collison nebulizer. The performance of the LSA was evaluated with disks of four different pore sizes, DP ¼ 0:2, 0.5, 2.0, and 10.0 mm. The porosity of these disks (percentage of
Fig. 2. Experimental setup.
ARTICLE IN PRESS G. Mainelis et al. / Atmospheric Environment 39 (2005) 3521–3533
surface covered by pores) is 20%, 25%, 35%, and 40%, respectively. The particle output from the LSA was measured for three suspension delivery rates QL ¼ 0:17, 0.6, and 2.0 mL min1 provided by a peristaltic pump (Masterflex C/L, Cole-Parmer, Inc., Vernon Hills, IL) and four aerosolization air flows QLSA ¼ 2; 5, 10, 20, and 30 L min1 delivered at the same pressure of 50 psi. The desired air flow rates were achieved using an adjustable flow meter. At the selected pressure of 50 psi, the pores of 0.2 and 0.5 mm restricted the air flow through the disks to maximum values of 10 L min1. To minimize fluctuations in the concentration of test particles delivered to the LSA via the peristaltic pump, the test particle suspension inside the particle reservoir was continuously mixed with a magnetic stirrer. When performing comparison experiments, the Collison nebulizer was operated at a constant pressure of 50 psi for all conditions and flowrates of QCOLL ¼ 2, 5, and 10 L min1. Flow rates higher than 10 L min1 were not achieved at a pressure of 50 psi and the flow rates less than 10 L min1 were set using an adjustable flow meter. A standard Collison nebulizer recycles 20 mL of suspension approximately every 6 s thus constantly mixing the suspension (May, 1973). Therefore, no additional mixing of the Collison nebulizer suspension was performed. The aerosolization flow rates for both the LSA and the Collison nebulizer were determined with the Buck calibrator (AP Buck, Inc., Orlando, FL). All experiments were conducted in a Class 2 Biological Safety Cabinet (Class II Type A/B3, NUAIRE, Inc., Plymouth, MN).
3525
Both the C INI and the C POST were determined using a light microscope at 400 magnification (Axioscope 20, Carl Zeiss, Inc., Thornwood, NY) and a 0.02 mm deep counting chamber (Hausser Scientific Partnership, Horsham, PA). Using these data, we also determined the percentage of liquid loss DV LIQ for each experimental condition. DV LIQ indicates what fraction of liquid (purified water in our case) delivered to the LSA was lost as vapor or liquid droplets during operation of the LSA. DV LIQ was calculated as follows: DV LIQ ¼ 1
V POST N POST V PSL , QL t N INI V PSL
(5)
where V PSL is the volume of a single PSL particle. 2.5. Aerosolization effect on bacterial viability
The N INI is the product of the PSL particle concentration in the test suspension, C INI , the liquid delivery flow rate, QL , and the time, t, for which the experiment was performed:
Analysis of microbial samplers and other bioaerosol studies sometimes require continuous microorganism aerosolization for prolonged periods of time. Stress introduced by continuous aerosolization and other factors may significantly affect microorganism viability during such experiments, a factor that may compromise study outcome. Thus, it is understandable that several studies analyzing the effect of aerosolization on bacterial viability have concentrated on negative effects accumulated over a certain time. Griffiths et al. (1996) investigated the effects of residence time in the spray and collection liquid on bioaerosol viability. Reponen et al. (1997) indicated that the metabolic injury to P. fluorescens dispersed with Collison nebulizer increases with aerosolization time. Therefore, in our preliminary analysis of the LSA’s effect on bacterial viability, we concentrated on viability change over time. One of the ways to investigate such change would be to collect airborne microorganisms and determine their relative recovery. However, the collection of already aerosolized microorganisms by commonly used methods, such as impaction, is known to affect microorganism viability by itself (Stewart et al., 1995) which may obscure the effect of aerosolization on viability. Thus, for both the LSA and the Collison nebulizer, we determined the concentration of culturable bacteria in the suspension, C CFU , and the concentration of aerosolized bacteria C AIR , at the beginning of the aerosolization ðt ¼ 0 minÞ and at the end of aerosolization ðt ¼ 90 minÞ. Change in C CFU =C AIR ratio over time indicates change in bacterial culturability, F CULT , over time:
N INI ¼ C INI QL t.
F CULT
2.4. Determination of the LSA generation efficiency The LSA generation efficiency, E G , was determined by comparing the number of PSL particles, N INI , delivered to the generator over a certain period of time, with the number of particles not aerosolized by the device and accumulated at the bottom of its glass vessel, N POST . The generation efficiency was defined as follows: EG ¼ 1
N POST . N INI
(2)
(3)
The total number of particles accumulated in the LSA at the end of the aerosolization, N POST , is a product of the accumulated liquid volume, V POST , and the particle concentration, C POST : N POST ¼ V POST C POST .
(4)
¼
C CFU =C AIR ðt ¼ 90Þ C CFU =C AIR ðt ¼ 0Þ 100%. C CFU =C AIR ðt ¼ 0Þ
ð6Þ The C CFU was determined by diluting and plating in triplicate the bacterial suspensions from each generator
ARTICLE IN PRESS 3526
G. Mainelis et al. / Atmospheric Environment 39 (2005) 3521–3533
on Petri dishes with trypticase soy agar, incubating the plates at 26 1C for 24 h, and then counting formed colonies. The concentration of bacteria in the air was determined using the optical particle counter (Grimm Technologies Inc.) as shown in Fig. 2. Overall, each test analyzing the LSA performance at different conditions was run at least three times and the data discussed below are presented by average values and standard deviations of those measurements.
3. Results and discussion 3.1. The effect of disk porosity on total particle output Fig. 3 summarizes the effect of disk pore size, DP , and aerosolization air flow, QLSA , on the total particle output by the LSA. A 0.99 mm PSL suspension of constant concentration was delivered to the LSA at a rate of QL ¼ 0:6 mL min1 for all experiments presented in this figure. For the used air pressure of 50 psi, the maximum aerosolization air flow for the 0.2 and 0.5 mm disks was limited to 10 L min1 due to the limited porosity of the disks. For air flows of QLSA ¼ 2, 5, and 10 L min1, the use of the disk with 0.5 mm pores resulted in the highest airborne particle concentration. For both 0.2 and 0.5 mm disks, the PSL particle output increased almost linearly with the increasing aerosolization airflow. It appears that the application of higher aerosolization flow rates would further increase aerosolized
particle concentration. However, as seen from the data obtained with 2.0 and 10.0 mm disks, the aerosolized particle concentration increased with increasing air flow, then leveled off and even decreased when the aerosolization flow rate was increased to 30 L min1. The 2.0 mm disk achieved a higher particle output than the 10 mm disk at air flows of 10 and 20 L min1, though its particle output was always lower than that of the 0.5 mm disk at flow rates of 2–10 L min1. For the 2 mm disk, the particle output at 20 L min1 was almost the same as that of 0.5 mm disk at 10 L min1. A possible explanation for the decreasing particle output at higher aerosolization flow rates is that high airflow and increased diameter of air streams emanating from the disk (due to larger pore size) disrupt the liquid film and fling it upwards and away from the disk before orderly bubbling can take place. Decrease and elimination of the bursting bubble action then lowers the total particle output of the LSA. Overall, it seemed that air sparging into the suspension film through the 0.5 mm disk created the best conditions for particle aerosolization at a particular aerosolization air flow; however, a very similar particle output was achieved with the 2 mm disk when the aerosolization flow rate was roughly doubled. 3.2. The effect of liquid delivery rate on total particle output The effect of liquid delivery rate on the total particle output of the LSA is shown in Fig. 4. The experiments in
Fig. 3. Effect of disk pore size on 0.99 mm PSL particle output at different aerosolization air flows, QLSA . PSL particle concentration in the liquid suspension was constant for all experiments and the suspension was delivered at a rate of QL ¼ 0:6 mL min1. Symbols represent averages of three repeats and error bars represent standard deviations. The data have been fitted with second- and third-order polynomials. For all curves, R2 40:97.
ARTICLE IN PRESS G. Mainelis et al. / Atmospheric Environment 39 (2005) 3521–3533
3527
Fig. 4. Effect of liquid delivery flow, QL , on the total 0.99 mm PSL particle output at different aerosolization air flows, QLSA . A disk with a 2 mm pore size was used in these experiments and the PSL particle concentration in the liquid suspension was constant for all experiments. Symbols represent averages of three repeats and error bars represent standard deviations. The data have been fitted with third-order polynomials. For all curves, R2 40:98.
this figure were performed using a disk with 2.0 mm pore size and with aerosolization air flows of QLSA ¼ 2, 5, 10, 20, and 30 L min1. The steepest increase in total particle output was between aerosolization airflows of 2 and 10 L min1 for all of the liquid delivery rates. When the aerosolization air flow QLSA was further increased from 10 to 20 L min1, the total particle output for the liquid delivery rates QL ¼ 0:6 and 2.0 mL min1 increased further, although less steeply. Also, when the aerosolization air flow for QL ¼ 0:6 and 2.0 mL min1 was increased to 30 L min1, the total particle output decreased to approximately the level observed at QLSA ¼ 10 L min1. For the lowest liquid delivery rate, QL ¼ 0:17 mL min1, the particle output decreased when the aerosolization flow rate was higher than 10 L min1. It can also be observed that for aerosolization flows of 10 L min1 and higher, increase in the liquid delivery rate resulted in higher particle output. Overall, it seems that each liquid delivery rate has an optimum aerosolization air flow: for the lowest investigated delivery rate of QL ¼ 0:17 mL min1, the optimal QLSA ¼ 10 L min1, while for higher QL of 0.6 and 2.0 mL min1 an aerosolization flow QLSA of 20 L min1 delivered the greatest total particle output. When the air flow is decreased from the optimal setting, the porous disk is overwhelmed with liquid, resulting in a thicker suspension film on the disk. The air streams emanating from the disk break up the suspension film less efficiently which results in reduction of the
total particle output. At the optimal setting, the bursting bubble action is stable and continuous. When the aerosolization air flow is increased above the optimal setting, the air streams overwhelm the force of gravity of the suspension film, rip it apart and fling the larger suspension fragments up and to the side without a steady bubbling process taking place, a condition that reduces particle output. This experiment illustrates that a balance between the liquid delivery rate and the aerosolization air flow results in the highest particle output. Each disk with different pore size is expected to have its own optimal settings. Such operating parameter settings can be estimated qualitatively by achieving circumstances that result in a steady particle output and produce a continuous and steady suspension ‘‘sizzle’’ on the disk. 3.3. A comparison of the total particle output by the Liquid Sparging Aerosolizer and the Collison nebulizer Fig. 5 presents a comparison of the total particle output by the LSA and the Collison nebulizer for 0.99, 2.93, and 5.09 mm PSL particles. The LSA liquid delivery flow for all the experiments was QL ¼ 0:6 mL min1. Each sub-figure uses a PSL suspension of constant concentration, though concentrations for different PSL sizes are not related. At our experimental setting, the Collison nebulizer had a maximum air flow below 20 L min1; thus, the highest air flow selected for experimentation was 10 L min1. For 0.99 mm PSL
ARTICLE IN PRESS 3528
G. Mainelis et al. / Atmospheric Environment 39 (2005) 3521–3533
Fig. 5. Comparison of total particle output between the Liquid Sparging Aerosolizer (LSA) and the Collison nebulizer when operating at different aerosolization flow rates and generating PSL particles of three different sizes: (A) 0.99 mm, (B) 2.93 mm, and (C) 5.09 mm. The LSA was operated with disks of two different porosities: 0.5 and 2.0 mm. Concentrations of PSL particles of each particular size were kept constant throughout the experiments and the LSA liquid delivery rate, QL , was 0.6 mL min1. Symbols represents averages of three repeats and error bars represent standard deviations. The data have been fitted with second- and third-order polynomials. For all curves, but one, R2 40:98. For the curve representing 5.09 mm particles aerosolized by the LSA with 2 mm disk, R2 ¼ 0:84.
particles, the Collison nebulizer produced a much greater particle output than the LSA, especially at 10 L min1. The difference decreased somewhat when the PSL size was increased to 2.93 mm. When aerosolizing 5.09 mm PSL, the use of the LSA with 2.0 mm poresize disk and aerosolization flow rates 10 L min1 and higher resulted in higher total particle output compared with the Collison nebulizer. It is anticipated that the 2.0 mm disk could produce an even higher total particle output at larger flows if the liquid delivery rate is increased, given the data from Fig. 3. Fig. 5 also shows that the LSA operated more efficiently with a 0.5 mm disk when aerosolizing 0.99 mm particles, but had a higher output with a 2.0 mm disk for 2.93 and 5.09 mm particles. To generalize this behavior, as the particle size increases, the size of the disk pores should also increase for optimal total particle output. When the disk pore size increases, so does the diameter of the multiple air streams emanating from the disk. Larger airstreams moving at higher velocities (increases with increasing QLSA ) are better able to pick up larger PSL particles from the suspension film and carry them away toward generator outlet. The ability of the Collison nebulizer to aerosolize larger particles is limited by the average size of droplets dispersed from the nozzle. May (1973)
indicated that the number frequency modal values when generating aqueous solution of fluorescein ranged from 1.8 to 2.7 mm when the Collison nebulizer operated at air pressures from 15 to 40 psi. The modal value could be somewhat higher for the 50 psi pressure used in our experiments; however, it still would be significantly below 5 mm. In general, the Collison nebulizer produces higher concentrations of smaller particles, but the LSA can be effectively used to aerosolize larger solid particles if the disk pore size and the aerosolization flow rate are properly selected. 3.4. Aerosol generation efficiency of the Liquid Sparging Aerosolizer The primary goal of this experiment was to investigate what fraction of the particles delivered to the LSA was actually aerosolized under different operating parameters. We also calculated the fraction of liquid (purified water in our case) lost as water vapor or aerosolized liquid mass. For this experiment, we selected disks with two pore sizes that showed the best results— 0.5 and 2.0 mm—and the extreme tested values of liquid delivery (0.17 and 2.0 mL). Aerosolization air flows of 5 and 10 L min1 were selected for the 0.5 mm disk, and
ARTICLE IN PRESS
4371 471 6271 771 4171 271 8571 1271 8471 6972 9471 7872 9371 8771 9171 3972 60 10 60 10 60 10 120 60 The data presented in this table are averages of three repeats7one standard deviation.
5.8 19.2 3.9 18.5 6.0 19.5 3.0 104.7 1.4370.02 5.470.17 0.5170.13 3.9170.30 0.6170.04 2.2570.16 1.6670.17 64.772.43 8.9570.59 17.571.16 8.9570.59 17.571.16 8.9570.59 17.570.12 17.970.12 105.376.94 5 5 10 10 5 5 30 30
0.2 2 0.2 2 0.2 2 0.2 2 0.5 0.5 0.5 0.5 2 2 2 2
10.15 19.91 10.15 19.91 10.15 19.91 20.31 119.47
Number of non-aerosolized particles accumulated in the LSA after aerosolization, NPOST ( 1010) Volume of Suspension delivered into the LSA, VINI (mL) (70.1) Number of particles delivered into the LSA, NINI ( 1010)
Fig. 7 presents the generation stability of the LSA when generating polydisperse NaCl particles. For these experiments, the LSA was fitted with a 0.5 mm disk, a
3529
Disk pore Aerosolization Liquid air flow, delivery size, DP (mm) QLSA (L min1) rate, QL (mL min1)
3.5. Generation stability over time
Table 1 Generation efficiency of the Liquid Sparging Aerosolizer (LSA)
flows of 5 and 30 L min1 were selected for the 2 mm disk. These values were selected to test what generation efficiency values could be expected even when the total particle output was not optimum, as shown for 2 mm disk in Fig. 4. All experiments here were performed with 0.99 mm PSL particles of the same concentration and the experimental details and results are presented in Table 1. Fig. 6 visualizes the generation efficiency, E G , of the LSA at different operational parameters. For the conditions investigated, the generation efficiency ranged from E G ¼ 39% to 94%. The highest efficiency occurred when using the 0.5 mm disk at an air flow of QLSA ¼ 10 L min1 and liquid delivery flow of QL ¼ 0:17 mL min1. The lowest observed efficiency occurred when using the 2.0 mm disk at QLSA ¼ 30 L min1 and QL ¼ 2:0 mL min1. The generation efficiency did not necessarily correlate with the produced aerosol output. In fact, the data show that for each set disk pore size and aerosolization flow rate, the aerosolization efficiency decreased when liquid delivery increased from 0.17 to 2 mL min1. As was shown in Fig. 4, an increase in liquid delivery rate increases the aerosol output. One can also observe that for the 0.5 mm disk, an increase in aerosolization flow rate increases the aerosolization efficiency when the liquid delivery rate stays constant. This trend for the 2 mm disk is reversed. Most likely it has something to do with the fact that for 2.0 mm disk, use of 30 L min1 results in less-efficient bubbling process, as shown in Fig. 4. Overall, the data show that for settings in which LSA is producing stable, continuously bursting bubbles, the aerosol efficiency is between 80% and 95%, i.e. a significant majority of particles delivered to LSA become airborne and do not accumulate at bottom of the glass vessel. The high particle generation efficiency numbers can be contrasted with lower liquid mass ‘‘generation efficiency’’ (carrier liquid loss) numbers as shown in Table 1. Such numbers mean that the LSA much more readily aerosolizes solid particles than the liquid, in which the particles are suspended. This is especially true when the suspension is delivered at a rate of 2 mL min1. When the suspension was delivered at a rate of 0.17 mL min1, the liquid loss rate varied from 85% for 2 mm disk and airflow of 30 L min1 to the range of 41–61% for other conditions. The higher liquid loss for the 2 mm disk and airflow of 30 L min1 is due to high desiccating capacity of 30 L min1 air flow and low liquid delivery rate of 0.17 mL min1, which increases the percentage of liquid that can be desiccated. As could be expected, higher aerosolization air flows resulted in more significant liquid loss.
Liquid loss during Aerosolization Generation Volume of liquid aeroso-lization, time, t (min) efficiency, E G , % collected in the EG ¼ 1NINI/NPOST DVLIQ (%) LSA after aerosolization, VPOST (mL) (70.1)
G. Mainelis et al. / Atmospheric Environment 39 (2005) 3521–3533
ARTICLE IN PRESS 3530
G. Mainelis et al. / Atmospheric Environment 39 (2005) 3521–3533
Fig. 6. Generation efficiency of Liquid Sparging Aerosolizer for different operational parameters. Symbols represent averages of three repeats and error bars represent standard deviations.
Fig. 7. Generation stability of the Liquid Sparging Aerosolizer using a polydisperse NaCl suspension of constant concentration. The LSA operated with a 0.5 mm disk, a liquid delivery flow of QL ¼ 0:6 mL min1 and an aerosolization air flow of QLSA ¼ 10 L min1. The indicated sizes represent average optical diameters of the aerosolized NaCl particles.
liquid delivery flow of QL ¼ 0:6 mL min1 and an aerosolization air flow of QLSA ¼ 10 L min1. As could be seen from the NaCl data, the concentration of all size fractions was stable during the period of 180 min of continuous generation. The standard deviation of total particle output for all size fractions was below 12%. Some variation in particle output may be due to variation in the aerosolization air flow which was caused by air pressure fluctuations in the air delivery system. Overall, the data presented here indicated that the LSA
is suitable for generating stable particle concentrations over extended periods of time. This may be very useful for extended exposure studies or evaluation of long-term efficiency of particle collectors. 3.6. The effect of the LSA on the stability and viability of a microorganism culture Bioaerosol generation using a bursting bubble principle is believed to be less damaging to sensitive
ARTICLE IN PRESS G. Mainelis et al. / Atmospheric Environment 39 (2005) 3521–3533
3531
Fig. 8. Changes in bacterial size distribution over time when aerosolizing with (A) Collison nebulizer, and (B) Liquid Sparging Aerosolizer. Operational conditions for Collison nebulizer: air pressure ¼ 50 psi, air flow QCOLL ¼ 5 L min1; operational conditions for LSA: disk pore size ¼ 0.5 mm, liquid supply rate QL ¼ 0:17 mL min1, air flow QLSA ¼ 5 L min1. Symbols represent averages of three repeats and error bars represent standard deviations.
microorganisms. In our preliminary analysis of the LSA’s effect on the microorganism stability and viability, we aerosolized sensitive P. fluorescens bacteria for 90 min using the LSA and the Collison nebulizer and then observed changes in bacterial size distribution and viability as described in the Materials and Methods section. In this experiment, the LSA was fitted with a 0.5 mm disk and operated at a liquid delivery rate of 0.17 mL min1 and an aerosolization air flow of 5 L min1. The Collison nebulizer was operated at an air flow of 5 L min1 and a pressure of 50 psi. As can be seen from Fig. 8, at the beginning of aerosolization, both generators produce bacteria with the number frequency mode of approximately 0.6 mm (location of particle peak). When aerosolizing bacteria, particles less than 0.5 mm in size are considered fragments (Mainelis et al., 2001). As could be expected from previous results (Fig. 5), the Collison nebulizer yielded a higher particle output. After a continuous 90 min generation period, the number of bacteria measured at 0.6 mm range doubled. Most likely this is due to the loss of liquid which concentrates the bacterial suspension inside the Collison nebulizer and also any background lysis from adding the bacteria to water. More significantly, the concentration of fragments emitted from the Collison nebulizer has increased about 3–3.5 times after continuous generation for 90 min. We believe that this is due to constant recycling of the suspension, which repeatedly subjects the bacteria to strong shear forces causing them to fragment. A similar fragmentation effect was observed by Mainelis et al. (2001) who also noted that fragmentation of cells caused an increase in a suspension’s electrical conductivity.
In contrast, the size distribution of P. fluorescens bacteria aerosolized with the LSA was virtually identical after 90 min of continuous generation. The difference in bacterial concentration was not statistically significant. The slight increase in the concentration of bacterial fragments could be attributed to natural lysis of the cells suspended in purified water without any nutrients. The change in bacterial viability after continuous aerosolization for 90 min is presented in Table 2. As could be seen, the viability of P. fluorescens bacteria aerosolized with the LSA on average has not changed, while the viability of the same bacteria aerosolized with the Collison nebulizer has decreased by over 50% after 90 min of aerosolization. The experiments performed here indicate that the LSA can retain the viability of sensitive microorganisms over extended aerosolization periods while maintaining constant bioaerosol concentration output. This feature should be useful in various bioaerosol investigations. More detailed analysis of effects on microorganism viability will be addressed in future studies. The LSA design described herein is based on a modified BGI Collison nebulizer. It has been further modified to allow for current production of BGI Collison nebulizers to be retrofitted and made into an LSA with a smaller ((3/4) in) sparging surface. The conversion process involves adding the sparging surface holder and reversing the path of the liquid and gaseous feed. Currently, a prototype of this adaptation is scheduled to be evaluated. In a second new prototype, the LSA was fitted with an enlarged (90 mm) sparging surface featuring pore size of 2 mm. This large unit, designed to operate at air flows of up to 100 L min1
ARTICLE IN PRESS G. Mainelis et al. / Atmospheric Environment 39 (2005) 3521–3533
3532
Table 2 Change in viability of Pseudomonas fluorescens bacteria over time when aerosolized with Collison nebulizer and the Liquid Sparging Aerosolizer as measured by plating Collison nebulizer
Liquid Sparging Aerosolizer
t ¼ 0 min Concentration of viable bacteria in the liquid suspension, CCFU ¼ NCFU/mLLIQ Concentration of total aerosolized bacteria, CAIR ¼ NTOT/LAIR Ratio of viable liquidborne vs. total airborne bacterial concentration, CCFU/CAIR Change in CCFU/CAIR over time, %
t ¼ 90 min
t ¼ 0 min
t ¼ 90 min
1.53 10 71.20 10
1.37 10 73.54 10
1.04 10 78.9 10
1.16 10973.11 107
1.77 10575.37 103
3.26 10579.22 103
5.67 10473.22 103
6.35 10474.96 103
8.6 10373.14 102
4.19 10371.22 102
1.83 10471.17 103
1.83 10471.44 103
9
—
8
9
7
51.374.3
9
—
7
0710.2
The data presented in this table are averages of three measurements7one standard deviation.
may find applications in large-scale simulations of aerosol release. Evaluation of this prototype is also ongoing.
4. Conclusions This study presented and evaluated a new aerosol generator that utilizes a bursting bubble principle and eliminates reuse of the particle suspension, thus creating a single-pass generator suitable for microorganism aerosolization. The obtained data demonstrated that this generator yields a stable particle output when aerosolizing polydisperse and monodisperse particles over extended periods of time. Our tests have shown that the use of disks with pore sizes of 0.5 and 2.0 mm resulted in the highest particle output and that the produced aerosol concentration depends on the air flow through the disk as well as the liquid delivery rate. Using these parameters, one can easily adjust the particle concentration and flow rate needed for a specific experiment. For most of the tested conditions, more than 70% of particles delivered to the LSA were aerosolized. At certain specific conditions, this percentage surpassed 90%. Preliminary experiments with a sensitive microorganism P. fluorescens showed that the bacterial distribution virtually did not change during 90 min of continuous aerosolization using the LSA. In fact, on average there was no (0%) viability loss, whereas the bacterial spectrum produced by a Collison nebulizer changed significantly over time and there was a 50% loss in viability after 90 min. This result demonstrates that low shear forces of bursting bubbles and absence of
suspension recycling maintain microorganism viability over prolonged generation. More exhaustive analysis of microorganism aerosol generation by the LSA will be presented in future studies. Overall, we believe that this new instrument could be used to generate bioaerosol for the evaluation of pathogen collection methods, inhalation and other studies where extended delivery of stable and undamaged biological aerosols is required.
Acknowledgments This study was supported by the Environmental and Occupational Health Sciences Institute (Piscataway, NJ, USA) and the CH Technologies Inc. (Westwood, NJ, USA). The authors are thankful for this support. The authors also appreciate the provision of stainless steel disks by the Mott Corporation (CT, USA).
References Chen, S.-K., Vesley, D., Brosseau, L.M., Vincent, J.H., 1994. Evaluation of single-use masks and respirators for protection of health care workers against mycobacterial aerosols. American Journal of Infection Control 22 (2), 65–74. Dennis, J.H., Stenton, S.C., Beach, J.R., Avery, A.J., Walter, E.H., Hendrick, D.J., 1990. Jet and ultrasonic nebulizer output: use of a new method for direct measurement of aerosol output. Thorax 45 (10), 728–732. Go´rny, R.L., Dutkiewicz, J., 1998. Evaluation of microorganisms and endotoxin levels of indoor air in living rooms occupied by cigarette smokers and non-smokers in Sosnowiec, Upper Silesia, Poland. Aerobiologia 14, 235–239.
ARTICLE IN PRESS G. Mainelis et al. / Atmospheric Environment 39 (2005) 3521–3533 Griffiths, W.D., DeCosemo, G.A.L., 1994. The assessment of bioaerosols: a critical review. Journal of Aerosol Science 25 (8), 1425–1458. Griffiths, W.D., Stewart, I.W., Reading, A.R., Futter, S.J., 1996. Effect of aerosolization, growth phase and residence time in spray and collection fluids on the culturability of cells and spores. Journal of Aerosol Science 27 (5), 803–820. Israeli, E., 1973. Effect of aerosolization and lyophilization on macromolecular synthesis in E. coli. VIth International Symposium on Aerobiology, Enschede, the Netherlands. Wiley, New York, Toronto. Jensen, P.A., Todd, W.F., Davis, G.N., Scarpino, P.V., 1992. Evaluation of eight bioaerosol samplers challenged with aerosols of free bacteria. American Industrial Hygiene Association Journal 53 (10), 660–667. Mainelis, G., Willeke, K., Baron, P., Grinshpun, S.A., Reponen, T., Go´rny, R.L., Trakumas, S., 2001. Electrical charges on airborne microorganisms. Journal of Aerosol Science 32, 1087–1110. May, K.R., 1973. The collison nebulizer: description, performance and application. Journal of Aerosol Science 4, 235–243.
3533
Palleroni, N.J., 1984. Bergey’s Manual of Systematic Bacteriology, vol. 1. Williams and Wilkins Co., Baltimore, p. 165. Reponen, T., Willeke, K., Ulevicius, V., Grinshpun, S.A., Donnelly, J., 1997. Techniques for dispersion of microorganisms into air. Aerosol Science and Technology 27, 405–421. Saint-Louis, R., Pelletier, E., 2004. Sea-to-air flux of contaminants via bubbles bursting. An experimental approach for tributyltin. Marine Chemistry 84 (3/4), 211–225. Stewart, S.L., Grinshpun, S.A., Willeke, K., Terzieva, S., Ulevicius, V., Donnelly, J., 1995. Effect of impact stress on microbial recovery on an Agar surface. Applied and Environmental Microbiology 61 (4), 1232–1239. Ulevicius, V., Willeke, K., Grinshpun, S.A., Donnelly, J., Lin, X., Mainelis, G., 1997. Aerosolization of particles from a bubbling liquid: characteristics and generator development. Aerosol Science and Technology 26 (2), 175–190. White, E., Lucke, W.E., 2003. Effects of fluid composition on mist composition. Applied Occupational & Environmental Hygiene 18 (11), 838–842.
Atmospheric Environment 39 (2005) 3521–3533 www.elsevier.com/locate/atmosenv
Design and performance of a single-pass bubbling bioaerosol generator Gediminas Mainelisa,, David Berrya, Hey Reoun Ana, Maosheng Yaoa, Kevin DeVoeb, Donna E. Fennella, Rudolph Jaegerc a
Rutgers University, Department of Environmental Sciences, 14 College Farm Rd., New Brunswick, NJ 08901, USA b BGI Inc., 58 Guinan Street, Waltham, MA 02451, USA c CH Technologies, Inc., 263 Center Avenue, Westwood, NJ 07675, USA Received 31 August 2004; accepted 8 February 2005
Abstract We describe and analyze a new particle generator that utilizes a bursting bubble principle and eliminates carrier fluid reuse. In this Liquid Sparging Aerosolizer (LSA), a suspension of particles or microorganisms is pumped at a flow rate of 0.2–2 mL min1 to the top surface of a porous stainless-steel disk where it forms a thin suspension film. Filtered air is then sparged through the disk into the film causing it to break into bubbles that subsequently burst, releasing particles into the air. The released particles are then captured by the sparging air stream and are carried away. Particles that impinge the glass vessel and liquid droplets not captured by the air stream drain to the bottom of the vessel and play no further role in the aerosolization process. We tested the LSA with disks of different pore sizes (0.2, 0.5, 2.0 and 10.0 mm) and different air flows (2–30 L min1) through the porous disks while generating polydisperse and monodisperse particles. Our tests showed that the use of 0.5 and 2.0 mm porosity disks resulted in the highest output of PSL particles in the desired size range, i.e., comparable to bacterial size. Each pore size seemed to have an optimal air flow rate; the produced aerosol concentration increased with increasing suspension delivery rate. The LSA also demonstrated stability of output concentration when aerosolizing particles over extended periods of time. In addition, the size distribution of injury-sensitive Pseudomonas fluorescens bacteria virtually did not change during 90 min of continuous aerosolization by the LSA. In fact, there was no (0%) viability loss, whereas the bacterial spectrum produced by a Collison nebulizer changed significantly over 90 min and there was a 50% loss in viability. The results indicate that the new instrument could be used to generate particles for the evaluation of pathogen collection methods, inhalation and other studies where extended delivery of stable and undamaged biological aerosols is required. r 2005 Elsevier Ltd. All rights reserved. Keywords: Microorganism aerosolization; Microorganism viability; Bursting bubble
1. Introduction
Corresponding author. Tel.: +1 732 932 7166;
fax: +1 732 932 8644. E-mail address: [email protected] (G. Mainelis).
Inhalation studies, instrument calibration, evaluation of pathogen collection methods and other aerosol investigations require stable and reliable aerosol and bioaerosol generators. Reliable generation of biological
1352-2310/$ - see front matter r 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.atmosenv.2005.02.043
ARTICLE IN PRESS 3522
G. Mainelis et al. / Atmospheric Environment 39 (2005) 3521–3533
aerosols is also needed to study transport and deposition of indoor bacteria and mold. Among several methods used to aerosolize microorganisms, pneumatic nebulization is probably the most commonly used tool (Chen et al., 1994; Jensen et al., 1992); however, nebulizing generators produce high shear forces that may affect microorganism viability (Griffiths and DeCosemo, 1994). Commercially available nebulizers, such as the Collison nebulizer, are able to produce high concentrations of aerosol, but have been suspected of injuring and fragmenting microorganisms (Reponen et al., 1997). Griffiths et al. (1996) used a glass atomizer to aerosolize Escherichia coli cells and Penicillium expansum spores and found that aerosolization reduced the culturable fraction of E. coli to near zero. Israeli (1973) also indicated that bacteria submitted to aerosolization lose their ability to form colonies on nutrient media. In addition, in a standard Collison nebulizer with 20 mL of water in the flask, most of the liquid solution is recirculated about every 6 s (May, 1973). Thus, because the suspension is ‘‘recycled’’ the bacteria are repeatedly subjected to shear force stress. It has also been shown that nebulizers produce large concentrations of aerosol in the 0.1–0.5 mm size range, an unimportant range for bio-aerosol generation (Dennis et al., 1990). When aerosolized microorganisms are used to evaluate the efficiency of various bioaerosol collection and control methods, preservation of the organisms’ structural and biological integrity is highly desired. Thus, there is a need for a bioaerosol generator that can not only efficiently produce a stable bioaerosol output, but also minimize injury and death to sensitive microorganisms. Ulevicius et al. (1997) proposed a generator in which particles, suspended in a liquid, are produced from a bubbling liquid. Injection of tangentially oriented drying jets dries the airborne droplets and carries the particles away from the bubbling source and out. The bursting bubble action has also been shown to produce a higher aerosol concentration than nebulization for the production of short-chain fatty acid mists (White and Lucke, 2003). The bubbling mechanism has been studied as a naturally occurring phenomenon and has been recognized as a significant factor in aerosolization of seawater and suspended contaminants from breaking waves (Saint-Louis and Pelletier, 2004). The generator proposed by Ulevicius et al. (1997) was shown to effectively aerosolize polystyrene latex (PSL) particles of 0.71–5.1 mm in size. In addition, it introduced a lower metabolic injury to Pseudomonas fluorescens bacteria compared to Collison nebulizer (Reponen et al., 1997). The microorganisms aerosolized using the generator were found to carry fewer electrical charges compared to those generated by Collison nebulizer (Mainelis et al., 2001), which indicates that liquid disruption is not as violent as in pneumatic nebulizers. However, even in this generator a certain fraction of the suspension was still
being recycled. It was also shown that the particle output depends on the distance between the drying jets and the liquid surface. Thus, as the particles and the liquid are being dispersed at a rate of approximately 8.7 g h1, the level of the suspension inside the generator is lowered, a process that increases the distance between the drying jets and the liquid surface resulting in a lower aerosol concentration. Also, the described two-flow generator requires two separate air flows—one for bubbling the liquid and one for drying the droplets—a requirement that may complicate the experimental setup. To overcome this complication, Ulevicius et al. (1997) also suggested a single-flow generator in which the bubbling air flow and the drying air flow are combined into one. However, the presented single-flow design was expected to have a lower utility than the twoflow design and was only minimally investigated. In the research described in this paper, we present and investigate a new aerosol and bioaerosol generator that not only utilizes the bubbling liquid principle to induce low shear stress to microorganisms, but also features no suspension ‘‘recycling’’ and a simple design made possible by the use of the same air flow for both bubbling and drying air. Use of the single-pass design was expected to minimize the output of fragments and non-viable microorganisms when producing biological aerosols. In our proposed Liquid Sparging Aerosolizer (LSA), a suspension of particles or microorganisms is pumped at a certain flow rate to the top surface of a porous stainless-steel disk where it forms a liquid film. The air is then sparged into the film through pores in the disk causing the liquid film to break into bubbles that subsequently burst, releasing particles into the air. The released particles are then captured by the same air stream and are carried out of the device. Particles and liquid not captured by the air stream coalesce, drain to the bottom of the vessel and play no further role in the aerosolization process. The primary goal of this investigation was to analyze the performance of the LSA as a function of the disk pore size, the suspension delivery rate, and the sparging (aerosolization) air flow rate. A separate set of experiments was performed to investigate the generation efficiency and stability as well as the possible effect of the proposed aerosolization device on sensitive microorganisms.
2. Materials and methods 2.1. Design of the Liquid Sparging Aerosolizer The LSA features a bursting bubble design that minimizes shear forces on particles being aerosolized (Fig. 1). The particle suspension used to generate an aerosol is delivered to the device via a peristaltic pump operating at liquid delivery rate, QL , from 0.17 to
ARTICLE IN PRESS G. Mainelis et al. / Atmospheric Environment 39 (2005) 3521–3533
3523
Fig. 1. Schematic representation of the Liquid Sparging Aerosolizer (LSA).
2 mL min1 (a syringe can be used instead of a peristaltic pump). The suspension travels downward through a narrow vertical needle and is dripped onto a 1 in (2.54 cm) porous stainless-steel disk (Mott Corporation, Farmington, CT) where it forms a thin liquid film. The disk is approximately 1 cm below the needle outlet. The aerosolization air flow, QLSA , is delivered under pressure to the underside of the porous disk via a separate airtight channel. The air is then forced by pressure upwards through the disk’s pores into the liquid (air sparges into the liquid). The multiple air streams emanating from the porous disk break the suspension film into bubbles within which particles are contained. An expanding pressure gradient between the inside and outside of the bubbles causes them to burst, aerosolizing the particles.
Visually, the process resembles boiling of fine bubbles. The same multiple air streams ðQLSA Þ then contribute to the drying of the airborne particles and liquid filaments, carrying the resulting particles and liquid residues out the exit port. Particles and liquid not captured by the QLSA air stream (as shown later they constitute a small fraction) are collected at the bottom of the glass vessel and play no further role in the aerosolization process. Thus, sensitive particles, such as microorganisms, participate in the generation process only once and there are no strong shear forces involved. Therefore, this generator should impart less damage to fragile elements compared to traditional pneumatic nebulization. The dimensions of the glass vessel are the same as those of the Collison nebulizer (Model MRE CN24/25, BGI Inc.,
ARTICLE IN PRESS 3524
G. Mainelis et al. / Atmospheric Environment 39 (2005) 3521–3533
Waltham, MA). The 1-in diameter porous disk facilitating the bubbling action can easily and quickly be replaced with other 1-in disks featuring different pore sizes. The liquid delivery flow rate, QL , and the aerosolization flow rate, QLSA , can also be easily adjusted for optimal performance. The effect of these parameters on the LSA performance was investigated in our research described below. 2.2. Test particles The performance of the LSA was investigated with monodisperse PSL particles (Bangs Laboratories, Fishers, IN) of three different sizes: 0.99, 2.93, and 5.09 mm in diameter. PSL suspensions were prepared by suspending aliquots of PSL particle stock using filtered (Milli-Q system, Millipore, Billerica, MA) water. Generation stability experiments were performed with polydisperse NaCl particles prepared by dissolving 1 g of reagent quality NaCl (Mallinckrodt Baker, Inc., Phillipsburg, NJ) into 500 mL of purified water. The effect of the LSA on viability and size distribution of sensitive bacteria was investigated using P. fluorescens vegetative cells. The rod-shaped gram-negative P. fluorescens bacteria are 0.7–0.8 mm in diameter and 1.5–3.0 mm in length (Palleroni, 1984) and are commonly found in ambient air (Go´rny and Dutkiewicz, 1998). P. fluorescens vegetative cells were grown in tryptic soy agar (Becton Dickinson Microbiological System, Sparks, MD) at 26 1C for 18 h. The bacteria were separated from the broth by centrifuging three times at 7000 rpm (BR-4, Jouan, Winchester, VA) for 7 min each time. The bacterial suspension was prepared by suspending the pellet in freshly filtered (Milli-Q system,
Millipore, Billerica, MA) water which had been sterilized. 2.3. Test system and test parameters The test system used in this research (Fig. 2) was designed to test the performance of the LSA under different experimental conditions. For certain experimental conditions, the LSA performance was compared to that of the Collison nebulizer (BGI Inc., Waltham, MA). Particles produced by the LSA and the Collison nebulizer at aerosolization flowrates of QLSA and QCOLL , respectively, entered a HEPA-filtered dry air flow, QDRY ¼ 50 L min1 , and were carried into the measurement chamber. The concentration, C AIR , and size of the test particles in the measurement chamber were isokinetically measured using an optical particle counter (Model 1.108, Grimm Technologies Inc., Douglasville, GA). This device measures particles from 0.3 to 20 mm in 15 size channels and operates at a flow rate QGRIMM ¼ 1:2 L min1 . Since both the LSA and the Collison nebulizer were operated at several different aerosolization air flows, which changes the total air flow amount at the measurement point, the number of particles produced by each device under different conditions was evaluated in terms of total particle output, N TOT , defined as N TOT ¼ C AIR ðQA þ QDRY Þ,
(1)
where QA is aerosolization air flow for the generator used: QLSA for the LSA or QCOLL for the Collison nebulizer. The performance of the LSA was evaluated with disks of four different pore sizes, DP ¼ 0:2, 0.5, 2.0, and 10.0 mm. The porosity of these disks (percentage of
Fig. 2. Experimental setup.
ARTICLE IN PRESS G. Mainelis et al. / Atmospheric Environment 39 (2005) 3521–3533
surface covered by pores) is 20%, 25%, 35%, and 40%, respectively. The particle output from the LSA was measured for three suspension delivery rates QL ¼ 0:17, 0.6, and 2.0 mL min1 provided by a peristaltic pump (Masterflex C/L, Cole-Parmer, Inc., Vernon Hills, IL) and four aerosolization air flows QLSA ¼ 2; 5, 10, 20, and 30 L min1 delivered at the same pressure of 50 psi. The desired air flow rates were achieved using an adjustable flow meter. At the selected pressure of 50 psi, the pores of 0.2 and 0.5 mm restricted the air flow through the disks to maximum values of 10 L min1. To minimize fluctuations in the concentration of test particles delivered to the LSA via the peristaltic pump, the test particle suspension inside the particle reservoir was continuously mixed with a magnetic stirrer. When performing comparison experiments, the Collison nebulizer was operated at a constant pressure of 50 psi for all conditions and flowrates of QCOLL ¼ 2, 5, and 10 L min1. Flow rates higher than 10 L min1 were not achieved at a pressure of 50 psi and the flow rates less than 10 L min1 were set using an adjustable flow meter. A standard Collison nebulizer recycles 20 mL of suspension approximately every 6 s thus constantly mixing the suspension (May, 1973). Therefore, no additional mixing of the Collison nebulizer suspension was performed. The aerosolization flow rates for both the LSA and the Collison nebulizer were determined with the Buck calibrator (AP Buck, Inc., Orlando, FL). All experiments were conducted in a Class 2 Biological Safety Cabinet (Class II Type A/B3, NUAIRE, Inc., Plymouth, MN).
3525
Both the C INI and the C POST were determined using a light microscope at 400 magnification (Axioscope 20, Carl Zeiss, Inc., Thornwood, NY) and a 0.02 mm deep counting chamber (Hausser Scientific Partnership, Horsham, PA). Using these data, we also determined the percentage of liquid loss DV LIQ for each experimental condition. DV LIQ indicates what fraction of liquid (purified water in our case) delivered to the LSA was lost as vapor or liquid droplets during operation of the LSA. DV LIQ was calculated as follows: DV LIQ ¼ 1
V POST N POST V PSL , QL t N INI V PSL
(5)
where V PSL is the volume of a single PSL particle. 2.5. Aerosolization effect on bacterial viability
The N INI is the product of the PSL particle concentration in the test suspension, C INI , the liquid delivery flow rate, QL , and the time, t, for which the experiment was performed:
Analysis of microbial samplers and other bioaerosol studies sometimes require continuous microorganism aerosolization for prolonged periods of time. Stress introduced by continuous aerosolization and other factors may significantly affect microorganism viability during such experiments, a factor that may compromise study outcome. Thus, it is understandable that several studies analyzing the effect of aerosolization on bacterial viability have concentrated on negative effects accumulated over a certain time. Griffiths et al. (1996) investigated the effects of residence time in the spray and collection liquid on bioaerosol viability. Reponen et al. (1997) indicated that the metabolic injury to P. fluorescens dispersed with Collison nebulizer increases with aerosolization time. Therefore, in our preliminary analysis of the LSA’s effect on bacterial viability, we concentrated on viability change over time. One of the ways to investigate such change would be to collect airborne microorganisms and determine their relative recovery. However, the collection of already aerosolized microorganisms by commonly used methods, such as impaction, is known to affect microorganism viability by itself (Stewart et al., 1995) which may obscure the effect of aerosolization on viability. Thus, for both the LSA and the Collison nebulizer, we determined the concentration of culturable bacteria in the suspension, C CFU , and the concentration of aerosolized bacteria C AIR , at the beginning of the aerosolization ðt ¼ 0 minÞ and at the end of aerosolization ðt ¼ 90 minÞ. Change in C CFU =C AIR ratio over time indicates change in bacterial culturability, F CULT , over time:
N INI ¼ C INI QL t.
F CULT
2.4. Determination of the LSA generation efficiency The LSA generation efficiency, E G , was determined by comparing the number of PSL particles, N INI , delivered to the generator over a certain period of time, with the number of particles not aerosolized by the device and accumulated at the bottom of its glass vessel, N POST . The generation efficiency was defined as follows: EG ¼ 1
N POST . N INI
(2)
(3)
The total number of particles accumulated in the LSA at the end of the aerosolization, N POST , is a product of the accumulated liquid volume, V POST , and the particle concentration, C POST : N POST ¼ V POST C POST .
(4)
¼
C CFU =C AIR ðt ¼ 90Þ C CFU =C AIR ðt ¼ 0Þ 100%. C CFU =C AIR ðt ¼ 0Þ
ð6Þ The C CFU was determined by diluting and plating in triplicate the bacterial suspensions from each generator
ARTICLE IN PRESS 3526
G. Mainelis et al. / Atmospheric Environment 39 (2005) 3521–3533
on Petri dishes with trypticase soy agar, incubating the plates at 26 1C for 24 h, and then counting formed colonies. The concentration of bacteria in the air was determined using the optical particle counter (Grimm Technologies Inc.) as shown in Fig. 2. Overall, each test analyzing the LSA performance at different conditions was run at least three times and the data discussed below are presented by average values and standard deviations of those measurements.
3. Results and discussion 3.1. The effect of disk porosity on total particle output Fig. 3 summarizes the effect of disk pore size, DP , and aerosolization air flow, QLSA , on the total particle output by the LSA. A 0.99 mm PSL suspension of constant concentration was delivered to the LSA at a rate of QL ¼ 0:6 mL min1 for all experiments presented in this figure. For the used air pressure of 50 psi, the maximum aerosolization air flow for the 0.2 and 0.5 mm disks was limited to 10 L min1 due to the limited porosity of the disks. For air flows of QLSA ¼ 2, 5, and 10 L min1, the use of the disk with 0.5 mm pores resulted in the highest airborne particle concentration. For both 0.2 and 0.5 mm disks, the PSL particle output increased almost linearly with the increasing aerosolization airflow. It appears that the application of higher aerosolization flow rates would further increase aerosolized
particle concentration. However, as seen from the data obtained with 2.0 and 10.0 mm disks, the aerosolized particle concentration increased with increasing air flow, then leveled off and even decreased when the aerosolization flow rate was increased to 30 L min1. The 2.0 mm disk achieved a higher particle output than the 10 mm disk at air flows of 10 and 20 L min1, though its particle output was always lower than that of the 0.5 mm disk at flow rates of 2–10 L min1. For the 2 mm disk, the particle output at 20 L min1 was almost the same as that of 0.5 mm disk at 10 L min1. A possible explanation for the decreasing particle output at higher aerosolization flow rates is that high airflow and increased diameter of air streams emanating from the disk (due to larger pore size) disrupt the liquid film and fling it upwards and away from the disk before orderly bubbling can take place. Decrease and elimination of the bursting bubble action then lowers the total particle output of the LSA. Overall, it seemed that air sparging into the suspension film through the 0.5 mm disk created the best conditions for particle aerosolization at a particular aerosolization air flow; however, a very similar particle output was achieved with the 2 mm disk when the aerosolization flow rate was roughly doubled. 3.2. The effect of liquid delivery rate on total particle output The effect of liquid delivery rate on the total particle output of the LSA is shown in Fig. 4. The experiments in
Fig. 3. Effect of disk pore size on 0.99 mm PSL particle output at different aerosolization air flows, QLSA . PSL particle concentration in the liquid suspension was constant for all experiments and the suspension was delivered at a rate of QL ¼ 0:6 mL min1. Symbols represent averages of three repeats and error bars represent standard deviations. The data have been fitted with second- and third-order polynomials. For all curves, R2 40:97.
ARTICLE IN PRESS G. Mainelis et al. / Atmospheric Environment 39 (2005) 3521–3533
3527
Fig. 4. Effect of liquid delivery flow, QL , on the total 0.99 mm PSL particle output at different aerosolization air flows, QLSA . A disk with a 2 mm pore size was used in these experiments and the PSL particle concentration in the liquid suspension was constant for all experiments. Symbols represent averages of three repeats and error bars represent standard deviations. The data have been fitted with third-order polynomials. For all curves, R2 40:98.
this figure were performed using a disk with 2.0 mm pore size and with aerosolization air flows of QLSA ¼ 2, 5, 10, 20, and 30 L min1. The steepest increase in total particle output was between aerosolization airflows of 2 and 10 L min1 for all of the liquid delivery rates. When the aerosolization air flow QLSA was further increased from 10 to 20 L min1, the total particle output for the liquid delivery rates QL ¼ 0:6 and 2.0 mL min1 increased further, although less steeply. Also, when the aerosolization air flow for QL ¼ 0:6 and 2.0 mL min1 was increased to 30 L min1, the total particle output decreased to approximately the level observed at QLSA ¼ 10 L min1. For the lowest liquid delivery rate, QL ¼ 0:17 mL min1, the particle output decreased when the aerosolization flow rate was higher than 10 L min1. It can also be observed that for aerosolization flows of 10 L min1 and higher, increase in the liquid delivery rate resulted in higher particle output. Overall, it seems that each liquid delivery rate has an optimum aerosolization air flow: for the lowest investigated delivery rate of QL ¼ 0:17 mL min1, the optimal QLSA ¼ 10 L min1, while for higher QL of 0.6 and 2.0 mL min1 an aerosolization flow QLSA of 20 L min1 delivered the greatest total particle output. When the air flow is decreased from the optimal setting, the porous disk is overwhelmed with liquid, resulting in a thicker suspension film on the disk. The air streams emanating from the disk break up the suspension film less efficiently which results in reduction of the
total particle output. At the optimal setting, the bursting bubble action is stable and continuous. When the aerosolization air flow is increased above the optimal setting, the air streams overwhelm the force of gravity of the suspension film, rip it apart and fling the larger suspension fragments up and to the side without a steady bubbling process taking place, a condition that reduces particle output. This experiment illustrates that a balance between the liquid delivery rate and the aerosolization air flow results in the highest particle output. Each disk with different pore size is expected to have its own optimal settings. Such operating parameter settings can be estimated qualitatively by achieving circumstances that result in a steady particle output and produce a continuous and steady suspension ‘‘sizzle’’ on the disk. 3.3. A comparison of the total particle output by the Liquid Sparging Aerosolizer and the Collison nebulizer Fig. 5 presents a comparison of the total particle output by the LSA and the Collison nebulizer for 0.99, 2.93, and 5.09 mm PSL particles. The LSA liquid delivery flow for all the experiments was QL ¼ 0:6 mL min1. Each sub-figure uses a PSL suspension of constant concentration, though concentrations for different PSL sizes are not related. At our experimental setting, the Collison nebulizer had a maximum air flow below 20 L min1; thus, the highest air flow selected for experimentation was 10 L min1. For 0.99 mm PSL
ARTICLE IN PRESS 3528
G. Mainelis et al. / Atmospheric Environment 39 (2005) 3521–3533
Fig. 5. Comparison of total particle output between the Liquid Sparging Aerosolizer (LSA) and the Collison nebulizer when operating at different aerosolization flow rates and generating PSL particles of three different sizes: (A) 0.99 mm, (B) 2.93 mm, and (C) 5.09 mm. The LSA was operated with disks of two different porosities: 0.5 and 2.0 mm. Concentrations of PSL particles of each particular size were kept constant throughout the experiments and the LSA liquid delivery rate, QL , was 0.6 mL min1. Symbols represents averages of three repeats and error bars represent standard deviations. The data have been fitted with second- and third-order polynomials. For all curves, but one, R2 40:98. For the curve representing 5.09 mm particles aerosolized by the LSA with 2 mm disk, R2 ¼ 0:84.
particles, the Collison nebulizer produced a much greater particle output than the LSA, especially at 10 L min1. The difference decreased somewhat when the PSL size was increased to 2.93 mm. When aerosolizing 5.09 mm PSL, the use of the LSA with 2.0 mm poresize disk and aerosolization flow rates 10 L min1 and higher resulted in higher total particle output compared with the Collison nebulizer. It is anticipated that the 2.0 mm disk could produce an even higher total particle output at larger flows if the liquid delivery rate is increased, given the data from Fig. 3. Fig. 5 also shows that the LSA operated more efficiently with a 0.5 mm disk when aerosolizing 0.99 mm particles, but had a higher output with a 2.0 mm disk for 2.93 and 5.09 mm particles. To generalize this behavior, as the particle size increases, the size of the disk pores should also increase for optimal total particle output. When the disk pore size increases, so does the diameter of the multiple air streams emanating from the disk. Larger airstreams moving at higher velocities (increases with increasing QLSA ) are better able to pick up larger PSL particles from the suspension film and carry them away toward generator outlet. The ability of the Collison nebulizer to aerosolize larger particles is limited by the average size of droplets dispersed from the nozzle. May (1973)
indicated that the number frequency modal values when generating aqueous solution of fluorescein ranged from 1.8 to 2.7 mm when the Collison nebulizer operated at air pressures from 15 to 40 psi. The modal value could be somewhat higher for the 50 psi pressure used in our experiments; however, it still would be significantly below 5 mm. In general, the Collison nebulizer produces higher concentrations of smaller particles, but the LSA can be effectively used to aerosolize larger solid particles if the disk pore size and the aerosolization flow rate are properly selected. 3.4. Aerosol generation efficiency of the Liquid Sparging Aerosolizer The primary goal of this experiment was to investigate what fraction of the particles delivered to the LSA was actually aerosolized under different operating parameters. We also calculated the fraction of liquid (purified water in our case) lost as water vapor or aerosolized liquid mass. For this experiment, we selected disks with two pore sizes that showed the best results— 0.5 and 2.0 mm—and the extreme tested values of liquid delivery (0.17 and 2.0 mL). Aerosolization air flows of 5 and 10 L min1 were selected for the 0.5 mm disk, and
ARTICLE IN PRESS
4371 471 6271 771 4171 271 8571 1271 8471 6972 9471 7872 9371 8771 9171 3972 60 10 60 10 60 10 120 60 The data presented in this table are averages of three repeats7one standard deviation.
5.8 19.2 3.9 18.5 6.0 19.5 3.0 104.7 1.4370.02 5.470.17 0.5170.13 3.9170.30 0.6170.04 2.2570.16 1.6670.17 64.772.43 8.9570.59 17.571.16 8.9570.59 17.571.16 8.9570.59 17.570.12 17.970.12 105.376.94 5 5 10 10 5 5 30 30
0.2 2 0.2 2 0.2 2 0.2 2 0.5 0.5 0.5 0.5 2 2 2 2
10.15 19.91 10.15 19.91 10.15 19.91 20.31 119.47
Number of non-aerosolized particles accumulated in the LSA after aerosolization, NPOST ( 1010) Volume of Suspension delivered into the LSA, VINI (mL) (70.1) Number of particles delivered into the LSA, NINI ( 1010)
Fig. 7 presents the generation stability of the LSA when generating polydisperse NaCl particles. For these experiments, the LSA was fitted with a 0.5 mm disk, a
3529
Disk pore Aerosolization Liquid air flow, delivery size, DP (mm) QLSA (L min1) rate, QL (mL min1)
3.5. Generation stability over time
Table 1 Generation efficiency of the Liquid Sparging Aerosolizer (LSA)
flows of 5 and 30 L min1 were selected for the 2 mm disk. These values were selected to test what generation efficiency values could be expected even when the total particle output was not optimum, as shown for 2 mm disk in Fig. 4. All experiments here were performed with 0.99 mm PSL particles of the same concentration and the experimental details and results are presented in Table 1. Fig. 6 visualizes the generation efficiency, E G , of the LSA at different operational parameters. For the conditions investigated, the generation efficiency ranged from E G ¼ 39% to 94%. The highest efficiency occurred when using the 0.5 mm disk at an air flow of QLSA ¼ 10 L min1 and liquid delivery flow of QL ¼ 0:17 mL min1. The lowest observed efficiency occurred when using the 2.0 mm disk at QLSA ¼ 30 L min1 and QL ¼ 2:0 mL min1. The generation efficiency did not necessarily correlate with the produced aerosol output. In fact, the data show that for each set disk pore size and aerosolization flow rate, the aerosolization efficiency decreased when liquid delivery increased from 0.17 to 2 mL min1. As was shown in Fig. 4, an increase in liquid delivery rate increases the aerosol output. One can also observe that for the 0.5 mm disk, an increase in aerosolization flow rate increases the aerosolization efficiency when the liquid delivery rate stays constant. This trend for the 2 mm disk is reversed. Most likely it has something to do with the fact that for 2.0 mm disk, use of 30 L min1 results in less-efficient bubbling process, as shown in Fig. 4. Overall, the data show that for settings in which LSA is producing stable, continuously bursting bubbles, the aerosol efficiency is between 80% and 95%, i.e. a significant majority of particles delivered to LSA become airborne and do not accumulate at bottom of the glass vessel. The high particle generation efficiency numbers can be contrasted with lower liquid mass ‘‘generation efficiency’’ (carrier liquid loss) numbers as shown in Table 1. Such numbers mean that the LSA much more readily aerosolizes solid particles than the liquid, in which the particles are suspended. This is especially true when the suspension is delivered at a rate of 2 mL min1. When the suspension was delivered at a rate of 0.17 mL min1, the liquid loss rate varied from 85% for 2 mm disk and airflow of 30 L min1 to the range of 41–61% for other conditions. The higher liquid loss for the 2 mm disk and airflow of 30 L min1 is due to high desiccating capacity of 30 L min1 air flow and low liquid delivery rate of 0.17 mL min1, which increases the percentage of liquid that can be desiccated. As could be expected, higher aerosolization air flows resulted in more significant liquid loss.
Liquid loss during Aerosolization Generation Volume of liquid aeroso-lization, time, t (min) efficiency, E G , % collected in the EG ¼ 1NINI/NPOST DVLIQ (%) LSA after aerosolization, VPOST (mL) (70.1)
G. Mainelis et al. / Atmospheric Environment 39 (2005) 3521–3533
ARTICLE IN PRESS 3530
G. Mainelis et al. / Atmospheric Environment 39 (2005) 3521–3533
Fig. 6. Generation efficiency of Liquid Sparging Aerosolizer for different operational parameters. Symbols represent averages of three repeats and error bars represent standard deviations.
Fig. 7. Generation stability of the Liquid Sparging Aerosolizer using a polydisperse NaCl suspension of constant concentration. The LSA operated with a 0.5 mm disk, a liquid delivery flow of QL ¼ 0:6 mL min1 and an aerosolization air flow of QLSA ¼ 10 L min1. The indicated sizes represent average optical diameters of the aerosolized NaCl particles.
liquid delivery flow of QL ¼ 0:6 mL min1 and an aerosolization air flow of QLSA ¼ 10 L min1. As could be seen from the NaCl data, the concentration of all size fractions was stable during the period of 180 min of continuous generation. The standard deviation of total particle output for all size fractions was below 12%. Some variation in particle output may be due to variation in the aerosolization air flow which was caused by air pressure fluctuations in the air delivery system. Overall, the data presented here indicated that the LSA
is suitable for generating stable particle concentrations over extended periods of time. This may be very useful for extended exposure studies or evaluation of long-term efficiency of particle collectors. 3.6. The effect of the LSA on the stability and viability of a microorganism culture Bioaerosol generation using a bursting bubble principle is believed to be less damaging to sensitive
ARTICLE IN PRESS G. Mainelis et al. / Atmospheric Environment 39 (2005) 3521–3533
3531
Fig. 8. Changes in bacterial size distribution over time when aerosolizing with (A) Collison nebulizer, and (B) Liquid Sparging Aerosolizer. Operational conditions for Collison nebulizer: air pressure ¼ 50 psi, air flow QCOLL ¼ 5 L min1; operational conditions for LSA: disk pore size ¼ 0.5 mm, liquid supply rate QL ¼ 0:17 mL min1, air flow QLSA ¼ 5 L min1. Symbols represent averages of three repeats and error bars represent standard deviations.
microorganisms. In our preliminary analysis of the LSA’s effect on the microorganism stability and viability, we aerosolized sensitive P. fluorescens bacteria for 90 min using the LSA and the Collison nebulizer and then observed changes in bacterial size distribution and viability as described in the Materials and Methods section. In this experiment, the LSA was fitted with a 0.5 mm disk and operated at a liquid delivery rate of 0.17 mL min1 and an aerosolization air flow of 5 L min1. The Collison nebulizer was operated at an air flow of 5 L min1 and a pressure of 50 psi. As can be seen from Fig. 8, at the beginning of aerosolization, both generators produce bacteria with the number frequency mode of approximately 0.6 mm (location of particle peak). When aerosolizing bacteria, particles less than 0.5 mm in size are considered fragments (Mainelis et al., 2001). As could be expected from previous results (Fig. 5), the Collison nebulizer yielded a higher particle output. After a continuous 90 min generation period, the number of bacteria measured at 0.6 mm range doubled. Most likely this is due to the loss of liquid which concentrates the bacterial suspension inside the Collison nebulizer and also any background lysis from adding the bacteria to water. More significantly, the concentration of fragments emitted from the Collison nebulizer has increased about 3–3.5 times after continuous generation for 90 min. We believe that this is due to constant recycling of the suspension, which repeatedly subjects the bacteria to strong shear forces causing them to fragment. A similar fragmentation effect was observed by Mainelis et al. (2001) who also noted that fragmentation of cells caused an increase in a suspension’s electrical conductivity.
In contrast, the size distribution of P. fluorescens bacteria aerosolized with the LSA was virtually identical after 90 min of continuous generation. The difference in bacterial concentration was not statistically significant. The slight increase in the concentration of bacterial fragments could be attributed to natural lysis of the cells suspended in purified water without any nutrients. The change in bacterial viability after continuous aerosolization for 90 min is presented in Table 2. As could be seen, the viability of P. fluorescens bacteria aerosolized with the LSA on average has not changed, while the viability of the same bacteria aerosolized with the Collison nebulizer has decreased by over 50% after 90 min of aerosolization. The experiments performed here indicate that the LSA can retain the viability of sensitive microorganisms over extended aerosolization periods while maintaining constant bioaerosol concentration output. This feature should be useful in various bioaerosol investigations. More detailed analysis of effects on microorganism viability will be addressed in future studies. The LSA design described herein is based on a modified BGI Collison nebulizer. It has been further modified to allow for current production of BGI Collison nebulizers to be retrofitted and made into an LSA with a smaller ((3/4) in) sparging surface. The conversion process involves adding the sparging surface holder and reversing the path of the liquid and gaseous feed. Currently, a prototype of this adaptation is scheduled to be evaluated. In a second new prototype, the LSA was fitted with an enlarged (90 mm) sparging surface featuring pore size of 2 mm. This large unit, designed to operate at air flows of up to 100 L min1
ARTICLE IN PRESS G. Mainelis et al. / Atmospheric Environment 39 (2005) 3521–3533
3532
Table 2 Change in viability of Pseudomonas fluorescens bacteria over time when aerosolized with Collison nebulizer and the Liquid Sparging Aerosolizer as measured by plating Collison nebulizer
Liquid Sparging Aerosolizer
t ¼ 0 min Concentration of viable bacteria in the liquid suspension, CCFU ¼ NCFU/mLLIQ Concentration of total aerosolized bacteria, CAIR ¼ NTOT/LAIR Ratio of viable liquidborne vs. total airborne bacterial concentration, CCFU/CAIR Change in CCFU/CAIR over time, %
t ¼ 90 min
t ¼ 0 min
t ¼ 90 min
1.53 10 71.20 10
1.37 10 73.54 10
1.04 10 78.9 10
1.16 10973.11 107
1.77 10575.37 103
3.26 10579.22 103
5.67 10473.22 103
6.35 10474.96 103
8.6 10373.14 102
4.19 10371.22 102
1.83 10471.17 103
1.83 10471.44 103
9
—
8
9
7
51.374.3
9
—
7
0710.2
The data presented in this table are averages of three measurements7one standard deviation.
may find applications in large-scale simulations of aerosol release. Evaluation of this prototype is also ongoing.
4. Conclusions This study presented and evaluated a new aerosol generator that utilizes a bursting bubble principle and eliminates reuse of the particle suspension, thus creating a single-pass generator suitable for microorganism aerosolization. The obtained data demonstrated that this generator yields a stable particle output when aerosolizing polydisperse and monodisperse particles over extended periods of time. Our tests have shown that the use of disks with pore sizes of 0.5 and 2.0 mm resulted in the highest particle output and that the produced aerosol concentration depends on the air flow through the disk as well as the liquid delivery rate. Using these parameters, one can easily adjust the particle concentration and flow rate needed for a specific experiment. For most of the tested conditions, more than 70% of particles delivered to the LSA were aerosolized. At certain specific conditions, this percentage surpassed 90%. Preliminary experiments with a sensitive microorganism P. fluorescens showed that the bacterial distribution virtually did not change during 90 min of continuous aerosolization using the LSA. In fact, on average there was no (0%) viability loss, whereas the bacterial spectrum produced by a Collison nebulizer changed significantly over time and there was a 50% loss in viability after 90 min. This result demonstrates that low shear forces of bursting bubbles and absence of
suspension recycling maintain microorganism viability over prolonged generation. More exhaustive analysis of microorganism aerosol generation by the LSA will be presented in future studies. Overall, we believe that this new instrument could be used to generate bioaerosol for the evaluation of pathogen collection methods, inhalation and other studies where extended delivery of stable and undamaged biological aerosols is required.
Acknowledgments This study was supported by the Environmental and Occupational Health Sciences Institute (Piscataway, NJ, USA) and the CH Technologies Inc. (Westwood, NJ, USA). The authors are thankful for this support. The authors also appreciate the provision of stainless steel disks by the Mott Corporation (CT, USA).
References Chen, S.-K., Vesley, D., Brosseau, L.M., Vincent, J.H., 1994. Evaluation of single-use masks and respirators for protection of health care workers against mycobacterial aerosols. American Journal of Infection Control 22 (2), 65–74. Dennis, J.H., Stenton, S.C., Beach, J.R., Avery, A.J., Walter, E.H., Hendrick, D.J., 1990. Jet and ultrasonic nebulizer output: use of a new method for direct measurement of aerosol output. Thorax 45 (10), 728–732. Go´rny, R.L., Dutkiewicz, J., 1998. Evaluation of microorganisms and endotoxin levels of indoor air in living rooms occupied by cigarette smokers and non-smokers in Sosnowiec, Upper Silesia, Poland. Aerobiologia 14, 235–239.
ARTICLE IN PRESS G. Mainelis et al. / Atmospheric Environment 39 (2005) 3521–3533 Griffiths, W.D., DeCosemo, G.A.L., 1994. The assessment of bioaerosols: a critical review. Journal of Aerosol Science 25 (8), 1425–1458. Griffiths, W.D., Stewart, I.W., Reading, A.R., Futter, S.J., 1996. Effect of aerosolization, growth phase and residence time in spray and collection fluids on the culturability of cells and spores. Journal of Aerosol Science 27 (5), 803–820. Israeli, E., 1973. Effect of aerosolization and lyophilization on macromolecular synthesis in E. coli. VIth International Symposium on Aerobiology, Enschede, the Netherlands. Wiley, New York, Toronto. Jensen, P.A., Todd, W.F., Davis, G.N., Scarpino, P.V., 1992. Evaluation of eight bioaerosol samplers challenged with aerosols of free bacteria. American Industrial Hygiene Association Journal 53 (10), 660–667. Mainelis, G., Willeke, K., Baron, P., Grinshpun, S.A., Reponen, T., Go´rny, R.L., Trakumas, S., 2001. Electrical charges on airborne microorganisms. Journal of Aerosol Science 32, 1087–1110. May, K.R., 1973. The collison nebulizer: description, performance and application. Journal of Aerosol Science 4, 235–243.
3533
Palleroni, N.J., 1984. Bergey’s Manual of Systematic Bacteriology, vol. 1. Williams and Wilkins Co., Baltimore, p. 165. Reponen, T., Willeke, K., Ulevicius, V., Grinshpun, S.A., Donnelly, J., 1997. Techniques for dispersion of microorganisms into air. Aerosol Science and Technology 27, 405–421. Saint-Louis, R., Pelletier, E., 2004. Sea-to-air flux of contaminants via bubbles bursting. An experimental approach for tributyltin. Marine Chemistry 84 (3/4), 211–225. Stewart, S.L., Grinshpun, S.A., Willeke, K., Terzieva, S., Ulevicius, V., Donnelly, J., 1995. Effect of impact stress on microbial recovery on an Agar surface. Applied and Environmental Microbiology 61 (4), 1232–1239. Ulevicius, V., Willeke, K., Grinshpun, S.A., Donnelly, J., Lin, X., Mainelis, G., 1997. Aerosolization of particles from a bubbling liquid: characteristics and generator development. Aerosol Science and Technology 26 (2), 175–190. White, E., Lucke, W.E., 2003. Effects of fluid composition on mist composition. Applied Occupational & Environmental Hygiene 18 (11), 838–842.