- Email: [email protected]
Accurate measurement of airborne biological particle concentration based on laser-induced fluorescence technique
Journal of Aerosol Science 117 (2018) 24–33
Contents lists available at ScienceDirect
Journal of Aerosol Science journal homepage: www.elsevier.com/locate/jaerosci
Accurate measurement of airborne biological particle concentration based on laser-induced fluorescence technique
T
⁎
Chenyang Lua,b, , Pei Zhanga, Guanghui Wanga, Jing Zhua, Xiaoyan Tangc, ⁎ Weiping Huangc, Shuanghong Chend, Xiongli Xud, Huijie Huanga, a
Shanghai Institute of Optics and Fine Mechanics, Chinese Academy of Sciences, Shanghai 201800, China University of Chinese Academy of Sciences, Beijing 100049, China c The Central Hospital of Shanghai Jiading, Shanghai 201800, China d The PLA Institution of Naval Medical, Shanghai 201606, China b
AB S T R A CT We develop a biological particle counter based on laser-induced fluorescence for accurate measurement of biological particle concentrations in the air. Pure water, NaCl particles, polystyrene latex spheres, and standard fluorescent particles are used as media to evaluate the performance of the counter. Staphylococcus aureus, Escherichia coli, and Pseudomonas aeruginosa are used as representative biological particles to evaluate the measurement accuracy of the counter. In experiments, the results measured by the counter are consistent with he results obtained with the culture method; For each bacterium, good linear agreement is observed between the two results, and the values of the coefficient R2 are all more than 0.97. Because of particle superposition errors, system error, etc., at low concentrations, the number measured by the counter divided by the number measured by the culture method, η = Ncounter/Nculture, is larger than that at high concentrations. For same type of test sample, although the distribution is different, the η values at same concentrations are similar. Finally, the repeatability of the two methods is tested. The results obtained using the counter are found to be more stable, with a relative standard deviation (RSD) of 8.14%; this is less than the RSD of 15% obtained using the culture method.
1. Introduction Bioaerosols are suspensions of airborne particles that contain living organisms or particles released from living organisms (Watches & Cox, 1995). Examples include bacteria, fungi, pollens, viruses, and protein allergens from animals or plants. Intrinsic fluorescence is an inherent characteristic of bioaerosols, and it is usually generated from organic substances [such as amino acids, riboflavin, and reduced nicotinamide adenine dinucleotide (NADH)] under excitation by ultraviolet (UV) light (Lakowicz, 2006). Since the discovery of the intrinsic laser-induced fluorescence (LIF) of bioaerosols, mixing techniques in combination with the single-particle scattering technique and LIF technique have been rapidly developed over the last two decades (Yong-Le et al., 2007). Many companies, colleges, and research institutions have studied the intrinsic fluorescence of various bioaerosols and developed systems and equipment for identification and classification of bioaerosols. In 1997, Hairston and his colleagues from TSI Inc. in the US designed a fluorescence detection system based on the APS-3310 particle counter using two lasers (Hairston, Ho, & Quant, 1997). In 1998, Pinnick et al. created an experimental device for detecting the fluorescence spectrum of a single biological aerosol particle,
⁎
Corresponding author at: Shanghai Institute of Optics and Fine Mechanics, Chinese Academy of Sciences, Shanghai 201800, China. E-mail address: [email protected] (H. Huang).
https://doi.org/10.1016/j.jaerosci.2017.12.010 Received 21 August 2017; Received in revised form 20 November 2017; Accepted 23 December 2017 Available online 26 December 2017 0021-8502/ © 2017 Elsevier Ltd. All rights reserved.
Journal of Aerosol Science 117 (2018) 24–33
C. Lu et al.
which could characterize biological particulates; this system has an additional two lasers (488 and 266 nm), unlike that of Hairston et al. The 488 nm laser is used not only for scattered light detection but also the detection of the intrinsic fluorescence. The 266 nm laser determines whether there are biological particles (Pinnick et al., 1998). In 2000, Brosseau designed and developed a fluorescence detector, the UV aerodynamic particle sizer. The instrument employs a pulsed, 355 nm laser rather than a continuous-wave UV laser such as that described by Hairston et al. Brosseau et al. (2000). Azbil BioVigilant Inc. introduced the IMD-A series, in which scattered light detection and fluorescence excitation are performed by the same laser, a semiconductor laser with a near-UV wavelength of 405 nm (Product Specifications of Azbil BioVigilant. Inc, 2013a, 2013b). The semiconductor laser can greatly reduce the cost and allow for miniaturization of the equipment. Evenstad et al. developed FLAPS-3317, a new fluorescence sensor based on a previous system. In this system, scattered light detection and fluorescence excitation are also performed by the same 405 nm semiconductor laser. However, the fluorescence is separated into two wave bands to identify bioparticles (Jim, Dahu, Hairston Peter, & Darrick, 2013). Many current systems and instruments resemble FLAPS-3317. These include the biological agent warning sensor (Primmerman, 2000) and the single-particle fluorescence analyzer (Eversole, Hardgrove, Cary, Choulas, & Seaver, 1999). Pan et al. designed an experimental device called the dual-excitation-wavelength particle fluorescence spectrometer. It has double-wavelength excitation and a multichannel receiver, and 16 varieties of atmospheric particles have been tested (Pan et al., 2011). In 2004, Kaye et al. introduced the wide issue bioaerosol sensor (WIBS) for networked deployment (Kayea et al., 2004). One year later, Kaye et al. developed WIBS-2 (Kaye et al., 2005). Several years later, Droplet Measurement Technologies developed a new instrument called WIBS-4. In 2012, Toprak and Schnaiter evaluated its performance in the laboratory and in the field (Toprak & Schnaiter, 2012). In recent years, Droplet Measurement Technologies developed a new instrument called WIBS-4A. WIBS uses pulsed UV xenon lamps for fluorescence excitation, and it generates two wave bands during operation. It also has a laser diode at 635 nm and is used for scattered light detection. The two wave bands for fluorescence excitation are at 280 and 370 nm, which correspond to the tryptophan and NADH spectral absorption peaks, respectively, and are used for intrinsic fluorescence detection (Product Specifications of Droplet Measurement Technologies). In addition, countries such as Germany, Switzerland, and Norway have also made great advances in the development of light-induced fluorescence detection technology (Bundke, Reimann, Nillius, Jaenicke, & Bingemer, 2010; Farsund, Rustad, Kasen, & Haavardsholm, 2010; Kiselev, Bonacina, & Wolf, 2011). Almost all of these studies focused on the identification and classification of bioaerosols. However, sometimes the concentration of microorganisms is more important than the identification and classification of bioaerosols. When the concentrations of infectious bacteria or viruses exceed a certain limit in the environment, these microorganisms cause diseases and adversely affect human health. Therefore, it is important to measure the concentration of bioaerosols accurately and promptly in various environments. In this paper, we present a counter with optical and detection systems, and conduct a series of experiments to evaluate its performance, focusing on its accuracy in measuring bioparticle concentrations. Finally, we compare the counter's results with those of the culture method. 2. Description of counter Many counters using LIF and light scattering are described in detail elsewhere. Our counter, with a new optical system and detection system, is described here. The exterior of the instrument is shown in Fig. 1. The dimensions of the device are 245 mm (width) × 220 mm (height)× 145 mm (depth) 2.1. Optical system The optical system is shown schematically in Fig. 2. The system consists of the input optics for the 405 nm near-UV laser diode beam used for both the scattered light and LIF signals. Receiving optics are also needed to collect light from the illuminated particles
Fig. 1. Exterior of instrument.
25
Journal of Aerosol Science 117 (2018) 24–33
C. Lu et al.
Fig. 2. Schematic diagram of the optical system. The system consists of a radiation source, paraboloidal mirror, light trap, two focusing lenses, and a dichroic mirror.
and selectively pass the appropriate wave bands to the photomultiplier tube (PMT) for both scattered light and fluorescence detection. A laser diode is inserted into the center of a paraboloidal mirror. The intersection of the laser beam and the aerosol flow is the focus point of the paraboloidal mirror. The laser irradiates the aerosol directly, and the light beam that passes through the aerosol particles is directed into a light trap. The scattered light and LIF are combined by the paraboloidal mirror and passed to the two focusing lenses, as shown in Fig. 2. After passing through the second lens, the mixed light converges. Then the scattered and fluorescence light are separated by a dichroic mirror. The wave band of the light reflected by the dichroic mirror is below 420 nm and includes the scattered light; the wave band of light that passes through the dichroic mirror is 420–650 nm, which corresponds to the wave band of the fluorescence. Finally, the scattered light and the fluorescence are focused onto separate PMTs. 2.2. Electronic and control systems We also combine scattered light detection and fluorescence detection to distinguish bioparticles from other particles. This counter uses the elastic light scattering detection and LIF techniques. These two technologies have proved successful in past research, so we omit a detailed description here. For each detected particle, the signal processing system measures its equivalent optical diameter from the scattered light signal, and then the fluorescence signal is used to determine whether it is a bioparticle. Fig. 3 shows the oscilloscope traces of the analog signals from both the scattered light and the fluorescence. If the particle is not a bioparticle, it will emit only the scattered light signal;
Fig. 3. Oscilloscope screenshots. The upper and lower lines are the pulse signals for the scattered light and the fluorescence, respectively. (a) Pulse signal of a nonfluorescent particle. (b) Pulse signal of a fluorescent particle.
26
Journal of Aerosol Science 117 (2018) 24–33
C. Lu et al.
Fig. 4. Process diagram for the detection system. This is the entire system process, which includes the transformation of various signals.
if it is a bioparticle, it will also generate the fluorescence signal. The light signals are then transmitted to a processor. Fig. 4 shows a diagram of the electronic functions. First, the PMT transforms the light signal into a current signal, which is transformed into a voltage signal by a transimpedance amplifier that has a low bias current and offset, and simultaneously amplifies the voltage signal. The voltage signal is amplified further by a noninverting amplifier to strengthen it. After the voltage signal is enhanced, the analog digital converter (ADC) converts the voltage signal into a digital signal, which is transferred to the fieldprogrammable gate array (FPGA). The FPGA's sampling rate is 10 Mps, and a time function control outputs the digital signals into a microcontroller unit (MCU). Finally, the MCU receives information regarding the number of bioparticles and other particles of different sizes and then determines the need to sound an alarm on the basis of the calculated results. 3. Experiments and analysis 3.1. Experimental apparatus and sample preparation To evaluate the performance of the biological particle counter, experiments were performed using representative samples: NaCl particles, polystyrene latex (PSL) spheres, Model B800 (a type of fluorescent microspheres from Thermo Scientific, USA), and samples of Staphylococcus aureus, Escherichia coli, and Pseudomonas aeruginosa. As shown in Fig. 5, an aerosol generator capable of producing aerosols with different concentrations was adopted to generate the target aerosol. The target aerosol then formed a stable environment in the buffer bottle. The biological particle counter was used to detect the target aerosol in the buffer bottle. There was a sampler between the outlet of the optical cavity and the inlet of the pump. The sampler had a dissolvable filter for collecting bioparticles. The details are shown in Fig. 6. First, we conducted a series of tests with pure water because all the samples were dissolved in pure water when the aerosols were generated. We verified that the water introduced no interference factors. In addition, we compared the results of the culture method to those obtained by the biological particle counter. The results produced by the biological particle counter were averaged over five cycles. The results of the culture method were also averaged. The results are shown in Table 1. The table shows that there were no bioparticles or fluorescent particles in pure water; therefore, it is suitable to use pure water as the solvent in other experiments. 3.2. Identification of nonfluorescent and fluorescent particles NaCl particles and PSL particles are representative nonfluorescent particles, whereas B800 microspheres are representative fluorescent particles. These samples were used to test the reliability of the biological particle counter. The standard Model B800 fluorescent microspheres were chosen as a substitute for bioparticles. The excitation and fluorescence spectra of these microspheres are similar to those of NADH, which is an important component of bioparticles. The mean diameter of the B800 microspheres is 0.8 µm, and the size uniformity (standard deviation) is less than 3%. The concentration of a B800 aerosol can be controlled by
Fig. 5. Experimental setup for performance evaluation of the monitor. The target aerosol is aerosolized by the aerosol generator, forming a uniform aerosol in the buffer bottle.
27
Journal of Aerosol Science 117 (2018) 24–33
C. Lu et al.
Fig. 6. Diagram showing the location of the sampler. This is only a schematic diagram, and some components are not displayed.
Table 1 Results of the biological particle counter and culture method. Test sample
Pure water
Number
1 2 3 4 5
Biological particle counter
Culture method
(≥0.5 µm) Average particle number (PPL)
Average number of biological particles (PPL)
Average number of biological particles (CFU/ L)
12.2 18.1 21.2 11.6 16.4
0 0 0 0 0
0 0 0 0 0
adjusting the aerosol generator. To avoid misdetection, we measured the target aerosol five times using the same concentration and considered the average of the five results in every group of experiments. For NaCl and PSL particles, we measured the results using three different concentrations. For B800 particles, there were six different concentrations. The results for the nonfluorescent particles are shown in Table 2, which shows that no fluorescence was detected. Therefore, the counter did not produce false detections when the test samples were NaCl and PSL particles. To test the counter's ability to detect fluorescence, we used B800 and a mixture of NaCl and B800 as test samples. The results for B800 and the mixture are displayed in Tables 3 and 4, respectively. In Table 4, the ratio of B800 particles to NaCl solution is a relative value, and in the experiment, all conditions are the same except the ratio of B800 particles to NaCl solution. According to the results in Table 3, although most of the detected B800 particles were fluorescent, a few particles were nonfluorescent. On the basis of the results in Table 1, we used pure water as the solvent, which also generated a few particles in the aerosol generator; consequently, the fluorescent particle numbers were slightly lower than the particle numbers. In the B800 test, the performance of the counter was adequate. The results in Table 4 show that the proportion of fluorescent particles increased in proportion to the ratio of B800 particles to NaCl solution. On the basis of the results in Tables 2–4, we can consider that the counter is able to identify the nonfluorescent and fluorescent particles accurately. 3.3. Accurate measurement of biological particle concentration Almost all of the studies of LIF technology have focused on the identification and classification of bioaerosols. However, our Table 2 Results as measured by the counter. The samples were NaCl and PSL particles. Test sample
(≥0.5 µm) Average particle number (PPL)
Number of fluorescent particles (PPL)
NaCl particles
12,250 502,45.4 114,236 1423.2 5056 12,138.4
0 0 0 0 0 0
0.5 µm PSL particles
28
Journal of Aerosol Science 117 (2018) 24–33
C. Lu et al.
Table 3 Results of fluorescence tests of B800 particles. Test sample
(≥0.5 µm) Average particle number (PPL)
Number of fluorescent particles (PPL)
Proportion of fluorescent particles
B800
193.6 386.4 689 1563 2349 3330
174.2 355.4 613 1400 2137 2930
90% 91.8% 89% 89.5% 91% 88%
Table 4 Results of fluorescence test of a mixture of NaCl and B800. Ratio of B800 to NaCl solution (one drop: liter)
(≥0.5 µm) Particle number (PPL)
Number of fluorescent particles (PPL)
Proportion of fluorescent particles
1:1
9450 9474 9520 9379 9508 9810 9708 9856 9997 9713 10,004 10,312 10,987 10,045 10,333
1087 1110 1092 1079 1121 2100 2212 2108 2245 2178 3302 3405 3298 3288 3345
11.5% 11.7% 11.4% 11.5% 11.8% 21.4% 22.8% 21.4% 22.4% 22.4% 33% 33% 30% 32.7% 32.4%
2:1
3:1
attention is focused on accurate measurement of bacterial concentrations. S. aureus, E. coli, and P. aeruginosa are used as representative biological particles to evaluate the measurement accuracy of the counter in the laboratory. S. aureus is a Gram-positive bacterium; E. coli and P. aeruginosa are Gram-negative bacteria. The results of the culture method for test samples of bacteria were compared with those of the biological particle counter to test the counter's reliability. The experimental process is outlined as follows: (a) First, the original saved frozen bacteria were activated in a thermostat. Two days later, the activated bacteria were placed in a centrifuge to separate them from the nutrients. (b) Then the bacteria, which were free from impurities, were dissolved in pure water to generate the target aerosol. (c) The aerosol generator was adjusted to control the concentration in the buffer bottle. When the concentration in the buffer bottle was stable, we ran the biological particle counter for five cycles. (d) We recorded the bioparticle detection results for each cycle. Then, five cycles later, we removed the dissolvable filter from the sampler and placed it in a prepared saline solution for dilution. The total volume was 10 mL (this value could be adjusted). (e) We distributed the bacteria homogeneously in the mixed liquor using a vortex oscillator to ensure that there are no clumps of two or more bacteria. (f) Then the sample was distributed into seven (or more) culture dishes. Two dishes contained 0.5 mL of mixed liquor; two contained 1.0 mL of mixed liquor; two contained 2.0 mL of mixed liquor, and the remaining dish contained 3.0 mL of mixed liquor. (g) Finally, we added a mixture of nutrient solution and agar to each dish. When the mixture began clotting, we left the culture dishes in a suitable environment. (h) The results of the culture method were obtained two days later. We counted the clumps on each dish. Then we verified whether the proportions among the seven dishes were correct. For example, the number of clumps in the dish containing 1.0 mL of mixed liquor should be twice that in the dish containing 0.5 mL of mixed liquor. We calculated the average over five cycles and compared it with the average results of the culture method. In the experiments, one cycle is 1 min, and the flow rate of the counter is 2 L/min, so the total air flow in five cycles is 10 L. The results of the two methods at various concentrations are shown as follows. Tables 5(a) and 5(b) show the results for S. Aureus, the S. Aureus suspension in Table 5(a) is diluted 100 times with original S. Aureus, the S. Aureus suspension in Table 5(b) is diluted 50 times with original S. Aureus. Table 5(c) shows the results for E. coli, and Table 5(d) shows the results for P. aeruginosa. Each measured result was the average of five cycles. According to Tables 5(a) and 5(b), 5(c), and 5(d), we can conclude the following:
29
Journal of Aerosol Science 117 (2018) 24–33
C. Lu et al.
Table 5(a) Results of biological particle counter and culture method for S. aureus (diluted 100 times with original S. aureus). Number
Result of biological particle counter (PPL) (Ncounter)
Result of the culture method (CFU/10 L) (Nculture)
η = Ncounter/Nculture
1 2 3 4 5 6 7 8 9 10 11 12
4 7.8 11 26 30 50 54 64.6 69.4 85 152.4 243.8
49 72 118 251 425 693 662 828 951 1183 2097 3803
81.6% 108.3% 93.2% 104% 71.4% 72.5% 81.8% 77.8% 72.9% 71.8% 73.04% 64.1%
Table 5(b) Results of biological particle counter and culture method for S. aureus (diluted 50 times with original S. aureus). Number
Result of biological particle counter (PPL) (Ncounter)
Result of culture method (CFU/10 L) (Nculture)
η = Ncounter/Nculture
1 2 3 4 5 6 7 8 9 10 11 12
5 8.8 14 22 33.6 39 46.2 53 65.4 92 143.2 210.8
61 101 169 223 462 549 633 680 871 1244 2106 3221
81.9% 88% 82.8% 98.6% 72.7% 71% 73% 78% 75.1% 73.9% 68% 65.4%
Table 5(c) Results of biological particle counter and culture method for E. coli. Number
Result of biological particle counter (PPL) (Ncounter)
Result of culture method (CFU/10 L) (Nculture)
η = Ncounter/Nculture
1 2 3 4 5 6 7 8 9 10 11 12
6 8.6 10 12.2 21.4 33.3 41.2 44.7 64.6 69.6 71.4 98
88 121 139 162 398 717 943 925 1154 1178 1472 1941
68.1% 71.1% 71.9% 75.3% 53.7% 46.4% 43.6% 48.3% 55.9% 55.9% 48.5% 50.5%
(1) The percentage η (= Ncounter/Nculture) is generally greater at low concentration than at high concentration: Tables 5(a) and 5(b) show that at concentrations of < 25 CFU/L, the η value of S. aureus is between 80% and 110%, and that at concentrations of > 25 CFU/L, the results are worse than those of the culture method, and η is between 64% and 80%. In addition, Tables 5(c) and 5(d) yield similar conclusions. (2) Tables 5(a) and 5(b) reveal that for the same type of test sample, although the distribution is different, the η values at same concentrations are similar. (3) A comparison of Table 5(a) with Tables 5(c) and 5(d) shows that for different types of test samples, η is different at different concentrations. But for each group,there is a good linear relationship. Conclusion (1) can be explained by the bacteria generated in the buffer bottle, which are composed of two or more particles but are regarded by the counter as a single particle. However, in the culture method, they are isolated, and at low concentrations, the probability that several S. aureus are clumped together is reduced. In addition, the intensity of the fluorescence excited by different S. aureus varies. Some S. aureus are inactive, and their fluorescence is too weak to be detected by the counter, but they are still 30
Journal of Aerosol Science 117 (2018) 24–33
C. Lu et al.
Table 5(d) Results of biological particle counter and culture method for P. aeruginosa. Number
Result of biological particle counter (PPL) (Ncounter)
Result of culture method (CFU/10 L) (Nculture)
η = Ncounter/Nculture
1 2 3 4 5 6 7 8 9 10 11 12
4.2 9 19.6 28 35.2 40.2 54.4 63.6 79.4 89.2 93.8 99
61 126 284 451 556 681 928 1022 1274 1335 1503 1601
68.8% 71.4% 69% 62.1% 63.3% 59.1% 58.7% 62.2% 62.3% 66.8% 62.4% 61.8%
cultivable. The relevance of the results of the two methods is also considered using Fig. 7. The fitted straight-line segment in Fig. 7(a) is y (CFU/L) = 1.54382*x (PPL) − 9.8974 , and the linearly dependent coefficient is 0.99. The fitted straight-line segment in Fig. 7(b) is y (CFU/L) = 1.53148*x (PPL) − 7.54765, and the linearly dependent coefficient is 0.99. The fitted straight-line segment in Fig. 7(c) is y (CFU/L) = 1.97899*x (PPL) − 2.34107 , and the linearly dependent coefficient is
Fig. 7. Relationship between the concentrations detected by the two methods for (a) S. aureus (diluted 100 times with original S. aureus), (b) S. aureus (diluted 50 times with original S. aureus), (c) E. coli, and (d) P. aeruginosa.
31
Journal of Aerosol Science 117 (2018) 24–33
C. Lu et al.
Table 6 Repeatability of the results measured by the two methods. Group
1
2
3
4
5
6
7
8
Average
Standard deviation
Relative standard deviation
Measured by counter (PPL) Measured by culture method (CFU/L)
56 48
47 47
57 50
44 66
54 55
50 59
53 68
54 66
51.9 57.375
4.226 8.62
0.0814 0.15
0.97. The fitted straight-line segment in Fig. 7(d) is y (CFU/L) = 1.59382*x (PPL) + 0.03379, and the linearly dependent coefficient is 0.99. These results show that the linear relationship between the concentrations detected by the two methods is adequate. Different bacteria may have different linear relationships; further research is needed here. Furthermore, we studied the repeatability of the results of both methods using S. aureus. We ensured that the concentration of the bacteria was stable and then performed the same experiment eight times. The results are shown in Table 6 and Fig. 8. The result shows that the averages of the two methods are similar. The standard deviation and relative standard deviation (RSD) of the two methods are reasonable. According to USP 1223, an RSD measured in the 15–35% range for the culture method would generally be acceptable. Many conventional microbiological methods are subject to sampling, dilution, plating, incubation, and operator errors (USP-, 1223). The RSD of 8.14% for the counter measurements is stable.
4. Conclusions This paper focused on accurate measurement of bacterial concentrations in air by a biological particle counter. The counter includes optical, electronic, and control systems. A series of experiments were presented to evaluate the performance of a single biological particle counter. Pure water, NaCl particles, PSL particles, and B800 microspheres were used to evaluate the performance of the counter. The results show that this counter can reliably distinguish between fluorescent and nonfluorescent particles. For target aerosols of S. aureus, E. coli, and P. aeruginosa, we compared the results measured by the counter with those of the culture method. We found that the percentage η (= Ncounter/Nculture) was higher at low concentration than at high concentration. When we compared the results for S. aureus at different concentrations, we found that for the same type of test sample, although there were different distributions, the η values at same concentrations were similar. A comparison of the result for S. aureus with those of E. coli and P. aeruginosa revealed that for different types of test samples, η was different at same concentrations. In addition, for each bacteria, good linear agreement was observed between the two results, and the coefficients (R2) were all greater than 0.97. The repeatability of the results measured by the counter and the culture method was also tested. The results measured by the counter were more stable, with an RSD of 8.14%, which is smaller than that of the culture method (15%). Studies of other bacteria were outside the scope of this work, so our research is limited. However, the research method, which compares the new method based on the LIF technique with the culture method, is feasible. In addition, this counter counted the number of fluorescent particles accurately, but it could not obtain the fluorescence spectrum. A good method to lower the false alarm rate is to analyze the fluorescence spectrum and develop a new algorithm. Another method is to use supplementary means such as measurement of the temperature characteristics of the bioparticles. In the future, more research is needed to ensure a low false alarm rate for detection and to achieve more precise identification of bioaerosols.
Fig. 8. Repeatability of the results measured by the two methods. The red line is the average. (a) Repeatability of the counter results. (b) Repeatability of the results of the culture method.
32
Journal of Aerosol Science 117 (2018) 24–33
C. Lu et al.
References Watches, C. M., & Cox, C. B. (1995). Bioaerosols handbook. Lewis Publishers. Lakowicz, J. R. (2006). Principles of fluorescence spectroscopy. Springer Science and Business Media. Pan Yong-Le, Eversole Jay. D., Kaye Paul. H., Foot Virginia, Pinnick Ronald. G., Hill Steven. C.,Mayo Michael. W., Bottiger Jerold. R., AlanHuston, VasanthiSivaprakasam, Chang Richard. K.,Bio-aerosol fluorescence detecting and characterising bio-aerosols via UV light-induced fluorescence spectroscopy, Optics of Biological Particles, 63–164. Hairston, P. P., Ho, J., & Quant, F. R. (1997). Design of an counter for real-time detection of bioaerosols using simultaneous measurement of particle aerodynamic size and intrinsic fluorescence. Elsevier Science Ltd. Pinnick, R. G., Hill, S. C., Nachman, P., Videen, G., Chen, G., & Chang, R. K. (1998). Aerosol fluorescence spectrum analyzer for rapid measurement of single micrometer-sized airborne biological particles. Aerosol Science and Technology, 28(2), 95–104. Brosseau, L. M., Vesley, D., Rice, N., Goodell, K., Nellis, M., & Hairston, P. (2000). Differences in detected fluorescence among several bacterial species measured with a direct-reading particle sizer and fluorescence detector. Aerosol Science and Technology, 32, 545–558. Product Specifications of Azbil BioVigilant.Inc (2013a). IMD-A® Series Rapid Biologic Detection System for Air. Product Specifications of Azbil BioVigilant.Inc (2013b). Comparison of IMD-A® 300 and 350 Models. Jim, E., Dahu, Q., Hairston Peter, P., & Darrick, N. (2013). Apparatus and system for characterizing an aerosol particle flow. PCT (043600 A1). Primmerman, C. (2000). A detection of biological agents. Lincoln Laboratory Journal, 12(1), 3–32. Eversole, J., Hardgrove, J., Cary, W., Choulas, D., & Seaver, M. (1999). Continuous, rapid biological aerosol detection with the use of UV fluorescence: Outdoor test results. Field Analytical Chemistry & Technology, 3(4‐5), 249–259. Pan, Y.-L., Hill, S. C., Pinnick, R. G., House, J. M., Flagan, R. C., & Dual, C. R. K. (2011). excitation-wavelength fluorescence spectra and elastic scattering for differentiation of single airborne pollen and fungal particles. Atmospheric Environment, 45(8), 1555–1563. Kayea Paul. H., Hirsta Edwin, Foot Virginia, Clark James. M., and Baxter Karen, A low-cost multi-channel aerosol fluorescence sensor for networked deployment, SPIE Vol. 5617. Kaye, P. H., Stanley, W. R., Hirst, E., Foot, E. V., Baxter, K. L., & Barrington, S. J. (2005). A single particle multichannel bio-aerosol fluorescence sensor. Optical Society of America. Toprak, E., & Schnaiter, M. (2012). Fluorescent biological aerosol particles (FBAPs) measured with the Waveband Integrated bioaerosol sensor WIBS-4: laboratory tests combined with a one year field study. Atmospheric Chemistry and Physics Discussion, 12, 17607–17656. Product Specifications of Droplet Measurement Technologies, WIDEBAND INTEGRATED BIOAEROSOL SENSOR−4A(WIBS-4A). Bundke, U., Reimann, B., Nillius, B., Jaenicke, R., & Bingemer, H. (2010). Development of a bioaerosol single particle detector (BIO IN) for the Fast Ice Nucleus CHamber FINCH. Atmospheric Measurement Techniques, 3(1), 263–271. Kiselev, D., Bonacina, L., & Wolf, J. P. (2011). Individual bioaerosol particle discrimination by multi-photon excited fluorescence. Optics Express, 19(24), 24516–24521. Farsund, O., Rustad, G., Kasen, I., & Haavardsholm, T. V. (2010). Required spectral resolution for bioaerosol detection algorithms using standoff laser-induced fluorescence measurements. IEEE Sensors Journal, 10(3), 655–661. USP-1223, Validation of alternative microbiological methods.
33
Contents lists available at ScienceDirect
Journal of Aerosol Science journal homepage: www.elsevier.com/locate/jaerosci
Accurate measurement of airborne biological particle concentration based on laser-induced fluorescence technique
T
⁎
Chenyang Lua,b, , Pei Zhanga, Guanghui Wanga, Jing Zhua, Xiaoyan Tangc, ⁎ Weiping Huangc, Shuanghong Chend, Xiongli Xud, Huijie Huanga, a
Shanghai Institute of Optics and Fine Mechanics, Chinese Academy of Sciences, Shanghai 201800, China University of Chinese Academy of Sciences, Beijing 100049, China c The Central Hospital of Shanghai Jiading, Shanghai 201800, China d The PLA Institution of Naval Medical, Shanghai 201606, China b
AB S T R A CT We develop a biological particle counter based on laser-induced fluorescence for accurate measurement of biological particle concentrations in the air. Pure water, NaCl particles, polystyrene latex spheres, and standard fluorescent particles are used as media to evaluate the performance of the counter. Staphylococcus aureus, Escherichia coli, and Pseudomonas aeruginosa are used as representative biological particles to evaluate the measurement accuracy of the counter. In experiments, the results measured by the counter are consistent with he results obtained with the culture method; For each bacterium, good linear agreement is observed between the two results, and the values of the coefficient R2 are all more than 0.97. Because of particle superposition errors, system error, etc., at low concentrations, the number measured by the counter divided by the number measured by the culture method, η = Ncounter/Nculture, is larger than that at high concentrations. For same type of test sample, although the distribution is different, the η values at same concentrations are similar. Finally, the repeatability of the two methods is tested. The results obtained using the counter are found to be more stable, with a relative standard deviation (RSD) of 8.14%; this is less than the RSD of 15% obtained using the culture method.
1. Introduction Bioaerosols are suspensions of airborne particles that contain living organisms or particles released from living organisms (Watches & Cox, 1995). Examples include bacteria, fungi, pollens, viruses, and protein allergens from animals or plants. Intrinsic fluorescence is an inherent characteristic of bioaerosols, and it is usually generated from organic substances [such as amino acids, riboflavin, and reduced nicotinamide adenine dinucleotide (NADH)] under excitation by ultraviolet (UV) light (Lakowicz, 2006). Since the discovery of the intrinsic laser-induced fluorescence (LIF) of bioaerosols, mixing techniques in combination with the single-particle scattering technique and LIF technique have been rapidly developed over the last two decades (Yong-Le et al., 2007). Many companies, colleges, and research institutions have studied the intrinsic fluorescence of various bioaerosols and developed systems and equipment for identification and classification of bioaerosols. In 1997, Hairston and his colleagues from TSI Inc. in the US designed a fluorescence detection system based on the APS-3310 particle counter using two lasers (Hairston, Ho, & Quant, 1997). In 1998, Pinnick et al. created an experimental device for detecting the fluorescence spectrum of a single biological aerosol particle,
⁎
Corresponding author at: Shanghai Institute of Optics and Fine Mechanics, Chinese Academy of Sciences, Shanghai 201800, China. E-mail address: [email protected] (H. Huang).
https://doi.org/10.1016/j.jaerosci.2017.12.010 Received 21 August 2017; Received in revised form 20 November 2017; Accepted 23 December 2017 Available online 26 December 2017 0021-8502/ © 2017 Elsevier Ltd. All rights reserved.
Journal of Aerosol Science 117 (2018) 24–33
C. Lu et al.
which could characterize biological particulates; this system has an additional two lasers (488 and 266 nm), unlike that of Hairston et al. The 488 nm laser is used not only for scattered light detection but also the detection of the intrinsic fluorescence. The 266 nm laser determines whether there are biological particles (Pinnick et al., 1998). In 2000, Brosseau designed and developed a fluorescence detector, the UV aerodynamic particle sizer. The instrument employs a pulsed, 355 nm laser rather than a continuous-wave UV laser such as that described by Hairston et al. Brosseau et al. (2000). Azbil BioVigilant Inc. introduced the IMD-A series, in which scattered light detection and fluorescence excitation are performed by the same laser, a semiconductor laser with a near-UV wavelength of 405 nm (Product Specifications of Azbil BioVigilant. Inc, 2013a, 2013b). The semiconductor laser can greatly reduce the cost and allow for miniaturization of the equipment. Evenstad et al. developed FLAPS-3317, a new fluorescence sensor based on a previous system. In this system, scattered light detection and fluorescence excitation are also performed by the same 405 nm semiconductor laser. However, the fluorescence is separated into two wave bands to identify bioparticles (Jim, Dahu, Hairston Peter, & Darrick, 2013). Many current systems and instruments resemble FLAPS-3317. These include the biological agent warning sensor (Primmerman, 2000) and the single-particle fluorescence analyzer (Eversole, Hardgrove, Cary, Choulas, & Seaver, 1999). Pan et al. designed an experimental device called the dual-excitation-wavelength particle fluorescence spectrometer. It has double-wavelength excitation and a multichannel receiver, and 16 varieties of atmospheric particles have been tested (Pan et al., 2011). In 2004, Kaye et al. introduced the wide issue bioaerosol sensor (WIBS) for networked deployment (Kayea et al., 2004). One year later, Kaye et al. developed WIBS-2 (Kaye et al., 2005). Several years later, Droplet Measurement Technologies developed a new instrument called WIBS-4. In 2012, Toprak and Schnaiter evaluated its performance in the laboratory and in the field (Toprak & Schnaiter, 2012). In recent years, Droplet Measurement Technologies developed a new instrument called WIBS-4A. WIBS uses pulsed UV xenon lamps for fluorescence excitation, and it generates two wave bands during operation. It also has a laser diode at 635 nm and is used for scattered light detection. The two wave bands for fluorescence excitation are at 280 and 370 nm, which correspond to the tryptophan and NADH spectral absorption peaks, respectively, and are used for intrinsic fluorescence detection (Product Specifications of Droplet Measurement Technologies). In addition, countries such as Germany, Switzerland, and Norway have also made great advances in the development of light-induced fluorescence detection technology (Bundke, Reimann, Nillius, Jaenicke, & Bingemer, 2010; Farsund, Rustad, Kasen, & Haavardsholm, 2010; Kiselev, Bonacina, & Wolf, 2011). Almost all of these studies focused on the identification and classification of bioaerosols. However, sometimes the concentration of microorganisms is more important than the identification and classification of bioaerosols. When the concentrations of infectious bacteria or viruses exceed a certain limit in the environment, these microorganisms cause diseases and adversely affect human health. Therefore, it is important to measure the concentration of bioaerosols accurately and promptly in various environments. In this paper, we present a counter with optical and detection systems, and conduct a series of experiments to evaluate its performance, focusing on its accuracy in measuring bioparticle concentrations. Finally, we compare the counter's results with those of the culture method. 2. Description of counter Many counters using LIF and light scattering are described in detail elsewhere. Our counter, with a new optical system and detection system, is described here. The exterior of the instrument is shown in Fig. 1. The dimensions of the device are 245 mm (width) × 220 mm (height)× 145 mm (depth) 2.1. Optical system The optical system is shown schematically in Fig. 2. The system consists of the input optics for the 405 nm near-UV laser diode beam used for both the scattered light and LIF signals. Receiving optics are also needed to collect light from the illuminated particles
Fig. 1. Exterior of instrument.
25
Journal of Aerosol Science 117 (2018) 24–33
C. Lu et al.
Fig. 2. Schematic diagram of the optical system. The system consists of a radiation source, paraboloidal mirror, light trap, two focusing lenses, and a dichroic mirror.
and selectively pass the appropriate wave bands to the photomultiplier tube (PMT) for both scattered light and fluorescence detection. A laser diode is inserted into the center of a paraboloidal mirror. The intersection of the laser beam and the aerosol flow is the focus point of the paraboloidal mirror. The laser irradiates the aerosol directly, and the light beam that passes through the aerosol particles is directed into a light trap. The scattered light and LIF are combined by the paraboloidal mirror and passed to the two focusing lenses, as shown in Fig. 2. After passing through the second lens, the mixed light converges. Then the scattered and fluorescence light are separated by a dichroic mirror. The wave band of the light reflected by the dichroic mirror is below 420 nm and includes the scattered light; the wave band of light that passes through the dichroic mirror is 420–650 nm, which corresponds to the wave band of the fluorescence. Finally, the scattered light and the fluorescence are focused onto separate PMTs. 2.2. Electronic and control systems We also combine scattered light detection and fluorescence detection to distinguish bioparticles from other particles. This counter uses the elastic light scattering detection and LIF techniques. These two technologies have proved successful in past research, so we omit a detailed description here. For each detected particle, the signal processing system measures its equivalent optical diameter from the scattered light signal, and then the fluorescence signal is used to determine whether it is a bioparticle. Fig. 3 shows the oscilloscope traces of the analog signals from both the scattered light and the fluorescence. If the particle is not a bioparticle, it will emit only the scattered light signal;
Fig. 3. Oscilloscope screenshots. The upper and lower lines are the pulse signals for the scattered light and the fluorescence, respectively. (a) Pulse signal of a nonfluorescent particle. (b) Pulse signal of a fluorescent particle.
26
Journal of Aerosol Science 117 (2018) 24–33
C. Lu et al.
Fig. 4. Process diagram for the detection system. This is the entire system process, which includes the transformation of various signals.
if it is a bioparticle, it will also generate the fluorescence signal. The light signals are then transmitted to a processor. Fig. 4 shows a diagram of the electronic functions. First, the PMT transforms the light signal into a current signal, which is transformed into a voltage signal by a transimpedance amplifier that has a low bias current and offset, and simultaneously amplifies the voltage signal. The voltage signal is amplified further by a noninverting amplifier to strengthen it. After the voltage signal is enhanced, the analog digital converter (ADC) converts the voltage signal into a digital signal, which is transferred to the fieldprogrammable gate array (FPGA). The FPGA's sampling rate is 10 Mps, and a time function control outputs the digital signals into a microcontroller unit (MCU). Finally, the MCU receives information regarding the number of bioparticles and other particles of different sizes and then determines the need to sound an alarm on the basis of the calculated results. 3. Experiments and analysis 3.1. Experimental apparatus and sample preparation To evaluate the performance of the biological particle counter, experiments were performed using representative samples: NaCl particles, polystyrene latex (PSL) spheres, Model B800 (a type of fluorescent microspheres from Thermo Scientific, USA), and samples of Staphylococcus aureus, Escherichia coli, and Pseudomonas aeruginosa. As shown in Fig. 5, an aerosol generator capable of producing aerosols with different concentrations was adopted to generate the target aerosol. The target aerosol then formed a stable environment in the buffer bottle. The biological particle counter was used to detect the target aerosol in the buffer bottle. There was a sampler between the outlet of the optical cavity and the inlet of the pump. The sampler had a dissolvable filter for collecting bioparticles. The details are shown in Fig. 6. First, we conducted a series of tests with pure water because all the samples were dissolved in pure water when the aerosols were generated. We verified that the water introduced no interference factors. In addition, we compared the results of the culture method to those obtained by the biological particle counter. The results produced by the biological particle counter were averaged over five cycles. The results of the culture method were also averaged. The results are shown in Table 1. The table shows that there were no bioparticles or fluorescent particles in pure water; therefore, it is suitable to use pure water as the solvent in other experiments. 3.2. Identification of nonfluorescent and fluorescent particles NaCl particles and PSL particles are representative nonfluorescent particles, whereas B800 microspheres are representative fluorescent particles. These samples were used to test the reliability of the biological particle counter. The standard Model B800 fluorescent microspheres were chosen as a substitute for bioparticles. The excitation and fluorescence spectra of these microspheres are similar to those of NADH, which is an important component of bioparticles. The mean diameter of the B800 microspheres is 0.8 µm, and the size uniformity (standard deviation) is less than 3%. The concentration of a B800 aerosol can be controlled by
Fig. 5. Experimental setup for performance evaluation of the monitor. The target aerosol is aerosolized by the aerosol generator, forming a uniform aerosol in the buffer bottle.
27
Journal of Aerosol Science 117 (2018) 24–33
C. Lu et al.
Fig. 6. Diagram showing the location of the sampler. This is only a schematic diagram, and some components are not displayed.
Table 1 Results of the biological particle counter and culture method. Test sample
Pure water
Number
1 2 3 4 5
Biological particle counter
Culture method
(≥0.5 µm) Average particle number (PPL)
Average number of biological particles (PPL)
Average number of biological particles (CFU/ L)
12.2 18.1 21.2 11.6 16.4
0 0 0 0 0
0 0 0 0 0
adjusting the aerosol generator. To avoid misdetection, we measured the target aerosol five times using the same concentration and considered the average of the five results in every group of experiments. For NaCl and PSL particles, we measured the results using three different concentrations. For B800 particles, there were six different concentrations. The results for the nonfluorescent particles are shown in Table 2, which shows that no fluorescence was detected. Therefore, the counter did not produce false detections when the test samples were NaCl and PSL particles. To test the counter's ability to detect fluorescence, we used B800 and a mixture of NaCl and B800 as test samples. The results for B800 and the mixture are displayed in Tables 3 and 4, respectively. In Table 4, the ratio of B800 particles to NaCl solution is a relative value, and in the experiment, all conditions are the same except the ratio of B800 particles to NaCl solution. According to the results in Table 3, although most of the detected B800 particles were fluorescent, a few particles were nonfluorescent. On the basis of the results in Table 1, we used pure water as the solvent, which also generated a few particles in the aerosol generator; consequently, the fluorescent particle numbers were slightly lower than the particle numbers. In the B800 test, the performance of the counter was adequate. The results in Table 4 show that the proportion of fluorescent particles increased in proportion to the ratio of B800 particles to NaCl solution. On the basis of the results in Tables 2–4, we can consider that the counter is able to identify the nonfluorescent and fluorescent particles accurately. 3.3. Accurate measurement of biological particle concentration Almost all of the studies of LIF technology have focused on the identification and classification of bioaerosols. However, our Table 2 Results as measured by the counter. The samples were NaCl and PSL particles. Test sample
(≥0.5 µm) Average particle number (PPL)
Number of fluorescent particles (PPL)
NaCl particles
12,250 502,45.4 114,236 1423.2 5056 12,138.4
0 0 0 0 0 0
0.5 µm PSL particles
28
Journal of Aerosol Science 117 (2018) 24–33
C. Lu et al.
Table 3 Results of fluorescence tests of B800 particles. Test sample
(≥0.5 µm) Average particle number (PPL)
Number of fluorescent particles (PPL)
Proportion of fluorescent particles
B800
193.6 386.4 689 1563 2349 3330
174.2 355.4 613 1400 2137 2930
90% 91.8% 89% 89.5% 91% 88%
Table 4 Results of fluorescence test of a mixture of NaCl and B800. Ratio of B800 to NaCl solution (one drop: liter)
(≥0.5 µm) Particle number (PPL)
Number of fluorescent particles (PPL)
Proportion of fluorescent particles
1:1
9450 9474 9520 9379 9508 9810 9708 9856 9997 9713 10,004 10,312 10,987 10,045 10,333
1087 1110 1092 1079 1121 2100 2212 2108 2245 2178 3302 3405 3298 3288 3345
11.5% 11.7% 11.4% 11.5% 11.8% 21.4% 22.8% 21.4% 22.4% 22.4% 33% 33% 30% 32.7% 32.4%
2:1
3:1
attention is focused on accurate measurement of bacterial concentrations. S. aureus, E. coli, and P. aeruginosa are used as representative biological particles to evaluate the measurement accuracy of the counter in the laboratory. S. aureus is a Gram-positive bacterium; E. coli and P. aeruginosa are Gram-negative bacteria. The results of the culture method for test samples of bacteria were compared with those of the biological particle counter to test the counter's reliability. The experimental process is outlined as follows: (a) First, the original saved frozen bacteria were activated in a thermostat. Two days later, the activated bacteria were placed in a centrifuge to separate them from the nutrients. (b) Then the bacteria, which were free from impurities, were dissolved in pure water to generate the target aerosol. (c) The aerosol generator was adjusted to control the concentration in the buffer bottle. When the concentration in the buffer bottle was stable, we ran the biological particle counter for five cycles. (d) We recorded the bioparticle detection results for each cycle. Then, five cycles later, we removed the dissolvable filter from the sampler and placed it in a prepared saline solution for dilution. The total volume was 10 mL (this value could be adjusted). (e) We distributed the bacteria homogeneously in the mixed liquor using a vortex oscillator to ensure that there are no clumps of two or more bacteria. (f) Then the sample was distributed into seven (or more) culture dishes. Two dishes contained 0.5 mL of mixed liquor; two contained 1.0 mL of mixed liquor; two contained 2.0 mL of mixed liquor, and the remaining dish contained 3.0 mL of mixed liquor. (g) Finally, we added a mixture of nutrient solution and agar to each dish. When the mixture began clotting, we left the culture dishes in a suitable environment. (h) The results of the culture method were obtained two days later. We counted the clumps on each dish. Then we verified whether the proportions among the seven dishes were correct. For example, the number of clumps in the dish containing 1.0 mL of mixed liquor should be twice that in the dish containing 0.5 mL of mixed liquor. We calculated the average over five cycles and compared it with the average results of the culture method. In the experiments, one cycle is 1 min, and the flow rate of the counter is 2 L/min, so the total air flow in five cycles is 10 L. The results of the two methods at various concentrations are shown as follows. Tables 5(a) and 5(b) show the results for S. Aureus, the S. Aureus suspension in Table 5(a) is diluted 100 times with original S. Aureus, the S. Aureus suspension in Table 5(b) is diluted 50 times with original S. Aureus. Table 5(c) shows the results for E. coli, and Table 5(d) shows the results for P. aeruginosa. Each measured result was the average of five cycles. According to Tables 5(a) and 5(b), 5(c), and 5(d), we can conclude the following:
29
Journal of Aerosol Science 117 (2018) 24–33
C. Lu et al.
Table 5(a) Results of biological particle counter and culture method for S. aureus (diluted 100 times with original S. aureus). Number
Result of biological particle counter (PPL) (Ncounter)
Result of the culture method (CFU/10 L) (Nculture)
η = Ncounter/Nculture
1 2 3 4 5 6 7 8 9 10 11 12
4 7.8 11 26 30 50 54 64.6 69.4 85 152.4 243.8
49 72 118 251 425 693 662 828 951 1183 2097 3803
81.6% 108.3% 93.2% 104% 71.4% 72.5% 81.8% 77.8% 72.9% 71.8% 73.04% 64.1%
Table 5(b) Results of biological particle counter and culture method for S. aureus (diluted 50 times with original S. aureus). Number
Result of biological particle counter (PPL) (Ncounter)
Result of culture method (CFU/10 L) (Nculture)
η = Ncounter/Nculture
1 2 3 4 5 6 7 8 9 10 11 12
5 8.8 14 22 33.6 39 46.2 53 65.4 92 143.2 210.8
61 101 169 223 462 549 633 680 871 1244 2106 3221
81.9% 88% 82.8% 98.6% 72.7% 71% 73% 78% 75.1% 73.9% 68% 65.4%
Table 5(c) Results of biological particle counter and culture method for E. coli. Number
Result of biological particle counter (PPL) (Ncounter)
Result of culture method (CFU/10 L) (Nculture)
η = Ncounter/Nculture
1 2 3 4 5 6 7 8 9 10 11 12
6 8.6 10 12.2 21.4 33.3 41.2 44.7 64.6 69.6 71.4 98
88 121 139 162 398 717 943 925 1154 1178 1472 1941
68.1% 71.1% 71.9% 75.3% 53.7% 46.4% 43.6% 48.3% 55.9% 55.9% 48.5% 50.5%
(1) The percentage η (= Ncounter/Nculture) is generally greater at low concentration than at high concentration: Tables 5(a) and 5(b) show that at concentrations of < 25 CFU/L, the η value of S. aureus is between 80% and 110%, and that at concentrations of > 25 CFU/L, the results are worse than those of the culture method, and η is between 64% and 80%. In addition, Tables 5(c) and 5(d) yield similar conclusions. (2) Tables 5(a) and 5(b) reveal that for the same type of test sample, although the distribution is different, the η values at same concentrations are similar. (3) A comparison of Table 5(a) with Tables 5(c) and 5(d) shows that for different types of test samples, η is different at different concentrations. But for each group,there is a good linear relationship. Conclusion (1) can be explained by the bacteria generated in the buffer bottle, which are composed of two or more particles but are regarded by the counter as a single particle. However, in the culture method, they are isolated, and at low concentrations, the probability that several S. aureus are clumped together is reduced. In addition, the intensity of the fluorescence excited by different S. aureus varies. Some S. aureus are inactive, and their fluorescence is too weak to be detected by the counter, but they are still 30
Journal of Aerosol Science 117 (2018) 24–33
C. Lu et al.
Table 5(d) Results of biological particle counter and culture method for P. aeruginosa. Number
Result of biological particle counter (PPL) (Ncounter)
Result of culture method (CFU/10 L) (Nculture)
η = Ncounter/Nculture
1 2 3 4 5 6 7 8 9 10 11 12
4.2 9 19.6 28 35.2 40.2 54.4 63.6 79.4 89.2 93.8 99
61 126 284 451 556 681 928 1022 1274 1335 1503 1601
68.8% 71.4% 69% 62.1% 63.3% 59.1% 58.7% 62.2% 62.3% 66.8% 62.4% 61.8%
cultivable. The relevance of the results of the two methods is also considered using Fig. 7. The fitted straight-line segment in Fig. 7(a) is y (CFU/L) = 1.54382*x (PPL) − 9.8974 , and the linearly dependent coefficient is 0.99. The fitted straight-line segment in Fig. 7(b) is y (CFU/L) = 1.53148*x (PPL) − 7.54765, and the linearly dependent coefficient is 0.99. The fitted straight-line segment in Fig. 7(c) is y (CFU/L) = 1.97899*x (PPL) − 2.34107 , and the linearly dependent coefficient is
Fig. 7. Relationship between the concentrations detected by the two methods for (a) S. aureus (diluted 100 times with original S. aureus), (b) S. aureus (diluted 50 times with original S. aureus), (c) E. coli, and (d) P. aeruginosa.
31
Journal of Aerosol Science 117 (2018) 24–33
C. Lu et al.
Table 6 Repeatability of the results measured by the two methods. Group
1
2
3
4
5
6
7
8
Average
Standard deviation
Relative standard deviation
Measured by counter (PPL) Measured by culture method (CFU/L)
56 48
47 47
57 50
44 66
54 55
50 59
53 68
54 66
51.9 57.375
4.226 8.62
0.0814 0.15
0.97. The fitted straight-line segment in Fig. 7(d) is y (CFU/L) = 1.59382*x (PPL) + 0.03379, and the linearly dependent coefficient is 0.99. These results show that the linear relationship between the concentrations detected by the two methods is adequate. Different bacteria may have different linear relationships; further research is needed here. Furthermore, we studied the repeatability of the results of both methods using S. aureus. We ensured that the concentration of the bacteria was stable and then performed the same experiment eight times. The results are shown in Table 6 and Fig. 8. The result shows that the averages of the two methods are similar. The standard deviation and relative standard deviation (RSD) of the two methods are reasonable. According to USP 1223, an RSD measured in the 15–35% range for the culture method would generally be acceptable. Many conventional microbiological methods are subject to sampling, dilution, plating, incubation, and operator errors (USP-, 1223). The RSD of 8.14% for the counter measurements is stable.
4. Conclusions This paper focused on accurate measurement of bacterial concentrations in air by a biological particle counter. The counter includes optical, electronic, and control systems. A series of experiments were presented to evaluate the performance of a single biological particle counter. Pure water, NaCl particles, PSL particles, and B800 microspheres were used to evaluate the performance of the counter. The results show that this counter can reliably distinguish between fluorescent and nonfluorescent particles. For target aerosols of S. aureus, E. coli, and P. aeruginosa, we compared the results measured by the counter with those of the culture method. We found that the percentage η (= Ncounter/Nculture) was higher at low concentration than at high concentration. When we compared the results for S. aureus at different concentrations, we found that for the same type of test sample, although there were different distributions, the η values at same concentrations were similar. A comparison of the result for S. aureus with those of E. coli and P. aeruginosa revealed that for different types of test samples, η was different at same concentrations. In addition, for each bacteria, good linear agreement was observed between the two results, and the coefficients (R2) were all greater than 0.97. The repeatability of the results measured by the counter and the culture method was also tested. The results measured by the counter were more stable, with an RSD of 8.14%, which is smaller than that of the culture method (15%). Studies of other bacteria were outside the scope of this work, so our research is limited. However, the research method, which compares the new method based on the LIF technique with the culture method, is feasible. In addition, this counter counted the number of fluorescent particles accurately, but it could not obtain the fluorescence spectrum. A good method to lower the false alarm rate is to analyze the fluorescence spectrum and develop a new algorithm. Another method is to use supplementary means such as measurement of the temperature characteristics of the bioparticles. In the future, more research is needed to ensure a low false alarm rate for detection and to achieve more precise identification of bioaerosols.
Fig. 8. Repeatability of the results measured by the two methods. The red line is the average. (a) Repeatability of the counter results. (b) Repeatability of the results of the culture method.
32
Journal of Aerosol Science 117 (2018) 24–33
C. Lu et al.
References Watches, C. M., & Cox, C. B. (1995). Bioaerosols handbook. Lewis Publishers. Lakowicz, J. R. (2006). Principles of fluorescence spectroscopy. Springer Science and Business Media. Pan Yong-Le, Eversole Jay. D., Kaye Paul. H., Foot Virginia, Pinnick Ronald. G., Hill Steven. C.,Mayo Michael. W., Bottiger Jerold. R., AlanHuston, VasanthiSivaprakasam, Chang Richard. K.,Bio-aerosol fluorescence detecting and characterising bio-aerosols via UV light-induced fluorescence spectroscopy, Optics of Biological Particles, 63–164. Hairston, P. P., Ho, J., & Quant, F. R. (1997). Design of an counter for real-time detection of bioaerosols using simultaneous measurement of particle aerodynamic size and intrinsic fluorescence. Elsevier Science Ltd. Pinnick, R. G., Hill, S. C., Nachman, P., Videen, G., Chen, G., & Chang, R. K. (1998). Aerosol fluorescence spectrum analyzer for rapid measurement of single micrometer-sized airborne biological particles. Aerosol Science and Technology, 28(2), 95–104. Brosseau, L. M., Vesley, D., Rice, N., Goodell, K., Nellis, M., & Hairston, P. (2000). Differences in detected fluorescence among several bacterial species measured with a direct-reading particle sizer and fluorescence detector. Aerosol Science and Technology, 32, 545–558. Product Specifications of Azbil BioVigilant.Inc (2013a). IMD-A® Series Rapid Biologic Detection System for Air. Product Specifications of Azbil BioVigilant.Inc (2013b). Comparison of IMD-A® 300 and 350 Models. Jim, E., Dahu, Q., Hairston Peter, P., & Darrick, N. (2013). Apparatus and system for characterizing an aerosol particle flow. PCT (043600 A1). Primmerman, C. (2000). A detection of biological agents. Lincoln Laboratory Journal, 12(1), 3–32. Eversole, J., Hardgrove, J., Cary, W., Choulas, D., & Seaver, M. (1999). Continuous, rapid biological aerosol detection with the use of UV fluorescence: Outdoor test results. Field Analytical Chemistry & Technology, 3(4‐5), 249–259. Pan, Y.-L., Hill, S. C., Pinnick, R. G., House, J. M., Flagan, R. C., & Dual, C. R. K. (2011). excitation-wavelength fluorescence spectra and elastic scattering for differentiation of single airborne pollen and fungal particles. Atmospheric Environment, 45(8), 1555–1563. Kayea Paul. H., Hirsta Edwin, Foot Virginia, Clark James. M., and Baxter Karen, A low-cost multi-channel aerosol fluorescence sensor for networked deployment, SPIE Vol. 5617. Kaye, P. H., Stanley, W. R., Hirst, E., Foot, E. V., Baxter, K. L., & Barrington, S. J. (2005). A single particle multichannel bio-aerosol fluorescence sensor. Optical Society of America. Toprak, E., & Schnaiter, M. (2012). Fluorescent biological aerosol particles (FBAPs) measured with the Waveband Integrated bioaerosol sensor WIBS-4: laboratory tests combined with a one year field study. Atmospheric Chemistry and Physics Discussion, 12, 17607–17656. Product Specifications of Droplet Measurement Technologies, WIDEBAND INTEGRATED BIOAEROSOL SENSOR−4A(WIBS-4A). Bundke, U., Reimann, B., Nillius, B., Jaenicke, R., & Bingemer, H. (2010). Development of a bioaerosol single particle detector (BIO IN) for the Fast Ice Nucleus CHamber FINCH. Atmospheric Measurement Techniques, 3(1), 263–271. Kiselev, D., Bonacina, L., & Wolf, J. P. (2011). Individual bioaerosol particle discrimination by multi-photon excited fluorescence. Optics Express, 19(24), 24516–24521. Farsund, O., Rustad, G., Kasen, I., & Haavardsholm, T. V. (2010). Required spectral resolution for bioaerosol detection algorithms using standoff laser-induced fluorescence measurements. IEEE Sensors Journal, 10(3), 655–661. USP-1223, Validation of alternative microbiological methods.
33